TRADE IN GOODS

This page intentionally left blank

 Trade in Goods
The GATT and the Other Agreements Regulating
Trade in Goods

PETROS C. MAVROIDIS

Edwin B. Parker Professor of Law, Columbia Law School, NY,

Professor at the University of Neuchâtel, Switzerland
Research Fellow at CEPR (Centre for Economic Policy Research)

1
 3
Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© P. Mavroidis, 2007
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
Crown copyright material is reproduced under Class Licence
Number C01P0000148 with the permission of OPSI
and the Queen’s Printer for Scotland
First published 2007
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose the same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Mavroidis, Petros C.
Trade in goods : the GATT and the other agreements regulating trade in goods /
Petros C. Mavroidis.
p. cm.
Includes bibliographical references and index.
ISBN 978–0–19–923903–0 (alk. paper)
1. Foreign trade regulation. 2. General Agreement on Tariff s and Trade (1947)
3. General Agreement on Tariff s and Trade (Organization) I. Title.
K4600.M393 2007
343⬘.087—dc22 2007034285
Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India
Printed in Great Britain
on acid-free paper by
Biddles Ltd., King’s Lynn, Norfolk

ISBN 978–0–19–923903–0

1 3 5 7 9 10 8 6 4 2
For Meritas
This page intentionally left blank
 Acknowledgements

Winston Churchill once said: ‘writing a book is an adventure. To begin with, it is

a toy and an amusement; then it becomes a mistress, and then it becomes a mas-
ter, and then a tyrant. The last phase is that just as you are about to be reconciled
to your servitude, you kill the monster, and fling him out to the public.’ A num-
ber of people should take credit for helping me to complete this book, although
they should not be held accountable for flinging it out to the public; this is my
sole responsibility.
This book is the natural extension (and substantial revision) of my Commentary
on the GATT (Oxford University Press, 2005). Rob Howse asked me to write on
that subject, and I thank him for showing confidence in me.
Bryce Bittner, Jorge Huerta-Goldman, Kevin Stemp, Anastasios Tomazos,
Arun Venkataraman, Edwin Vermulst, and Jasper-Martijn Wauters read parts
of the original manuscript, and corrected my misunderstandings on a host of
issues. Heinz Opelz shared with me his invaluable experience as Director of mar-
ket access issues of the GATT. At the WTO, Cato Adrian, Marc Bacchetta, Tessa
Bridgeman, Willie Chatsika, Edwini Kessie, Patrick Allison Low, Juan-Alberto
Marchetti, Julie Pain, Serge Stamnas, Gretchen Stanton, Luigi Stendardo, Joelle
Vuillemenot, and Erik Wijkstrom, never tired of responding to my (ever increas-
ing number of ) questions. Rhian-Mary Wood-Richards helped me identify doc-
uments important to my research, and bring the various chapters to their present
status. The same goes for Sarah Sladen who has helped me enormously to prepare
this book, and had to brush up her German on the way. My computer skills, or
lack of them, have betrayed me a number of times: Angay Vijayakumar, Ashley
Pineda, and Boris Niyazov, thank you so much for being such a help on this
score. My students on both sides of the Atlantic, through their comments, ques-
tions and overall class-participation, greatly contributed to the manner in which
I deal with the issues here. Teaching at CLS about the WTO for the past years
with my dear friends and colleagues Kyle Bagwell, George Bermann, and Jagdish
Bhagwati, has been a very enriching experience for me. At Neuchâtel, I benefited
from discussions with Yvan Fauchère, Pauline Lièvre, Panagiotis Delimatsis, and
Laurenz Sigismondi. Yvan, in particular, not only generously shared with me his
expertise on SPS issues, but helped me bring this book to its current shape.
I benefited from a series of talks with Marc Bacchetta, Chad Bown, Marc
Busch, Meredith Crowley, Bill Davey, Jeff Dunoff, Claus-Dieter Ehlermann, Bill
Ethier, Crawford Falconer, Joe Francois, Gary Horlick, Aaditya Mattoo, Niall
Meagher, Mattia Melloni, Damien J. Neven, Federico Ortino, Joost Hugo Bart
Pauwelyn, Ernst-Ulrich Petersmann, Tom Prusa, Eric Reinhardt, Kamal Saggi,
viii Acknowledgements

Simon Schropp, T.N. Srinivasan, Mike Trebilcock, and Joseph H.H. Weiler who
generously shared their unique knowledge on GATT issues with me.
Bernard Hoekman, Doug Irwin, André Sapir, and Joel Trachtman read re-drafts
of my chapters and gave me very useful comments. They also largely influenced
the approach taken in this book: Doug persuaded me that we have a lot to learn
by placing the GATT in its appropriate historical context, and I have certainly
immensely enjoyed my time reading the preparatory work of the GATT. Joel’s
comments were, as always, to the point. He has always been particularly generous
with his time and I have benefited a lot from my interaction with him over the
years. As is the case with the other people I thank in this note, there is unfortu-
nately no reciprocity involved, so far at least. Bernard and André are uniquely
positioned to discuss the law and economics of the world trading system. They
have shared with me their vast knowledge of the institutions and the political
economy considerations of the key actors. They have also corrected my many
misconceptions on issues of their expertise.
Ivan Crowley read the whole manuscript, and deserves a very particular men-
tion. His comments have been outstanding as they have been detailed: he spot-
ted numerous errors and inconsistencies in previous drafts, and brought them
to my attention; through his comments, he opened up new avenues for me. He
responded beyond the call of duty.
David Palmeter has been a great friend and mentor over the years. My under-
standing of the GATT has been deeply influenced by his work and our cooper-
ation. His continuous interest in trade issues and his ever-innovating approach is
a source of inspiration for me.
Lance Liebman put a wonderful group together (Kyle Bagwell, Gene
Grossman, Henrik Horn, Bob Staiger, Alan Sykes) to work on the American Law
Institute (ALI) project on WTO law, and asked me to participate in it. I have
learned so much participating in this group, it is simply impossible for me to
thank my colleagues and friends there enough.
I have already expressed elsewhere that my interest in the study of the GATT
is largely due to Frieder Roessler. He hired me for the GATT legal service a few
years ago, and through our discussions inspired my interest in the field. I hope
that he will get some pleasure reading this book. Bernard Hoekman and Henrik
Horn picked up from where Frieder had left off. Bernard’s knowledge of the
world trading system is unique in so many respects. What is also unique is his
approach. He is a true social scientist. Henrik has been my steady co-author for
over 10 years. I owe him much more than I will ever be able to express, and hope
that, when reading this book, he will have the sense that his (ongoing) tutoring
was (is) not in vain.
While writing this book, I had the opportunity to revisit some of Bob Hudec’s
work and realize, yet again, that he had already assessed (in a much superior man-
ner) most of the issues I am dealing with. It is a pity he is not with us any more;
I would have loved to share this book with him.
 Acknowledgements ix

My wife, Suja Rishikesh-Mavroidis, has still not given up trying to educate

me on market access issues, her area of expertise. She also undertook most of our
family tasks, providing me with precious time to work on this book. I will never
be in a position to thank her enough for all she has done for our family and me
over the years. This book, as is the case of all my work, is for her and our daugh-
ters, Meera-Natalia, Riya-Valentina, and Tara-Eleni.
Petros C. Mavroidis
Commugny, Switzerland
April 27, 2007
This page intentionally left blank
 Summary Contents

Acknowledgements vii
List of Abbreviations xix
Table of Cases xxiii
Table of Legislation xxvii

1. From GATT 1947 to GATT 1994 1

2. Disciplines on Trade Instruments 39
3. Domestic Instruments 190
4. State Contingencies 320
5. Institutional Provisions 388
6. The GATT, Then and Now 447

Appendices 455
References 478
Index 499
This page intentionally left blank
 Contents

1. From GATT 1947 to GATT 1994 1

1 The Genesis of the GATT 2
2 Why the GATT? 6
2.1 Why ask the question? 6
2.2 The negotiators say . . . 7
2.3 Economic theory suggests 12
2.4 And some unexplored avenues 18
2.5 To conclude (?) 20
3 The GATT in the GATT era: An Agreement, and an Institution
as Well 20
4 The GATT Rounds of Trade Liberalization 23
5 Expanding the GATT: The Emergence of the Codes 24
6 The GATT in the WTO era: Just an Agreement, Albeit a Pivotal One 25
7 The Modern GATT discipline in a Nutshell 30
7.1 The GATT in its WTO context 30
7.2 The legal relation of the GATT with the WTO Agreement 31
7.3 The substantive content of modern GATT 34

2. Disciplines on Trade Instruments 39

1 Introduction 42
2 The Treatment of QRs 42
2.1 The legal text 42
2.2 What is a QR? 43
2.2.1 The attribution of practices to governments 45
2.2.2 The measures covered 50
2.2.3 The standard of review: no effects test and no intent test 55
2.2.4 The relationship between Art XI and other GATT provisions 58
2.2.4.1 Art III of the GATT 58
2.2.4.2 Art VI of the GATT 62
2.2.4.3 Art XX of the GATT 63
2.2.4.4 Art XXI of the GATT 63
2.2.4.5 Infant industry protection (Art XVIIIc) 63
2.2.4.6 Arts XII and XVIII of the GATT 63
2.2.4.7 Art XIII of the GATT 63
2.3 Trade in textiles 67
2.4 Discriminatory legal QRs 67
2.5 The non-segmentation nature of Art XI of the GATT 68
xiv Contents

3 Tariff Protection in the GATT 68

3.1 The legal text 68
3.2 The basic discipline with respect to tariff protection 70
3.3 Why yes to duties and no to QRs? 71
3.4 The mechanics of binding the duties 72
3.4.1 The Harmonized System (HS) 72
3.4.2 What is the tariff promise? Ordinary customs duties,
other duties and charges 74
3.4.3 The typology of ordinary customs duties 78
3.4.4 Multilateral trade negotiations (rounds) 79
3.4.5 The treatment of export tariffs (taxes) under GATT 84
3.4.6 Certification of schedules 86
3.5 Interpreting tariff commitments 87
3.5.1 The legal relevance of HS 87
3.5.2 Scheduling of concessions must be WTO consistent 88
3.5.3 The (non) impact of legitimate expectations 93
3.6 Changes in tariff protection after their consolidation 93
3.6.1 Switching between different types of duties 94
3.6.2 Withdrawing concessions from members leaving the WTO 95
3.6.3 Renegotiating the tariff protection 97
3.6.3.1 INR holders 99
3.6.3.2 PSI countries 100
3.6.3.3 SI countries 102
3.6.3.4 MFN trade is the basis for defining PSIs and SIs 102
3.6.3.5 New products 103
3.6.3.6 The mechanics of the negotiation 103
3.6.4 Rectifications and modifications of schedules 112
3.7 Fees and charges for services rendered 114
3.8 Customs Valuation (CV) 118
3.9 The Agreement on Pre-shipment Inspection (PSIA) 119
3.10 The MFN 120
3.10.1 The MFN discipline in a nutshell 123
3.10.2 The measures coming under the purview of the MFN 124
3.10.2.1 Any advantage . . . 124
3.10.2.2 . . .granted to products originating in
any country . . . 125
3.10.2.3 . . .irrespective whether it discriminates de jure
or de facto . . . 126
3.10.3 . . .must be extended to like products 126
3.10.4 The rules of origin 129
3.10.5 The MFN treatment must be extended immediately
and unconditionally 132
3.10.6 No rebalancing permitted 135
3.10.7 The standard of review: no effect and no intent test 136
3.11 Special and differential treatment for developing countries 137
3.11.1 The Enabling clause 137
 Contents xv

3.11.2 Some historical features 138

3.11.3 The Enabling clause in the WTO legal order 143
3.11.4 The test for compliance with the Enabling clause 144
3.11.5 Is the candle worth the flame? (criticism of the
Enabling clause) 145
3.12 Preferential Trade Agreements (PTAs) 148
3.12.1 PTAs in the WTO: FTAs and CUs 148
3.12.2 Globalization yes, preferences yes as well 148
3.12.3 What matters when discussing preferential trade? 150
3.12.4 PTAs in the GATT 152
3.12.4.1 No per se inconsistency 152
3.12.4.2 The Art XXIV of the GATT test in a nutshell 153
3.12.4.3 Designed at home, approved (?) in Geneva 153
3.12.4.4 The multilateral track 156
3.12.4.5 The bilateral track 168
3.12.4.6 Why not litigate more? 171
3.12.5 What is the permissible extent from MFN
deviation when a PTA is formed? 177
4 Export Subsidies 179
4.1 Rigidity required? 179
4.2 From discouraging to outlawing export subsidies 181
4.3 Counteracting prohibited subsidies 184

3. Domestic Instruments 190

1 National Treatment and its Rationale 193
2 The Coverage 195
2.1 Wide coverage 195
2.2 Subsidies 195
2.2.1 What is a (domestic) subsidy? 195
2.2.1.1 Financial contribution 196
2.2.1.2 Benefit to recipient 198
2.2.1.3 The specificity requirement 200
2.2.2 The classification of subsidies 200
2.2.3 Counteracting subsidies 201
2.2.4 Remedies 201
2.2.5 Farm subsidies 202
2.2.5.1 The regulation of farm trade in the GATT era 202
2.2.5.2 The WTO Agreement in a nutshell 206
2.2.5.3 AG over SCM 208
2.3 Government procurement 210
2.3.1 What is government procurement? 210
2.3.2 The basic discipline: non-discrimination 211
xvi Contents
2.3.3 The past and present of GPA membership 211
2.4 The coverage of Art III of the GATT—revisited 212
3 National Treatment, and Negative Integration 215
4 National Treatment: The Legal Discipline 216
4.1 Fiscal measures 216
4.1.1 DCS products 217
4.1.1.1 Defining DCS 217
4.1.1.2 SATAP 221
4.1.2 Like products 234
4.1.2.1 Defining like products 234
4.1.2.2 Taxation in excess 236
4.2 Non-fiscal measures 237
4.2.1 A measure . . . 237
4.2.2 . . . affecting sale, offering for sale, . . . 238
4.2.3 . . . of like products 238
4.2.4 Less favourable treatment 242
4.3 A critique of the case-law concerning NT 247
4.4 General exceptions 254
4.4.1 Coverage 254
4.4.2 An exception to which provisions? 254
4.4.3 Regulatory diversity 259
4.4.4 A two-tier test for compliance with Art XX of the GATT 260
4.4.5 Public morals (Art XX(a) of the GATT) 261
4.4.5.1 Coverage 261
4.4.5.2 The necessity requirement 262
4.4.6 Human, animal and plant life (Art XX(b) of the GATT) 262
4.4.6.1 Coverage 262
4.4.6.2 The necessity requirement 262
4.4.6.3 No science required 263
4.4.7 Domestic laws and regulations (Art XX(d) of the GATT) 264
4.4.7.1 Coverage 264
4.4.7.2 The necessity requirement 264
4.4.8 Exhaustible natural resources (Art XX(g) of the GATT) 266
4.4.8.1 Coverage 266
4.4.8.2 Jurisdictional limits 268
4.4.8.3 Measure must be relating to the objective pursued 269
4.4.8.4 Measures restricting domestic production or
consumption 270
4.4.9 The chapeau of Art XX of the GATT 271
4.4.9.1 The scope of the chapeau 271
4.4.9.2 Complying with the chapeau 271
4.4.9.2.1 Arbitrary or unjustifiable discrimination 271
4.4.9.2.2 Disguised restriction of trade 274
4.4.10 A critique of the case-law 276
4.4.10.1 The discontents of an exhaustive list 276
4.4.10.2 Extra-territoriality 277
 Contents xvii

4.4.10.3 A two-tier test? 285

4.4.10.4 The standard of review 285
4.5 A specific exception: films 286
5 Dealing with NTBs 287
5.1 The TBT and SPS Agreements 287
5.2 The legal relationship between GATT, TBT, and SPS 288
5.3 TBT 288
5.3.1 Coverage 288
5.3.2 International standards 289
5.3.3 Unilateral technical regulations 295
5.3.4 Domestic standards 296
5.4 The WTO SPS Agreement 297
5.4.1 Coverage 297
5.4.2 The SPS Agreement in a nutshell 298
5.4.3 International standards 299
5.4.4 Unilateral measures based on scientific evidence 299
5.4.5 Measures adopted on the basis of precaution 309
6 State Trading Enterprises (STEs) 315
6.1 STEs: definitional issues 315
6.2 The legal obligation in a nutshell 316

4. State Contingencies 320

1 Introduction 321
2 National Security (Art XXI of the GATT) 322
2.1 The legal discipline 322
2.2 Practice in the GATT years 322
2.3 Practice in the WTO era 330
3 BoP 331
4 Exchange Restrictions 335
5 Antidumping (Art VI of the GATT) 338
5.1 An unfair (?) price discrimination 338
5.2 Finding dumping 343
5.3 Establishing injury 349
5.4 Sunset reviews 352
5.5 Price differentiation in antitrust statutes 358
5.6 What does the ‘anti’ in AD refer to? 361
5.6.1 Discriminatory pricing 361
5.6.2 Predation 362
5.6.3 Strategic dumping 363
5.6.4 AD as a safeguard mechanism 364
6 Countervailing Duties (CVD) 365
7 Safeguards 365
7.1 The regulatory framework 365
7.2 The typology of safeguards 366
7.3 Safeguards imposed by PTAs 367
xviii Contents
7.4 Conditions for lawful application of safeguards 368
7.4.1 Initiation of investigation 368
7.4.2 Unforeseen developments 370
7.4.3 Increased imports 371
7.4.4 Injury 371
7.4.5 Are safeguards truly non-discriminatory? 375
7.4.6 The issue of compensation 376
7.4.7 The duration of safeguards 378
7.5 Safeguards and voluntary export restraints (VERs) 380
7.6 Safeguards: a safe harbour too far? 381
8 The Notion of Market Access Revisited 385

5. Institutional Provisions 388

1 Introductory Remarks 389
2 Participation 390
2.1 Contracting Parties to the GATT, WTO members 390
2.2 Accession 391
2.3 Non-application 394
3 Decision Making 395
3.1 Consensus, the default rule 395
3.2 Special procedures 396
3.2.1 Interpretations 396
3.2.2 Amendment of the GATT 397
4 Dispute Settlement 398
4.1 The early GATT years 398
4.2 The late GATT years 400
4.3 The WTO DSU 402
4.4 The WTO: exclusive forum for adjudication of trade disputes 404
4.5 A closer look into procedures 406
4.5.1 Request for establishment of a panel 406
4.5.2 Burden of proof 407
4.5.3 Discovery powers 408
4.5.4 Standard of review 410
4.5.5 The process before the AB 411
4.5.6 The types of complaints before panels and the AB 411
4.5.6.1 Violation complaints 412
4.5.6.2 NVCs 412
4.5.6.3 Situation complaints 415
4.5.7 Recommendations and suggestions by panels and the AB 416
4.6 Enforcing WTO obligations 417
4.6.1 The process in a nutshell 417
4.6.2 The types of complaint matter for enforcement purposes 418
4.6.3 Compliance within a reasonable period of time 418
4.6.4 Compliance panels 419
4.6.5 Compensation 420
4.6.6 Suspension of concessions 421
 Contents xix

4.6.6.1 Procedural issues 421

4.6.6.2 Quantifying suspension of concessions 422
4.7 Monitoring compliance 424
4.8 The value of precedent in the WTO 425
4.9 Does WTO enforcement work? 426
5 Waivers 435
6 Transparency 436
6.1 Introductory remarks 436
6.2 Trade Policy Review Mechanism (TPRM) 437
6.3 Art X of the GATT 440
6.3.1 The discipline in a nutshell 440
6.3.2 The coverage 441
6.3.3 The standard of review 442
6.3.4 Laws of general application 443
6.3.5 Uniform, reasonable, and impartial administration 443
6.3.5.1 Uniform administration 444
6.3.5.2 Reasonable administration 444
6.3.5.3 Impartial administration 445
6.3.6 The obligation to maintain independent tribunals 445
7 Relation to the Havana Charter 446

6. The GATT, Then and Now 447

1 The GATT 1947: A Solid Text (with some compromises) 447
2 The Evolution of the GATT: No Keeping up with the Joneses 448
3 Rule of Reason 451

Appendices 455
References 478
Index 499
 List of Abbreviations
AB Appellate Body
ACP African, Caribbean and Pacific group
AD Antidumping
ADICMA Association of Industrial Producers of Leather, Leather Manufactures
and Related Products (Argentina)
AG Agreement on Agriculture
AGOA Africa Growth and Opportunity Act (US)
AMS Aggregate measurement of support
ANZCERTA Australia-New Zealand Closer Economic Relations Trade Agreement
Art Article
ATC Agreement on Textiles and Clothing
BFA Bananas Framework Agreement
BIA Best information available
BISD Basic Instruments and Selected Documents (GATT publication)
BoP Balance of payments
BTN Brussels Convention on Nomenclature for the Classification of
Goods in Customs Tariffs
CAP Common Agricultural Policy (EC)
CCCN Customs Cooperation Council Nomenclature
CDSOA Continued Dumping and Subsidies Offset Act
CEA Council of Economic Advisors
CFI European Court of First Instance
CITES Convention on the International Trade of Endangered Species
CLC Contingent liberalization commitments
CRTA Committee on regional trade arrangements
CTD Committee on trade and development
CTG Council for Trade in Goods
CTS Consolidated Tariff Schedules
CU Customs union
CUSFTA Canada/US free trade area
CV Customs Valuation
CVD Countervailing duty
DDG Deputy Director-General
DG Director-General
Doc Document
DOC Department of Commerce (US)
DCS Directly competitive or substitutable products
DSB Dispute Settlement Body
DSU Dispute Settlement Understanding
EBA Everything but arms
EC European Community
 List of Abbreviations xxi

ECR European Court Reports

ECT European Community Treaty
ECJ European Court of Justice
EFTA European Free Trade Association
ESTO European Scientific Technology Observatory
EU European Union
FAO Food and Agriculture Organization (UN)
FOREX Foreign exchange contracts
FSC Foreign Sales Corporation
FTA Free trade area
FTAA Free trade area of the Americas
GAAP Generally accepted accountancy principles
GATS General Agreement on Trade in Services
GATT General Agreement on Tariffs and Trade
GC General Council
GDP Gross domestic product
GIR General Interpretative Rule (HS)
GMO Genetically modified organism
GN Geneva Nomenclature
GNP Gross national product
GPA Government Procurement Agreement
GSP Generalized system of preferences
G 90 Group of 90 less developed WTO Members
G 20 Group of 20 leading developing WTO Members
HHI Herfindhal-Hirschman Index
HS Harmonized System
ICITO Interim Commission for the International Trade Organization
ICJ International Court of Justice
IEC International Electrotechnical Commission
ILA Agreement on Import Licensing Procedures
ILC International Law Commission
IMF International Monetary Fund
INR Initial negotiating right
IT Information technology
ITA Information Technology agreement
ITC International Trade Centre
ITCom International Trade Commission (US)
ISO International Organization for Standardization
ITLOS International Tribunal on the Law of the Sea
ITO International trade organization
LFN Least favoured nation
LFT Less favourable treatment
LLDCs Least developed countries
MAS Mutually agreed solution
MEA Multilateral environmental agreement
MERCOSUR Free trade area between Argentina, Brazil, Paraguay, and
Uruguay (Mercado del sur)
xxii List of Abbreviations
MFA Multi-fibre Agreement
MFN Most-favoured-nation
MITI Ministry of International Trade and Investment (now, METI)
MRA Mutual Recognition Agreement
NAFTA North America Free Trade Agreement
NAMA Non-agricultural market access
NGO Non-governmental organization
NME Non-market economies
NT National treatment
NTB Non-tariff barrier
NV Normal value
NVC Non-violation complaint
OCD Ordinary customs duty
ODC Other duties and charges
OECD Organization for Economic Cooperation
and Development
OPEC Organization of Petroleum Exporting Countries
PCIJ Permanent Court of International Justice
PFC Production flexibility contract
PGE Permanent Group of Experts
POI Period of investigation
PPA Protocol of provisional application
PSE Producer Subsidy Equivalent
PSI Principal Supplying Interest
PSIA Agreement on Pre-shipment Inspection
PTA Preferential Trade Agreement
R&D Research and development
RTAA Reciprocal Trade Agreements Act
QR Quantitative restriction
RBP Restrictive business practice
ROO Agreement on Rules of Origin
RPT Reasonable period of time
SATAP So as to afford protection
SCM Agreement on Subsidies and Countervailing Measures
SCt Supreme Court (US)
SG Agreement on Safeguards
SG&A Administrative, selling, and general expenses (AD)
SGS Société Générale de Surveillance
SI Substantial interest
SPS Agreement on the Application of Sanitary and
Phyto-sanitary Measures
STE State trading enterprise
TBT Agreement on Technical Barriers to Trade
TEC Treaty Establishing the European Community
TEDs Turtle excluding devices
TEE Taxes of equivalent effect
TMB Textiles Monitoring Body
 List of Abbreviations xxiii

TNC Trade Negotiating Committee

TPA Trade Promotion Authority
TPRM Trade Policy Review Mechanism
TRIMs Agreement on Trade-related Investment Measures
TRIPs Agreement on Trade Related Intellectual Property Rights
TRQ Tariff quota
UK United Kingdom
UN United Nations
UNCTAD United Nations Conference on Trade and Development
US United States
USSR Union of Soviet Socialist Republics
VER Voluntary export restraint
VCLT Vienna Convention on the Law of Treaties
WB World Bank
WCO World Customs Organization
WHO World Health Organization
WP Working party
WTO World Trade Organization
This page intentionally left blank
 Table of Cases
Please note: this table makes use of short titles of GATT Panel, WTO Panel and AB Reports. Full
report titles and citations can be found in the Appendices (pp. 456–64).

Accession of Hungary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

Akzo v EC Commission [1991] ECR I-3359. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Argentina—Footwear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94, 117–18, 367, 371
Argentina—Hides and Leather . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49–50, 56–7, 440, 442–5
Argentina—Poultry Antidumping Duties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Argentina—Preserved Peaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
Australia—Automotive Leather II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182–3, 422
Australia—Salmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .299, 305, 306–7, 307

Belgian Family Allowances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125, 133

Belgian Tax Allowances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
Border Tax Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213, 234, 240, 241, 253, 277, 280
Brazil—Aircraft. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171, 185–6
Brooke Group Ltd v Brown & Williamson Tobacco Corp (1993) 509 US 209 . . . . . . . . . . . . . 360

Canada—Aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182, 183–4, 185, 188–9, 409, 419–20

Canada—Autos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126, 133, 134, 182
Canada—Dairy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Canada—FIRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59, 339
Canada—Wheat Exports and Grain Imports . . . . . . . . . . . . . . . . . . . 243, 265, 315, 316, 317–19
Canada/EC—Article XXVIII Rights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106–7
Chile—Alcoholic Beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .226–7, 228, 229, 231–3
Chile—Price Band System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74–5

Dominican Republic—Import and Sale of Cigarettes . . . . . . . . . . . . . . . . . . . . . . . 77, 115, 335–8

EC—Agreements with Portugal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

EC—Animal Feed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126–7
EC—Asbestos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33, 128, 239–43, 252, 253, 257, 262, 263,
275–6, 279, 284, 288, 307, 414, 453
EC—Association with African and Malagasy States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165–6
EC—Bananas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
EC—Bananas III . . . . . . . . 33, 34, 64, 65–6, 89–92, 113, 125, 136–7, 243, 423, 424, 435–6, 441
EC—Bed Linen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
EC—Chicken Cuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87–8
EC—Citrus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
EC—Commercial Vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .211, 405
EC—Computer Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
EC—Hormones . . . . . . . . . . . . . . . . . . . . . . . . 299, 300, 301, 302, 305–6, 308–9, 312, 409, 410
EC—Imports of Beef . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
EC—Minimum Import Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
EC—Oilseeds I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412–13
xxvi Table of Cases

EC—Pipe Fittings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .342, 349, 351

EC—Poultry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57, 66, 125, 441, 443
EC—Sardines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33–4, 288, 292–4
EC—Selected Customs Matters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441–2, 444, 445–6
EC—Tariff Preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129, 133–4, 137, 143, 144–5,
147, 279, 280, 284, 453
EC Commission v Netherlands [2004] ECR I-11375. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312–13
EFTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

India—Autos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51–2, 59
India—Patents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
India—Quantitative Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63, 155, 334
Indonesia—Autos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

Japan—Agricultural Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

Japan—Agricultural Products II . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265, 299, 310, 311, 312, 409
Japan—Alcoholic Beverages I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234, 275
Japan—Alcoholic Beverages II. . . . . . . . . . . . . . . . . . . . . . . . . . .194–5, 217–18, 219, 221–2, 223,
224–5, 227, 230, 231, 236
Japan—Apples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301, 302–3, 310, 311
Japan—Film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50, 414, 415
Japan—Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45–8, 51, 54, 55, 62, 238
Japan—SPF Dimension Lumber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Japan—Taxes on Alcoholic Beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27, 28

Korea—Alcoholic Beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217, 218–20, 225–6, 227

Korea—Beef . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63, 244–7, 255, 264–5, 286, 315, 316–17
Korea—Dairy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370

Lotus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

Mexico—Antidumping Measures on Beef and Rice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348

Mexico—Corn Syrup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419

Spain—Unroasted Coffee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113–14, 124, 127–8

Tax Legislation Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28, 29

Tetrapak v EC Commission [1995] ECR II-762 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359–60
Turkey—Textiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31–2, 169–70, 173, 177–8

US—1916 Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424

US—AD Measures on OCTG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356, 357
US—Carbon Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
US—Certain EC Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
US—Corrosion–Resistant Steel Sunset Review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355, 357, 358
US—Cotton Yarn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
US—Countervailing Duty Investigation on DRAMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197–8
US—Countervailing Measures on Certain EC Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
US—Customs User Fee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116–17, 118, 124
US—FSC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27, 182, 186–8, 196–7, 237–8, 238, 243, 419
US—Fur Felt Hats. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
 Table of Cases xxvii
US—Gambling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265–6
US—Gasoline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255–6, 260, 261, 269, 270, 271–2, 274–5
US—Hot Rolled Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
US—Lamb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371–3, 374, 375, 411
US—Lead and Bismuth II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
US—Line Pipe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .158, 170–1, 173, 175, 338, 369, 373, 375
US—Malt Beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230, 234–5, 236
US—Nicaraguan Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322–30
US—Non–Rubber Footwear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124, 125, 135–6
US—OCTG Sunset Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356, 425
US—Offset Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339339
US—Restrictions on Imports of Sugar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76, 90
US—Section 110(5) Copyright Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
US—Section 337 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
US—Shrimp . . . . . . . . . . . . . . . . . . . . 27, 257, 259–60, 266–74, 277, 279, 280, 281, 284–5, 425
US—Softwood Lumber IV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197, 199, 200, 425
US—Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
US—Steel Safeguards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .370, 371, 378, 432
US—Superfund . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55–6, 230, 236
US—Taxes on Automobiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .230, 234, 235, 252
US—Textiles Rules of Origin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
US—Tuna (Mexico) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257, 259
US—Underwear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
US—Upland Cotton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .209, 210
US—Wheat Gluten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .367, 371
US—Wool Shirts and Blouses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
US—Zeroing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
This page intentionally left blank
 Table of Legislation
Agriculture Agreement (AG) . . . . 90, 91, 206,
207, 208, 210
Annex 2 . . . . . . . . . . . . . . . . . . . . . . . . .207
Art 4.1 . . . . . . . . . . . . . . . . . . . . . . . . 90, 91
Art 4.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Art 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . .87
Art 6.3 . . . . . . . . . . . . . . . . . . . . . . . . . .209
Art 6.5 . . . . . . . . . . . . . . . . . . . . . . . . . .207
Art 13 . . . . . . . . . . . . . . . . . . . . . . . 91, 208
Art 20 . . . . . . . . . . . . . . . . . . . . . . . . . .206
Art 21.1 . . . . . . . . . . . . . . . . . . 90, 208, 209
Antidumping Code (Agreement on
Implementation of Article VI
of GATT) . . . . . . . . . 24, 37, 321, 339,
342, 344, 350, 352,
361, 368, 372, 374,
375, 387, 448
Art 2 . . . . . . . . . . . . . . . . . . . . . . . 357, 358
Art 2.2 . . . . . . . . . . . . . . . . . . . . . . 345, 347
Art 2.2.1 . . . . . . . . . . . . . . . . . . . . . . . . 345
Art 2.2.1.1 . . . . . . . . . . . . . . . . . . . . . . .346
Art 2.2.2 . . . . . . . . . . . . . . . . . . . . . . . .346
Art 2.4 . . . . . . . . . . . . . . . . . . . . . .344, 346
Art 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
Art 3.2 . . . . . . . . . . . . . . . . . . . . . . . . . . 350
Art 3.3 . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Art 3.4 . . . . . . . . . . . . . . . . . . . . . . 349, 352
Art 3.5 . . . . . . . . . . . . . . . . . . . . . . . . . .349
Art 4.1 . . . . . . . . . . . . . . . . . . . . . . . . . .343
Art 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Art 5.4 . . . . . . . . . . . . . . . . . . . . . . . . . .343
Art 5.8 . . . . . . . . . . . . . . . . . . 347, 349, 356
Art 5.10 . . . . . . . . . . . . . . . . . . . . . . . . .344
Art 6 . . . . . . . . . . . . . . . . . . . 340, 347, 358
Art 6.1 . . . . . . . . . . . . . . . . . . . . . . . . . .340
Art 6.2 . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 6.3 . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 6.4 . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 6.5 . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 6.8 . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 6.9 . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 6.10 . . . . . . . . . . . . . . . . . . . . . 349, 358
Art 6.11 . . . . . . . . . . . . . . . . . . . . . . . . .340
Art 6.13 . . . . . . . . . . . . . . . . . . . . . . . . . 341
Art 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
Art 9.2 . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Art 9.5 . . . . . . . . . . . . . . . . . . . . . . . . . .348
Art 11.2 . . . . . . . . . . . . . . . . . . . . . . . . . 353
Art 11.3 . . . . . . . . . . . . . . . . . 352, 356, 358
Art 11.4 . . . . . . . . . . . . . . . . . . . . . 353, 358
Art 17.6 . . . . . . . . . . . . . . . . . 341, 342, 354
Art 18.1 . . . . . . . . . . . . . . . . . . . . . . . . . 339
Cobden-Chevalier Treaty 1860 . . . . . . . . .120
Convention on the International Trade of
Endangered Species (CITES). . . . . 285
Customs Valuation Code (Agreement
on Implementation of Article VII
of GATT and Protocol to the
Agreement) . . . . . . . . . . . 24, 119, 135
Dispute Settlement Understanding
(DSU) . . . . . . . . . . .22, 402, 407, 425,
428, 431, 432
Appendix 3 . . . . . . . . . . . . . . . . . . . . . .408
Art 3.2 . . . . . . . . . . . . . . . . . . . . . . 199, 268
Art 3.5 . . . . . . . . . . . . . . . . . . . . . . . . . .424
Art 3.6 . . . . . . . . . . . . . . . . . . . . . . . . . .424
Art 4.11 . . . . . . . . . . . . . . . . . . . . . . . . .402
Art 5 . . . . . . . . . . . . . . . . . . . . . . . . . . .405
Art 6.2 . . . . . . . . . . . . . . . . . . . . . . 406, 419
Art 10. . . . . . . . . . . . . . . . . . . . . . . . . . .406
Art 11. . . . . . . . . . . . . . . 341, 342, 409, 410
Art 12.12 . . . . . . . . . . . . . . . . . . . . . . . . 331
Art 13 . . . . . . . . . . . . . . . . . . . . . . . . . .408
Art 13.1 . . . . . . . . . . . . . . . . . . . . . . . . .409
Art 13.2 . . . . . . . . . . . . . . . . . . . . . . . . .408
Art 19. . . . . . . . . . . . . . . . . . . . . . . 416, 422
Art 21. . . . . . . . . . . . . . . . . . . . . . . . . . .403
Art 21.3(c) . . . . . . . . . . . . . . . . . . . . . . . 418
Art 21.5 . . . . . . . . . . . . . 404, 419, 420, 425
Art 21.6 . . . . . . . . . . . . . . . . . . . . . 424, 425
Art 22 . . . . . . . . . . . . . . . . . . . . . . . . . .423
Art 22.1 . . . . . . . . . . . . . . . . . . . . . 417, 428
Art 22.4 . . . . . . . . . . . . . . . . . 185, 422, 430
Art 22.6 . . . . . . . . . . . . . . . . . . . . . . . . . 108
Art 22.7 . . . . . . . . . . . . . . . . . . . . . . . . . 421
Art 22.8 . . . . . . . . . . . . . . . . . . . . . . . . . 425
Art 23 . . . . . . . . . . . . . . . . . . . . . . . . . .405
Art 23.1 . . . . . . . . . . . . . . . . . . . . .404, 405
Art 23.2 . . . . . . . . . . . . . . . . . . 92, 404, 405
Art 23.2(a) . . . . . . . . . . . . . . . 404, 405, 435
xxx Table of Legislation
Art 25 . . . . . . . . . . . . . . . . . . . . . . . . . .405
Art 26 . . . . . . . . . . . . . . . . . . . . . . . . . . 412
Art 26.1(c) . . . . . . . . . . . . . . . . . . . . . . . 418
Art 26.1(d) . . . . . . . . . . . . . . . . . . . . . . . 418
Art 26.2 . . . . . . . . . . . . . . . . . . . . . 416, 418
EC Treaty
Art 28 . . . . . . . . . . . . . . . . . . . . . . 248, 250
General Agreement on Tariffs and
Trade (GATT) . . . . . . . .26–7, 30, 218,
372, 390, 447–8
Art I . . . . . . . . . . . . . .52, 56, 64, 74, 122–3,
125, 126, 127, 129, 136,
137, 143, 211, 256, 331,
398, 399, 448
Art I.1 . . . . 124–8, 133, 134, 135, 136, 144
Art I.2 . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Art I(b)(iv) . . . . . . . . . . . . . . 27, 28, 29, 143
Art II . . . . . . . . . . 16, 23, 38, 51, 56, 68–70,
74, 81, 84, 92, 108, 1
13, 114, 332, 394
Art II.1(a) . . . . . . . . . . . . . . . . . . . . . . . . .87
Art II.1(b) . . . . . . . . . . . . 75, 76, 77, 78, 84,
87, 88, 89, 115, 336
Art II.2. . . . . . . . . . . . . . . . . . . . . . . . . . 117
Art II.2(c). . . . . . . . . . . . . . . . .115, 116, 117
Art II.7. . . . . . . . . . . . . . . . . . . . . . . . . . .90
Art III . . . . . . . . . . . . 44, 52, 55, 56, 58, 59,
60, 61, 68, 127, 195, 213,
215, 221, 227–34, 239,
244, 248, 249, 251, 252,
253, 255, 256, 257, 258,
262, 285, 287, 288, 307,
308, 331, 359, 386, 448
Art III.1 . . . . . . . . . . . . .216, 221, 222, 223,
228, 239, 240, 242, 251
Art III.2 . . . . . . . . . . . . . .56, 125, 212, 213,
216, 219, 221–6, 228–36,
238, 239, 240, 243,
251, 252
Art III.4 . . . . . . . . . . . . . . 34, 125, 214, 235,
237–44, 246, 247, 248,
255, 256, 286, 287
Art III.5 . . . . . . . . . . . . . . . . . . . . . . . . . 195
Art III.8 . . . . . . . . . . . . . . . . . . . . . 195, 211
Art IV . . . . . . . . . . . . . . . . . . . . 52, 53, 286
Art V . . . . . . . . . . . . . . . . . . . . . . . . 68, 331
Art VI . . . . . . . . . . . 35–6, 38, 44, 62, 180,
321, 338, 339, 449
Art VII . . . . . . . . . . . . . . . . . . . . . . 118, 119
Art VII.2 . . . . . . . . . . . . . . . . . . . . . . . . 118
Art VIII . . . . . . . . . . . . . . .114–15, 116, 118
Art VIII.1 . . . . . . . . . . . . . . . . . . . . . . . 116
Art VIII.1(a). . . . . . . . . . . . . . . . . . 116, 117
Art IX . . . . . . . . . . . . . . . . . . . . 124–5, 135
Art IX.1 . . . . . . . . . . . . . . . . . . . . . . . . . 125
Art X . . . . . . . . . . . . . 258, 436, 440–5, 449
Art X.1 . . . . . . . . . . . . . . . . . . 441, 442, 445
Art X.3. . . . . . . . . . . 34, 440, 443, 444, 445
Art X.3(a). . . . . . . . . . . . 441, 442, 443, 445
Art X.3(b) . . . . . . . . . . . . . . . . . . . 445, 446
Art XI . . . . . . . . . 35, 37, 38, 42–6, 49–62,
64, 68, 85, 159, 203, 237,
254, 259, 286, 322,
331, 381, 443
Art XI.1 . . . . . . . . . . . . . . 44, 47, 48, 51, 56,
57, 62, 202, 334
Art XI.2 . . . . . . . . . . 45, 138, 202, 203, 448
Art XI.3 . . . . . . . . . . . . . . . . . . . . . . . . .202
Art XII . . . . . . . . . . . . .37, 60, 331, 335, 384
Art XII.4a . . . . . . . . . . . . . . . . . . . . . . . 332
Art XIII . . . . . . . . . . . 63, 64, 65, 67, 72, 91,
211, 331, 436
Art XIII.1. . . . . . . . . . . . . . . . . . . . . . . . .64
Art XIII.2 . . . . . . . . . . . . . . . . . . . . . . . .64
Art XIII.2(d) . . . . . . . . . . . . . . . . . . . . . . 65
Art XIII.3(a) . . . . . . . . . . . . . . . . . . . . . . 65
Art XIV . . . . . . . . . . . . . . . . . . . . . . 67, 276
Art XV . . . . . . . . . . . . . . . 37, 321, 332, 335,
337, 338, 384
Art XV.2 . . . . . . . . . . . . . . . . . . . . . . . . 337
Art XV.6 . . . . . . . . . . . . . . . . . . . . . . . . 332
Art XV.9 . . . . . . . . . . . . . . . . . . . . 335, 337
Art XVI . . . . . . . . . . . . 37, 91, 181, 196, 321
Art XVI.1 . . . . . . . . . . . . . . . . . . . . . . . . 29
Art XVI.3 . . . . . . . . . . . . . . . . . . . . . . .203
Art XVI.4 . . . . . . . . . . . . . . . . . . . . . 28, 29
Art XVII . . . . . . . . . . . . . . . . . 210, 315–19
Art XVII.1 . . . . . . . . . . . . . . . . . . . . . . . 317
Art XVII.1(a) . . . . . . . . . . . . . . . . . . . . . 316
Art XVII.1(b) . . . . . . . . . . . . . .317, 318, 319
Art XVII.2 . . . . . . . . . . . . . . . . . . . . . . . 316
Art XVIII . . . . 37, 60, 63, 81, 321, 332, 334
Art XVIII.11 . . . . . . . . . . . . . . . . . . . . . 334
Art XVIII.12a . . . . . . . . . . . . . . . . . . . . 332
Art XVIII(b) . . . . . . . . . . . . . 331, 332, 334
Art XVIII(c). . . . . . . . . . . . . . . . . . . . . . . 63
Art XIX . . . . . . . . . . . . . . 37, 166, 321, 365,
366, 369, 371, 380,
381, 382, 384, 385
Art XX . . . . . . . . . . . . 37, 38, 44, 54, 61, 63,
67, 179, 228, 241, 247, 248,
249, 251, 253–61, 263, 265,
266, 271, 274, 275, 276, 277,
281, 282, 285, 286, 302, 307
 Table of Legislation
Art XX(a) . . . . . . . . . . . . 257, 261, 262, 276
Art XX(b) . . . . . . . . . . . 259, 261, 262, 263,
266, 267, 275
Art XX(d) . . . . . . . 247, 255, 263, 264, 265
Art XX(g) . . . . . . . . . . . . 54, 261, 266–70,
273, 279, 285
Art XX(i) . . . . . . . . . . . . . . . . . . . . . . . . .54
Art XX(j) . . . . . . . . . . . . . . . . . . . . . . . . .54
Art XXI . . . . . . . . . . . 37, 38, 44, 60, 61, 63,
67, 166, 282, 321–6,
328, 329, 330
Art XXI(b)(i)-(ii) . . . . . . . . . . . . . . . . . . . 67
Art XXI(b)(iii) . . . . . . . . . . . . 322, 324, 325
Art XXII . . . . . . . . . . . 21, 63, 77, 154, 172,
321, 398, 401
Art XXII.1 . . . . . . . . . . . . . . . . . . . . . . . 399
Art XXIII. . . . . . . . . . 21, 77, 154, 172, 328,
398, 401, 412
Art XXIII.1 . . . . . . . . . . . . . . . . . . . . . . 399
Art XXIII.1(a) . . . . . . . . . . . . 325, 326, 411
Art XXIII.1(b) . . . . . . . . 237, 238, 326, 327,
411, 412, 414, 415
Art XXIII.1(c) . . . . . . . . . . . . .327, 411, 415
Art XXIII.2 . . . . . . . . . . . . . . . . . . .326–30
Art XXIV. . . . 31, 32, 64, 81, 132, 148, 152,
153, 154, 156, 157, 158,
160, 161, 165–72, 174,
175, 177, 178, 368, 451
Art XXIV.4 . . . . . . . . . . . . . . . . . . . . . . 152
Art XXIV.5 . . . . . . .132, 153, 161, 168, 177
Art XXIV.5(a) . . . . . . . 32, 160–4, 170, 177
Art XXIV.5(b) . . . . . . . . 159, 160, 162, 171
Art XXIV.5(c) . . . . . . . . . . . . . . . . . . . . 171
Art XXIV.6 . . . . . . . . . . 106, 153, 162, 163
Art XXIV.7 . . . . . . . . . . . . . . . . . . 153, 156
Art XXIV.7(a) . . . . . . . . . . . . . . . . . . . . 157
Art XXIV.7(b) . . . . . . . . . . . . . . . . . . . . 157
Art XXIV.8 . . . . . . . . . . 132, 153, 164, 166,
167, 171, 177, 178
Art XXIV.8(a) . . . . . . . . . . . . . . . . 170, 177
Art XXIV.8(b) . . . . . . . . . . . . . . . . . . . . 171
Art XXV . . . . . . . . . . . . . . . . . . . . . . . . 395
Art XXV.1 . . . . . . . . . . . . . . . . . . . . . . . 395
Art XXV.3 . . . . . . . . . . . . . . . . . . . . . . . 395
Art XXV.4 . . . . . . . . . . . . . . . . . . . 390, 395
Art XXV.5 . . . . . . . . . . . . . . . . . . . . . . . 328
Art XXVI . . . . . . . . . . . . . . . . . . . . . . .390
Art XXVI.5(c) . . . . . . . . . . . . . . . . . .390–1
Art XXVII . . . . . . . . . . . . . . . . 81, 94, 95–6
Art XXVIII . . . . . . . . . . 71, 81, 94, 97–104,
106, 108, 109, 110, 112,
163, 384, 428
xxxi
Art XXVIII bis . . . . . . . . . . . 71, 79, 84, 203
Art XXVIII.1 . . . . . . . . . . . . . . . . . 104, 110
Art XXVIII.2 . . . . . . . . . . . . . . . . . 98, 104
Art XXVIII.3. . . . . . . . . . 99, 104, 107, 109
Art XXVIII.3b. . . . . . . . . . . . . . . . . . . . .99
Art XXVIII.4. . . . . . . . . . . . . . . . . 110, 111
Art XXVIII.5 . . . . . . . . . . . . . . . . . . . . . 109
Art XXIX. . . . . . . . . . . . . . . 5–6, 446, 449
Art XXIX.2 . . . . . . . . . . . . . . . . . . . . . . . . 5
Art XXX . . . . . . . . . . . . . . . . . . . . 112, 397
Art XXXI. . . . . . . . . . . . . . . . . . . . . . . . 393
Art XXXII . . . . . . . . . . . . . . . . . . . . . . . 391
Art XXXIII . . . . . . . . . . . . . . . . . . . . . . 391
Art XXXV . . . . . . . . . . . . . . . . . . . 394, 395
Art XXXVI . . . . . . . . . . . . . . . . . . 140, 141
Art XXXVI.8 . . . . . . . . . . . . . . . . . . . . 142
Art XXXVII . . . . . . . . . . . . . . . 140, 141–2
Art XXXVIII . . . . . . . . . . . . . . . . . 140, 142
London Draft
Annex 10 . . . . . . . . . . . . . . . . . . . . . . . 11
Part I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Part II . . . . . . . . . . . . . . . . . . . . . . . . . . 4, 5
Part III . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Part IV . . . . . . . . . . . . . . . . . . 138, 140, 142
Preamble . . . . . . . . . . . . . . . . . . . . . . . . . 11
Schedules . . . . . . . . . . . . . . . . . . . . . . 77, 90
General Agreement on Trade in Services
(GATS)
Annex . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Art I . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Art I.1 . . . . . . . . . . . . . . . . . . . . . . . . . .238
Art III . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Art V . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Art VI . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Art XVI . . . . . . . . . . . . . . . . . . . . . . . . . 331
Art XVII . . . . . . . . . . . . . . . . . . . . . . . . 331
Geneva Final Act . . . . . . . . . . . . . . . . . . . . . . 4
Protocol of Provisional Application . . . . . . 4
Government Procurement Agreement
(GPA). . . . . . . . . . . . 24, 210, 211, 212
Art V.1 . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Art V.3 . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Art XXIV.2 . . . . . . . . . . . . . . . . . . . . . . 211
Harmonized System Convention
Art 3.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Art 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Havana Charter for an International Trade
Organization . . . . . . . . . . . 3, 5, 6, 446
Art 18. . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Art 93.1(b)-(c) . . . . . . . . . . . . . . . . . . . . 327
Import Licensing Procedures Agreement
(ILA) . . . . . . . . . . . . . . . 24, 33, 41, 65
xxxii Table of Legislation
Art 1.4 . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Art 1.9 . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Art 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Art 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Art 3.2 . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Art 3.5 . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Art 3.5(1) . . . . . . . . . . . . . . . . . . . . . . . . .66
Information Technology Agreement . . . 82–4
Lomé Convention . . . . . . . . . . . . . . . . 64, 169
Montreal Rules . . . . . . . . . . . . . . . . . . 22, 401
Multi-fibre Agreement (MFA) . . . . . . . 67, 203
Pre-shipment Inspection
Agreement . . . . . . . . . . . . 119–20, 135
Art 4 . . . . . . . . . . . . . . . . . . . . . . . . . . .120
Art 4(h) . . . . . . . . . . . . . . . . . . . . . . . . .120
Rules of Origin Agreement . . . . . . . . . . . . 132
Art 1.2 . . . . . . . . . . . . . . . . . . . . 130, 131–2
Art 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Art 2(b) . . . . . . . . . . . . . . . . . . . . . . . . . 131
Art 2(d) . . . . . . . . . . . . . . . . . . . . . 130, 131
Art 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Safeguards Agreement
(SG) . . . . . . . . . . . . . . 37, 91, 321, 350,
365, 368, 369, 373,
374, 384, 387, 449
Art 2 . . . . . . . . . . . . . . . . . . . 366, 367, 369
Art 2.1 . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Art 2.2 . . . . . . . . . . . . . . . . . . . . . . . . . .380
Art 3.1 . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Art 4.1(c) . . . . . . . . . . . . . . . . . . . . . . . . 374
Art 4.2 . . . . . . . . . . . . . . . . . . 373, 374, 375
Art 4.2(a) . . . . . . . . . . . . . . . . . . . . . . . . 411
Art 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Art 5.1 . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Art 5.2(a) . . . . . . . . . . . . . . . . . . . . . . . . 375
Art 5.2(b) . . . . . . . . . . . . . . . . . . . . . . . . 376
Art 7.1 . . . . . . . . . . . . . . . . . . . . . . . . . . 378
Art 7.3 . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Art 7.5 . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Art 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
Art 8.1 . . . . . . . . . . . . . . . . . . . . . . 376, 377
Art 8.3 . . . . . . . . . . . . . . . . . . . . . . 377, 378
Art 11.2 . . . . . . . . . . . . . . . . . . . . . . . . . 381
Art 12.3 . . . . . . . . . . . . . . . . . . . . . . . . .377
Sanitary and Phyto-sanitary Measures
Agreement (SPS) . . . . . . . . .24, 34, 61,
252, 262, 287, 288,
291, 297, 298, 301,
308, 386
Annex A . . . . . . . . . . . . . . . . . 297, 298, 301
Art 1 . . . . . . . . . . . . . . . . . . . . . . . . . . .297
Art 2.2 . . . . . . . . . . . . . . . . . . 299, 302, 303
Art 2.3 . . . . . . . . . . . . . . 298, 305, 308, 309
Art 3.2 . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Art 3.3 . . . . . . . . . . . . . . . . . . . . . . . . . .309
Art 3.4 . . . . . . . . . . . . . . . . . . . . . . . . . .299
Art 5.1 . . . . . . . . . . . . . . 299, 300, 302, 307
Art 5.2 . . . . . . . . . . . . . . . . . . . . . . . . . .299
Art 5.5 . . . . . . . . . . . . . . . . 304–9, 312, 314
Art 5.6 . . . . . . . . . . . . . . 304, 312, 314, 409
Art 5.7 . . . . . . . . . . . 309, 310, 311, 312, 314
Art 13 . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Subsidies and Countervailing
Measures Agreement (SCM)
(Agreement on Interpretation
of Articles VI, XVI and
XXII of GATT) . . . . . . . . . .24, 37, 42,
196, 201, 208, 210,
321, 368, 372,
374, 375, 387
Annex 1 . . . . . . . . . . . . . . . . . . . . . . . . . 181
Art 1 . . . . . . . . . . . . . . . . . . . . . . . 196, 197
Art 1.1(a)(2) . . . . . . . . . . . . . . . . . . . . . . 198
Art 1.1(b) . . . . . . . . . . . . . . . . . . . . . . . . 198
Art 2.1(b) . . . . . . . . . . . . . . . . . . . . . . . .200
Art 3 . . . . . . . . . . . . . . . . . . . . . . . 181, 200
Art 3.1 . . . . . . . . . . . . . . . . . . . . . . 182, 183
Art 3.1(a) . . . . . . . . . . . . . 29, 183, 419, 420
Art 3.1(b) . . . . . . . . . . . . . . . . . . . . . . . .209
Art 3.1(c) . . . . . . . . . . . . . . . . . . . . . . . . 181
Art 4 . . . . . . . . . . . . . . . . . . . 185, 186, 201
Art 4.5 . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Art 4.7 . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Art 4.10 . . . . . . . . . . . . . . . . . 183, 185, 187
Art 5 . . . . . . . . . . . . . . . . . . . . . . .200, 202
Art 6.3 . . . . . . . . . . . . . . . . . . . . . . . . . .202
Art 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Art 7.9 . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Art 8 . . . . . . . . . . . . . . . . . . . . . . . . . . .200
Art 14. . . . . . . . . . . . . . . . . . . . . . . 198, 199
Art 14(d) . . . . . . . . . . . . . . . . . . . . . . . . 199
Art 31. . . . . . . . . . . . . . . . . . . . . . . . . . .200
Technical Barriers to Trade
Agreement (TBT) . . . . . . . 24, 33, 34,
61, 252, 262, 287,
288, 291, 295,
298, 299, 386
Annex 1 . . . . . . . . . . . . . . . . . . . . . . . . . 292
Annex 3 . . . . . . . . . . . . . . . . . . . . . . . . .296
Art 1.5 . . . . . . . . . . . . . . . . . . . . . . . 34, 288
 Table of Legislation
Art 2.1 . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Art 2.2 . . . . . . . . . . . . . . . . . . . . . . . 34, 295
Art 2.4 . . . . . . . . . . . 34, 291, 292, 294, 296
Art 2.5 . . . . . . . . . . . . . . . . . . . . . . 295, 296
Art 4.1 . . . . . . . . . . . . . . . . . . . . . . . . . .296
Art 4.2 . . . . . . . . . . . . . . . . . . . . . . . . . .296
Textiles and Clothing Agreement (ATC) . . . 67
Art 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Trade in Civil Aircraft Agreement . . . . . . . .24
Trade Related Intellectual Property
Agreement (TRIPS) . . . . . . . 150, 362,
398, 449
Trade-related Investment Measures
Agreement (TRIMS) . . . . . . . . . . . . 33
Understanding on Notification,
Consultation, Dispute
Settlement and Surveillance . . . . 21, 22
Annex . . . . . . . . . . . . . . . . . . . . . . . . . 21–2
United Nations Charter
Art 102. . . . . . . . . . . . . . . . . . . . . . . . . . .86
Vienna Convention on the Law of
Treaties . . . . . . . . . . . . . . 33, 88, 256,
342, 343
Art 31. . . . . . . . . . . . . . . . . . . . . . 6, 31, 275
Art 31.2 . . . . . . . . . . . . . . . . . . . . . . . . . .87
Art 70 . . . . . . . . . . . . . . . . . . . . . . . . . . 393
xxxiii
World Trade Organization Agreement
(Marrakesh Agreement) . . . . 22, 31–4,
431, 449
Annex 1A . . . . . . . . . . . . . . . 32, 33, 90, 288
Annex 3 . . . . . . . . . . . . . . . . . . . . . . . . .437
Art II . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Art IX . . . . . . . . . . . . . . . . . . 390, 396, 435
Art IX.2 . . . . . . . . . . . . . . . . . . . . . 396, 397
Art X . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Art X.2. . . . . . . . . . . . . . . . . . . . . . . .397–8
Art XI.1 . . . . . . . . . . . . . . . . . . . . . . . . . 391
Art XII . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Art XIII . . . . . . . . . . . . . . . . . . . . . 282, 395
Art XIV . . . . . . . . . . . . . . . . . . . . . . . . . 394
Art XV . . . . . . . . . . . . . . . . . . . . . . . . . . .97
Art XVI . . . . . . . . . . . . . . . . . . . . . . 27, 213
US legislation
Africa Growth and Opportunity Act . . . 131
Agricultural Adjustment Act 1933 . . . . .448
Continued Dumping and Subsidies
Offset Act . . . . . . . . . . . . . . . . . . . 339
Cuban Liberty and Democratic
Solidarity Act . . . . . . . . . . . . . . .330–1
Smoot-Hawley Act . . . . . . . . . . . . . . . . . . . 8
Webb Pomerene Act 1918. . . . . . . . . . . . .54
This page intentionally left blank
 1
From GATT 1947 to GATT 1994

1 The Genesis of the GATT 2

2 Why the GATT? 6
2.1 Why ask the question? 6
2.2 The negotiators say . . . 7
2.3 Economic theory suggests 12
2.4 And some unexplored avenues 18
2.5 To conclude (?) 20
3 The GATT in the GATT era: An Agreement, and an
Institution as Well 20
4 The GATT Rounds of Trade Liberalization 23
5 Expanding the GATT: The Emergence of the Codes 24
6 The GATT in the WTO era: Just an Agreement,
Albeit a Pivotal One 25
7 The Modern GATT discipline in a Nutshell 30
7.1 The GATT in its WTO context 30
7.2 The legal relation of the GATT with the WTO Agreement 31
7.3 The substantive content of modern GATT 34

Overview

The General Agreement on Tariffs and Trade (GATT) is part of the wider post-
Second World War effort to establish key multilateral institutions that would
contribute to more harmonious relations across nations. Its humble origin not-
withstanding, or maybe, because of its humble origin, the GATT managed to
foster un-paralleled cooperation among trading nations. In this chapter, follow-
ing a very brief account of the advent of the GATT, we ask the question why was
the GATT needed in the first place. A response to this question does not have a
mere historical value; it is highly informative about the objectives that the various
instruments committed to the GATT purport to achieve. To address this ques-
tion we first review the GATT preparatory work, and then, economic theory.
The preparatory work of the GATT is quite informative about what various
2 From GATT 1947 to GATT 1994

trading nations had in mind, but is not particularly helpful when it comes to
explaining the various instruments committed to the GATT edifice. Economic
theory has been struggling with this question, and has provided many insights
that help us better understand why the GATT was needed in the first place. We
still lack nevertheless, an internally coherent theory which will explain GATT
from A to Z. It is probably quite appropriate to think of the GATT as an instru-
ment that combats international beggar thy neighbour policies. The next question
is how should we define beggar thy neighbour coming under the purview of the
GATT, and what does the GATT do to address it. We will thus briefly discuss, in
this chapter, the GATT discipline which could be summarized as follows: since
any policy might affect trade, the GATT binds both trade and domestic instru-
ments, albeit at varying degrees. The GATT outlaws quotas. Tariffs, in turn, will
be negotiated, bound, applied on a non-discriminatory basis, and, eventually,
reduced in the context of trade liberalizing rounds. The GATT has concluded
eight such rounds so far. Its success in reducing tariff barriers around the world
has been hailed as monumental. By virtue of the non-discrimination obligation,
it requires that World Trade Organization (WTO) members do not afford pro-
tection to their domestic production through the use of domestic instruments. It
follows, that the only form of protection that WTO members can offer to their
domestic production is through tariffs. WTO members can, nevertheless, raise
the originally negotiated protection, in case one of the state contingencies specif-
ically provided for in the GATT contract occurs.

1 The Genesis of the GATT

Adlai Stevenson is reported to have said that trade is quite boring, and that its greatest
need was for fresh clichés.¹ Viewed from the perspective of a statesman, this is probably
a wise statement: trade talks until recently, with the emergence of the anti-globalization
movement, would not make the headlines the way other events do. This was probably
even more the case when the GATT was created amidst other towering achievements
in the post-Second World War era, such as the advent of the United Nations, the World
Bank (WB), and the International Monetary Fund (IMF). The GATT came into life
lame, since its institutional umbrella never saw the light of the day. Yet its initial low key
character has proved to be a key ingredient of its institutional success. For the GATT is
a success story, an important one. The reputed US statesman would probably react in a
different way, had he been asked about trade today.
The GATT is an off-shoot of a broader and more ambitious project, the
International Trade Organization (ITO).² The ITO was intended to serve as the

¹ See Aaronson (1999).

² See, for example, Dam (1970), Gardner (1980), Hoekman and Kostecki (2001), Hudec (1975),
Jackson (1969), Matsushita et al (2006), Trebilcock and Howse (2005), Van den Bossche (2005),
 The Genesis of the GATT 3

institutional foundation to administer the Havana Charter for an International

Trade Organization, which was drafted following a series of negotiations of the
so-called Preparatory Committee for the ITO, held between 1946 and 1948 in
London, New York, Geneva and finally in the Cuban capital bearing its name.³
The Havana Charter was supposed to regulate trade (eg trade in goods and com-
modity agreements), and the use of public policy as related to domestic employ-
ment, economic development and restrictive business practices.⁴
The ITO was negotiated subsequent to the Bretton Woods negotiations, which
led to the creation of the WB, and the IMF. Although the literature discuss-
ing the formation of the Bretton Woods institutions sometimes suggests that the
ITO was supposed to constitute the third pillar of Bretton Woods, there is lack
of evidence in its negotiating history to support this view.⁵ Nevertheless, it is
certainly true that all three institutions shared one overarching goal: a genuine
effort to help promote international economic cooperation during the post-Sec-
ond World War period. The design and aim of these three institutions distin-
guished them from the major institutional arrangement of that time, the United
Nations, which was intended to be the overarching international organization to
help avoid another world war, without however addressing the underlying rea-
sons for conflict. Their different function notwithstanding, the three institutions
complemented the United Nations: promoting trade among nations,⁶ providing
technical assistance to address development-related issues, and financing devel-
opment policies were meant to be the international community’s response to the
ongoing quest of how best to address the causes for conflict.⁷

Wilcox (1949), and Zeiler (1999). The enactment of the US Reciprocal Trade Agreements Act
(RTAA) in 1934 greatly facilitated this process. Irwin (2005, pp 204ff ) very convincingly makes
the point that the RTAA ‘fundamentally changed US trade politics by shifting tariff authority
from Congress, which proved very responsive to domestic import-competing industries, to the
executive branch, which was more apt to consider the national interest and use tariff negotiations
in the service of foreign policy objectives.’ The RTAA, however, was met with a lot of skepticism
in some quarters. The prevailing view at the time, as Aaronson (1999) reports, was that trade pro-
tectionism protects US jobs. RTAA was thus portrayed as a mere temporary measure, and not as
permanent feature in the US legal arsenal.
³ Meade (1942) had persuasively argued in favour of an international trade organization. Meade
participated in the GATT/ITO negotiations as a member of the UK delegation.
⁴ The ITO was definitely a far-reaching agreement. For example, Art 7 discussed what now-
adays has been termed social dumping. Charnovitz (1995) rightly points out that social dumping
in the ITO framework could provide the legal cause for an action on nullification and impairment
(whereby, the affected trading partner will be requesting some form of compensation). In today’s
GATT, the only clause which might suggest that impermissibly low cost-products can be legitim-
ately excluded from a given market, is Art XX(e) of the GATT which allows WTO members to
block imports of products produced through prison labour.
⁵ See the discussion in Jackson (1969), and Palmeter and Mavroidis (2004).
⁶ Dam (2004) reports that as late as 1916, the majority of US congressmen considered tariffs to
be a domestic issue. A notable exception among them was Cordell Hull, a genuine free-trader, who
believed that trade ‘dovetailed with peace’, see Hull (1948, pp 81ff ).
⁷ Henry Cabot Lodge used to say that the UN system was designed to avoid hell, but it was no
guarantee for heaven.
4 From GATT 1947 to GATT 1994

The GATT, a less ambitious (than the ITO) project was being negotiated in
Lake Success, New York—a village in northwest Long Island which had also
been the temporary headquarters of the United Nations from 1946 to 1951.
Between January and February 1947, state representatives negotiated the GATT,
a legal instrument aimed at relaxing government-mandated trade protection.
A relatively speaking homogeneous group (with few outliers) prepared the first
draft of the GATT.⁸ The GATT was not designed in any shape or form to be an
institution; rather, it was simply a trade agreement which was eventually to come
under the aegis of the ITO, once that would-be institution would be ratified and
enter into force. Whereas no reservations were allowed in the final text, reserva-
tions were quite appropriate during the drafting stage.⁹
The negotiations on the GATT text were complemented by negotiations on
tariff reductions: between April and October 1947, the state representatives,
negotiating under the institutional umbrella of the Preparatory Committee for
the ITO, conducted a round of tariff negotiations at the European office of the
United Nations in Geneva, Switzerland. In effect, this was the very first round
of multilateral trade negotiations.¹⁰ The outcome of these negotiations coupled
with the document negotiated at Lake Success together constituted the Geneva
Final Act, which included the Protocol of Provisional Application (PPA). Under
the terms of the PPA, the governments that participated in the negotiations
undertook to fully apply Part I of the GATT (dealing with tariff concessions and
the most-favoured-nation, or most-favoured-nation (MFN) clause) and Part III
of the GATT (containing provisions dealing with administrative issues). These
governments further undertook to apply Part II of the GATT (the heart of the
GATT, covering national treatment, antidumping, subsidies, safeguards, bal-
ance of payments, prohibition of quantitative restrictions, general exceptions to
the obligations assumed and dispute settlement) ‘to the fullest extent not incon-
sistent with existing legislation’.¹¹ The GATT entered into force on 1 January
1948 by virtue of the PPA. Its original 23 members were: Australia, Belgium,
Brazil, Burma, Canada, Ceylon, Chile, Republic of China, Cuba, Czechoslovak

⁸ The participants were: Australia, Belgium, Brazil, Canada, Chile, China, Cuba,
Czechoslovakia, France, India, Lebanon, Luxembourg, Netherlands, New Zealand, Norway,
Union of South Africa, Union of Soviet Socialist Republics (USSR), United Kingdom, United
States (see Drafting Committee of the Preparatory Committee of the UN Conference on Trade
and Employment, Draft GATT, E/PC/T/C.6/85 of 15 February 1947). The USSR ultimately did
not participate in the GATT/ITO endeavour. Penrose (1953, pp 12ff, and 90ff ) highlights the role
that some eminent economists, such as John Meynard Keynes, James Meade, and Lionel Robbins,
played in drafting the GATT. The final text was drafted by John Leddy, a US public official who
had also been trained in economics. Indeed, one important difference between the negotiation of
the GATT and, say, the Uruguay round agreements, is the identity of the delegates: the compos-
ition of national delegations has been increasingly politicized over the years.
⁹ See § 2 of the Drafting Committee of the Preparatory Committee of the UN Conference on
Trade and Employment, Draft GATT, E/PC/T/C.6/85 of 15 February 1947.
¹⁰ See 55 UNTS 187 (1947). 45,000 tariff concessions were negotiated, that is roughly $10
billion worth of trade, see Drache (2003).
¹¹ See UNITS 308, 1 (a) and (b) (1947).
 The Genesis of the GATT 5

Republic, France, India, Lebanon, Luxembourg, Netherlands, New Zealand,

Norway, Pakistan, Southern Rhodesia, Syria, South Africa, the United Kingdom,
and the United States.
The application of Part II only to the extent consistent with existing domestic
legislation is often referred to as ‘grandfather rights’.¹² Those member states which
had existing legislation inconsistent with provisions of the GATT at the time it
entered into force were allowed to continue applying these GATT inconsistent
measures. The inconsistency of domestic legislation of some member states with
provisions of the GATT did not create many problems since the grandfathering
provision was meant to be temporary as set out in Art XXIX: 2 of the GATT:
Part II of this Agreement shall be suspended on the day on which the Havana Charter
comes into force.
The entry into force of the GATT was provisional.¹³ As explained in Jackson
(1969), the GATT could have entered into force on a definitive basis, once a given
number of countries representing a high percentage of international trade would
have agreed to do so. It was decided, nevertheless, not to proceed in this way in
order to avoid creating discrepancies between those who had accepted the GATT
on a provisional basis, and those who would be accepting it definitively so. The
GATT also entered into force independently of the ITO: as explained in more
detail below, the adoption of the Havana Charter, which would signal the advent
of the ITO, was expected to occur within a fairly short period following the entry
into force of the GATT, and the plan was that the GATT would come under the
aegis of the ITO. In the meantime, GATT Art XXIX explicitly governed the
legal relationship between the GATT and the ITO in the following terms:
Article XXIX
The Relation of this Agreement to the Havana Charter
1. The contracting parties undertake to observe to the fullest extent of their executive
authority the general principles of Chapters I to VI inclusive and of Chapter IX of
the Havana Charter pending their acceptance of it in accordance with their constitu-
tional procedures.
2. Part II of this Agreement shall be suspended on the day on which the Havana Charter
enters into force.
3. If by September 30, 1949, the Havana Charter has not entered into force, the contract-
ing parties shall meet before December 31, 1949, to agree whether this Agreement
shall be amended, supplemented or maintained.

¹² A good example is the US countervailing duty law which did not require a determination
of material injury as required by Art VI, Part II of the GATT, until the United States agreed to
include such a requirement for signatories to the so-called 1979 Tokyo Round Subsidies Code or
for other countries that entered into comparable bilateral agreements with the US.
¹³ Eric Wyndham-White was the first Executive Secretary of the GATT heading a small secre-
tariat. Eventually, in subsequent years, his post was renamed as Director-General of the GATT.
6 From GATT 1947 to GATT 1994
4. If at any time the Havana Charter should cease to be in force, the CONTRACTING
PARTIES shall meet as soon as practicable thereafter to agree whether this Agreement
shall be supplemented, amended or maintained. Pending such agreement, Part II of
this Agreement shall again enter into force; Provided that the provisions of Part II
other than Article XXIII shall be replaced, mutatis mutandis, in the form in which
they then appeared in the Havana Charter; and Provided further that no contracting
party shall be bound by any provisions which did not bind it at the time when the
Havana Charter ceased to be in force.
5. If any contracting party has not accepted the Havana Charter by the date upon which
it enters into force, the CONTRACTING PARTIES shall confer to agree whether,
and if so in what way, this Agreement in so far as it affects relations between such
contracting party and other contracting parties, shall be supplemented or amended.
Pending such agreement the provisions of Part II of this Agreement shall, notwith-
standing the provisions of paragraph 2 of this Article, continue to apply as between
such contracting party and other contracting parties.
6. Contracting parties which are Members of the International Trade Organization
shall not invoke the provisions of this Agreement so as to prevent the operation of
any provision of the Havana Charter. The application of the principle underlying this
paragraph to any contracting party which is not a Member of the International Trade
Organization shall be the subject of an agreement pursuant to paragraph 5 of this
Article.
Since the Havana Charter, as we will see in more detail infra, never came to fru-
ition, the GATT was applied for 47 years oxymoronically pending the advent
of the ITO. Before we continue with the history of the GATT however, it is
probably appropriate to ask at this point the question why was the advent of the
GATT necessary? We already alluded above that the ITO was thought of as part
and parcel of a wider construction that would foster international cooperation in
the post-Second World War era. Let us add some detail to this discussion now.

2 Why the GATT?

2.1 Why ask the question?

The WTO judge must ask this question: by virtue of Art 31 of the Vienna
Convention on the Law of Treaties (VCLT), the WTO judge must, when inter-
preting the GATT, take into account the object and purpose of the treaty.¹⁴ Absent

¹⁴ Dispute Settlement Understanding (DSU), Art 3.2 obliges the WTO judge to refer to cus-
tomary rules of interpretation, when interpreting the WTO. The Appellate Body (AB), in its very
first report (US—Gasoline) held that reference to customary rules of interpretation obliged the
WTO judge to have recourse to the VCLT-system: the judge must, according to Art 31 of the
VCLT, interpret an agreement in good faith. To this effect, the judge, will interpret the agreement
in accordance with the ordinary meaning of its terms, in their context, taking into account the
object and purpose of the treaty, and any subsequent practice or agreement relating to the same
subject-matter, as well as any other relevant rule of public international law. If need be, and only in
 Why the GATT? 7

some knowledge as to what is being sought through the advent of the GATT, it
will be difficult for the judge to establish a coherent understanding of the instru-
ments committed: the various instruments committed to the GATT serve a par-
ticular purpose, the purpose sought by the founding fathers of the GATT, and not
any purpose. Understanding the rationale for the GATT is, thus, the necessary
first step towards evaluating (and interpreting) its various instruments.
There is an intellectual reason as well, justifying the asking of the question:
we can learn what the objectives and purposes of the founding fathers were. A
response to this question will elucidate us as to the original design of the GATT.
Absent an understanding of the rationale behind the advent of the GATT, it will
be difficult, if not irresponsible altogether, to suggest what needs to be changed in
the current edifice.
Trade liberalization can serve different purposes; it is not as if one can hit one
bird only with this stone. In Lewis Carroll’s oft quoted passage from Alice in
Wonderland, Alice asks ‘where do we go from here?’, only to get the response ‘it
depends where you want to go’. Delving into the question ‘why a GATT?’ is the
question a wondering Alice would ask the founding fathers of the GATT.
The natural place to entertain this question is the historical account of the
GATT. As we will try to show in what follows, a coherent response to this ques-
tion is hardly provided by the negotiating history of the GATT: we know few
things about the motives of countries other than the United States and the
United Kingdom, and we also know that these two countries, the main negotiat-
ing partners of the GATT, did not have a common understanding of the ration-
ale for the GATT; probably for this reason, the common intention of the parties
as reflected in the GATT preamble, is at a level of generality that makes it simply
uninformative, when it comes to using such knowledge in order to interpret the
various instruments committed. Economic theory has attempted to provide a
response to this question. Following a quick survey of the literature offered infra,
our conclusion is that we are still lacking a framework that explains in a coherent
manner the rationale for the GATT. Theory, nevertheless, is still evolving. And,
anyway, the existing level of research has a lot to offer, in terms of analytical tools,
to the trade practitioner (be it a negotiator or a judge) aiming at understanding
the GATT, and eventually, making it a better institution.

2.2 The negotiators say . . .

One, probably natural, way to start this discussion is to look into the (official)
records explaining the intent of the founding fathers of the GATT. There are
two revelation mechanisms: national and multilateral. Under the former, one
could extend research to a wide range of possible sources, ranging from official

accordance with the limited conditions of Art 32 of the VCLT, the judge will further refer to the
preparatory work of the agreement.
8 From GATT 1947 to GATT 1994

constitutional records to unofficial representations. Historians will, for good rea-

sons, normally privilege the former, a choice followed here as well.¹⁵ The prepara-
tory work of the work of the GATT exhausts the multilateral sources.
Researching into national sources is, of course, informative about national per-
ceptions but not necessarily about common intentions: any time, for example, we
observe variance between a nationally defined objective and the end result of the
international negotiation, we can legitimately cast doubt on the validity of the
national intention to act as explanatory variable for the common understanding.
We should probably start this discussion by pointing out that we do not know
much about national positions of the original participants, other than the United
Kingdom and the United States. Hart (2002) offers a very comprehensive account
of Canada’s priorities. His main conclusions are that Canada essentially played
second fiddle to the UK and US negotiators who monopolized the negotiations,
and was quite happy with the introduction of MFN which, in its view (and cor-
rectly so), was a means to curtail the dominance of the then two super-powers.¹⁶
Capling (2001) discusses the Australian participation. Once again we are left
with the impression that the Australian participation did not heavily influence
the negotiated outcome. In Capling’s account, Australia was largely pre-occupied
with sectoral negotiations, and even threatened to walk out of the negotiating
room when the United States initially refused to reduce its duties on wool and
wool products.
The overwhelming majority of historical accounts focuses on the US and UK
negotiating positions.
Irwin (1996a, and 2005) delves into the US constitutional process. He aptly
describes the context in which US trade policy was being shaped in order to
describe the goals sought by the US administration. The United States was com-
ing out from the Great Depression. Legislative initiatives, such as the Hawley-
Smoot Act, although they have been over-dramatized, help us understand the
prevailing attitude towards crisis-management in that period of time. In Irwin’s
account, the United States was looking to reduce tariffs around the world. US
Congress nevertheless, the US administration felt, was looking for product by
product tariff reductions, and this was a constraint imposed on US negotiators.
The US negotiators were also unequivocally for the dismantling of preferences
(largely practiced by the UK government at that time); crucially, the US admin-
istration was looking for a reduction of the role of the state, taking the view that
trade would be safer in private hands.¹⁷ It should be noted that, at that time, the
US policy-makers could count on visionaries, such as Cordell Hull, who genu-
inely believed that they were responding to a call of history and were contributing

¹⁵ This approach is also followed because it is not my aim to conduct extensive research on this
score, but, to a large extent, to rely on existing research, which is quite rich anyway.
¹⁶ See Hart (2002, pp 133ff ).
¹⁷ Miller (2000) takes exactly the same view. See also Gardner (1980), Jackson (1969), Dam
(1970), and Zeiler (1999) on this score.
 Why the GATT? 9

to world peace by establishing a world trade institution. Moving the competence

on trade issues from Congress to the Executive, as noted supra, reinforced the
voice of people like Cordell Hull.
The same author (Irwin) also explains the UK attitude which was quite dif-
ferent, in that it was quite pragmatic: the United Kingdom, the only other state
with considerable bargaining power at the time, sought tariff reductions, but also
the maintenance of a system of preferences that would allow it to trade with its
Commonwealth partners on a discriminatory basis. The UK government, largely
influenced by the thinking of John Maynard Keynes, who was acting as policy
adviser, believed that the UK government would have less to offer to US tariff reduc-
tions, if it had unilaterally removed its imperial preferences. In a way, reducing its
imperial preferences was a quid pro quo for US tariff reductions. In Miller’s (2000)
account, a large part of the negotiations concerned the exchange between UK
imperial preferences against US tariff reductions.¹⁸ Keynes must also be behind
the UK negotiating position, that the role of state in the post-war era should not
be put into question. The two delegations thus had a different approach on the
issue of state trading, cartels, and quotas: Miller (2000) explains that the UK
delegation, in favour of a strong role for government in the post-Second World
War era, insisted on allowing for quotas for balance of payments (BoP) reasons.
In Miller’s (2000) account, the pragmatic, government-friendly attitude of the
UK delegation prevailed in four key issues: preferences (which were not elimi-
nated), quotas (which were allowed for BoP reasons), state trading (which could
legitimately persist, provided that it operated on non-discriminatory terms), and
cartels (that were not disallowed).¹⁹
There was a substantial overlap as well. The British were not opposed to the
idea of creating a trade institution, which they viewed as means to curtail US
dominance. Both governments seem to believe that trade expansion (through
the signing of the GATT) was an appropriate means to fight unemployment
that crippled the US economy in the pre-New Deal era and was looming all over
Europe in the aftermath of the Second World War.²⁰ It should come as no sur-
prise that a specific provision in an earlier GATT draft was dedicated to this
issue.²¹ However, whereas the US delegates viewed opening of trade as an antidote,

¹⁸ Recall that, because of Hawley-Smoot, the US was practicing abnormally high tariff s imme-
diately after the Second World War and had thus a lot to offer in terms of tariff reductions.
¹⁹ The provisions on cartels (restrictive business practices, RBPs) were included in Chapter V of
the ITO Charter. The non-advent of the latter meant that the only legal relevance of this provision
was through the rather innocuous Art XXIX of the GATT reflected supra.
²⁰ Gardner (1980) underscores this point. Modern research casts doubt on the usefulness of trade
instruments to solve unemployment problems, see, inter alia, Krugman (1995), and Dewatripont
et al (1999) for analyses concerning the effects of trade on unemployment on the two sides of the
Atlantic. This is not to deny, of course, that trade and employment policies do interact, and that
the formulation of coherent policies can be beneficial. Jansen and Lee (2007) offer a survey of the
most recent empirical literature on this score which questions some of the orthodoxies of the past.
²¹ GATT, Art XVI of the New York Draft entitled Maintenance of domestic employment recog-
nized the right of every government to aim to achieve full employment, and imposed an obligation
10 From GATT 1947 to GATT 1994

the UK government had a more cautious approach on this issue: arguably influ-
enced by Keynes’ thinking, the UK delegates did not deny that trade could
indeed be beneficial on this score; they seem to privilege nevertheless, short-term
macroeconomic solutions to address unemployment. In the dominant UK view,
a drop in tariffs would not in and of itself help achieve the GATT objective to
end unemployment; wealthy domestic markets (with little if any unemployment)
would in turn drive trade expansion.²²
One can not nevertheless, maintain the view that fighting unemployment
was the only concern: many factors (other than trade) are better suited to fight
unemployment, and this much was known at that time as well. Trade expansion
could probably be of some help in this endeavour,²³ but trade alone, as suggested
above, could not be the answer to the problems the US society was facing. It
follows that a look into national constitutional processes is not very informative
about a common rationale for the GATT: the two most important players shared
(to some extent) the view that trade expansion would play an important role in
achieving full employment and increase welfare, although they disagree as to how
trade expansion should be achieved; they further buy into the idea that, at that
juncture, promotion of multilateral cooperation should be pursued, and that an
international trade organization is part of a wider cooperative game. The cost of
MFN was probably higher for the United States: before the GATT/ITO negotia-
tions, the US government did not practice widespread MFN in its trade relations;
in what seems to be a fairly rational decision, the US government practised condi-
tional MFN, the condition being the bargaining power of the other signatory.²⁴
In negotiating the GATT/ITO, not only did the US accept, counter-intuitively
to some extent, substantive MFN, it also accepted procedural MFN (multilateral-
ism), since it accepted to essentially negotiate, rather than act unilaterally. At
the same time, as we saw above, elimination of imperial preferences was a major
request by the US government, whereas keeping them as negotiating ploy was
very much the UK response.
What governments were after, it seems, is to curtail the potential for unilat-
eral action. They seem to realize that unilateral actions by other countries can
have negative (external) effects on their own welfare: a high US tariff could lead

on states striving for full employment to avoid creating balance of payment problems on their
trading partners. According to the companion provision (Art XVII), governments which aimed
at promoting the establishment or reconstruction of particular industries would be free to do so,
provided that they negotiated a settlement with the affected states, see UN Economic and Social
Council Doc E/PC/T/C.6/85 of 15 February 1947.
²² Th is is very much a Keynesian view. Keynes cared a lot about short-run effects and tended to
pay less attention to long run: a drop in tariff s could, for example, in the short run lead to domestic
unemployment, and such effects pre-occupied him. It is him after all who famously used to say ‘in
the long run we are all dead’.
²³ This could be the case, were the United States to open up its import market in sectors where
no domestic production existed, while requesting, in return, the opening up of foreign markets in
goods where US production existed (assuming no capacity constraints).
²⁴ See Hawkins (1951, pp 81ff ), and Miller (2000, pp 10ff ) on this score.
 Why the GATT? 11

to increased unemployment in the United Kingdom. We have little evidence,²⁵

nevertheless, as to precisely which external effects (besides fighting unemploy-
ment) they had in mind.
The second revelation (of the intentions) mechanism is the GATT itself, and
more precisely, negotiating documents and the GATT preamble. The GATT
preamble is informative with respect to the common aspirations of the contract-
ing parties, but not that helpful when it comes to using the common aspirations
as guidance in order to interpret the various GATT (committed) instruments.
Raising standards of living, ensuring full employment, making full use of the
resources of the world, are mentioned as the objectives sought through the
GATT. The objectives sought are thus expressed at a very high level of generality:
they cannot serve as compass that will guide us on the direction to choose when
interpreting the various instruments committed to the GATT; the objectives,
as expressed in the GATT preamble, are hard to implement through the use of
the instruments committed. They are general aspirations that can be achieved
through various government (and private) actions, trade liberalization being at
best, a contributing factor.²⁶
Nevertheless, the preamble is not the only place to look at. Looking into various
negotiating documents is quite informative as well. There are of course, hundreds
of negotiating documents relating to the GATT that were circulated between
1946 and 1948 during the London Conference, the New York Conference, the
Geneva Conference, and the Havana Conference. It is probably appropriate to
concentrate on the documents explaining the heart of the GATT, the tariff bar-
gain. Rationally, negotiators focused on barriers which constituted the main
market access impediment at the time and were probably hiding the relevance
of NTBs, assuming they were already important at that time. Recall that the
institutional provisions in the current GATT text were added only in an effort to
ensure that GATT’s life until the entry into force of the ITO would be smooth.
Had the ITO entered into force simultaneously with the GATT, there would
have been, presumably, no institutional provisions in the latter. The tariff bargain
is the heart of the GATT, and it is best explained in Annexure 10 of the London
Draft:²⁷ negotiations should take place under strict reciprocity (in the sense,
that countries should not be expecting one way preferences), and following the

²⁵ Absent, of course, the rather generic references to unemployment and macro-economic

stability.
²⁶ Irwin (2005) rightly underscores that, prior to the late forties, at least in the United States,
there was no statement or generally held view that the government was responsible for maintaining
full employment, and to avoid recessions. The Council of Economic Advisors (CEA), established
by the US Employment Act of 1946, was a watershed: the government was explicitly assuming
responsibility for macro-economic management. The preamble is broad, and thus unhelpful when
it comes to using it as a compass that will help us orientate the various GATT instruments, but it
does reflect a new view of the government’s role, and thus, very much the mindset of the founding
fathers.
²⁷ See E/PC/T/33.
12 From GATT 1947 to GATT 1994

principal supplier rule. The stated rationale for the latter is that, countries granting
a concession to principal suppliers would be in a position to extract the maximum
concession as counterpart to their offer. Reciprocity and principal supplier rule
are key ingredients of the terms of trade explanation of the GATT, as we will see
in what follows. Although no explicit reference to this theory is made in the vari-
ous negotiating documents, one should not outright exclude its relevance during
negotiations. At the very least, there is coincidence between what happened dur-
ing the negotiations and the manner in which economic theory tried to explain
the rationale for the GATT in subsequent years.²⁸
In a nutshell, a survey of the revelation mechanisms suggests that signing a
trade agreement (the GATT) was perceived as a means to constrain the behav-
iour of trading partners, and not of domestic lobbies, with the (predominantly
US) idea that the GATT would be a contribution towards establishing world
peace.

2.3 Economic theory suggests

Our walk through the revelation mechanisms discussed above was not counter-
productive: if anything, it seems that trade liberalization was viewed as a response
to unilateral behaviour. We are still in the dark nevertheless, as to which unilat-
eral behaviour trade agreements are meant to address. So, a detour to economic
theory is warranted in order to form an opinion on this score. Economists will,
usually, ask the following questions:
(1) are there gains from trade liberalization?
(2) if yes, why are countries not pursuing trade liberalization anyway, that is, in
the absence of a written commitment (agreement) to do so?
Assuming a positive response to (1) above, it is the response to (2) that will take us
close to the response to the question: what is the rationale for trade agreements?
One will be hard-pressed to find, among economists, people arguing that there
are no gains from trade liberalization. Indeed, as explained in numerous writings
from Smith’s classic The Wealth of Nations onwards, all nations have a compara-
tive advantage in producing particular goods, and gains from trade result from
specialization.²⁹ The modern understanding of international trade policy is, to a

²⁸ See infra in section 2.3 of this chapter.

²⁹ For a very simple, yet very clear explanation, of gains from trade, see Krugman and Wells
(2005). Rose (2004) in a provocative paper, argued that there is not much evidence that the GATT/
WTO had contributed to trade liberalization: in his view, nations would have behaved more or less
in the same way even in the absence of an institutional framework such as the GATT/WTO. Irwin
(2005) first criticized Rose’s approach, rightly pointing that the author himself accepts the impact
of the WTO when fi xed effects have been taken into account (as they should). Subramanian and
Wei (2007) have shown that the WTO has had a strong positive impact on trade, amounting to
120% additional world trade (or $8 trillion in 2000 alone). They also show that trade promotion
is uneven and that it depends on what do countries do with their membership, who they negotiate
 Why the GATT? 13

large extent, still based on the notion of comparative advantage, as further devel-
oped by Ricardo. The obvious next question is why then does not trade liberal-
ization³⁰ take place unilaterally? Why, in other words, do we need the GATT in
order to do what we would have been anyway doing (even without the GATT)?
Two theories have been advanced to answer this question: the commitment,
and the terms of trade theory. The commitment theory has been developed in
various research papers. The focus of this approach is the relationship between
government and its private sector: a government will choose its trade policy and
will commit it in an international agreement signed to this effect; the private
sector will act accordingly. The gain for the government is that investment deci-
sions are forestalled; it will lose, however, contributions by the various lobbies.³¹
There are many good arguments in support of this view: for a government facing
unpopular (for segments of the society) choices, the argument ‘I have tied my
hands to the mast’³² is an easy way out that minimizes the political costs associ-
ated with such unpopular choices.³³ In this understanding, national governments
will use GATT as an excuse for unpopular choices.³⁴
A number of authors have taken issues with important features of the com-
mitment theory: Srinivasan (2005), for example, asks why is there need for an
international commitment, if the source of the distortion is domestic? One could
respond that the commitment theory does not pretend that it offers the first best
response to a domestic distortion: it offers one plausible response. Is nevertheless
the response plausible? The validity of this theory largely depends on the cred-
ibility of the commitment. How much of a safe-box is the GATT? Is it true that
once a commitment has been made, there is no way back? No, is the short answer.
First, the GATT contains a number of loopholes (safeguards, antidumping (AD),
etc) which allow for cheap exit from the GATT obligations; second, violations of

with and what products are covered in their negotiations. Goldstein et al (2007) report similar
results.
³⁰ Grandmont and McFadden (1972) have shown that, loosely speaking, global competitive
free market equilibrium is Pareto optimal to the extent that raising the welfare of one single player
will negatively affect the welfare of someone else. Th is theorem relies on several assumptions, some
of which are quite strong and unrealistic. For example, the presence of production externalities
(eg sub-standard production of steel which leads to environmental pollution) casts doubt on the
validity of the theorem.
³¹ See Bagwell and Staiger (2002, pp 32–4).
³² See Irwin (1996a), Maggi (1999) and Maggi and Rodriguez-Clare (1998) who discuss the
time inconsistency issue.
³³ Economic theory distinguishes between Pareto and Kaldor Hicks efficiency. Trade liberal-
ization is Kaldor Hicks efficient, that is, while overall the society wins from market opening, some
of its segments (domestic producers facing foreign competition) might (and often do) lose. Overall,
however, that is, when the losses are measured against the gains of other segments of the society
(users of cheaper inputs, consumers), the society will, in the usual case, benefit.
³⁴ This is the well known time inconsistency issue: to deter pressure from lobbies, a government
might be willing to ‘tie’ its policies against the GATT mast. As we will see, however, in what fol-
lows, the GATT mast is not an absolute safe box, and if at all (at least some) governments retain
discretion to deal with this issue at their discretion. Th is issue is linked to the remedies issue, which
we discuss in Chapter 5. On this score, more generally, see Maggi and Rodriguez-Clare (1998).
14 From GATT 1947 to GATT 1994

the GATT contract can also serve as temporary safeguard, since GATT remedies
are, as a matter of practice, prospective only; third, many domestic legal orders
have adopted a hostile attitude towards WTO law anyway. In light of all of the
above, it is at best doubtful that we can speak of a credible commitment, when
we try to explain the GATT through the lens of the commitment theory. If the
commitment is not credible, why commit in the first place?
Critically, as things stand, commitment theory also suffers from the lack of research
explaining the various GATT instruments in light of commitment theory. This does
not mean, however, that it is a totally unhelpful theory. Commitment theory might
be crucial in explaining specific behaviour: for example, a government might be will-
ing to take a dispute and lose before the GATT, only to turn then to its domestic con-
stituency and explain that its hands are indeed tied, or, that, at the very least, it would
be now quite costly to disregard the GATT ruling. This is what Hudec called ‘use
GATT as an excuse’. One might have difficulty in proving that this is indeed what
happened in particular cases, although one might legitimately speculate along these
lines. In other words, the fact that commitment theory does not explain the GATT
from A to Z, does not mean that its relevance as a tool to help us understand specific
transactions should be dismissed altogether. Commitment theory may be of more
relevance to explain recent accessions to the WTO: the contract covers more policy
space nowadays, that is, a larger number of instruments are being committed.
Regan (2006) has recently revisited this literature, and has forcefully argued that
the very purpose of the GATT (and the WTO) is to counter-attack protectionism.
In this view, absent the GATT, trading nations will have an incentive to continue
protecting domestic producers. His view is not the classic commitment story: in his
story, there is a divide between government and country preferences. Trade agree-
ments aim to bridge this gap and make governments act in accordance with the
country interest. Hence, in his story the game is not between government and the
private sector. There is however, an important overlap: both in the commitment
and the protectionism stories, the role for trade agreement is essentially to address
a domestic problem and not an international externality. The problem with this
explanation for the GATT is that it remains unclear why governments would nego-
tiate such an agreement in the first place, unless their true aim was to maximize
social welfare (in which case, the country-government divide does not exist).
The terms of trade theory³⁵ differs from the two mentioned above, in that it
traces the rationale for trade agreements (and hence, the GATT) not in domestic
distortions but in international externalities.³⁶ This theory has recently found its

³⁵ Bagwell (2006) traces this theory to the work of early and mid-nineteenth century econo-
mists such as Mill and Torrens. Harry Johnson (1953–4) is credited with an elegant formalization
of the theory.
³⁶ As we saw, there is some historical support for this view. Drache (2003) further reports a
meeting between Churchill and Roosevelt in 1941 in Newfoundland, where the two leaders com-
monly held the view that international beggar thy neighbour policies must be addressed in the
post-Second World War era within the framework of a (trade) institution.
 Why the GATT? 15

advocate in the scholarship of Bagwell and Staiger (2002). Bagwell and Staiger
(2002) use terms of trade theory to interpret the various GATT institutions.
Their starting point is that it is simply not the case that unilateral trade liberaliza-
tion is the first best instrument for all: countries that can influence the terms of
trade, have a strong incentive to set tariffs and influence the terms under which
particular goods are being traded on the world stage to their advantage: optimal
tariff theory, from which they borrow in this respect, suggests that governments
behaving in accordance with their self-interest should not necessarily unilaterally
abolish tariffs.³⁷ Hence, tariffs can and do occur for good reasons. Unilateral
setting of tariffs however, can lead to terms of trade externalities: a country
setting tariffs will choose its tariff rates by calculating the welfare implications
to its domestic producers and consumers. It will also be imposing an externality
on foreign producers of the product or commodity affected by the tariffs, which
in all likelihood will not be internalized.³⁸ Consequently, unilateral tariff-setting
can spiral into a Prisoner’s Dilemma type³⁹ of situation, where countries behave
in a non-cooperative manner and impose externalities on each other. In other
words, in this vein, absent a trade agreement, different countries will not behave
cooperatively.⁴⁰ A trade agreement is the means to internalize these externalities
and accordingly, a means to escape from the Prisoner’s Dilemma. Through inter-
national negotiations, trading partners will aim at reducing tariffs to the polit-
ically optimum-level (that is, at the level where, tariff imposition does not result
in external effects for foreign producers). In this vein, reciprocity is the driving
force. The terms of trade theory rests on only two key assumptions: ceteris pari-
bus, a government preference is such that it suffers a welfare loss when its terms of
trade deteriorate; second, assuming a country is large, an increase in an import

³⁷ In doing that, they take issue with Krugman (1991), who summarizes GATT-think in three
sentences: (a) exports are good; (b) imports are bad; (c) other things equal, an equal increase in
imports and exports is good. The last step is obviously assumed by reciprocity-driven international
negotiations. On the other hand, in Krugman’s explanation, a government needs to be mercantilist
to behave in the manner he describes. Th is is not necessarily so: a politically motivated government,
and probably a government that wants to affect terms of trade to its own advantage, could behave in
a similar manner. It thus follows that Krugman (1991) does not believe that there is enough empir-
ical evidence to support optimal tariff-theory.
³⁸ Absent extraordinary circumstances, the country setting the tariff level will have little
incentive to internalize this externality. In Bagwell and Staiger’s terminology, such will not be a
politically optimum tariff, that is, a tariff which takes adequately into consideration the interests of
domestic producers and consumers, but imposes no external effect on foreign producers. One can
of course take the issue with the terminology here: since the (negative) effect on foreign producers
travels through the price mechanism, one should not speak of an externality. This is, however, pure
semantics.
³⁹ Fudenberg and Tirole (1991) provide an explanation of the Prisoner’s Dilemma and how one
gets out of it by changing the pay-offs, pp 9–10 and 110–12. See also, Axelrod (1984).
⁴⁰ A prima facie similar view is taken in Krugman (1991). However, whereas Krugman argues
that free trade is the best unilateral policy, Bagwell and Staiger pay attention to optimal tariff
theory, and view reciprocal negotiations as the necessary mechanism to liberalize trade. Krugman
accepts the optimal tariff-argument as a matter of theory, but dismisses its practical importance
(less so, in his later writings).
16 From GATT 1947 to GATT 1994

tariff results in terms of trade gain for the importing country,⁴¹ and terms of
trade loss for the exporting country.⁴² Both assumptions are of course, quite rea-
sonable. It is important to note that, under these circumstances at least, there
is in many cases, a one-to-one relationship between what Bagwell and Staiger
call terms of trade externalities and what trade negotiators call market access. By
imposing an externality, tariff-setting countries are restricting foreign producers’
access to their own market. Reciprocal tariff negotiations take place to intern-
alize such externalities, that is, to guarantee market access. As demonstrated by
Bagwell and Staiger (2002), it is important to also emphasize that the terms of
trade theory helps explain key GATT concepts, such as reciprocity and MFN.
Recently, the terms of trade theory has come under critique. Regan (2006)
argues that terms of trade manipulation does not happen simply because there is
market power. In his view, terms of trade manipulation is a sophisticated policy
which should not be assumed light-heartedly. Ethier (2004) has taken issue with
the terms of trade theory as well, albeit for different reasons. His attack is twofold:
first, he has argued that trade agreements like the GATT do not in fact prevent
countries from influencing the terms of trade. For instance, the GATT does not
include disciplines on export taxes. To be precise, the GATT does not include a
provision as elaborate as Art II of the GATT that would be applicable to export
taxes. This does not mean, however, as we shall see in Chapter 2, that export
taxes cannot be negotiated and bound in the context of GATT trade liberalizing

⁴¹ Indeed the classic work of Edgeworth (1894) and Bickerdike (1907) suggest that as long as a
foreign country’s offer curve was not perfectly elastic, a country could gain from a tariff (the offer
curve shows the quantity of one product that an agent will export for each quantity of another
product that it will import). Countries with large import markets will be in a position to affect
terms of trade in a more meaningful manner (than small import markets) and thus will have more
to gain through tariff imposition. Economists have viewed the arguments on optimal tariff with
considerable scepticism: Krugman and Obstfeld (1997) observe that at any rate this argument is
of little practical importance since small countries (which constitute the majority of the WTO
membership) cannot affect terms of trade anyway. Fenstra (2004) further noted that there was
little empirical knowledge about the elasticity of foreign export supply. Empirical evidence came
from the work of Broda et al (2006). In this paper, the authors show that countries that are not
members of the WTO systematically set higher tariff s on goods that are supplied inelastically. Th is
paper supplies considerable empirical proof to the optimal tariff theory. It does not, however, as the
authors themselves acknowledge, address the question whether entering the WTO leads countries
to lower their tariffs in a way that no longer reflects the terms of trade motive.
⁴² By relying on these two assumptions only, as Bagwell and Staiger (2002) have elegantly
shown, the terms of trade theory is robust to a wide range of government preference specifications,
including specifications that allow for political motivations and constraints. The authors examine
the political economy of the whole endeavour in which governments have an enlarged (and prob-
ably more realistic) objective function so as to include political considerations. They also discuss the
‘commitment’ approach, whereby governments, unless they credibly commit to an international
regime the reform of their trade policies, their instruments, might have an incentive to renege on
their promises, and private economic operators, anticipating this behaviour, might not undertake
policies necessary to reform trade policies. As noted above, Srinivasan (2005) is not convinced by
the ‘commitment’ approach, arguing instead that when the source of distortion is domestic a trade
agreement is probably an ill-conceived instrument to address it. On this issue see also the argu-
ments advanced by Downs and Rocke (1997).
 Why the GATT? 17

rounds. As a matter of fact though, there are very few export taxes bound into
national schedules of concessions. Still, his argument, slightly rephrased, in prin-
ciple holds: why do not governments use (unbound) export duties to influence
terms of trade?⁴³ The answer in part probably lies in the fact that it is not an
easy exercise to design the export taxes that will affect terms of trade; manipula-
tion of terms of trade is easier through import duties. One could, probably, add
that we end up with a similar critique of the terms of trade theory if we replace
export taxes by domestic instruments: an import tariff can be decomposed into
a domestic tax on consumers and a subsidy on producers. In principle, domes-
tic taxes must be non-discriminatory (across domestic and foreign products).
Thus, it would be quite costly for a GATT contracting party to affect terms of
trade using a domestic rather than a trade instrument. Nevertheless, the original
GATT subsidy rules were quite user-friendly: the subsidizing country might have
had to consult and, eventually, reduce the amount of subsidization, if it affected
important interests of its trading partners.⁴⁴ There was however, no obligation to
abolish subsidization.⁴⁵ Ethier has a second point: in his view, the terms of trade
theory does not explain why small countries, which cannot affect terms of trade,
participate in trade agreements. Bagwell and Staiger (2002) respond that devel-
oping countries would be worse off outside the GATT, since they would not be
in position to profit from non-discriminatory access to most world markets. Th is
much is true. As Ethier, however, points out this is a different (than that offered
by the terms of trade theory) rationale for joining in a trade agreement.
The terms of trade explanation for trade agreements, its intellectual appeal
notwithstanding, probably confounds two separate issues: the manner in which
externalities spill-over when trade policies are unilaterally set, and the ration-
ale for trade agreements. Undeniably, there are terms of trade externalities when
policies are unilaterally set (there are other externalities as well, a point to which
we will come back shortly). This does not mean however that the terms of trade
externalities exhaust the rationale for trade agreements, because trade agreements
are not the only possible reaction to terms of trade externalities. Why would a
powerful country facing terms of trade externalities rush to an agreement and not
respond in the same manner, that is, through unilateral behaviour? A unilateral
response might even, on occasion, be more effective. Some additional elements

⁴³ Interestingly, empirically speaking it is often the case that most of a country’s terms of trade
market power is on the export side, eg oil and Organization of Petroleum Exporting Countries
(OPEC), copper and Chile, phosphates and Morocco, Russia/South Africa and diamonds, etc.
⁴⁴ Note that in the European Community context, a scheme whereby the proceedings from
domestic taxation are used to finance domestic producers is outright illegal. There is standing
case-law to this effect, see, for example, Case 77–72 Carmine Capolongo v Azienda Agricole Maya
[1973] ECR 611; Case 77–76 Flli Cucchi v Avez SpA [1977] ECR 987; Case C-266/91 Celulose Beira
Industrial SA v Fazenda Pública [1993] ECR I-4337; Case C-347/95 Fazenda pública v União das
Cooperativas Abastecedoras de Leite de Lisboa, UCRL (UCAL) [1997] ECR I-4911.
⁴⁵ This analysis also shows the limits of the bindingess of tariff bindings, at least during the
original GATT years, before the Uruguay round disciplines on subsidies show the light of day.
18 From GATT 1947 to GATT 1994

are probably needed to explain why trading partners agree to be party to a trade
agreement instead of reacting unilaterally. At the same time, one should not neg-
lect that the terms of trade theory explains why countries which can affect terms
of trade in particular product-markets sign the GATT.
The terms of trade theory, on the other hand, has a net advantage when it
comes to explaining some GATT institutions: the work of Bagwell and Staiger,
especially, has demonstrated that it can be used to explain the design of the basic
GATT instruments, that is, MFN, reciprocity, subsidies, and safeguards.⁴⁶ It is
difficult, however, to explain, as Regan (2006) points out, other GATT institu-
tions, such as AD.
The end result is that economic theory has not, as of yet, come up with a com-
prehensive explanation for the GATT, as we know it. This does not mean that we
should disregard it, far from it. For a start, the terms of trade theory explains in
a satisfactory manner the heart of the tariff bargain and the various legal instru-
ments committed to this endeavour. It might not be very helpful in explaining
some other GATT legal instruments, such as AD for example, but this does not
mean that we should not be using it where appropriate. Moreover, as we dis-
cussed supra, it might have underpinnings in the negotiation of the GATT as
well. Economic theory, more generally, has been particularly helpful in helping
us understand the gains from trade liberalization: we know that welfare gains
are greater multilaterally than unilaterally, and that the multilateral system helps
invigorate domestic export lobbies to add political weight to trade liberalization.
What we still lack is an internally consistent theory that we can use as guidance
to understand and interpret all of the GATT instruments.

2.4 And some unexplored avenues

The discussion above leaves us with more questions than answers. Where else
then should one look for a rationale? One possible rationale for trade agreements
could lie in the sheer amount of information per se (and its administration) that is
required for a well-functioning agreement. The complexity of the subject-matter
does not facilitate collusive behaviour;⁴⁷ tacit cooperation is facilitated when the
subject-matter of cooperation is simple, such as a cartel agreeing on prices or out-
put restrictions.⁴⁸ This is not the case in trade relations, where anything can, in
principle at least, affect trade. Some formal agreement is needed in the first place,
because tacit cooperation is not a viable alternative: countries will be at a loss as
to what has been agreed with respect to which of the dozens of instruments that
can affect trade. They will further need an agreement in order to explain to their

⁴⁶ We will be reviewing this literature in the various chapters of this book.

⁴⁷ Endless discussions and at least two unfinished papers with Henrik Horn on this score are
acknowledged.
⁴⁸ See on this score Schelling (1960).
 Why the GATT? 19

economic operators what is permitted and what is not.⁴⁹ It is the subject-matter

of the GATT itself, in other words, that argues in favour of a formal agreement.
At the heart of the terms of trade theory lies the realization that trade agree-
ments will be signed because of international externalities rather than domes-
tic problems. Such externalities do not necessarily have to affect terms of trade;
it could be the case, for example, that a small country is facing, for whatever
reasons, high duties in markets of its interest, and is requesting accession to the
GATT in order to benefit from the lower MFN duties.⁵⁰ Such countries do not
affect terms of trade, yet there is a good argument why they would rather be
inside and not outside the GATT club: they want to avoid being the victims of
unilateral behaviour, to the extent that such behaviour has international spill-
over. This last condition (international spill-over) will be fairly easy to meet in
a world which, in Friedman’s (2005) expression becomes increasingly ‘flat’. The
constraint that multilateral rules impose on unilateral behaviour is what prob-
ably motivates their request for participation. The argument here is that:
(1) any unilateral behaviour can, in principle at least, affect trade, and
(2) there is an interest to constrain unilateralism; but
(3) at the same time, tacit agreement on all behaviour affecting trade is difficult
to administer.
This is probably the reason why trade agreements like the GATT, come about in
the first place.
On the other hand, more classic international relations-type of explanations
should not be discarded, since they could offer additional (contributing) expla-
nations for the GATT. It would seem, for example, that yet another reason why
the GATT was created was as a response to Soviet expansionism (Eichengreen,
2006). Such explanations insist on the link between the negotiation of the GATT
and the wider Bretton-Woods institutions umbrella.⁵¹ It could also be that the
rationale is not the same across WTO members. Some countries simply cannot
threaten in a credible manner. For the small, vulnerable economies participation
in a trade agreement is, as discussed above, a necessity, otherwise, they will be
facing high, volatile tariffs (among all sorts of impediments to trade). Powerful
countries, on the other hand, might be using the offer for cooperation on trade
issues (coupled with the explicit promise not to react unilaterally) as a carrot
in order to extract cooperation from their trading partners in other fields. It is
important to stress the seemingly obvious point that trade is but a (small) sub-set
of international relations. Countries interact on various fronts and cooperative

⁴⁹ This point was made to me orally by André Sapir.

⁵⁰ In private discussions with Henrik Horn, the conclusion has been that one should underscore
the unilateral character of the spill-over in the terms of trade-theory, rather than the outcome (terms
of trade manipulation) itself.
⁵¹ See Irwin (1996a and 1998b).
20 From GATT 1947 to GATT 1994

reputation could be credibly built by being cooperative in a field where one can
aff ord to be non-cooperative.⁵²

2.5 To conclude (?)

The more plausible rationale is that the GATT is probably there in order to con-
strain unilateralism. The various GATT instruments are there to guarantee that
the multiple negative effects stemming from unilateral definition of trade policies
will be addressed. This is the approach that I will be following when discussing
the various instruments committed to the GATT in the following chapters of
this book.

3 The GATT in the GATT era: An Agreement,

and an Institution as Well
We can return to the historical account of the GATT. Following the unwilling-
ness⁵³ of the US Congress to ratify the ITO Charter on 6 December 1950, the
GATT was left to operate in an institutional vacuum devoid of a de jure institu-
tional foundation. The GATT contains only a few, scattered provisions of insti-
tutional nature. Nonetheless, it evolved into a de facto international institution

⁵² Economic theory tends to disregard such extra-GATT explanations for a number of reasons.
It is difficult to assess the impact of this disregard.
⁵³ The Congress never got beyond early hearings. The Truman administration withdrew the
ITO when it was clear that there was not going to be Congressional support for this project. On the
account of Miller (2000), both Brown (1950), and Wilcox (1949), two prominent US negotiators,
published their written version of the GATT/ITO negotiations in an effort to persuade the US
public to support the multilateral endeavour. Their efforts were not, alas, crowned with success.
Commentators do not agree on the reason why the US Congress failed to ratify the ITO Charter.
Gardner (1980, pp 373ff ) argues that protectionist forces in the United States were at least a con-
tributing factor: many lobbies requested that President Truman draw back from asking the Senate
to ratify. The wide and probably too ambitious coverage of the Havana Charter (including a full
range of economic issues ranging from commodity agreements to economic development and even
employment) and the threat it represented to national sovereignty was probably the other import-
ant contributing factor. On this score, see also Dam (2004). Diebold (1952, pp 20ff ) offers a dif-
ferent explanation: the US Council of the International Chamber of Commerce, a powerful lobby
by any standard, opposed the ITO because it did not go far enough: many policies of economic
nationalism were left intact. Odell and Eichengreen (1998) offer support for the thesis that the rea-
son why the ITO project failed has to do with the failure of the Truman administration to mobilize
public support on its behalf. At any rate, the ITO could not have realistically entered into force
without the participation of the United States, the leading trading nation at that time. Finally,
Irwin (1995) takes the view that, independently of the reasons leading to it, failure was probably a
blessing in light of the many arcane provisions in the ITO Charter. What is definitely true is that
the early GATT experience did not do the (eventual) ratification of the ITO any favours. Leddy
(1958) reports that, in the eyes of the Congress, the GATT ‘was not repulsive, but neither was
she fair’.
 The GATT in the GATT era 21

in a practical attempt to fill the gap that had been left by the non-establishment
of the ITO.
The GATT integration process benefited and suffered from the lack of an
institutional foundation. On the negative side, the countries participating in
the GATT could not plan far ahead; any institutional innovation was more of a
practical response to an observed need, rather than a springboard which would
accommodate future challenges. On the positive side, however, this ‘functional
institutionalism’ á la GATT gained legitimacy precisely because the edifice
was built on perceived and not imaginary needs. The fact that all decisions in
the GATT years were taken by consensus added to the legitimacy of the whole
endeavour: in the absence of explicit opposition proposals would be voted on
with tacit acquiescence sufficing for decisions to be adopted.⁵⁴
Over the years, the GATT developed an institutional infrastructure that
would adequately cope with its ever-increasing needs. Although the participants
in the GATT process were formally called ‘Contracting Parties’,⁵⁵ they behaved
as members of an institution, operating under a vague ‘institutional’ umbrella.
A number of now well-known GATT institutions came into being through
a series of decisions reflecting the consensus of the GATT Contracting Parties.
Dispute settlement provides a most appropriate illustration of this practice: the
two GATT provisions which explicitly refer to the settlement of disputes (ie Arts
XXII and XXIII) contain no specific procedures, let alone a reference to GATT
panels.⁵⁶ Over the years, procedures for dispute settlement evolved as GATT
Contracting Parties adopted a practice of submitting disputes to ad hoc panels of
experts which would be the adjudicators of the dispute. These experts would then
submit the report (ie GATT-speak for the judgment) to the CONTRACTING
PARTIES—the ultimate decision-making body in the GATT composed of all
those countries signatories to the GATT—which was responsible for adopting
the report submitted. Some formality was added at the conclusion of the Tokyo
round with the adoption of the Understanding on Notification, Consultation,
Dispute Settlement and Surveillance of 28 November 1979. This understanding
included an annex setting out an Agreed Description of the Customary Practice
of the GATT in the Field of Dispute Settlement, which provided a description of

⁵⁴ A number of authors lamented the GATT pragmatic process. Dam (1970) and Hudec (1975)
have been notable exceptions in this discussion.
⁵⁵ The highest decision-making body in the GATT is the CONTRACTING PARTIES (in
capital letters), which refers to the entire GATT membership adopting decisions by consensus.
In many ways, the term ‘CONTRACTING PARTIES’ is appropriate since it was governments
(and customs territories) which signed on to a contract liberalizing trade (the GATT), rather than
adhering to become members of an institution. Evidently, the latter would have been the case had
the ITO been established.
⁵⁶ An early decision by the CONTRACTING PARTIES held that consultations under Art
XXII: 1 would be considered as fulfi lling the consultation requirements of Art XXIII:1. See GATT
Doc BISD 9S/20.
22 From GATT 1947 to GATT 1994

the way in which dispute settlement procedures had evolved since the inception
of the GATT:
Panels set up their own working procedures. The practice for the panels has been to hold
two or three formal meetings with the parties concerned. The panel invited the parties
to present their views either in writing and/or orally in the presence of each other. The
panel can question both parties on any matter which it considers relevant to the dispute.
Panels have also heard the views of any contracting party having a substantial interest
in the matter, which is not directly party to the dispute, but which has expressed in the
Council a desire to present its views. Written memoranda submitted to the panel have
been considered confidential, but are made available to the parties to the dispute. Panels
often consult with and seek information from any relevant source they deem appropriate
and they sometimes consult experts to obtain their technical opinion on certain aspects
of the matter. Panels may seek advice or assistance from the secretariat in its capacity as
guardian of the General Agreement, especially on historical or procedural aspects. The
secretariat provides the secretary and technical services for panels.⁵⁷
Three years later, acting at a Ministerial Conference,⁵⁸ the CONTRACTING
PARTIES reaffirmed the 1979 Understanding and added more detail, including
a requirement that:
The contracting party to which such a recommendation [ie to bring a challenged measure
into conformity with GATT] has been addressed, shall report within a reasonable speci-
fied period on action taken or on its reasons for not implementing the recommendation
or ruling by the CONTRACTING PARTIES.⁵⁹
Further minor steps were taken in a Decision on Dispute Settlement Procedures
on 30 November 1984.⁶⁰ Although panels had not been initially included in
the original text of the GATT, over time they became ‘institutionalized’ in the
GATT sense as an institutional response to the perceived need to better admin-
ister the settlement of disputes. Eventually the Montreal Rules were adopted⁶¹
during the Uruguay round, and the panel process began to take its current form.
Panels became institutionalized in 1995 by virtue of their explicit addition in the
text of the DSU, annexed to the Marrakesh Agreement Establishing the WTO.
This illustration evidences, in quite characteristic manner, how one of the most
important GATT institutions came to life: the institutional design reflects a learn-
ing by doing approach which, importantly, had the backing of the total membership.

⁵⁷ See the Agreed Description of the Customary Practice of the GATT in the Field of Dispute
Settlement (Art XXIII:2), GATT Doc BISD 26S/215, 217, ¶ 6 (iv).
⁵⁸ Ministerial conferences were convened periodically throughout the GATT years to inaug-
urate and conclude a trade round. Sometimes they were also convened to unblock stalled negotia-
tions. In practice ministers almost never conducted the day-to-day negotiations themselves.
⁵⁹ See GATT Doc BISD 29S/13, 15, ¶ (viii).
⁶⁰ See GATT Doc BISD 31S/9.
⁶¹ Essentially, the right to establish a panel and the move towards compulsory third-party
adjudication was introduced through these procedures along with other modifications of less
importance.
 The GATT Rounds of Trade Liberalization 23

In this manner, not only the eventual design corresponded to actual needs, but it
also enjoyed wide legitimacy. The positive spill-over is that such experiences most
likely increased the confidence of the trading nations to continue the institutional-
ization of the GATT and, eventually, to decide the advent of the WTO.

4 The GATT Rounds of Trade Liberalization

The GATT provides for specific disciplines on both trade and domestic instru-
ments affecting the flow of trade between nations. As we will see in more detail
in Chapter 2, the GATT’s recipe for liberalization is to try to narrow down the
options for protecting domestic producers to tariffs only. The GATT aims to
discipline tariff protection by:
(1) reducing tariff volatility, and
(2) reducing the general level of tariffs altogether.
The former is reflected in the obligation (Art II of the GATT) not to impose duties
above the negotiated levels. The latter is the outcome of a series of multilateral
rounds of negotiations initially focusing on the reduction of tariffs, and, later, on
the regulation of non-tariff barriers.⁶² During the GATT years, the Contracting
Parties conducted eight rounds of multilateral negotiations (see Table 1.1):

Table 1.1 Negotiating Rounds

Name of the Round Chronology Number of Participants

Geneva 1947 23
Annecy 1949 13
Torquay 1950 38
Geneva 1956 26
Dillon 1960–61 26
Kennedy 1962–67 62
Tokyo 1973–79 102
Uruguay 1986–94 123⁶³

The Doha round which was launched in that Middle Eastern city in the fall
of 2001 is the ninth trade round since the inception of the GATT and the fi rst
since the advent of the WTO. Following the accession of Tonga in July 2007, 151
countries participate in the negotiations.

⁶² By this, we do not mean that domestic production can be protected only through tariff s and
quantitative restrictions (QRs). Ostensibly, the GATT founding fathers were quite aware of the
fact that instruments seemingly unrelated to trade can indeed affect trade flows. We will examine
the disciplines imposed on trade affecting instruments in Chapter 2.
⁶³ Source: Understanding the WTO, The WTO: Geneva, p 15.
24 From GATT 1947 to GATT 1994

5 Expanding the GATT: The Emergence of the Codes

Up until the Kennedy round, trade negotiators were essentially preoccupied

with the reduction of tariff barriers. The Kennedy round began the shift towards
examining the role that non-tariff barriers (NTBs)⁶⁴ play in restricting the
cross-border trade in goods. Originally, negotiations on NTBs focused on barri-
ers imposed for economic reasons (AD, countervailing, safeguards); eventually,
however (first through the negotiation of the Agreement on Technical Barriers to
Trade (TBT) during the Tokyo round, and then through the renegotiation of the
TBT as well as the negotiation of the Agreement on the application of Sanitary
and Phyto-sanitary measures (SPS) during the Uruguay round), trading nations
began to negotiate NTBs adopted ostensibly for public policy reasons (such as the
protection of human health). In principle, these NTBs were unconnected to the
competitive position of domestic industries. The agreements concluded during
the Tokyo round dealing with NTBs were the following:

(1) Agreement on Implementation of Article VI of the General Agreement on

Tariffs and Trade (Antidumping Code (AD));
(2) Agreement on Interpretation and Application of Articles VI, XVI and XXIII
of the General Agreement on Tariffs and Trade (Subsidies Code (SCM));
(3) Agreement on Import Licensing Procedures (ILA);
(4) Agreement on Trade in Civil Aircraft;
(5) Agreement on Technical Barriers to Trade (TBT);
(6) Agreement on Implementation of Article VII of the General Agreement on
Tariffs and Trade and Protocol to the Agreement (Customs Valuation Code
(CV));
(7) Agreement on Government Procurement (GPA).

Unlike the single undertaking approach, that reigned during the Uruguay round
for negotiations of the multilateral trade agreements (whereby, with few excep-
tions, WTO members were requested to adhere to all agreements concluded),
the Tokyo round codes, sometimes referred to ‘GATT à la carte’, allowed GATT
contracting parties the option of deciding whether or not to participate as a sig-
natory to the various codes. The WTO members added, through the successful
conclusion of the Uruguay round, new agreements, like the SPS, and a series of

⁶⁴ In Robert Baldwin’s (1970) formidable expression, trade negotiations resemble a tide: when
the water flows back (with the reduction of tariffs), trading partners could better appreciate the
other barriers to trade (NTBs).
 The GATT in the WTO era 25

understandings clarifying GATT provisions.⁶⁵ The agreements mentioned above

and the SPS are now annexed to the WTO Agreement.⁶⁶

6 The GATT in the WTO era: Just an Agreement,

Albeit a Pivotal One
The WTO, product of successful negotiations during the Uruguay round, pro-
vided the new institutional arrangement for trade liberalization with its advent
on 1 January 1995. Negotiators agreed to provide for an institutional umbrella
that would administer all agreements concluded during the round. Art II of the
Agreement Establishing the WTO (WTO Agreement) states:
Article II
Scope of the WTO
1. The WTO shall provide the common institutional framework for the conduct of trade
relations among its Members in matters related to the agreements and associated legal
instruments included in the Annexes to this Agreement.
2. The agreements and associated legal instruments included in Annexes 1, 2 and 3
(hereinafter referred to as “Multilateral Trade Agreements”) are integral parts of this
Agreement, binding on all Members.
3. The agreements and associated legal instruments included in Annex 4 (hereinafter
referred to as “Plurilateral Trade Agreements”) are also part of this Agreement for
those Members that have accepted them, and are binding on those Members. The
Plurilateral Trade Agreements do not create either obligations or rights for Members
that have not accepted them.
4. The General Agreement on Tariffs and Trade 1994 as specified in Annex 1A (here-
inafter referred to as “GATT 1994”) is legally distinct from the General Agreement
on Tariffs and Trade, dated 30 October 1947, annexed to the Final Act Adopted at
the Conclusion of the Second Session of the Preparatory Committee of the United
Nations Conference on Trade and Employment, as subsequently rectified, amended
or modified (hereinafter referred to as “GATT 1947”).
The establishment of the WTO relegated the GATT to its original design: a trade
agreement under the institutional umbrella of a trade organization. The WTO

⁶⁵ We will be referring to these understandings whenever appropriate in this book. The Uruguay
round also distinguished between multilateral agreements (where participation is compulsory
for all WTO members) and plurilateral agreements (where participation is optional). The Civil
Aircraft and the Government Procurement agreements, both annexes to the WTO Agreement, are
the only remaining plurilateral agreements.
⁶⁶ According to standard case-law, annexes are integral part of the agreement. By virtue of lex
specialis, a WTO adjudicating body faced with an issue coming under the disciplines of both the
GATT and one of the annex 1A agreements, will first review the consistency of the transaction at
hand under the annex. The GATT remains legally relevant to the extent that an issue has not been
specifically addressed in an annex.
26 From GATT 1947 to GATT 1994

in other words, takes over from the GATT institution and provides the neces-
sary institutional coverage for the GATT agreement.⁶⁷ The establishment of the
WTO also marks the end of the waiting period for the advent of the ITO.
The GATT agreement is no longer the same agreement that was signed back
in 1947. It is a new agreement with a substantially modified content though
the negotiators during the Uruguay round agreed to add to the original text all
adopted decisions by the GATT CONTRACTING PARTIES since 1947. The
new agreement has been named the GATT 1994 and it comprises the following
elements:
1. The General Agreement on Tariffs and Trade 1994 (“GATT 1994”) shall consist of:
(a) the provisions in the General Agreement on Tariffs and Trade, dated 30 October
1947, annexed to the Final Act Adopted at the Conclusion of the Second Session
of the Preparatory Committee of the United Nations Conference on Trade and
Employment (excluding the Protocol of Provisional Application), as rectified,
amended or modified by the terms of legal instruments which have entered into
force before the date of entry into force of the WTO Agreement;
(b) the provisions of the legal instruments set forth below that have entered into
force under the GATT 1947 before the date of entry into force of the WTO
Agreement:
(i) protocols and certifications relating to tariff concessions;
(ii) protocols of accession (excluding the provisions (a) concerning provi-
sional application and withdrawal of provisional application and (b) pro-
viding that Part II of GATT 1947 shall be applied provisionally to the
fullest extent not inconsistent with legislation existing on the date of the
Protocol);
(iii) decisions on waivers granted under Article XXV of GATT 1947 and still
in force on the date of entry into force of the WTO Agreement⁶⁸;
(iv) other decisions of the CONTRACTING PARTIES to GATT 1947;
(c) the Understandings set forth below:
(i) Understanding on the Interpretation of Article II:1(b) of the General
Agreement on Tariffs and Trade 1994;
(ii) Understanding on the Interpretation of Article XVII of the General
Agreement on Tariffs and Trade 1994;
(iii) Understanding on Balance-of-Payments Provisions of the General
Agreement on Tariffs and Trade 1994;
(iv) Understanding on the Interpretation of Article XXIV of the General
Agreement on Tariffs and Trade 1994;

⁶⁷ See on this score, von Bogdandy (2003).

⁶⁸ The waivers covered by this provision are listed in footnote 7 on pages 11 and 12 in Part II
of document MTN/FA of 15 December 1993 and in MTN/FA/Corr.6 of 21 March 1994. The
Ministerial Conference shall establish at its first session a revised list of waivers covered by this
provision that adds any waivers granted under GATT 1947 after 15 December 1993 and before the
date of entry into force of the WTO Agreement, and deletes the waivers which will have expired
by that time.
 The GATT in the WTO era 27
(v) Understanding in Respect of Waivers of Obligations under the General
Agreement on Tariffs and Trade 1994;
(vi) Understanding on the Interpretation of Article XXVIII of the General
Agreement on Tariffs and Trade 1994; and
(d) the Marrakesh Protocol to GATT 1994.
The term ‘other decisions’ featuring in Art 1(b)(iv) of the GATT 1994 is by no
means a model of clarity. The panel in its report on Japan—Alcoholic Beverages
II was called upon to address the issue of whether GATT panel reports that
had been adopted by way of a decision of the GATT CONTRACTING
PARTIES are in fact ‘decisions’ of the CONTRACTING PARTIES to GATT
1947 within the meaning of Art 1(b)(iv) of the GATT 1994. The panel was of
the view that adopted panel reports had the status of any other decision of the
CONTRACTING PARTIES. Consequently, the panel held that adopted panel
reports form an integral part of GATT 1994 as they are ‘other decisions of the
Contracting Parties to GATT 1947 within the meaning of Article 1(b)(iv) of
GATT 1994’ (§ 6.10). The AB disagreed with the panel and held that the ‘deci-
sion’ to adopt a panel report is not a ‘decision’ within the meaning of Art 1(b)
(iv) of the GATT 1994, although it did acknowledge that adopted reports are
‘an important part of the GATT acquis’ (p 15). The term ‘GATT acquis’ is a cre-
ation of the AB, which only clarified the meaning of this concept subsequently in
US—Shrimp (Art 21.5—Malaysia). We quote from §§ 108–9:
‘In this respect, we note that in our Report in Japan—Taxes on Alcoholic Beverages, we
stated that:
Adopted Panel Reports are an important part of the GATT acquis. They are often con-
sidered by subsequent panels. They create legitimate expectations among WTO Members,
and, therefore, should be taken into account where they are relevant to any dispute.
This reasoning applies to adopted Appellate Body Reports as well. Thus, in taking into
account the reasoning in an adopted Appellate Body Report—a Report, moreover, that was
directly relevant to the Panel’s disposition of the issues before it—the Panel did not err. The
Panel was correct in using our findings as a tool for its own reasoning. Further, we see no
indication that, in doing so, the Panel limited itself merely to examining the new measure
from the perspective of the recommendations and rulings of the DSB. (original emphasis)
Hence, it turns out that what the AB meant by the term acquis was the legitimate
expectations of WTO members to see that relevant prior case-law will duly be
taken into account in future disputes, even though there is no legal obligation to
follow the findings and conclusions of GATT panels. This issue arose again in the
context of the dispute that led to the panel report on US—FSC, where the panel
was of the view that decisions to adopt reports should come under Art XVI of the
WTO Agreement. Such decisions are not binding on subsequent panels as Art
XVI of the WTO Agreement itself provides:
the WTO shall be guided by the decisions, procedures and customary practices followed
by the CONTRACTING PARTIES to GATT 1947.
28 From GATT 1947 to GATT 1994

Consequently, the legal effect of adopted GATT reports is not to bind subsequent
panels dealing with the same issue, but simply to provide ‘guidance’. On appeal,
the AB in its report on US—FSC followed a rather convoluted reasoning even
though it ended up ultimately following the panel’s conclusion (§§ 108–15):
In our Report in Japan—Alcoholic Beverages, we stated that not every decision of the
CONTRACTING PARTIES to the GATT 1947 is an “other decision” within the mean-
ing of paragraph 1(b)(iv) of the language incorporating the GATT 1994 into the WTO
Agreement. In that respect, we disagreed with the view that “adopted panel reports in them-
selves constitute ‘other decisions of the CONTRACTING PARTIES to GATT 1947’ for
the purposes of paragraph 1(b)(iv) of the language of Annex 1A incorporating the GATT
1994 into the WTO Agreement.” (emphasis added) The reason for this conclusion was that
adopted panel reports “are not binding, except with respect to resolving the particular dispute
between the parties to that dispute”. (emphasis added) As we said there, the decision to adopt
a panel report was not intended by the GATT 1947 CONTRACTING PARTIES to “con-
stitute a definitive interpretation of the relevant provisions of GATT 1947.” (emphasis added)
The opening clause of the 1981 Council action states: “The Council adopts these reports
on the understanding that with respect to these cases, and in general . . .”. The 1981 Council
action is, therefore, somewhat equivocal in tenor. On the one hand, it is clear from the
text that the 1981 Council action relates specifically to the Tax Legislation Cases and is
an integral part of the resolution of those disputes. This would suggest that, consistently
with our Report in Japan—Alcoholic Beverages, the Council action is binding only on the
parties to those disputes, and only for the purposes of those disputes.
On the other hand, we note that the opening clause of the 1981 Council action also
prefaces the substance of the statement with the words “in general”. The United States
argues that these words indicate that the 1981 Council action was an “authoritative inter-
pretation” of Article XVI:4 of the GATT 1947 that has “general” application and that,
therefore, bound all the contracting parties. The European Communities counters that
the 1981 Council action formed part of the resolution of the Tax Legislation Cases and
that, in adopting that decision, the GATT 1947 Council was acting in dispute settle-
ment “mode”. The European Communities contends further that disputes are resolved
on the basis of the generally applicable rules that are, first, interpreted “in general” and
then applied to the facts of a specific dispute. It is in this limited sense that the European
Communities contends that the GATT 1947 Council meant the term “in general”.
The remainder of the text of the 1981 Council action embodies the substantive statement
of the GATT 1947 Council on Article XVI:4 of the GATT 1947 and does not, in our view,
shed any additional light on whether that statement bound all the contracting parties or only
the parties to the Tax Legislation Cases. We, therefore, share the Panel’s view that the text of the
1981 Council action alone does not resolve the ambiguity highlighted by the conflicting argu-
ments of the United States and the European Communities. Thus, we consider that the Panel
was correct to examine the circumstances surrounding the 1981 Council action.
When the 1981 Council action was adopted, the Chairman of the GATT 1947
Council stated, inter alia, that “the adoption of these reports together with the under-
standing does not aff ect the rights and obligations of contracting parties under the General
Agreement.” In our view, if the contracting parties had intended to make an authoritative
interpretation of Article XVI:4 of the GATT 1947, binding on all contracting parties,
they would have said so in reasonably recognizable terms. We think it most unlikely that
 The GATT in the WTO era 29
the Chairman would have stated that the action did “not aff ect the rights and obligations
of contracting parties”, if it represented an authoritative interpretation of Article XVI:4 of
the GATT 1947. In our view, an authoritative, and generally binding, interpretation of
Article XVI:4 would, in all probability, have been perceived by the contracting parties
as affecting their rights and obligations and would not, therefore, have been accompan-
ied by such a statement. Thus, we are of the view that the statement of the GATT 1947
Council Chairman is consistent with a reading of the 1981 Council action which views
that action as an integral part of the resolution of the Tax Legislation Cases, binding only
the parties to those disputes.
As the Panel observed, it is also noteworthy that, in the report of the GATT 1947
Council to the CONTRACTING PARTIES on its actions during that year, the 1981
Council action was addressed under the heading “Recourse to Articles XXII and XXIII”.
This tends to support the view that the 1981 Council action was a part of the resolution
of the Tax Legislation Cases and not an authoritative interpretation of Article XVI:4 of the
GATT 1947, binding on all the contracting parties.
In light of these surrounding circumstances, we conclude that the Panel was correct to
find, in paragraph 7.85 of the panel report, that the 1981 Council action is not an “other
decision” under paragraph 1(b)(iv) of the language incorporating the GATT 1994 into
the WTO Agreement, and does not form part of the GATT 1994.
We recognize that, as “decisions” within the meaning of Article XVI:1 of the WTO
Agreement, the adopted panel reports in the Tax Legislation Cases, together with the 1981
Council action, could provide “guidance” to the WTO. The United States believes that
the “guidance” to be drawn from the 1981 Council action, through footnote 59, is that the
FSC measure is not an “export subsidy”. The present dispute involves the interpretation
and application of Article 3.1(a) of the SCM Agreement and the question of whether the
FSC measure involves export subsidies under that provision. In contrast, the 1981 Council
action addresses the interpretation and application of Article XVI:4 of the GATT 1947.
The “guidance” that the 1981 Council action might provide, therefore, depends, in part,
on the relationship between these different provisions. (original emphasis)
Consequently, case-law has relegated the legal relevance of GATT reports to
guidance, without however explaining what this term precisely means. At one end
of the spectrum, one could understand that panels have absolute discretion and
could even refrain from referring to a GATT adopted report to the extent that they
do not deem it provides any guidance. At the other end of the spectrum, panels
should religiously follow prior rulings. None of these two viewpoints is reasonable.
Panels, for good reasons that have been abundantly explained in literature, should
refer to, but should not be obliged to follow prior relevant case-law: in presence of
distinguishing (factual) factors, and/or a new theory relevant to resolving a dispute,
deviations should be not only permissible, but rather welcome.⁶⁹
We should note the absence of the concept of stare decisis in WTO law, and
for that matter, public international law. Previous decisions by GATT panels are

⁶⁹ Advances in industrial organization research for example, could (and did) justify changes
in antitrust analysis on both sides of the Atlantic. In the WTO era, practice has shown many
instances where panels and the AB itself have followed prior rulings by the AB, and some instances
(very few indeed) where they have not. We revert to this discussion in Chapter 5.
30 From GATT 1947 to GATT 1994

usually referred to as support for findings already reached. This trend is observ-
able in WTO practice. Case-law in the GATT/WTO context, has not always
developed in linear fashion. As a result, we often observe cases with, at best, select-
ive references to past case-law.⁷⁰ Assuming the cited sample stays the same, there
is case-law evolution. The cited sample nevertheless, as we will see in the follow-
ing chapters, has not remained the same over the years and this has occasionally
lent to confusion of the status of WTO law.⁷¹
Panels in the WTO era have even cited and followed the reasoning of
unadopted⁷² GATT reports to the extent they found it persuasive in a particular
dispute. Unadopted reports cannot conceivably come under GATT 1994 since
they cannot appropriately fall under any of the categories mentioned. They have
not been accepted as part of the GATT acquis either. Still, the AB report on
Japan—Alcoholic Beverages II held that:
a panel could nevertheless find useful guidance in the reasoning of an un-adopted panel
report that it considered to be relevant. (p16)⁷³
Ultimately, the guidance that a GATT panel report provides to subsequent WTO
panels will depend on its persuasiveness (as appreciated by subsequent panels of
course). One could of course, delve into endless debates about the proper legal
status of prior reports. What matters at the end of the day, is that, although their
legal value is uncertain or imprecise, prior reports will invariably constitute valu-
able guidance for subsequent panels dealing with the same or comparable issue.
They do not, however, form an integral part of GATT 1994.

7 The Modern GATT discipline in a Nutshell

7.1 The GATT in its WTO context

When a country (or customs territory, such as Hong Kong before its accession to China)
accedes to the WTO it eventually undertakes three different sets of obligations:
(1) obligations reflected in the multilateral agreements—that is, obligations
which a WTO member must respect because of its accession to the WTO. In
other words, there is no possible opt-out from this set of obligations;

⁷⁰ There are also cases where panels (and the AB) refer to all prior case-law exhaustively. It is to
be expected with the accumulation of case-law, that adjudicating bodies will refer to leading cases
only which establish a jurisprudential trend.
⁷¹ In an incomplete contract, like the WTO, linear evolution of case-law is quite an important
factor in securing predictability as to transaction costs (legal security).
⁷² In the GATT era, a panel report would be adopted only if the losing party had consented to
its adoption. Conversely, in the WTO era the ‘winner’s’ natural acceptance suffices for the adop-
tion of the report.
⁷³ In this vein, the panel report on US—Lamb looked at both adopted and unadopted GATT
reports to support one of its findings (§ 7.78).
 The Modern GATT discipline in a Nutshell 31

(2) obligations reflected in the plurilateral agreements—that is, obligations

which a WTO member may wish to assume by becoming a signatory to the
particular WTO plurilateral agreement;
(3) ad hoc obligations—that is obligations which bind the acceding country and
govern its legal relationship with the WTO membership. These are sui generis
obligations which are assumed by the acceding WTO member and govern in
a specific manner its legal relations with the incumbent WTO members.
Category (3) is idiosyncratic and its subject-matter depends on the negotiations
between the acceding country and the incumbent WTO Members.⁷⁴ The GATT
is part of the multilateral obligations under (1) above.

7.2 The legal relation of the GATT with the WTO Agreement
The GATT is one of the many multilateral trade agreements that WTO members
signed at the end of the Uruguay round or afterwards at the time of accession to
the WTO. The relationship of the GATT with the WTO Agreement and the
many understandings has been resolved by WTO jurisdrudence rather than by
legislative fiat. Conversely, the relationship of the GATT with the other annexes
has been resolved by legislative means.
The AB has consistently held that the WTO Agreement is one agreement and
that the GATT is one of its annexes. The legal implication is that the GATT
disciplines have to be construed in a manner that takes into account the WTO
Agreement (in the Art 31 of the VCLT sense of the term⁷⁵).
A series of understandings were negotiated during the Uruguay round which
elaborate on specific GATT provisions. The AB, in its report on Turkey—Textiles
showed total deference to the Understanding on Art XXIV of the GATT, and

⁷⁴ For example, during the negotiations leading up to the accession of the People’s Republic
of China (‘China’) to the WTO, a special safeguard mechanism was concluded which contrary
to the terms of the WTO Safeguards Agreement allows WTO members the possibility to impose
discriminatory (ie country-specific) safeguards against China. See on this score, Spadi (2000). For
similar GATT practice, see the accession protocol of Hungary, GATT Doc BSID 20S/3 at § 4.
More generally, such obligations can represent either additional obligations to those assumed by
the WTO membership (‘WTO plus’, like the obligation accepted by China), or even a lessening
of the obligations assumed by incumbents (‘WTO minus’). For an excellent discussion of GATT/
WTO practice on this score, see Charnovitz (2006). For instance, Lithuania acceded to the WTO
on 31 May 2001, but was given until 31 December 2005 to bring its excise taxes on beer and mead
into conformity with Art III of the GATT. See the Report of the Working Party on the Accession
of Lithuania to the WTO, WTO Doc WT/ACC/LTU/52 of 7 November 2000 at § 66. ‘WTO
minus’ obligations hence, have a function comparable to grandfathering of obligations that the
original GATT signatories practiced (see Art I of the GATT). The Uruguay round marked an end
to the grandfathering practice. ‘WTO-minus’ obligations reintroduced this practice from the back
door, so to speak. So far, however, such reintroduction has been agreed for the short run only.
⁷⁵ The VCLT is a treaty itself that sets out the general rule of interpretation of international
treaties for which the AB in its very first report (US—Gasoline) acknowledged that WTO adjudi-
catory bodies are required to observe by virtue of Art 3.2 of the DSU.
32 From GATT 1947 to GATT 1994

applied without any further discussion its clarification of the term ‘general inci-
dence of duties’. It considered that the understanding at hand completed the
information provided in Art XXIV of the GATT, by elaborating on the param-
eters of the term ‘general incidence of duties’. We quote from § 53 of the report:
With respect to “duties”, Article XXIV:5(a) requires that the duties applied by the con-
stituent members of the customs union after the formation of the customs union “shall
not on the whole be higher . . . than the general incidence” of the duties that were applied by
each of the constituent members before the formation of the customs union. Paragraph 2
of the Understanding on Article XXIV requires that the evaluation under Article XXIV:5(a)
of the general incidence of the duties applied before and after the formation of a customs
union “shall . . . be based upon an overall assessment of weighted average tariff rates and
of customs duties collected.” Before the agreement on this Understanding, there were
different views among the GATT Contracting Parties as to whether one should consider,
when applying the test of Article XXIV:5(a), the bound rates of duty or the applied rates of
duty. This issue has been resolved by paragraph 2 of the Understanding on Article XXIV,
which clearly states that the applied rate of duty must be used. (original emphasis)
The AB will thus, defer to the legislative solution advanced in an understanding
to the extent that it regulates an issue left unregulated or obscure in the GATT.⁷⁶
The relationship between the GATT and the other multilateral agreements on
the trade in goods comprising Annex 1A to the WTO Agreement is trickier. The
GATT and 12 other agreements (AD, Agreement on Agriculture (AG), Agreement
on Textiles and Clothing (ATC), CV, ILA, Principal Supplying Interest Agreement
(PSI), Agreement on Rules of Origin (ROO), SCM, Agreement on Safeguards
(SG), SPS, TBT, Agreement on Trade-related Investment Measures (TRIMs)) are
all classified as multilateral trade agreements in goods or Annex 1A agreements.⁷⁷
The General Interpretative Note to Annex 1A establishes the legal relationship
between the GATT on the one hand and the 12 agreements on the other:
In the event of conflict between a provision of the General Agreement on Tariff s and
Trade 1994 and a provision of another agreement in Annex 1A to the Agreement
Establishing the World Trade Organization (referred to in the agreements in Annex 1A
as the ‘WTO Agreement’), the provision of the other agreement shall prevail to the extent
of the conflict.
The term ‘conflict’ is key to understanding the relationship between the GATT
and the other Annex 1A agreements. Earlier case-law suggests that the term
should be interpreted stricto sensu: in the absence of a true conflict (in the sense

⁷⁶ One could cast doubt on this conclusion arguing that it suffers from selection bias.
Nonetheless, it should be expected that the AB follows this approach any time it is called to inter-
pret an understanding. There is nothing special about the understanding on the interpretation of
Art XXIV of the GATT.
⁷⁷ Two of the Tokyo round codes, the GPA and the Civil Aircraft have been transformed into
plurilateral agreements in the WTO era. Seven new agreements have been concluded: AG, ATC,
PSI, ROO, SG, SPS, and TRIMs. The ROO, however, falls short of establishing common (harmo-
nized) rules of origin.
 The Modern GATT discipline in a Nutshell 33

that there is a legal impossibility to respect both provisions simultaneously, that

is, a provision of the GATT, and a provision of another Annex 1A agreement), a
WTO adjudicating body could start its analysis from the review of the GATT,
instead of the provisions of the specialized agreements. An appropriate illustra-
tion of this trend in case-law is offered by the panel report on EC—Bananas
III, where the panel decided to apply cumulatively the disciplines of GATT,
ILA, and TRIMs, starting nonetheless from the provisions of the GATT itself
(§§ 7.158–7.163). The cumulative application of the GATT and Annex 1A
agreements is in many ways the natural by-product of the strict interpretation of
the term ‘conflict’.⁷⁸
Subsequent case-law construes the term ‘conflict’ more in line with the lex
specialis principle: although this principle, as such, is not set out in the VCLT,
WTO adjudicating bodies use it with increasing frequency. Its use can only be
applauded because the lex specialis principle is a neccesary (but not sufficient) step
towards observing the ut regis valeat quaem pereat (effective treaty interpretation)
principle—the cornerstone of the VCLT-edifice.⁷⁹
In § 77 of its report on EC—Asbestos, the AB concluded that a measure which
revealed the characteristics of a technical regulation which simultaneously fell
under both the TBT and the GATT, should have been examined under the TBT
rather than under the GATT. In holding for this proposition, the AB distanced
itself from the panel’s findings in opting for a review of the measure under the
GATT in the absence of a true conflict between the GATT and the TBT dis-
ciplines. The panel report on EC—Sardines went the extra mile, so to speak,

⁷⁸ Unfortunately, this earlier era did not pass without causing problems. The most notable ‘dis-
tortion’ has probably been the interpretation of the relationship between the SG Agreement and Art
XIX of the GATT. SG, Art 2 does not include unforeseen developments among the pre-conditions
for lawful application of safeguards. This criterion is reflected in Art XIX of the GATT. One could
argue that the negotiators willingly omitted references to unforeseen developments when drafting
the SG Agreement, precisely because of the ambiguity surrounding this concept, a point that Sykes
(2003c) has emphasized. Nevertheless, the AB in its report on Argentina—Footwear in § 81, held:
Therefore, the provisions of Article XIX of the GATT 1994 and the provisions of the
Agreement on Safeguards are all provisions of one treaty, the WTO Agreement. They
entered into force as part of that treaty at the same time. They apply equally and
are equally binding on all WTO Members. And, as these provisions relate to the
same thing, namely the application by Members of safeguard measures, the Panel
was correct in saying that “Article XIX of GATT and the Safeguards Agreement
must a fortiori be read as representing an inseparable package of rights and disciplines
which have to be considered in conjunction.” Yet a treaty interpreter must read all
applicable provisions of a treaty in a way that gives meaning to all of them, harmoni-
ously. And, an appropriate reading of this “inseparable package of rights and disci-
plines” must, accordingly, be one that gives meaning to all the relevant provisions of
these two equally binding agreements. (original emphasis)
This case-law led to reintroduction of unforeseen developments among the prerequisites for a
lawful imposition of safeguards, probably contrary to the intentions of the negotiators.
⁷⁹ According to this principle, an interpreter should ensure that at the end of the interpret-
ative exercise all provisions retain their meaning. In other words, the interpreter should avoid that
another provision of the same treaty is reduced to redundancy.
34 From GATT 1947 to GATT 1994

and explicitly clarified why, in its view, the TBT disciplines should be examined
before those of the GATT in the order of analysis for measures which can fall
simultaneously under the GATT and the TBT. We quote from §§ 7.14–7.16:
[. . .] [I]f we were to determine that the EC Regulation is not inconsistent with the provi-
sions of the TBT Agreement invoked by Peru, it requests that we examine its claims in
respect of Article III:4 of the GATT 1994.
In addressing the issue of the order of analysis, we have taken into account earlier con-
siderations of this question. We recall the Appellate Body’s statement in EC—Bananas III
which stated that the panel “should” have applied the Licensing Agreement first because
this agreement deals “specifically, and in detail” with the administration of import
licensing procedures. The Appellate Body noted that if the panel had examined the meas-
ure under the Licensing Agreement first, there would have been no need to address the
alleged inconsistency with Article X:3 of the GATT 1994. The Appellate Body suggests
that where two agreements apply simultaneously, a panel should normally consider the
more specific agreement before the more general agreement.
Arguably, the TBT Agreement deals “specifically, and in detail” with technical reg-
ulations. If the Appellate Body’s statement in EC—Bananas III is a guide, it suggests
that if the EC Regulation is a technical regulation, then the analysis under the TBT
Agreement would precede any examination under the GATT 1994. Moreover, Peru, as
the complaining party, requested that we first examine its claim under Article 2.4 of the
TBT Agreement followed by Article 2.2 if we find that the EC Regulation is consistent
with Article 2.4. And similarly, only if we were to find that the EC Regulation is con-
sistent with Article 2.2 does Peru ask us to consider its claim under Article 2.1. In the
event that we were to find that the EC Regulation is consistent with the TBT Agreement,
Peru requests that we examine its claim under Article III:4 of the GATT 1994. We note
that the European Communities did not contest Peru’s request regarding this sequencing
analysis.
On appeal, the panel’s approach was, implicitly at least, upheld by the AB (§ 195).
Finally, the relationship between two multilateral Annex 1A agreements (other
than the GATT) is sometimes prescribed by legislative means (see for example,
Art 1.5 of the TBT, which clarifies the relationship between the TBT and the SPS
Agreements) and for the rest, it is left to the discretion of the WTO adjudicating
bodies.

7.3 The substantive content of modern GATT

The GATT founding fathers can be credited with a remarkable document that
has withstood the test of time for 60 years. The recipe could be described as
follows: since both trade and domestic instruments can affect trade, the discip-
lining of both sets of instruments is necessary, otherwise any obligations imposed
could be easily circumvented. The disciplining of trade instruments could be
described, in summary form, in this way: WTO members negotiate the level of
tariff s to be imposed on imported products; let us call such tariffs, default tariffs.
 The Modern GATT discipline in a Nutshell 35

Such tariffs serve as ceilings (or bindings, in WTO parlance) and WTO members
cannot impose higher tariffs than what they promised to their trading partners.⁸⁰
Through bindings, one avoids tariff-volatility, and thus, unpredictability as to
transaction costs.⁸¹ Crucially, those WTO members which can affect the terms
of trade remove one important weapon from their arsenal.⁸² Unilateral tariff
reductions are rare in practice, as countries which can affect terms of trade will
usually demand that their counterparts also reduce the level of tariffs on products
of their particular export interest; reciprocity, hence, drives international nego-
tiations. On the other hand, there are many examples of countries which cannot
affect the terms of trade and refuse to bind tariff lines, preferring instead to use
them as negotiating chips in future rounds (of dubious efficiency or simply, as
a means to provide their producers with some breathing space, if need be) even
though they might be de facto applying lower tariffs.⁸³ Irrespective of whether
tariffs have been bound or not, WTO members are required to apply their tariff
protection in a non-discriminatory manner. This is the very essence of the MFN
clause. Importantly, WTO members cannot treat non-WTO members better
than WTO members. MFN is thus the carrot towards entry into the WTO:
acceding countries know that their products will receive, in principle, the best
possible treatment in the markets of the WTO incumbents.⁸⁴ Some WTO mem-
bers (donors) will apply to a sub-set of the WTO membership, the developing
countries, a different set of default tariffs, preferential tariffs. Although there is
nothing exceptional about being a developing country, indeed the majority of the
WTO membership are developing countries, preferential tariffs applied to devel-
oping countries are legally considered to be an exception to MFN default tariffs.
Another form of preferential tariffs constitute a genuine exception to MFN: those
imposed across members of a preferential trade agreement, a customs union or a
free trade area.
In principle, WTO members are also required not to apply import and export
quotas pursuant to Art XI of the GATT. Finally, they have also agreed not to use
export subsidies.⁸⁵ Tariffs, import and export quotas, and export subsidies exhaust
the realm of trade instruments.

⁸⁰ In GATT legalese duties are said to be ‘bound’. It is not the case that all goods of all WTO
members are bound.
⁸¹ WTO members can always raise the tariff protection above the bound level, assuming the
requirements of Art XXVIII of the GATT have been met: in principle, an adequate compensation
must be agreed with the affected parties (see Chapter 2).
⁸² Schelling (1960) has shown how the decision not to use some instruments can facilitate
cooperation among states acting in their own self interest.
⁸³ On the relevance of terms of trade in international negotiations, see the excellent and compre-
hensive analysis by Bagwell and Staiger (2002). Bhagwati (2002) offers very insightful arguments
in favour of relaxed reciprocity, while Irwin (2002) presents a very comprehensive account on the
origins of reciprocity in US trade history.
⁸⁴ And, of course, as the number of WTO members increases, so does the size of the carrot.
⁸⁵ Pursuant to Art 4 of the SCM, an Annex 1A agreement to the WTO Agreement, export sub-
sidies are prohibited. The intellectual legitimacy of the prohibition of export subsidies has been
36 From GATT 1947 to GATT 1994

Consequently, assuming a reasonable understanding of the term ‘protection’,⁸⁶

the GATT sanctions only tariffs as a permissible form of protection. Moreover,
tariffs are transitional to the extent they are continuously reduced and ultimately
perhaps eliminated as a result of trade liberalizing rounds.
WTO members also agree, by virtue of adherence to the GATT, not to dis-
criminate between domestic and foreign products through domestic instru-
ments. Disciplining domestic instruments as well is necessary in order to achieve
trade liberalization. Trade instruments can, by virtue of the equivalence proposi-
tions, be decomposed to domestic instruments: for example, instead of imposing
a tariff, a state can impose a tax on its consumers and provide a subsidy to its
producers. Beggar thy neighbour policies can thus result from the use of domes-
tic instruments. In the absence of commitments on all instruments, commit-
ments on trade instruments can be circumvented through the use of domestic
instruments. The consequential GATT discipline is that, with respect to domes-
tic instruments, WTO members must ensure that they will not use them in a dis-
criminatory manner, that is, in a manner that confers an advantage to domestic
(over foreign) production. Policies (environmental, competition, human rights,
etc) will be unilaterally defined by WTO members and will be applied in a non-
discriminatory manner to all products legally circulating in a given market, that
is, domestic and foreign products that have paid their ticket to entry (tariffs).
Our discussion above also reveals that the GATT is a negative integration
contract: WTO members are essentially free to unilaterally define their policies
which might affect trade. Therefore, they are under no constraint at all by the
GATT to follow a particular antitrust policy, environmental policy, labour pol-
icy, and so forth.
WTO members further negotiate state contingent protection (which could take
the form of tariffs, or of a quantitative restriction): they can increase default tar-
iffs above the ceiling level, assuming an agreed contingency has occurred. WTO
members are permitted to increase the level of (import) tariffs above the bound
level when a domestic investigating authority has found that exporters have been
dumping their particular products into the importing market and the domestic
industry has suffered injury up to the margin of dumping—that is, the diffe-
rence between the dumped and the ‘normal price’ (Art VI of the GATT and the

eloquently put into question by Bagwell and Staiger (2006). They advance the argument that it
is at best counter-intuitive to prohibit subsidies (an instrument arguably expanding trade) in the
context of a trade agreement the object of which is to expand trade. It is important to emphasize
that what is being outlawed is not limited to predatory type export subsidies, but rather any export
subsidy must be immediately withdrawn.
⁸⁶ Economic theory has various definitions to offer, see, for example, Bagwell and Staiger
(2002). There is, however, nothing like an operational definition of protection for which we could
take off the shelf and use it in order to distinguish wheat from chaff. For the purposes of our dis-
cussion here, let us accept that protection equals the willful (intended) bestowal of an advantage
effectively granted to a domestic producer. This definition which captures intent and effect is very
close to the one advanced in Regan (2006).
 The Modern GATT discipline in a Nutshell 37

AD Agreement). In a similar vein, members can increase the level of (import)

tariffs for products subsidized by the exporting member up to the amount of the
subsidy granted to exporters (Art XVI of the GATT and the SCM Agreement).
Additionally, they can either impose higher tariffs or simply import quotas any
time the requirements of Art XIX of the GATT and the Safeguards Agreement
have been met. These are the three more frequently used state contingencies.
There are other state contingencies negotiated in the GATT contract: import
quotas can be imposed whenever a state is facing balance of payments problems
(Arts XII and XVIII of the GATT) and quotas can also been imposed when a
state invokes the security exception (Art XXI of the GATT).
Default protection is being gradually reduced, through multilateral negotiating
rounds, while recourse to contingent protection is being gradually rationalized.
The relationship between default and state contingent protection is not neces-
sarily a rule-exception relationship. The rule-exception relationship in law in gen-
eral, and in WTO law as well, signals a shift in the burden of proof,⁸⁷ and also
circumscribes the evidence (that is, the totality of facts) that the complaining
party should submit to a panel. While the contract itself might prejudge this issue
(that is, whether a state contingency is an exception), this is not necessarily always
the case. In the GATT context, the legislator has indicated a preference by intro-
ducing, for example, the term exception in the title of Arts XX and XXI of the
GATT. It would be quite odd for the WTO judge to disregard such clear legisla-
tive guidance. Assume for example, a WTO member imposes an import quota
on toxic waste.⁸⁸ In that case all the complaining party will have to demonstrate
is that it is prevented from exporting the toxic waste product by the imposition of
this quota in contravention with Art XI of the GATT. The burden of proof will
then shift to the importing member which will have to justify the import quota by
invoking, for example, one of the general exceptions under Art XX of the GATT.
The complainant will in all likelihood rebut the arguments made by the defend-
ant under Art XX of the GATT, but it will not have to invoke it first. Therefore,
legally speaking, Art XX of the GATT is an exception to Art XI of the GATT. A
similar solution will result in case a WTO member wants to avail itself of the pos-
sibility to invoke Art XXI (security exception), Arts XII, and XVIII of the GATT
(balance of payments), or Art XV of the GATT (exchange restrictions): assuming
demonstration that Art XI of the GATT has been violated, the intervening
state will carry the burden of proof to demonstrate that it has fulfilled the require-
ments of the provision it has invoked.
The security exception is a state contingency which does signal a shift in the
burden of proof: assuming a threat to national security, a WTO member can

⁸⁷ The term is employed here as a synonym to the burden of production, (standard of review)
which will depend on the letter of the law or the jurisprudential interpretation of the provision
at hand.
⁸⁸ Assume for the sake of the argument that the case at hand is not a case subject to Art III of the
GATT by virtue of its Interpretative Note (domestic measure enforced at the border).
38 From GATT 1947 to GATT 1994

increase its default protection. Importantly, the complainant will not be asked
to demonstrate why the defendant did not act in conformity with Art XXI of
the GATT; its burden is restricted to a demonstration that, say, Art XI of the
GATT has been violated. This is not the case with all state contingencies: AD,
safeguards, and countervailing, are self-standing provisions: a complaining party
arguing, for example, that AD duties have been wrongfully imposed cannot sim-
ply invoke a violation of Art II or Art XI of the GATT; rather, it must demon-
strate that the requirements of Art VI of the GATT (and the relevant Uruguay
round agreement) have not been met.⁸⁹ That is, these are state contingencies that
do not entail a shift in the burden of proof.
It is difficult to discern the theory of the court when deciding whether a state
contingency is an exception or not (and thus, when allocating the burden of proof.
The allocation of burden of proof is judge-made law in the WTO-context).⁹⁰ One
appropriate criterion could be an inquiry into the question whether the initiative
to act depends on a prior action by another trading partner or not. Following this
route, one would put in AD, safeguards, countervailing, and the security excep-
tion in one basket. But where should one classify, for example, Art XX of the
GATT? Some health externalities giving rise to regulatory intervention might be
trans-boundary, while others could be domestic. Private information per se does
not seem to solve the issue either: one would imagine two different allocations
of burden of proof when dealing with regulation to address health externalities
(Art XX of the GATT), depending on whether the externality is domestic, or
trans-boundary. The solemnity of the process is the likeliest candidate: WTO law
imposes that a public procedure (investigation) is followed prior to the impos-
ition of AD, countervailing and/or safeguards. This is what distinguishes AD,
safeguards and countervailing from other state contingencies: affected traders
have the right to participate in this procedure. Consequently, not only will they
know about the actions of the intervening WTO member, they will also have the
possibility to defend themselves during the process, by virtue of the due process
clauses included therein.

⁸⁹ Assume a complainant facing AD duties argues that the country imposing them has violated
its obligations under Arts I and II of the GATT. The respondent would argue that it has the right to
do so once a dumping contingency has occurred, provided of course that dumping has led to injury.
Requesting from the respondent to demonstrate that it has conformed to all requirements included
in the AD Agreement is tantamount to imposing a disproportionate burden, contrary to due pro-
cess considerations. The defendant will not know what to defend. Practice is in line with the view
expressed here and in antidumping litigation; the complainant will be called to demonstrate why
the defendant has violated one or more of the provisions of the agreement. The same is true for all
other state contingencies discussed here.
⁹⁰ Panels and the AB have not explained the rationale behind their decisions on this score.
 2
Disciplines on Trade Instruments

1 Introduction 42
2 The Treatment of QRs 42
2.1 The legal text 42
2.2 What is a QR? 43
2.2.1 The attribution of practices to governments 45
2.2.2 The measures covered 50
2.2.3 The standard of review: no eff ects test and no intent test 55
2.2.4 The relationship between Art XI and other
GATT provisions 58
2.2.4.1 Art III of the GATT 58
2.2.4.2 Art VI of the GATT 62
2.2.4.3 Art XX of the GATT 63
2.2.4.4 Art XXI of the GATT 63
2.2.4.5 Infant industry protection (Art XVIIIc) 63
2.2.4.6 Arts XII and XVIII of the GATT 63
2.2.4.7 Art XIII of the GATT 63
2.3 Trade in textiles 67
2.4 Discriminatory legal QRs 67
2.5 The non-segmentation nature of Art XI of the GATT 68
3 Tariff Protection in the GATT 68
3.1 The legal text 68
3.2 The basic discipline with respect to tariff protection 70
3.3 Why yes to duties and no to QRs? 71
3.4 The mechanics of binding the duties 72
3.4.1 The Harmonized System (HS) 72
3.4.2 What is the tariff promise? Ordinary customs duties,
other duties and charges 74
3.4.3 The typology of ordinary customs duties 78
3.4.4 Multilateral trade negotiations (rounds) 79
3.4.5 The treatment of export tariff s (taxes) under GATT 84
3.4.6 Certification of schedules 86
3.5 Interpreting tariff commitments 87
3.5.1 The legal relevance of HS 87
3.5.2 Scheduling of concessions must be WTO consistent 88
3.5.3 The (non) impact of legitimate expectations 93
40 Disciplines on Trade Instruments
3.6 Changes in tariff protection after their consolidation 93
3.6.1 Switching between diff erent types of duties 94
3.6.2 Withdrawing concessions from members leaving
the WTO 95
3.6.3 Renegotiating the tariff protection 97
3.6.3.1 INR holders 99
3.6.3.2 PSI countries 100
3.6.3.3 SI countries 102
3.6.3.4 MFN trade is the basis for defining PSIs and SIs 102
3.6.3.5 New products 103
3.6.3.6 The mechanics of the negotiation 103
3.6.4 Rectifications and modifications of schedules 112
3.7 Fees and charges for services rendered 114
3.8 Customs Valuation (CV) 118
3.9 The Agreement on Pre-shipment Inspection (PSIA) 119
3.10 The MFN 120
3.10.1 The MFN discipline in a nutshell 123
3.10.2 The measures coming under the purview of the MFN 124
3.10.2.1 Any advantage. . . 124
3.10.2.2 . . . granted to products originating in any
country . . . 125
3.10.2.3 . . . irrespective whether it discriminates de jure
or de facto . . . 126
3.10.3 . . . must be extended to like products 126
3.10.4 The rules of origin 129
3.10.5 The MFN treatment must be extended immediately
and unconditionally 132
3.10.6 No rebalancing permitted 135
3.10.7 The standard of review: no eff ect and no intent test 136
3.11 Special and differential treatment for developing countries 137
3.11.1 The Enabling clause 137
3.11.2 Some historical features 138
3.11.3 The Enabling clause in the WTO legal order 143
3.11.4 The test for compliance with the Enabling clause 144
3.11.5 Is the candle worth the flame? (criticism of the
Enabling clause) 145
3.12 Preferential Trade Agreements (PTAs) 148
3.12.1 PTAs in the WTO: FTAs and CUs 148
3.12.2 Globalization yes, preferences yes as well 148
3.12.3 What matters when discussing preferential trade? 150
3.12.4 PTAs in the GATT 152
3.12.4.1 No per se inconsistency 152
3.12.4.2 The Art XXIV of the GATT test in a nutshell 153
3.12.4.3 Designed at home, approved (?) in Geneva 153
3.12.4.4 The multilateral track 156
3.12.4.5 The bilateral track 168
 Overview 41
3.12.4.6 Why not litigate more? 171
3.12.5 What is the permissible extent from MFN deviation
when a PTA is formed? 177
4 Export Subsidies 179
4.1 Rigidity required? 179
4.2 From discouraging to outlawing export subsidies 181
4.3 Counteracting prohibited subsidies 184

Overview

In this chapter, we discuss the manner in which the GATT binds trade instru-
ments, that is, instruments affecting either imported or domestic goods, but not
both. The GATT imposes disciplines on import/export QRs, tariffs, and export
subsidies. Import/export QRs are abolished. Historically, nevertheless, this has
not always been the case: for many years, trade in textiles and farm goods stayed
outside the disciplines of the GATT in this respect. While the discipline on
import/export QRs is straightforward, the term import/export QR itself is prone
to varying interpretations. Case-law suggests a wide interpretation, whereby, any
measure which has a QR-effect should come under the purview of the GATT
discipline. This interpretation can lend to very problematic applications of the
discipline. The ILA regulates administration of import/export QRs in a man-
ner that guarantees no extra costs for international trade. Tariffs, on the other
hand, do not have to be abolished; they will be bound in the context of trade
liberalizing rounds, and eventually reduced. There is thus a preference of the
GATT for the use of tariffs over QRs, for reasons having to do with the admin-
istration of the two instruments, as is explained in more detail infra. The use
of tariffs is nevertheless, transitional, since gradual reduction will ultimately
lead to their elimination. Tariffs will be applied on a non-discriminatory basis
(MFN), that is, irrespective of the origin of the good. The rationale for MFN
is to avoid concession erosion. There are two important legal exceptions from
MFN: preferential trade agreements (PTAs), and special and differential treat-
ment for developing countries. The GATT has adopted a very benign attitude
towards the former. Economic theory has traditionally viewed such schemes with
scepticism, because of their potential to create trade diversion (and the ensuing
welfare losses). Economists have more recently insisted on the preference erosion
argument: members of such schemes become the enemies of multilateral liberal-
ization, since MFN liberalization ipso facto leads to erosion of PTA preferences.
It is, nevertheless, hard to establish the counterfactual, that is, whether MFN
liberalization would have occurred in the absence of the PTA option. Developing
42 Disciplines on Trade Instruments

countries have benefited from lower than MFN tariffs for their products. Practice
in this field is very often against the spirit and sometimes against the letter of the
law: donor countries have managed to reduce the scope for preferences. On the
other hand, empirical evidence suggests that developing countries that have not
benefited from preferences have developed faster than those that did. Finally,
the preference erosion argument is quite prevalent in this context as well. Export
subsidies are illegal by virtue of the Uruguay round SCM Agreement. One can
question the wisdom of prohibiting an instrument which does not restrict but
expands trade. In any event, at the end of the day the emerging picture concern-
ing the disciplines imposed on trade instruments suggests that trading nations
can protect domestic producers’ income only through import tariffs. They can
further affect terms of trade through export tariffs (taxes) which are explicitly
excluded from the scope of import/export QRs. Export tariffs can be bound and
even eliminated in the context of trade rounds.

1 Introduction

Section 2 focuses on the prohibition of QRs under the GATT. In section 3 of this
chapter, we discuss tariff protection in the GATT context. In section 4 we will
survey the GATT disciplines on export subsidies.

2 The Treatment of QRs

2.1 The legal text

GATT, Art XI reads:
Article XI
General Elimination of Quantitative Restrictions
1. No prohibitions or restrictions other than duties, taxes or other charges, whether
made effective through quotas, import or export licenses or other measures, shall be
instituted or maintained by any contracting party on the importation of any product
of the territory of any other contracting party or on the exportation or sale for export
of any product destined for the territory of any other contracting party.
2. The provisions of paragraph 1 of this Article shall not extend to the following:
(a) Export prohibitions or restrictions temporarily applied to prevent or relieve crit-
ical shortages of foodstuffs or other products essential to the exporting contracting
party;
(b) Import and export prohibitions or restrictions necessary to the application of
standards or regulations for the classification, grading or marketing of commodities
in international trade;
 The Treatment of QRs 43
(c) Import restrictions on any agricultural or fisheries product, imported in any form,
necessary to the enforcement of governmental measures which operate:
(i) to restrict the quantities of the like domestic product permitted to be mar-
keted or produced, or, if there is no substantial domestic production of the
like product, of a domestic product for which the imported product can be
directly substituted; or
(ii) to remove a temporary surplus of the like domestic product, or, if there is no
substantial domestic production of the like product, of a domestic product
for which the imported product can be directly substituted, by making the
surplus available to certain groups of domestic consumers free of charge or at
prices below the current market level; or
(iii) to restrict the quantities permitted to be produced of any animal product the pro-
duction of which is directly dependent, wholly or mainly, on the imported com-
modity, if the domestic production of that commodity is relatively negligible.
Any contracting party applying restrictions on the importation of any product pursu-
ant to subparagraph (c) of this paragraph shall give public notice of the total quantity or
value of the product permitted to be imported during a specified future period and of any
change in such quantity or value. Moreover, any restrictions applied under (i) above shall
not be such as will reduce the total of imports relative to the total of domestic production,
as compared with the proportion which might reasonably be expected to rule between
the two in the absence of restrictions. In determining this proportion, the contracting
party shall pay due regard to the proportion prevailing during a previous representative
period and to any special factors which may have affected or may be affecting the trade in
the product concerned.

2.2 What is a QR?

The term ‘quantitative restriction’ (QR)¹ is not self-interpreting. The founding
fathers of the GATT clearly, nonetheless, excluded from its coverage ‘duties, taxes
or other charges’. At the same time, the phrase ‘. . . whether made effective through
quotas, imports or export licenses or other measures . . .’ seems to suggest that the
drafters intended to include in this provision not only de jure, but also de facto
QRs. The question then arises where the line should be drawn when the scope of
a provision shifts from de jure to de facto?² In other words, should any measure
which might have an effect of a QR be covered in the prohibition or conversely,
should an intent test be applied so as to prohibit only measures that were intended
to protect domestic producers? Then there is the issue whether only import export
QRs should be banned, or, conversely, whether domestic QRs enforced at the

¹ It is worthwhile mentioning that the term ‘quantitative restriction’ appears only in the head-
ing of Art XI of the GATT, while the terms ‘restriction’ and ‘prohibition’ are employed in the body
of the article. In light of the choice of terms, it seems reasonable to conclude (as case-law has already
done) that the three terms can be used interchangeably. Th is is the manner in which they are being
used in this volume.
² We will face the same problem when we discuss de facto discrimination in the context of the
disciplines imposed on domestic instruments.
44 Disciplines on Trade Instruments

border should also come under the purview of Art XI of the GATT. The word-
ing of Art XI of the GATT suggests that this provision deals with import and
export quotas only, and not with all quotas. Domestic quotas enforced at the border
should come under the purview of Art III of the GATT. This is why, film quotas,
for example, which are regulated in Art IV of the GATT, are considered to be an
exception to Art III, and not to Art XI of the GATT.³ There has been, neverthe-
less, no case so far that confirms this view. The repercussions (of this confirm-
ation) are quite substantial for a number of practices that have QR-effects.⁴
Although some issues still remain unresolved, case-law has provided some
clarifications to fill the gaps left opened by the text of the GATT:
(1) since the GATT is a contract between sovereign states (or customs unions)
only QRs attributable to governments can be challenged—that is, private
practices are, in principle, excluded from coverage.⁵ However, the degree of
intervention necessary to attribute a measure to a government is not specified
in the text of Art XI of the GATT. Case-law has provided some important
clarifications in this respect (section 2.2.1);
(2) the typology of measures covered is not specified in the text of Art XI of the
GATT. Case-law has interpreted the term ‘measure’ broadly to encompass
measures that go beyond just laws, regulations and the like (section 2.2.2);
(3) the law is unclear as to the standard of review applicable by adjudicating bod-
ies in cases coming under the purview of Art XI of the GATT. Earlier case-law
had found that an intent test is unwarranted in this context, and an effects test
is equally inappropriate. More recently, on one occasion, a panel seems to have
distanced itself from this seemingly inflexible interpretation (section 2.2.3);
(4) the relationship between Art XI of the GATT and other GATT provisions
has also been addressed. It is clear that both Arts XX and XXI of the GATT
are exceptions to Art XI of the GATT, whereas Art XI of the GATT is no
exception to Art VI of the GATT. The relationship with Art III of the GATT
has also been examined by adjudicating bodies, but the end result here is more
ambiguous. A similar observation is warranted when discussing the relation-
ship between Art XI of the GATT and Art XIII of the GATT (section 2.2.4).
The jurisprudential evolution briefly mentioned above concerns the interpret-
ation of Art XI.1 of the GATT. The significance of the exceptions mentioned in

³ See the relevant discussion in Chapter 3.

⁴ GATT case-law has yet to establish a criterion on this score. Keep in mind that some domes-
tic quotas (like fi lm quotas) affect both domestic and imported products, whereas some other
domestic quotas (like production quotas) might affect only domestic products. Should both types
of domestic quotas be treated in the same way? We advance some thoughts on this score infra when
we discuss the OPEC.
⁵ For instance, if a country has no antitrust laws and the distributors of automobiles agree not
to distribute foreign cars, such practices would not be covered by the prohibition of Art XI of the
GATT.
 The Treatment of QRs 45

Art XI.2 of the GATT has been severely reduced (if not eliminated altogether for
that matter), as a result of the WTO Agreement on Agriculture.⁶

2.2.1 The attribution of practices to governments

The attribution of a practice to government is a threshold issue, to the extent
that Art XI of the GATT becomes irrelevant if the QR does not emanate from
the authority of government. The GATT panel on Japan—Semi-conductors is the
first dispute that dealt with the conditions under which a practice involving a QR
should be attributed to government. The panel in this dispute was called upon
to address a situation in which the government of Japan, as a result of an agree-
ment it had reached with the United States, adopted a series of measures which
induced Japanese companies producing semi-conductors to raise their prices when
exporting to the European Community (EC) market. As a result of the increase
of prices of Japanese semi-conductors, the European Community suffered a net
welfare loss as there were no domestic producers of the same product during
that period in the mid to late 1980s, and consumers of semi-conductors were
required to pay a higher price for the same product. The European Community
challenged the Japanese incentive-mechanism, arguing that, exports of Japanese
semi-conductors to the EC market were substantially reduced, as a result of the
price increase. In the complainant’s view, the practice was attributable to govern-
ment because Japanese producers would not have raised their export prices (and
as a result, the volume of exports would not have been reduced), in the absence of
the incentive-mechanism provided by the Japanese government. Although Japan
did not deny that it had issued an administrative guidance to the relevant indus-
try, it maintained that these companies were under no legal compulsion to follow
the suggested prices. A description of the measures adopted by the Japanese gov-
ernment is summarized in §§ 112–14 of the report in the following manner:
To begin with, the Panel noted the Japanese Government’s own description of its measures
as provided to the United States in its Position Paper of April 1987, notably that “Japan
exercised administrative guidance to achieve production cutbacks and adopted more
stringent export licensing practises” and that “actions have been taken aimed at reducing
supplies and squeezing out grey market transactions.” It referred also to the measures
taken as “recently-ordered production cut-backs,” and that “the measures (i.e. those relat-
ing to production and export administration) taken by the Japanese Government have as
their exclusive purpose and effect avoiding below cost sales of semi-conductors in third-
country markets.
The Panel further examined the structure and elements of the measures adopted. It
noted that Japanese producers were required to submit detailed information on costs on
a regular basis. It also noted the importance of the statutory requirement for exporters
to supply information on export prices and of the heavy penalties attached for failure to

⁶ We will therefore not address Art XI.2 of the GATT. The WTO AG is discussed in
Chapter 3.
46 Disciplines on Trade Instruments
comply with that requirement. The objective of identification in the monitoring meas-
ures was clear. For instance, in cases where the exporter was not a producer, the origin
of the transaction had to be declared and identified. The Panel noted that this gave the
Japanese Government a comprehensive basis for precise identification of the source of any
below-cost pricing. It also observed that any producer or exporter would have been aware
that the Japanese Government would be in a position to have this information. The pre-
paredness of the Japanese Government to request, and to continue requesting, for below
cost sales to cease was also evident.
The Panel examined the operation of the supply and demand forecasts. It noted that
MITI had instituted regular meetings of the Supply and Demand Forecasts Committee,
involving producers, upon which its forecasts were drawn up. The Panel considered that
the Government of Japan played a decisive role in the entire operation. Indeed it was stated
by Japan that “the Japanese Government, in consideration of large inventories of products,
made an attempt to restore balance in supply and demand.” Thus in the first and second
quarters of 1987, the Government of Japan compiled the supply and demand forecasts” to
get production levels reflective of actual demand.” The Panel recalled the statement quoted
in paragraph 112 above concerning the production cut-backs and the avoidance of below
cost sales of semi-conductors in third country markets. On the basis of these, the Panel
considered that the Government of Japan had intervened to facilitate the reduction of the
production levels of semi-conductors through the operation of the supply and demand fore-
casts. The Panel further considered that if Japanese producers and exporters were subject to
any measure restricting the exportation or sale for export of semi-conductors, they would
have to adjust their production levels accordingly. The Panel therefore considered that the
operation of the supply and demand forecasts had facilitated the reduction of the produc-
tion levels, strengthening the effectiveness of the other measures adopted.
It is important to note that the fines imposed were not for failure to follow sug-
gested prices, but for failure to notify practised prices. Japan did not impose prices
on its producers; rather, it used administrative guidance in order to avoid pricing
below cost.⁷ The adoption of administrative guidance is by no means tantamount
to legal compulsion. The panel report stands for the proposition that providing
incentives to private parties to act in a manner inconsistent with Art XI of the
GATT (or, more appropriately, disincentives to act in manner consistent with Art
XI of the GATT) in certain circumstances suffices for a measure to be attributed
to government. The illegal act was the higher (as compared to the prices practised
before the issuance of administrative guidance) price, and the resulting restric-
tion on exports. Since, in this case the measure was attributed to the Japanese
government, it was deemed to be GATT inconsistent (ie a violation of the obliga-
tion set out in Art XI of the GATT). We quote from §§ 109–11 and 114–17:
In order to determine this, the Panel considered that it needed to be satisfied on two essen-
tial criteria. First, there were reasonable grounds to believe that sufficient incentives or

⁷ It is worth mentioning that pricing above cost ensures no ‘dumping’ will be found in the con-
text of an AD investigation. ‘Dumping’ can occur when the price of the good(s) investigated in the
home market exceeds the price sold in the export market.
 The Treatment of QRs 47
disincentives existed for non-mandatory measures to take effect. Second, the operation of
the measures to restrict export of semi-conductors at prices below company-specific costs
was essentially dependent on Government action or intervention. The Panel considered
each of these two criteria in turn. The Panel considered that if these two criteria were met,
the measures would be operating in a manner equivalent to mandatory requirements
such that the difference between the measures and mandatory requirements was only one
of form and not of substance, and that there could be therefore no doubt that they fell
within the range of measures covered by Article XI.1.
...
The Panel considered that, in the above circumstances, the Japanese Government’s
measures did not need to be legally binding to take effect, as there were reasonable
grounds to believe that there were sufficient incentives or disincentives for Japanese pro-
ducers and exporters to conform. The Panel did not consider that these circumstances
were, of themselves, sufficient to ensure compliance. Indeed, events showed that despite
the existence of the Arrangement, a certain number of Japanese producers and exporters
had pursued their original course of production and sales. What was required to ensure
compliance were additional Government measures.
...
The Panel then considered whether the operation of the measures was essentially
dependent on Government action. The complex of measures was, in the Panel’s view,
so dependent. The period between September 1986 and January 1987 gave an interest-
ing indication of how Japanese firms were disposed to operate where they were subject
to less constraint. It was apparent that they had been prepared to produce and sell up
to a quantity which included what was later termed “false demand” in the context of
the revised supply or demand forecast in February 1987. The Panel considered that the
disposition to produce and sell was what the Government of Japan by its complex of
measures intended to control, by the strengthening of the monitoring measures, lowering
of the minimum export amount requiring an export licence to 50,000 yen, requests to
producers not to export at prices below company-specific costs, and the revisions of the
supply and demand forecasts.
The Panel also considered that the series of statements quoted in paragraph 112 above
were relevant in this context. In addition to these, the Panel noted that Japan had stated
in the proceedings of the Panel that “although monitoring by MITI was limited in scope,
it was still meaningful because MITI represented a neutral and objective figure over-
seeing the entire industry while taking into account costs and prices among competing
companies in Japan. Monitoring also helped to stamp out suspicion among companies
that others were cheating or resorting to dumping.” Japan had further stated that “if the
semi-conductor manufacturers were to pursue their own profits and ignore MITI’s con-
cern, the whole dumping prevention mechanism would collapse,” and that “the adminis-
tration presents (firms) with objective facts and considerations and others that are usually
not obtainable by one firm alone.” The Panel considered that these statements concerning
the way in which the Government exercised its authority were a further confirmation of
the fact that the Government’s involvement was essential to the prevention of sales below
company-specific costs.
All these factors led the Panel to conclude that an administrative structure had been
created by the Government of Japan which operated to exert maximum possible pressure
48 Disciplines on Trade Instruments
on the private sector to cease exporting at prices below company-specific costs. This was
exercised through such measures as repeated direct requests by MITI, combined with the
statutory requirement for exporters to submit information on export prices, the system-
atic monitoring of company and product-specific costs and export prices and the insti-
tution of the supply and demand forecasts mechanism and its utilization in a manner to
directly influence the behaviour of private companies. These measures operated further-
more to facilitate strong peer pressure to comply with requests by MITI and at the same
time to foster a climate of uncertainty as to the circumstances under which their exports
could take place. The Panel considered that the complex of measures exhibited the ration-
ale as well as the essential elements of a formal system of export control. The only dis-
tinction in this case was the absence of formal legally binding obligations in respect of
exportation or sale for export of semi-conductors. However, the Panel concluded that
this amounted to a difference in form rather than substance because the measures were
operated in a manner equivalent to mandatory requirements. The Panel concluded that
the complex of measures constituted a coherent system restricting the sale for export of
monitored semi-conductors at prices below company-specific costs to markets other that
the United States, inconsistent with Article XI:1.⁸
Subsequent GATT and WTO case-law has consistently referred to this ruling,
when deciding whether a particular measure should be attributed to government.
In short, Japan—Semi-conductors has been understood as:
(1) opening the door to the possibility of challenging private practices to the
extent they can be somehow attributed to a government;⁹
(2) finding that incentives suffice to attribute behaviour of private actors to a
government.
Point (2), as practiced in this case, probably casts the net too wide. There is nothing
wrong, as a matter of theoretical proposition, to attribute to government practices by
private parties, when the government has provided private parties with the incentive
to adopt a particular behaviour. However, the panel should have explained better why
the incentive provided by the Japanese government was so strong. The panel should
have compared a situation where Japanese producers charge the recommended higher
prices (and thus probably lose some income) against the situation where they con-
tinue to charge prices as they did before the US/Japan Semi-conductor pact was signed.
Recall, that the economic operators did not get punished for practicing a particular
price, but for failing to notify the practiced prices. At the very least, some information
concerning the percentage of Japanese producers practicing higher prices (compared
to the period before the advent of the US/Japan Semi-conductor pact) would have

⁸ MITI is the acronym for the Japanese Ministry of Trade and Investment, which is now known
as METI (Ministry of Economy, Trade and Industry).
⁹ One can draw a parallel here with the criteria for attribution under the SCM Agreement.
WTO case-law has made it clear that private bank loans to private entities at non-market rates
can still be attributed to government if it can be shown that that such rates were requested by
the government. It is dubious, however, whether providing only incentives to provide non-market
rates should in and of itself suffice to characterize such a scheme a subsidy.
 The Treatment of QRs 49

strengthened the persuasiveness of the panel report. As things stand, the whole report
stands on the dubious proposition that Japanese producers might, as a result of the
measures adopted by the Japanese government, have had the incentive to raise their
prices. Whether they actually did that is an open issue.
Case-law thus shows a tendency to avoid Type II errors (by including transac-
tions that could reasonably have been left outside of the coverage of Art XI of
the GATT).¹⁰ This tendency is reinforced by the unwillingness of the panel in
Japan—Semi-conductors to determine in more conclusive fashion the intensity
of the link between the incentives provided and the trade outcome. The failure
of the panel to make this connection has led to the question as to whether any
incentives suffice for behaviour to be deemed government action that amounts
to a violation of Art XI of the GATT. For example, would a ‘buy national’ cam-
paign be considered a violation of Art XI of the GATT?¹¹ Would a campaign
informing the consumer of the merits of domestic products also be considered as
a violation of Art XI of the GATT? Assuming that it is crystal-clear that Art XI
of the GATT does not cover domestic quotas, an issue to which we will return
later, then such schemes should lie outside its purview. What if, however, the gov-
ernment dissuades importers from importing goods under similar slogans? As it
stands, the Japan—Semi-conductors panel report can be reasonably understood
to support affirmative rather than negative responses to such a question.
Although it did not explicitly distance itself from Japan—Semi-conductors, the
panel report on Argentina—Hides and Leather attempted to limit the scope of
the ruling. In this case, the question before the panel was whether the Argentine
law that allowed for the presence of delegates of the downstream industry (lea-
ther products) at the customs-clearance of hides was inconsistent with Art XI of
the GATT. The European Community argued that Argentine producers of hides
might be unwilling to export their produce to Europe, for fear that they will be
earmarked by the domestic downstream industry which was thus in position to
know of all exporting transactions. We read in § 11.18:
Furthermore, and notwithstanding Argentina’s assertion to the contrary, Resolution
2235 is, in our view, a legally binding governmental measure. It is well-established in

¹⁰ A Type I error is what is referred to as a false positive: in our setting, a GATT panel examines
a particular transaction under Art XI of the GATT and pronounces its inconsistency with that
provision when it should not have done so. A Type II error is a false negative: the same GATT
panel refuses to subject a particular transaction under Art XI of the GATT when it should have
done so. The costs associated with each type of error can vary significantly depending on various
considerations.
¹¹ Buy national campaigns have been declared inconsistent with GATT and domestic legis-
lation. In the United States, in Hawaii v Ho, 41 Hawaii 565 (1957), a US court found that a law,
which obliged all traders selling foreign eggs to use a placard ‘We sell foreign eggs’, to be incon-
sistent with Art III.4 of the GATT (see Chapter 3), see the more detailed discussion in Jackson
et al, (2002, p 525). The negotiating history of the GATT suggests, nevertheless, that there was no
unanimous willingness to outlaw Buy National policies. In fact, the opposite is true: most negotia-
tors believed that such schemes should not be judged GATT inconsistent, see E/PC/T/33, pp 28ff,
and EP/C/T/C.6/55, pp 43ff.
50 Disciplines on Trade Instruments
GATT/WTO jurisprudence that only governmental measures fall within the ambit of
Article XI:1. This said, we recall the statement of the panel in Japan—Measures Affecting
Consumer Photographic Film and Paper to the effect that:
[P]ast GATT cases demonstrate that the fact that an action is taken by private par-
ties does not rule out the possibility that it may be deemed governmental if there is
sufficient governmental involvement with it. It is difficult to establish bright-line
rules in this regard, however. Thus, that possibility will need to be examined on a
case-by-case basis.
Moreover, the panel on Argentina—Hides and Leather went on to explain that a
government measure providing incentives to act in a manner inconsistent with
Art XI of the GATT should not be associated with a measure which does not
eliminate all potentiality for such behaviour. We quote from § 11.19:
. . . we do not think that it follows either from that panel’s statement or from the text or
context of Article XI:1 that Members are under an obligation to exclude any possibility
that governmental measures may enable private parties, directly or indirectly, to restrict
trade, where those measures themselves are not trade-restrictive.
Although no bright lines have been established by this report either, it is safe to
say that a reasonableness criterion has fortunately crept into the panel’s analysis
in this context. This is a later in time report and should, for this reason only, be
given considerable attention. One should not forget on the other hand, that the
AB had no opportunity so far to pronounce on this score.

2.2.2 The measures covered

Case-law has, thus, established, so far, that only a sub-set of all measures having
a QR effect, come under the purview of Art XI of the GATT: those attributed to
government (ie members of the WTO).¹² However, this is not the end of the road:
the set of measures that come under the purview of Art XI of the GATT must
be ‘prohibitions or restrictions other than duties, taxes or other charges, whether
made effective through quotas, import or export licenses or other measures’. This
is hardly a self-interpreting term. It nevertheless does, by negative inference,
inform us that a measure that is not considered a prohibition or a restriction does
not come under the purview of Art XI of the GATT. So the question is how
should we understand the terms ‘restriction’ or ‘prohibition’?
GATT, Art XI does not reflect an exhaustive list of the measures that come under
its purview. It does not even contain an indicative list.¹³ It is clear, as mentioned

¹² It is worth recalling that customs territories—to the extent that they enjoy sovereignty in the
conduct of their trade policy—can also become members of the WTO. A measure should be attrib-
uted to such entities under the same conditions as it is attributed to a government.
¹³ Indicative lists serve two purposes: (1) they help avoid Type II errors to the extent that a judge
will know which law should apply when facing one of the transactions explicitly mentioned in the
list, and (2) on the other hand, they help inform the judge as to the type of other transactions not
explicitly mentioned that would come under the purview of a particular legal provision.
 The Treatment of QRs 51

above, that ‘duties, taxes, or other charges’ are not covered by the blanket prohib-
ition of Art XI.1 of the GATT. Indeed, the opposite would be running counter to
the principle of effective treaty interpretation, as Art II of the GATT provides WTO
members with the institutional possibility to consolidate their tariff lines as a per-
missive form of protection. There is no obligation to abolish duties as reflected in
Art II of the GATT. Of course, by virtue of the equivalence propositions,¹⁴ the same
outcome can result through the use of either import duties or QRs. In other words,
tariffs have their QR equivalent. However, to understand the disciplines on QRs as
extending to also cover tariffs would lead to confusion. It cannot be the case that one
GATT provision (Art II) allows for tarriffs, whereas another (Art XI) outlaws them.
Let us start this discussion by looking at the measures that case-law has
included in the coverage of Art XI of the GATT. The WTO panel report on
India—Autos was called upon to address an Indian measure (the trade balancing
condition, discussed infra) which affected the volume of trade to and from India.
Recalling the earlier Japan—Semi-conductors jurisprudence, the panel inter-
preted the term ‘prohibitions or restrictions’ set out in Art XI of the GATT as
covering any action which results in a QR (either on the export, or on the import
side). We quote from §§ 7.246–7.250:
Article XI:1 refers to restrictions “made effective through quotas, import or export
licenses or other measures.” This formulation, which includes a “broad residual category”
of “other measures”, suggests a broad scope to the types of measures which can be consid-
ered to fall within the meaning of Article XI:1.
The panel recalls in particular the conclusion of the panel in Japan—
Semi-conductors that:
Article XI:1, unlike other provisions of the General Agreement, did not refer to laws or
regulations but more broadly to measures. This wording indicated clearly that any meas-
ure instituted or maintained by a contracting party which restricted the exportation or
sale for export of products was covered by this provision, irrespective of the legal status of
the measure.
The panel thus understood the Japan—Semi-conductors GATT panel report as
elevating substance above form to the extent that any measure that has an effect
which results in a QR is prohibited by Art XI of the GATT. Unsurprisingly, the
same panel subsequently moved on to hold that the term ‘quantitative restriction’
does not only cover cases where a numerical target has been laid down, but any
cases where a de facto QR is the outcome of some governmental activity. With
this in mind, the panel addressed the Indian trade balancing condition: through

¹⁴ Lerner first explained that terms of trade can be affected equivalently through import duties
or export taxes (Lerner theorem). Various equivalence propositions have been added to the Lerner
theorem ever since: one can, for example, achieve the same result by using different trade instru-
ments (a tariff in place of a QR), or even by de-composing a trade instrument, say a tariff, into two
domestic instruments (a production subsidy and a tax on consumption).
52 Disciplines on Trade Instruments

this measure, India was essentially imposing a threshold on the amount of exports
that each manufacturer could expect to make, which in turn would determine
the amount of imports that could be made. This measure, in the panel’s view
amounted to an import restriction (see §§ 7.254, 7.257–7.260, 7.264, 7.268,
7.270–7.272 and 7.276–7.277).
The next question is whether the insistence of WTO adjudicating bodies to
subject any measure which has a QR effect to the disciplines of Art XI of the
GATT also covers domestic quotas, or whether a carve-out, as briefly alluded to
supra, should be introduced in this respect.¹⁵ The GATT text, prima facie sug-
gests that a carve-out to this effect is indeed appropriate.
First, the wording of Art XI of the GATT suggests a rather restrictive under-
standing of its coverage in this respect: it refers to restrictions on the importation/
exportation, and not, for example, in connexion with importation/exportation, as
is the case with Art I of the GATT. If the intention of the parties was to allow for
any quota to come under the purview of Art XI of the GATT, they could have
chosen a more appropriate (to this effect) wording. A literal interpretation of this
provision would suggest that domestic quotas are GATT consistent, since such
quotas are not imposed on the the exportation (or on the sale for exportation).
Second, the only case of domestic quota regulated in the GATT concerns laws
and regulations applying to both domestic and imported goods: Art IV of the
GATT, discussed in more detail in Chapter 3, allows WTO members to impose
screen quotas aimed at favouring the screening of domestic films. GATT, Art IV
is an exception to Art III of the GATT, and not to Art XI of the GATT.
Still, although good arguments could be advanced to this effect, it is ques-
tionable whether the treatment of film quotas in the GATT provides the rule for
treatment of all domestic quotas. The reason is the following: as briefly alluded
to above, some domestic quotas affect both domestic and imported goods, and
some do not. Take the film quotas as an illustration: assume that a WTO mem-
ber imposes a requirement that 50% of all films shown in national movie theatres
are domestic films (however defined). Such a requirement ipso facto regulates and
the number of domestic and the number of imported films that can be shown at
any time. The fact that this requirement has a border effect (since, arguably, less
than otherwise, ceteris paribus, foreign films will now be imported) is immaterial
for the logic of Art IV of the GATT. This remains a domestic measure. The law
(Art IV of the GATT), nevertheless, does not clearly explain whether film quotas
are considered to be domestic quotas because:
(1) they are intrinsically a domestic measure (film quotas in favour of national
films) which might have a border effect; or
(2) they concern both domestic and imported goods.

¹⁵ The response is affirmative in the EC context (Art 28 of the European Community Treaty
(ECT)).
 The Treatment of QRs 53

If the former, then all domestic quotas escape the coverage of Art XI of the
GATT. But if the latter, then the question is raised whether we should treat all
domestic quotas in the same manner irrespective whether they concern domestic
only, or domestic and imported goods as well.
Take a production quota: it might concern both domestic and imported goods
depending on how liberal the investment regime of the country at hand is, and
the relevant rules of origin. Some production quotas, nevertheless, could be con-
cerning only domestic goods. Assume for example, that the OPEC uses domes-
tic technology to extract petrol, and the countries participating therein impose
a ceiling on the quantity of petrol to be extracted. In this case, the production
quota concerns domestic goods only. It is thus a different type of quota than the
film quotas we know from Art IV of the GATT. Should we nonetheless, reserve
to both of them the same treatment?
The answer to this question depends on our overall understanding of the func-
tion of the GATT: does it only outlaw cases where an advantage has been granted
to the domestic product which competes with a foreign like/DCS¹⁶ product? In
this case, for example, a WTO member violates the GATT when it taxes say
domestic cars at 10% and imported cars at 20% (as we will see in detail in Chapter 3).
Or, conversely, should we construe the GATT as outlawing cases where an advan-
tage is provided to the domestic market (as opposed to the domestic competing
good) as well? In this case, a WTO member would be violating the GATT if it
stopped exporting steel necessary for the production of cars: its domestic car mar-
ket would benefit as a result, at the detriment of its international competition. In
other words, do the beggar thy neighbour policies covered by the GATT concern
direct benefits to the domestic production, or do they extend to indirect benefits
as well? And what is the place of Art XI of the GATT, the instrument we are
focusing on in this section, in this discussion?
The GATT imposes disciplines on both inbound trade and outbound trade. As
we briefly saw in Chapter 1, and we will see in more detail in this and the follow-
ing chapters, there is an absolute parallelism on the disciplines imposed: tariffs
(import and domestic) are permitted, whereas quotas (import and export) are
not. The functionality of the various instruments committed to this effect, never-
theless, is not the same. Assume that a government wants to maximize domestic
producers’ welfare: an import quota shields the domestic producer of the com-
petitive product from foreign competition, conferring thus a direct benefit; an
export quota is costly to domestic producers of the non-exported commodity,
but could be providing a benefit to, for example, the downstream industry that
uses this commodity as input to its production, conferring thus an indirect bene-
fit (to the downstream industry). The GATT system, by outlawing export quo-
tas, takes a stance against segmentation of markets and declares its intention to

¹⁶ We explain these terms in Chapter 3.

54 Disciplines on Trade Instruments

outlaw policies which favour the domestic markets, and not only policies that
favour domestic (competitive) products.
Do production quotas favour the domestic market (the downstream produc-
ers)? Not necessarily. It could be the case, for example, that all of the limited
domestic production is sold to foreign processors only. Since export quotas are
not permitted, domestic production quotas do not necessarily segment two markets.
But even if we assume, quod non, that the discipline embedded in Art XI of the
GATT should be extended to cover all domestic quotas as well, various para-
graphs of Art XX of the GATT¹⁷ make it clear that, to the extent that the down-
stream domestic producer has not benefited from the limited output, no violation
of the GATT has occurred:
(1) GATT, Art XX(g) allows for measures relating to the conservation of
exhaustible natural resources, if such measures are made effective in conjunc-
tion with restrictions on domestic production or consumption;
(2) GATT, Art XX(i) allows for measures involving restrictions on exports of
domestic materials, if such measures are part of a stabilization plan, and do
not operate to increase the exports of or the protection afforded to the domes-
tic industry;
(3) GATT, Art XX(j) allows for measures essential to the acquisition of products
in general or local short supply, provided that all contracting parties are enti-
tled to an equitable share of the international supply of that product.
The inescapable conclusion here is that quotas which do not confer an advan-
tage to domestic downstream producers, like a simple production quota of raw
material which can serve as input for the production of goods, are not in violation
of the GATT. Irrespective hence, how wide or narrow one interprets the terms
of Art XI of the GATT, the conclusion with respect to simple domestic quotas
should be that they do not violate the GATT.
Export cartels are an idiosyncratic case. Following the de facto acceptance of
the effects doctrine worldwide, an export cartel will be prosecuted in the mar-
ket that it cartelizes (and not its home market). Anyway, its home market has
little, if any, interest to regulate its activities, since it will not be affected by it.
In this vein, for example, the US Webb Pomerene Act of 1918 exempts (par-
tially) export cartels from antitrust prosecution. Is the United States violating
Art XI of the GATT through this law? The likelier scenario would be that an
export cartel restricts output or increases prices. The question however, would be
whether this QR effect should be attributed to the government. Recall, that fol-
lowing Japan—Semi-conductors, providing incentives is enough for behaviour
to be attributed to a government. Is the US government providing incentives
to its economic operators to cartelize the world market by partially exempting

¹⁷ Th is provision includes a list of exceptions to all GATT provisions. We discuss it in detail in

Chapter 3.
 The Treatment of QRs 55

them from antitrust prosecution?¹⁸ A negative response seems more appropriate.

First, the ultimate decision lies with the economic operators themselves: the US
government is not imposing cartelization. Second, contrary to what is the case in
Japan—Semi-conductors, the US government is not incentivizing its economic
operators so that they adopt a particular behaviour. Third, economic opera-
tors know that they might, anyway, have to face a foreign antitrust authority,
if they decide to cartelize foreign markets. So far, nevertheless, the consistency
of export cartels with the WTO rules has not been discussed before a WTO
adjudicating body.

2.2.3 The standard of review¹⁹: no eff ects test and no intent test

The GATT panel report on US—Superfund has for years been considered to be
the leading case when it comes to responding to the question of what is the stand-
ard of review that should be applied in litigation coming under the purview of
Art XI of the GATT. It stands for the proposition that Art XI of the GATT pro-
tects expectations as to a particular behaviour that WTO members are required to
comply with, and, consequently, there is no need for the successful complainant
to demonstrate trade eff ects: that is, there is no need for the complainant to dem-
onstrate the actual trade damage sustained as a result of the challenged measure.
In this particular case, the complainants were arguing that a US internal tax,
which granted domestic products (petroleum products) slightly better treatment
than the imported like products, was GATT inconsistent. The complainants
maintained that there is no requirement first to demonstrate trade effects for a
violation of Art III or Art XI of the GATT to be established. The United States
disagreed, and argued that effects matter; it further argued that, in light of the
minimal difference in tax treatment, no trade effects could be shown, and con-
sequently, no violation of the GATT could be established. The panel disagreed
with the argument of the United States and upheld the claims of the complaining
parties. It follows that a complainant challenging a measure on the basis of Art XI
of the GATT need not demonstrate any trade effects resulting from the contested
measure in order for its claim to succeed. We quote from § 5.2.2:
The general prohibition of quantitative restrictions under Article XI, which the Panel on
Japanese Measures on Imports of Leather examined, and the national treatment obliga-
tion of Article III, which Canada and the EEC invoked in the present case, have essen-
tially the same rationale, namely to protect expectations of the contracting parties as to
the competitive relationship between their products and those of the other contracting
parties. Both articles are not only to protect current trade but also to create the predictability

¹⁸ On this score, and more generally, on the interplay between trade and competition policies,
see Fox (1997).
¹⁹ Throughout this book, the term, ‘standard ’ of review is used as synonymous with burden of
persuasion: how much evidence the complaining party must adduce in order to absolve its burden
of proof (a term encompassing both burden of production—who must submit the evidence—and
burden of persuasion).
56 Disciplines on Trade Instruments
needed to plan future trade. That objective could not be attained if contracting parties
could not challenge existing legislation mandating actions at variance with the General
Agreement until the administrative acts implementing it had actually been applied to
their trade. Just as the very existence of a regulation providing for a quota, without it
restricting particular imports, has been recognized to constitute a violation of Article
XI:1, the very existence of mandatory legislation providing for an internal tax, without it
being applied to a particular imported product, should be regarded as falling within the
scope of Article III:2, first sentence.
This report has probably been misconstrued over the years. A conservative read-
ing of this report suggests that laws that have not produced any effects, but which
are of mandatory nature (in that they leave no discretion as to the behaviour that
must be followed) are still violating Art XI of the GATT. This, however, does not
mean that laws that have produced trade effects can be successfully challenged
without a demonstration of negative trade effects. Nonetheless, this is precisely
how US—Superfund has been understood in subsequent practice.²⁰ Case-law
also suggests that the complaining party does not need to show (protectionist)
intent either, in order to establish a violation of Art XI of the GATT. Although
GATT/WTO adjudicating bodies did not have to face the question as such, this
conclusion seems appropriate in light of the fact that they have outlawed practices
without inquiring into the intent behind the policy-choice.²¹
More recently, the panel on Argentina—Hides and Leather seems to distance
itself from this line of reasoning. The panel distinguished between de jure and
de facto QRs. With respect to the latter, the panel was of the view that for a
successful legal challenge to be mounted, the complainant must demonstrate a
causal link between the challenged measure and the (reduced) level of imports or
exports, whatever the case may be. Without elaborating further on the instance
of a de jure QR, the panel implicitly articulated that the evidentiary standard
should be lower in such cases (§§ 11.21 and 11.22):
Finally, as to whether Resolution 2235 makes effective a restriction, it should be recalled
that Article XI:1, like Articles I, II and III of the GATT 1994, protects competitive oppor-
tunities of imported products, not trade flows. In order to establish that Resolution 2235
infringes Article XI:1, the European Communities need not prove actual trade effects.
However, it must be borne in mind that Resolution 2235 is alleged by the European
Communities to make effective a de facto rather than a de jure restriction. In such circum-
stances, it is inevitable, as an evidentiary matter, that greater weight attaches to the actual
trade impact of a measure.
Even if it emerges from trade statistics that the level of exports is unusually low, this
does not prove, in and of itself, that that level is attributable, in whole or in part, to

²⁰ For instance, in Japan—Semi-conductors the panel did not examine evidence relating to
reduced exports to the EC market and satisfied itself in reviewing evidence demonstrating that
Japan put in place an administrative guidance which would discourage exports.
²¹ In Japan—Semi-conductors the panel did not address the issue of the intent behind the meas-
ure as discussed above, nor have any other of the reports dealing with Art XI of the GATT.
 The Treatment of QRs 57
the measure alleged to constitute an export restriction. Particularly in the context of an
alleged de facto restriction and where, as here, there are possibly multiple restrictions, it is
necessary for a complaining party to establish a causal link between the contested meas-
ure and the low level of exports. In our view, whatever else it may involve, a demonstra-
tion of causation must consist of a persuasive explanation of precisely how the measure at
issue causes or contributes to the low level of exports. (original emphasis)
Importantly, in a footnote to § 11.22, the panel noted:
The Appellate Body in European Communities—Measures Aff ecting the Importation
of Certain Poultry Products similarly required of the complaining party in that case a
demonstration of a causal relationship between the imposition of an EC licensing pro-
cedure and the alleged trade distortion. See the Appellate Body Report on European
Communities—Measures Aff ecting the Importation of Certain Poultry Products (here-
after “European Communities—Poultry”), adopted on 23 July 1999, WT/DS69/AB/R,
at paras. 126–127. While this interpretation related to a claim under the Agreement on
Import Licensing Procedures, it is not apparent why the logic should be any different in
the case of a claim under Article XI:1 of the GATT 1994. (original emphasis)
As already explained above, the European Community had claimed that the
mere presence of representatives of the Argentinean downstream domestic (lea-
ther goods) industry during customs clearance procedures was sufficient in
establishing a QR, on the grounds that it was in the interest of the domestic
upstream (hides) industry not to reveal their identity as exporters of hides from
the Argentine market, since such revelation might be detrimental to their inter-
ests to sell hides in the domestic market as well: in the EC view, the upstream
industry would have little incentive to continue exporting since it could be facing
retaliatory action by the domestic downstream industry. The panel rejected this
claim and found that the presence of representatives of the domestic industry was
insufficient (or rather, the causal link too remote) for establishing a violation of
Art XI of the GATT. The evidence submitted by the European Community was
judged inadequate to support its claim under the following grounds (§ 7.35):
We agree that it is unusual to have representatives from a downstream consuming indus-
try involved in the Customs process of export clearance. As noted above, it seems to us that
the levels of exports of raw hides from Argentina may be low. The European Communities
has stated the matter to us in the form of a rhetorical question—what other purpose could
these downstream industry representatives have in this government process of export
clearance than restricting exports? However, it is up to the European Communities to
provide evidence sufficient to convince us of that. In this instance, we do not find that
the evidence is sufficient to prove that there is an export restriction made effective by the
mere presence of tanners’ representatives within the meaning of Article XI.
At the moment of writing, it remains to be seen to what extent the standard
of review (ie establishing of causal link between the challenged measure and
the level of exports), as reflected in the panel report on Argentina—Hides and
Leather, will be followed in future cases. The AB has yet to pronounce on this
58 Disciplines on Trade Instruments

issue. It seems, nevertheless, reasonable to conclude that, absent an effects test,

or, at the very least, some form of theoretical construction demonstrating that
restrictive effects are the necessary consequence of the presence of Argentine
representatives at customs clearance, it will be impossible to meet this test.

2.2.4 The relationship between Art XI and other GATT provisions

2.2.4.1 Art III of the GATT
The GATT, as we saw in Chapter 1, disciplines both domestic and border meas-
ures. One should keep in mind, however, that the legal disciplines imposed on
these two sets of measures are not identical. As explained above, Art III of the
GATT deals with domestic measures and imposes a discrimination test, whereby
the complainant (a foreign country—that is, a country other than the country
imposing the measure, let us call it F) has to show that a measure not only restricts
market access for its products, but further, that it does so in a manner more oner-
ous than domestic (D) products. GATT, Art XI on the other hand, deals only with
border measures (QR) and imposes a prohibition—that is, the complainant has to
demonstrate that a measure restricts market access for its products (F), irrespect-
ive of how the trade of D products is regulated in that market. The legal threshold
for establishing a violation under the GATT is thus less demanding for border
measures than it is for domestic measures. This is the natural consequence of the
imperative of the legislator (ie the drafters of the GATT) that, in principle, outlaws
the use of quotas (Art XI of the GATT), but allows for the use of domestic instru-
ments to the extent that they are non-discriminatory (Art III of the GATT).
The relationship between these two provisions is partly addressed in the
Interpretative Note ad Art III of the GATT, which relevantly reads:
Any internal tax or other internal charge, or any law, regulation or requirement of the
kind referred to in paragraph 1 which applies to an imported product and to the like
domestic product and is collected or enforced in the case of the imported product at the
time or point of importation, is nevertheless to be regarded as an internal tax or other
internal charge, or a law, regulation or requirement of the kind referred to in paragraph 1,
and is accordingly subject to the provision of Article III.
Hence, measures which are applicable to both domestic and foreign products,
even if applicable at the border, will continue to be covered by the discipline
included in Art III of the GATT. The Interpretative Note, however, does not
explain which measures precisely should be considered import/export QRs and
which should be considered domestic instruments. The GATT does not decide
which instruments (trade or domestic) WTO members should choose.²² Th is is

²² The GATT, as we will see below, allows for only two discriminatory QRs: those imposed
addressing BoP problems (Art XIV of the GATT) and those imposed to address threats to national
security (Art XXI of the GATT). All other QRs can only be justified through recourse to Art
XX of the GATT. The chapeau of Art XX of the GATT includes a non-discrimination test in
the application of the provisionally justified measure under one of the grounds set forth in the
 The Treatment of QRs 59

left to the discretion of the WTO members. Absent a clear definition of the set of
measures that must apply to D/F and of the set of measures that must apply to F
only (assuming that when doing so we have ipso facto established absence of over-
lap between the two sets of measures), the complainant should logically always
have an incentive to claim that the challenged measure is a border measure when
in doubt. On the flip side, the defendant’s incentive is to couch regulatory inter-
ventions in language that prima facie makes them look like domestic measures in
anticipation of the (eventual) complainant’s line of attack.
The GATT panel in its report on Canada—FIRA suggested that with the
exception of truly unique circumstances, such as state trading companies which
often operate both as importers and distributors, a dividing line must be drawn
between measures covered by Art III of the GATT and by Art XI of the GATT.
In the WTO era this issue was revisited. Although the panel in its report on
India—Autos did not distance itself from the outcome reached in the Canada—
FIRA report, it nevertheless did make the point that a priori simultaneous appli-
cation of both Art III and Art XI of the GATT cannot be discarded.²³
Subsequent WTO case-law has not taken a clear stance in favour of an either/
or approach: ie a measure is either covered by Art III or by Art XI of the GATT.
From a legal perspective, however, a measure should be subject to either Art III or
Art of the XI GATT, but not simultaneously to both. The legal test for a finding
of a violation of Art III of the GATT differs from that of a violation of Art XI of
the GATT. If the same measure can fall under both provisions, then a measure
could be legal and illegal at the same time (in the case of a non-discriminatory
quota). This cannot be the case neither as a matter of logic, nor as a matter of legal
methodology (ut legis valeat quaem paereat).²⁴

sub-paragraphs of Art XX of the GATT. Hence with the exception of QRs for BoP and for national
security grounds, QRs can be justified only if they are non-discriminatory (in the sense that they
apply to both domestic and foreign goods). In other words, other than two categories mentioned,
for a QR to be justified under the GATT, it must not be an import or export quota (in the sense that
it affects trade of either domestic or foreign goods); it must simultaneously affect the commerce of
both domestic and foreign goods. Th is does not mean, however, that the GATT forbids the use of
QRs for reasons other than BoP or national security. WTO members can still go ahead and impose
QRs on such other grounds. They will carry the burden of proof though to demonstrate that their
measure affects trade of both domestic and foreign goods: that it is a QR, but not an import or
export QR.
²³ What the panel had in mind is a situation where different facets of the same measure could
be regarded as border measures and some as internal measures. The level of (dis-)aggregation of
a ‘measure’ is nowhere defined in the GATT. As a result, depending on the level of (dis-) aggre-
gation, it is not unthinkable that various sub-measures belonging to the whole come under dif-
ferent GATT provisions. One should not read too much into a sentence in the panel report on
India—Autos that a priori simultaneous application of both Art III and Art XI of the GATT
cannot be excluded.
²⁴ Th is does not mean that domestic measures do not have a QR effect. Indeed, all regulations
might have such an effect but cannot be the exclusive criterion for distinguishing between transac-
tions falling under Art III, and under Art XI of the GATT.
60 Disciplines on Trade Instruments

One wonders whether it is sensible to ask the (tough) question: what should be
the criterion for distinguishing between these two provisions? It would be diffi-
cult to respond to this question without dipping into Pandora’s box concerning
the competing jurisdiction of national sovereignty and international disciplines.
We can instead ask a more benign question: does the law, as it stands, allow for
abuses? Can WTO members invoke these provisions in a manner that will allow
them to circumvent their obligations? It is important to recall that, at the end of
the day, what matters most is for WTO members not to afford protection to their
domestic production through means other than tariffs.
GATT, Arts III and XI could be viewed as incomplete stipulations, in the
sense that there is no exhaustive list of the instruments envisaged. A WTO adju-
dicating body at a subsequent stage will complete the contract through infor-
mation provided at the dispute settlement-stage. A panel or the AB should
(indeed, must) be guided by the following embedded (in the contract) legislative
foundation:
(1) first, we know that ‘duties, taxes or other charges’ are not covered by Art XI of
the GATT—that is, fiscal measures are not covered by Art XI of the GATT,
even if they have (as they undeniably do) a QR effect. Consequently, Art XI
of the GATT does not extend to cover transactions coming under the pur-
view of Arts II and VIII of the GATT;²⁵
(2) second, we know that there is a legislative preference in favour of a broad
interpretation of Art III of the GATT and a narrow construction of Art XI
of the GATT.²⁶ It follows that not all border measures are covered by Art XI
of the GATT.
(3) third, we also know that there are only two permissible discriminatory QRs:
those imposed to address balance of payments (BoP, Arts XII and XVIII
of the GATT) problems, and those imposed on national security grounds
(Art XXI of the GATT). Whereas the former, by definition, concern foreign
products,²⁷ the latter could concern both domestic and foreign products.²⁸
Empirically speaking, discriminatory QRs are no longer much of an issue today:
BoPs are rarely, if ever, imposed, and WTO members have had very infrequent

²⁵ GATT, Art XI thus aims to ensure that the judge avoids Type I errors (false positives) by
subjecting to Art XI of the GATT transactions that should come under the purview of other
provisions.
²⁶ This is what the Interpretative Note ad Art III of the GATT discussed in more detail infra,
essentially amounts to.
²⁷ To the extent that a reduction in the amount of foreign exchange leaving the country is being
requested. This is, of course, not a highly recommendable instrument since countries facing such a
situation could, for example, devalue their currency and thus provide a ‘boost’ to their exports.
²⁸ Th is would be the case, for example, when the intervening state imposes an embargo on sales
of products of interest to the country threatening its national security.
 The Treatment of QRs 61

recourse to Art XXI of the GATT.²⁹ Hence, what really matters are non-
discriminatory QRs—that is, QRs which will have to be justified by recourse
to Art XX of the GATT.³⁰ Then the question is how soon will the burden of proof
pass on to the defendant? Soon, if the legal basis of the complaint is Art XI of the
GATT, more slowly, if it is Art III of the GATT.³¹
So now we move on to discuss cases where abusive use of Art III of the GATT
has been made: cases where, under the guise of non-discrimination (D/F), a
WTO member is only hitting foreign products (F) through a domestic instru-
ment. The incentive formally to use an internal measure when only foreign prod-
ucts are being concerned by the measure at hand has been addressed above. The
current legislative insurance policy against such abuses is three fold:
(1) first, even if a measure does not hit actual domestic production, it can hit
potential domestic production. A prohibition of sales of construction mater-
ial containing asbestos will dissuade interested parties from investing in this
market;³²
(2) second, the directly competitive or substitutable products (DCS) category³³
in and of itself is a disincentive to have abusive recourse to Art III of the
GATT. WTO members cannot indirectly protect a substitutable product:
a country producing bananas but not strawberries cannot tax the latter with
a prohibitive tax, since it will have to tax the former by a similar tax as well.
Thus, the GATT has armed itself against the classic understanding of protec-
tion (through the use of a trade/domestic instrument a WTO member pro-
vides an advantage to the domestic producer of a substitute);
(3) third, the advent of the TBT and SPS Agreements signals yet another step
towards reducing the risk of abuses. Even in the absence of domestic pro-
duction, WTO members cannot intervene through the imposition of mar-
ket access restrictions of products by measures covered by the TBT/SPS
Agreements, unless they conform with the requirements set forth in those
two agreements. For instance, a country which does not produce genetic-
ally modified organisms (GMOs) cannot make the access to their market
more onerous by imposing an internal tax of say 100% ad valorem (and

²⁹ There are two panel reports issued to this effect, and one more case is reported (a conflict
between Honduras and Nicaragua) which never found its way to WTO adjudication.
³⁰ If not, they will be violating Art XI of the GATT in any event
³¹ Assuming of course that Art XX of the GATT is accepted as a legitimate exception to Art III
of the GATT. Whereas the position taken in this book is that this should not be the case, case-law
has so far moved in the opposite direction.
³² There is of course nothing wrong with using an internal measure when no domestic produc-
tion exists. Internal measures do not have to continuously hit domestic (D) and foreign (F) prod-
ucts. Actually, they might be hitting F only because there is no more D production as a result of the
very measure in place and the possible regulatory diversity on this issue across WTO members.
³³ As we will see in sufficient detail in Chapter 3, the DCS category includes products that are
directly competitive or substitutable, that is, loosely speaking, participating in the same relevant
product market.
62 Disciplines on Trade Instruments

thus finance its social policy), unless it meets the requirements of the SPS
Agreement. Note that in this case, it is simply irrelevant whether the WTO
member concerned produces a competing non-GMO farm product or not.³⁴
In the absence of domestic production of a substitute, the current regime out-
laws beggar thy neighbour policies that extend beyond the classic understand-
ing of protection as described above.
It seems that the existing regime has in large part taken care of the incentive to
cheat and use internal measures (raising thus the burden of proof for potential
complainants)³⁵ when only foreign products are being hit.³⁶ It follows that when
WTO members choose an internal measure, there is a legitimate presumption
that they act in good faith, that is, that they intended a discipline to apply to both
domestic and foreign products.

2.2.4.2 Art VI of the GATT

Recall that, in Japan—Semi-conductors, Japan was found to be in violation of its
GATT obligations, because it provided its economic operators with incentives not to
dump. Japan argued before the panel that it was acting in accordance with its GATT
obligations since, dumping is condemned by Art VI of the GATT. The panel was
thus led to discuss the relationship between Art VI and Art XI of the GATT. It held
that Art VI of the GATT was silent on actions by exporting countries; it was quite
vocal, however, on actions by importing countries, since it allowed them to impose
AD duties in order to counteract dumping. It then went on to conclude that (§ 120):
. . . Article VI did not provide a justification for measures restricting the exportation or
sale for export of a product inconsistently with Article XI:1.

³⁴ Assuming of course that say corn and GMO corn are like or DCS products, a difficult issue in
itself as reflected by WTO jurisprudence.
³⁵ Note that the current allocation of burden of proof is not relatively unfavourable to the
complainant to the extent it does not have to show either (discriminatory) effects or intent to dis-
criminate when challenging the legitimacy of a measure with Art III of the GATT. But even the
introduction of a more stringent, intent-based test would not alter our conclusions here. The three
built-in safeguards discussed above are valid even were such a test to be introduced.
³⁶ Note that the EC regime does not face this problem since Art 28 of the ECT covers both Art
III and Art XI of the GATT. The question we raise here is an issue in EC law with respect to tax-
ation: whereas domestic taxation is legal if non-discriminatory, customs duties and taxes of equiva-
lent effect (TEE) are illegal. Case-law in the EC legal order has made two important contributions
when addressing the issue whether a measure should be considered an internal tax or a TEE: the
form of the law does not matter, hence it is irrelevant if a country calls a TEE an internal imposition;
de facto discriminatory internal impositions (that is, impositions hitting almost exclusively foreign
products) will be considered internal measures if they form part of a general taxation scheme, see
on this score Joined Cases 2 and 3–69 Sociaal Fonds voor de Diamandarbeiders v SA Ch Brachfeld &
Sons and Chougol Diamond Co [1969] ECR 211, Case 78–76 Steinike & Weinling v Federal Republic
of Germany [1977] ECR 595, Arachides (Case 158/82 Commission of the European Communities v
Kingdom of Denmark [1983] ECR 3573), Case C-343/90 Manuel José Lourenço Dias v Director da
Alfândega do Porto [1992] ECR I-4673. Still the ECJ has consistently held that the intention of
the regulator does not matter, see Joined cases C-441/98 and C-442/98 Kapniki Michaïlidis AE v
Idryma Koinonikon Asfaliseon (IKA) [2000] ECR I-7145.
 The Treatment of QRs 63

2.2.4.3 Art XX of the GATT

Assuming that a violation of Art XI of the GATT has been established, the WTO
member concerned can seek justification by invoking one of the grounds men-
tioned in Art XX of the GATT (see Chapter 3).

2.2.4.4 Art XXI of the GATT

WTO members can lawfully impose QRs in order to safeguard their national
security interests (see Chapter 3).

2.2.4.5 Infant industry protection (Art XVIIIc)

GATT, Art XVIII(c) allows developing country WTO members to deviate from
their obligations in order to protect an infant industry. Developing countries
have not recently made use of this provision, although such practices were quite
popular in the sixties and the seventies.³⁷ Theoretically, one could make the case
in favour of protection during the early stages of development: Lucas (1988), for
example, offers a skill acquisition model of endogenous growth and suggests that,
by allowing countries to establish a comparative advantage in the production of
high learning goods, the erection of trade barriers during early stages of devel-
opment may enhance their long-term growth. The manner in which such meas-
ures have been practiced, nevertheless, has been very often counter-productive.
Gradually, at least some developing countries have been abandoning this policy
tool.
The India—Quantitative Restrictions panel³⁸ made it clear that an invocation
of Art XVIII(c) of the GATT is justiciable. At the time of writing, the only such
restriction in place is by Bangladesh.³⁹

2.2.4.6 Arts XII and XVIII of the GATT

The panel and the AB reports on Korea—Various Measures on Beef (discussed
in Chapter 3) confirmed that, by virtue of Arts XII and XVIII of the GATT,
WTO members can lawfully impose QRs in order to safeguard their balance of
payments.

2.2.4.7 Art XIII of the GATT

The title of Art XIII of the GATT is ‘Non discriminatory administration of QRs’.
Logically, obligations regarding the administration of QRs kick in only after the
consistency of a QR with the GATT has first been established. For example, if a
QR has been justified on BoP grounds, it must also, in principle, comply with Art

³⁷ See the discussion on special and differential treatment, infra, in this chapter.
³⁸ This report is discussed in Chapter 4 under ‘3 BoP’.
³⁹ See WTO Doc G/C/7 of 16 January 2002. The United States first (WTO Doc G/C/8 of
18 February 2002) and the EC subsequently (WTO Doc G/C/9 of 20 February 2002) requested
consultations with the government of Bangladesh. No subsequent notification ever occurred.
64 Disciplines on Trade Instruments

XIII of the GATT.⁴⁰ Case-law, however, is rather confusing in this respect, in the
sense that, it has failed to clearly articulate the ‘QR justification-administration’
sequence. The issue arose in the EC—Bananas III dispute concerning the cover-
age of the MFN requirement reflected in Art XIII.1 of the GATT. The European
Community had in place two QRs, one applicable to bananas originating in the
African, Caribbean and Pacific group (ACP)⁴¹ countries, and another originating
in the rest of world. The tariff rate for the former was much lower than that for the
latter, and, as a result, non-ACP producers had suffered an important trade loss.
The European Community further provided preferential treatment for bananas
originating in countries that had signed the so-called Framework Agreement with
it (which the European Community had attached in its schedule). The panel, and
the AB subsequently, upheld the claim advanced by the complainants (to the effect
that the MFN requirement should cover all imports of bananas, the like prod-
uct at hand, § 191 of the AB report). In the same paragraph, the AB held that the
formation of a PTA in conformity with Art XXIV of the GATT can provide a
legitimate exception to Art I of the GATT and consequently, to Art XIII of the
GATT as well. In this case, nonetheless, Art XXIV of the GATT was inapplic-
able; both the panel and the AB held that a one-way preference scheme (like the
Lomé Convention) cannot be lawfully characterized as a PTA. Still, the AB went
on to find that the EC regime for bananas was in violation of Art XIII of the GATT,
even though neither the panel nor the AB made the point that the EC regime should
have been found to be GATT inconsistent on the basis that the restrictions them-
selves were illegal.⁴² But of course, such a finding is unwarranted absent a prior find-
ing that Art XI of the GATT has been violated. In other words, the ambit of Art XIII
of the GATT should not be extended so as to cover issues of substantive consistency
of a QR with the GATT; it should be narrowed down to issues of consistency in the
administration of an otherwise GATT consistent QR.
As stated above, the title of Art XIII of the GATT leaves us with the impres-
sion that, what is required from this provision is that otherwise lawful QRs are
being administered on a non-discriminatory basis. The title of Art XIII of the
GATT is misleading, insofar as Art XIII of the GATT does not operate in an his-
torical vacuum. GATT, Art XIII.2 states that:
. . . contracting parties shall aim at a distribution of trade . . . approaching as closely as
possible the shares that the various contracting parties might be expected to obtain in the
absence of such restrictions.

⁴⁰ See, the excellent analysis in Jackson (1969) pointing to the sequence between Arts XI, XII
and XIII.
⁴¹ The group is composed of numerous countries that are former colonies of some members of
the European Community that had concluded agreements with the European Community that
included preferential trade provisions (eg Lomé Convention and the Cotounou Agreement).
⁴² Th is should have been the outcome since the claim under Art XXIV of the GATT had been
rejected. Hence, the QR was illegal in the first place. Under the circumstances, the panel and the
AB should not even have entertained the question of the administration of the (illegal) QR.
 The Treatment of QRs 65

Pursuant to Art XIII.2 (d) of the GATT, a WTO member lawfully imposing a
QR is required to allocate quotas to various suppliers in a manner that respects
their pre-QR market shares. A reference period (usually the previous 3–5 years)
will be used as benchmark for the calculation of market shares.⁴³ Hence, it is
not non-discriminatory administration of quotas that is privileged through Art
XIII of the GATT; it is respect of historic market shares, a rather discriminatory
policy (since new aggressive suppliers will not be put at equal footing with old
suppliers).
GATT, Art XIII does not mention the period of time that QRs can remain
in place. Ostensibly, the length of time a QR can be imposed depends on the
rationale behind the QR: for instance, if a member experiences BoP problems for
2 years, a QR should logically be in place for the same period of time.⁴⁴
Governments imposing QRs might have the incentive to auction off import
licenses to interested importers. The price to acquire an import license will
depend on the difference between the price of the (unlimitedly sold) domestic
good and that of the (limited) imported good. WTO members are subject to
transparency-obligations as to the administration of import licenses under Art
XIII.3 (a) of the GATT.
During the Tokyo round, the ILA⁴⁵ was concluded which has since been
superseded by the ILA negotiated during the Uruguay round. The objectives of
the ILA are to simplify and make import licensing procedures transparent; to
ensure their fair and equitable application and administration; and to prevent
procedures such as the granting of import licenses from having restrictive or
distorting effects on imports.
Import licensing does not necessarily have to apply only when a quantitative
restriction is in place. In fact, there are two forms of licensing provided for in
the ILA: automatic licensing (Art 2 of the ILA), which occurs in the absence
of a restriction in place; and non-automatic licensing (Art 3 of the ILA), which
takes place when a lawful restriction is in place. A WTO member will usually⁴⁶
have recourse to automatic licensing in order to monitor imports into its market.
Non-automatic licensing is, in practice, the more frequently used type of licens-
ing. Non-automatic licensing can take place irrespective of the type of restriction
in place. The AB in its report on EC—Bananas III, for example, accepted the

⁴³ Consequently, WTO members with anticipated falling market shares can profit from a QR
in their export markets.
⁴⁴ Safeguards, on the other hand, can take the form of a QR (as is often the case) but cannot be
imposed for more than 8 years and once the safeguards are lifted, they cannot be re-imposed on the
same product until an equal period of time has passed during which the safeguards had been previ-
ously applied. See Arts 7.3 and 7.5 of the SG.
⁴⁵ Recall from Chapter 1, that a series of agreements had been concluded during the Tokyo
round, the notorious Tokyo round Codes. We use the two terms (agreements, codes) interchange-
ably when referring to the Tokyo round agreements.
⁴⁶ The use of automatic licensing is infrequent because it usually corresponds to some public
order concern, such as monitoring imports. There are, however, substitute mechanisms available on
the spot (that is, at the customs office).
66 Disciplines on Trade Instruments

legality of imposing a licensing scheme when a tariff quota was in place (§ 193).
In practice, the scenario envisaged in EC—Bananas III is quite typical: import
licensing schemes in the majority of cases coincide with a restriction, be it a QR
or a tariff quota (TRQ). Import restriction measures will normally push the price
of the imported commodity higher than it would otherwise be, and a govern-
ment might have the incentive to auction the quotas in place. A licensing scheme
can serve this purpose.
In its report on EC—Poultry, the AB was of the view that WTO members
must ensure that no distorting effects will be caused either to the trade covered
or not covered by the scheme at hand, in the application of the non-automatic
import licensing schemes (§§ 120ff ). The quintessential obligation here is the one
enshrined in Art 3.2 of the ILA: there should be no trade-restrictive or distort-
ing effects stemming from the application of the licensing scheme. How should
one understand the terms ‘trade-restrictive’ or ‘ distorting eff ects’ on imports? At
one extreme, one could interpret them so as to outlaw any auctioning scheme,
since arguably, imposing a premium on an import license might dissuade the
marginal investor from purchasing one. Such interpretation should, however, be
discarded: the very purpose of auctioning is precisely for the state to make some
money. Indeed, none of the obligations included in Art 3.5 of the ILA (see infra)
outlaws such practices. This provision should rather be interpreted as outlawing
practices that purposefully discourage the use of the quotas.
Additional obligations regarding non-automatic import licensing are set out in
Art 3.5 of the ILA: first, the period of application should not be excessively long;
second, the period of validity of the license should not be unreasonably limited in
time; third, members should not discourage full utilization of quotas; fourth, the
desirability for issuing licenses for products in economic quantities must be taken
into account by the WTO member issuing an import license; fifth, when allo-
cating import licenses, the previous import performance of the petitioner must be
taken into account;⁴⁷ sixth, in the case where a quota is administered through
a license which is not country of origin specific (quotas can be administered
through licenses which are either country of origin specific or not), importers
are free to choose the source of imports; seventh, in case of variations between the
amount designated in the license and the amount actually imported (which can
occur, as Art 1.4 of the ILA itself acknowledges), Art 3.5(l) of the ILA calls for
compensatory adjustments to ensure that trade flows continue unimpeded; finally,
WTO members undertake, in accordance with Art 1.9 of the ILA, to make for-
eign exchange available to petitioners for an import license under the same terms
as importers of goods for which no import licensing scheme is in place.

⁴⁷ It would seem that some importers might want to purchase licenses for strategic reasons and
contribute towards restricting or eliminating market access of imports to their domestic market.
This could be the case, for example, if the importer is also the producer of a substitutable (to the
imported) product.
 The Treatment of QRs 67

2.3 Trade in textiles

Before the entry into force of the Uruguay round agreements, trade in textile prod-
ucts was regulated in the Multi-fibre Agreement (MFA).⁴⁸ This agreement was
signed in 1974 between some developed and 31 developing countries, and essen-
tially imposed a worldwide system of bilateral quotas. The Agreement on Textiles
and Clothing (ATC) was signed along with the other Uruguay round agreements,
and replaced the MFA. WTO members had to notify to the Textiles Monitoring
Body (TMB) their quotas and abolish them within 10 years, that is, by 1 January
2005. ATC, Art 9 makes it plain that the whole agreement was of transitional
value: its purpose was to ensure that all quantitative restrictions would be elimi-
nated and/or replaced by tariffs, as is the case with most other products (with the
exception of farm products, the regime for which is being discussed in Chapter 3).
As a result, the TMB has also ceased to exist as of 1 January 2005.⁴⁹

2.4 Discriminatory legal QRs

As is elaborated on in Chapter 3, QRs can be justified through recourse to one
of the general exceptions set forth in Art XX of the GATT, provided that these
measures do not discriminate across countries where the same conditions prevail;
that is, in principle, for an otherwise prohibited QR to be justified, it must apply
to both domestic and foreign goods, assuming comparable conditions across the
domestic and foreign markets prevail. However, as we have seen, the GATT does
tolerate two types of QRs that apply only to either domestic or foreign goods:
QRs imposed for BoP and national security reasons.
If a quota has been lawfully imposed, WTO members, by virtue of Art XIV
of the GATT, can legitimately deviate from the obligation to apply it in a non-
discriminatory manner, as prescribed in Art XIII of the GATT. The following
two conditions are required:
(1) the quota must have been imposed to address problems relating to BoP;
(2) the discriminatory quota must affect a small part of its trade.
With respect to national security exceptions, Art XXI of the GATT does not
contain a non-discrimination clause similar to that included in the chapeau of Art
XX of the GATT.⁵⁰ Moreover, the examples set out in Art XXI (b)(i) and (ii) of
the GATT are examples of discriminatory measures.

⁴⁸ See the very informative discussion in Trebilcock and Howse (2005, pp 482ff ).
⁴⁹ The TMB was an interesting institution. It was composed of delegates by both exporting and
importing countries and had the power to recommend ways to end disputes brought to its atten-
tion (Arts 8.8–10 of the ATC). It was passage obligé for all WTO members and, the complainant
could submit the dispute to the DSB (see Chapter 5), only in case the TMB recommendations had
not been observed. Conciliation procedures exist in other Annex 1A agreements as well (see, for
example, Art 17 of the AD), they reveal, however, a predominantly bilateral character.
⁵⁰ See Chapter 3.
68 Disciplines on Trade Instruments

2.5 The non-segmentation nature of Art XI of the GATT

Since most QRs must be applied in a non-discriminatory manner (otherwise they
will be illegal), a question could be raised as to the ultimate usefulness of this pro-
vision. In other words, what would the GATT system lose if Art XI of the GATT
did not exist and all measures were subject to the national discrimination prin-
ciple as we know it under Art III of the GATT? One could answer this question
by way of illustration. Assume that country A possesses a car and a tyre indus-
try, and the two are not vertically integrated. Assume further that A imposes
an export restriction on tyres to protect its car industry. As a result, country B’s
car industry sustains a loss, assuming that B does not produce tyres. B cannot
attack such measures under Art III of the GATT since no discrimination can
be shown.⁵¹ GATT, Art V is also a highly inadequate legal basis because it only
covers goods in transit (not the case in our example). For B to mount a successful
legal challenge against such practices, an Art XI-type legal instrument is indeed
required; otherwise it would be impossible successfully to challenge such policies.
GATT, Art XI thus, is also an instrument to counter market segmentation, since
ultimately, the discrimination sanctioned by this provision is that in favour of the
domestic market and not in favour of the restricted product.⁵²

3 Tariff Protection in the GATT

3.1 The legal text

GATT, Art II reads:
Article II
Schedules of Concessions
1. (a) Each contracting party shall accord to the commerce of the other contracting par-
ties treatment no less favourable than that provided for in the appropriate Part of
the appropriate Schedule annexed to this Agreement.
(b) The products described in Part I of the Schedule relating to any contracting party,
which are the products of territories of other contracting parties, shall, on their
importation into the territory to which the Schedule relates, and subject to the
terms, conditions or qualifications set forth in that Schedule, be exempt from
ordinary customs duties in excess of those set forth and provided therein. Such
products shall also be exempt from all other duties or charges of any kind imposed
on or in connection with the importation in excess of those imposed on the date

⁵¹ Th is provision outlaws discrimination across a pair of products (domestic, foreign) that are
substitutes and not complements as in this illustration.
⁵² Recall, nevertheless, that markets can be lawfully segmented through export taxes, since
Art XI of the GATT outlaws quotas but not taxes.
 Tariff Protection in the GATT 69
of this Agreement or those directly and mandatorily required to be imposed there-
after by legislation in force in the importing territory on that date.
(c) The products described in Part II of the Schedule relating to any contracting party
which are the products of territories entitled under Article I to receive preferential
treatment upon importation into the territory to which the Schedule relates shall,
on their importation into such territory, and subject to the terms, conditions or
qualifications set forth in that Schedule, be exempt from ordinary customs duties
in excess of those set forth and provided for in Part II of that Schedule. Such prod-
ucts shall also be exempt from all other duties or charges of any kind imposed on
or in connection with importation in excess of those imposed on the date of this
Agreement or those directly or mandatorily required to be imposed thereafter by
legislation in force in the importing territory on that date. Nothing in this Article
shall prevent any contracting party from maintaining its requirements existing
on the date of this Agreement as to the eligibility of goods for entry at preferential
rates of duty.
2. Nothing in this Article shall prevent any contracting party from imposing at any time
on the importation of any product:
(a) a charge equivalent to an internal tax imposed consistently with the provisions of
paragraph 2 of Article III in respect of the like domestic product or in respect of
an article from which the imported product has been manufactured or produced
in whole or in part;
(b) any anti-dumping or countervailing duty applied consistently with the provisions
of Article VI;
(c) fees or other charges commensurate with the cost of services rendered.
3. No contracting party shall alter its method of determining dutiable value or of con-
verting currencies so as to impair the value of any of the concessions provided for in
the appropriate Schedule annexed to this Agreement.
4. If any contracting party establishes, maintains or authorizes, formally or in effect, a
monopoly of the importation of any product described in the appropriate Schedule
annexed to this Agreement, such monopoly shall not, except as provided for in that
Schedule or as otherwise agreed between the parties which initially negotiated the
concession, operate so as to afford protection on the average in excess of the amount
of protection provided for in that Schedule. The provisions of this paragraph shall not
limit the use by contracting parties of any form of assistance to domestic producers
permitted by other provisions of this Agreement.
5. If any contracting party considers that a product is not receiving from another con-
tracting party the treatment which the first contracting party believes to have been
contemplated by a concession provided for in the appropriate Schedule annexed to
this Agreement, it shall bring the matter directly to the attention of the other contract-
ing party. If the latter agrees that the treatment contemplated was that claimed by the
first contracting party, but declares that such treatment cannot be accorded because
a court or other proper authority has ruled to the effect that the product involved
cannot be classified under the tariff laws of such contracting party so as to permit the
treatment contemplated in this Agreement, the two contracting parties, together with
any other contracting parties substantially interested, shall enter promptly into fur-
ther negotiations with a view to a compensatory adjustment of the matter.
70 Disciplines on Trade Instruments
6. (a) The specific duties and charges included in the Schedules relating to contract-
ing parties members of the International Monetary Fund, and margins of pref-
erence in specific duties and charges maintained by such contracting parties, are
expressed in the appropriate currency at the par value accepted or provisionally
recognized by the Fund at the date of this Agreement. Accordingly, in case this par
value is reduced consistently with the Articles of Agreement of the International
Monetary Fund by more than twenty per centum, such specific duties and charges
and margins of preference may be adjusted to take account of such reduction; pro-
vided that the CONTRACTING PARTIES (i.e., the contracting parties acting
jointly as provided for in Article XXV) concur that such adjustments will not
impair the value of the concessions provided for in the appropriate Schedule or
elsewhere in this Agreement, due account being taken of all factors which may
influence the need for, or urgency of, such adjustments.
(b) Similar provisions shall apply to any contracting party not a member of the Fund,
as from the date on which such contracting party becomes a member of the Fund
or enters into a special exchange agreement in pursuance of Article XV.
7. The Schedules annexed to this Agreement are hereby made an integral part of Part I of
this Agreement.

3.2 The basic discipline with respect to tariff protection

Customs duties can, by virtue of Art II of the GATT, be bound: by binding, we
understand the process whereby a WTO member incurs the obligation not to
levy duties beyond a certain multilaterally agreed level. The extent of bindings
varies from one WTO member to another, and is the outcome of trade negotia-
tions. Once the WTO membership is satisfied with the content of negotiations,⁵³
the WTO Secretariat will circulate the members’ schedules of concessions (or, as
it is sometimes referred to, list of commitments) for verification and certification.
Customs duties do not have to be bound. Irrespective, however, whether they have
been bound or are left unbound, they have to be applied on a non-discriminatory
basis (MFN).
Since customs duties concern only imported products, MFN entails no dis-
crimination among two imported products. Furthermore, a WTO member can-
not, by virtue of the MFN, provide imports from countries that are not members
of the WTO, a more preferential tariff treatment than that accorded to imports

⁵³ It is slightly more complicated in the real world. Assume A, B, and C negotiate the import
duty on wheat that A will be applying. Assume further that B manages to extract from A a promise
of 10% ad valorem. B and A will notify the outcome of their negotiation to the WTO Secretariat.
Their agreement, however, remains confidential until the end of negotiations. Nothing, in the
meantime, stops C from negotiating with A and extracting a 7% promise. C will also notify the
bilateral agreement it reached with A. At the end of the negotiations, A will have to impose, by vir-
tue of the MFN obligation, a 7% import duty on all wheat imported into its market, irrespective
of its origin. The promise of 10% it gave to B, nevertheless, is not void of consequences: as we will
see infra, assuming A wishes, subsequent to the original negotiation, to revise its duty on wheat
upwards, B might be recognized as initial negotiating right (INR) holder (if A wishes to impose say
a 12% import duty, see the discussion on Art XXVIII of the GATT).
 Tariff Protection in the GATT 71

of like products originating in WTO members. On the other hand, if a WTO

member in practice imposes duties below the bound level (which, in GATT par-
lance, are known as applied duties), it incurs the obligation not to discriminate
across products of different WTO origin, with respect to applied duties as well.
To ensure that the net is not cast too wide (so as to eventually eliminate all tariff
protection altogether), the obligation not to discriminate is assumed with respect
to imported like products.
Negotiations on tariff liberalization are based on reciprocal commitments.
Reciprocity is what drives international negotiations. Reciprocity is often referred
to as a GATT legal principle.⁵⁴ It does not, however, amount to a legal principle
exonerating WTO members from observing the contract when another WTO
member has failed to do so: if A violates the contract, B cannot, by invoking non
adimplenti contractus (reciprocity) unilaterally stop honouring its own commit-
ments. B can at most challenge A’s practices before the WTO adjudicating bodies,
and A will, if found guilty, have to implement the rulings. In the meantime B will
have to continue observing the contract. Only in the case when A refuses to imple-
ment the multilateral rulings, B can request authorization to stop honouring its
own commitments, on a transitional basis, until A starts respecting the contract
once again. Reciprocity further provides the rationale for the rights and obligations
of parties when tariffs are being renegotiated, assuming parties cannot reach an
agreement on the compensation to be paid (Art XXVIII of the GATT).
Duties that have been bound cannot be unilaterally revised upwards; the multi-
lateral process embedded in Art XXVIII of the GATT has to be observed. In prin-
ciple, compensation has to be agreed between the WTO member wishing to revise
upwards its bound duties and a sub-set of WTO members particularly affected by the
decision to revise bound duties. The new duties will reflect a new agreed equilibrium
and be applied on an MFN basis. Reciprocity is very much behind the procedure
established in Art XXVIII of the GATT, the idea being that the balance of rights and
obligations established in the original negotiation should be preserved in subsequent
negotiations, irrespective of the agreed level of duties. In sum, there is one key diffe-
rence between a trade liberalizing round (Art XXVIII bis) and the procedure under
Art XXVIII of the GATT: the former leads to a new agreed lower level of duties;
the latter, to an agreed rebalancing of duties, whereby some duties have been revised
upwards, and some downwards. In both scenarios, reciprocity is the key.

3.3 Why yes to duties and no to QRs?

The discussion above shows that the GATT allows tariffs, at least on a transitional
basis, while, as we saw supra, it disallows QRs. Yet, according to the equivalence

⁵⁴ Recall our discussion in Chapter 1, where we highlighted the fact that, during the original
negotiation of the GATT, strict reciprocity was the key negotiating principle, see E/PC/T/33,
Annexure 10.
72 Disciplines on Trade Instruments

propositions, tariffs can be expressed in QR-terms and vice versa. What then jus-
tifies the differential treatment between tariffs and QRs? The answer is twofold:
tariffs are easier to administer, and QRs are more distorting.
Tariffs are set at a level and, assuming that the MFN has been respected,⁵⁵ do not
affect the competitive advantage of nations. A tariff will impose proportionately the
same burden on all exporters, say 10% ad valorem, while those charging the highest
price (who ceteris paribus would not be gaining large market shares anyway) will
be paying the highest mark-up in absolute terms. A QR, however, will have to be
continuously renegotiated in order to ensure that the level at which it is distributed
across exporters takes adequate account of the competitive advantage of each; one
of them; the negotiating costs should not be underestimated. Were one to arbitrar-
ily ease administrative costs, by setting QRs, on a yearly basis, at a fi xed percent-
age of the volume of exports made during the previous year (say 80% of the total
exports), it risks neglecting the competitive advantage of the trading nations; such
a scheme would have to function à la Art XIII of the GATT, where historic quotas
are being preserved. That is, without necessarily respecting comparative advantage.
The bottom line is that, while tariffs respect competitive advantage, QRs do not
necessarily do so, or might do it at a very high (negotiating) cost.
What we have described above does not, however, exhaust the differences by
far. For instance, and of significant importance for many countries, QRs will
need to be allocated somehow across producers, and depending on how this
is done, one might end up with allocations that might further impair trade.⁵⁶
Quota allocation is also likely to give rise to various political economy problems.
All these distortions are avoided when tariffs are being used.

3.4 The mechanics of binding the duties

3.4.1 The Harmonized System (HS)
To bind duties, WTO negotiators first had to agree on a common language to
describe goods, and, based on this common language, exchange concessions.⁵⁷ There
is a strong transaction-costs argument in favour of harmonizing the description of
goods: the same good is not necessarily expressed in identical terms in two different
jurisdictions, yet the same concession is relevant for all WTO members. Clarity as to
what precisely is being negotiated is the quintessence of any negotiation.⁵⁸

⁵⁵ As we will see later in this book, this is a generous assumption. A large percentage of today’s
trade is not conducted on an MFN basis. Yet this observation does not put into question the sound-
ness of the remarks made here.
⁵⁶ Th is would be the case, for example, if the importer who has been allocated a set quantity is
also a producer of a competing good.
⁵⁷ The HS classification is of course relevant for unbound duties as well.
⁵⁸ Unless, of course, parties believe that they can profit from non-clarity; it is difficult however,
to see how non-clarity could simultaneously work to the benefit of all negotiators in a repeated
game with various partners when there is a strong likelihood that today’s advantages resulting from
non-clarity could haunt them in the future.
 Tariff Protection in the GATT 73

The Harmonized System (HS) supplies the common language. The HS is a

document elaborated in the World Customs Organization (WCO), an inter-
national organization with headquarters in Brussels, Belgium. Its function is to
describe goods in a multilaterally agreed manner. It comprises about 5,000 com-
modity groups, each identified by a six-digit code. Goods’ descriptions are also
expressed in two and four digits: the lower the number of digits, the more generic
the product category (for example, at the two-digit-level one might find the term
‘vehicles’); the higher the number of digits, the more specific the product cat-
egory (for example, at the six-digit-level, one might find the term ‘passenger cars
of less than 2 tons’).
What the HS ensures is that any given good will come under the same tariff
line (heading) across national jurisdictions. This will be the fi xed component;
the variable component will be the tariff treatment. For example, bumpers will
come under the heading 8708.10 which reads ‘Bumpers and Parts Thereof ’ in all
WTO members alike. What might change is the tariff treatment of bumpers: the
United States might impose a 2% import duty, whereas Pakistan 10%. They will
both, however, subject bumpers to heading 8708.10.
The system is used by more than 190 countries as the basis for their customs
tariffs. Not all WTO members have formally adhered to the HS; however, either
for legal reasons (formal adherence to the HS) or de facto, all WTO members fol-
low the HS classification up to the six-digit-level.⁵⁹
From the six-digit-level onwards, WTO members are free to ‘shape’ their con-
cessions to their liking. HS Convention, Art 3.3 (Art 3.3 of the HS) reads in this
respect:
Nothing in this Article shall prevent a Contracting Party from establishing, in its Customs
tariff or statistical nomenclatures, subdivisions classifying goods beyond the level of the
Harmonized system, provided that any such subdivision is added and coded at a level
beyond that of the six-digit numerical code set out in the Annex to this Convention.
However, by adding digits, WTO members can only confer advantages. An
example from the US Tariff Schedule can help explain this point: Chapter 87 of
the HS is entitled ‘Vehicles other than Railway Rolling-Stock, and Parts and accessor-
ies Thereof ’. 8708 is entitled ‘Parts and Accessories of the Motor Vehicles of Headings
8701 to 8705’ (the two categories corresponding to tractors, motor vehicles for

⁵⁹ The HS originates from the Geneva Nomenclature (GN), which came into being on 1
July 1937. The GN was, in turn, replaced by the Brussels Convention on Nomenclature for the
Classification of Goods in Customs Tariffs (BTN) in 1959. The BTN was replaced in 1974 by the
Customs Cooperation Council Nomenclature (CCCN) in 1974 which was replaced by the HS in
1988. The HS has been amended three times since 1988 (largely in order to account for changes in
the technology): 1992, 1996, and 2002. The WTO Committee on Tariff Concessions established
simplified procedures to implement these changes and any future changes in the HS relating to
GATT concessions (see GATT Doc BISD 39S/300). The HS is administered by the HS commit-
tee, which is composed of representatives from each HS contracting party. The HS committee has
the legal power to amend the HS, to prepare explanatory notes and, in general, to provide interpret-
ative guidance as far as the HS is concerned.
74 Disciplines on Trade Instruments

the transport of ten or more persons and motor cars principally designed for the
transport of persons). 8708.10 reads ‘Bumpers and Parts Thereof ’. 8708.10.60
reads ’Bumpers’ (ie stampings). The United States bound its tariffs in Chapter 87
at the eight-digit-level at 2.7%. This means that, if it enters a new category at the
10- or 12-digit-level, it will be able to impose a maximum duty of 2.7%. Eight-
digit descriptions can be challenged before the WTO for, for example, violating
Art I of the GATT (MFN).⁶⁰
The language used in the various nomenclatures is quite generic in light of
the fact that the intention was to subsume to each one of them a large number
of transactions. As a result, disputes might arise as to the appropriate classifica-
tion of a particular transaction. HS, Art 10 reflects the dispute settlement provi-
sions available to its signatories. Disputes will be submitted to the HS committee
which will consider the dispute and make recommendations to the parties. The
HS committee will base its decision on the text of the HS convention, but also on
the various General Interpretative Rules (GIR) that it has adopted over the years,
as well as their Explanatory Notes. GIR 1, for example, subjects classification
according to the terms of a heading, whereas GIR 3(a) stipulates that the heading
which provides the most specific description shall be preferred to headings pro-
viding a more general description. There are specific GIRs for composite goods
(GIR 3(b)).

3.4.2 What is the tariff promise? Ordinary customs duties,

other duties and charges
Based on the HS classification, trading nations will exchange, on a recipro-
cal basis, tariff concessions on various goods. GATT, Art II explains the legal
obligations assumed in this context. The key terms in Art II of the GATT are
‘ordinary customs duty’ (OCD), and ‘other duties and charges’ (ODC). Essentially,
those two terms circumscribe the obligation assumed under Art II of the GATT,
that is, to provide a ceiling that cannot be exceeded. Neither of the two terms is
detailed any further in this provision. GATT, Art II, however, suggests that a
dividing line must be drawn between the two terms, since different obligations
are assumed with respect to each one of them. During a recent dispute between
Chile and Argentina (Chile—Price Band System), both the panel and the AB con-
firmed this view. The issue before the WTO adjudicating bodies was the notori-
ous price band system that Chile practiced: Chile did not impose a fi xed rate of
duty on some imports but, instead, imposed a duty which was calculated on the
basis of the world price of a given commodity, and the corresponding domestic
price. Argentina complained that, when applying this system, Chile sometimes
imposed duties beyond its bound level. The question was whether the Chilean

⁶⁰ As we will see infra, the process of certification of schedules of concessions does not at all
immunize them from subsequent legal challenges. The AB EC—Bananas III report made this
point abundantly clear. On this issue, that is, the consistency of eight-digit classifications, see infra
the discussion on the AB report on EC—Tariff Preferences.
 Tariff Protection in the GATT 75

system qualified as an OCD. The panel first, and the AB later (§§ 264–278)
confirmed that a measure cannot simultaneously come under the purview of the
term OCD and ODC. In the AB’s view, all that is required for something to con-
stitute an OCD is for it to be expressed in a particular form.
The term ODC was clarified in the Understanding on the Interpretation of
Art II.1(b) GATT, which was negotiated and concluded during the Uruguay
round. There, the trading partners aimed to introduce some clarity with respect
to ODCs: there was nothing like an exhaustive list reflecting all of them; there
was substantial ambiguity as to their legal status; and, at the same time, there was
increasing awareness of their damaging effects on international trade transac-
tions. A number of delegations tabled concrete proposals aiming at ‘disciplining’
ODCs. The GATT Secretariat was requested to produce a document⁶¹ which
summarized the submitted proposals and provide a road map towards the solu-
tion of this problem. The Secretariat document made a number of very important
points:

(1) ODCs should be applicable only to imports:

2. It has been established in the Council decision on the Introduction of a Loose-
Leaf System for the Schedules of Tariff Concessions that “such duties or charges
are in principle only those that discriminate against imports” (BISD 27S, p.22)—
i.e., they do not include charges applied to imports and domestic goods alike.
(2) there was nothing like an agreed definition of ODCs:
4. The definition of ODCs falling under the purview of Article II:1(b) can only be
done by exclusion—i.e. by reference to those categories of ODC not covered by
it. It would be impossible, and logically fallacious, to draw up an exhaustive list
of ODCs which do fall under the purview of Article II:1(b), since it is always pos-
sible for governments to invent new charges. Indeed, an attempt to provide an
exhaustive list would create the false impression that charges omitted from it, or
newly invented, were exempt from the II:1(b) obligation.
(3) the applicable date concerning the binding of an ODC:
6. The applicable date is the date as of which Other Duties and Charges are bound.
This is currently accepted as being the date of the instrument by which the tar-
iff concession on the item in question was first incorporated into the General
Agreement, i.e. the date of the first binding of the item.
7. The text of Article II:1(b) relevant for the purpose of determining the applicable
date for bindings of ODCs reads as follows:
. . . Such products shall also be exempt from all other duties or charges of any
kind imposed on or in connection with importation in excess of those imposed
on the date of this Agreement or those directly and mandatorily required to
be imposed thereafter by legislation in force in the importing territory on that
date.

⁶¹ See GATT Doc MTN.GNG/NG7/W/53 of 2 October 1989.

76 Disciplines on Trade Instruments
8. For those bindings undertaken at the time of the institution of GATT, contract-
ing parties are obliged not to raise the level of applicable ODCs beyond that
prevailing at that date. For countries which acceded to the GATT subsequently
the applicable date for bindings undertaken at the time of accession is usually
specified as the date of accession itself. When new concessions are added or old
ones renegotiated under rounds of tariff negotiations the protocols have speci-
fied the applicable date for bindings. The Protocol Embodying the Results of
the 1960–61 Tariff Conference (“The Dillon Round”) specified that the applic-
able date in respect of each product which was the subject of a concession in
that Round would be the date of the protocol, provided the product in question
was not already subject to a binding provided for in the same part or section
of a schedule to the General Agreement. In other words the date of the Dillon
Round Protocol was only the applicable date for concessions negotiated for the
first time in that Round: for earlier concessions renegotiated in the Dillon Round
the applicable date remained that of their first incorporation into the General
Agreement.
12. In the case of items bound for the first time in the Uruguay Round the nature
and level of ODCs to be inscribed in schedules would be those in force at the
date of the Uruguay Round tariff protocol.
(4) the legal effect of inscribing an ODC in a schedule of concessions:
18. In itself the inscription of ODCs in schedules would have no legal effect. The
binding commitment under Article II:1(b) already exists; the inscription of
bound levels would simply make existing obligations transparent.
20. Taking up the second category of questions, it has been asked in the Group
whether the inscription of ODCs in the schedules as called for by the proposal
would imply recognition of their legality; or, in other words, would not the pos-
sibility of legalizing GATT illegal charges be created. A recent panel—US—
Restrictions on Imports of Sugar (L/6514) ruling is relevant to this question.
The panel concluded “that the restrictions on the importation of certain sugars
maintained by the United States under the authority of the Headnote in the
Tariff Schedule of the United States are inconsistent with Article XI:1 and can-
not be justified under the provisions of Article II:1(b).” The panel argued that
the wording of the Article suggests that contracting parties cannot incorporate
provisions in their schedules to qualify their obligations under other Articles of
the General Agreement. This would indicate that the mere inscription of ODCs
in the schedules (even if they are certified) cannot place them beyond legal chal-
lenge on grounds of inconsistency with obligations under other Articles of the
General Agreement. If a panel were to find a particular charge illegal it could
recommend the elimination of the charge and its removal from the schedule.
As a further safeguard, New Zealand’s latest proposal suggests that any duty or
charge claimed would be recorded without prejudice to its legal status.

Many of these observations found their way into the Understanding on the
Interpretation of Art II:1(b) of the GATT, concluded during the Uruguay
round. The travaux preparatoires of the understanding reveal the willingness of
 Tariff Protection in the GATT 77

negotiators to define ODCs by exclusion; it was agreed that it would have been
a daunting task to attempt to draw an exhaustive list of ODCs, a point that the
panel, in its report on Dominican Republic—Import and Sale of Cigarettes, fully
endorsed (§7.114). According to the Understanding, not only OCDs, but also
ODCs imposed on imported products, will have to be reflected in schedules of
concessions. Their inclusion, however, does not ipso facto guarantee their con-
formity with the GATT. WTO members are free to challenge such duties and
charges before the WTO adjudicating bodies.
The Understanding reads:
Members hereby agree as follows:
1. In order to ensure transparency of the legal rights and obligations deriving from para-
graph 1(b) of Article II, the nature and level of any “other duties or charges” levied on
bound tariff items, as referred to in that provision, shall be recorded in the Schedules
of concessions annexed to GATT 1994 against the tariff item to which they apply. It
is understood that such recording does not change the legal character of “other duties
or charges”.
2. The date as of which “other duties or charges” are bound, for the purposes of Article
II, shall be 15 April 1994. “Other duties or charges” shall therefore be recorded in the
Schedules at the levels applying on this date. At each subsequent renegotiation of a
concession or negotiation of a new concession the applicable date for the tariff item
in question shall become the date of the incorporation of the new concession in the
appropriate Schedule. However, the date of the instrument by which a concession on
any particular tariff item was first incorporated into GATT 1947 or GATT 1994 shall
also continue to be recorded in column 6 of the Loose-Leaf Schedules.
3. “Other duties or charges” shall be recorded in respect of all tariff bindings.
4. Where a tariff item has previously been the subject of a concession, the level of “other
duties or charges” recorded in the appropriate Schedule shall not be higher than the
level obtaining at the time of the first incorporation of the concession in that Schedule.
It will be open to any Member to challenge the existence of an “other duty or charge”,
on the ground that no such “other duty or charge” existed at the time of the original
binding of the item in question, as well as the consistency of the recorded level of any
“other duty or charge” with the previously bound level, for a period of three years after
the date of entry into force of the WTO Agreement or three years after the date of
deposit with the Director-General of the WTO of the instrument incorporating the
Schedule in question into GATT 1994, if that is a later date.
5. The recording of “other duties or charges” in the Schedules is without prejudice to
their consistency with rights and obligations under GATT 1994 other than those
affected by paragraph 4. All Members retain the right to challenge, at any time, the
consistency of any “other duty or charge” with such obligations.
6. For the purposes of this Understanding, the provisions of Articles XXII and XXIII
of GATT 1994 as elaborated and applied by the Dispute Settlement Understanding
shall apply.
7. “Other duties or charges” omitted from a Schedule at the time of deposit of the
instrument incorporating the Schedule in question into GATT 1994 with, until
the date of entry into force of the WTO Agreement, the Director-General to the
78 Disciplines on Trade Instruments
CONTRACTING PARTIES to GATT 1947 or, thereafter, with the Director-
General of the WTO, shall not subsequently be added to it and any “other duty or
charge” recorded at a level lower than that prevailing on the applicable date shall not
be restored to that level unless such additions or changes are made within six months
of the date of deposit of the instrument.
The decision in paragraph 2 regarding the date applicable to each concession for the pur-
poses of paragraph 1(b) of Article II of GATT 1994 supersedes the decision regarding the
applicable date taken on 26 March 1980 (BISD 27S/24).
Consequently, with respect to ODCs, we can conclude that:
(1) ODCs have to be recorded in schedules of concessions otherwise they are
ipso facto GATT inconsistent. However, since ODCs are linked to customs
duties, and since only bound customs duties appear in a schedule, in practice,
ODCs applicable to unbound duties do not have to be included in schedules
of concessions;⁶²
(2) ODCs are not by virtue of their inclusion in a schedule of concession GATT
consistent. WTO members can, for a period of three years after the entry
into force of the WTO, challenge ODCs reflected in a schedule of conces-
sions, either because an ODC did not exist at the time of the original binding
of the item in question, or because it exceeded its prior level (§ 4);
(3) Even after this period of three years, WTO members can still challenge
ODCs on grounds other than those mentioned above, that is, challenge them
for being GATT inconsistent (§ 5).
It follows that nowadays ODCs are being negotiated just like OCDs. As is the
case with OCDs, ODCs additional to those existing before cannot be entered;
their introduction to schedules of concessions makes ODCs transparent. Through
negotiations, they will be reduced, if not eliminated altogether.

3.4.3 The typology of ordinary customs duties

Consolidated concessions (that is, bound import duties) are expressed in:
(1) ad valorem duties;
(2) specific duties;
(3) compound duties;
(4) alternative duties (or mixed duties); and
(5) technical duties.

⁶² The argument, however, could be made that unbound ODCs must observe the transparency
requirement laid down in Art X of the GATT. Assuming this view is upheld in case-law it does
not necessarily eliminate all problems that might arise in practice. Interested traders will have to
inquire into (sometimes arcane and hard to localize) national laws, instead of simply checking the
national schedule of concessions notified to the WTO.
 Tariff Protection in the GATT 79

A duty is ad valorem when it is a percentage of the value of the imported product

(eg 15% of a fax machine the import price of which is $400). A duty is specific
when it is related to the weight, volume, surface, etc of the good at hand (eg $20
per ton of imported wheat). A duty is compound when it comprises an ad valorem
duty to which the customs authority adds or from which it subtracts a specific duty
(eg 10% on the import price of wheat plus $2 per imported kg of wheat). A duty
is alternative or mixed when it ensures a minimum or maximum tariff protection
through the choice between, in most cases, an ad valorem and a specific duty (eg
10% on the import price of or $2 per imported kg of wheat whichever is the max-
imum). Finally, particularly where agricultural products are concerned, a technical
duty is determined by complex technical factors such as alcohol or sugar content.
Other duties or charges are usually expressed in ad valorem terms.⁶³

3.4.4 Multilateral trade negotiations (rounds)

As already mentioned in Chapter 1, concessions will be exchanged in the con-
text of multilateral trade negotiations, the so-called trade rounds. GATT, Art
XXVIII bis makes it clear that, there is not a pre-determined kick off for any
round; it is the WTO membership that will decide, when warranted, to open up
multilateral talks. At the moment of writing, we are in the middle of the ninth
round since the inception of the GATT, the Doha round.⁶⁴ All WTO members
participate in a trade round.
Different alliances have been formed during rounds. Some of them have been
quite ephemeral,⁶⁵ but some like the CAIRNS group have managed to withstand,
so far at least, the test of time.⁶⁶ During the Doha round, negotiations have been
taking place across three groups:
(1) the OECD countries;
(2) the G 20, comprising the leading developing countries, such as Argentina,
and Brazil;

⁶³ ODCs feature under column 6 of the Uruguay round schedules. Only Botswana, Côte
d’Ivoire, Namibia, Swaziland, and South Africa have used compound ODCs.
⁶⁴ Rounds, with few exceptions, tend to get their name from the place where negotiations are
launched (exceptions include the Kennedy or the Dillon round). Notorious across rounds are the
so-called Green room meetings, where chosen delegations only will participate in the Director-
General’s office (which used to be green). The objective is to restrict access to few key delegations
in order to favour agreement. Over the years the number of participants has grown. Then, there are
the meetings for Heads of Dels (Heads of Delegations). The Trade Negotiating Committee (TNC),
where representatives of all WTO members participate, and is usually headed by the Director-
General of the GATT/WTO oversees the progress in any multilateral round.
⁶⁵ The Quad, for example, used to be the all-powerful group comprising the European
Community, the United States, Canada, and Japan. During the Doha round, the new Quad, where
Brazil and India have replaced Canada and Japan, has been leading the negotiations. Negotiation-
specific groupings have also seen the light of day: Brazil, India, and South Africa, for example,
head the NAMA 11 group, which has been very active during the non-agricultural market access
NAMA negotiations of the Doha round.
⁶⁶ The CAIRNS group has been pushing for liberalization of farm trade, see infra in Chapter 3.
80 Disciplines on Trade Instruments

(3) the G 90, comprising the LLDCs, as well as the remaining developing
countries.
Most of the negotiations of the Doha round concern demands that G 20 has
been formulating and are directed towards the Organization for Economic
Development and Cooperation (OECD) group (and some counter-demands by
the latter directed to the former).⁶⁷ G 90 have been to a large extent absent from
the negotiating scene, partly motivated from declarations by EC officials that
they should have the round for free (that is, without making any concessions).⁶⁸
MFN liberalization nevertheless (by the OECD and the G 20 groups) has been
seen with scepticism by the G 90: the countries in this group fear that this process
will lead to erosion of the preferences they have been enjoying. Preference erosion
has been discussed in different fora during the Doha round, and comprehensively
in the Aid for Trade group.⁶⁹
WTO members will have to, of course, faithfully implement the results of
the negotiations. This is what pacta sunt servanda amounts to. Those who fol-
low negotiations nevertheless, must have noticed the ongoing debates about the
US fast-track. This is a particularity of the US constitutional procedures which
should have no bearing on the negotiations, at least from a legal perspective (since
the US government is bound by the pacta sunt servanda principle, just like any
other WTO member). By agreeing to a fast-track procedure, the US Congress
will approve or disapprove the final product of negotiations, without being in a
position to modify their content (which it should be doing anyway, by virtue of
pacta sunt servanda). This is what the Trade Promotion Authority (TPA), the legal
basis for fast-track, essentially amounts to. The TPA is accorded for a fi xed period
of time and can be renewed. To avoid trouble, the negotiating community strives
to conclude a round while the TPA is still in force.
At the end of the negotiating process, some schedules will reflect one (the
MFN), and some, two tariff rates: the MFN rate applicable by default, and the
preferential rate, applicable to a sub-set of the WTO membership (the developing

⁶⁷ Th is is not odd. According to WTO statistics, the European Community represents approxi-
mately 20% of all world imports and 20% of all world exports (in value); the United States, 15%
of world exports, and 25% of world imports (hence, the trade deficit). The big two hence, represent
more or less 45% of all world imports, and it is quite normal that they have been receiving the
majority of demands for market access. Efforts to open up the Chinese market are increasing.
China represents 15% of world exports and 13% of world imports. At first glance, such percent-
ages sound quite reasonable. Moreover, one should not neglect that China has a high savings rate
(approximately 50% of its gross domestic product (GDP) goes to savings), mainly as a result of the
absence of a national social security scheme: working people substitute pension (when they retire)
with savings while they are still employed. The WTO mandate does not extend to social policy
issues: requiring thus from China to introduce such social schemes, is not an issue that appropri-
ately finds its place in the context of multilateral trade negotiations.
⁶⁸ As we will see infra, nevertheless, when we discuss special and diff erential treatment for devel-
oping countries, this is not a good strategy: those who do not participate in a round do not have a
say on the items where donors will make concessions, and will not undertake the (sometimes) neces-
sary adjustment. Such proposals are thus of highly questionable value for developing countries.
⁶⁹ At the time of writing, the form of the antidote for preference erosion had not been agreed.
 Tariff Protection in the GATT 81

countries, and the countries with which the WTO member making the tariff
promise has entered into a preferential trade agreement). Over subsequent
rounds, WTO members will make new concessions, that is, they will agree to
lower bound tariffs. This does not mean, however, that entries prior to the most
recent become legally irrelevant. In fact, the opposite is true: in order to obtain a
complete picture of the situation regarding the status of a WTO member’s sched-
ule of concessions in the area of goods, it is sometimes necessary to look, not only
at the most recently negotiated concessions, but also to the situation before. The
country, for example, that initially negotiated the concession has the legal right
to participate in an Art XXVII of the GATT process: assuming a country that
made a tariff promise wishes at a subsequent stage to revise its promise, it will
have to negotiate compensation, inter alia, with the country that initially negoti-
ated the concession.⁷⁰ On the other hand, it could be the case that, in the most
recent round, a mere amendment was introduced to a concession, or an amend-
ment was introduced in the interim (that is, between multilateral trade rounds).
These amendments may be termed renegotiations or modifications resulting from
action under the GATT Arts II (schedules of concession), XVIII (governmental
assistance to economic development), XXIV (PTAs: customs unions (CU), and
FTAs, XXVII (withholding or withdrawal of concessions), or XXVIII (modifi-
cation of schedules), or even rectifications to correct errors found in the sched-
ules. The value of the amendment will be hard to ascertain absent a look into the
history of the concession. The status of Canada’s commitments in 1980 offers an
appropriate illustration of this point:
Canada—Schedule V
Part I—MFN
— Geneva 1947, amended in PR1/48, PR3/49, PR4/50, PRM5/55, PRM6/57,
PRM7/57, BR/58;
— Annecy 1949, amended in PR4/50, PRM5/55, PRM6/57, PRM7/57;
— Torquay 1951, amended in PRM/5, PRM6/57, PRM7/57;
— Japanese Protocol 1955, Geneva 1956, amended in PRM6/57, PRM7/57;
— Swiss declaration 1958, Geneva 1962, amended in the Third Certification of
Rectification and Modifications 1967;
— Israeli Protocol 1962;
— Portuguese Protocol 1962;
— Spanish Protocol 1963;
— Kennedy round 1964–67, amended in First Certification of Changes 1969;
— Geneva 1979 Protocol;
— Protocol Supplementary to Geneva 1979 Protocol
Part II—Preferential
— Geneva 1947, amended in PRM5/55, PRM6/57, PRM7/57;
— Annecy 1949, amended in PR4/50, PRM5/55, PRM6/57;

⁷⁰ More on this, later in this chapter.

82 Disciplines on Trade Instruments
— Torquay 1951, amended in PRM5/55, PRM6/57, PRM7/57;
— Geneva 1956;
— Kennedy round 1964–67.
GATT Docs. SECRET/180 and Add.1, SECRET/244 and Add. 1–4 (XXVIII.5),
SECRET/183 and Add. 1 (XXVIII.4), SECRET 193 (II.5) and SECRET/224/Add. 4
(XXVIII.3) also reflect Canada’s concessions.
Concessions have, on occasion, been exchanged also in mini-rounds, where only
a sub-set of the WTO membership participated: for instance a few WTO mem-
bers have taken the initiative and liberalized trade in information technology
goods; this is the story of the Information Technology Agreement (ITA).
The ITA was concluded at the Singapore Ministerial Conference⁷¹ in
December 1996. At that time 29 (including the then 15 member states of the
European Union) countries or separate customs territories signed the declar-
ation. However, it was still unclear at that time whether the provisions of the
Declaration would come into effect, as the Declaration stipulated that partici-
pants representing approximately 90% of world trade would have to notify their
acceptance of the ITA by 1 April 1997. The original 29 signatories did not reach
this 90% trade coverage criteria, since they collectively accounted for only 83%
of world trade in information technology (IT) products. However, in the ensu-
ing months after the Singapore Ministerial Conference and leading up to 1 April
1997, a number of other countries expressed an interest in becoming participants
in the ITA and notified their acceptance. Thus, the 90% criterion was met, and
the ITA entered into force with the first staged reduction in tariffs occurring on
1 July 1997. Following ITA I, an ITA II was successfully concluded. The WTO
homepage (<http://www.wto.org>) describes the ITA in the following terms:
The ITA is solely a tariff cutting mechanism. While the Declaration provides for the
review of non-tariff barriers (NTBs), there are no binding commitments concerning
NTBs. There are three basic principles that one must abide by to become an ITA partici-
pant: 1) all products listed in the Declaration must be covered, 2) all must be reduced to a
zero tariff level, and 3) all other duties and charges (ODCs) must be bound at zero. There
are no exceptions to product coverage, however for sensitive items, it is possible to have
an extended implementation period. The commitments undertaken under the ITA in the
WTO are on an MFN basis, and therefore benefits accrue to all other WTO Members.
Normal implementation period
Staging of concessions—the Declaration provides in principle for the staging-in of
concessions in equal rate reductions based on the following timeframe:
1st 1 July 1997
2nd 1 January 1998
3rd 1 January 1999
4th Complete elimination of duties no later than 1 January 2000

⁷¹ According to Art IV of the WTO Agreement, a Ministerial Conference is composed of rep-

resentatives of all members and meets at least once every 2 years. The Ministerial Conference can
take decisions on all matters under any of the multilateral trade agreements.
 Tariff Protection in the GATT 83
Extended staging/developing countries
The Declaration provides that “unless otherwise agreed by the participants” each par-
ticipant would follow the staged reductions listed above. However, there have been a
number of countries that have requested extended staging (beyond the year 2000) dur-
ing the course of negotiations, and which have been agreed to by the other participants.
This is reflected in their schedule of commitments. In general, developing countries have
requested and received extended staging for at least some products in their schedule. In
no case, however, is the extended staging period beyond the year 2005.
...
The Committee of Participants on the Expansion of Trade in Information Technology
Products
After the ITA came into being, the participants saw the need to establish a formal
Committee under the WTO to carryout the provisions of the Declaration. The document
on Implementation of the Ministerial Declaration on Trade in Information Technology
Products (hereinafter “Implementation” document, G/L/160) sets out the establishment
of the Committee. The Committee held its first meeting on 29 September 1997, and
there have been approximately five formal meetings per year of the Committee since that
time. The Committee has established rules of procedure (WTO Doc. G/IT/3) which are
similar to other WTO bodies and typically has a diverse agenda of matters relating to
the ITA.
Ongoing Work of the Committee
In addition to the topic of product coverage, mentioned below, the Committee has
worked on a number of issues since its inception. These include the examination of classi-
fication divergences, consultations on non-tariff barriers, invoking new participants, and
discussing implementation matters.
ITA II
The Ministerial Declaration and the Implementation document provides that partici-
pants will periodically review the product coverage (often termed ITA II) specified in the
Attachments to the Declaration. The first review commenced in October 1997 when par-
ticipants were invited to submit lists of additional information technology products for
possible additional tariff concessions. In 1998, participants continued to discuss the lists,
provide clarification, exchange views, and negotiate with a view to making a decision on
whether to revise the Attachments by 30 June 1998. However, no agreement could be
reached at that time, and the process continued throughout 1998. At subsequent meet-
ings of the Committee, still no agreement could be reached and to date there has been
no products added to the original coverage. Consultations between delegations on the
review of product coverage continue. (excerpt from the WTO homepage, <http://www.
wto.org>)
All results of the ITA negotiations were subsequently ‘multilateralized’, and are
now applied in a non-discriminatory (ie across the WTO membership) manner.⁷²
The ITA has now 69 members (counting the European Community as 25), out
of which 35 are developing countries. Together, the 69 members of the ITA
represent approximately 97% of world trade in IT products. Mexico, South Africa,

⁷² Hence, it is wrong to understand the ITA as a plurilateral agreement. As a result of the multi-
lateralization of the results, no MFN issue (see infra) arose.
84 Disciplines on Trade Instruments

and Brazil are the most important non-ITA participants, accounting for more or
less 3% of world ITA trade. After its peak in 2000 (16.5%), the ITA now counts
for approximately 12% of total world exports; the share of ITA products in world
trade exceeds (in 2006) that of farm products. With the outcome of the Doha
round still pending, the ITA is the most important MFN trade liberalization ini-
tiative after the Uruguay round.

3.4.5 The treatment of export tariff s (taxes) under GATT

The discussion so far was dedicated exclusively to the treatment of import tariffs
under GATT. Export tariffs, however, do occur, albeit not with the same fre-
quency as import tariffs. In fact, countries might have the same incentives to use
import or export tariffs since, depending on their market power in a given prod-
uct, they can affect terms of trade to their favour through the use of either of these
two instruments (Lerner theorem). GATT, Art XXVIII bis makes it clear that,
not only import concessions, but export concessions as well can be negotiated.
The relevant part reads:
The contracting parties recognize that customs duties often constitute serious obstacles
to trade; thus negotiations on a reciprocal and mutually advantageous basis, directed to
the substantial reduction of the general level of tariffs and other charges on imports and
exports and in particular to the reduction of such high tariffs as discourage the import-
ation even of minimum quantities, and conducted with due regard to the objectives
of this Agreement and the varying needs of individual contracting parties, are of great
importance to the expansion of international trade. The CONTRACTING PARTIES
may therefore sponsor such negotiations from time to time. (emphasis added)
It follows that nothing stops WTO members from negotiating export conces-
sions: they can promise, for example, that they will not impose an export tax
on commodities that are scarce. The process for consolidating concessions is not
spelled out in the GATT. The wording of Art II of the GATT leaves no doubt
that it was designed to be the institutional vehicle for the consolidation (binding)
of import tariffs. There is no Art II of the GATT equivalent for export tariffs.
This does not mean, however, that export concessions cannot be negotiated on an
ad hoc basis. What matters is that at the end of the day the schedule of concessions
of a WTO member reflects its commitment not to impose an export tariff, and
not which specific procedure has been followed to this effect. Empirically speak-
ing, there are few export concessions that have been negotiated.
Lack of practice is probably the single most important reason explaining
why the founding fathers did not spend time and effort designing a mechanism
for negotiation of export tariffs à la Art II of the GATT. We read, for example,
from the interim report of the (US) Special Committee on Relaxation of Trade
Barriers:⁷³

⁷³ Report of 8 December 1943, International Trade Files, Lot File 57D-284, p 34.
 Tariff Protection in the GATT 85
Understanding on the Interpretation of Art II:1(b) GATT
Except during wartime, governmentally-imposed export duties, and prohibitions and
quantitative restrictions on exports have had relatively little influence in limiting the
over-all movement of commodities in world trade, although they have seriously affected
the movement of specific products. Export taxes and quantitative restrictions on exports
have been instituted for a variety of reasons. Some, such as export taxes on coffee in cer-
tain Latin American countries, have been imposed for revenue purposes. Some have been
imposed for indirect protective reasons: for example, the United States prohibition on
commercial exports of tobacco seed for the purpose of preventing the cultivation abroad
of American types of tobacco. In a different category are the Mexican export taxes, which
are used for revenue purposes and, in combination with an export tax-rebate system to
enforce membership in export cooperatives. Some, such as the United States control of
helium exports, have been imposed for security reasons. Some have been imposed pur-
suant to international agreements; for example, the undertaking by Cuba, in connection
with the trade agreement with the United States, to prohibit the exportation of avocados
to the United States except during the months of July through September . . .
In the same report, the committee believes that, the lack of actual trade inter-
est notwithstanding, it would still be sensible to regulate export taxes in light of
their potential interest. The committee believed that provision should be made,
in the context of a trade agreement,⁷⁴ to abolish all objectionable export taxes. In
its view, export taxes for revenue purposes, or enforced pursuant to international
agreements, or imposed under the conditions of famine or severe domestic short-
age in the exporting country, or even designed to regulate the trade in military
supplies under specified conditions, should not be regarded objectionable.⁷⁵ The
final compromise nevertheless, was different: whereas export quantitative restric-
tions were disallowed, this was not the case with export taxes. As explained supra,
export taxes are not covered by Art XI of the GATT, that is, they are not QRs in
the GATT sense of the term. Consequently, WTO members are, absent a com-
mitment to this effect, free to impose export taxes to their liking.⁷⁶

⁷⁴ The GATT was the envisaged trade agreement.

⁷⁵ See pp 35ff of the report of the committee, op cit. See, also, Dam (1977) discussing the US
southern states’ opposition to export taxes.
⁷⁶ Anecdotally, it probably bears mention that an initiative was launched in the Doha round
to subsume export taxes to the general prohibition under Art XI of the GATT. Developing coun-
tries staunchly opposed it and the issue was dropped. Developing countries have on occasion had
recourse to export taxes in order to confer an advantage to their domestic processing industry.
Through export taxes one can, of course, influence terms of trade of a particular commodity to
their advantage. Assuming they have market power (as, for example, in the case of petrol), export-
ing countries can influence the terms of trade of a particular commodity through an export tax.
A typical example is export taxes on phosphates by Morocco: in a recently concluded bilateral free
trade agreement between the United States and Morocco, the latter undertook the obligation not
to impose export taxes on phosphates (Annex 2c of the agreement). For most commodities, absent
collusion, exporters have no market power. Hence, tariff escalation schemes whereby the input is
taxed by an export tax but not the output have not delivered the expected benefits. Note that devel-
oping countries typically keep imports of raw materials (inputs) at duty free level, while imposing
tariffs on the final product to protect the value added domestic industry: cotton and clothing offer
an appropriate illustration.
86 Disciplines on Trade Instruments

3.4.6 Certification of schedules

At the end of a round, schedules will be certified (validated) through a multilat-
eral process. Usually, schedules will be annexed to a protocol which will be open
for signature to all WTO members.⁷⁷ Clarifications might be sought and, even-
tually, the Protocol shall be registered in accordance with the provisions of Art
102 of the UN Charter.⁷⁸ The decisions by which protocols are adopted are very
ambiguous as to their legal effects. The question namely arises whether WTO
members, by accepting the protocol have, ipso facto, accepted that all schedules
annexed are GATT consistent. The response to this question has been provided
in case-law, and, as we will see infra, is negative.⁷⁹
Reducing tariff protection has often been hailed by many as the crowing
achievement of the GATT, and there is a lot of truth in it, even though, as Irwin
(1996a and 1998a) has argued, one has probably exaggerated the size of the suc-
cess. Irwin has argued that (average) duties calculated on dutiable imports stood
at 19.34% in 1947 and 13.87% in 1948. However, were one to include duty free
trade, the ration of duties collected to total imports amounted to 7.55% in 1947
and 5.71% in 1948. The average tariffs on dutiable imports (weighted with the
1939 trade values) in 1947 was 32.2% and 25.4% in 1948. As 60% of US imports
were duty free in 1947, the average tariff for total imports was much lower. In
addition, were tariff averages to be weighted with the 1947 trade weights (and
prices and not with 1939 values) then the post-Geneva average tariff on dutiable
imports drops from 25.4% to 15%. Including duty free imports, the pre-Annecy
(GATT 1948) tariff average of all US imports was estimated by the US Tariff
Commission to amount to 5.9%. This number fits unusually well with the ration
of duties calculated to total imports (free and dutiable) of 5.97% in 1950.
The number of tariff lines bound is not identical across WTO members. It is
usually the case that big, rich markets (such as the European Community and
the United States) have made the maximum number of tariff commitments.
Developing countries have on average, made fewer commitments.⁸⁰

⁷⁷ See for example, GATT Doc L/7463 of 24 May 1994, reflecting the Geneva (1994) Protocol
to the GATT to which the schedules of concessions made during the Uruguay round were annexed.
See also GATT Doc PC/M/2 of 2 June 1994 at p 3. Hoda (2001) includes in pp 227ff all protocols
signed at the end of each GATT round since the inception of the GATT and until (and including)
the Uruguay round.
⁷⁸ The GATT practice is described in Jackson (1969, pp 211ff ).
⁷⁹ See infra, section 3.5.3 for a discussion of the AB EC—Bananas III report.
⁸⁰ Th is is not value judgment on who is and who is not a free market, since market access can be
impaired through other measures as well, through contingent tariffs (such as AD), or even through
NTBs. On the other hand, some researchers have cast doubt to the idea that high tariff s are uni-
versally detrimental to growth, arguing that policy prescriptions (such as low tariffs) designed to
promote growth within developed economies may not be appropriate for universal adoption, see on
this score, Rodriguez and Rodrik (2001), and more recently, de Jong and Ripoli (2006) who use a
panel data set comprising 60 countries and spanning 1975–2000.
 Tariff Protection in the GATT 87

3.5 Interpreting tariff commitments

A series of issues relating to tariff commitments have given rise to litigation over
the years. Case-law has in a sense completed the GATT contract, although not
always in a linear manner. There is still substantial ambiguity as a result of some
of the judicial interventions in this context. One classification of the case-law in
this context could be as follows:
(1) recourse to the HS rules in order to interpret a tariff concession is a matter of
legal compulsion (section 3.5.1);
(2) the scheduling of concessions must be WTO consistent (section 3.5.2);
(3) finally, in case of ambiguity as to what has been agreed, the legitimate expec-
tations of one of them are immaterial for the interpretation of the concession
(section 3.5.3).

3.5.1 The legal relevance of HS

The legal status of the HS was clarified in recent case-law. The panel report on
EC—Chicken Cuts first held the view that the HS provides context, in the VCLT-
sense of the term (Art 31.2), for the schedules of concessions submitted by WTO
members.⁸¹ As such, panels must always take it into account when the issue of
interpretation of a particular commission arises. The AB confirmed this finding
(§ 199).
In EC—Chicken Cuts, the panel dealt comprehensively for the first time with
the interpretation of a concession. The relevant facts are adequately described in
§§ 7.2 and 7.3 of the report:
The EC Schedule provides for a tariff of 102.4€/100kg/net for products covered by sub-
heading 0207.14.10 and allows the European Communities to use special safeguard
measures under Article 5 of the Agreement on Agriculture in respect of such products.
The EC Schedule provides for a tariff of 15.4% ad valorem for products covered by sub-
heading 0210.90.20 and there is no reservation for the use of special safeguard measures
under Article 5 of the Agreement on Agriculture in respect of such products.
Brazil and Thailand (the complainants) submit that less favourable treatment has been
accorded to frozen boneless salted chicken cuts in violation of Article II:1(a) and Article
II:1(b) of the GATT 1994 because, through the relevant EC measures, the European
Communities changed its customs classification so that those products, which had pre-
viously been classified under subheading 0210.90.20 and were subject to an ad valorem
tariff of 15.4%, are now classified under subheading 0207.14.10 and are subject to a tariff
of 102.4€/100kg/net as well as being potentially subject to special safeguard measures
pursuant to Article 5 of the Agreement on Agriculture. (original emphasis)

⁸¹ Van Damme (2007) has advanced the view that, although schedules of commitments are
undeniably treaty language (and, consequently, it is only appropriate that the starting-point of
interpretation is the VCLT) some adjustments are in order, in light of their special characteristics.
She imports notions of public international law that will help the WTO judge to develop future
case-law in this field in a linear manner.
88 Disciplines on Trade Instruments

Salted meat exported from Brazil was a product which suffered from the
change in classification. Note, however, that this is not, at least in the European
Community’s argument, a case of unilateral modification: the argument was that
salted meat, if not salted for preservation purposes, could not benefit from the lower
tariff since this tariff line was intended to cover meat salted for preservation pur-
poses only. This was not the case of Brazil’s exports.⁸² The panel rejected the
arguments advanced by the European Community, and agreed with Brazil.⁸³
In its view, nothing in the HS description conditioned the classification of salted
meat under 02.10 on the purpose of salting (that is, preservation). In doing that,
however, it went through all interpretative elements of the VCLT system, pay-
ing particular attention to the HS system since it considered it to be context of
the negotiation (§§ 7.104ff ). The AB upheld the panel’s findings in this respect
(§§ 199ff ).⁸⁴

3.5.2 Scheduling of concessions must be WTO consistent

GATT, Art II.1b states that goods will be imported in a market:
subject to the terms, conditions or qualifications set forth in that Schedule.

⁸² Note that Brazil did not even argue that this was the case. Its argument was simply that noth-
ing in the tariff line indicated that only salted for preservation purposes meat was covered; in its
arguments any salted meat, that is, even meat containing a salt percentage that falls short of guar-
anteeing preservation, was covered.
⁸³ Remarkably, in this case, as in EC—Computer Equipment before it, the parties to the dis-
pute defended before the panel the view that their dispute did not concern a classification issue
(although this is precisely what this case was all about). The panel did not disagree with the parties.
Such cases are evidence of an emerging trend that the passage to compulsory adjudication (with the
advent of the Uruguay round) has a counter-weight: panels will accept artificial claims and will not
disturb them, even if they might disagree with them.
⁸⁴ The AB took a slightly different view on the use of the interpretative elements but did not
disturb the panel’s findings on the proper classification of salted meat at all. The parties to the dis-
pute could have submitted their dispute to the HS dispute settlement procedures explained supra,
which they did not. On the other hand, the panel could have used its powers under Art 13 of the
DSU and request an expert testimony from HS officials. It did not, however. It did address ques-
tions to the HS committee, but exercised its own discretion when evaluating the responses granted.
It seems that the responses by the HS committee were rather favourable to the claims made by the
European Community. Still, the panel did not side with the HS committee. As a result, the panel
adopted an interpretation that does not, it seems, coincide with the preferred interpretation of the
HS Committee. Th is is unfortunate. So, while the HS now provides the context, panels retain dis-
cretion when interpreting it. The amount of discretion here should not be underestimated. Recall
the discussion of the GIR supra: they are hardly self-interpreting. One could imagine different
alternatives to deal with this issue in a more satisfactory manner: panels could make more use of
their discovery powers (Art 13 of the DSU), or even, panels could interrupt their process and defer
the issue to the HS dispute settlement. The latter requires an amendment of the current agree-
ments, so it is not on the cards, at least in the short run. The former requires no amendment at all:
while panels retain the authority to weigh expertise, the Art 13 of the DSU process is more sol-
emn and the provided expertise will be easier scrutinized. The panel on EC—Chicken Cuts regret-
tably opted for a less transparent questionnaire addressed to the HS committee, rather than a more
formal invitation to come and provide testimony before the panel during the proceedings.
 Tariff Protection in the GATT 89

The question arises whether the terms, conditions or qualifications included in a

schedule supersede the GATT provisions or not. During the GATT years, grand-
fathering was a well-known process:⁸⁵ trading nations would exempt from GATT
obligations aspects of their trade policy that they did not want to subject to the
GATT disciplines. The advent of the WTO signalled the end of grandfathering.
WTO members wishing to exclude aspects of their trade policy from the GATT
disciplines were essentially left with one option: request a waiver. During the
notorious EC—Bananas III dispute between the European Community and vari-
ous banana-exporting countries, the question arose, inter alia, whether a WTO
member can specify the conditions of import in its schedule of concessions in a
discriminatory, that is, a GATT inconsistent manner,⁸⁶ depending on the origin
of the imported good. The European Community had included in its schedule of
concessions a condition, according to which, some WTO members (which had
voluntarily concluded with the European Community the so-called Framework
Agreement), could profit from a preferential rate on bananas, whereas the remain-
ing WTO members would not. A series of banana-exporting WTO members
complained claiming that the measure at hand was inconsistent with the MFN
rule. The European Community argued in response that its regime was consist-
ent with Art II.1b of the GATT, that it had been notified to the WTO member-
ship and had gone through the certification process as well, without any WTO
member raising an issue as to its legality. The question thus before the panel could
be phrased as follows: do WTO members, when scheduling their concessions, have
to observe the GATT disciplines? Or, conversely, whether GATT obligations are
legally relevant only after schedules have been certified.
The wording of Art II.1b of the GATT⁸⁷ suggests that the EC-approach should
be accepted: if at all, the treatment of imported goods is subjected to conditions,
etc included in the schedule; nowhere does Art II of the GATT mention that such
conditions, etc should be GATT consistent. By virtue of the in dubio mitius legal
principle,⁸⁸ one would expect that an adjudicating body would find that, in the
absence of explicit, unambiguous transfer of sovereignty in this respect, trading
nations could conclude whatever they please:⁸⁹ this is what contractual autonomy
amounts to. The panel first, and the AB then, recalling the earlier Headnote juris-
prudence, rejected the argument of the defendant. In their view, a WTO member
can, through conditions attached to its schedule, grant other WTO members

⁸⁵ See the relevant passage in Jackson (1969).

⁸⁶ See the discussion on MFN infra.
⁸⁷ The panel and the AB discussed the present litigation against the background of Art 4.1 of
the AG which reflects a standard very comparable to that of Art II.1b of the GATT. The fact that,
as the quoted passage underscores, the AB based its findings on the headnote jurisprudence which
is an Art II of the GATT jurisprudence, leaves no room for doubt that, in the eyes of the two WTO
adjudicating bodies, there is absolute parallelism between Art II.1b of the GATT on the one hand,
and, Art 4.1 of the AG on the other.
⁸⁸ Note that the AB, like many international courts, has adhered to this principle.
⁸⁹ To the extent, of course, that they do not violate peremptory norms of public international law.
90 Disciplines on Trade Instruments

rights but it cannot, through this means, diminish its obligations. We quote from
§§ 154–8:
The market access concessions for agricultural products that were made in the Uruguay
Round of multilateral trade negotiations are set out in Members’ Schedules annexed to
the Marrakesh Protocol, and are an integral part of the GATT 1994. By the terms of the
Marrakesh Protocol, the Schedules are ‘Schedules to the GATT 1994’, and Article II:7 of
the GATT 1994 provides that ‘Schedules annexed to this Agreement are hereby made
an integral part of Part I of this Agreement’. With respect to concessions contained in
the Schedules annexed to the GATT 1947, the panel in United States—Restrictions on
Importation of Sugar (‘United States—Sugar Headnote’) found that:
. . . Article II permits contracting parties to incorporate into their Schedules acts
yielding rights under the General Agreement but not acts diminishing obligations
under that Agreement.
This principle is equally valid for the market access concessions and commitments for
agricultural products contained in the Schedules annexed to the GATT 1994. The ordin-
ary meaning of the term ‘concessions’ suggests that a Member may yield rights and grant
benefits, but it cannot diminish its obligations. This interpretation is confirmed by para-
graph 3 of the Marrakesh Protocol, which provides:
The implementation of the concessions and commitments contained in the sched-
ules annexed to this Protocol shall, upon request, be subject to multilateral examin-
ation by the Members. This would be without prejudice to the rights and obligations of
Members under Agreements in Annex 1A of the WTO Agreement. (emphasis added)
The question remains whether the provisions of the Agreement on Agriculture allow mar-
ket access concessions on agricultural products to deviate from Article XIII of the GATT
1994. The preamble of the Agreement on Agriculture states that it establishes ‘a basis for
initiating a process of reform of trade in agriculture’ and that this reform process ‘should
be initiated through the negotiation of commitments on support and protection and
through the establishment of strengthened and more operationally effective GATT rules
and disciplines’. The relationship between the provisions of the GATT 1994 and of the
Agreement on Agriculture is set out in Article 21.1 of the Agreement on Agriculture:
The provisions of GATT 1994 and of other Multilateral Trade Agreements in Annex
1A to the WTO Agreement shall apply subject to the provisions of this Agreement.
Therefore, the provisions of the GATT 1994, including Article XIII, apply to market
access commitments concerning agricultural products, except to the extent that the
Agreement on Agriculture contains specific provisions dealing specifically with the same
matter.
Article 4.1 of the Agreement on Agriculture provides as follows:
Market access concessions contained in Schedules relate to bindings and reductions
of tariffs, and to other market access commitments as specified therein.
In our view, Article 4.1 does more than merely indicate where market access conces-
sions and commitments for agricultural products are to be found. Article 4.1 acknow-
ledges that significant, new market access concessions, in the form of new bindings
and reductions of tariffs as well as other market access commitments (i.e. those made
as a result of the tariffication process), were made as a result of the Uruguay Round
 Tariff Protection in the GATT 91
negotiations on agriculture and included in Members’ GATT 1994 Schedules. These
concessions are fundamental to the agricultural reform process that is a fundamental
objective of the Agreement on Agriculture.
Th at said, we do not see anything in Article 4.1 to suggest that market access con-
cessions and commitments made as a result of the Uruguay Round negotiations on
agriculture can be inconsistent with the provisions of Article XIII of the GATT 1994.
There is nothing in Articles 4.1 or 4.2, or in any other article of the Agreement on
Agriculture, that deals specifically with the allocation of tariff quotas on agricultural
products. If the negotiators had intended to permit Members to act inconsistently with
Article XIII of the GATT 1994, they would have said so explicitly. The Agreement on
Agriculture contains several specific provisions dealing with the relationship between
articles of the Agreement on Agriculture and the GATT 1994. For example, Article 5 of
the Agreement on Agriculture allows Members to impose special safeguards measures
that would otherwise be inconsistent with Article XIX of the GATT 1994 and with
the Agreement on Safeguards. In addition, Article 13 of the Agreement on Agriculture
provides that, during the implementation period for that agreement, Members may
not bring dispute settlement actions under either Article XVI of the GATT 1994 or
Part III of the Agreement on Subsidies and Countervailing Measures for domestic sup-
port measures or export subsidy measures that conform fully with the provisions of the
Agreement on Agriculture. With these examples in mind, we believe it is significant that
Article 13 of the Agreement on Agriculture does not, by its terms, prevent dispute settle-
ment actions relating to the consistency of market access concessions for agricultural
products with Article XIII of the GATT 1994. As we have noted, the negotiators of the
Agreement on Agriculture did not hesitate to specify such limitations elsewhere in that
agreement; had they intended to do so with respect to Article XIII of the GATT 1994,
they could, and presumably would, have done so. We note further that the Agreement
on Agriculture makes no reference to the Modalities document or to any ‘common
understanding’ among the negotiators of the Agreement on Agriculture that the mar-
ket access commitments for agricultural products would not be subject to Article XIII
of the GATT 1994.
For these reasons, we agree with the Panel’s conclusion that the Agreement on
Agriculture does not permit the European Communities to act inconsistently with the
requirements of Article XIII of the GATT 1994. (original emphasis)
It is remarkable that the AB, a body with an exaggerated tendency to adopt text-
ual interpretations, this time paid lip-service (if at all) to the wording of Art 4.1
of the AG. It is equally remarkable that the AB did not even summarily discuss
the legal value of protocols of accession in this context: it refers to the Marrakesh
Protocol, but does not even discuss the legal value of the multilateral review of
annexed schedules. And, importantly, it did not clarify the legal value of the
Marrakesh Protocol itself. Instead, it moved to dismiss the claims presented by
the European Community based, for all practical purposes, on the Headnote jur-
isprudence. This jurisprudence, however, provides no response to the question
we asked: why do GATT schedules of concessions have to observe the GATT
disciplines, rather than sequencing the latter to the former? It provides simply an
assertion: they must do that. This reasoning is not satisfactory.
92 Disciplines on Trade Instruments

The outcome, however, although legally questionable, is not intellectually

indefensible. There are both pros and cons associated with the approach privi-
leged by the AB: removing the potential for discrimination helps reduce trade
diversion; on the other hand, of course, introducing flexibility is a good argument
in favour of inducing more meaningful commitments.
Were one to introduce a comprehensive legal review of all schedules at the end
of a negotiation, the process would be probably disproportionately burdened: in
the absence of centralized control of legality in the WTO, it would be essentially
incumbent upon WTO members to check each other’s schedule from a GATT
consistency perspective. Since Art 23.2 of the DSU has outlawed unilateral qual-
ifications of illegality, one could potentially imagine dozens of panels introduced
at this stage. This is not a mere theoretical possibility: recall that eight-digit level
classifications, the GATT consistency of which cannot be ex ante guaranteed,
are very much state practice today. In light of the above, it seems reasonable to
defend, from a policy-perspective at least, the (implicit) thesis advocated by the
AB: the multilateral review occurring with the exchange of schedules is limited to
verification of the accuracy of commitments; the multilateral review has no bearing
on their legality (GATT consistency).
The problem, of course, is that this approach assumes that there is no deal at
the end of a round, since part of the deal can be undone at a subsequent stage
through adjudication.⁹⁰ One could, for example, legitimately ask the question
whether the European Community would be equally satisfied with the outcome
of the Uruguay round if the Framework Agreement it had concluded with some
bananas exporters was not part of the package.
To a large extent, this discussion is now water under the bridge. Assuming
case-law progresses in a linear manner in this area, as has been the case with the
headnote and EC—Bananas III jurisprudence so far, WTO members have been
warned that the scheduling of concessions must be in conformity with the GATT
obligations, and especially, with the obligation not to discriminate.⁹¹ Essentially,
the AB, through EC—Bananas III, closed the door to MFN deviations that the
European Community was opening up: MFN deviations cannot be the subject-
matter of negotiations that lead to inter se agreements. Deviations from the MFN
principle will have to be based on provisions of the GATT. In essence, the AB
implicitly adopted a contextual reading of the term conditions appearing in Art II
of the GATT.⁹²

⁹⁰ This is in and of itself problematic. Arguably, one could make the argument that, Mexico
and the other complainants in EC—Bananas III, should be estopped from raising a challenge
against the Framework Agreement since they signed a deal which included this very instrument.
The estoppel-doctrine, however, has not found application in WTO law as yet, see Matsushita et al
(2006, pp 82ff ).
⁹¹ A legislative amendment to this effect would be most welcome and would settle this issue
once and for all.
⁹² The AB confirmed this ruling in its report on EC—Poultry (§§ 98–99).
 Tariff Protection in the GATT 93

3.5.3 The (non) impact of legitimate expectations

The LAN dispute (EC—Computer Equipment) between the United States and
the European Community involved a disagreement between the two parties as
to the proper classification of certain computer equipment. The particular com-
modity (LAN equipment) could, conceivably, come under two different HS
classifications; in light of the substantial discrepancy as to the tariff bindings
committed under each tariff classification by the European Community, the
United States had a strong trade interest in seeing it classified under the heading
with the lower import duty. To make matters even more complicated, there was
divergent practice with respect to the classification of LAN equipment among
individual European Community member states.⁹³ The panel was asked whether
the legitimate expectations of the WTO member which had negotiated the con-
cession (the United States) could have an impact on the clarification of the issue:
in the case at hand, the US (legitimate) expectations stemmed from the fact that,
at least some, EC member states had been practicing the lower import duties and
the associated belief that they would continue to do so. The panel agreed with the
United States, finding that legitimate expectations, in case of ambiguity, matter.
The AB however, in its report, rejected the panel’s interpretation. In its view, if at
all, it should be the legitimate expectations of the WTO membership in toto that
could be relevant. In the absence of such demonstration, the legitimate expecta-
tions of the WTO member negotiating a concession, are not determinative, in
case of ambiguity, as to what precisely has been negotiated. In its view, such a
standard would greatly undermine the security and predictability of the system;
it is the legitimate expectations of the whole WTO membership, and not of indi-
vidual Members that matters (§§ 80–96).⁹⁴

3.6 Changes in tariff protection after their consolidation

Once tariffs have been consolidated (bound), WTO members cannot impose
duties above their level: they can apply lower duties, but not higher than the
bound level.⁹⁵ Since all WTO members have to observe is the tariff ceiling they

⁹³ No claim by the United States was advanced, to the effect that such divergent practice might
be inconsistent with the statutory requirements for a GATT consistent CU.
⁹⁴ Th is decision, however, did not contribute to the resolution of the dispute. The AB did not
explain at all what, in cases like this where a quixotic test is used to establish the common inten-
tions of the WTO membership, is the appropriate legal criterion to resolve the ambiguity. It should
be noted nevertheless, that the parties to the dispute, by insisting that their dispute was not a tariff
classification issue (which it actually was), deprived the WTO adjudicating bodies from the possi-
bility to address the ‘heart’ of the claim. Th is is probably another good example underscoring the
need for closer cooperation between the WTO adjudicating bodies and the HS committee.
⁹⁵ There is often discrepancy (sometimes substantial) between bound and applied rates.
Governments might prefer to keep high bound rates in order to keep a negotiating chip in trade
negotiations. For a theoretical explanation on such motives, see Maggi and Rodriquez-Clare
(2006), where the two authors construct a model whereby governments enter into a trade agree-
ment in order to address terms of trade externalities and to face pressure by domestic lobbies.
94 Disciplines on Trade Instruments

have negotiated, they retain the power to switch from one type of tariff protec-
tion (say, specific duty) to another (say, ad valorem), provided that, by doing so,
they do not violate the negotiated ceiling of protection. This much has been con-
firmed in case-law (section 3.6.1). They can also withdraw concessions negotiated
with WTO members which have since withdrawn from the institution provided
that the requirements of Art XXVII of the GATT are being observed (section
3.6.2). WTO members retain the right to renegotiate their tariff protection, pro-
vided that the legislative requirements of Art XXVIII of the GATT have been
respected (section 3.6.3). Finally, WTO members cannot introduce unilateral
modifications to their schedule of concessions, unless if, when doing so, they
respect the decision concerning procedures for modification and rectification of
schedules of tariff concessions (section 3.6.4).

3.6.1 Switching between different types of duties

The question of whether WTO members can switch from one particular type of
import tariffs (say ad valorem) to another (say specific duties), following consoli-
dation of tariff protection, was the central focus of the AB report on Argentina—
Footwear. Argentina was imposing on imports of footwear either an ad valorem
duty of 35%, or a specific duty which was calculated on the basis of the world price.
It had, nevertheless, bound its duties on footwear, during the Uruguay round, as
ad valorem duties only. It later decided to revert back to the previously practised
form (that is, specific duties).⁹⁶ The complainant argued that such a change was
not permissible (essentially because, in its view, through binding, WTO mem-
bers agreed both not to impose tariffs beyond the ceiling imposed, and not to
change the type of duties as well). The panel upheld its claim. Overturning the
panel finding in this respect, the AB held the view that switching from one form
of protection to another is perfectly legitimate as long as the overall ceiling of pro-
tection is not violated (§§ 44–55). The AB insisted very much on this aspect, and
upheld Argentina’s right to switch. It did nevertheless, find that Argentina had
indeed violated its obligations since, following the conversion, the resulting duty
was higher than the negotiated ceiling.
The ruling by the AB risks being highly problematic, absent a monitoring
scheme. There are two distinct issues here. First, the timing issue: the AB refrained
from explaining when the conversion (from say, ad valorem to specific duties)
should take place in order to establish whether the ceiling has been respected.
Arguably, one could end up with different results if the conversion takes place

⁹⁶ Argentina did not feel that it had to follow the notification procedures included in the 1980
decision (see infra, section 3.6.4). In the absence of a complaint to this effect, the panel felt that it
did not have to deal with this issue either. As a result, it is not clear, as a matter of case-law, whether
switching between forms (types) of duties (say, from ad valorem to specific), requires WTO mem-
bers to formally notify the WTO about the change. The better arguments are in favour of an
affirmative response to this question in light of the potential that the value of the concession might
be affected through such a change.
 Tariff Protection in the GATT 95

at different points in time (say, when the concession was originally negotiated,
or when the litigation occurred). Second, when duties are bound at an ad val-
orem rate, a promise is given that a certain rate will be applied independently of
the price fluctuations. It is expected that to the extent prices decrease, imported
goods will be burdened less. Consequently, when duties are being negotiated and
an ad valorem duty is agreed, the eventual decreasing burden of the duty is some-
thing contracting partners will reasonably have in mind (and arguably, pay for,
when they reciprocally agree their own concessions). Specific duties on the other
hand, are delinked from the price of the imported item. As a result, the advan-
tage mentioned above (that is, the decreasing burden of the duty) is lost any time
duties are expressed in terms of specific duties. Allowing for conversion might
thus amount to undoing the balance of rights and obligations as struck by the
trading nations.⁹⁷

3.6.2 Withdrawing concessions from members leaving the WTO

GATT, Art XXVII reads:
Article XXVII
Withholding or Withdrawal of Concessions
Any contracting party shall at any time be free to withhold or to withdraw in whole
or in part any concession, provided for in the appropriate Schedule annexed to this
Agreement, in respect of which such contracting party determines that it was initially
negotiated with a government which has not become, or has ceased to be, a contract-
ing party. A contracting party taking such action shall notify the CONTRACTING
PARTIES and, upon request, consult with contracting parties which have a substantial
interest in the product concerned.
There has been some practice in this context. A very representative example is
offered by the United Kingdom/Palestine negotiations. Upon termination of the
League of Nations mandate of the United Kingdom with respect to Palestine, the
successor state could not be regarded as being bound by obligations under the
GATT. The Declaration adopted by the CONTRACTING PARTIES on 9 May
1949 provides:
Whereas the Government of the United Kingdom in the course of the negotiations
leading up to the drawing up of the General Agreement on Tariffs and Trade in 1947,

⁹⁷ This mode of calculating presupposes that a choice as to the date (as argued above) when the
calculation of the conversion is made. Were one to pick the date when the concession was made, it
will have to come up with a coefficient that will gradually be reducing the impact of the tariff. Such
calculation relies heavily on the assumption that prices are being gradually reduced. Whether this
has indeed been the case is of course a matter of empirical observation. Whether price reduction
was reasonably anticipated is a different, highly complicated issue. Such arguments would argue
against the AB’s decision and in favour of rigidity: no change of the type of duties should take place
unilaterally, as was the case in Argentina—Footwear. The easy way out would be to subject requests
to change the type of duties to the Art XXVIII of the GATT process on renegotiation of duties, see
infra section 3.6.3.
96 Disciplines on Trade Instruments
negotiated on behalf of the mandated territory of Palestine for concessions to be accorded
to products originating in such territory and for concessions to be accorded to the prod-
ucts of other contracting parties entering such territory, and
Whereas the Government of the United Kingdom ceased to be responsible for the man-
dated territory of Palestine on 15 May 1948,
The CONTRACTING PARTIES
Declare that, since the United Kingdom ceased, as from 15 May 1948, to be a contract-
ing party in respect of the territory formerly included in the Palestine mandate,

1. Section E shall be deemed to be no longer part of Schedule XIX; and

2. Any contracting party shall, in accordance with Article XXVII of the General
Agreement, be free to withhold or withdraw, in whole or in part, any concession pro-
vided for in the appropriate schedule annexed to the GATT which such contract-
ing party determines was initially negotiated with the United Kingdom on behalf of
Palestine, provided that the contracting party taking such action shall give notice to
all other contracting parties and, upon request, consult with the contracting parties
which have a substantial interest in the product concerned.⁹⁸
There have been other cases, namely:

Table 2.1

Concessions Granted by Initially Negotiated with Reference

Australia China, Syria/Lebanon, Philippines L/1266

Benelux: Section A— China, Syria/Lebanon, L/674
Metropolitan Territories Liberia, Philippines
Benelux: Section C— China L/658 and Add.1
Netherlands New Guinea
Canada China, Liberia, L/553
Philippines, Korea
Ceylon China L/1102, L/1505
Chile Colombia L/3191
Czechoslovakia Palestine GATT/CP/23
Finland China L/659
France China, Syria/Lebanon, L/460 and L/1269
Philippines
Germany, Fed Rep Philippines L/1264
India China, Colombia, Philippines G/77 and L/1430
Pakistan China L/1293
Sweden Colombia, Philippines, China L/950
United Kingdom China L/786
United States China GATT/CP/115 and Add.3
Uruguay China, Colombia L/1613

In recent years, there have been no cases concerning application of Art XXVII
of the GATT. Although there is a specific provision in the WTO Agreement

⁹⁸ See the discussion on Art XXVII of the GATT in the GATT Analytical Index.
 Tariff Protection in the GATT 97

regulating withdrawal from the WTO (Art XV), there are no reported cases of
withdrawal. The closest we came, was the ‘freezing’ of Yugoslavia’s participa-
tion. During the late eighties/early nineties, the Yugoslav Federation slowly col-
lapsed and new state entities appeared in the international scene. One of them,
the Federal Republic of Yugoslavia claimed to be the successor of the Federation,
and requested to be acknowledged as such, and preserve its GATT Contracting
Party status. However, other GATT Contracting Parties reacted and through
collective action ‘froze’ the participation of the Federal Republic of Yugoslavia in
the GATT:⁹⁹
1. Yugoslavia
—Status as a contracting party (L/7000, L/7002, L/7007, L/7008, L/7009, L/7022)
The Chairman said that the break-up of the former Socialist Federal Republic of
Yugoslavia had posed the question of its status as a contracting party. While the delega-
tion speaking in the name of the Federal Republic of Yugoslavia (FRY) had laid claim to
the status of successor to the former Socialist Federal Republic of Yugoslavia (L/7000),
this claim had been contested by some contracting parties and some others had reserved
their position on the issue. Some contracting parties had also suggested that the dele-
gation claiming to represent the FRY as a successor to the Socialist Federal Republic of
Yugoslavia (SFRY) in GATT should not participate in GATT activities until the FRY
had sought fresh membership, while others held the view that its participation should be
without prejudice to the FRY’s claim to successor status. He had held extensive informal
consultations with contracting parties and believed there was agreement that this issue
would need consideration by the Council. In these circumstances, without prejudice to
the question of who should succeed the former SFRY in the GATT, and until the Council
considered this issue, he proposed that the representative of the FRY should refrain from
participating in the business of the Council.
The Council so agreed.
What exactly was meant by the phrase ‘that the representative of the FRY should
refrain from participating in the business of the Council’ was made clear in sub-
sequent practice: the Federal Republic of Yugoslavia did not participate in the
General Council (GC) meetings. Later, with the advent of the WTO, Serbia and
Montenegro (a state entity corresponding geographically to the Federal Republic
of Yugoslavia) requested its accession anew to the WTO and a working party is
currently in place to this effect.¹⁰⁰

3.6.3 Renegotiating the tariff protection

A WTO member can increase its bound protection on a given item, provided
that the multilateral process included in Art XXVIII of the GATT has been

⁹⁹ See the discussion before the GATT Council in GATT DocC/M/257 of 10 July 1992.
¹⁰⁰ Subsequently, Montenegro seceded and the two states (Serbia, Montenegro) are currently
negotiating their accession to the WTO. In the meantime they have both obtained observer
status.
98 Disciplines on Trade Instruments

followed: in the typical case, the member wishing to raise its duties on a bound
item will negotiate and agree compensation (that is, a lowering of protection
on items other than the one where an increase of protection is being requested)
with a sub-set of the WTO membership that has been more severely affected by
the tariff change. The agreed compensation will be applied on an MFN basis.
According to Art XXVIII.2 of the GATT, the WTO members participating in
the renegotiation of the concession:
shall endeavour to maintain a general level of reciprocal and mutually advantageous con-
cessions not less favourable to trade than that provided for in this Agreement prior to
such negotiations.
GATT, Art XXVIII epitomizes the idea that the GATT is about maintaining a
level of reciprocally negotiated concessions: as long as WTO members stay within
a cooperative equilibrium, it is simply immaterial what the actual level of duties
is. In that, this provision is the single clearest pronouncement of reciprocity, one
of the cornerstones of the GATT edifice. Tariff adjustments might be needed
for a variety of reasons over time. For example, for political economy reasons, a
WTO member might be willing to redistribute wealth among its constituencies,
by overexposing some producers to international competition (this is what the
Art XXVIII of the GATT compensation essentially amounts to), while sheltering
others from it (this is what the modification of the level of tariff protection will
lead to). There are further good arguments supporting the view that, allowing
for this possibility by introducing flexibility in the GATT (in the form of Art
XXVIII of the GATT), will induce governments to make more meaningful tariff
concessions in the first place.¹⁰¹
GATT, Art XXVIII can be invoked independent of the occurrence of an
agreed contingency (as is for example the case in AD or safeguards). Indeed, for
an Art XXVIII of the GATT negotiation to be launched, political expediency
by and large suffices. Crucially, the initiative to launch this procedure rests solely
with the requesting state.
Apart from the requesting WTO member, a renegotiation will involve:
(1) WTO Members holding INRs;
(2) the WTO members that qualify as PSI members; and
(3) the WTO members having a substantial interest (SI).
The first two categories of countries have a right to participate in the negotiations.
INR-holders, PSIs and the applicant WTO member are defined in Art XXVIII
of the GATT as the primarily concerned members. SIs will be consulted but have
no legal right to participate in the negotiations.

¹⁰¹ See on this issue Bagwell and Staiger (2002), Finger (1993), Horn and Mavroidis (2003),
and Sykes (2003c).
 Tariff Protection in the GATT 99

In case no agreement is reached, the requesting WTO member will still be free
to increase its tariff protection. INR-holders, PSIs and SIs have, according to Art
XXVIII.3 of the GATT, the right to withdraw substantially equivalent conces-
sions. The text of the agreement leaves unanswered the question whether such
withdrawal of concessions will be on a bilateral or on an erga omnes basis. SIs, on
the other hand, have the legal right to withdraw substantially equivalent conces-
sions if they are not satisfied with the offer of the applicant WTO member, even
if the primarily concerned WTO members have reached an agreement between
them (Art XXVIII.3b of the GATT).

3.6.3.1 INR holders

A WTO member is an INR holder, if it has originally negotiated a specific con-
cession with another WTO member. The form of international negotiations how-
ever, can prejudge the manner in which INR holders will be identified. During
some trade rounds, concessions were negotiated and concluded on a bilateral basis
(and then, of course, applied on a MFN basis). In such cases, the identification of
the INR holder is a simple factual question: these are the fixed INRs. Fixed INRs
are sub-divided into historical INRs and current INRs. Take the example of a
concession on fax machines. Assume that during the Tokyo round, India binds
the duties on fax machines at 15%, as a result of a negotiation that it conducted
with Japan. During the Uruguay round, India agrees to reduce the duty on fax
machines to 10%. This time however, it negotiates the concession with Korea.
Japan is the historical INR holder and Korea is the current INR holder (until of
course, in the context of a new round, India agrees to further reduce its customs
duty on fax machines and then negotiates with a third country). Historical INR-
holders retain their legal right to participate in an Art XXVIII of the GATT
negotiation, only if the proposed tariff increase exceeds the level of the duty at its
historical level (eg in the example of fax machines, India requests, after the con-
clusion of the Uruguay round, an increase of its duty for fax machines at a level
higher than 15%). Otherwise, they will not be participating.
During some other trade rounds (eg the Kennedy round) concessions were not
negotiated bilaterally. Trading nations agreed tariff cuts in a multilateral manner
(eg they all agreed to reduce tariff protection on industrial goods by 50% inde-
pendently of the level of protection on these goods in each individual country).
In such cases, it is impossible to adopt the fixed INRs formula. Instead, the term
used in those cases is ‘floating’ INRs, defined in the same way as PSI countries
(see infra).
Because concessions may involve INRs at different levels of customs protection
for different WTO members, identification of INR holders is not always easy:
INR holders were previously identified in the bilateral agreements deposited with
the WTO Secretariat and in various informal working documents. However, it
has been the case that historical INRs have not always been incorporated into suc-
cessive schedules of concessions. To reduce uncertainty in this respect, the WTO
100 Disciplines on Trade Instruments

membership has undertaken a series of initiatives. For example, the Council for
Trade in Goods (CTG) adopted the Decision on the Establishment of Loose-
Leaf Schedules in November 1996 (WTO Doc G/L/138):
Each Member shall include in its schedule all INRs at the current bound rate. Other
Members may request the inclusion of any INR that had been granted to them. Historical
INRs different from the current bound rate not specifically identified shall remain valid
where a Member modifies its concession at a rate different from the rate at which the INR
was granted.
The situation has further improved with the finalization of the Consolidated
Tariff Schedules (CTS) Database which provides consolidated information on
the schedules of concessions of members.

3.6.3.2 PSI countries

There are two definitions of a PSI. First, the Interpretative Note ad Art XXVIII of
the GATT defines in point 4 under §1 a PSI as a member which:
has had, over a reasonable period of time prior to the negotiations, a larger share in the
market of the applicant contracting party than a contracting party with which the con-
cession was initially negotiated or would, in the judgment of the CONTRACTING
PARTIES, have had such a share in the absence of discriminatory quantitative restric-
tions by the applicant contracting party. (emphasis added)
Second, the same Interpretative Note ad Art XXVIII of the GATT in point 5
under §1 states that:
the CONTRACTING PARTIES may exceptionally determine that a contracting party
has a principal supplying interest if the concession in question affects trade which consti-
tutes a major part of the local exports of such contracting party. (emphasis added)
During the Uruguay round negotiations, an Understanding on the Interpretation of
Art XXVIII of the GATT was adopted which, inter alia, defines PSIs in § 1:

For the purposes of modification or withdrawal of a concession, the WTO member which has
the highest ratio of exports affected by the concession (i.e., exports of the product to the market of
the Member modifying or withdrawing the concession) to its total exports shall be deemed to
have a principal supplying interest if it does not already have an initial negotiating right or a prin-
cipal supplying interest as provided for in paragraph 1 of Article XXVIII. (emphasis added)
This clause, as made explicit in the understanding, would be reviewed within
5 years in order to check whether it has functioned in a satisfactory manner or,
whether it would have to be amended. An amendment would be required if it was
felt that the clause had not contributed to its objective function, which is defined
in the same understanding as follows:
. . . with a view to deciding whether this criterion has worked satisfactorily in securing
a redistribution of negotiating rights in favour of small and medium-sized exporting
Members.
 Tariff Protection in the GATT 101

This clause was discussed, as it was supposed to, in early 2000. In early 2000, the
CTG requested the committee on market access to undertake the review and the
committee in its report noted that:
. . . at this stage, there was no basis to change the criterion contained in paragraph 1 of the
aforementioned Understanding.¹⁰²
It stems from the discussions in the committee that no use of this possibility was
made between 1995–2000. However, developing countries were keen in keeping
this clause in place, in light of its potential use:
4.6 The Secretariat stated that according to information available to it, this criterion
had not been invoked officially. However, as Members were aware, Article XXVIII nego-
tiations took place in various ways and not everything that took place in these negotia-
tions was made known to the Secretariat.
4.7 The representative of Mexico stated that this provision concerned the interests of
small and medium-sized exporting Members, but in particular also of developing coun-
tries. An initial negotiating right could be granted in those cases where a country might
have an export product which represented a high percentage of its total exports. When
that product was affected by renegotiations by a third country, the exporting country
would have the right to declare itself as a country with principal supplying interests.
Taking into account the information from the Secretariat, namely that it was not aware
of an invocation of this criterion, his view would be that developing countries could look
into this matter and see whether there was some way of improving the criterion or not. In
fact, the review called upon Members to improve the criterion if necessary or if possible.
(original emphasis) ¹⁰³
This item was discussed again in the next committee meeting (WTO Doc
G/MA/M/24 of 20 July 2000). No action was taken and, as a result, the clause
was neither amended nor abolished.
The rationale for including PSIs in the negotiation is explained in the
Interpretative Note ad Art XXVIII of the GATT which in point 4 under para-
graph 1 relevantly provides:
The object of providing for the participation in the negotiation of any contracting party
with a principal supplying interest, in addition to any contracting party with which the
concession was originally negotiated, is to ensure that a contracting party with a larger
share in the trade affected by the concession than a contracting party with which the
concession was originally negotiated shall have an effective opportunity to protect the
contractual right which it enjoys under this Agreement.
The same document adds that:
It would . . . not be appropriate for the CONTRACTING PARTIES to determine that
more than one contracting party, or in those exceptional cases where there is near equal-
ity more than two contracting parties, had a principal supplying interest.

¹⁰² See WTO Doc G/MA/111.

¹⁰³ See WTO Doc G/MA/M/23 of 12 May 2000.
102 Disciplines on Trade Instruments

So, the inclusion of PSIs is justified on the grounds that they are the WTO mem-
bers that will suffer most from the change in tariff s. Limiting this to two mem-
bers has to do with the concern to facilitate the negotiating process; it is based on
the (valid) assumption that an outcome (agreement) is more likely the lower the
number of participants. The Interpretative Note ad Art XXVIII of the GATT
recognizes that much (§ 1.4):
On the other hand, it is not intended that the scope of the negotiations should be such as
to make negotiations and agreement under Article XXVIII unduly difficult nor to create
complications in the application of this Article in the future to concessions which result
from negotiations thereunder.
It follows that Art XXVIII of the GATT aims to strike a balance between the old
and the new market situation: the inclusion of INRs is justified on the grounds
that historically, at least they had a strong export interest in this particular mar-
ket. PSIs take a seat around the table because the legislator wanted to account for
the new market situation, and ensure that those which suffer the larger damage
are those who will be participating in the negotiated settlement.

3.6.3.3 SI countries
The Interpretative Note ad Art XXVIII of the GATT in point 7 under paragraph
1 states:
The expression ‘substantial interest’ is not capable of a precise definition and accordingly
may present difficulties for the CONTRACTING PARTIES. It is however, intended to
be construed to cover only those contracting parties which have, or in the absence of dis-
criminatory quantitative restrictions affecting their exports could reasonably be expected
to have a significant share in the market of the contracting party seeking to modify or
withdraw the concession.
In practice, WTO members having 10% or more of the market of the WTO
member seeking to modify the concession have been considered as having sub-
stantial interest in the concession. A report of the committee on tariff concessions
dating from July 1985 confirms that the 10% share has been generally applied for
the definition of SI countries.¹⁰⁴ Hence, SIs, contrary to PSIs, do not have to have
a market share larger than that of INRs in order to claim a right to participate in
the negotiation.

3.6.3.4 MFN trade is the basis for defining PSIs and SIs
The Uruguay round Understanding on the Interpretation of Art XXVIII of the
GATT states in § 3:
In the determination of which Members have a principal supplying interest . . . or sub-
stantial interest, only trade in the affected product which has taken place on a MFN basis

¹⁰⁴ See GATT Doc TAR/M/16.

 Tariff Protection in the GATT 103
shall be taken into consideration. However, trade in the affected product which has taken
place under non-contractual preferences shall also be taken into account if the trade in
question has ceased to benefit from such preferential treatment, thus becoming MFN
trade, at the time of the negotiation for the modification or withdrawal of the concession,
or will do so by the conclusion of that negotiation.

WTO members that have some form of preferential arrangement with the
requesting WTO member can thus never become PSIs or SIs.

3.6.3.5 New products

The Uruguay round Understanding on the Interpretation of Art XXVIII of the
GATT states in § 4:
When a tariff concession is modified or withdrawn on a new product (i.e., a product for
which three years’ trade statistics are not available) the Member possessing initial negotiat-
ing rights on the tariff line where the product is or was formerly classified shall be deemed
to have an initial negotiating right in the concession in question. The determination of
principal supplying and substantial interests and the calculation of compensation shall take
into account, inter alia, production capacity and investment in the affected product in the
exporting Member and estimates of export growth, as well as forecasts of demand for the
product in the importing Member. For the purposes of this paragraph, ‘new product’ is
understood to include a tariff item created by means of a breakout from existing tariff line.
(original emphasis)

So for new products, in the absence of trade statistics, recourse to other proxies
such as production capacity and investment will be made in order to identify PSI
and SI WTO members.

3.6.3.6 The mechanics of the negotiation

GATT, Art XXVIII distinguishes between three different forms of negotiation:
in two of them, the requesting WTO member (the applicant Contracting Party, in
the Art XXVIII of the GATT terminology) does not need to secure the approval
of the WTO membership before it enters into negotiations; in one, it does. The
common elements in all three negotiations are:

(1) the limited number of participants : in all three types of negotiation, it is the
primarily concerned members which will negotiate. The INRs and PSIs derive,
by virtue of their participation, an advantage: they can, in principle, deter-
mine the areas where tariff reductions will occur (promoting thus, their own
export interests). The requesting state will identify the commodity, the tariff
treatment of which it wishes to modify; the PSIs and INRs, either acting
separately or collectively, will identify the commodity (or list of commod-
ities) where compensation will be paid;
104 Disciplines on Trade Instruments

(2) there should be equivalence between the damage suffered because of the
modification, and the compensation offered: Art X XVIII.2 of the GATT
reads:
In such negotiations and agreement, which may include provision for compensatory
adjustment with respect to other products, the contracting parties concerned shall
endeavour to maintain a general level of reciprocal and mutually advantageous conces-
sions not less favourable to trade than that provided for in this Agreement prior to such
negotiations.
Moreover, the Interpretative Note ad Art XXVIII of the GATT reads in § 4.6:
It is not intended that provision for participation in the negotiations of any contracting
party with a principal supplying interest, and for consultation with any contracting party
having a substantial interest in the concession which the applicant contracting party is
seeking to modify or withdraw, should have the effect that it should have to pay compen-
sation or suffer retaliation greater than the withdrawal or modification sought, judged in
the light of the conditions of trade at the time of the proposed withdrawal or modifica-
tion, making allowance for any discriminatory quantitative restrictions maintained by
the applicant contracting party.
GATT, Art XXVIII distinguishes between procedures that require prior approval
by the CTG, and procedures that do not do so. We start from the latter.

(a) Procedures where no prior approval is required GATT, Art XXVIII dis-
tinguishes between two sub-categories. The first category is described in Art
XXVIII.1 of the GATT. In this case, the requesting WTO member, must initiate
negotiations during a specified period (July to October), in any 3-year-period
starting on 1 January 1958. The requesting WTO member will notify the CTG
of its interest to initiate negotiations, and the CTG will determine the identity of
the other primarily concerned members.¹⁰⁵
Assuming that, at the end of the negotiations, an agreement has been reached
between the participants, the requesting WTO member will notify its new
schedule of concessions to the WTO, which will be applied on an MFN basis. It
could however be that no agreement has been reached. In this case, the request-
ing WTO member can go ahead and unilaterally modify its concessions. If it
decides to exercise this option, the WTO member runs the risk of facing retali-
ation, not only from the members participating in the negotiation, but from the
rest of the WTO membership as well. GATT, Art XXVIII.3 relevantly reads in
this respect:
(a) If agreement between the contracting parties primarily concerned cannot be reached
before 1 January 1958 or before the expiration of a period envisaged in paragraph
1 of this Article, the contracting party which proposes to modify or withdraw the

¹⁰⁵ GATT, Art XXVIII states that it is the CONTRACTING PARTIES that will be entrusted
with this task. However, following the advent of the WTO, this task has been confined to the CTG.
 Tariff Protection in the GATT 105
concession shall, nevertheless, be free to do so and if such action is taken any con-
tracting party with which such concession was initially negotiated, any contracting
party determined under paragraph 1 to have a principal supplying interest and any
contracting party determined under paragraph 1 to have a substantial interest shall
then be free not later than six months after such action is taken, to withdraw, upon
the expiration of thirty days from the day on which written notice of such withdrawal
is received by the CONTRACTING PARTIES, substantially equivalent conces-
sions initially negotiated with the applicant contracting party.
(b) If agreement between the contracting parties primarily concerned is reached but any
other contracting party determined under paragraph 1 of this Article to have a sub-
stantial interest is not satisfied, such other contracting party shall be free, not later
than six months after action under such agreement is taken, to withdraw, upon the
expiration of thirty days from the day on which written notice of such withdrawal is
received by the CONTRACTING PARTIES, substantially equivalent concessions
initially negotiated with the applicant contracting party.
This paragraph suggests that two categories of WTO members can react (in case
the requesting WTO member has decided to modify its concession, notwith-
standing the collapse of negotiations): both the participants in the negotiations
(PSIs and INRs), and the SIs can withdraw substantially equivalent concessions,
provided that they do so on goods initially negotiated with the requesting WTO
member. There are still, however, important questions left open:
(1) What if the two categories of WTO members have no concessions initially
negotiated with the requesting WTO member? Should they lose their right
to retaliate in such a case?
(2) Should the retaliation by those entitled to it, be on a bilateral or on an erga
omnes basis?
(3) It could be the case that WTO members, other than INRs, PSIs or SIs, are
also affected by the unilateral modification of the concession. Should they
not be entitled to react?
Let us take each question in turn. GATT/WTO practice¹⁰⁶ seems to suggest
that the requirement under (1) above, can prove to be highly problematic in prac-
tice. Indeed, the number of WTO members nowadays makes it quite likely¹⁰⁷
that affected primarily concerned members and/or SIs have no initially negotiated
concessions with the requesting WTO member. Practice (like the Canada/EC
dispute, discussed infra), seems to suggest that retaliating WTO members indi-
cate the goods where they purport to increase tariffs in retaliation, irrespective
of whether they have initially negotiated concessions on these goods with the

¹⁰⁶ See Hoda (2001) for a comprehensive, and highly informative, study on Art XXVIII of the
GATT negotiations.
¹⁰⁷ It is, of course, to be expected that, the smaller the number of participants, the higher the
likelihood that PSIs and/or SIs are also INRs.
106 Disciplines on Trade Instruments

applicant state or not. Although such practice runs counter the explicit wording
of Art XXVIII of the GATT, no formal challenge against it by affected WTO
members has taken place as yet. Practice suggests that this condition has, de
facto, been relaxed.¹⁰⁸
With respect to the second question: although Art XXVIII of the GATT is
not explicit in this respect, it should be the case that retaliating WTO members
should do so only on a bilateral basis, that is, they should be barred from raising
their tariffs in retaliation erga omnes. Otherwise, we might end up with a coun-
ter-retaliation by affected innocent states, counter-counter-retaliation, and so on
and so forth. However, although probably counter-intuitively so, practice suggests
a different path: there are at least two traceable instances where the WTO member
reacting to a unilateral modification, threatened to do so on an erga omnes basis.
First, the 1990 award by the Arbitrator on Canada/European Communities—Article
XXVIII rights, notes with respect to Canada’s threat to withdraw concessions sub-
stantially equivalent to those modified by the European Community, that:
should Canada exercise her right to withdraw concessions, she undertakes obligations to
compensate third countries having negotiating rights in respect of Canada for the prod-
ucts on which concessions would be withdrawn.¹⁰⁹
Canada of course would incur no obligations to compensate anyone, had its retali-
ation taken place on a bilateral basis. Implicitly hence, this passage suggests that
Canada would be raising duties, not only vis-à-vis the European Community,
but also vis-à-vis the rest of its trading partners (on an MFN basis). This view has
been confirmed in subsequent practice. Canada and the European Community
run into the same argument in the context of the European Community’s sub-
sequent enlargement (with the accession of Austria, Finland, and Sweden). The
three acceding countries had to raise their duties with respect to some products,
from their prior unilateral level to the new harmonized, European Community
level. According to Art XXIV.6 of the GATT (see infra), in such cases, the mem-
bers of the preferential trade agreement, and the outsiders, enter into Art XXVIII
of the GATT negotiations the subject-matter of which is to determine the level of
compensation due to outsiders. When discussing the amount of compensation,
‘built-in’ compensation, that is, compensation already paid since the acceding
countries had to, with respect to certain goods, lower their tariffs to meet the CU,
will be taken into account. This is where Canada and the European Community
disagreed: the former believed that the ‘built-in’ compensation was not adequate;
the latter believed that the opposite was the case. Canada threatened to retaliate:
As a result of tariff modifications which became effective January 1, 1995, the access of
Canadian exporters to the acceding countries has been impaired. Consequently, Canada

¹⁰⁸ Assuming, however, a legal challenge against such practice it is hard to imagine how a WTO
panel will neglect the explicit wording of Art XXVIII of the GATT in this respect.
¹⁰⁹ See GATT Analytical Index, p 947.
 Tariff Protection in the GATT 107
wishes to notify Members that it will exercise its rights under Article XXVIII:3 to withdraw
substantially equivalent tariff concessions on products of interest to the European Union
as outlined in the attached table. These modifications shall take effect 30 days after distri-
bution of this notification to WTO members.
Members may wish to note that in this respect, these modifications will not affect the
application of rates under the Generalized System of Preferences.
Any Member that believes it has supplier rights which are affected by this action is
invited to inform the Permanent Representative of Canada to the WTO.¹¹⁰
The last quoted sentence indicates that Canada’s purported retaliation was sup-
posed to take place on an erga omnes basis. Although Canada targeted prod-
ucts originating in the European Community, a negative trade impact on other
WTO members producing the same goods could not be avoided. The parties to
the dispute finally agreed to a settlement, and as a result, Canada never enacted
its threat.¹¹¹ This practice suggests that, for some WTO members at least, erga
omnes retaliation as response to unilateral modifications is very much on the
cards. This is regrettable. From a pure policy perspective, erga omnes retaliation
might lead to an endless circle of counter-retaliation by WTO members. The
incentive of the retaliating member is, absent agreed compensation, to overshoot
the amount of suspended concessions, in order to provide the applicant state with
an incentive not to modify unilaterally the concession in the first place. There is
no mechanism to ex ante ensure that retaliation will be kept within the statutory
limits.¹¹² Affected parties are between a rock and a hard place: if they choose the
legal route, they will have to challenge the retaliation before a panel (the subject-
matter of which will be to review whether suspended concessions have indeed
been substantially equivalent to the trade damage resulting from the modifica-
tion of the concession); if they decide to act illegally, and adopt counter-retaliatory
measures, they open up themselves to counter-counter-retaliation or legal chal-
lenges, as the case may be, and so on and so forth.¹¹³

¹¹⁰ See WTO Doc G/SECRET/1 of 1 March 1995.

¹¹¹ See WTO Docs G/SECRET/1 Add. 1 of 7 March 1995, and Add 2 of 5 February 1998.
¹¹² Th is term (‘substantially equivalent concessions’) appears in Art 22.4 of the DSU as well.
According to this provision, retaliation must be substantially equivalent to the damage suffered
by the commission of the illegal act. The working hypothesis of this provision is the commission of
an illegality. It is difficult to sustain that a unilateral modification of a concession, the absence of
agreed compensation notwithstanding, amounts to an illegality: Art XXVIII of the GATT allows
WTO members to do it. So one can only wonder as to the rationale behind introducing the same
standard of compensation, to address two diametrically opposed situations. Still, in light of the
identical wording across the two provisions, one would expect to see the case-law under Art 22.4
of the DSU to be of direct relevance to the calculation of compensation under Art XXVIII of the
GATT.
¹¹³ The calculation of compensation can be very problematic. Let us stick to the Canada/EC
dispute, and assume that Canada includes cars in its list and that Korea produces steel and cars.
Korea will see its market share in Canada’s cars import market drop if its steel gets exported to the
European Community. This loss cannot even in part be compensated since, if at all, trade diver-
sion will favour Canadian cars (since the Canadian retaliatory tariff is MFN and thus hits Korean
cars as well). The Arbitrator’s report on EC—Bananas III (Ecuador) (Article 22.6—EC) refused to
108 Disciplines on Trade Instruments

On the other hand, from a purely legal perspective, it is at least debatable that
such practice is GATT consistent: assume that Canada carried out its threat and
suspends concessions erga omnes. In this scenario, innocent bystanders have to pay
the price because the European Community has unilaterally modified its conces-
sion. This is however, an outright illegality (a violation of Art II of the GATT)
and it seems highly unlikely that Art XXVIII of the GATT was intended to con-
done such practices.¹¹⁴ Finally, from a purely practical perspective, the current
institutional framework is ill-equipped to deal with the scenario of MFN retali-
ation: assume that Canada retaliates on an MFN basis, and that it then invites
affected parties to negotiate their compensation with it. Under what GATT pro-
vision would such negotiations take place? Arguably, Art XXVIII of the GATT
is not applicable in this context since none of the three statutory procedures for
renegotiation applies here. Hence negotiations, if at all, will take place within an
institutional void. One way to avoid trouble, would be to amend the existing Art
XXVIII of the GATT by introducing two amendments to the current text:
(1) first, that in case of no agreement, retaliation, if at all, shall be bilateral and
not erga omnes; and
(2) the amount of retaliation, in the case of disagreement, shall be fi xed through
recourse to binding arbitration (as per Art 22.6 of the DSU).¹¹⁵

entertain claims on indirect benefits: the United States claimed that it should be compensated for
lost profits resulting from the EC bananas import regime. In its view, the European Community, by
blocking imports into its market of bananas originating in Mexico, was ipso facto blocking exports
to the Mexican market of fertilizers originating in the United States. In other words, in the US
view, Mexico would have little need for US fertilizers in light of the reduced export opportunities
of bananas to the EC market. The Arbitrators decided against the US claim in this respect. In their
view, the European Community could be held liable for trade in bananas lost by Mexican exporters,
but not for trade in fertilizers lost by US exporters as a result of Mexico’s decision to reduce imports
of the said commodity. DSU, Art 22 consequently, must be construed so as to disallow the inclusion
of indirect benefits, when calculating the amount of countermeasures. To what extent this case-law
will be followed in the context of Art XXVIII of the GATT is an open question.
¹¹⁴ One could make the argument that, facing the prospect of an MFN retaliation would pro-
vide WTO members with an incentive not to modify unilaterally in the first place: if they did, they
would be credited with spiraling countermeasures across the board. Consequently, in light of the
reputation costs that they would suffer, WTO members, by backwards induction would rather
refrain from modifying unilaterally. Such arguments are, nonetheless, unpersuasive: for one, the
text of Art XXVIII of the GATT itself states that they remain free to modify unilaterally, hence,
they are in fact exercising their rights under the GATT. On the other hand, it is difficult to sustain
that the WTO is a relational contract where reputation costs matter. Th is was probably the case
in the original GATT, it is hardly the case anymore.
¹¹⁵ Voices of this type have already been raised. During the Uruguay round negotiations,
Switzerland, in a communication to the GATT (GATT Doc MTN.GNG/NG7/W/65 of
23 December 1989) first observed that a consensus was emerging that retaliation could indeed be
erga omnes, and then went on to propose ‘the introduction of an interpretative note to paragraph 3
of Article XXVIII explicitly authorizing contracting parties to take retaliatory action on a bilateral
basis against the contracting party which originally withdraws its concession. To avoid possible
misuse, the implementation of retaliatory measures should be subject to the prior approval of the
CONTRACTING PARTIES.’ Its proposal, unfortunately, was not met with enthusiasm by its
trading partners.
 Tariff Protection in the GATT 109

What about the rights of WTO members which do not belong to any of the two
categories envisaged in Art XXVIII.3 of the GATT? The current text does not
explicitly acknowledge that they have a legal right to retaliate under this provi-
sion. On the other hand, they will, in all likelihood, be affected by a modifica-
tion of the schedule, irrespective of whether an agreement between the primarily
concerned parties has been reached, or not. What can be done if such an occasion
arises? The most reasonable way out would be to acknowledge to such WTO
members the right to take a non-violation complaint (NVC, see Chapter 5)¹¹⁶
against the requesting WTO member. This solution seems warranted, in light of
the explicit acknowledgement in Art XXVIII.3 of the GATT, that the request-
ing WTO member has the right unilaterally to modify its schedule of conces-
sions, even in the absence of agreed compensation; hence, such behaviour cannot
be deemed to be illegal. The only question remaining in this context is whether
the right to file an NVC can be exercised irrespective of whether an agreement
between the primarily concerned parties has been reached. In the absence of rele-
vant practice, we tend to respond to this question in the affirmative. From the
perspective of the affected state, it could be totally immaterial whether an agree-
ment occurred or not.¹¹⁷
Lastly, recall that SIs are free to react, if they are not happy with the outcome
of the negotiations, regardless of whether an agreement between the primarily
concerned Members has been successfully negotiated or not.
The second category of procedures where no prior approval is required is
described in Art XXVIII.5 of the GATT: WTO members can by virtue of this
provision, reserve their right to re-negotiate, and eventually exercise this right at
a later date. It reads:
Before 1 January 1958 and before the end of any period envisaged in paragraph 1 a con-
tracting party may elect by notifying the CONTRACTING PARTIES to reserve the
right, for the duration of the next period, to modify the appropriate Schedule in accord-
ance with the procedures of paragraph 1 to 3. If a contracting party so elects, other con-
tracting parties shall have the right, during the same period, to modify or withdraw,
in accordance with the same procedures, concessions initially negotiated with that con-
tracting party.
Practice in this context constitutes the majority of Art XXVIII of the GATT
negotiations. The discussion above is mutatis mutandis applicable in this context.

(b) Procedures where prior approval is required The procedure explained supra
has one important downside: the right can be exercised only within a particular
time-period. WTO members which have not reserved their right to renegotiate,

¹¹⁶ Such complaints are possible, when benefits to trading partners are being nullified or impaired
as a result of GATT consistent behaviour.
¹¹⁷ It is immaterial, if compensation, for example, is paid in commodities of no interest to the
state at hand.
110 Disciplines on Trade Instruments

or who wish to negotiate outside the period prescribed in Art XXVIII.1 of

the GATT, can do so only if they have first secured the authorization of the
WTO membership (following the procedure included in Art XXVIII.4 of the
GATT). To this effect, the WTO member concerned will submit its request
to the CTG, the competent organ to decide. A request for authorization must
include the elements specified in the Interpretative Note ad Art XXVIII of the
GATT (§ 4.1):
Any request for authorization to enter into negotiations shall be accompanied by all rele-
vant statistical and other data. A decision on such request shall be made within thirty
days of its submission.
The time-period within which the negotiation must be completed should be short
(60 days), but could be extended. The Interpretative Note ad Art XXVIII of the
GATT relevantly provides in this respect (§ 4.3):
It is expected that negotiations authorized under paragraph 4 for modification or with-
drawal of a single item, or a very small group of items, could normally be brought to a
conclusion in sixty days. It is recognized, however, that such a period will be inadequate
for cases involving negotiations for the modification or withdrawal of a larger number
of items and in such cases, therefore, it would be appropriate for the CONTRACTING
PARTIES to prescribe a longer period.
Assuming that negotiations have been successfully concluded, the new schedule
of concessions notified by the requesting WTO member will enter into force.
It could, however, be the case that no agreement can be reached. In this case,
according to Art XXVIII.4 of the GATT:
the applicant contracting party shall be free to modify or withdraw the concession, unless
the CONTRACTING PARTIES determine that the applicant contracting party has
unreasonably failed to offer adequate compensation.
Hence, in case of disagreement between the negotiating partners, the CTG will
determine whether adequate compensation has been offered. The Interpretative
Note ad Art XXVIII of the GATT states the deadline by which the CTG must
decide on this issue (§ 4.4):
The determination referred to in paragraph 4 (d) shall be made by the CONTRACTING
PARTIES within thirty days of the submission of the matter to them unless the applicant
contracting party agrees to a longer period.
The Interpretative Note ad Art XXVIII of the GATT also provides some useful
information as to the elements that the CTG should take into account, when
determining whether adequate compensation had indeed been offered (§ 4.5):
In determining under paragraph 4 (d) whether an applicant contracting party has unrea-
sonably failed to offer adequate compensation, it is understood that the CONTRACTING
PARTIES will take due account of the special position of a contracting party which has
 Tariff Protection in the GATT 111
bound a high proportion of its tariffs at very low rates of duty and to this extent has less
scope than other contracting parties to make compensatory adjustment.
If the CTG determines that adequate compensation has indeed been offered, the
modified concession will be allowed to stand. In the opposite case, the applicant
WTO member might still go ahead and unilaterally modify the concession. In
this case, INRs, PSIs, and SIs have the right to suspend substantially equivalent
concessions. GATT, Art XXVIII.4 reads:
If such action is taken, any contracting party with which the concession was initially
negotiated, any contracting party determined under paragraph 4 (a) to have a principal
supplying interest and any contracting party determined under paragraph 4 (a) to have a
substantial interest, shall be free, not later than six months after such action is taken, to
modify or withdraw, upon the expiration of thirty days from the day on which written
notice of such withdrawal is received by the CONTRACTING PARTIES, substantially
equivalent concessions initially negotiated with applicant contracting party.
In light of the rule that:
(1) an authorization is required before the applicant WTO members enters into
an Art XXVIII.4 of the GATT type negotiation;
(2) the authorization will be granted following a consensus decision¹¹⁸ to this
effect; and
(3) the agreement of potential INRs, PSIs, SIs is essential, indeed the conditio
sine qua non for a favourable decision,
it is to be expected that the negotiation on the level of compensation, in such
cases, effectively takes place before the authorization has been granted. This
might explain why some requests are being withdrawn (in the absence of agree-
ment on the compensation).¹¹⁹
For SIs, as noted above, an agreement between the primarily concerned parties
on the compensation to be paid can be totally irrelevant: they can go ahead and
suspend substantially equivalent concessions if they do not find the agreement
satisfactory.¹²⁰ With respect to WTO members who are neither INRs or PSIs, nor
SIs, the NVC seems to be the only viable route. Here too, the presence or absence
of agreement between the primarily concerned parties is simply immaterial, since
they could be negatively affected even when an agreement has been reached.

¹¹⁸ As explained in detail in Chapter 5, in practice, all GATT decisions, with the exception of
waivers, are taken by consensus.
¹¹⁹ Grenada for example, withdrew its request, arguing that it could not be present during
the discussions (see WTO Doc G/C/M/59 of 22 March 2002). Subsequently, it re-introduced its
request and managed successfully to conclude the negotiations (see WTO Doc G/SECRET/16
and Add. 1).
¹²⁰ This could be the case if, for example, the compensation is paid on items which are not of
export interest to the SI. Assuming the suspension of concessions by the SI is erga omnes, affected
parties will be invited to negotiate their compensation.
112 Disciplines on Trade Instruments

3.6.4 Rectifications and modifications of schedules

Until 1959 changes on the tariff level were incorporated into the GATT and its
schedules through a series of protocols. There were five protocols of rectification in
total, one protocol of modification, and nine protocols of rectification and modi-
fication: the term ‘rectification’ essentially captures mundane changes (where, for
example, the scope of the concession is not being affected), whereas the term ‘modi-
fication’ refers to substantive changes, such as those resulting from an Art XXVIII
of the GATT renegotiation of duties. In 1959, the CONTRACTING PARTIES
agreed to adopt a procedure of certification and discontinue the practice of pre-
paring protocols of rectifications and modifications, subject to the revised text of
Art XXX GATT being accepted. The protocol amending Art XXX of the GATT
lapsed in 1967, and in November 1968 the CONTRACTING PARTIES never-
theless decided to establish certification procedures for modification and rectifi-
cation.¹²¹ Six collective certifications were adopted under this decision. Another
decision was taken in 1980¹²² to revise the procedures for rectification and modi-
fication whereby a system of individual certifications was introduced. Although
the 1980 Decision refers to modifications resulting from actions under specified
Articles, in practice these procedures have also been used to include schedules of
unilateral commitments made pursuant to, for example, the ITA. In the light of
prior practice, the 1980 Decision¹²³ distinguishes between two situations:
(1) modifications;¹²⁴
(2) rectifications.
Modifications imply substantive change of the concession (§1 of the 1980
Decision), whereas rectifications imply no such change (§2 of the 1980 Decision).
We discussed above the, in practice, most important procedure for modifica-
tions (Art XXVIII of the GATT).¹²⁵ Rectifications of schedule involve unilateral
action, whereby a WTO member sends a new description of its concession. Recall
that some modifications might not be the outcome of an agreement (a WTO
member can and might modify its concession even if it has reached no agreement
with the affected parties). By the same token, a WTO member might, under the
guise of rectification, actually hide a modification. Some multilateral review was
thus deemed necessary, and this is precisely what the 1980 Decision does. Both

¹²¹ See GATT Doc BISD 16S/16–17.

¹²² Decision of 26 March 1980, L/4962, see GATT Doc BISD 27S/25–26, hereinafter the 1980
Decision.
¹²³ Although there has never been an official pronouncement to this effect, there should be no
doubt that this decision is part of GATT 1994 by virtue of Art 1(b)(iv).
¹²⁴ GATT, Art XXVIII negotiations come under this procedure. So do modifications under
Art II, Art XVIII (BoP), Art XXIV, and Art XXVII of the GATT.
¹²⁵ As we will see infra, Art XXIV of the GATT negotiations come under the purview of Art
XXVIII of the GATT, since the members of say a CU will negotiate with INRs, and PSIs the com-
pensation to be paid in case duties at the CU-level are higher (for at least one member of the CU)
than the corresponding duties before the formation of the CU.
 Tariff Protection in the GATT 113

modifications and rectifications are certified and communicated to the Director-

General of the WTO (the former within 3, the latter within 6 months). The new
(modified or rectified) schedules will be definitive if no objection has been raised
by a WTO member within 3 months (§ 3 of the 1980 Decision).
For reasons already explained supra (EC—Bananas III), it seems safe to con-
clude that consenting to a modified schedule means that the WTO membership
has conceded that the schedule at hand is accurate, but it does not necessarily also
mean that it has consented to its legality. WTO members have not, by virtue of
the 1980 Decision, given up their rights to challenge the consistency of modified
(or rectified) schedules with the multilateral rules. Indeed, this is exactly what
happened in EC—Bananas III, where a host of trading nations challenged the
certified EC schedule, which accurately described the concessions made by the
European Community; its legality, however, was a matter of contention.
This issue arose for the first time, during the proceedings that led to the GATT
panel report on Spain—Un-roasted Coff ee. Spain had originally made a tariff
concession on coffee. It then modified its negotiated concession by adding in its
schedule a distinction between roasted and un-roasted coffee, and, as a result,
provided different tariff treatment to the two categories of coffee. Brazil com-
plained since, as a result of this intervention, its coffee exports to the Spanish
market, were being negatively affected. The panel outlawed the Spanish measure
(§ 4.4):
The Panel found that there was no obligation under the GATT to follow any particular
system for classifying goods, and that a contracting party had the right to introduce in its
customs tariff new positions or sub-positions as appropriate. ‘A footnote to this paragraph
reads: Provided that a reclassification subsequent to the making of a concession under
the GATT would not be a violation of the basic commitment regarding that concession
(Article II:5).’ (emphasis added)
The value of a concession, however, can be negatively affected even if the duties
are set at a lower than the bound level. Assume that the litigation between Spain
and Brazil occurs today, and assume further that the entry coff ee appears at the
six-digit level, and that Spain has bound it at 5%.¹²⁶ Now Brazil starts exporting
roasted coffee paying a 5% import duty. Spain introduces an eight-digit classi-
fication whereby roasted coffee continues to be subjected to a 5% duty, whereas
un-roasted coffee, the new category, pays only 2%. Assuming a substitution
effect (in this case, consumers view, in principle, roasted and un-roasted coffee
as interchangeable products) because of the price difference, one would observe
trade diversion from Brazil to the countries exporting un-roasted coffee to Spain.
So Spain now formally abides by its Art II obligation, but de facto has conferred
an advantage to a sub-set of the WTO membership (from which it might have

¹²⁶ In today’s world, it would of course be the European Community, Spain being a member
of it.
114 Disciplines on Trade Instruments

extracted an advantage). Crucially, Brazil has suffered trade damage while it has
paid Spain a reciprocal concession in order to extract the promise on coffee. In
cases like this, Brazil can anyway raise a non-violation complaint and request
compensation.¹²⁷ It might also succeed in a violation complaint, assuming it can
prove that roasted and un-roasted coffee are like products.¹²⁸ If Spain introduces
a new sub-division and imposes duties higher than 5%, then it would of course be
violating Art II of the GATT.¹²⁹

3.7 Fees and charges for services rendered

In what has preceded we have discussed the tariff promise of WTO members.
OCDs, and ODCs, however, do not exhaust the burdens on imported products
upon importation. WTO members remain free to adopt additional measures
related to importation of goods, to the extent that they respect the requirements
of Art VIII of the GATT. Measures such as statistical taxes, or public health-
related controls may come under the purview of this provision.
GATT, Art VIII reads:
Article VIII
Fees and Formalities connected with Importation and Exportation
1. (a) All fees and charges of whatever character (other than import and export duties
and other than taxes within the purview of Article III) imposed by contracting
parties on or in connection with importation or exportation shall be limited in
amount to the approximate cost of services rendered and shall not represent an
indirect protection to domestic products or a taxation of imports or exports for
fiscal purposes.
(b) The contracting parties recognize the need for reducing the number and diversity
of fees and charges referred to in subparagraph (a).
(c) The contracting parties also recognize the need for minimizing the incidence and
complexity of import and export formalities and for decreasing and simplifying
import and export documentation requirements.
2. A contracting party shall, upon request by another contracting party or by the
CONTRACTING PARTIES, review the operation of its laws and regulations in the
light of the provisions of this Article.
3. No contracting party shall impose substantial penalties for minor breaches of cus-
toms regulations or procedural requirements. In particular, no penalty in respect of
any omission or mistake in customs documentation which is easily rectifiable and

¹²⁷ See Chapter 6.

¹²⁸ See the discussion on MFN, infra in this chapter.
¹²⁹ A similar to the Spain—Un-roasted Coff ee case arose recently: Panama eliminated from its
schedule item 1901.10.10 (modified milk) to which it applied a 5% duty. It then created two new
tariff items: 1901.10.11 (infant milk formula) with an import tariff of 0%, and 1901.10.19 (other)
with an import tariff of 65%. Mexico complained since, in its view, this unilateral modification
violated, inter alia, Art II of the GATT (WTO Doc WT/DS329/1). The parties to the dispute
reached a mutually agreed solution (MAS) that they notified to the WTO, whereby Panama agreed
to reduce the tariff on item 1901.10.19 to 5% (WTO Doc WT/DS329/2).
 Tariff Protection in the GATT 115
obviously made without fraudulent intent or gross negligence shall be greater than
necessary to serve merely as a warning.
4. The provisions of this Article shall extend to fees, charges, formalities and require-
ments imposed by governmental authorities in connection with importation and
exportation, including those relating to:
(a) consular transactions, such as consular invoices and certificates;
(b) quantitative restrictions;
(c) licensing;
(d) exchange control;
(e) statistical services;
(f ) documents, documentation and certification;
(g) analysis and inspection; and
(h) quarantine, sanitation and fumigation.
GATT, Art II.2(c) allows WTO members to impose:
fees or other charges commensurate with the cost of services rendered.
It thus opens the door to the possibility of imposing such fees. GATT, Art
VIII regulates in more detail the conditions under which they will be imposed.
Impositions under Art VIII of the GATT should not overlap with ODCs. The
panel, in its report on Dominican Republic—Import and Sale of Cigarettes con-
firmed this point of view. When discussing the consistency of a foreign exchange
fee imposed by the Dominican Republic with the GATT, it first attempted to
classify it under one of the possible categories of impositions featuring in the
GATT. In this process, it made clear that it understood Art VIII of the GATT
impositions as distinct from ODCs in the following terms (§ 7.115):
The foreign exchange fee is imposed on imported products only and it is not an ordinary
customs duty. It is computed on the value of imports, not on the cost of the services ren-
dered by the customs authorities. Consequently, it is not a fee or charge that falls under
Article VIII of the GATT. It is obviously not an anti-dumping or countervailing duty.
Therefore, it is a border measure in the nature of an ODC within the meaning of Article
II:1(b).¹³⁰
In practice, however, there are reported cases where it is difficult to establish an
absolute ‘firewall’ between ODCs and measures coming under Art VIII of the
GATT: Honduras bound all its customs duties during the Uruguay round at 35%
ad valorem. It also included an additional 3% ODC. In its schedule, Honduras
explains that the ODC will be reduced to 1.5% by 31 December 1995 and then
further modified in order to comply with the requirements of Art VIII of the
GATT. So an ODC will be transformed to an Art VIII of the GATT imposition.
There are other similar cases which, as this example shows, are of transitional
value.

¹³⁰ This finding was not appealed.

116 Disciplines on Trade Instruments

The GATT panel report US—Customs User Fee provided a series of clarifica-
tions on the ambit of Art VIII of the GATT. The facts of the case that are crucial
to our study are summarized in § 7 of the report:
The term ‘customs user fee’ refers to a number of fees imposed by the United States for
the processing by the US Customs Service of passengers, conveyances and merchandise
entering the United States. Only one of these fees is at issue in this dispute. It is the ‘mer-
chandise processing fee,’ an ad valorem charge imposed for the processing of commercial
merchandise entering the United States. (original emphasis)
The panel went on to explain that the services rendered, for which a charge will be
imposed, do not have to be requested by the traders (§ 77):
In referring to these customs-related government activities as ‘services rendered,’ the
drafters of Articles II and VIII were clearly not employing the term ‘services’ in the eco-
nomic sense. Granted that some government regulatory activities can be considered as
‘services’ in an economic sense when they endow goods with safety or quality character-
istics deemed necessary for commerce, most of the activities that governments perform
in connection with the importation process do not meet that definition. They are not
desired by the importers who are subject to them. Nor do they add value to the goods
in any commercial sense. Whatever governments may choose to call them, fees for such
government regulatory activities are, in the Panel’s view, simply taxes on imports. It must
be presumed, therefore, that the drafters meant the term ‘services’ to be used in a more art-
ful political sense, i.e., government activities closely enough connected to the processes of
customs entry that they might, with no more than the customary artistic licence accorded
to taxing authorities, be called a ‘service’ to the importer in question. No other interpret-
ation can make Articles II:2(c) and VIII:1(a) conform to their generally accepted meaning.
On the other hand, the panel also explained that the service rendered must be
linked to a particular transaction and not to the cost of service as such, otherwise,
the discipline of Art VIII.1 of the GATT would become void of any reasonable
meaning (§ 81):
The United States interpretation, by contrast, presented serious difficulties. Granted
that the terms ‘commensurate with’ and ‘approximate’ were intended to confer a cer-
tain degree of flexibility in the requirement that fees not exceed costs, the range of fees
permitted under the US merchandise processing fee could by no stretch of language be
considered a matter of mere flexibility. Moreover, the United States contention that ‘cost
of services rendered’ referred only to the total cost of the relevant government activities
would leave Articles II:2(c) and VIII:1(a) without any express standard for apportioning
such fees among individual importers, thereby committing the issue of apportionment,
at best, to an implied requirement of equitable (or non-protective) apportionment that
would be neither predictable nor capable of objective application. Finally, if ‘cost of serv-
ices rendered’ meant the total cost of customs operations, the ‘fiscal purposes’ criterion of
Article VIII:1(a) would be rendered largely redundant.
Based on this analysis, the panel went on to find that the US system was GATT
inconsistent, since, by virtue of its nature (ad valorem), the duty imposed was not
 Tariff Protection in the GATT 117

linked to the cost of the provided service: minor value transactions would pay less
than major value transactions for the same service (§§ 84–86):
The Panel was of the view, however, that the interpretation proposed by the United States
presented an equally serious problem with regard to the policy objectives of the General
Agreement. The problem was that the United States interpretation would enlarge the
‘service fee’ authority granted by Articles II:2(c) and VIII:1(a), more importantly the
former. Article II:2(c) is a rather extraordinary exception. It authorizes governments
to impose new charges on imports in excess of the ceiling established by a tariff bind-
ing. Given the central importance assigned by the General Agreement to protecting the
commercial value of tariff bindings, any such exceptions would require strict interpret-
ation. The exception stated in Article II:2(c) requires particularly strict interpretation,
however, because it does not conform to the policy justification normally given for such
exceptions. In the words of an explanation of Article II:2 contained in a 1980 proposal
by the Director-General (27S/24), the policy justification for the three types of border
charges permitted by Article II:2 was that they did not ‘discriminate against imports.’ If
the import fees authorized by Article II:2(c) were in fact fees for beneficial services, this
justification would be valid. But given the reality that most such fees are simply an ordin-
ary tax on imports, it cannot be said that such fees do not disadvantage imports vis-à-vis
domestic products. In simple terms, Article II:2(c) authorizes governments to impose
new protective charges in addition to the bound tariff rate. As such, it is an exception
which should be doubly guarded against enlargement by interpretation.
In the Panel’s view, the interpretation advocated by the United States would expand
the scope of Articles II:2(c), as well as VIII:1(a). It would permit a broader variety of
import fees to be imposed, and the greater availability and convenience of such fees
would, the Panel believed, lead to an increase in both the number and the level of such
fees. The Panel was convinced that the attainment of GATT policy objectives would not
be furthered by such an interpretation. Thus, even though the requirement that import
fees not exceed the cost of individual entries might increase the protective effect of such
fees in a particular case, the Panel was unable to accept the United States argument that
such consequences justified a more flexible interpretation. The Panel was satisfied that
the text of the General Agreement did impose such a requirement, and that it would not
promote the objectives of the General Agreement to relax it in the manner proposed by
the United States.
The Panel concluded that the term ‘cost of services rendered’ in Articles II:2(c) and VIII:1(a)
must be interpreted to refer to the cost of the customs processing for the individual entry in
question, and accordingly that the ad valorem structure of the United States merchandise pro-
cessing fee was inconsistent with the obligations of Articles II:2(c) and VIII:1(a) to the extent
that it caused fees to be levied in excess of such costs. (original emphasis)

The analysis in US—Customs User Fee has not been put into question in subse-
quent practice. WTO panels that had to deal with this issue, borrowed from it
and explicitly referred to this case-law, when dealing with a legal claim under Art
VIII of the GATT. The panel report on Argentina—Footwear reflects a quasi-
identical ruling on a similar issue: Argentina had an ad valorem charge for serv-
ices rendered in connection with the importation of goods. The panel, applying
118 Disciplines on Trade Instruments

the same logic as in US—Customs User Fee, found the Argentine measure to
be GATT inconsistent. It went on to address an argument by Argentina, that
this imposition was in line with the obligation it had assumed (by virtue of its
contract with the IMF). The panel rejected this argument as well, finding that
nothing in the contractual arrangement between the WTO and the IMF could
support the Argentine argument (§§ 6.74–6.80). This panel confirms that ad val-
orem schemes are inconsistent with Art VIII of the GATT, while leaving the door
open to justify them under the WTO/IMF arrangement, assuming that they
have been recommended by the IMF.¹³¹

3.8 Customs Valuation (CV)

When duties are shipped from one country to the other, they will form the object
of a CV. CV is a customs procedure applied to determine the customs value of
imported goods. CV procedures are applicable only when imported items have
to face ad valorem duties. In such transactions, the customs value is essential to
determine the duty to be paid on an imported good. GATT, Art VII stipulates
that the value for customs purposes of imported merchandise should be based on
the actual value of the imported merchandise on which the duty is assessed, and
should not be based on the value of merchandise of national origin or on arbitrary
or fictitious values. GATT, Art VII.2 reflects a definition of actual value in the
following terms:
(a) The value for customs purposes of imported merchandise should be based on the
actual value of the imported merchandise on which duty is assessed, or of like mer-
chandise, and should not be based on the value of merchandise of national origin or
on arbitrary or fictitious values.
(b) ‘Actual value’ should be the price at which, at a time and place determined by the
legislation of the country of importation, such or like merchandise is sold or offered
for sale in the ordinary course of trade under fully competitive conditions. To the
extent to which the price of such or like merchandise is governed by the quantity
in a particular transaction, the price to be considered should uniformly be related
to either (i) comparable quantities, or (ii) quantities not less favourable to import-
ers than those in which the greater volume of the merchandise is sold in the trade
between the countries of exportation and importation.
(c) When the actual value is not ascertainable in accordance with subparagraph (b) of
this paragraph, the value for customs purposes should be based on the nearest ascer-
tainable equivalent of such value.
GATT, Art VII condones regulatory diversity, and permits the use of different
methods for CV. During the Tokyo round, a CV agreement was adopted, which
introduced new disciplines. This agreement has since been superseded by the

¹³¹ The panel report on US—Certain EC Products is yet another illustration that the analysis in
US—Custom User Fee is still good law (§ 6.69). See infra in Chapter 4 for a more detailed discus-
sion on the deference to IMF recommendations that panels have shown.
 Tariff Protection in the GATT 119

Uruguay round Agreement on Customs Valuation. This agreement applies only

to the valuation of imported goods for the purpose of levying ad valorem duties.
It does not contain obligations concerning valuation for purposes of determin-
ing export duties or quota administration based on the value of goods, nor does
it lay down conditions for the valuation of goods for internal taxation or for-
eign exchange control. The new CV Agreement, following the approach already
adopted in Art VII of the GATT, makes it clear that CV shall, with the exception
of circumstances mentioned in the agreement, be based on the actual price of the
goods to be valued, which generally appears on the invoice. This price, adjusted
for certain elements featured in Art 8 of the CV, constitutes the transaction value.
The transaction value is the most important method of valuation referred to in
the agreement.
The transaction value, however, is not the only method that can be used by
customs authorities for CV purposes. The agreement provides for five alternative
methods to be applied in the prescribed hierarchical order in cases where there
is no transaction value, or where the transaction value is not acceptable because
the price has been distorted as a result of certain conditions. These other meth-
ods are: transaction value of identical goods, transaction value of similar goods,
deductive method, computed method and, finally, fall-back method, all of which
are defined in the CV Agreement.

3.9 The Agreement on Pre-shipment Inspection (PSIA)

The PSIA is one of the agreements introduced with the advent of the WTO as a
result of the successful conclusion of the Uruguay round. Th rough pre-shipment
inspection, one aims to ensure that the quantity and quality of the goods to be
exported conform to the specifications reflected in a sales contract. Therefore,
while the CV Agreement deals with the value of goods at the point of import-
ation, the PSIA deals with similar issues at the point of exportation.¹³² Although
PSIA inspections are government-contracted, they are performed by private
companies: the government of an importing WTO member will, thus, contract
a private company (PSI entity, like the Société Générale de Surveillance (SGS))
and ask it to perform an evaluation. PSI entities (companies) exercise essentially
two functions: first, they ensure conformity of goods with the terms of the sales
contract, and, second, they verify the invoice price.
Government contracts with PSI companies have been either foreign exchange
contracts (FOREX), where the basic objective of the government is to pre-
vent exodus of capital through over-invoicing, and/or customs contracts, where
the main aim is to ensure that there is no loss in customs revenue as a result of

¹³² The two agreements have thus a complementary function. On PSIA, see the very compre-
hensive account of Low (1995) who explains the rationale behind its enactment as well its basic
institutions.
120 Disciplines on Trade Instruments

under-valuation, or mis-classification of the good. A PSI company might also

provide a number of subsidiary services which include, inter alia, the verification
of origin of the product, maintenance of data for statistical purposes, technical
assistance and training. Expertise by PSIA companies is not binding on customs
authorities. They might decide to neglect it, and perform their own evaluation.
Disputes might arise between the PSI entities and exporters concerning the
evaluation by the former. Note that such disputes are between private parties,
and not between state entities. In such cases, the complaining party can refer
the dispute to the so-called Independent Review Entity (Art 4 of the PSIA). The
Independent Review Entity is a three-person body, including a member nomi-
nated by the PSI entities, a member nominated by the exporters’ organization,
and an independent trade expert. Its decisions are binding on their addressees
and cannot be appealed (Art 4(h) of the PSIA).¹³³

3.10 The MFN

The MFN obligation is relevant only for one of the three trade instruments:
tariffs.¹³⁴ Tariffs must, in principle, not discriminate across goods because of
their origin. There is no discrimination test in the context of QRs, and, as we will
see in section 4 of this chapter, there can be no discrimination test in the context
of export subsidies since they concern domestic goods only.
Hudec (1988) writes that, even in Medieval times, the city of Mantua (Italy)
obtained from the Holy Roman Emperor the promise that it would always bene-
fit from any privilege granted by the Emperor to ‘whatsoever other town’. Jackson
(1997, p158) notes that the term as such appears for the first time at the end of the
seventeenth century. During the nineteenth century, the provision appeared in a
number of treaties across European states. For instance, the Cobden-Chevalier
Treaty of 1860, liberalizing trade between Great Britain and France, included an
MFN clause guaranteeing that a signatory would not be treated worse than any
other state with which the other signatory had, or would assume, trade relations.
Such schemes, however, were not tantamount to worldwide non-discriminatory
trade; if at all, non-discriminatory trade existed between a sub-set of states, ie
those that had entered into a similar contractual arrangement espousing the
MFN clause.
Since the classic work of Viner (1924), economists have come a long way
towards explaining why non discriminatory (MFN) trade is the practically most
efficient way to go about organizing trade relations. Schwarz and Sykes (1998)
have advanced the idea that absent MFN there is risk of concession erosion: the
value (to A) of concessions negotiated between A and B, risks being eroded

¹³³ At the time of writing, two such decisions had been issued, see WTO Docs G/PSI/IE/R/1
and 2.
¹³⁴ This explains its placement immediately after the discussion on tariff s in this book.
 Tariff Protection in the GATT 121

through subsequent negotiations between B and C (assuming B has ‘conceded’ to

C more than it did to A).¹³⁵ Concession erosion provides a disincentive to nego-
tiate and liberalize in the first place. MFN is the insurance policy against con-
cession erosion. The relationship between concession erosion and terms of trade
externalities is interesting: although, in theory, concession erosion can exist irre-
spective of terms of trade effects, it becomes all the more meaningful in presence
of such effects. An illustration is appropriate on this score: if B, in the example
above, is quite small, then its tariff cut (favourable to C) will not change the world
price for any given good. A and C will be facing the same export price as before
and will continue to supply the same amount as before. B will be procuring from
C, but A’s loss will be probably infinitesimal. Moving, however, to a scenario with
terms of trade externalities, the situation could be presented as follows: assume
that B initially offered a 10% tariff to A and, then, offered a 5% tariff to C. Now,
since B is large, a lot of A’s exports will be deflected into the world market (B’s
consumers privileging products originating in C). The world price of A’s exports
will drop (assuming of course that A is big; if A is small, there will be no or little
effect on world prices). Thus, A’s reduced market access goes hand in hand with
A’s terms of trade loss. Assuming thus, that the country granting the concession
(B, in our example) can affect terms of trade, and that the country suffering con-
cession erosion is an important producer of the commodity at hand, concession
erosion and terms of trade losses go hand in hand.¹³⁶
Ethier (2004) persuasively argues that the role of MFN becomes all the more
important as the number of participants in trade deals increases. Assume that
A and B have signed an agreement and included an MFN clause. A subsequently,
negotiates with C and knows that whatever benefit it will grant to C will have
to be automatically extended to B as well. B in other words, will be free-riding
in A’s negotiation with C.¹³⁷ Any subsequent negotiation involves additional
negotiating costs for A, and additional advantages only for B. A and B might
thus be dissuaded from entering into subsequent negotiations. They would rather
invite all interested parties to participate in the initial negotiation, where they
all make commitments and all profit, in a non-discriminatory manner, from the
concessions entered.¹³⁸

¹³⁵ Close to this concept is the concept of bilateral opportunism advanced by Bagwell and Staiger
(2002). Other alternative explanations have been offered as well: recall the discussion on Ethier
(2004) in Chapter 1. Ethier takes the view that MFN is the instrument to internalize political
externalities. Horn and Mavroidis (2001) provide a survey of the economic thinking with the
aim of explaining MFN.
¹³⁶ Numerous discussions on this point with Kyle Bagwell and Henrik Horn are
acknowledged.
¹³⁷ It is irrelevant if B also negotiates with X, Y and Z, since B will have the same disincentive to
negotiate with them and let A free-ride on its negotiations. The problem, of course, is more acute if
A only negotiates subsequent deals.
¹³⁸ Incidentally, Hawkins (1951, pp81ff ) reports that the unwillingness to compensate free-
riders is what motivated the US government to grant MFN under the RTAA only to the countries
that it considered principal suppliers of a particular commodity.
122 Disciplines on Trade Instruments

The MFN obligation, the cornerstone of the GATT as it is often called, is

included in Art I of the GATT:
Article I
General Most-Favoured-Nation Treatment

1. With respect to customs duties and charges of any kind imposed on or in connection
with importation or exportation or imposed on the international transfer of payments
for imports or exports, and with respect to the method of levying such duties and
charges, and with respect to all rules and formalities in connection with importation
and exportation, and with respect to all matters referred to in paragraphs 2 and 4 of
Article III, any advantage, favour, privilege or immunity granted by any contracting
party to any product originating in or destined for any other country shall be accorded
immediately and unconditionally to the like product originating in or destined for the
territories of all other contracting parties.
2. The provisions of paragraph 1 of this Article shall not require the elimination of
any preferences in respect of import duties or charges which do not exceed the lev-
els provided for in paragraph 4 of this Article and which fall within the following
descriptions:
(a) Preferences in force exclusively between two or more of the territories listed in
Annex A, subject to the conditions set forth therein;
(b) Preferences in force exclusively between two or more territories which on July 1,
1939, were connected by common sovereignty or relations of protection or suzer-
ainty and which are listed in Annexes B, C and D, subject to the conditions set
forth therein;
(c) Preferences in force exclusively between the United States of America and the
Republic of Cuba;
(d) Preferences in force exclusively between neighbouring countries listed in Annexes
E and F.
3. The provisions of paragraph 1 shall not apply to preferences between the countries for-
merly a part of the Ottoman Empire and detached from it on July 24, 1923, provided
such preferences are approved under paragraph 5¹³⁹, of Article XXV which shall be
applied in this respect in the light of paragraph 1 of Article XXIX.
4. The margin of preference on any product in respect of which a preference is permit-
ted under paragraph 2 of this Article but is not specifically set forth as a maximum
margin of preference in the appropriate Schedule annexed to this Agreement shall not
exceed:
(a) in respect of duties or charges on any product described in such Schedule, the
difference between the most-favoured-nation and preferential rates provided for
therein; if no preferential rate is provided for, the preferential rate shall for the
purposes of this paragraph be taken to be that in force on April 10, 1947, and, if no
most-favoured-nation rate is provided for, the margin shall not exceed the diffe-
rence between the most-favoured-nation and preferential rates existing on April
10, 1947;

¹³⁹ The authentic text erroneously reads ‘subparagraph 5 (a)’.

 Tariff Protection in the GATT 123
(b) in respect of duties or charges on any product not described in the appropri-
ate Schedule, the difference between the most-favoured-nation and preferential
rates existing on April 10, 1947.
In the case of the contracting parties named in Annex G, the date of April 10, 1947,
referred to in subparagraph (a) and (b) of this paragraph shall be replaced by the respect-
ive dates set forth in that Annex.

3.10.1 The MFN discipline in a nutshell

The MFN is the carrot offered to outsiders: by acceding to the WTO, a country
knows that it will benefit from the best possible treatment granted by incum-
bents to imported products, with respect to, in principle, all measures affecting
trade.¹⁴⁰ The promise, more specifically, consists of the following elements:
(1) with respect to, in principle, all measures which affect trade either de jure or
de facto, any advantage granted to goods originating anywhere in the world
(section 3.10.2);
(2) must be extended to the like products (section 3.10.3);
(3) originating in any WTO member (section 3.10.4);
(4) immediately and unconditionally (section 3.10.5).
The standard of review, that has been applied by WTO adjudicating bodies in
Art I of the GATT cases, is quite favourable to the complainant: there is no need
to demonstrate intent to discriminate, and there is no need to demonstrate the
resulting trade effects either (section 3.10.6).
In today’s world, however, MFN trade becomes more of an illusion than a real-
ity. As a series of empirical works have demonstrated,¹⁴¹ most trade is conducted
nowadays on non-MFN terms. The proportion of non-MFN trade in the overall
volume of trade is ever increasing, in light, inter alia, of the renewed enthusiasm
for PTAs.¹⁴²
GATT, Art I.2 reflects the so-called grandfathering clause: during the GATT
years, those acceding could grandfather pre-existing schemes which were MFN
inconsistent, and immunize them from legal challenges. This provision is of mere
historical interest nowadays, since, following the advent of the WTO, no grand-
fathering is permissible anymore.

¹⁴⁰ Hence, the coverage of MFN extends beyond trade instruments; it covers domestic instru-
ments as well. Th is is why, although legally relevant, we will avoid references to MFN in Chapter 3
where we discuss the disciplining of domestic instruments. The analysis presented here is also
pertinent for domestic instruments as well.
¹⁴¹ See, inter alia, Sapir (1998), Schott (1989) and compare with more recent studies such as
Limão (2006c).
¹⁴² Bhagwati in his inimitable expression often refers to MFN as LFN, the least favoured nation
clause. The title LFN is probably appropriate in light of the manifold preferences that WTO mem-
bers often grant to each other.
124 Disciplines on Trade Instruments

3.10.2 The measures coming under the purview of the MFN

3.10.2.1 Any advantage . . .
A textual reading of Art I.1 of the GATT supports a wide interpretation of the
term ‘advantage’: the word any, appearing before the word advantage, invites
the WTO adjudicating bodies to extend the coverage of the MFN clause to any
measure which, in their view, qualifies as an advantage. The advantages in the Art
I of the GATT sense of the term are further divided into three categories:
(1) customs duties and charges of any kind imposed on or in connection with
importation and exportation;
(2) rules and formalities in connection with importation and exportation;
(3) internal measures.

(a) Customs duties and charges of any kind GATT, Art I.1 does not distinguish
between goods where protection has been consolidated (bound duties), and
goods where protection has not been consolidated. As a result, the MFN clause
is equally applicable to both bound and to unbound customs duties. The GATT
panel on Spain—Unroasted Coff ee, discussed supra, stated as much in the follow-
ing terms (§ 4.3):
Having noted that Spain had not bound under the GATT its tariff rate on unroasted coffee,
the Panel pointed out that Article I:1 equally applied to bound and unbound tariff items.
It follows that, in the case applied rates are different to bound rates, WTO mem-
bers, by virtue of the MFN clause, must ensure that such duties are applied in a
non-discriminatory manner (since a lower applied rate is undeniably an advan-
tage). GATT practice suggests that consular fees (GATT Doc BISD II/12) and
customs user fees (US—Customs User Fee) fall into this category.
GATT practice also shows that not only actions, but also omissions, to the
extent that they confer an advantage are covered by the discipline laid down in
Art I.1 of the GATT: the GATT panel on US—Customs User Fee held for the
proposition that an exemption from the imposition of a customs fee should be
considered to be an advantage in the sense of Art I.1 of the GATT.
GATT, Art I.1 also states that the MFN clause does not extend only to duties
and charges as such, but also to the methods of levying them. GATT Doc. L/3149
(29 November 1968) makes it plain that AD duties are covered by this term.
This issue is of course now moot, in light of the specific provision to this effect,
enshrined in the body of the WTO Agreement on Antidumping. The GATT
panel report on US—Non-Rubber Footwear makes it clear (§ 6.8) that counter-
vailing duties are also covered by this term.

(b) Rules and formalities in connection with importation and exportation The
GATT explicitly mentions one such formality: Art IX of the GATT deals with
marks of origin. It essentially imposes one obligation: that marks of origin be
 Tariff Protection in the GATT 125

applied on an MFN basis (Art IX.1 of the GATT). In its second paragraph, it
reflects a best endeavours clause: laws and requirements relating to marks of ori-
gin should be reduced to the minimum, without however, putting into question
the regulatory objectives pursued. As to this last point, Art IX of the GATT rec-
ognizes that marks of origin can be an effective tool against fraudulent or mis-
leading indications which run counter to consumers’ interests. The rest of the
provision contains various illustrations of the best endeavours clause.
Case-law has indentified a number of other rules and formalities that come
under the purview of Art I of the GATT. The leading case in this respect is the
EC—Bananas III litigation discussed above. In this panel report, the use of a
less complicated licensing procedure (§§ 7.188ff ), the incentive given to operators
to purchase bananas of a particular origin (§ 7.194), the issuance of a license to
import bananas of a particular origin upon the economic activity performed by
the economic operator requesting the license (§§ 7.220ff ), the granting of licenses
to operators representing producers from certain countries only (§§ 7.251ff ), and
the imposition on certain bananas of the in-quota tariff rate provided that they
originate in particular countries (§§ 7.235ff ), were all considered to be advan-
tages within the meaning of Art I.1 of the GATT. These findings were confirmed
by the AB (§§ 206ff ).
The panel report on US—Non-Rubber Footwear from Brazil found that the
automatic backdating of the revocation of a countervailing duty order without
the need to have an injury review conducted in this respect is an advantage in the
sense of Art I.1 of the GATT.
The AB report on EC—Poultry in §§ 96ff understood the MFN obligation
to cover tariff quotas as well. That is, when deciding on rates within and outside
the tariff quota, WTO members must ensure that they have adhered to the non-
discrimination principle.

(c) Internal measures GATT, Art I.1 also makes it clear, by virtue of the explicit
reference to Art III.2 and III.4 of the GATT, that the MFN clause extends to bor-
der and internal measures alike. The first confirmation of this point came with
the GATT panel report on Belgian Family Allowances, where tax exemptions for
products purchased by public bodies (ie internal measures) were found to be cov-
ered by Art I.1 of the GATT. This case-law has been adhered to with no exception
ever since.

3.10.2.2 . . . granted to products originating in any country . . .

GATT, Art I.1 when referring to:
any advantage, favour, privilege or immunity granted by any contracting party to any
product originating in or destined for any other country shall be accorded immediately
and unconditionally to the like product originating in or destined for the territories of all
other contracting parties. (emphasis added)
126 Disciplines on Trade Instruments

imposes on WTO members the obligation to treat goods originating in a WTO

member at least as well as they treat goods originating in a non-WTO mem-
ber (hence the name, MFN clause). Adherence to the WTO amounts then to a
guarantee that a country will receive the best trade treatment in the markets of
incumbent members.

3.10.2.3 . . . irrespective whether it discriminates de jure or de facto . . .

GATT, Art I does not explicitly refer to either de jure or de facto discrimination.
The AB, in its report on Canada—Autos, held the view that both forms of dis-
crimination are covered by the prohibition included in Art I.1 of the GATT.
Consequently, not only measures discriminatory on their face (eg a law whereby
only goods originating in a sub-set of the WTO membership benefit from the
lower, in comparison to the bound, applied rate), but measures which on their
face seem origin-blind can also be successfully challenged to the extent that they
de facto discriminate among different sources of supply. We quote from § 78:
In approaching this question, we observe first that the words of Article I:1 do not restrict
its scope only to cases in which the failure to accord an ‘advantage’ to like products of all
other Members appears on the face of the measure, or can be demonstrated on the basis of
the words of the measure. Neither the words ‘de jure’ nor ‘de facto’ appear in Article I:1.
Nevertheless, we observe that Article I:1 does not cover only ‘in law’, or de jure, discrim-
ination. As several GATT panel reports confirmed, Article I:1 covers also ‘in fact’, or de
facto, discrimination. Like the Panel, we cannot accept Canada’s argument that Article
I:1 does not apply to measures which, on their face, are ‘origin-neutral.’
The problem with de facto discrimination is that it effectively amounts to opening
Pandora’s box. The definition of like products, just like the definition of uncon-
ditional treatment, will have an important bearing on the judgment whether de
facto discrimination has occurred.

3.10.3 . . . must be extended to like products

Likeness is, of course, far from being a self-interpreting term. GATT/WTO jur-
isprudence has had recourse to various criteria to establish likeness. Tariff classi-
fication emerges as probably the dominant criterion in the context of Art I of the
GATT.¹⁴³ First, the 1978 panel report on EEC—Animal Feed Proteins moved
towards this criterion. In this case, the question arose as to whether, by treating
different protein products in different ways, the European Community was vio-
lating its obligations. There was little doubt that the products at hand could be
regarded as substitutable, that is, as belonging to the same relevant product mar-
ket. The question, however, was whether such a criterion (substitutability) was

¹⁴³ See on this issue the comprehensive analysis of Davey and Pauwelyn (2000).
 Tariff Protection in the GATT 127

still appropriate to decide likeness under Art I of the GATT.¹⁴⁴ The panel decided
that this should not be the case noting that (§ 4.20):
. . . such factors as the number of products and tariff items carrying different duty rates and
tariff bindings, the varying protein contents and the different vegetable, animal and syn-
thetic origins of the protein products before the Panel—not all of which were subject to the
EEC measures. Therefore, the Panel concluded that these various protein products could
not be considered as ‘like products’ within the meaning of Articles I and III (§ 4.2).
...
The Panel noted that the general most-favoured-nation treatment provided for in
Article I:1 . . .did not mention directly competitive or substitutable products. In this
regard the Panel did not consider animal, marine and synthetic proteins to be products
like those vegetable proteins covered by the measures.
The panel report on Japan—SPF Dimension Lumber went even further and pro-
vided an explicit acknowledgement of the relevance of tariff classification as the
dominant criterion to establish likeness (§§ 5.11–5.12):
. . . if a claim of likeness was raised by a contracting party in relation to the tariff treat-
ment of its goods on importation by some other contracting party, such a claim should be
based on the classification of the latter, i.e., the importing country’s tariff.
The Panel noted in this respect that ‘dimension lumber’ as defined by Canada was
a concept extraneous to the Japanese Tariff . . . nor did it belong to any internationally
accepted customs classification. The Panel concluded therefore that reliance by Canada
on the concept of dimension lumber was not an appropriate basis for establishing ‘like-
ness’ of products under Article I:1 of the General Agreement.¹⁴⁵
GATT panels have even gone so far and dismissed the relevance of factors other
than tariff classification, arguing that they are irrelevant for the purposes of defin-
ing likeness. The GATT panel on Spain—Un-roasted Coffee, for example, set aside
the relevance of process-based distinctions in defining likeness when dealing with
a complaint by Brazil to the effect that a Spanish classification of un-roasted cof-
fee between Colombian mild, other mild, unwashed Arabica, robusta and other,
which accorded to the first two categories a duty free treatment, to the last three a
7% import duty, while the duty for roasted coffee was left un-bound; this practice,
the panel found, was inconsistent with Art I of the GATT. It noted (§§ 4.7–4.10):
The Panel examined all arguments that had been advanced during the proceedings for
the justification of a different tariff treatment for various groups and types of un-roasted

¹⁴⁴ As we will see in Chapter 3, substitutability is a key criterion to decide likeness in the context
of the discipline applied to domestic instruments.
¹⁴⁵ Note that panels dealt in both cases with goods such as protein products and dimension
lumber. Such products do not come under a two- or four-digit level. Although case-law has not pro-
vided us with a number, it seems safe to conclude that the higher the number of digits involved, the
easier it will be to show likeness. There should be no doubt that the six-digit level provides detailed
enough classifications. A similar criterion (detailed classifications) has been privileged by the AB
to decide on likeness under Art III.2 of the GATT in the Japan—Alcoholic Beverages II report, see
infra in Chapter 3.
128 Disciplines on Trade Instruments
coffee. It noted that these arguments mainly related to organoleptic differences result-
ing from geographical factors, cultivation methods, the processing of the beans, and the
genetic factor. The Panel did not consider that such differences were sufficient reason to
allow for a different treatment. It pointed out that it was not unusual in the case of agri-
cultural products that the taste and aroma of the end-product would differ because of one
or several of the above-mentioned factors.
The Panel furthermore found relevant to its examination of the matter that un-roasted
coffee was mainly, if not exclusively, sold in the form of blends, combining various types
of coffee, and that coffee in its end-use, was universally regarded as a well-defined and
single product intended for drinking.
The Panel noted that no other contracting party applied its tariff regime in respect of
un-roasted, non-decaffeinated coffee in such a way that different types of coffee were sub-
ject to different tariff rates.
In light of the foregoing, the Panel concluded that un-roasted, non-decaffeinated cof-
fee beans listed in the Spanish Customs Tariff . . . should be considered as like products
within the meaning of Article I:1. (original emphasis)
The question nevertheless, arises whether any tariff classification can serve as
criterion to decide on likeness. Recall that the harmonized tariff classification
extends up to the six-digit level. WTO members can, nonetheless, as noted supra,
shape their tariff bindings using eight-digit classifications. Through such classifi-
cations, they could be providing advantages to a sub-set of goods coming under a
six-digit classification. Importantly, they could do as much by introducing proc-
ess-based distinctions: assuming for example that at the six-digit level the item
construction material is listed and goods are bound at 10% import duty, a WTO
member could introduce at the eight-digit level a distinction between asbestos-
containing and asbestos-free construction material, imposing a 10% import duty
on the former and a 0% import duty on the latter. Are such distinctions GATT
consistent? The GATT panel on Spain—Un-roasted Coff ee responds in the nega-
tive. Is this still good law?
This issue is particularly interesting if viewed in the context of the commit-
ment theory, briefly discussed in Chapter 1. Recall that under this approach, one
rationale for trade agreements (and, consequently, the GATT) is the willingness
of governments to ‘tie themselves to the mast’ of an international contract and
thus resist pressure from political lobbies.¹⁴⁶ As we will see in Chapter 3, such dis-
tinctions, when operated through domestic instruments, have been judged GATT
consistent: this is what the notorious EC—Asbestos jurisprudence amounts to for
all practical purposes. So why then would a government prefer to choose a trade
instead of a domestic instrument to achieve the same objective?
Assuming that such distinctions are GATT consistent irrespective of the
instrument used to attain them, a WTO member might wish to commit itself to

¹⁴⁶ Hudec used to call this approach ‘use GATT as an excuse’. The excuse was of course neces-
sary in order to adopt policies that were maximizing national welfare but not the narrower produ-
cer welfare.
 Tariff Protection in the GATT 129

border measures in order to avoid subsequent pressures to enact domestic laws.

A government which promises a 10% duty on asbestos-containing materials can
convincingly thwart subsequent pressures to ban such imports by pointing to
the commitment made. On the other hand, making a tariff commitment entails
a quid pro quo negotiation: the government will be extracting a promise from a
trading partner (and at the same time it will be satisfying a domestic lobby) in
order to make a tariff commitment.
As things stand, however, for legal reasons, it is uncertain whether WTO
members can use trade and domestic instruments interchangeably to advance
social (process based) preferences. Whereas case-law with respect to domestic
instruments is settled, and is favourable to such distinctions, there are no cases
that squarely address this issue in the context of trade instruments. The only case
from which one could draw useful references is the AB report on EC—Tariff
Preferences.¹⁴⁷ There the AB judged that an EC scheme conferring tariff prefer-
ences to a sub-set of developing countries upon satisfaction of criteria unilaterally
set by the European Community¹⁴⁸ was GATT consistent, to the extent that the
criteria set were objective. There are two reasons, however, that should mitigate
the ‘sooners’ who would rush to adopt a ‘lock, stock, and barrel’ transposition of
this case-law to Art I of the GATT:
(1) the AB left it for later to decide what an objective criterion is;
(2) the dispute came under the purview of the enabling clause and not Art I of
the GATT.¹⁴⁹
Assuming that the criterion unilaterally used has been accepted by WTO adjudi-
cating bodies as an objective criterion, and that panels and the AB will be willing
to apply the reasoning followed in EC—Tariff Preferences in Art I of the GATT
litigation as well,¹⁵⁰ then eight or more digit classifications can also serve as
criteria to decide on likeness.

3.10.4 The rules of origin

WTO members have to extend any advantage granted to a product originating
in one country, to all like goods originating in all other WTO members. The

¹⁴⁷ For a more detailed discussion of this case, see infra section 3.11 in this chapter.
¹⁴⁸ In the case at hand, as we will see in more detail in section 3.11, the criteria established by the
European Community were unrelated to production process: developing countries that pursued
policies aiming at combating drug trafficking were being compensated through additional (to what
applied elsewhere) tariff preferences.
¹⁴⁹ Point (2) might prove the biggest obstacle in transposing this case-law to Art I of the GATT.
The AB insisted on the statutory language in the enabling clause which, in its view, allowed for
distinctions across developing countries. No equivalent language can be found in the context of
Art I of the GATT.
¹⁵⁰ The latter assumption should be less problematic than the former. The enabling clause, like
the MFN, serves a similar function: it must be applied on a non-discriminatory basis to a sub-set of
the WTO membership.
130 Disciplines on Trade Instruments

WTO ROO does not impose a harmonized set of rules that WTO members
must observe when it comes to conferring origin. Instead, Art 2(d) of the ROO
condones regulatory diversity in this respect.¹⁵¹ It reads:
Until the work programme for the harmonization of rules of origin set out in Part IV is
completed, Members shall ensure that:

...
(d) the rules of origin that they apply to imports and exports are not more stringent than
the rules of origin they apply to determine whether or not a good is domestic and
shall not discriminate between other Members, irrespective of the affi liation of the
manufacturers of the good concerned.
The ROO deals with MFN rules of origin (Art 1.2 of the ROO). Besides non-
discrimination, WTO members must also ensure that their rules of origin are
transparent, that they do not have restricting effects on international trade, and
that they are administered in a consistent, uniform, impartial and reasonable
manner. In the long run, the WTO ROO aims to establish harmonized rules of
origin among all WTO members (Art 9 of the ROO). The work was due to end
in July 1998, but several deadlines have been missed. It is being conducted by the
WTO Committee on Rules of Origin, and a Technical Committee operating
under the auspices of the WCO.
It follows that, until a harmonized regime sees the light of the day, regulatory
diversity will persist.¹⁵² Regulatory diversity¹⁵³ is not inconsequential, for rules
of origin can be used as a trade instrument and restrict trade. For a start, rules
of origin do not arise at all if a product is wholly obtained or produced in one
country. It is an issue only if more than one country is involved in a products’
production, and its origin has to be determined. This is a very likely scenario
in today’s world, where trade in tasks is widely practiced through off-shoring
and outsourcing. In such cases, the product will usually have the origin of the
country where its last substantial transformation occurred.¹⁵⁴ Some of the rules
employed to decide whether substantial transformation has occurred leave sub-
stantial discretion to the administering authority and, as a result, disputes arise

¹⁵¹ See on this score, Vermulst (1995).

¹⁵² The main objective of the ROO is, as per its Art 9, to harmonize non-preferential rules of
origin. Th is requires work on 5,000 tariff lines. Some progress has been done but a lot still needs to
be done. The rules of origin applied to antidumping, countervailing, safeguards, SPS, TBT, and
labelling issues figure among the thorniest issues where no consensus has been reached as yet.
¹⁵³ Vermulst (1992) offers a very comprehensive comparative study in this respect, with examples
from state practice across various jurisdictions.
¹⁵⁴ To decide whether substantial transformation occurred, countries will have recourse to cri-
teria such as the percentage criterion (where the question is whether a certain percentage of value
added has been added), the change in tariff heading (as a result of the transformation), and the
so-called technical criterion, which prescribes certain production or sourcing processes that may
(positive technical criterion) or may not (negative technical criterion) confer originating status.
 Tariff Protection in the GATT 131

frequently.¹⁵⁵ Nevertheless, because the WTO rules of origin condone regulatory

diversity and, essentially, hinge on non-discrimination (a criterion that is easy
to satisfy since rules of origin can be decided so as to accommodate domestic
lobbies’ pressures, and even be modified at any time in the future to the extent
that they continue to be applied in a non-discriminatory basis) disputes will be
very infrequently submitted to the WTO. The only reported case so far is the
US—Textiles Rules of Origin, where the Panel in §§ 6.23–24 underscored the
wide discretion that WTO members enjoy when designing their rules of origin in
the following terms:
With regard to the provisions of Article 2 at issue in this case—subparagraphs (b)
through (d)—we note that they set out what rules of origin should not do: rules of origin
should not pursue trade objectives directly or indirectly; they should not themselves cre-
ate restrictive, distorting or disruptive effects on international trade; they should not pose
unduly strict requirements or require the fulfilment of a condition unrelated to manufac-
turing or processing; and they should not discriminate between other Members. These
provisions do not prescribe what a Member must do.
By setting out what Members cannot do, these provisions leave for Members them-
selves discretion to decide what, within those bounds, they can do. In this regard, it
is common ground between the parties that Article 2 does not prevent Members from
determining the criteria which confer origin, changing those criteria over time, or apply-
ing different criteria to different goods.
There are preferential rules of origin as well: in fact, there is a menu of preferential
rules of origin; different rules are applied depending on the percentage of value
added, the overall integration process pursued, etc.¹⁵⁶ Preferential rules of ori-
gin are, in general, more restrictive than non-preferential rules. Vermulst (1992,
pp 37ff ) cites numerous EC court judgments in support of this claim. There is also
empirical evidence to this effect: Brenton and Manchin (2003) investigate EC
preferential schemes (such as ‘Everything But Arms’), and suggest that for them
to generate substantial improvements for developing countries, the European
Community would have to reconsider its current rules of origin. Mattoo et al
(2003) estimate the medium-term benefits for African exporters stemming from
the US Africa Growth and Opportunity Act (AGOA), and argue that the gains
would be five times as much, were the United States to relax the current stringent
rules of origin.¹⁵⁷ There is, further, ample empirical evidence that rules of origin
have contributed to trade diversion.¹⁵⁸
The current disciplines on preferential rules of origin are very much a subject of
debate. As already stated, the ROO Agreement (Art 1.2) covers:

¹⁵⁵ Some of the most notorious disputes are reported in Vermulst and Waer (1990).
¹⁵⁶ See Inama (2003), and Krueger (1997) on this issue.
¹⁵⁷ Similar evidence concerning the US trade with other American countries is provided in
Estevadeordal and Garay (1996).
¹⁵⁸ For a survey of the literature not only on this score, but on all other aspects concerning rules
of origin, see Krishna (2005).
132 Disciplines on Trade Instruments
all rules of origin used in non-preferential commercial policy instruments. (emphasis
added)
On its face, this provision would suggest that preferential rules of origin are not
covered by the ROO.¹⁵⁹ Does exclusion from the ROO coverage amount to
exclusion from the WTO disciplines altogether? An interpretation of the WTO
contract based on lex specialis would unavoidably lead to this conclusion. An
interpretation in accordance with lex posterior would lend further support to this
approach: the ROO came into life on 1 January 1995, that is, following many
years of GATT practice where the consistency of preferential rules of origin had
been discussed (even tangentially sometimes) in the context of Art XXIV of the
GATT. Working parties examined the consistency of a PTA with the GATT
rules:¹⁶⁰ rules of origin could conceivably come under the terms other regula-
tions of commerce (Art XXIV.5 of the GATT), or restrictive regulations of commerce
(Art XXIV.8 of the GATT), and their consistency with these two provisions was
the subject of many inconclusive discussions. Some confusion stems from the fact
that discussion on preferential rules of origin continues even in the post-Uruguay
round era, that is, after the advent of the ROO Agreement. The discussion so far,
nevertheless, has stopped short of addressing the consistency of preferential rules
of origin with the WTO rules, as those reviewing the consistency of a PTA with
the multilateral rules have been requesting factual information only. As we will
see infra, nevertheless, the absence of multilateral disciplines on preferential rules
of origin is in itself quite problematic.
Imaginative proposals have been tabled aiming to reduce the costs in the
meantime, that is, while awaiting the advent of a comprehensive legal regime.¹⁶¹
There is, in fact, a strong transaction costs argument in favour of harmonizing the
existing regime. There is an associated belief that, through harmonization, rules
of origin will be rationalized: indeed, it is very much the case that the existing
regime leaves ample room for introducing protectionist measures under the guise
of rules conferring origin. The ongoing Doha round negotiations have failed, so
far, to produce tangible results on this score.

3.10.5 The MFN treatment must be extended immediately

and unconditionally
By virtue of the MFN obligation, WTO members must extend any advantage
(as understood above) immediately and unconditionally to all WTO members.
The term ‘immediately’ seems to suggest that no time should lapse between the
granting of an advantage in the first instance, and its extension to all like products

¹⁵⁹ ROO, Annex II includes a common declaration with regard to preferential rules of origin.
In essence, this Annex imposes a transparency obligation.
¹⁶⁰ See infra.
¹⁶¹ See, for example, Lloyd (1993) who proposes the introduction of a tariff equivalent that will
be negotiable.
 Tariff Protection in the GATT 133

originating in WTO members. The term ‘unconditionally’, on the other hand,

calls for an interpretation whereby no conditions should be attached when an
advantage is being extended. The difference in content notwithstanding, WTO
adjudicating bodies will typically review the two terms in tandem. The AB report
on Canada—Autos is a good illustration of this point (§§ 75–86).
Most of the interpretative problems have to do with the understanding of
the term ‘unconditionally’. The issue that has occupied panels is, to what extent
should conditions only additional to those which are necessary for the granting
of an advantage in the first place be relevant for the interpretation of the term
‘unconditionally’, or, conversely, if no conditions at all should be imposed in the
first instance? It seems that the latter reading should be irrelevant for the inter-
pretation of the term ‘unconditionally’: if at all, such a discussion should take
place in the context of the likeness-determination. Nevertheless, much of the
case-law relating to the interpretation of the term ‘unconditionally’ blurs this
distinction. There is a first category of cases (panel and working party reports),
which interpret the term ‘unconditionally’ as equivalent to outlawing any condi-
tions imposed by the importing WTO member. These reports for all practical
purposes do not compare two situations to see whether additional commitments
have been imposed in a two-stage game, for the granting of the same advantage.
They outlaw the imposition of a condition irrespective of discrimination across
two transactions:
(1) the GATT panel report on Belgian Family Allowances, concerning tax exemp-
tions for products purchased by public bodies made conditional on the exist-
ence of a certain system of family allowances to be in force in the exporting
country, were found to be inconsistent with Art I.1 of the GATT;
(2) the GATT panel report on EEC—Imports of Beef reflects the view that condi-
tioning a duty waiver upon certification by a particular government violates
the obligation to grant MFN unconditionally;
(3) the working party report on Accession of Hungary¹⁶² reflects the view that to
condition a tariff treatment upon the prior acceptance of a cooperation agree-
ment is a violation of the requirement imposed by Art I.1 of the GATT to
grant MFN unconditionally;
(4) the WTO panel report on Indonesia—Autos found that Indonesian prac-
tices granting tax advantages to Korean companies which had entered into
arrangements with Indonesian companies were inconsistent with the obliga-
tion under Art I.1 of the GATT to grant MFN unconditionally;
(5) perhaps, even more dramatically, the panel report on EC—Tariff Preferences
in §§ 7.59 and 7.60 adopts an interpretation of the term ‘unconditionally’,

¹⁶² Report adopted on 30 July 1973, see GATT Doc BISD 20S/34.
134 Disciplines on Trade Instruments

whereby it becomes impossible to attach any conditions, not even when the
advantage is granted in the first place:

In the Panel’s view, moreover, the term ‘unconditionally’ in Article I:1 has a broader
meaning than simply that of not requiring compensation. While the Panel acknowl-
edges the European Communities’ argument that conditionality in the context of trad-
itional MFN clauses in bilateral treaties may relate to conditions of trade compensation
for receiving MFN treatment, the Panel does not consider this to be the full meaning
of ‘unconditionally’ under Article I:1. Rather, the Panel sees no reason not to give that
term its ordinary meaning under Article I:1, that is, ‘not limited by or subject to any
conditions.’
Because the tariff preferences under the Drug Arrangements are accorded only on the
condition that the receiving countries are experiencing a certain gravity of drug prob-
lems, these tariff preferences are not accorded ‘unconditionally’ to the like products ori-
ginating in all other WTO members, as required by Article I:1. The Panel therefore finds
that the tariff advantages under the Drug Arrangements are not consistent with Article
I:1 of GATT 1994.
There is a second batch of cases which do not take such an absolute approach.
These reports compare two situations and essentially try to ascertain to what
extent additional conditions have been imposed when extending an already
granted (probably, following the fulfillment of specified conditions) advantage:
(1) the GATT panel report on EEC—Minimum Import Prices dealt with the fol-
lowing issue: the EC authorities required a payment deposit from all coun-
tries that could not guarantee a specified minimum import price. However,
since the payment of the deposit was requested by all exporting countries
falling into this category, the EC scheme was not considered to be a violation
of Art I.1 of the GATT;
(2) the WTO panel report on Canada—Autos held the view that the term
‘unconditionally’ does not mean that all conditions are prohibited. Rather,
unconditionally refers, in the panel’s view, to the notion that MFN treatment
towards another WTO member shall not be conditional on reciprocal con-
duct by that other WTO member. Thus, conditions that are non-discrimina-
tory across two transactions involving like goods originating in two different
WTO members do not violate Art I of the GATT (§§ 10.22 and 10.24):
In our view, whether an advantage within the meaning of Article I:1 is accorded ‘uncon-
ditionally’ cannot be determined independently of an examination of whether it involves
discrimination between like products of different countries.
...
In this respect, it appears to us that there is an important distinction to be made between,
on the one hand, the issue of whether an advantage within the meaning of Article I:1 is
subject to conditions, and, on the other, whether an advantage, once it has been granted
to the product of any country, is accorded ‘unconditionally’ to the like product of all
other Members. An advantage can be granted subject to conditions without necessarily
 Tariff Protection in the GATT 135
implying that it is not accorded ‘unconditionally’ to the like product of other Members.
More specifically, the fact that conditions attached to such an advantage are not related to
the imported product itself does not necessarily imply that such conditions are discrim-
inatory with respect to the origin of imported products. We therefore do not believe that,
as argued by Japan, the word ‘unconditionally’ in Article I:1 must be interpreted to mean
that making an advantage conditional on criteria not related to the imported product
itself is per se inconsistent with Article I:1, irrespective of whether and how such criteria
relate to the origin of the imported products.
It is submitted that the first string of cases rests on a misguided rationale. The
test embedded in Art I of the GATT is about discrimination, not about deregula-
tion: states are allowed to introduce legislation they wish to the extent that they
do not discriminate across like goods originating in different WTO members.
With respect to tariff-treatment only, it is true that no conditions can be attached
beyond what already exists in the classification.¹⁶³ But it should be perfectly legit-
imate for a WTO member to condition access to its market only upon prior veri-
fication that, for example, a good’s origin is the one declared in its accompanying
documents. Indeed, Art IX of the GATT says as much. The question is whether
the GATT, as we know it, has taken care of all legitimate concerns that import-
ing states might have. The advent of the CV and the PSIA Agreements (see infra)
is evidence enough that the original contract had not done so. The ongoing nego-
tiation on trade facilitation is a further argument in support of this thesis. If at
all, past experience is evidence that there is no reason to believe that the contract
has been completed in this respect. Importing states might still be willing to con-
dition access upon the supply of information. One would have, however, to buy
an insurance policy against abuses: WTO members could, for example, impose
onerous conditions, which had not been negotiated before and, thus, undercut
the value of their (negotiated) tariff concessions. Ideally, the request for supply
of information should correspond to a legitimate public order rationale. The CV,
and PSIA Agreements, as well as the ongoing discussions on trade facilitation, are
multilateral initiatives aiming to satisfy legitimate rationales. WTO adjudicating
bodies should have the authority to acknowledge that unilateral measures, which
extend beyond multilateral initatives, might also pursue legitimate public order
objectives.¹⁶⁴

3.10.6 No rebalancing permitted

The panel report on US—Non-Rubber Footwear dealt with the following situ-
ation: the United States claimed that it could legitimately rebalance (offset) the

¹⁶³ This observation is good law with respect to HS-classifications only. As discussed above,
there is no ex ante guarantee that eight-digit classifications are GATT consistent.
¹⁶⁴ Assuming the discrimination test is the same in Art I, and Art XX of the GATT (quite
reasonable assumption), then Art XX of the GATT cannot be construed as exception to Art I of
the GATT. In Chapter 3, we discuss the relationship between Art III, and Art XX of the GATT.
The same rationale applies to the relationship between Art I, and Art XX of the GATT.
136 Disciplines on Trade Instruments

damage done to some imports through more favourable treatment under another
measure. In its view, in other words, it is permissible under Art I.1 of the GATT
to discriminate negatively with respect to some measures as long as some positive
discrimination is provided through other measures. What matters, in the US
view, is the total treatment of imports, that is, the treatment with respect to all of
the legislation applied to a particular import transaction. The panel disagreed. In
its view, no rebalancing is permissible under Art I of the GATT. It is the consist-
ency of specific measures coming under the purview of Art I.1 of the GATT with
the multilateral disciplines that matters (§§ 6.10ff ).

3.10.7 The standard of review: no eff ect and no intent test

The panel in its report on EC—Bananas III held the view that all that the com-
plainant has to demonstrate in order to observe its burden of proof under Art I of
the GATT, is that the defendant has created more favourable competitive oppor-
tunities for some WTO members only. There is no need in its view, to also dem-
onstrate trade damage (effects) or (protectionist) intent (§ 7.239):
The requirement to match EC import licenses with BFA export certificates means that
those BFA banana suppliers who are initial holders of export certificates enjoy a commer-
cial advantage compared to banana suppliers from other third countries. We note that
it is not possible to ascertain how many of the initial BFA export certificate holders are
BFA banana producers or to what extent the tariff quota rent share that accrues to initial
holders of BFA export certificates is passed on to the producers of BFA bananas in a way
to create more favourable competitive opportunities for bananas of BFA origin. However,
we also note that the possibility does exist to pass on tariff quota rent to BFA banana
producers in such a way, whereas there is no such possibility in respect of non-BFA third-
country banana producers. Thus, the EC’s requirement affects the competitive relation-
ship between bananas of non-BFA third-country origin and bananas of BFA origin. It is
certainly true that Article I of GATT is concerned with the treatment of foreign products
originating from different foreign sources rather than with the treatment of the suppli-
ers of these products. In this respect, we note that the transfer of tariff quota rents which
would normally accrue to initial holders of EC import licenses to initial holders of BFA
export certificates does occur when bananas originating in Colombia, Costa Rica and
Nicaragua are, at some point, traded to the EC. Therefore, in our view, the requirement
to match EC import licenses with BFA export certificates and thus the commercial value
of export certificates are linked to the product at issue as required under Article I. In prac-
tice, from the perspective of EC importers who are Category A or C operators, bananas
of non-BFA third-country origin appear to be more profitable than bananas of BFA ori-
gin. This is confirmed by the fact that EC import licenses for non-BFA third-country
bananas and Category B licenses for BFA bananas are typically oversubscribed in the
first round of license allocations, while Category A and C licenses for BFA bananas are
usually exhausted only in the second round of the quarterly license allocation procedure.
The EC argues that the fact that licenses allowing the importation of non-BFA bananas
at in-quota tariff rates are usually exhausted in the first round amounts to an advantage
for bananas of Complainants’ origin. While we do not endorse the EC’s view, even if this
 Tariff Protection in the GATT 137
were to constitute an advantage, we note ‘that Article I:1 does not permit balancing more
favourable treatment under some procedure against a less favourable treatment under
others.’ (original emphasis)¹⁶⁵
This test (no need to show intent, effect) has been upheld by the AB and applied
in all cases coming under the purview of Art I of the GATT so far.¹⁶⁶

3.11 Special and differential treatment for developing countries

3.11.1 The Enabling clause
Developing countries felt, over the course of the years, that they could not com-
pete for export markets on an equal basis with developed countries. In their view,
the MFN tariff rate amounted to an impediment in that it provided for non-dis-
criminatory access to export markets irrespective of the level of development of
the exporting country. A new mechanism was needed and this mechanism was,
initially, a 10-year waiver allowing for preferential rates applicable to imports
from developing countries only; this was subsequently replaced by the Enabling
clause. The panel report on EC—Tariff Preferences recounts the inception of the
Enabling clause. Since this is an official narrative, we revert to it to elucidate the
rationale behind and the mechanics of the enabling clause:
During the Second Session of UNCTAD, on 26 March 1968, a Resolution was adopted
on expansion and Diversification of Exports and Manufactures and semi-manufactures of
Developing Countries’ (Resolution 21 (II)). In this Resolution, UNCTAD agreed to the
‘early establishment of a mutually acceptable system of generalized, non-reciprocal and
non-discriminatory preferences which would be beneficial to the developing countries’
and established a Special Committee on Preferences as a subsidiary organ of the Trade
and Development Board, with a mandate to settle the details of the GSP arrangements. In
1970, UNCTAD’s Special Committee on Preferences adopted Agreed Conclusions which
set up the agreed details of the GSP arrangement. UNCTAD’s Trade and Development
Board took note of these Agreed Conclusions on 13 January 1970. In accordance with the
Agreed Conclusions, certain developed GATT contracting parties sought a waiver for
the GSP from the GATT Council. The GATT granted a 10-year waiver on 25 June 1971.
Before the expiry of this waiver, the CONTRACTING PARTIES adopted a decision on
‘Differential and More Favourable Treatment, Reciprocity and Fuller Participation of
Developing Countries’ (the ‘Enabling Clause’) on 28 November 1979.
The Enabling clause is thus, the decision of the GATT CONTRACTING
PARTIES which allowed deviations from the MFN rate in favour of developing
countries to become a permanent feature of the GATT and now the WTO legal
order. It is through the Enabling clause that WTO members can accord tariff

¹⁶⁵ BFA stands for Bananas Framework Agreement, an agreement that the European
Community signed with several banana producing and exporting countries.
¹⁶⁶ Th is test can prove quite problematic, in practice. In Chapter 3, we explain in detail why an
intent test is probably required when evaluating claims that discrimination has been afforded.
138 Disciplines on Trade Instruments

preferences to developing countries, through their national generalized system of

preferences (GSP) schemes.

3.11.2 Some historical features

Part IV of the GATT reflects the original institutional attempt to allow for spe-
cial and differential treatment for developing countries. Part IV does not appear
in the first edition of the GATT. Actually, the story of trade and development
deserves a short narrative. The trading partners, for a host of reasons, ranging
from the desire to make the GATT attractive to outsiders to a genuine belief that
something should be done to favour the weaker partners, began to debate which
tools the trading system should put at the disposal of less developed nations.¹⁶⁷
Although one could trace sporadic initiatives in the very early years of the
GATT as well, the first time a comprehensive discussion on trade and develop-
ment took place in the GATT was in 1958 with the circulation of the Haberler
report. Haberler was requested by the GATT CONTRACTING PARTIES to
examine the issue, and ended up concluding that claims by the less developed
trading partners that the existing rules on trade liberalization would not neces-
sarily work to their advantage were not entirely unjustified.¹⁶⁸
In this report, both the short- and the long-term trends in commodity prices
and the factors influencing them were examined. The report concludes, inter alia,
that existing protectionist policies in the farm sector by developed (industrial-
ized) nations, as well as tariff escalation practices by many developed nations
were contributing factors to growth of developing countries. It is probably worth
recounting in this respect, that the United States obtained a waiver in 1955, which
allowed it to grossly subsidize its farm production over the subsequent years.
The European Community, on the other hand, followed a policy which made
its farm market practically impenetrable: by adopting the notorious ‘dual price’
system, any imported product, when imported into the European Community,
would be burdened with a customs duty which equaled the difference between its
world and its European price (variable levies). As a result, exports to the lucrative
European market were discouraged.
The consistency of the EC regime was doubtful at best. Some of its aspects were
challenged and many remain unchallenged. The reasons for the latter vary: the
European Community could in part hide behind Art XI.2 of the GATT which

¹⁶⁷ Trade of course, is only part of a development strategy and there are inherent limits to how
much development can be achieved through trade liberalization. On this score, T.N. Srinivasan
accurately, in a recent speech, criticized the characterization of the ongoing round as development
round: if it succeeds, in his view its results, by definition almost, will be limited. And if it fails,
it will give the world community the wrong signals about the WTO and what can be achieved
through the trade liberalization process. The flip-side of this argument is, of course, that absent
reforms of domestic policies, developing countries should not be expecting that development is
a realistic prospect. Tupy (2005) eloquently stated that ‘trade liberalisation as a cure for African
poverty is often over-emphasized. The main causes of African impoverishment are internal’.
¹⁶⁸ GATT (1958), Trends in International Trade, Geneva. Gottfried Haberler, of Harvard
University, was one of the best trade economists of his time.
 Tariff Protection in the GATT 139

allows for limited exceptions from the general prohibition to ban QRs; it could,
in part, hide behind the indulgence of its trading partners who did not want to
put into question the integration process of an entity whose key members were
largely responsible for two world wars in the previous 40 years. The reputation
costs that the United States—in light of its immunity from prosecution because
of the waiver they had previously obtained—would incur, if it were to decide to
challenge the European farm policy, should not be underestimated either. As a
result, the two most prominent markets kept their doors closed to imports of
farm products from the rest of the world.
The Haberler report made a series of recommendations to address the issue:
reduction of the existing protectionism was one of the measures suggested.
Importantly, it sensitized the trading partners to the fact that not all gain alike
from the existing regime; something needed to be done to address the concerns of
those who were being left behind.
On the other hand, many developing nations, influenced by the writings of
Prebish and Singer, adopted import substitution policies during the same period
(fifties and sixties), whereby they encouraged domestic production of goods that
would substitute imported goods. The argument for import substitution was jus-
tified as the adequate response to what was then termed ‘terms of trade pessimism’:
the idea that exports of developing countries were progressing at a slower pace
than total exports. A related idea was what became known as ‘elasticity pessimism’:
devaluation will improve trade balance assuming the Marshall/Lerner condition
holds, that is, the sum of import and export demand elasticities exceeds one in
absolute value. If elasticities are too low, other means (possibly QR) are needed to
change an adverse trade balance. There is almost no evidence that the elasticities
are so low, but that was the post-war fear of many developing countries, see Lal
(2000). At the same time, in that era, liberal market economies in the eyes of many
were discredited, and a strong argument in favour of government-driven economies
was being promoted by economists, including, among others, Prebish and Singer.
In this view, development essentially equaled industrialization, and under the
influence of terms of trade and elasticity pessimism, a strong argument in favour
of preferential access to developed countries’ markets was being advanced: absent
such a mechanism, developing countries could not secure the income needed to
buy capital goods that they could not manufacture by themselves. So, the recipe
in a nutshell was a combination of preferential access to developed countries’ mar-
kets coupled with closure to imports markets for developing nations.¹⁶⁹ The end
result would be an increase in income for developing nations that could, in turn,

¹⁶⁹ Recourse to protectionism was boosted, oddly enough, by the GATT institutional design.
Many developing countries were facing BoP problems. Now, the relatively speaking better
response to a BoP crisis would be devaluation (which, in turn might have beneficial effects on
the quantity of exported goods). However, in a world of fi xed parities, devaluation was not an
option. On the other hand, in part, Art XVIII of the GATT allowed GATT members to have
recourse to import restrictions to address their balance of payments. The conditions established in
this provision were fairly easy to meet. As a result, many GATT members privileged this option
140 Disciplines on Trade Instruments

purchase the capital goods necessary for their development. Developing countries,
persuaded by such arguments, started submitting their requests for a negotiation
on preferential tariff rates for developing countries only.
The next step comes in 1961, when the CONTRACTING PARTIES adopted a
Declaration on the Promotion of trade of less-developed countries. In this declar-
ation, the CONTRACTING PARTIES introduce many of the elements/conclu-
sions of the Haberler report. There is an increasing recognition that liberalization of
the trade in farm products is very much a concern for developing countries, as the
first comprehensive studies on the welfare implications of such a demarche start to
appear. At the same time, there are increasing voices to the effect that there is an
asymmetry in trade liberalization: in areas of interest to developed nations (manu-
factured products), liberalization (in the form of tariff reductions) was progressing
fast; in areas of interest to developing nations (farm products), this was hardly the
case. As stated above, there are limits to how much an international setting can do
to counteract asymmetries in bargaining power between its constituents. And of
course, those with an advantage have little incentive to cede their power.
During the Kennedy round of international trade negotiations (1962–7) the
Committee on Legal and Institutional Framework of GATT in Relation to Less-
Developed Countries (one of the negotiating groups), worked on a chapter on Trade and
Development. This chapter was finalised in a Special Session of the CONTRACTING
PARTIES, held from 17 November 1964 to 8 February 1965, and was added by
virtue of an amending protocol to the GATT where it now appears as Part IV. Part IV
came into effect on 27 June 1966. Part IV contained three new legal provisions:
(1) Principles and objectives (Art XXXVI of the GATT);
(2) Commitments (Art XXXVII of the GATT); and
(3) Joint action (Art XXXVIII of the GATT).
A look at the wording of each provision makes it clear that they are ‘best endeavours’
clauses, whereby GATT contracting parties recognize the validity of the claim that
something must be done to help developing countries, and a sub-set of the whole group,
the developed nations, promise to undertake in the future concrete action to this effect.
The concrete action came some years later through the establishment of the preference
schemes.
Let us take a brief look at the three provisions. GATT, Art XXXVI is a formal
recognition that market access for products of export interest to developing coun-
tries¹⁷⁰ has to be improved. GATT, Art XXXVI does not prescribe the measures
that should be adopted to this effect. It does, however, lay down one important

and ended up affecting the treatment of imports only. Had they devalued instead of invoking
Art XVIII of the GATT, their exports could have enjoyed the benefits resulting from devaluation.
¹⁷⁰ The degree of homogeneity across developing countries, as understood in those years, was
relatively higher in the early sixties. It was thus easier for them to reach consensus among them, and
formulate common demands that corresponded more or less to similar needs.
 Tariff Protection in the GATT 141

principle: non-reciprocity. To avoid any misunderstandings as to what this prin-

ciple should capture, § 8 of the provision pertinently reads:
The developed contracting parties do not expect reciprocity for commitments made by
them in trade negotiations to reduce or remove tariffs and other barriers to the trade of
less-developed contracting parties.
The Interpretative Note ad Art XXXVI of the GATT sheds some additional light
on this issue:
It is understood that the phrase “do not expect reciprocity” means, in accordance with
the objectives set forth in this Article, that the less-developed contracting parties should
not be expected, in the course of trade negotiations, to make contributions which are
inconsistent with their individual development, financial and trade needs, taking into
consideration past trade developments.
Later, during the final stages of the Kennedy round, this provision was further
interpreted as follows:
There will, therefore, be no balancing of concessions granted on products of interest to
developing countries by developed participants on the one hand and the contribution
which developing participants would make to the objective of trade liberalization on the
other and which it is agreed should be considered in the light of the development, finan-
cial and trade needs of developing countries themselves. It is, therefore, recognized that
the developing countries themselves must decide what contributions they can make.¹⁷¹
In a nutshell, Art XXXVI of the GATT should be understood as embodying
the principle that, from now, it should be expected that concessions by develop-
ing countries will first and foremost be compatible with their development pro-
cess.¹⁷² One would expect that, to the extent that they cannot influence terms
of trade, it would not make much economic sense for developing countries to
impose trade restrictions, absent some public order justification. As it turns out,
this instrument was more a means to provide developing countries with some
extra flexibility to be used for, essentially, any reasons they deemed appropriate.
GATT, Art XXXVII was some sort of a general clause recommending various
actions that developed states should undertake in order to help promote issues of
interest to developing countries: chief among them, the incitation to reduce the
gap between (high) barriers on processed goods, and (low) barriers on primary
products. This is the often quoted ‘tariff escalation’ argument. The validity of this
argument can be put into question: the problem seems to be the high tariff on

¹⁷¹ See GATT, COM.TD/W/37, p 9.

¹⁷² Rodrik (1999) perceptively observes that the negotiations in the fifties and the sixties coin-
cide with the excessive faith, in many developing countries, in import-substitution strategies. There
was a widespread belief that government intervention could achieve a lot in terms of promoting
development policies. The same author (pp 4ff ) laments that the excessive faith in government
intervention was substituted by excessive faith in what market openness can accomplish, marginal-
izing thus the role for useful government intervention in many developing countries.
142 Disciplines on Trade Instruments

processed goods, rather than the gap in protection between primary and proc-
essed goods. Hence, this item should not be placed in the agenda in terms of a
relative difference (gap in protection between primary and processed goods), but,
instead, in absolute terms (high protection on processed goods).¹⁷³
The remaining part of Art XXXVII of the GATT deals with issues that were
further detailed in subsequent agreements: for example, developed countries, when
imposing countervailing or AD duties, or introducing safeguard measures, were
to ‘have special regard to the trade interests’ of developing countries and ‘explore
all possibilities of constructive remedies before applying such measures’. In the
Uruguay round AD agreement, this provision has become a binding legal obliga-
tion. In the context of AD for example, it has been interpreted as an obligation to
examine the feasibility of introducing price undertakings on dumped imports ori-
ginating in developing countries, before AD duties are eventually imposed.
GATT, Art XXXVIII was meant to provide the institutional vehicle that
would make the best endeavours-clauses reflected in the two aforementioned
provisions happen: institutional arrangements for furthering the objectives of
Part IV should be made, collaboration to this effect with the United Nations and
its organs and agencies was envisaged, and some monitoring of the rate of growth
of the trade of developing countries should be introduced.
Besides the introduction of these provisions, the GATT CONTRACTING
PARTIES agreed on the establishment of the Committee on Trade and
Development (CTD), which was established in 1964. Its mandate was to review
the application of the provisions of Part IV. Also in 1964, the International Trade
Centre (ITC) was established, with the aim of promoting trade of developing
countries. The ITC became later a joint agency of United Nations Conference on
Trade and Development (UNCTAD) and GATT.
The feeling among developing countries has been that Part IV has fallen short
of substantively contributing to the development policies pursued. First, the
importance of trade should be viewed in terms of its relative impact: it is simply
not the case that trade liberalization is a substitute for development policies in
toto. Depending on the importance of international trade on the gross national
product (GNP), it can be an important or a relatively unimportant aspect of a
wider development strategy. On the other hand, even within these narrow(er)
parameters, Part IV could not have much of an impact in light of the fact that it
was, for all practical purposes, a set of best endeavour clauses, a list of ‘I wish to
do’ items, deprived of binding language in any direction.
Concrete steps are taken for the first time with the advent of the GSP schemes,
and their facilitation through the Enabling clause. Still, Part IV exercised some
influence to subsequent developments: the Enabling clause essentially reproduces
the non-reciprocity idea, first embedded in Art XXXVI.8 of the GATT.¹⁷⁴

¹⁷³ A series of discussions with Henrik Horn on this issue are acknowledged.
¹⁷⁴ On the history of preferences and the way forward, see, inter alia, Inama (2003), Keck and Low
(2003), Michalopoulos (2001), Srinivasan (1998), Wang and Winters (2000) and Whalley (1989).
 Tariff Protection in the GATT 143

In a way, Part IV could be seen as the instrument that first operated as the
institutional awareness that something needed to be done to address the concerns
expressed by developing countries, and then paved the way to concrete reforms of
the world trading system.

3.11.3 The Enabling clause in the WTO legal order

The AB, in its report on EC—Tariff Preferences, noted that the Enabling clause
has become an integral part of the GATT, by virtue of Art 1(b)(iv) of GATT
1994, since, as noted above, the Enabling clause was adopted as a decision of the
CONTRACTING PARTIES. The AB in its report on EC—Tariff Preferences,
notes that since the Enabling clause enables WTO members to grant tariff prefer-
ences to a subset of the WTO membership (namely the developing countries),¹⁷⁵
it constitutes a legal exception to Art I of the GATT (§ 99). The legal implication,
in the AB’s view, is that the Enabling clause takes precedence over Art I of the
GATT (§ 102). The next question that the WTO adjudicating bodies faced was
the allocation of the burden of proof. The AB, reversing the panel holdings in this
respect, held for the proposition that it is insufficient for a complaining party,
when challenging a measure taken pursuant to the Enabling clause, to claim
violation of Art I of the GATT only (§ 110). Due process considerations (§ 113)
require that the complaining party:
identify those provisions of the Enabling Clause with which the scheme is allegedly
inconsistent, without bearing the burden of establishing the facts necessary to support
such inconsistency. (§ 115, original emphasis)
Identification, of course, does not amount to an obligation to respect the standard
of review associated with a claim under the Enabling clause. All a complainant has
to do, arguably, is refer to the Enabling clause once the complainant anticipates
that this will be the legal defense privileged by the defendant. The soundness of
this approach can of course, be questioned. It would, probably, have made bet-
ter sense for the AB to go all the way and construct the Enabling clause not as an
exception to Art I of the GATT, but as a self-standing obligation: WTO members
apply one set of tariffs to imports from developed nations and another on imports

¹⁷⁵ The term ‘developing countries’, is nowhere defined in the GATT. A sub-set of developing
countries, the least developed countries (LLDCs) is defined for the purposes of the SCM Agreement,
in an annex to the SCM Agreement. This list reproduces the UN list on LLDCs, and is now being
used for purposes other than serving the SCM Agreement. LLDCs are by definition developing
countries. As far as the remaining membership of the developing countries group is concerned,
GATT/WTO practice reveals a practice where, by virtue of the self-election principle (in itself an
expression of the principle of sovereignty), WTO members will choose for themselves whether or
not they belong to this group. Depending on the criteria that one chooses to distinguish between
the two groups, we can note some abusive expressions of the self-election principle to this effect,
which have remained largely unchallenged. On the other hand, there are some abusive omissions of
developing countries from national GSP schemes as well. Th is issue is largely unresolved as a matter
of legal determination, and to a large extent practice becomes the guiding principle in distinguish-
ing developed from developing countries. Practice suggests that almost all OECD members are
considered to be developed countries, and thus do not benefit from inclusion into GSP schemes.
144 Disciplines on Trade Instruments

from developing countries. Complainants will carry a similar burden of proof irre-
spective of whether they attack violations of MFN, or of the Enabling clause since
the two provisions do not, in principle, allow for distinctions based on the origin
of the good for each sub-set of the WTO membership (developed, developing)
that they are concerned with. More than anything else, the functionality of the
AB innovation in the allocation of the burden of proof is hard to grasp.

3.11.4 The test for compliance with the Enabling clause

The panel report on EC—Tariff preferences dealt with the following situation:
India and Pakistan both benefited from the European Community’s GSP.
Pakistan, however, received extra preferences because it qualified under the
so-called Drug Arrangements, a scheme aimed at compensating those WTO
members adopting active policies against drug production and trafficking. India
complained that, by discriminating in favour of Pakistani imports, the European
Community was in violation of Art I:1 of the GATT, a claim upheld by the panel
(§ 7.60). The panel went on to examine to what extent recourse to the Enabling
clause could be offered as justification. In the panel’s view, the Enabling clause
requires that developed countries, absent a priori limitations, must, by virtue of
the term ‘non discriminatory’ featuring in footnote 3 of the Enabling clause, give
identical tariff preferences under GSP schemes to all developing countries with-
out any differentiation.
The AB reversed the panel’s findings in this respect. It started its analysis
(§ 157) by pointing to the terms used in § 3(c) of the Enabling clause, which
specifies that ‘differential and more favourable treatment’ provided under the
Enabling clause:
. . . shall in the case of such treatment accorded by developed contracting parties to devel-
oping countries be designed and, if necessary, modified, to respond positively to the
development, financial and trade needs of developing countries.
In its view, this paragraph made it plain that developments needs are not neces-
sarily shared to the same extent by all developing countries (§ 162),¹⁷⁶ and,
responding to such needs would consequently require that a GSP scheme may
be ‘non-discriminatory’ even if ‘identical’ tariff treatment is not accorded to all

¹⁷⁶ The AB suggests that the drafters could easily have inserted the term ‘all ’ before developing
countries, if they really wanted to drive home the point that no discrimination across developing
countries is permitted. Grossman and Sykes (2005) take issue with this statement arguing that the
AB has treated silence in a very inconsistent manner in its case-law. They argue that, by the same
token, the drafters could have inserted the term ‘certain’ before developing countries, if they wanted
to allow for discrimination. The fact that they did not is probably equally relevant for their intent.
They take the view that based not on silence, but on actual expression, the Enabling clause makes
one distinction only between developing and LLDCs. This is the only relevant distinction, in their
view. Howse (2003) has defended the view that neither the panel nor the AB should have entered
into any discussion of the issue at all since, in his view, the Enabling clause is not justiciable.
 Tariff Protection in the GATT 145

GSP beneficiaries (§ 165). As a result, additional preferences cannot be outright

excluded (§ 169). It went on to rule that:
in granting such differential tariff treatment, however, preference-granting countries are
required, by virtue of the term ‘non-discriminatory’, to ensure that identical treatment is
available to all similarly-situated GSP beneficiaries, that is, to all GSP beneficiaries that
have the ‘development, financial and trade needs’ to which the treatment in question is
intended to respond. (§ 173)
Applying its test to the specific case, the AB found that the Drug Arrangements
were not WTO consistent, since the European Community laid down a closed
list of beneficiaries (§§ 180 and 187). For its scheme to be deemed WTO con-
sistent, the European Community, in the AB’s view, would have to modify its
current Regulation so as to ensure that it reflects:
criteria or standards to provide a basis for distinguishing beneficiaries under the Drug
Arrangements from other GSP beneficiaries. (§ 188)
Accordingly, WTO members can distinguish between recipients of preferences
between developing countries, provided that they have established criteria to this
effect. The AB opened thus the door to extra preferences, but did not provide any
principles which could serve as benchmark to distinguish between acceptable and
(eventually) unacceptable criteria. Recall that the criteria for extra preferences in
this case were established unilaterally. It is at best debatable whether the donors
have incentives to adopt criteria that will promote development of the recipients
and not simply their own social preferences which might upset the prioritization
of development options for the recipients.¹⁷⁷

3.11.5 Is the candle worth the flame? (criticism of the Enabling clause)¹⁷⁸
Preferential schemes have undeniably had beneficial effects for developing coun-
tries. They are increasingly, however, becoming the object of criticism. In short,
the debate has shifted from an unambiguous ‘trade not aid’ perspective in the
eighties to ‘aid rather than (preferential) trade’ nowadays. The idea is that assum-
ing that one can target aid to the intended beneficiaries (and thus, eliminate
corruption, administrative costs), it could probably prove a more efficient tool
to increase the income of developing nations (the ultimate goal of preferential
schemes) than preferential rates. This is the first and probably more serious criti-
cism of the existing GSP schemes.
Second, because of preference erosion, developing countries have less of an
incentive to liberalize trade on an MFN-basis (as § 3 of the Enabling clause would
request them to do). In this vein, Limão (2006c) has persuasively argued that

¹⁷⁷ Diverging views have been expressed on this score, see, inter alia, Brenton (2003), Hoekman
(2004), Hoekman et al (2004), Rodrik (2002), Stevens (2002).
¹⁷⁸ This title is borrowed from Grossman and Sykes (2005).
146 Disciplines on Trade Instruments

beneficiaries of GSP schemes have little incentive to see their preferences eroded
as a result of MFN liberalization. Hence, absent some refocusing of the GSP
schemes, one risks seeing developing countries blocking MFN liberalization in
trade rounds.¹⁷⁹ By the same token, Hoekman et al (2001) ran the following
simulation: they assume that any tariff rate greater than 15% should be treated as
a tariff peak. If LLDCs benefit from 0% tariff rate on all tariff peaks, their exports
will increase by 11%; if all GSP beneficiaries benefit from 0% tariff rate on all
tariff peaks, then LLDCs exports will increase by 5% only; finally, if the MFN
rate for tariff peaks drops to 5%, then LLDCs exports will not increase at all. It
becomes obvious that LLDCs have little incentive to favour MFN liberalization.
Third, there is increasing empirical evidence pointing to the direction that the
benefits from GSP schemes should not be exaggerated. Grossman and Sykes (2005),
citing abundant and internally-consistent empirical evidence to this effect,¹⁸⁰
essentially conclude that the flame is not worth the candle. There is little support to
the proposition that GSP schemes have had substantial positive welfare effects on
recipients. Indeed, the GSP schemes by both the European Community and the
United States can hardly be considered to conform to the idea of generalized prefer-
ences: the EC-scheme distinguishes between non-sensitive, semi-sensitive, sensitive,
and very-sensitive products. Grossman and Sykes (2005)¹⁸¹ calculate that develop-
ing countries roughly receive tariff reductions of 100% (for non-sensitive), 65% (for
semi-sensitive), 30% (for sensitive), and 15% (for very sensitive products) from the
usual MFN rate for goods in each category. The export interest of most developing
countries is concentrated on the very-sensitive category of products.¹⁸²

¹⁷⁹ Erosion of preferences seems to have motivated the actions by many developing countries in
the NAMA group of the Doha round. Limão (2006) has argued in favour of introducing subsidies
in favour of developing countries to address the preference erosion issue.
¹⁸⁰ See Sapir and Lundberg (1984), Karsenty and Laird (1986), MacPhee and Oguledo (1991),
Brown (1987 and 1989).
¹⁸¹ Pp 7ff.
¹⁸² Th is is how the EC webpage describes the EC GSP in a nutshell: ’For the period 01.01.2006–
31.12.2008, there are three types of arrangement in force for beneficiary countries, under the EU’s
GSP in Regulation (EC) No 980/2005: all beneficiary countries enjoy the benefit of the general
arrangement; the special incentive arrangement for sustainable development and good govern-
ance (the “GSP+”) provides additional benefits for countries implementing certain international
standards in human and labour rights, environmental protection, the fight against drugs, and good
governance (see Commission Decision 2005/924/EC for the list of GSP+ beneficiary countries);
the special arrangement for the least-developed countries (LLDCs), also known as the “Everything
But Arms” (EBA) initiative, provides for the most favourable treatment of all, in the aim of grant-
ing the LLDCs “duty-free and quota-free” access to the EU’s market. Given the great number of
developing countries, differences between them—in terms of level of development—are huge. The
rationale of the GSP is that developing countries cannot compete with developed countries. At pre-
sent, some developing countries cannot even face the competition of other developing countries.
Thus, there is a need to target the tariff preferences available under the GSP to these least developed
countries, which need them most. In February 2001, the Council adopted the so-called “EBA
(Everything But Arms) Regulation” (Regulation (EC) 416/2001), granting duty-free access to
imports of all products from least developed countries without any quantitative restrictions, except
to arms and munitions. At present, 49 developing countries belong to the category of LLDC’s. The
provisions of the EBA Regulation (Council Regulation (EC) No 416/2001 of 28 February 2001)
 Tariff Protection in the GATT 147

By the same token, the US GSP scheme excludes communist countries from
its coverage, as well as countries that withhold the supply of vital commodity
resources (aimed at OPEC), countries that injure US commerce by affording
preferences to other developed countries, countries that do not enforce arbitral
awards in favour of US citizens, countries that aid terrorism, countries that do
not protect internationally recognized workers’ rights, and countries that do
not address child labour.¹⁸³ Under these circumstances, the beneficial welfare
implications of GSP schemes could be frustrated by arbitrary decisions based
on ‘impressions’ (likely driven by political/social preferences) on the part of the
developed country.
Fourth, Özden and Reinhardt (2003), in an empirical study, point to yet
another disturbing factor linked to GSP schemes: countries that gradually extri-
cated themselves from GSP schemes, subsequently undertook greater liberaliza-
tion than those that chose to retain their eligibility to participate in them. In a
similar vein, Sachs and Warner (1995) conclude that developing countries with
more liberal trade policies achieve higher rates of growth and development than
countries that are more protectionist.¹⁸⁴
Fifth, Grossman and Sykes (2005) correctly point out yet another salient fea-
ture: when discussing the EC—Tariff Preferences dispute, they demonstrate that
the European Community was essentially paying Pakistan (and all beneficiar-
ies of the Drug Arrangements) through India’s money, since trade diversion is
potentially very prevalent when extra preferences are being introduced. Now that
GSP+ benefits seem to be WTO consistent (post-EC—Tariff Preferences) trade
diversion looms as a very likely prospect with all negative externalities associated
with it.¹⁸⁵
Sixth, the complexities associated with administration of preferential schemes
should not be underestimated either. Most of these complexities are linked to the
preferential rules of origin (no cumulation, diagonal or total cumulation) which
make evidence of the origin of a product a particularly cumbersome exercise.

have been incorporated into the GSP Regulation (Council Regulation (EC) No 2501/2001). Only
imports of fresh bananas, rice and sugar are not fully liberalized immediately. Duties on those
products will be gradually reduced until duty free access will be granted for bananas in January
2006, for sugar in July 2009 and for rice in September 2009. In the meantime, there will be duty
free tariff quotas for rice and sugar (see the latest regulations for sugar quotas No 1381/2002 and
rice quotas 1401/2002 in the list of legislation). The EBA Regulation foresees that the special
arrangements for LLDC’s should be maintained for an unlimited period of time and not be subject
to the periodic renewal of the Community’s scheme of generalised preferences. Therefore, the date
of expiry of Council Regulation (EC) No 2501/2001 does not apply to its EBA provisions’.
¹⁸³ See 19 USC § 2461 and 2462 and the discussion in Grossman and Sykes (2005, pp 4ff ).
¹⁸⁴ In literature one often sees references to Korea and Chile as the two most prominent exam-
ples of countries which, post-exclusion from GSP schemes, unilaterally liberalized trade and
boosted their growth rates.
¹⁸⁵ One could legitimately make the argument, for example, that the bargaining position of
developing countries as a group will be substantially weaker since they will be all fighting each
other for export income from developed nations (donors).
148 Disciplines on Trade Instruments

Hudec took the view back in 1987 that GSP schemes were not that benefi-
cial to developing countries: they essentially blocked beneficiaries from adopting
changes beneficial to them;¹⁸⁶ in his view, developing countries would have been
better off by simply abandoning such schemes in exchange for non-discrimina-
tory access in the agricultural and textiles markets of donors. Hudec’s intuitions
have sadly been largely confirmed by subsequent research.

3.12 Preferential Trade Agreements (PTAs)

3.12.1 PTAs in the WTO: FTAs and CUs
The term ‘regional integration’, which is often used in literature, is probably mis-
leading: in essence, what the term aims to capture are preferential schemes that
deviate from the obligation not to discriminate. Not all such schemes are regional
in the sense of geographic proximity. One third of FTAs currently under investi-
gation are among countries that are not in geographic proximity: the number of
cross-regional schemes, has risen from six in 1995 to 25 in 2002.¹⁸⁷ The European
Community—Mexico, the Singapore—New Zealand, and the Mexico—Japan
FTAs are appropriate illustrations. We will be using the term ‘preferential trade
agreements’ (PTAs), which is used in the social sciences, and more appropriately
reflects the objective function of such schemes.
The WTO regime does not treat PTAs in a coherent manner: whereas the
GATT, in Art XXIV, distinguishes between FTAs and CUs, Art V of the General
Agreement on Trade in Services (GATs) only mentions one form of preferential
scheme, entitled economic integration. A GATS economic integration resembles
a GATT FTA, although there are some important differences between the two
schemes.
Turning to the GATT again, there is an overlap between a CU and an FTA:
in both, parties to the agreement will have to liberalize substantially all trade
between them, for their scheme to be deemed to be GATT consistent. The dif-
ference between these two forms is that in the case of a CU (contrary to what
happens in an FTA context) a common external commercial policy is also
established.

3.12.2 Globalization yes, preferences yes as well

Sapir (1998, p 718) mentions that when the WTO was established all but three of
its original members were parties to at least one of the 62 PTAs still in force (the
exceptions being Japan, Hong Kong, China, and Korea). Two of them (Japan,
and Korea) have been quite actively negotiating PTAs ever since. This seems
quite paradoxical in two ways: on the one hand, one would expect that, following

¹⁸⁶ A point confirmed by the subsequent study undertaken by Özden and Reinhardt (2003).
¹⁸⁷ Source: (2003) World Trade Report 2003, WTO: Geneva.
 Tariff Protection in the GATT 149

the substantial reduction of tariffs at the multilateral level, there would not be
a persuasive argument for going regional: in a world of low tariff protection, there
is not much to gain for countries that want to further reduce existing protec-
tion.¹⁸⁸ On the other hand, there is an obvious tension between globalization—
notwithstanding the abusive manner in which the term is currently used—and
preferential trade: it cannot be that countries simultaneously pursue the global
and the preferential perspective.¹⁸⁹
The increasing number of PTAs is a rather recent phenomenon. In abso-
lute terms, most of the PTAs notified to the WTO came into being after 1993.
Historically, the European Community championed the establishment of PTAs.
Many reasons help to explain this trend, but two stand out as probably the
dominant explanations:
(1) some of the partners were candidates for accession (and a preferential scheme
was thought to be the antechamber to the European Community), or
ex-colonies of individual European Community members, with which, it
was felt, some form of preferential trade should be established;
(2) trade policy was the only ‘genuine’ common policy of the European
Community in the realm of international relations: signing trade deals with
various partners carried an inherent positive externality, in that it affirmed
the European persona in the eyes of the world. A number of constituencies
favoured this demarche.¹⁹⁰
The example of the European Community was, in the nineties, emulated by the
United States which signed a series of agreements with states in the American
continent: it started with the FTA with Canada (CUSFTA), it moved to NAFTA
and has culminated in negotiations for the FTAA (Free Trade Area of the
Americas).¹⁹¹ Then it was South East Asia’s turn. Most of the preferential schemes
negotiated nowadays are between countries in that geographic area.
One explanation of the rush towards preferential trade could possibly be
that PTAs are not pursued solely for commercial (trade) purposes.¹⁹² It could
very well be the case that regionalism is the expression of the political will to
‘lock in’ policies and take distance from past ‘sinful’ behaviour (although this
has not necessarily always been the case). Thus, going preferential could serve as

¹⁸⁸ See on this score, Limão (2006a and 2006b).

¹⁸⁹ The official WTO website (<http://www.wto.org>) shows that PTAs have almost doubled
since the coming into being of the WTO.
¹⁹⁰ On EC regionalism, see the excellent analysis in Sapir (2000a and 2000b). Sapir explains the
ever-changing EC rationales for going preferential.
¹⁹¹ FTAA covers all states in the American continent from Canada to Argentina, with the excep-
tion of Cuba. The negotiations, at the time of writing, were ongoing. See the excellent analysis of
Limão (2006a) who discusses the implications of the negotiation for the United States.
¹⁹² Although, as P. Krishna (2005) perceptively has argued, before the Uruguay round, the slow
progress of multilateral talks in and of itself might have provided many countries with an incentive
to pursue the ‘regional’ path.
150 Disciplines on Trade Instruments

a signaling mechanism. Th is is what, at least partially, pushed Mexico to join

NAFTA. Such an explanation is not at odds with the current shape of the WTO
contract: the GATT/WTO regime does not provide a safe ‘lock in’ for domestic
policies: with the exception of Agreement on Trade Related Intellectual Property
Rights (TRIPs), the WTO remains essentially a negative integration type of
contract which does not impose common policies on its membership.¹⁹³ In a
broader sense, PTAs could be endogenous.¹⁹⁴,¹⁹⁵

3.12.3 What matters when discussing preferential trade?

The amount of PTAs has undeniably been increasing: recently, as a result of the
proliferation of preferential schemes, the preferential trade share of intra-PTAs
trade in merchandise imports rose (at the world level) from 43.2% to 51.2% (from
2000 to 2005).¹⁹⁶ However, unless the welfare implications of PTAs have been
identified, economists will not speak of a problem. Economists initially viewed
regional integration with scepticism: following the early work of Viner (1950), it
has been shown both as theoretical proposition and in terms of empirical proof
that, such schemes both create and divert trade, and, for a number of reasons
(ranging from political motives to adverse selection), empirical evidence sup-
ported the thesis that the latter, in practice, outweighed the former.¹⁹⁷ This does
not mean, however, that all PTAs are welfare reducing.¹⁹⁸ First, it could be the
case that the welfare implications are positive: indeed as the Kemp-Wan theorem
shows, trade diversion can be eliminated by reducing external tariffs so as to keep
trade with non-members to a PTA unchanged (keeping, in other words, prices
constant).¹⁹⁹ Second, the perspective matters: it could be that schemes which are

¹⁹³ Indeed, if trade diversion was the only motive for going regional, it would seem that in a
world where customs duties are quite low, there would be less incentive for governments to exercise
this option. In practice, however, the proliferation of PTAs coincides with the period of (relatively)
lower tariff protection. It could be, of course, that through PTAs WTO members address NTBs,
or, conversely, that trade diversion takes place through NTBs. There is evidence of the latter (Serra
et al, 1997 who concludes that rules of origin represent a formidable obstacle to trade creation),
and, except for the European Community, almost no evidence of the former.
¹⁹⁴ See Fontagne et al (2005) on this issue. See also the excellent analysis in Baldwin (1997) and
Winters (1996). Mattli (1999) provides a very comprehensive and analytical account of European
regionalism in its historical dimension.
¹⁹⁵ PTAs on the other hand are not signed across homogeneous countries. See the World Trade
Report 2003, op cit, pp 46ff. Many of these PTAs are among transition economies following the
collapse of the Soviet block.
¹⁹⁶ See the World Trade Report 2003, op cit, p 48.
¹⁹⁷ See the empirical evidence reflected in the World Trade Report 2003, op cit, pp 52ff.
¹⁹⁸ For an excellent analysis on this score, see Schiff and Winters (2003).
¹⁹⁹ The result in the Kemp-Wan theorem applies in a set of given circumstances (and it is not
clear how large the set is). It is not a universally true result in any sense. On the other hand, the
Kemp-Wan theorem is not a passage oblige in order to support a claim that PTAs are not necessarily
welfare reducing. It does not make much sense for a group of countries though to lower their exter-
nal tariff sufficiently to avoid trade diversion. So while Kemp-Wan rests on sound theory, its prac-
ticability is at best doubtful.A trading bloc, as Krugman (1991, p 10) observes ‘will normally have
more monopoly power in world trade than any of its members alone. The standard theory of the
 Tariff Protection in the GATT 151

not welfare maximizing from a world perspective, nevertheless make those par-
ticipating in them better off. More importantly, the quantification presupposes
some benchmark. Typically, the way to measure trade diversion is by comparing
the ex post (creation of the PTA) to the ex ante situation, or even to a situation
where the lower, preferential rate is applied on a MFN basis. But one might legit-
imately take the view that the alternative to trade diversion is not necessarily the
status quo that arises when PTAs are permitted. It could be the case that the coun-
tries concerned would not have the same MFN rate if members were not allowed
to have PTAs. Worse still, it could be the case that a country would even be reluc-
tant to join the WTO in the first place if PTAs could not be formed under broadly
the same conditions that WTO members have practiced over the years.²⁰⁰
Crucial to the establishment of the appropriate benchmark is to understand
why WTO members enter into preferential schemes, that is, an enquiry into their
motives. The motives for entering into such schemes are not easily discerned.
Hypothesizing about them, one can end up with various explanations: it could
be that some countries, frustrated with the slow pace of liberalizing trade at a
multilateral level, use PTAs as a second best; it could also be that such schemes
are ‘signaling mechanisms’: Mexico, as discussed above, by joining NAFTA, also
showed the world that it was abandoning its policies of the past and was espous-
ing a different model. It cannot be excluded that recourse to PTAs is privileged
because a country sees a PTA as a way to get investment/access to foreign technol-
ogy that can increase its income, or access to low-cost production of inputs that
can increase its ability to export certain products; or even that trade is a compo-
nent of a wider public policy package. Finally, it is probably worth recalling that,
at least initially, PTAs were truly regional. Geographic distance and cultural
factors might deter trade, and there are good reasons to believe that geographic
proximity coupled with cultural affinity (a rather frequent combination among
regional partners) may go in the opposite direction.²⁰¹ The appropriate counter-
factual (how would international trade look like absent a given PTA) cannot be

optimal tariff tells us that the optimal tariff for a country acting unilaterally to improve its terms
of trade is higher, the lower the elasticity of world demand for its exports. So for a trading bloc
attempting to maximize the welfare of its residents, the optimal tariff rate will normally be higher
than the optimal tariff rates of its constituent countries acting individually’.
²⁰⁰ Moreover, trade diversion proponents often assume that trade is diverted to the less efficient
source. Assuming that this is not the case, however, assuming that is that those with the initiative to
go preferential choose their preferential trading partners ‘correctly’, no trade diversion will result.
²⁰¹ Krugman (1991) and Summers (1991) have argued that PTAs among countries in geographic
proximity should be encouraged, whereas PTAs among countries which are not neighbours (in a
geographic sense) should be discouraged. In their analysis, the former are more likely to avoid the
adverse possibility of welfare reduction and to lead to a larger improvement in welfare. In this line
of thinking, PTAs among geographically proximate countries is ‘natural’, and the potential losses
from trade diversion are limited, and the potential gains from trade creation, large. It is question-
able how strong this claim is since, there is empirical evidence that there are substantial border
effects. Anderson and van Wincoop (2003), for example, have estimated that national borders
reduce trade between industrialized countries by moderate amounts of 20%–50%.
152 Disciplines on Trade Instruments

established absent such an enquiry. Consequently, to decide whether PTAs are

stumbling blocks or not (as Bhagwati and Panagariya (1999) put it), one needs
to enquire on a case-by-case basis, having first established the appropriate bench-
mark (counterfactual) for quantification of the welfare implications stemming
from the creation of a PTA.
Does the legal test enshrined in Art XXIV of the GATT do that? No, is the
short answer. The GATT legal system does not enquire into the motives of coun-
tries entering into PTAs; nor does it care about the welfare implications. It sim-
ply aims to make departures from MFN onerous; the welfare implications of
PTAs will only tangentially, if at all, be discussed when the consistency of a PTA
with the multilateral rules is being evaluated.²⁰² On the other hand, as will be
explained in detail, no serious enforcement of the relevant GATT provision has
ever occurred. One could thus characterize the economist’s and the lawyer’s look
at PTAs as two ships passing by in the night.
In this section, we do not deal with PTAs formed to promote development
of developing countries either (usually referred to as ‘south—south cooperation’).
This is a possibility under § 2c of the enabling clause (L/4903 of 28 November
1989), and 21 such schemes have so far been notified to the WTO.²⁰³ The rest
of the paper is relevant for the study of the 112 notified PTAs (101 FTAs, and 11
CUs).²⁰⁴

3.12.4 PTAs in the GATT

3.12.4.1 No per se inconsistency
GATT, Art XXIV.4 makes it plain that PTAs can happily co-exist within the
multilateral (WTO) edifice. It reads in this respect:
The contracting parties recognize the desirability of increasing freedom of trade by the
development, through voluntary agreements, of closer integration between the econ-
omies of the countries parties to such agreements. They also recognize that the purpose
of a customs union or of a free-trade area should be to facilitate trade between the con-
stituent territories and not to raise barriers to the trade of other contracting parties with
such territories.

²⁰² Limao (2006a) and (2006b) suggests that even if Art XXIV of the GATT could provide a
legal test that stopped all trade-diverting PTAs from being formed, and maybe abide by some kind
of Kemp-Wan ideal rule, the interest created to preserve preferences once granted, would push
against further MFN liberalization. So, maybe, an ideal Art XXIV of the GATT could stop PTAs
that caused a lot of static trade diversion, and in this way prevent PTAs that eroded the benefits
of existing MFN liberalization; but even this ideal Art XXIV of the GATT would not be expected
to be able to stop PTAs that would slow down the process of further MFN liberalization. Similar
suggestions had been voiced by Krishna (1998 and 2005).
²⁰³ According to the information provided in the WTO official website (<http://www.wto.org>).
Such schemes are notified to the CTD and not to the Committee on Regional Trade Agreements
(CRTA).
²⁰⁴ Source: <http://www.wto.org>.
 Tariff Protection in the GATT 153

Hence, this provision opens the way for GATT consistent PTAs. The specific
conditions for satisfying consistency with the multilateral rules are laid down in
Arts XXIV.7, XXIV.5/6 and XXIV.8 of the GATT.

3.12.4.2 The Art XXIV of the GATT test in a nutshell

GATT, Art XXIV distinguishes between two forms of PTAs: CU and FTA.
GATT, Art XXIV imposes three obligations on WTO members wishing to enter
into a PTA, the first of a procedural, and the latter two of a substantive nature:
(1) an obligation to notify the PTA. Since regional integration essentially
amounts to an exception from the basic obligation to treat international trade
in a non-discriminatory manner, the resulting legal consequence is that WTO
members wishing to enter into a PTA, and thereby deviate from the obliga-
tion to treat trade from all other WTO members in a non-discriminatory
manner, will have the burden of proof to state that they have complied with
the relevant multilateral rules. The first step towards meeting this obligation
comes with the notification of the scheme;
(2) an obligation to liberalize among constituents of the PTA substantially all
trade (internal requirement); and
(3) an obligation not to raise the overall level of protection and make access of
products of third parties not participating in the PTA more onerous (external
requirement).

3.12.4.3 Designed at home, approved (?) in Geneva

There are two possibilities available in the GATT to review the consistency of
a PTA with the multilateral rules. First, the multilateral track: WTO members
have to notify PTAs they enter into to the WTO CRTA (Track I).²⁰⁵ Second,
the bilateral track: WTO members may challenge the consistency of a PTA
with the multilateral rules through panel proceedings under the DSU (Track
II). The substantive requirements for reviewing consistency of a PTA are iden-
tical under Track I and Track II. GATT, Art XXIV does not put into question
the rationale for going regional: this is considered a sovereign, political decision;
only the conditions under which a PTA can legitimately be formed are put into
question.
In the past, voices have been raised as to whether Track II should be available
at all. The most eloquent proponent of a negative response to this question has
been Roessler (2000), who argues that, for reasons having to do with the institu-
tional balance of the WTO, a limited judicial review by WTO adjudicating bod-
ies is the most appropriate one. A similar argument (if not identical, altogether)

²⁰⁵ The CRTA represents the ‘institutional consolidation’ of the notorious Art XXIV working
parties which would be established on an ad hoc basis to review the consistency of any given PTA
with the GATT rules. Hereafter, when referring to CRTA practice, I will include the Art XXIV
working parties practice as well.
154 Disciplines on Trade Instruments

was raised by India in the context of its litigation mentioned infra. India argued
that the question as to whether a restriction can be justified on BoP grounds is
inherently political (borrowing from the political question doctrine, known in
some legal orders) and this is why such issues have been entrusted to committees
and not to panels. In other words, the argument could be raised that the consist-
ency of a PTA with the WTO is of political nature, and thus not justiciable. The
AB rejected the argument by India essentially on textual grounds.²⁰⁶
The wording of the WTO Understanding on Art XXIV of the GATT (adopted
during the Uruguay round) tends to support the view espoused by the AB. In
§ 12 it reads:
The provisions of Articles XXII and XXIII of GATT 1994 as elaborated and applied by
the Dispute Settlement Understanding may be invoked with respect to any matters aris-
ing from the application of those provisions of Article XXIV relating to customs unions,
free-trade areas or interim agreements leading to the formation of a customs union or a
free-trade area. (emphasis added)
Hence, on its face, the understanding seems to grant a large review power to WTO
adjudicating bodies.²⁰⁷ WTO practice confirms this view. Having said that, one
can legitimately ask, whether WTO panels are well-equipped to deal with such
complicated issues. Beyond expertise, time constraints alone might prove an obs-
tacle for panels wishing to perform a comprehensive review of a PTA. As will be
shown infra, some GATT panels opted for a limited review of PTAs by panels, while
it is very difficult (in light of the very limited practice) to confirm whether WTO
panels are eager to follow the same path. The AB, at any rate, seems to espouse the
view that a comprehensive review of a PTA by panels is not unthinkable.
In sum, the two tracks are available, although the extent of review in the
bilateral track is probably debatable. There remains one question to ask: what
is the interplay between Tracks I and II? Assume, for instance, that a PTA is
simultaneously before a panel and the CRTA. Can this happen? In theory, yes,
since there is no obligation to suspend panel proceedings while an issue is being
discussed before a WTO committee.²⁰⁸ If the CRTA concludes by consensus

²⁰⁶ See India—Quantitative Restrictions at §§ 98ff. As noted above, the genuinely political
question (why opt for a PTA?) is not put into question at all by Art XXIV of the GATT. On the
other hand, it is not accidental that the body of Art XXI of the GATT which deals with the security
exception—probably, the only genuine political question—reads in a manner that makes it obvi-
ous that the margin of discretion rests primarily with the state invoking the exception. Th is is not
the case of Art XXIV of the GATT which acknowledges the discretion of multilateral organs to
decide whether a given PTA is in conformity with WTO law.
²⁰⁷ Roessler (2000) argues that the terms of the understanding lean towards a restrictive under-
standing of its scope: the reference made is to the application of Art XXIV and not to Art XXIV of
the GATT as such. However, even if such a dichotomy were true, a WTO panel adjudicating on an
application of Art XXIV of the GATT, will, inevitably, still have to provide its understanding of
the term reflected in this provision.
²⁰⁸ To complete the picture here, Art 12.12 of the DSU allows the complaining party to request
suspension of panel proceedings. This is a right bestowed upon the complainant, and not an
 Tariff Protection in the GATT 155

on the (in-) consistency of the notified PTA with the multilateral rules (a rather
improbable scenario in light of past experience), there is good reason to believe
that the panel subsequently dealing with the issue will follow the opinion reflected
in the CRTA. To the extent that it can serve as guidance, we find it plausible that
panels try to emulate the attitude of panel on India—Quantitative Restrictions
on this score, which, while dealing with a similar issue (eg to what extent a panel
dealing with an issue which had already been decided by the WTO Balance of
Payments Committee) relevantly provided in § 5.94:
. . . we see no reason to assume that the panel would not appropriately take those conclu-
sions into account. If the nature of the conclusions were binding . . . a panel should respect
them.
The same should be true for panels dealing with PTA-related issues: the compre-
hensiveness of the review undertaken before the CRTA is a good reason argu-
ing in favour of a thorough examination by the panel of the reasoning and the
outcome provided by the CRTA. There is, however, no legal compulsion for the
panel to follow a CRTA decision.
Should panels stop short of deciding whether a PTA is WTO consistent,
in the instance when the CRTA has not yet pronounced on its consistency (lis
pendens)?²⁰⁹ Such an approach, in light of today’s institutional realities, is hardly
something to recommend. If panels were to behave in this way, they would risk
depriving WTO members of their MFN rights: the CRTA will invariably take
a long time to reach consensus, and, as practice shows, the consensus will, in all
likelihood, reflect an agreement to disagree.²¹⁰ On the other hand, should the
CRTA be bound by a panel’s (and or AB’s) decision on the consistency of a PTA
with the relevant WTO rules? This seems to be a likelier scenario in light of the
time constraints that panels have to adhere to and the absence of such constraints
when the CRTA reviews a scheme. The formal answer has to be, once again, no.
The legal effect of the judiciary’s decision is not such that it acknowledges the force
of res judicata (binding any discretion of the CRTA to subsequently deviate from
its reasoning/outcome).
Hence, in principle, one cannot exclude the possibility that the CRTA and a
panel can reach different conclusions on the same issue. Because the former, as

obligation to behave in this way assuming certain contingencies (eg discussions of the issue before
a WTO committee) have been met.
²⁰⁹ Th is is not the same question as the one raised by India in India—Quantitative Restrictions.
In that case, India put into question the competence of the panel to adjudicate the dispute altogether,
arguing that, was the panel to be acknowledged jurisdiction over that dispute, the institutional bal-
ance of the WTO would be undone. Here we assume that panels are competent to adjudicate a dis-
pute and ask the question whether they should defer judgment until the ratione materiae competent
WTO organ has pronounced on this score.
²¹⁰ For a relatively recent expression of a typical disagreement, see §§ 31–8 of the Working
Party report on the Free Trade Agreement between EFTA and Turkey, adopted on 17 December 1993
(L/7336).
156 Disciplines on Trade Instruments

will be explained in what follows is unlikely to end up with a unanimous vote,

this possibility has never, so far, materialized in practice.²¹¹

3.12.4.4 The multilateral track

WTO members deciding to enter into a regional integration scheme have to
notify the WTO of their intention to do so (Art XXIV.7 of the GATT). The
CRTA is the WTO ratione materiae competent organ which will receive notifica-
tions and examine the compatibility of the notified scheme with the multilateral
rules.
The CRTA²¹² is the successor to the Art XXIV of the GATT working parties.
Regional schemes were routinely reviewed during the GATT years as well: an
ad hoc Art XXIV working party would be established, to perform the task now
entrusted to the CRTA. Participation in working parties was open to all inter-
ested GATT Contracting Parties. As is the case with the CRTA, Art XXIV of the
GATT working parties decide by consensus. The CRTA was established through
a decision by the WTO General Council on 7 February 1996.²¹³ All WTO mem-
bers can participate in the CRTA. The mandate of the CRTA is described in the
decision of 7 February 1996:
(a) to carry out the examination of agreements in accordance with the procedures and
terms of reference adopted by . . . and thereafter present its report to the relevant body
for appropriate action;
...
(a) to consider the systemic implications of such agreements and regional initiatives
for the multilateral trading system . . .
The CRTA decides by consensus. Rule 33 of the Rules of Procedure for Meetings
of the Committee on Regional Trade Agreements,²¹⁴ stipulates that:
Where a decision cannot be arrived at by consensus, the matter at issue shall be referred,
as appropriate, to the General Council, the Council for Trade in Goods, the Council for
Trade in Services or the Committee on Trade and Development.
In principle, the CRTA has wide powers. GATT, Art XXIV.7 relevantly provides
that it has the power:
. . . to make such reports and recommendations to contracting parties as they may deem
appropriate. (emphasis added)
In principle, hence, one cannot exclude the possibility that the CRTA concludes
that a notified PTA is GATT inconsistent (although such a conclusion seems a

²¹¹ To a large extent, this is not a PTA-specific issue: the WTO contract nowhere regulates the
status of ‘secondary’ law. This is not unusual, since the WTO Agreement has not been conceived by
its founding fathers to be of constitutional order.
²¹² For a more detailed discussion on the inception of the CRTA, see Mavroidis (2005).
²¹³ See WTO Doc WT/L/127.
²¹⁴ See WTO Doc WT/REG/1 of 14 August 1996.
 Tariff Protection in the GATT 157

priori, highly unlikely, in light of the fact that WTO members participating in
the PTA will be voting as well). This conclusion is underscored by the explicit
wording of Art XXIV.7(b) of the GATT which explains the powers of the CRTA
when it reviews an interim agreement leading to the establishment of a CU or
an FTA:
If . . . the CONTRACTING PARTIES find that such agreement is not likely to result
in the formation of a customs union or of a free-trade area . . . the CONTRACTING
PARTIES shall make recommendations to the parties to the agreement. The parties shall
not maintain or put into force, as the case may be, such agreement if they are not prepared to
modify it in accordance with these recommendations. (emphasis added)
GATT CONTRACTING PARTIES have, however, never in the history of the
GATT, reached (by consensus) a decision that a notified scheme was inconsist-
ent with the multilateral rules—a practice continued in the WTO era. CRTA
reports are adopted irrespective of whether CRTA members have unanimously
decided on the (in-) consistency of the PTA reviewed; it is normally the case that
an adopted report reflects divergent views on this issue. In fact, this is the usual
case, as the record examined below demonstrates.
GATT, Art XXIV :7(a) reads :
Any contracting party deciding to enter into a customs union or free-trade area, or an
interim agreement leading to the formation of such a union or area, shall promptly notify
the CONTRACTING PARTIES and shall make available . . . such information . . . as
will enable them to make such reports and recommendations to contracting parties as
they may deem appropriate. (emphasis added)
The language of Art XXIV of the GATT suggests that prospective action is being
notified. The language stops short of mentioning that Art XXIV of the GATT
operates as a ‘green light’, which must be complied with, for a PTA to be GATT
compatible. It could, of course, be legitimately argued that since what is noti-
fied is prospective action and since the consistency of prospective action is the
subject-matter of multilateral review, WTO members should refrain from oper-
ationalizing their PTAs before they are given the ‘green light’ to do so. The CRTA
was probably intended to serve a function similar to that of an antitrust authority
when examining a merger: the merger cannot be consummated absent a ‘green
light’.²¹⁵
Most of the time, however, especially recently, regional integration schemes
have been notified ex post facto (after their establishment), contrary to the wording
of Art XXIV:7(a). For example, the NAFTA was signed on 17 December 1992,

²¹⁵ In July 2006, the General Council adopted a Transparency Mechanism for PTAs aimed,
as the title suggests, at enhancing the transparency for PTAs: all PTAs would be notified as early
as possible and, anyway, immediately following their ratification by participants. The WTO
Secretariat (members of the Trade Policy Review Mechanism (TPRM) Division) would then pre-
pare a factual presentation of the PTA that would be circulated to all WTO members, see WTO
Doc TN/RL/18 of 13 July 2006.
158 Disciplines on Trade Instruments

entered into force on 1 January 1994 and a working party to examine its consist-
ency with the GATT rules was established only on 23 March 1994. The respect-
ive dates for the EC—Visegrad Agreements (FTAs with Hungary, Poland, and
the Czech and Slovak Federal Republic Interim Agreement) are 16 December
1991, and 30 April 1992.²¹⁶
Consequently, working parties (and now the CRTA), have often been pre-
sented with a fait accompli: at the time of writing, for seven PTAs that have already
entered into force, no request for examination of their consistency with Art XXIV
of the GATT has been submitted; 38 PTAs in force still await the beginning of
the factual review; 17 of them are under factual examination; finally, on five of
them (including NAFTA that came into force over 10 years ago) consultations on
the draft report are being held at the time of writing.²¹⁷
What was originally supposed to be an ex ante review became slowly an ex post
review with all the problems that such a shift in time might entail. Delayed notifi-
cations have a domino effect: indeed, what would the remedy be in the case where
a PTA in force is found to be WTO inconsistent?²¹⁸ Since PTAs represent, in
principle, a permanent deviation from MFN, members of a GATT inconsistent
PTA would have to compensate continuously their trading partners for all trade
diversion (assuming correct measurement, that is, establishment of the proper
counterfactual) caused by the PTA. On the other hand, how realistic is it to see a
suggestion by a WTO panel to dissolve an existing PTA? Delayed notifications,
it seems, are a contributing factor towards tolerance of PTAs of unknown and
probably doubtful GATT consistency.²¹⁹
Hence, with respect to the question whether a green light by the CRTA
is necessary for a PTA to lawfully enter into force, practice strongly suggests
that the answer is no. One WTO panel report (US—Line Pipe) has sided
with this view. Hence, the value of a CRTA report, if any is to be realized, is in
providing transparency as to the views that WTO members hold on the inter-
pretation of Art XXIV of the GATT.²²⁰ We turn to this issue in what immedi-
ately follows.

²¹⁶ A survey of the respective dates of all PTAs (some information to this effect appears in the
WTO homepage, <http://www.wto.org>) amply supports this conclusion.
²¹⁷ Source: <http://www.wto.org>.
²¹⁸ Since, even if they meet the substantive requirements, they have failed to meet the procedural
requirement.
²¹⁹ Note that, during the Doha round, the WTO members reaffirmed their preference for early
notification. The Negotiating Group on Rules adopted a decision (WTO Doc JOB(06)/59/Rev.5 of
29 June 2006) obliging WTO members acceding to a PTA to notify their PTA immediately upon
its ratification (Rule B3). The same decision provides in an annex that WTO members are under
a duty to notify concessions made in the context of the PTA at a tariff line-level (at the eight, 10 or
more digit level, depending on national classifications), any preferential rules of origin applied, as
well as the MFN rates and import statistics applied before the advent of the PTA.
²²⁰ Although time inconsistencies should not be excluded on this score.
 Tariff Protection in the GATT 159

(a) The external requirement Broadly speaking, by virtue of the external

requirement, WTO members should not, when entering into a PTA, raise their
protection vis-à-vis their (non participants in the PTA) trading partners. Because
entering into a CU entails the establishment of a common external policy,
the requirements for meeting the external requirement are not the same for FTAs
and CUs.
With respect to FTAs, Art XXIV.5(b) of the GATT reads:
. . . duties and other regulations of commerce . . . shall not be higher or more restrictive
than the corresponding duties and other regulations of commerce existing in the same
constituent territories prior to the formation of the free-trade area . . .
GATT, Art XXIV.5(b) does not explicitly state that WTO members participat-
ing in an FTA cannot modify their external protection when joining the FTA.
An imperative to this effect could have been dictated by the very nature of an
FTA: it aims only at liberalizing trade within its constituents, without address-
ing the question of external protection at all. Hence, WTO members retain their
sovereignty with respect to external protection, when joining an FTA: there is no
reason to see changes in this respect. As QRs are prohibited anyway by virtue of
Art XI of the GATT, and tariffs will continue to be unilaterally defined by mem-
bers of the FTA, the heart of the obligation embedded in Art XXIV.5(b) of the
GATT hinges on the interpretation of the term ‘other regulations of commerce’.
Note that the obligation enshrined in this legal provision does not refer to
duties and other regulations of commerce as a whole, but to individual instru-
ments: the obligation assumed thus, is not, in principle, to come to more or less the
pre-FTA situation by rebalancing various instruments; it is to ensure that each
and every individual trade instrument will not become more restrictive, post-
establishment of the FTA.
GATT, Art XXIV.5(b) does not contain an exhaustive list of regulations of
commerce other than duties. In practice, the instrument most likely to be changed
as a result of the establishment of a PTA is the rules of origin. Recall that there
is no WTO discipline, besides a transparency obligation, regarding preferential
rules of origin. Rules of origin, nevertheless, are of particular interest in an FTA
context: unless goods circulating through an FTA are accompanied by a certifi-
cate of origin, exporters will have an incentive to ship to the cheapest port of entry
(since external protection remains an issue of national sovereignty and it could
very well be the case that there are asymmetries as to the level of customs duties
among members of an FTA). On the other hand, by adopting for example, prefer-
ential rules of origin, members of an FTA can drastically affect trade flows.
Since PTAs are meant to allow the conferral of WTO consistent preferences,
even in the absence of disciplines on preferential rules of origin, the rules of origin
contracted in the context of a PTA should be more favourable (or at the very least,
not more burdensome) than those applied on an MFN basis. Adopted reports
evidence no discussion of the term ‘preferential rules of origin’ (and the ensuing
160 Disciplines on Trade Instruments

obligations) from a legal perspective, although factual information on preferen-

tial rules of origin might occasionally be provided. Is this a problem? Well, yes.
Serra et al (1997), as well as the studies mentioned supra, show that rules of origin
are one of the most important factors of trade diversion. On such an issue, we
have had no benefit of a thorough discussion in the GATT context: we have thus,
anecdotal evidence suggesting that preferential rules of origin are anything but
preferential, but no conclusive discussion on the consistency of such rules with
the spirit of Art XXIV of the GATT.
With respect to CUs, the external requirement is distinguished into two spe-
cific obligations: an obligation not to raise the overall level of protection above a
certain threshold and a specific obligation to compensate in cases where customs
duties in some constituents of a CU had to be raised to match the CU-level.
Suppose country A and country B choose to enter into a CU. Country A has a
system of chicken quotas for imports of poultry from country C, but B does not
have any such restrictions. To achieve a common external commercial policy, A
and B need to harmonize these divergent trade policies towards C. If country A
is unwilling to give up its chicken quotas, then the result of harmonization will
be the imposition of a new trade barrier by country B against country C, ie one
that did not exist before the CU was formed. Moving towards a common exter-
nal policy might involve negotiations where re-adjustments to such policies take
place. GATT, Art XXIV.5(a) adopts a flexible approach on this score, allowing
for readjustments:
. . . duties and other regulations of commerce . . . shall not on the whole be higher or more
restrictive than the general incidence of the duties and regulations of commerce applicable
in the constituent territories prior to the formation of such union . . . (emphasis added)

The italicized words mark the difference between the text of Art XXIV.5(b) of the
GATT and that of Art XXIV.5(a) of the GATT. At first, the words seem to sug-
gest the inevitable need for adjustment. On the whole and general incidence invite
a comparison of the general (and not item by item) situation before and after the
formation of the CU. Incidentally, this seems to have been the intention of the
drafters. Quoting from the preparatory materials:
The phrase ‘on the whole’ . . . did not mean that an average tariff should be laid down in
respect of each individual product, but merely that the whole level of tariffs of a customs
union should not be higher than the average overall level of the former constituent terri-
tories», (GATT Doc. EPCT/C.II/38 at p. 9 reproduced in the GATT Analytical Index:
Guide to GATT Law and Practice, Updated 6th Edition (1995) at p. 803).
The Sub-Committee recommended that the words ‘average level of duties’ be replaced
by ‘general incidence of duties’ in paragraph 2(a) of the new Article. It was the intention
of the Sub-Committee that this phrase should not require a mathematical average of
customs duties but should permit greater flexibility so that the volume of trade may be
taken into account. (Havana Reports reproduced in the GATT Analytical Index, op cit,
at p 803)
 Tariff Protection in the GATT 161

Subsequent practice seems to accept that in principle an item-by-item approach

is unwarranted in the context of Art XXIV.5(a) of the GATT, but there is dis-
agreement as to the precise level on which comparisons will take place. The 1983
working party report on Accession of Greece to the European Communities contains
the view expressed by the European Community, that:
Article XXIV.5 required only generalized, overall judgment on this point.²²¹
By the same token, the 1988 working party report on Accession of Portugal and
Spain to the European Communities contains the view of the EC that:
Article XXIV.5 only required an examination on the broadest possible basis.²²²
This view, however, failed to convince other members of the working party. The
same report, though, contains the view of a member of the working party which:
could not accept the Communities’ contention that the extension of the tariff of the
EC/10 to the EC/12 was compatible with their obligations under Article XXIV.5(a)
regardless of the effect on the tariffs of Spain and Portugal. Article XXIV.5(a) required
a comparison with the pre-accession tariffs of the constituent territories and the relative
size of those territories was not a relevant factor.²²³
The report (at p 311) also contains the view of a member of the working party
which:
could not accept the Communities’ contention that the extension of the tariff of the
EC/10 to the EC/12 was compatible with their obligations under Article XXIV.5(a)
regardless of the effect on the tariffs of Spain and Portugal. Article XXIV.5(a) required
a comparison with the pre-accession tariffs of the constituent territories and the relative
size of those territories was not a relevant factor.
Disagreements appeared often among working party members as to whether
bound or applied rates should be used in the context of Art XXIV.5(a) of the
GATT.²²⁴
The picture seems much clearer now on this score, with the entry into force of
the WTO Understanding on the Interpretation of Article XXIV of the General
Agreement on Tariffs and Trade 1994:
The evaluation under paragraph 5(a) of Article XXIV of the general incidence of the duties
and other regulations of commerce applicable before and after the formation of a customs

²²¹ See BISD 30S/168, p 184.

²²² See BISD 35S/293, pp 295–6.
²²³ Idem, p 311.
²²⁴ See for example the discussions of the working party examining the compatibility of the
EEC with Art XXIV, GATT Doc SR.18/4, pp 46–54 and also in C/M/8, SR.19/6–7; see the
Working Party report on Accession of Greece to the European Communities, op cit, p 175; see also
the 1991 working party report on Free Trade Agreement Between Canada and the United States,
BISD 38S/47, p 66. As noted above, such disagreements have been put to rest with the advent of the
Understanding on Art XXIV of the GATT.
162 Disciplines on Trade Instruments
union shall in respect of duties and charges be based upon an overall assessment of weighted
average tariff rates and of customs duties collected. This assessment shall be based on import
statistics for a previous representative period to be supplied by the customs union, on a
tariff-line basis and in values and quantities, broken down by WTO country of origin. The
Secretariat shall compute the weighted average tariff rates and customs duties collected in
accordance with the methodology used in the assessment of tariff offers in the Uruguay
Round of Multilateral Trade negotiations. For this purpose, the duties and charges to be
taken into consideration shall be the applied rates of duty. It is recognized that for the pur-
pose of the overall assessment of the incidence of other regulations of commerce for which
quantification and aggregation are difficult, the examination of individual measures, regu-
lations, products covered and trade flows affected may be required. (emphasis added)²²⁵
So it is the applied rate that matters. Note, however, that the understanding does not
impose the same obligation (to use applied rates) with respect to FTAs. This should
mean that, when entering into an FTA, a WTO member should be in a position to
raise its protection from the applied to the bound level. It will thus have, by virtue of
Art XXIV.5(b) of the GATT, to ensure that its customs duties in the FTA era are not
higher than its bound duties before. This flexibility gives some extra leeway to accord
preferences in an FTA setting in an era where duties have been decreasing substantially.
There is an additional provision that only members of a CU must respect, in
order to satisfy the external requirement. GATT, Art XXIV.6 reads:
If, in fulfilling the requirements of subparagraph 5(a), a contracting party proposes to
increase any rate of duty inconsistently with the provision of Article II, the procedure
set forth in Article XXVIII shall apply. In providing for compensatory adjustment, due
account shall be taken of the compensation already afforded by the reduction brought
about in the corresponding duty of the other constituents of the union.
GATT, Art XXIV.6 deals only with customs duties. In this regard it is lex specialis to
Art XXIV.5(a) of the GATT: the additional obligation laid down in this provision
covers only customs duties and not other regulations of commerce. An example can help
illustrate the function of Art XXIV.6 of the GATT: assume that A, B, and C decide
to enter into a CU. Assume that before the formation of the CU the tariff protection
(to avoid any unnecessary complications, assume equivalence between bound and
applied rates) of the automotive sector in the three countries was the following:
A 20%
B 30%
C 40%

²²⁵ It is interesting that the WTO understanding focuses on applied as opposed to bound duties.
By adopting this focus, it is going further than simply stating that WTO members cannot use a CU
to undo tariff obligations that were previously bound; it is also stating that WTO members can-
not use a CU to jointly raise applied tariffs. This is interesting because one prediction of the theory
would be that a CU would have the incentive to set higher external tariff s than the members would
acting individually (ie before the CU) and that this would be one bad thing about the CU in terms
of its multilateral effects. So from the multilateral perspective, this rule makes sense.
 Tariff Protection in the GATT 163

A, B, and C, bind customs duties on cars at 30% (CU level). Arguably, they
have met their obligations under Art XXIV.5(a) of the GATT. They have not
necessarily met their obligations under Art XXIV.6 of the GATT, though. As
will be shown in what follows, if Art XXIV.5(a) of the GATT is violated, Art
XXIV.6 of the GATT will be ipso facto violated; however, compliance with
Art XXIV.5(a) of the GATT does not automatically lead to compliance with
Art XXIV.6 GATT. In other words, compliance with Art XXIV.5(a) of the
GATT is a necessary, but not sufficient condition for compliance with Art
XXIV.6 of the GATT.
GATT, Art XXIV.6 comes into play because A had to raise its pre-CU duty
from 20% to 30%. In such cases, Art XXVIII of the GATT type negotiations
will kick in. This means that WTO members which have initial negotiating rights,
principal supplying interest or substantial interest will participate in the negotia-
tions with the members of the CU; such negotiations aim to compensate those
WTO members which will have more difficult access to A’s market as a result of
the formation of the CU. The second sentence of Art XXIV.6 of the GATT makes
it clear that built-in compensation will be taken into account: new market oppor-
tunities, resulting from the fact that other constituents of the CU were forced
to lower their pre-CU duties to the common rate, will be taken into account. In
the end, an obligation to compensate will exist, only if the built-in compensation
does not suffice to take care of the injury suffered, as a result of A’s new, higher
duties. Let us go through two factually different scenarios to illustrate this point.
First scenario: A is a low per capita income small country, whereas C is a high
per capita income large country. Neither A nor C produces cars (or, their domes-
tic production allows for substantial amount of imports). The fact that C lowers
its duties from 40% to 30%, in all likelihood, over-compensates the fact that
A raised its own duties from 20% to 30%. This is the notion of built-in compen-
sation. C will import more cars than before, and hence, exporters will be com-
pensated for their losses (resulting from fewer exports to A).
Second scenario: A is the high per capita income large country, whereas C is
the low per capita income small country. In this case, the amount of trade lost
because A had to raise its duties is, most likely, not compensated by the fact that
C lowered its own duties. In such cases, there is nothing like sufficient built-in
compensation. Hence, something needs to be done. GATT, Art XXIV.6 calls
for compensation which will be offered to the WTO members, following an Art
XXVIII of the GATT negotiation.²²⁶

²²⁶ As argued above, a CU has an incentive to raise rather than lower its tariff s, if it wants to
improve its terms of trade. As Krugman (1991, p 9) observes, ‘even if formation of customs union
does not involve any increase in external tariffs, it can still in effect be beggar thy neighbour policy’,
since the terms of trade will have been affected to its favour anyway (with corresponding losses for
its trading partners).
164 Disciplines on Trade Instruments

(b) The internal requirement²²⁷ The internal requirement is common for FTAs
and CUs alike. According to Art XXIV.8 of the GATT, WTO members wish-
ing to enter into a CU or an FTA, will have to eliminate duties and other restrict-
ive regulations of commerce with respect to substantially all trade in products
originating in the constituents of the regional integration scheme. Duties are a
self-interpreting term. This is not the case, however, with respect to the other
two terms. Contrary to what happened in the context of Art XXIV.5(a) of the
GATT, no clarification of these two terms took place during the Uruguay round
negotiations.²²⁸
It has been suggested that the term ‘substantially all trade’ has both a
quantitative as well as a qualitative component, in the sense that it covers a certain
percentage of trade and at the same time no major sector of a national economy
can be excluded.²²⁹ However, as is probably always the case, whereas the former
(quantitative criterion) facilitates bargaining, the latter (qualitative criterion) will
invariably increase transaction costs. This is what happened in a series of work-
ing parties dealing with this issue. Probably the most appropriate way to sum up
practice in this field is offered by the working party report on EC—Agreements
with Portugal²³⁰ where the EC noted that:
there is no exact definition of the expression referring to the term ‘substantially all trade’.
The opinion has been expressed in the EEC working party that it is:
inappropriate to fi x a general figure of the percentage of trade which would be subjected
to internal barriers.²³¹
In the same working party, EC members expressed the view that:
a free-trade area should be considered as having been achieved for substantially all trade
when the volume of liberalized trade reached 80 per cent of total trade.²³²
The working party report on EFTA on the other hand, records the view that:
the percentage of trade covered, even if it were established to be 90 per cent, was not con-
sidered to be the only factor to be taken into account.²³³
Other working party reports reflect the view that the exclusion of a whole sec-
tor, no matter what percentage of trade involved, is contrary to the spirit of both

²²⁷ See the comprehensive analysis of Mathis (2002) on this score.

²²⁸ The negligible incentive that outsiders have to clarify this provision is probably the reason
why no agreement has been made possible over the years. With few exceptions (complementary
products), outsiders are better off the less PTA partners integrate their markets. Their incentive is
rather to clarify the external requirement, which they did, through the WTO understanding we
discussed supra.
²²⁹ The Analytical Index for example, reflects reports discussing this distinction at pp 824–5.
²³⁰ See BISD 20S/171, § 16.
²³¹ See the working party report on EEC, GATT Doc BISD 6S/100, § 34.
²³² See GATT Doc BISD 6S/70, § 30.
²³³ See GATT Doc BISD 96/83, § 48.
 Tariff Protection in the GATT 165

Art XXIV of the GATT, and the GATT itself.²³⁴ Nothing changed in this
respect in more recent years.²³⁵ Discussions in the context of the CRTA are
hardly illuminating. The GATT Analytical Index (vol 2) on p 824 in footnote
162 provides an exhaustive list of working party reports dealing with this issue
where the outcome is still the same: the term ‘substantially all trade’ has not
been clearly defined in relevant GATT practice. In a series of papers that the
WTO Secretariat prepared for the committee,²³⁶ the conclusion was inescapable:
50 years of practice notwithstanding, WTO members have failed to come up
with a workable definition of the term.
Australia tabled a proposal, after the conclusion of the Uruguay round agree-
ments, on the clarification of the term, which is at least worth discussing.²³⁷
Australia parts company with the usually mentioned but nebulously defined idea
that the term ‘substantially all trade’ reflects both a quantitative and a qualita-
tive element. In Australia’s view, there is only a quantitative element that can
come under substantially all trade and future negotiations should concentrate on
putting a number next to the concept. Australia specifically proposed that sub-
stantially all trade should be defined as coverage by a free-trade agreement or an
agreement establishing a CU of 95% of all the six-digit tariff lines listed in the
HS. In its response to questions by other WTO members,²³⁸ Australia accepted
that the 95% figure is an arbitrary benchmark, but intended to move negotiations
out of a deadlock and provide a workable and reasonable rule of thumb. Australia
was also mindful of the fact that in the case where trade is concentrated in only a
few products, the 95% figure could exempt sizeable trade flows. This is why it also
proposed an assessment of prospective trade flows under an arrangement at vari-
ous stages. So far, however, this proposal has provoked no meaningful discussion.
As a result, there is nothing like a consensus (not even close) on what ‘substantially
all trade’, another key term in the context of Art XXIV of the GATT means.

(c) Other restrictive regulations of commerce Disagreements on the understand-

ing of this term focus around two issues. First, the coverage of the term as such:
starting with the 1970 working party report on EEC—Association with African
and Malgasy States, one notes the opinion of members of the working party,
which were not members of the notified PTA, that trade had not been substan-
tially liberalized, in view of the continued imposition by certain parties to the
Convention (the Association of EEC with African and Malgasy States) of fiscal

²³⁴ See the working party report on EEC—Agreements with Finland, GATT Doc BISD 29S/79,
§ 12.
²³⁵ See the working party report on Free Trade Area between Canada and the US, GATT Doc
BISD 38S/73, § 83.
²³⁶ See WTO Docs WT/REG/W/17 (31 October 1997); WT/REG/W/17/Add 1 (5 November
1997); WT/REG/W/17/Corr. 1 (15 December 1997); WT/REG/W/17/Rev. 1 (15 February
1998).
²³⁷ See WTO Doc WT/REG/W/18 (17 November 1997).
²³⁸ See WTO Doc WT/REG/W/22/Add. 1 of 24 April 1998.
166 Disciplines on Trade Instruments

charges on imports from other members. At the same time, the members of the
PTA were declaring that:
the provisions of Article XXIV, concerning the concept of a free-trade area concerned only
protective measures. The taxes referred to were of a fiscal character, not protective . . . ²³⁹
Where should the line between ‘protective’ and ‘fiscal’ be drawn, was not dis-
cussed (and explained) any further. The cited passage unambiguously evidences
the disagreement between the members of this working party as to the coverage
of the term ‘other restrictive regulations of commerce’.
Second, the issue arises as to whether the list of provisions reflected in the par-
enthesis²⁴⁰ in Art XXIV.8 of the GATT should be deemed to be of exhaust-
ive character or not: the inferences to draw from the omission of Art XXI of
the GATT (the security exception) from the list reflected in the parenthesis was
discussed in the working party report on the EEC. The view of the (then) EEC
member states was that:
it would be difficult, however, to dispute the right of contracting parties to avail them-
selves of that provision which related, inter alia, to traffic in arms, fissionable materials,
etc., and it must therefore be concluded that the list was not exhaustive.²⁴¹
Other working party reports reflect a series of discussions on this issue, but no
definitive agreement across countries.²⁴²
Similar discussions are reported with respect to the relationship between Art
XIX of the GATT, and Art XXIV of the GATT: what inference should one draw
from the fact that the former has been omitted from the list reflected in the par-
enthesis? During the Uruguay round negotiations, a draft decision was tabled to
clarify once and for all the relationship between the two legal provisions. It read:
When an Article XIX action is taken by a member of a customs union or free-trade area,
or by the customs union on behalf of a member, it [need not] [shall not] be applied to other
members of the customs union or free-trade area. However, when taking such action it
should be demonstrated that the serious injury giving rise to the invocation of Article
XIX is caused by imports from non-members; any injury deriving from imports from
other members of the customs union or free-trade area shall not be taken into account in
justifying the Article XIX action.²⁴³
Had this proposal been accepted, it would have provided a much needed clarifi-
cation on this score. The proposal was, alas, rejected.
Subsequent practice has not helped clarify this issue either: first the AB and
thereafter a panel were faced with the question as to whether a member of a PTA

²³⁹ See BISD 18S/133, pp 135–7.

²⁴⁰ GATT, Art XXIV.8 outlaws any restrictive regulations of commerce other than those
reflected in parenthesis in the body of this provision.
²⁴¹ See BISD 6S/70, p 97.
²⁴² See the Analytical Index, pp 820ff.
²⁴³ See WTO Doc WT/REG/W/17/Rev. 1, p 4.
 Tariff Protection in the GATT 167

(a CU in the case of Argentina²⁴⁴ and an FTA in the case of the United States²⁴⁵)
could impose safeguards against other members of the PTA. Note that on both
occasions, the WTO adjudicating bodies were requested to interpret the WTO
Safeguards Agreement, and not Art XXIV of the GATT per se. However, through
their interpretation they were clearly ‘influencing’ the question as to whether the
items figuring in the parenthesis of Art XXIV.8 of the GATT form an exhaustive
list or not. On both occasions, the WTO adjudicating bodies held that mem-
bers of a PTA can impose safeguards against other members of a PTA, provided
that they respect a parallelism: they can do so if they have counted PTA imports
when assessing injury; they cannot do so, however, if they have not counted PTA
imports when assessing injury.
Following these events, it is clear now that the list in the parenthesis of Art
XXIV.8 of the GATT is not exhaustive.²⁴⁶ What else, besides safeguards, should
one include? Should one by inference be brave enough and apply the stated rea-
soning in all forms of ‘contingent protection’? Only future experience will tell.
Art XXI is, of course, a candidate.

(d) The result of non-clarity Schott (1989) identifies four cases where PTAs were
judged broadly consistent with the GATT. Since his study there has been one case
where there has been a definitive and unambiguous acceptance, at the CRTA
level, that the notified PTA was GATT consistent: the CU between the Czech
and the Slovak republics. We are simply in the dark as to the consistency of the
remaining 99% of all PTAs currently in place.
Hudec (1972, p 1362) notes:
The seeming collapse of the MFN rules is probably the single most important cause of the
present day pessimism about the GATT substantive rules.
Hudec, 20 years later (1993b, p 154), remarks:
the GATT’s somewhat benign attitude toward RAs is merely one part of this larger toler-
ance toward departures from MFN in general.

²⁴⁴ See Argentina—Footwear (EC).

²⁴⁵ See US—Wheat Gluten.
²⁴⁶ The AB had the opportunity to pronounce on the legal status of GATT adopted panel
reports: according to its case-law, such reports form part of the so-called GATT acquis and provide
useful guidance to subsequent panels dealing with the same issue. It had never had the opportun-
ity so far to pronounce on the legal status of WTO adopted reports. From a political perspective
GATT adopted reports seem to carry enhanced legitimacy because of the imprimatur given to
them through positive consensus of the losing party (which is not the case in the WTO era with
the passage to negative consensus). From a legal perspective, however, it would seem untenable to
support the thesis that WTO adopted reports do not enjoy at least the same status as their GATT
counterparts. So although there is nothing like stare decisis in WTO law (indeed, there is nothing
like stare decisis in public international law) adopted WTO reports exercise, as modern practice
shows, considerable influence on subsequent reports dealing with the same issue. Following prior
decisions when appropriate, and motivating departures when needed is probably the surest way for
WTO adjudicating bodies to acquire legitimacy.
168 Disciplines on Trade Instruments

Roessler (1993, p 321) seems in agreement:

The record under the current procedures is not encouraging. During the past three dec-
ades about 50 working parties have been established to examine RIAs. None of them
was able to reach a unanimous conclusion on the GATT-consistency of the agreement
examined . . .

3.12.4.5 The bilateral track

Where does all of this leave us? The CRTA has never gone so far as to pronounce
unanimously on the incompatibility of a notified PTA with Art XXIV of the
GATT.²⁴⁷ Hence, CRTA reports, with five exceptions so far, essentially reflect
disagreements across members as to the consistency of the notified PTA with the
multilateral rules. Track II, as established above, is also available. Since 1995,
recourse to Track II has been greatly facilitated through the passage to nega-
tive consensus, where, as we will see in more detail in Chapter 5, the will of the
complainant suffices for a panel to be established (its report adopted, and coun-
termeasures, assuming non-compliance in case of victory for the complainant,
authorized). One would expect that since disagreements on the consistency of
a PTA with the WTO reflect disagreements about the interpretation of the key
terms discussed above, and since panels are established to decide on the clarifica-
tion/interpretation of the WTO contractual terms, panels would be quite busy in
this area. Yet, practice demonstrates the contrary. Before, however, we move on
to discuss the possible reasons why this has been the case, let us first take a look at
panel practice in this area.
During the GATT years (1948–94), three panels were established to exam-
ine claims relating to the consistency of a PTA with the multilateral rules.²⁴⁸
Two reports were issued and they both remain unadopted. The first of these, the
EC—Citrus panel report argues for an examination of individual measures only
by panels and refuses to pronounce on the overall consistency of a PTA with the
multilateral rules. The panel does not see its role as a surrogate to the (then) Art
XXIV working parties:
The Panel noted that at the time of the examination of the agreements entered into by the
European Community with certain Mediterranean countries, there was no consensus
among contracting parties as to the conformity of the agreement with Article XXIV.5 . . .

²⁴⁷ Schott (1989, p 25) mentions only four unanimous decisions. In all four cases, working par-
ties actually admitted that the notified PTA was GATT consistent. The same was true in the only
other case decided by consensus ever since: the CU between the Czech and the Slovak Republics.
²⁴⁸ The first, after a request by Canada in 1974 in connection with the accession to the European
Community of Denmark, Ireland, and the United Kingdom (GATT Doc C/W/250) was not
activated because the parties to the dispute reached an agreement (GATT Doc C/W/259). The
second, led to an unadopted panel report in EC—Tariff Treatment on Imports of Citrus Products
from Certain Countries in the Mediterranean Region, GATT Doc L/5776 (hereafter the EC—Citrus
panel report). The third report is on EEC—Import Regime of Bananas, GATT Doc DS38/R of
11 February 1994 (hereafter the EEC—Bananas panel report) which also remains unadopted.
 Tariff Protection in the GATT 169
The agreements had not been disapproved, nor had they been approved. The Panel found
therefore that the question of conformity of the agreements with the requirements of
Article XXIV and their legal status remained open.²⁴⁹
This report remains unadopted, and hence, of limited legal relevance.²⁵⁰
EEC—Bananas is the second report. This report made one important inter-
pretative contribution by holding that one way preferential arrangements are per
se inconsistent with Art XXIV of the GATT; obligations to liberalize must be
assumed by all participants. § 159 of the report relevantly reads in this respect:
This lack of any obligation of the sixty-nine ACP countries to dismantle their trade barri-
ers, and the acceptance of an obligation to remove trade barriers only on imports into the
customs territory of the EEC, made the trade arrangements set out in the Convention
substantially different from those of a free trade area, as defined in Article XXIV:8(b).
Unsurprisingly, the same panel went on to conclude in § 164 that the Lomé
Convention (the agreement between the European Community and a series of
African, Caribbean, and Pacific states) did not correspond to the type of agree-
ments which Art XXIV of the GATT covers. Th is report remains unadopted
as well and, although the view expressed in the cited passage is sound, the legal
value of the report is minimal.
During the WTO era, practice in this area continues to be scarce. However,
the resulting case-law has far-reaching implications. The Turkey—Textiles panel
report²⁵¹ records the first time that the issue of the consistency of a PTA with the
GATT was discussed. India argued that it suffered damage as a result of Turkey’s
decision to erect new barriers to its textiles exports, following the signature and
the entry into force of the European Community—Turkey CU. India argued
that its MFN rights were being impaired and the ball shifted to Turkey’s camp to
justify its measures. Turkey invoked its CU with the European Community, and
the panel first had to address whether it was competent to discuss the consistency
of a PTA with the GATT. It held the view that WTO adjudicating bodies are
competent to examine PTA related issues, but should stop short of providing an
overall assessment of consistency of a PTA with the WTO contract. The view of
the panel on this issue is reflected in §§ 9.52 and 9.53 of the report. We quote:
As to the second question of how far-reaching a panel’s examination should be of the
regional trade agreement underlying the challenged measure, we note that the Committee
on Regional Trade Agreements (CRTA) has been established, inter alia, to assess the
GATT/WTO compatibility of regional trade agreements entered into by Members, a very
complex undertaking which involves consideration by the CRTA, from the economic,

²⁴⁹ See GATT Doc L/5776, dated 7 February 1985 at § 4.6 and at § 4.10.
²⁵⁰ See on this issue the conclusions of the WTO AB in Japan—Taxes on Alcoholic Beverages,
briefly discussed in Chapter 1.
²⁵¹ See the panel report on Turkey—Restrictions on imports of textile and clothing products, WTO
Doc WT/DS34/R of 31 May 1999.
170 Disciplines on Trade Instruments
legal and political perspectives of different Members, of the numerous facets of a regional
trade agreement in relation to the provisions of the WTO. It appears to us that the issue
regarding the GATT/WTO compatibility of a customs union, as such, is generally a
matter for the CRTA since, as noted above, it involves a broad multilateral assessment of
any such custom union, i.e. a matter which concerns the WTO membership as a whole.
...
As to whether panels also have the jurisdiction to assess the overall WTO compati-
bility of a customs union, we recall that the Appellate Body stated that the terms of ref-
erence of panels must refer explicitly to the “measures” to be examined by panels. We
consider that regional trade agreements may contain numerous measures, all of which
could potentially be examined by panels, before, during or after the CRTA examination,
if the requirements laid down in the DSU are met. However, it is arguable that a customs
union (or a free-trade area) as a whole would logically not be a “measure” as such, subject
to challenge under the DSU. (footnotes omitted)
In the panel’s view, for reasons having to do more with the administrative bur-
den, the CRTA is the more appropriate forum to review consistency of notified
PTAs. The panel report was appealed. The AB report holds for a different prop-
osition. The Art XXIV of the GATT defence, in the AB’s view, holds only if two
conditions are met:
First, the party claiming the benefit of this defense must demonstrate that the meas-
ure at issue is introduced upon the formation of a customs union that fully meets the
requirements of sub-paragraph 8(a) and 5(a) of Article XXIV. And second, that party
must demonstrate that the formation of that customs union would be prevented if it were
not allowed to introduce the measure at issue. . . . We would expect a panel, when exam-
ining such a measure, to require a party to establish that both of these conditions have
been fulfilled. It may not always be possible to determine whether the second of the two
conditions has been fulfilled without initially determining whether the first condition
has been fulfilled.²⁵²
Hence, in the AB’s view, WTO adjudicating bodies must request from parties
raising the PTA defence to first establish that they have fulfilled the conditions to
raise such a defence. It remains to be seen to what extent the cited obiter dictum of
the AB will be followed in future experience.
More recently, the panel report on US—Line Pipe²⁵³ faced an argument by
the United States that, since it is a member of NAFTA, it was entitled to treat
imports from NAFTA differently than imports from non-NAFTA sources when
imposing a tariff quota. The panel first addressed the issue of burden of proof
(burden of production):
As the party seeking to rely on an Article XXIV defense . . . the onus is on the United
States to demonstrate compliance with these conditions. (§ 7.142 of the report)

²⁵² See WTO Doc WT/DS34/AB/R of 22 October 1999, §§ 58–9.

²⁵³ WTO Doc WT/D S202/R of 29 October 2001.
 Tariff Protection in the GATT 171

The same report addressed the issue of the quantum of proof (burden of persua-
sion) that the party carrying the burden of proof has to provide in order to estab-
lish a prima facie case of the consistency of a PTA with the multilateral rules.
§ 7.144 of the report reads in this respect:
In our view, the information provided by the United States in these proceedings, the
information submitted by the NAFTA parties to the Committee on Regional Trade
Agreements (“CRTA”) (which the United States has incorporated into its submissions
to the Panel by reference), and the absence of effective refutation by Korea, establishes a
prima facie case that NAFTA is in conformity with Article XXIV:5(b) and (c), and with
Article XXIV:8(b).
The information provided by the United States in the proceedings is reflected
in § 7.142 of the report and is reflected in a statement that duties on 97% of
the NAFTA parties’ tariff lines will be eliminated within 10 years, whereas,
with respect to ‘other regulations of commerce’, a reference to ‘the principles
of national treatment, transparency, and a variety of other market access rules’
is made, plus the information that NAFTA participants had to submit to the
CRTA. Obviously, in the panel’s view, the submitted information was enough to
make a prima facie case of consistency of NAFTA with Art XXIV of the GATT.
The panel seems to have paid particular attention to the fact that the complain-
ant did not refute the evidence provided. The panel, however, did not entertain a
comprehensive review of the evidence before it. This is the only instance (not con-
doned yet by the AB) where the issue of the burden of persuasion was discussed.
As regards the legal significance of the fact that the CRTA had not issued its
report at the time the dispute was submitted to the panel, § 7.144 reads:
Concerning Article XXVIII:8(b), we do not consider the fact that the CRTA has not yet
issued a final decision that NAFTA is in compliance with Article XXIV:8 is sufficient
to rebut the prima facie case established by the United States. Korea’s argument is based
on the premise that a regional trade arrangement is presumed inconsistent with Article
XXIV until the CRTA makes a determination to the contrary. We see no basis for such a
premise in the relevant provisions of the Agreements Establishing the WTO.
Hence, the panel here applied, without referring to it, the legal principle in dubio
mitius, and concluded that there should be no presumption that the notified
PTA is WTO inconsistent. The panel, thus, refused to construe Art XXIV of the
GATT as disallowing the operation of a PTA which had not first been cleared by
the CRTA. This is additional evidence that, in WTO practice, the role of CRTA
is not equated to that of a competition authority when discussing the approval of
a notified merger.

3.12.4.6 Why not litigate more?

The inescapable conclusion, following a review of practice under both the bilat-
eral and the multilateral tracks, is that a largely incomplete contract (Art XXIV
172 Disciplines on Trade Instruments

of the GATT) has only partially been completed through subsequent practice.
Importantly, the key terms reflected in the body of Art XXIV of the GATT have
not been interpreted in a manner that represents the consensus of the WTO
membership. As a result, except for some obvious cases, we simply cannot tell
whether the numerous PTAs currently in practice are GATT consistent or not.
Trade practice basically constitutes inertia, an unwillingness to clarify the terms
of a 60-year-old agreement.
The EC delegate is recorded stating in the 1978 working party report on the
Agreement between the EEC and Egypt that:
. . . as regards the possibility of consultations with the contracting parties concerning the
incidence of the Agreement on their trade interests . . . nothing prevented these countries
from invoking the relevant provisions of the General Agreement, such as Articles XXII
and XXIII.²⁵⁴
The EC delegate probably knew that subjecting the European Community’s
numerous PTAs to meticulous scrutiny was not much of a credible threat. His
expectation was probably rational, at that time at least, since in the GATT years
the defendant had the right to even block the establishment of a panel. Have
things changed ever since? Unambiguously, yes. On the one hand, there is a pro-
liferation of PTAs. On the other hand, there are changes in both the law and the
case-law that would prima facie argue in favour of more activity before WTO
adjudicating bodies:
first, the substantive law is now, in part, clearer (Understanding on Art XXIV of
the GATT) and, it is also clear, that panels have the competence to adjudicate
claims relating to the consistency of a PTA with the multilateral rules;
second, panels will be established at the sole request of the complainant;
third, the original burden of proof of the complainant is easy to meet (the com-
plainant would be required to demonstrate deviation from MFN and, upon such
demonstration, the burden of proof would shift to the defendant), whereas the
defendant will have to show that its PTA is GATT consistent.
This latter element represents no change from the prior regime. Complainants
that is, will face the same burden as before but this time, they can base their argu-
ments on a clearer legislative framework, and do not have to secure the defend-
ant’s consensus for a panel to be established. They are definitely, from a legal
perspective at least, in a substantially more favourable position than they were in
the pre-Uruguay round era.
It seems that law and case-law have opened a huge avenue to potential com-
plainants aiming to challenge before the WTO judiciary the consistency of such
PTAs with the multilateral rules.

²⁵⁴ See GATT Analytical Index, p 781.

 Tariff Protection in the GATT 173

Yet, there is not much judicial activity in this respect. Except for the two cases
mentioned (Turkey—Textiles, and US—Line Pipe), there is no other reported
case in this context. So why have WTO members stayed idle? Is the problem not
a real one?²⁵⁵ There is no clear cut answer to this question. In what follows, I try
to advance some potential avenues for future research since, as things stand, this
is a largely unresearched area:²⁵⁶ I lay out some possible reasons that could help
explain why, in light of the discussion above, rational behaviour would support
very few, if any, challenges against PTAs before WTO panels. There is no rank-
ing among the various possible explanations. I look into explanations depending
on whether a WTO member is an outsider or an incumbent, that is, whether it
participates or not in a PTA, since the former (much narrower) and the latter
(much wider) category of countries might for different reasons rationally choose
not to challenge the consistency of a PTA with the multilateral rules.
Take the case of a WTO member which does not participate in any PTA. The
list is of course shrinking all the time. A risk-averse WTO member would ration-
ally choose not to challenge a PTA, even though it does not run the risk that its
actions might provoke legal challenges on the same score,²⁵⁷ and this for a num-
ber of good reasons:
(1) the contract is highly incomplete and risk-averse panels might rationally
choose not to pronounce the inconsistency of a PTA, unless if they stand on
very firm grounds;
(2) the cost of litigation should not be underestimated;
(3) cost-benefit analysis might suggest inaction;
(4) reasons having to do with not restraining their own future behaviour might
argue against a challenge;
(5) other policy/political reasons might point to the same perspective.
I will take each of the elucidated grounds in turn. A panel is usually composed
of trade delegates, who are career diplomats serving in Geneva.²⁵⁸ When facing

²⁵⁵ See Limáo (2006b) on this score. See also the analysis in Levy and Srinivasan (1999).
²⁵⁶ This part is largely based on Mavroidis (2006).
²⁵⁷ One of course cannot exclude that a legal challenge under Art XXIV of the GATT might
provoke a legal challenge against the complainant under a different legal provision. Indeed, there is
increasing empirical evidence that action by A against B in antidumping, increases the likelihood
that B might attack A in the same field. Martin and Vergote (2004) find that AD is being increas-
ingly used in a strategic, retaliatory fashion. In the Martin-Vergote setting, retaliation actually
occurs in equilibrium as part of a cooperative relationship among privately informed governments.
Zanardi (2004) provides additional supporting evidence: US producers are among the hardest-hit
(through AD) producers around the world. Yet, the US administration chooses not to act as com-
plainant before the WTO (that is, to represent their interests) even though sometimes the chances
of victory might be realistic. It probably wants to discourage AD litigation in general, since the
United States itself has ample recourse to AD duties to protect its own market.
²⁵⁸ On the function of panels and, more specifically, the decisive role of WTO Secretariat,
see the excellent analysis in Nordström (2005b) which supports the conclusions reached here
174 Disciplines on Trade Instruments

a question relating to the consistency of a PTA with the WTO, they know that
they are not prejudging just one transaction (as is the case, for example, when they
face a challenge against an allegedly illegal imposition of AD duties). Through
their response, they will be putting into question all subsequent trade-related
behaviour of the PTA hopefuls. They also know that:
(1) no WTO adjudicating body before them ever moved to pronounce on the
inconsistency of a PTA with the WTO, except in very obvious cases (such as
the EC/ACP arrangement quoted above);
(2) that they have little legislative guidance to draw from, since the terms of Art
XXIV of the GATT, with the few notable exceptions mentioned supra, have
not been further elaborated/interpreted either through legislative activity or
through subsequent (CRTA) practice; and that
(3) there is probably a relationship between (1) and (2), that is, wisely, previous
panels chose to abstain from far-reaching pronouncements in the absence of
a clear legislative document.
They are also aware that PTAs are being tolerated in the WTO. As Schott (1989,
p 25) mentions:
Besides the ambiguity of its provisions, political considerations have often outweighed
other factors in decisions to accede to the terms of the agreements. In addition, affected
third countries have been reticent to criticize preferential deals because the majority of
GATT members participate in such agreements.
To a large extent, WTO members have learned to live in a world where PTAs will
not be challenged and, by now, they might rationally be expecting to receive a ‘no
challenge’ status when they enter into PTAs.
Unless the PTA is on very shaky grounds, panels might rationally choose to
abstain from making a definitive pronouncement on the question asked. In light
of the uncertainty surrounding the precise legal frontiers between consistency
and inconsistency with the multilateral rules, it is unlikely that a panel will find
itself on firm legal grounds to provide a definitive statement on the law itself,
except for exceptional cases. It does not have to hide behind non liquet to do that;
it can communicate the message that, for example, the burden of proof has not
been met, or evidence provided has not been effectively refuted and ask for add-
itional homework from the complainant.
The cost of WTO litigation should not be underestimated either. Bown and
Hoekman (2005) provide some numbers in this regard and point to the fact that
the numbers they cite in and of themselves might dissuade potential complain-
ants from launching complaints. Nordström (2005a) concludes along the same

(the WTO Secretariat, aware of the extent of unresolved issues before the CRTA would hardly be
pressing for definitive interpretations in the context of Art XXIV of the GATT).
 Tariff Protection in the GATT 175

lines and his empirical study cautions against those who tend to underestimate
the cost of a WTO litigation. Of course, a potential complainant will measure
the cost of litigation against the probability of prevailing and the expected prof-
its if it does prevail. As already stated above, the original burden of proof for a
complainant in such a case is quite low. However, the burden of proof might
shift back to the complainant, in a case where panels adopt an attitude similar
to that of the panel report on US—Line Pipe discussed above. In such a case, the
complainant will have to address technical arguments in an area of law that does
not tend to lend itself to yielding predictable outcomes in a litigation case. More
(and more technical) issues equal higher litigation costs. Remember, there has
not been one single panel so far that has taken one positive step towards interpret-
ing the controversial terms appearing in the body of Art XXIV of the GATT and
highlighted supra. Yes, the original burden of proof seems to be quite low, but this
is not the end of the story.
On the other hand, the potential complainant will actually, through such
a complaint, be subsidizing all other WTO members. It is highly likely that a
collective action problem exists here where everybody expects someone else to
take the lead and free riding is very much on the cards. Past experience eminently
supports the conclusion that in this area GATT/WTO practice is akin to the ‘if I
don’t do it, somebody else will’ mentality.
It further appears counter-intuitive that WTO members would be willing to
enforce the internal requirement: if less trade liberalization exists among con-
stituents of a PTA, less trade diversion will ensue. Assuming that trade diversion
goes against the interests of a particular exporter (who will thus be losing a mar-
ket), the less members of a PTA liberalize their trade in cars, the better export
opportunities outsiders will have to export cars to the PTA at hand. Of course,
the opposite could also be the case. An example with complementary products
illustrates this point: the more members of a PTA liberalize trade in cars, the
likelier it would be for WTO members who are non-parties to the PTA to sell
wheels in the PTA market. This example serves the purpose of showing that chal-
lenging a PTA is not necessarily the best way to receive compensation for trade
diversion. Staying idle might work just as well. At best, the argument as to why
non-parties would be willing to enforce the internal requirement is less clear than
with respect to the external requirement.
Outsiders of today might be the incumbents of tomorrow. Successfully arguing
a PTA-related case might come at a cost, since, in the event that today’s successful
complainant might decide to go ‘preferential’ tomorrow, it might face the music
it helped, through its pleadings, to compose. A non-rigorous understanding of
the existing regime might, on the other hand, come in handy, assuming a change
of heart on this issue. One cannot a priori exclude the possibility that inaction
might reflect a side payment for favours by members of a PTA in other fields of
government activity. There is some off the record (hard to prove) evidence to this
effect.
176 Disciplines on Trade Instruments

Take the case of the European Union. Today’s level of integration is a far cry
from the picture in the late fifties. GATT Contracting Parties, however, chose
(rationally) not to challenge the European integration process at the time. Indeed,
there was too much at stake: destabilizing the process through a legal challenge
before the GATT could have had immeasurable side-effects. Who in the GATT
would be eager to take the credit for successfully outlawing a GATT inconsist-
ent CU, knowing that, at the same time, it would be also taking the blame for
destabilizing a quintessential piece of the post-Second World War peace process?
I am not suggesting that all consummated PTAs reflect the very same specifici-
ties of the European integration process. But it is entirely plausible that the ele-
ments of the decision to go preferential could well extend beyond trade concerns.
Knowledge of such elements, or indeed, even uncertainty about them, might
prove enough to deter a legal challenge.
The data discussed above (112 CUs and FTAs) and the identity of players par-
ticipating in the preferential game supports the intuition that we are probably
facing a mutual deterrence scenario. Nowadays, all important players participate
in one or more PTAs. Most of them are of dubious WTO consistency. They have
probably naturally moved to a cooperative equilibrium, where no one challenges
anyone else’s PTA and live happily within a regime of tolerance. Deviating from
the current situation could open up a large number of legal challenges that no one
may wish to face at this stage.
At the same time, in light of the legal uncertainty as to what is considered a
WTO consistent PTA, even those who might believe they have complied with
the legislative requirements when entering into a PTA (assuming such a group
exists), if they are sufficiently risk-averse, they could be unwilling to mount the
first challenge for fear of reciprocal activity against them. The results by Martin
and Vergotte (2004) and Zanardi (2004) are probably relevant here: an incum-
bent challenging a PTA, if successful, is probably ipso facto shooting itself in the
foot. Any interpretation/clarification of the requirements reflected in Art XXIV
of the GATT, in a given case, assuming that the case-law aspect of such a pro-
nouncement will not be neglected in future litigation, could mean that a mem-
ber risks facing the achieved interpretation in challenges against its own PTA.
Any interpretation in other words, could become the benchmark to (re-) discuss
the consistency of a member’s PTA with the multilateral rules. In such cases,
there is probably substantial uncertainty ex ante as to the content of the interpret-
ation; there is, however, quasi certainty that, whatever the interpretation might
be, it will, eventually, be applicable against the original complainant’s PTA.
Consequently, the complainant will have to choose between a strategy of passiv-
ity, where its own PTA is not being challenged, and a strategy of active behaviour,
where it risks paying the price for any ephemeral victory before a WTO adjudi-
cating body.
The existing framework does not look very promising. Assuming there is agree-
ment that something needs to be done, WTO members would probably need
 Tariff Protection in the GATT 177

to reflect on a new, more workable test to discuss the consistency of PTAs with
the multilateral rules; otherwise they risk seeing no change in the current prac-
tice. On the other hand, changing the rules of the game requires some additional
thinking on the alternatives to the current loose PTA review. Through PTAs,
countries integrate in other, not necessarily WTO covered areas. Trade could
simply be a component of a wider scheme and, hence, one would need to measure
the impact of introducing PTA busters in the WTO regime. In a nutshell: yes,
there is a problem with the current inaction, but we might be creating other, even
more significant problems by introducing active surveillance. It seems that all the
pieces of the puzzle have not yet been put in place. It is for the WTO membership
to decide whether inaction is preferable or not to action in a context of imperfect
information on this issue.²⁵⁹

3.12.5 What is the permissible extent from MFN deviation

when a PTA is formed?
Increasingly, PTAs include elements which go beyond the measures explicitly
covered by Art XXIV of the GATT. The US—Morocco FTA, for example, con-
tains a clause whereby the latter promises not to impose export taxes on phos-
phates.²⁶⁰ The treatment of export taxes is not explicitly addressed in Art XXIV
of the GATT. Undoubtedly, however, it is a duty and even if the term ‘duties’
appearing in Art XXIV.8 of the GATT is considered to cover import duties only,
export taxes should come under the term ‘restrictive regulations of commerce’
appearing in the same provision: such regulations must be abolished with respect
to substantially all trade, as required by Art XXIV.8 of the GATT. Assuming that
trade in phosphates is not infinitesimal, then Morocco would be required to abol-
ish the export tax, otherwise its FTA with the United States does not meet the
statutory requirement of Art XXIV.8 of the GATT. Meeting this requirement
is necessary for the FTA to be judged GATT consistent. This last condition (the
necessity requirement) has been acknowledged by the AB in its Turkey—Textiles
report in the following terms (§ 58):
Accordingly, on the basis of this analysis of the text and the context of the chapeau of
paragraph 5 of Article XXIV, we are of the view that Article XXIV may justify a measure
which is inconsistent with certain other GATT provisions. However, in a case involving
the formation of a customs union, this “defence” is available only when two conditions
are fulfilled. First, the party claiming the benefit of this defence must demonstrate that
the measure at issue is introduced upon the formation of a customs union that fully meets
the requirements of sub-paragraphs 8(a) and 5(a) of Article XXIV. And, second, that
party must demonstrate that the formation of that customs union would be prevented if

²⁵⁹ Note in this respect, that Baldwin (2006) has argued that an eventual interconnection of the
many PTAs will eventually lead to global free trade.
²⁶⁰ See Art 10.2 and Annex 2C.
178 Disciplines on Trade Instruments
it were not allowed to introduce the measure at issue. Again, both these conditions must
be met to have the benefit of the defence under Article XXIV.
It follows (a contrario) that elements that are not necessary for the formation of
a PTA are not exempted from the MFN obligation. What is the benchmark to
judge necessity? In principle, any (trade and domestic) instrument can affect
trade, and, consequently, in principle, any regulation can be restrictive of com-
merce. On the other hand, depending on how narrowly or widely one interprets
the necessity-requirement, one might end up prejudging the extent of the per-
missible deviation from MFN (and thus, the trade creation/trade diversion ratio).
Suppose, for example, that two countries participating in a PTA recognize each
other’s environmental standards as equivalent to each other, and sign a Mutual
Recognition Agreement (MRA) to this effect. Their MRA might of course sub-
stantially liberalize trade. Were one to accept that this MRA is necessary for the
PTA to come into being, then the parties to the MRA are under no obligation
to extend their MRA to goods from other countries which can demonstrate that
they fulfil the conditions for such extension. Were one, however, to take the
opposite view, then, assuming other countries can satisfy the regulatory require-
ments of the signed MRA, they can legitimately claim extension of the MRA to
their benefit.²⁶¹
This question has not been answered in case-law so far. Practice, as discussed
above, is highly uninformative about the meaning of the terms ‘substantially all
trade’ and/or ‘restrictive regulations of commerce’.
There are good arguments, nevertheless, to support the thesis that the term
‘restrictive regulations of commerce’ should be confined to trade instruments only.
The purpose of Art XXIV of the GATT is to reduce protection for a sub-part
of the WTO membership, the countries participating in a PTA (in legal terms,
members of a lawful PTA can treat each other in deviation from MFN). The
only permissible (legal) protection in GATT is protection through trade instru-
ments, tariff protection. It follows that the only advantage that WTO members
can give each other when forming a PTA is a tariff advantage. The rest of their
policies have to respect the MFN obligation. But restrictive regulations of com-
merce cannot be equated to tariffs: the term ‘duties’ (tariffs) appears as such in
Art XXIV.8 of the GATT; restrictive regulations of commerce must mean some-
thing else. What else? Tariffs (duties) is shorthand (sometimes clumsy) for more
than one instrument: as we saw supra in this chapter, WTO members negotiate
not only OCD, but also ODCs. The latter are certainly a restrictive regulation of
commerce since, other things equal, they affect the volume of goods entering a
market. By the same token, rules of origin legitimately come in the picture: PTA

²⁶¹ See Mathis (2006), who expressed similar thoughts on this score, as well as various other
contributions in Bartels and Ortino (2006).
 Export Subsidies 179

members want to ensure the authenticity of the origin of goods benefiting from
preferential treatment.²⁶²
Domestic instruments are not meant to protect domestic production: Art III
of the GATT, as we will see in Chapter 3 is unambiguous in this respect. If they
are not meant to protect, then it is impossible that a member of a PTA should
be allowed to give one sub-set of the WTO membership (the PTA members)
an advantage that it does not extend to the remaining membership. The inclu-
sion of Art XX of the GATT in the list of measures that do not have to observe
the requirement to liberalize substantially all trade provides support for this
argument.²⁶³

4 Export Subsidies

4.1 Rigidity required?

Economic theory informs us that subsidies are not as distorting as other classic
trade instruments: whereas subsidies affect one margin (ie producers) of trade,
other trade instruments (eg quantitative restrictions or tariffs) have an impact on
two margins (ie producers and consumers). Subsidies are viewed by economists as
an instrument whereby governments, at least in a perfect competition model,²⁶⁴
subsidize foreign consumers: a reduction in the marginal cost of foreign firms
(which may be the consequence of a subsidy) in general reduces prices for the
domestic firms. Hence, many economists have made the point that instead of
imposing countervailing measures to offset subsidization, governments affected
by foreign subsidies should be sending a ‘thank you note’ to the subsidizing
government.²⁶⁵
Perfect competition, nevertheless, is an unrealistic scenario. In practice, sub-
sidies often constitute the preferred instrument of industrial policy in most

²⁶² This is the good faith argument for including rules of origin in the category of restrictive
regulations of commerce. In practice, as signaled elsewhere in this chapter, rules of origin also
function as means to reduce the promised benefits.
²⁶³ Thus, to go back to the example used supra, members of a PTA concluding an MRA will
have to extend it to non-PTA WTO members which satisfy the regulatory requirements of the
concluded MRA.
²⁶⁴ It bears emphasizing here that in non-perfect competition markets subsidies granted by
governments to monopolies may not necessarily benefit consumers wherever they may be, as the
beneficiary-monopolist may use the subsidy (perversely and rationally at the same time) to maxi-
mize its profit by setting the price somewhere between the market price and the reservation price
(ie monopoly price). Ostensibly, even when there are few industries competing on the world stage,
these firms may still have incentives to price their products above the market price without entering
into any tacit or explicit agreement.
²⁶⁵ In the absence of any other distortion and if—and this is a big if—such a subsidy will reduce
welfare in the two countries combined. The foreign country may lose or gain. The domestic econ-
omy may also lose or gain. It all depends on their strategic interaction.
180 Disciplines on Trade Instruments

jurisdictions.²⁶⁶ The positive externalities stemming from subsidization (eg in

consumer welfare terms) are not overlooked in economic theory or empirical
work. With the notable exception of predatory subsidization,²⁶⁷ economists
generally caution against far-reaching disciplines on subsidies. Th is does not
mean, however, that subsidization should remain unregulated. In fact, one can
come up with good arguments in favour of regulation of subsidies. Strategic
trade theory in the 1980s has been interpreted by some policy-makers as a
recommendation for active subsidization in order to induce enhanced prod-
uctivity.²⁶⁸ As Brander and Spencer (1985) nevertheless explain, an agreement
to reduce or even eliminate subsidies makes sense in a world of strategic trade
policy: true, by reducing its subsidies, country A will be losing something; at
the same time, however, it might be willing to do that if country B does the
same as well.
This is definitely a sound argument in favour of disciplining subsidization. The
problem is that disciplining subsidization might have negative external effects:
recent research (Bagwell and Staiger, 2006) suggests that WTO members might
be reluctant to make further tariff commitments if subsidies are over-disciplined.
Hence they argue against rigidity when disciplining subsidies; they also explain,
however, that having no disciplines at all might be equally problematic: WTO
members might lose market access opportunities as a result of subsidization, and
even the incentive to liberalize, particularly if subsidization continuously under-
mines their market expectations, for which, of course, they have paid a reciprocal
concession.
Too much restriction is counterproductive, and so is too little. What to do
then? Different views have been expressed on this score, failing however, so far, to
provide a comprehensive framework for dealing with subsidies. Faute de mieux,
Sykes (2003a) argues in favour of abandoning regulation of domestic subsidies.
He acknowledges nevertheless, that this is is not an unproblematic scenario.
The attitude of the world trading system towards subsidies has demonstrably
evolved since the inception of the GATT. The GATT initially followed a much
more lenient approach to disciplining Contracting Parties’ use of subsidies.²⁶⁹
GATT, Art VI implicitly recognized that subsidies may be a legitimate instru-
ment of domestic public policy by specifying that countervailing duties to off-
set the subsidy can only be imposed if the effect of subsidization has resulted in
material injury to a domestic industry. The wholesale change stems from strin-
gent rules adopted by WTO members with respect to export subsidies: while

²⁶⁶ See Grossman and Helpman (2001).

²⁶⁷ Theoretically, it could be the case that subsidies be predatory and thus hurt competition, and
not just competitors. In the real world, this looks like a highly unlikely scenario.
²⁶⁸ See Krugman (1983) and d’Andrea Tyson (1992).
²⁶⁹ For a comprehensive analysis, see Sykes (2003a). See also, Janow and Staiger (2003) on the
same score. The more policy-type perspective of this issue is provided in an authoritative manner in
Hoekman and Kostecki (2001) as well as Trebilcock and Howse (2005).
 Export Subsidies 181

Art XVI of the GATT discourages the use of subsidies conditioned on exports
(‘export subsidies’) on ‘non-primary’ products, the SCM Agreement outlaws
them.
The subtlety with which the GATT framers dealt with the issue of subsid-
ies reflects an understanding among the Contracting Parties that subsidies are a
widely used instrument of government policy which is by no means necessarily
economically inefficient (eg positive externalities and public goods, which are not
internalized in market prices for particular products, may justify subsidization).
At the same time, however, subsidies can also be seen as undermining the market
access expectations that result from bargained tariffs and other trade concessions:
bargaining down to a low tariff rate for, say steel imports, may be a pyrrhic vic-
tory if subsidies are granted in the importing country that are set so high as to
prevent any realistic possibility of import competition. In a nutshell, while the
GATT aimed to ensure that subsidies did not remove the incentive to make tar-
iff concessions, the WTO pursued other policy objectives, and tried to close off
other protectionist loop-holes following the successful overall reduction of tariffs
over a period of almost 50 years.²⁷⁰

4.2 From discouraging to outlawing export subsidies

Export subsidies are a trade instrument which expand trade and negatively affect
the terms of trade for the commodity to which they are tied. GATT, Art XVI,
as briefly discussed above, imposes two obligations on WTO members: to notify
their exports subsidies and to limit them, in order to address the concerns of their
affected trading partners. SCM, Art 3²⁷¹ goes much further and outlaws two
types of subsidies: export subsidies and local content requirements.²⁷²
An Illustrative List annexed to the SCM Agreement sets out a non-exhaus-
tive list of export subsidies. The list in Annex I of the SCM Agreement contains
12 illustrative types of export subsidies. The AB report on Brazil—Aircraft
(Art 21.5—Canada) clarified that a subsidy that falls under the purview of the
Illustrative List is ipso facto prohibited: there is no additional requirement to dem-
onstrate that the subsidy is contingent upon export performance under Art 3.1
(c) of the SCM.

²⁷⁰ According to Horlick and Clarke (1994), the SCM was possible thanks to a last minute
compromise between the United States and the European Community, which following years of
pursuing diametrically opposed opinions on this issue, lived a period of certain rapprochement
during the Clinton administration era.
²⁷¹ SCM, Art 3 is an Annex 1A agreement. Recall our discussion in Chapter 1: it prevails over
the GATT to the extent of the conflict. Standing case-law suggests that the SCM disciplines pre-
vail over Art XVI of the GATT.
²⁷² Local content requirements are, of course, not a trade instrument. It is a domestic subsidy,
hence a domestic instrument. It is treated, however, differently than any other domestic subsidy: it
is illegal, whereas all other domestic subsidies, as we will see in Chapter 3, are simply actionable.
182 Disciplines on Trade Instruments

Assuming a scheme is not reproduced in the Illustrative List, the complainant

must demonstrate that the subsidy at hand is either de jure or de facto an export
subsidy: the letter of Art 3.1 of the SCM (‘either in law or in fact’) leaves no doubt
that subsidies can be prohibited either de jure or de facto. De facto export con-
tingency operates as an anti-circumvention provision against attempts by WTO
members to link benefits to exports without explicitly stating in the law that this
has indeed been the case. The AB in its report on Canada—Aircraft discussed the
different evidentiary standards required to demonstrate the existence of a de jure
or a de facto subsidy. It explained why, in its view, the latter was a more demand-
ing standard in the following terms:
In our view, the legal standard expressed by the word ‘contingent’ is the same for both
de jure and de facto contingency. There is a difference, however, in what evidence may
be employed to prove that a subsidy is export contingent. De jure export contingency
is demonstrated on the basis of the words of the relevant legislation, regulation or legal
instrument. Proving de facto export contingency is a much more difficult task. There is
no single legal document which will demonstrate, on its face, that a subsidy is ‘contin-
gent . . . in fact. . . . Upon export performance.’ Instead, the existence of this relationship of
contingency, between the subsidy and the export performance, must be inferred from the
total configuration of the facts constituting and surrounding the granting of the subsidy,
none of which on its own is likely to be decisive in any given case. . . . We note that satis-
faction of the standard for determining de facto export contingency set out in footnote
4 requires proof of three different substantive elements: first, ‘the granting of a subsidy’;
second, ‘is . . . tied to . . .’; and third, ‘actual or anticipated exportation or export earnings.’
(original emphasis)
The AB report on US—FSC (Art 21.5—EC), citing prior relevant case-law,
explained its understanding of the evidentiary standard associated with proof
that a de jure export subsidy indeed occurred (§ 112):
We recall that in Canada—Autos, we stated:
. . . a subsidy is contingent “in law” upon export performance when the existence of
that condition can be demonstrated on the basis of the very words of the relevant legis-
lation, regulation or other legal instrument constituting the measure. . . . [F]or a subsidy
to be de jure export contingent, the underlying legal instrument does not always have to
provide expressis verbis that the subsidy is available only upon fulfilment of the condition
of export performance. Such conditionality can also be derived by necessary implication
from the words actually used in the measure. (original emphasis)
The evidentiary standard associated with a demonstration of a de facto export
subsidy was discussed at length in the panel report on Australia—Automotive
Leather II (§§ 9.36–9.66). In this case the panel found that a subsidy was de facto
an export subsidy based on the following factors:
(1) Australia agreed to pay Howe (a private economic operator) 30 million
Australian dollars in three installments, if Howe were to meet certain sales
and investment targets;
 Export Subsidies 183

(2) The terms of the contract between Australia and Howe did not require Howe
to export, though it provided the latter with incentives to do so;
(3) The government’s awareness at the time the contract was concluded that
Howe earned the majority of its income from exports was crucial in the
panel’s evaluation;
(4) For Howe to meet the targets set, exporting was passage obligé, since the
Australian market was too small to absorb its production.

However, the AB in Canada—Aircraft took a narrower view on this issue when

dealing with a subsidy paid by TPC (a Canadian entity) to Canadian aircraft pro-
ducers. The AB paid particular attention to footnote 4 of the SCM, and held that
mere knowledge that the beneficiary is exporting does not suffice for the de facto
threshold to be met. Something more is required. Footnote 4, which interprets
the term ‘subsidies contingent [. . .] in fact’ set out in Art 3.1 of the SCM reads:
This standard is met when the facts demonstrate that the granting of a subsidy, without
having been made legally contingent upon export performance, is in fact tied to actual
or anticipated exportation or export earnings. The mere fact that a subsidy is granted to
enterprises which export shall not for that reason alone be considered to be an export sub-
sidy within the meaning of this provision.
We quote from §§ 172—4 of the report:
The second substantive element in footnote 4 is “tied to”. The ordinary meaning of
“tied to” confirms the linkage of “contingency” with “conditionality” in Article 3.1(a).
Among the many meanings of the verb “tie”, we believe that, in this instance, because the
word “tie” is immediately followed by the word “to” in footnote 4, the relevant ordinary
meaning of “tie” must be to “limit or restrict as to . . . conditions”. This element of the
standard set forth in footnote 4, therefore, emphasizes that a relationship of condition-
ality or dependence must be demonstrated. The second substantive element is at the very
heart of the legal standard in footnote 4 and cannot be overlooked. In any given case,
the facts must “demonstrate” that the granting of a subsidy is tied to or contingent upon
actual or anticipated exports. It does not suffice to demonstrate solely that a government
granting a subsidy anticipated that exports would result. The prohibition in Article 3.1(a)
applies to subsidies that are contingent upon export performance.
We turn now to the third substantive element provided in footnote 4. The dictionary
meaning of the word “anticipated” is “expected”. The use of this word, however, does not
transform the standard for “contingent . . . in fact” into a standard merely for ascertaining
“expectations” of exports on the part of the granting authority. Whether exports were
anticipated or “expected” is to be gleaned from an examination of objective evidence.
This examination is quite separate from, and should not be confused with, the examin-
ation of whether a subsidy is “tied to” actual or anticipated exports. A subsidy may well
be granted in the knowledge, or with the anticipation, that exports will result. Yet, that
alone is not sufficient, because that alone is not proof that the granting of the subsidy is
tied to the anticipation of exportation.
184 Disciplines on Trade Instruments
There is a logical relationship between the second sentence of footnote 4 and the “tied
to” requirement set forth in the first sentence of that footnote. The second sentence of
footnote 4 precludes a panel from making a finding of de facto export contingency for the
sole reason that the subsidy is “granted to enterprises which export”. In our view, merely
knowing that a recipient’s sales are export-oriented does not demonstrate, without more,
that the granting of a subsidy is tied to actual or anticipated exports. The second sentence
of footnote 4 is, therefore, a specific expression of the requirement in the first sentence to
demonstrate the “tied to” requirement. We agree with the Panel that, under the second
sentence of footnote 4, the export orientation of a recipient may be taken into account as
a relevant fact, provided that it is one of several facts which are considered and is not the
only fact supporting a finding. (original emphasis)
Unfortunately, the AB did not complete its analysis in this respect. As a result,
we now know that awareness per se is not enough to establish de facto export-
contingency, but we are in the dark as to what else is required. Although the AB has
insisted that this is not the case, such case-law implicitly suggests an intent test:
assuming the complainant can demonstrate that the government intended the
challenged scheme to be an export subsidy, it will be quite hard for the defendant
to refute proof of violation of Art 3 of the SCM. Such an intent standard although
not required, is probably the safest way to establish export contingency.
Panels can request the help of the Permanent Group of Experts (PGE), estab-
lished in order to evaluate whether a scheme is a prohibited subsidy. The PGE
comprises trade experts and, if requested to pronounce on a case, its opinion
binds the panel (Art 4.5 of the SCM). So far, panels have refused to make use of
this institutional facility.²⁷³

4.3 Counteracting prohibited subsidies

Assuming a WTO member successfully challenges a prohibited subsidy, the
panel and/or the AB will, by virtue of Art 4.7 of the SCM, recommend that the
subsidy be withdrawn without delay.²⁷⁴ Panels that are asked to pronounce on
such issue are requested, in deviation of their usual procedures, to reach a speedy
judgment within 90 days (Art 4.6 of the SCM).
In case the recommendation to withdraw the subsidy is not followed, the
injured party can have recourse to appropriate countermeasures (Art 4.10 of the
SCM). A footnote to Art 4.10 of the SCM explains that appropriate means not
disproportionate (sic).²⁷⁵ Consequently, one observes a variance between the remedy

²⁷³ Seriously inquiring into the motives of panels goes beyond the purposes of this book. The
refusal to do so, one could presume, should be WTO Secretariat motivated, the fi xed element
across all panel compositions (see Chapter 5).
²⁷⁴ Th is is the only case where the content of a recommendation (see Chapter 5) is specific; nor-
mally, WTO adjudicating bodies will recommend that the concerned partu bring its measures into
compliance (Art 19 of the DSU).
²⁷⁵ Footnotes 9 and 10 provide as follows with respect to the term ‘appropriate
countermeasures’:
 Export Subsidies 185

in the case a prohibited subsidy is maintained, and the remedy in all other viola-
tions of the WTO Agreement: Art 22.4 of the DSU (which is applicable in all
other cases) uses the term ‘equivalent’ to capture the relationship between the
damage inflicted and the amount of permissible counteraction against a persist-
ing violation. Similarly, Art 7.9 of the SCM provides that in case of actionable
subsidies, countermeasures ‘commensurate with the degree and nature of the
adverse effects determined to exist’ may be authorized: the damage suffered by
the complainant forms the benchmark for determining the level of countermeas-
ures in both cases. In the case of prohibited subsidies that have not withdrawn
following a Dispute Settlement Body (DSB) recommendation, however, the
countermeasures may be appropriate, as long as they are not disproportionate. The
choice of words cannot be accidental.
The term ‘appropriate countermeasures’ is far from being self-interpreting.
Based on the notion that the case of prohibited subsidies is the only instance
where the WTO Agreement not only explicitly outlaws a practice (whereas, in all
other cases outlawing a practice is the privilege of the WTO adjudicating body),
but also modifies the substantive content of a panel’s (or AB’s) recommendation,
WTO adjudicating bodies have consistently (so far) held the view that the pun-
ishment of prohibited subsidies through countermeasures should be harder than
the punishment of any other breach of the WTO Agreement.
The report on Brazil—Aircraft (Art 22.6—Brazil) had the opportunity to
clarify the ambit of appropriate countermeasures and explain the relationship
between Art 4.10 of the SCM and Art 22.4 of the DSU.²⁷⁶ It linked the amount
of the countermeasures to the amount of the subsidy. This case (and its ‘twin’
dispute, Canada—Aircraft) concerned (export) subsidization by Canada and
Brazil of their respective national (regional) aircraft producers. A duopoly was in
place whereby both parties, confirming Brander and Spencer’s predictions, saw
an interest in subsidizing their domestic producer. To base its finding that the
quantification of appropriate countermeasures should be linked to a benchmark
other than the damage suffered by the complainant (as is the case under Art 22.4
of the DSU for all violations of the WTO Agreement and Art 7.9 of the SCM in
the case of actionable subsidies), the Arbitrators first explained the difference they
saw in the function of the remedy against a prohibited subsidy, as opposed to rem-
edies to address any other nullification or impairment of WTO members’ rights.
Important in their reasoning was the fact that they considered that the purpose of
Art 4 of the SCM is to achieve the withdrawal of the prohibited subsidy (§ 3.48):
. . . the purpose of Article 4 is to achieve the withdrawal of the prohibited subsidy. In
this respect, we consider that the requirement to withdraw a prohibited subsidy is of
a different nature than removal of the specific nullification or impairment caused to a

‘This expression is not meant to allow countermeasures that are disproportionate in

light of the fact that the subsidies dealt with under these provisions are prohibited’.
²⁷⁶ Report of the Arbitrators, Brazil—Aircraft (Art 22.6—Brazil), paras 3. 42–3.60.
186 Disciplines on Trade Instruments
Member by the measure. The former aims at removing a measure which is presumed
under the WTO Agreement to cause negative trade effects, irrespective of who suffers
those trade effects and to what extent. The latter aims at eliminating the effects of a meas-
ure on the trade of a given Member; the fact that nullification or impairment is estab-
lished with respect to a measure does not necessarily mean that, in the presence of an
obligation to withdraw that measure, the level of appropriate countermeasures should be
based only on the level of nullification or impairment suffered by the Member requesting
the authorisation to take countermeasures.²⁷⁷
They then rejected arguments by Brazil to the effect that the proposed bench-
mark, the amount of the subsidy, was not reasonable. The Arbitrators argued
that the subsidy benchmark was not too onerous since, in all likelihood, Brazil
gained much more from its subsidies than it actually invested. They also rejected
an argument to the effect that their benchmark amounted to (an impermissible
under WTO practice) punitive damages. We quote from §§ 3.54 and 3.55:
Our interpretation of the scope of the term ‘appropriate countermeasures’ in Article 4
of the SCM Agreement above shows that this would not be the case. Indeed, the level
of countermeasures simply corresponds to the amount of subsidy which has to be with-
drawn. Actually, given that export subsidies usually operate with a multiplying effect (a
given amount allows a company to make a number of sales, thus gaining a foothold in a
given market with the possibility to expand and gain market shares), we are of the view
that a calculation based on the level of nullification or impairment would, as suggested
by the calculation of Canada based on the harm caused to its industry, produce higher
figures than one based exclusively on the amount of the subsidy. On the other hand, if the
actual level of nullification or impairment is substantially lower than the subsidy, a coun-
termeasure based on the actual level of nullification or impairment will have less or no
inducement effect and the subsidizing country may not withdraw the measure at issue.
Brazil also claimed that countermeasures based on the full amount of the subsidy
would be highly punitive. We understand the term ‘punitive’ within the meaning given to
it in the Draft Articles. A countermeasure becomes punitive when it is not only intended
to ensure that the State in breach of its obligations bring its conduct into conformity with
its international obligations, but contains an additional dimension meant to sanction the
action of that State. Since we do not find a calculation of the appropriate countermeas-
ures based on the amount of the subsidy granted to be disproportionate, we conclude
that, a fortiori, it cannot be punitive.²⁷⁸
The same logic (to link the amount of countermeasures to the amount of subsidy
paid) was followed in the Arbitrators’ report on US—FSC (Art 22.6—US). The
Arbitrators, extensively referring to public international law and the International
Law Commission (ILC) reports on State responsibility,²⁷⁹ held that the European
Community (complainant) should be authorized to adopt countermeasures up
to US$ 4,043 million, that is, the amount of subsidies paid by the United States to

²⁷⁷ Report of the Arbitrators, Brazil—Aircraft (Art 22.6—Brazil), para 3. 48.

²⁷⁸ Report of the Arbitrators, Brazil—Aircraft (Art 22.6—Brazil), §§ 3.54–3.55.
²⁷⁹ Report of the Arbitrators, US—FSC (Art 22.6—US), paras 5.30–5.62.
 Export Subsidies 187

its national producers (beneficiaries under the Foreign Sales Corporation (FSC)
scheme).²⁸⁰ One should, however, add a caveat here: the Arbitrators’ claim that
following calculation of the elasticities, they would have ended up with a similar
number, had they used the EC trade eff ects as the benchmark.²⁸¹ The Arbitrators
clarified that trade effects are not a priori ruled out as a benchmark. They were
simply of the view that Art 4.10 of the SCM does not require a trade effects test to
be used, nor did it limit the countermeasures in this way.²⁸²
To allow one complaining member to take countermeasures of an amount
equal to the full amount of the subsidy may prove problematic in cases of sequen-
tial enforcement, where more members decide to challenge the same measure in
subsequent WTO proceedings. This report, US—FSC (Art 22.6—US), added
a few words to address the (hypothetical) situation where, subsequent to the EC
challenge, another WTO member decided to attack the same US measure (the
FSC scheme):
This is an appropriate point at which to underline that there is one matter that is particu-
lar to the circumstances of this case and is material to this conclusion, yet has not—up to
this point—been expressly dealt with.
In the circumstances of this case, the European Communities is the sole complainant
seeking to take countermeasures in relation to this particular violating measure. That
is also, in our view, a relevant consideration in our analysis. Had there been multiple
complainants each seeking to take countermeasures in an amount equal to the value of
the subsidy, this would certainly have been a consideration to take into account in evalu-
ating whether such countermeasures might be considered to be not “appropriate” in the
circumstances. That is not, however, the situation before us.
The reasoning we have followed above could be construed—in a purely abstract man-
ner—to be as inherently applicable to any other Member as to the complainant in this
case viz. the European Communities. We would simply underline, in this regard, that
in this case, we were not presented with a multiple complaint but a complaint by one
Member. Thus we have not been obliged to consider whether or how the entitlement to
countermeasures based on our reasoning above should be allocated across more than one
complainant. Thus to the extent that there would be an issue of allocation, as it were, it
need not—and did not—enter into consideration as an element to otherwise “discount”
the European Communities’ entitlement to countermeasures in this particular case.
Understandably, it would be our expectation that this determination will have the
practical effect of facilitating prompt compliance by the United States. On any hypothesis
that there would be a future complainant, we can only observe that this would give rise
inevitably to a different situation for assessment. To the extent that the basis sought for

²⁸⁰ Report of the Arbitrators, US—FSC (Art 22.6—US), paras 6.1–6.30. With respect to the
term ‘not disproportionate’, the arbitrators considered that ‘the entitlement to countermeasures is
to be assessed in light of the legal status of the wrongful act and the manner in which the breach of
that obligation has upset the balance of rights and obligations as between Members. It is from that
perspective that the judgment as to whether countermeasures are disproportionate is to be made’,
Arbitrator, US—FSC (Art 22.6—US), para 5.24.
²⁸¹ Report of the Arbitrators, US—FSC (Art 22.6—US), para 6.57.
²⁸² Report of the Arbitrators, US—FSC (Art 22.6—US), paras 6.33–6.34.
188 Disciplines on Trade Instruments
countermeasures was purely and simply that of countering the initial measure (as opposed
to, e.g., the trade effects on the Member concerned) it is conceivable that the allocation
issue would arise (although due regard should be given to the point made in footnote 84
above). We take note, on this point, of the statement by the European Communities:
. . . it may well be that the European Communities would be happy to share the task
of applying countermeasures against the United States with another member and
voluntarily agree to remove some of its countermeasures so as to provide more scope
for another WTO member to be authorized to do the same. This will be another fact
that future arbitrators could take into consideration.
It must be stressed, however, that there is no mechanical automaticity to this. The
essence of such assessments is that it is a matter of judging what is appropriate in the case
at hand. There could well be other factors to take into account in their own right, e.g., if
for instance the matter of bilateral trade effects were essentially at issue.²⁸³
Consequently, on both occasions where a WTO adjudicating body had to define
the term ‘appropriate countermeasures in the context of prohibited subsidies’, it did
so by linking it to the amount of subsidy paid rather than the trade effects caused.
From a legal perspective, the foundation would be that the damage, in case
recourse to a prohibited subsidy is being made, is not the trade effects caused, but
rather the act of providing prohibited export/import substitution subsidization
itself.
Interestingly, the Arbitrators in Canada—Aircraft Credits and Guarantees (Art
22.6—Canada), saw force in the argument that there was need to induce Canada
to comply, in light of its statements before the panel that it did not intend to
do so. The Arbitrators used the amount of the subsidy as the benchmark²⁸⁴ and
calculated the amount of countermeasures to be US$206,497,305.²⁸⁵ They then
continued, however, to examine whether any adjustments needed to be made to
this amount to make it ‘an appropriate level of countermeasures’. In their view, an
upward adjustment of this amount was justified in order to induce compliance,

²⁸³ Report of the Arbitrators, US—FSC (Art 22.6—US), paras. 6.26–6.30. The Arbitrators’
claim that that they would have ended up with the same amount, had they used the EC trade effects
as benchmark to quantify the appropriateness of countermeasures, and this passage seem hard to rec-
oncile. The Arbitrators calculated total trade effects (something which is discernible from the report):
then if the trade effects calculation is correct, this is a case where (total) trade effects yield a number
as high as the amount of subsidy paid. However, since the number chosen is a number within a range
of possibilities, we simply do not know if the EC injury is within the lower or the higher ebb of the
range. In other words, the European Communities might have been over- or under-compensated
depending on the placement of its injury within the range calculated in the Arbitrators’ report. Be
it as it may though, Esserman and Howse (2004) voiced their dissatisfaction with this report, argu-
ing that the ultimate remedy was clearly disproportionate, in violation of the standard enshrined in
Art 4.10 of the SCM.
²⁸⁴ Report of the Arbitrators, Canada—Aircraft Credits and Guarantees (Art 22.6—Canada),
para 3.51. The amount of the subsidy was calculated on the basis of the benefit to the recipient,
ie the benefit conferred by the loan, rather than the amount of the loan as such. Report of the
Arbitrators, Canada—Aircraft Credits and Guarantees (Art 22.6—Canada), para 3.60.
²⁸⁵ Report of the Arbitrators, Canada—Aircraft Credits and Guarantees (Art 22.6—Canada),
para 3.90.
 Export Subsidies 189

in light of Canada’s statements that it would not withdraw the subsidy.²⁸⁶ So the
Arbitrators added 20% to the level of the countermeasures in order to induce
compliance:
Recalling Canada’s current position to maintain the subsidy at issue and having regard
to the role of countermeasures in inducing compliance, we have decided to adjust the
level of countermeasures calculated on the basis of the total amount of the subsidy by an
amount which we deem reasonably meaningful to cause Canada to reconsider its current
position to maintain the subsidy at issue in breach of its obligations. We consequently
adjust the level of countermeasures by an amount corresponding to 20 per cent of the
amount of the subsidy as calculated in Section III.E above, i.e.:
US$206,497,305 x 20% (US$41,299,461) = US$247,796,766.
As we have noted in paragraph 3.120, adjustments such as the one we are making can-
not be precisely calibrated. There is no scientifically based formula that we could use to
calculate this adjustment. In that sense, the adjustment might be viewed as a symbolic
one. Even so, we are convinced that it is a justified adjustment in light of the circum-
stances of this case and, in particular, the need to induce compliance with WTO obliga-
tions. Without such an adjustment, we would not be satisfied that an appropriate level of
countermeasures had been established in this case.²⁸⁷
It is probably warranted that we do not see too much in this precedent, in light
of the rather idiosyncratic facts associated with this litigation (Canada’s repeated
statements before the panel that it did not intend to comply with any recom-
mendation). This is, however, the only genuine case of punitive damages, that is
damages dissociated from the legal wrong: no matter what the standard is, trade
effects, or the payment of an illegal subsidy, the 20% mark-up is certainly dissoci-
ated from either benchmark.

²⁸⁶ Report of the Arbitrators, Canada—Aircraft Credits and Guarantees (Art 22.6—Canada),
para 3.107.
²⁸⁷ Report of the Arbitrators, Canada—Aircraft Credits and Guarantees (Art 22.6—Canada),
§§ 3.121–3.122.
 3
Domestic Instruments

1 National Treatment and its Rationale 193

2 The coverage 195
2.1 Wide coverage 195
2.2 Subsidies 195
2.2.1 What is a (domestic) subsidy? 195
2.2.1.1 Financial contribution 196
2.2.1.2 Benefit to recipient 198
2.2.1.3 The specificity requirement 200
2.2.2 The classification of subsidies 200
2.2.3 Counteracting subsidies 201
2.2.4 Remedies 201
2.2.5 Farm subsidies 202
2.2.5.1 The regulation of farm trade in the GATT era 202
2.2.5.2 The WTO Agreement in a nutshell 206
2.2.5.3 AG over SCM 208
2.3 Government procurement 210
2.3.1 What is government procurement? 210
2.3.2 The basic discipline: non-discrimination 211
2.3.3 The past and present of GPA membership 211
2.4 The coverage of Art III of the GATT—revisited 212
3 National Treatment, and Negative Integration 215
4 National Treatment: The Legal Discipline 216
4.1 Fiscal measures 216
4.1.1 DCS products 217
4.1.1.1 Defining DCS 217
4.1.1.2 SATAP 221
4.1.2 Like products 234
4.1.2.1 Defining like products 234
4.1.2.2 Taxation in excess 236
4.2 Non-fiscal measures 237
4.2.1 A measure . . . 237
4.2.2 . . . aff ecting sale, off ering for sale, . . . 238
4.2.3 . . . of like products 238
4.2.4 Less favourable treatment 242
4.3 A critique of the case-law concerning NT 247
4.4 General exceptions 254
4.4.1 Coverage 254
 Domestic Instruments 191
4.4.2 An exception to which provisions? 254
4.4.3 Regulatory diversity 259
4.4.4 A two-tier test for compliance with Art XX
of the GATT 260
4.4.5 Public morals (Art XX(a) of the GATT) 261
4.4.5.1 Coverage 261
4.4.5.2 The necessity requirement 262
4.4.6 Human, animal and plant life
(Art XX(b) of the GATT) 262
4.4.6.1 Coverage 262
4.4.6.2 The necessity requirement 262
4.4.6.3 No science required 263
4.4.7 Domestic laws and regulations (Art XX(d)
of the GATT) 264
4.4.7.1 Coverage 264
4.4.7.2 The necessity requirement 264
4.4.8 Exhaustible natural resources (Art XX(g)
of the GATT) 266
4.4.8.1 Coverage 266
4.4.8.2 Jurisdictional limits 268
4.4.8.3 Measure must be relating to the
objective pursued 269
4.4.8.4 Measures restricting domestic production or
consumption 270
4.4.9 The chapeau of Art XX of the GATT 271
4.4.9.1 The scope of the chapeau 271
4.4.9.2 Complying with the chapeau 271
4.4.9.2.1 Arbitrary or unjustifiable discrimination 271
4.4.9.2.2 Disguised restriction of trade 274
4.4.10 A critique of the case-law 276
4.4.10.1 The discontents of an exhaustive list 276
4.4.10.2 Extra-territoriality 277
4.4.10.3 A two-tier test? 285
4.4.10.4 The standard of review 285
4.5 A specific exception: films 286
5 Dealing with NTBs 287
5.1 The TBT and SPS Agreements 287
5.2 The legal relationship between GATT, TBT, and SPS 288
5.3 TBT 288
5.3.1 Coverage 288
5.3.2 International standards 289
5.3.3 Unilateral technical regulations 295
5.3.4 Domestic standards 296
5.4 The WTO SPS Agreement 297
5.4.1 Coverage 297
5.4.2 The SPS Agreement in a nutshell 298
5.4.3 International standards 299
5.4.4 Unilateral measures based on scientific evidence 299
192 Domestic Instruments
5.4.5 Measures adopted on the basis of precaution 309
6 State Trading Enterprises (STEs) 315
6.1 STEs: definitional issues 315
6.2 The legal obligation in a nutshell 316

Overview
The GATT disciplines domestic instruments as well since, by virtue of the equiva-
lence propositions, trade instruments (say a tariff ) can be decomposed into domes-
tic instruments (production subsidy cum consumption tax), and, as a result, absent
commitment on domestic instruments, tariff concessions could be eroded. Since
tariff protection is gradually becoming meaningless (with the notable exception
of the two tariff peaks, farm and textile goods), the discipline on domestic instru-
ments becomes all the more important. The identification of domestic instruments
is a daunting task: regulatory diversity and change in social preferences increase
substantially the negotiating costs, assuming willingness to complete the contract
during the negotiation stage. Wisely, the GATT founding fathers did not identify
one by one the domestic instruments coming under the purview of the GATT;
they simply stated that all instruments affecting trade are covered by the non-
discrimination obligation, which, with respect to domestic instruments, is called
national treatment. The GATT provides some specificity though: it explicitly
mentions that local content requirements come under the purview of the obligation
imposed, and excludes from its coverage subsidies and government procurement.
The Uruguay round SCM, another Annex 1A agreement (like the GATT) regu-
lates subsidies. Starting from the perspective that subsidies hurt trading partners
(by adopting an injury to competitors test), the Uruguay round agreement outlaws
some subsidies, while allowing trading partners to counteract the effect of all non-
prohibited subsidies. On the other hand, some WTO members have signed a
plurilateral agreement on government procurement whereby they essentially
re-introduced the non-discrimination obligation in their inter se relations. We
will review the main features of both agreements in this chapter. Going back to
the GATT discipline, we should note that contract incompleteness is not, in and
of itself, problematic. It becomes problematic, however, if the interested parties
behave in an opportunistic manner. The WTO members have such an incen-
tive: since, in the majority of cases, they will possess private information about
their intended objectives, they could cheat and reveal that, through their pol-
icies, they are pursuing GATT consistent objectives. The instrument to verify
what indeed a domestic instrument was after, is national treatment. National
treatment is legalese for non-protectionism, and, there is nothing like an oper-
ational definition of protection. Recourse to proxies is inevitable. Case-law shows
a remarkable inertia to come up with a coherent, intelligible approach in this
 National Treatment and its Rationale 193

context. If at all, it has made matters worse, substantially worse: by adopting a

no effects, no intent standard of review, case-law has been routinely punishing
tax differentials which could occur for a number of reasons including, but not
necessarily confined to, protectionism. Case-law thus confuses ends (protection-
ism) and means (tax differentials). National treatment also means that the GATT
is about negative integration with respect to domestic instruments: policies will
be defined unilaterally and (some) external effects will be internalized through
the national treatment obligation. National treatment, in other words, does not
exhaust the potential for beggar thy neighbour; it just makes it costlier. There are
exceptions to national treatment: unambiguously, film quotas are such an excep-
tion; case-law has added many more, although it is quite questionable whether the
case-law additions should be regarded as exceptions to national treatment. New
generation agreements, like the TBT and the SPS, have adopted a more intelligent
approach towards combating protectionism originating in domestic instruments:
they contain proxies, additional to non-discrimination, which can help illuminate
the uninformed party about the intentions of the regulating entity: measures must
not be more restrictive than necessary to attain the objective sought; recourse to
international standards, when appropriate, is warranted; in principle, recourse to
scientific evidence is compulsory (a requirement under SPS only); measures must
be internally consistent, that is levels of protection should not be arbitrarily dis-
similar, so as to hide beggar thy neighbour behaviour (a requirement under SPS
only). Their coverage nevertheless, is limited to a few domestic instruments only.
Case-law in the context of TBT/SPS has not been problem-free: the use of inter-
national standards and the science/no science divide have been handled in a rather
clumsy manner. In this chapter, besides national treatment, TBT and SPS, we
will also review the application of non-discrimination to a transaction explicitly
mentioned in the GATT: state trading enterprises (marketing boards). Section 1
discusses the rationale for national treatment, while section 2 focuses on its cover-
age. In section 3, we explain why because of national treatment, the GATT should
be viewed as a negative integration contract with respect to domestic instruments.
An analysis of the legal discipline (national treatment) is presented in section 4
where, we also discuss the exceptions to national treatment. Section 5 deals with
TBT and SPS, whereas section 6 deals with state trading enterprises.

1 National Treatment and its Rationale

The GATT discipline imposed on all domestic instruments affecting trade is

national treatment (NT): WTO members cannot afford protection to domestic
production through internal measures. As a result, all protection¹ in the GATT

¹ Protection, as briefly argued in Chapter 1, is an elusive notion. It bears repetition that,

although there are various approaches and suggested definitions in economic theory, we still
lack an operational definition of this term. Recall that this term is understood in this book as
194 Domestic Instruments

regime is a matter of negotiation and can take one form only: tariff protection.
In other words, once imported products have paid their ticket to entry (in the
form of tariffs) into a particular market, they should be assimilated to domestic
products, and be subjected to a regulatory regime identical to that applied to
domestic products.
It becomes gradually obvious that NT serves the same purpose as MFN:
concession erosion. Recall our discussion about the equivalence propositions: a
tariff can, for example, be decomposed in a domestic tax (on consumers) and
a domestic subsidy (on producers). It follows that beggar thy neighbour policies
can be pursued through domestic instruments as well. Note, however, that this
time concession will not be eroded because a larger advantage will be, subsequent
to the original negotiation, granted to another foreign producer; they will be
eroded because an advantage will be now granted to the domestic producer. NT
is thus insurance policy against the risk of seeing tariff commitments becoming
meaningless, by virtue of unilaterally defined internal policies.²
An insurance policy against concession erosion is necessary for WTO mem-
bers to have the incentive to continue negotiating and further liberalize trade;
this is in essence the rationale for the inclusion of NT in the GATT:³ NT is
supposed to guarantee that the end-product of a multilateral negotiation will not
be undone through unilateral, subsequent actions that affected trading partners
cannot influence.⁴
The AB in its report on Japan—Alcoholic Beverages II (p 16) affirmed this
understanding of NT in the following terms:

suggesting intended conferral of an advantage to the domestic producer. We will see in what pre-
cisely it translates in GATT/WTO case-law in this chapter. On this issue, see Bagwell and Staiger
(1995), Bagwell et al (2002), Grossman and Helpman (2002), Horn (2006), Neven (2001), Regan
(2006), and Srinivasan (2005).
² It is reasonable to assume that WTO members, when negotiating concessions, have in mind
the productivity not only of their own sector, but also of their rivals. So when A requests say a 10%
import duty from B on wheat, A should have made its calculations translating the 10% promise
into concrete market access in B’s market. B’s productivity in wheat is part of that calculation.
If B is allowed, subsequent to the negotiation, to impose an internal consumption tax of 1% on
domestic wheat and 30% on imported wheat, it will be, other things equal, upsetting B’s calcula-
tions and eroding the promise given. NT obliges A to impose either (in my example) 1% or 30%
on both A and B wheat and thus, stops B from accepting the competitive relationship between the
two products.
³ Research in the context of national treatment is still in statu nascendi. Horn (2006) and Horn
et al (2006) stand out as the more important contributions in this area. These papers focus more on
the incompleteness of Art III of the GATT, and borrow from contract theory in order to explain
why it is probably wise that Art III of the GATT has been drafted in its current form (we return to
this issue in what immediately follows in section 2). Saggi and Sara (2006) employ a model with
two asymmetric countries and one good to show that the high quality country benefits from NT,
if the size of their markets is equal. The size of the market can act as a countervailing force. NT, in
other words, is probably not a great idea for small countries producing low quality goods.
⁴ This is of course the case by virtue of domestic constitutional laws. Domestic laws will be
defined through domestic constitutional processes and, absent express transfer of sovereignty, are
not a matter of international negotiation.
 The Coverage 195
. . . The broad and fundamental purpose of Article III is to avoid protectionism in the
application of internal tax and regulatory measures. More specifically, the purpose of
Article III ‘is to ensure that internal measures not be applied to imported or domestic
products so as to afford protection to domestic production.’ Toward this end, Article III
obliges Members of the WTO to provide equality of competitive conditions for imported
products in relation to domestic products.⁵

2 The Coverage

2.1 Wide coverage

NT extends to regulatory interventions of both fiscal, and non-fiscal nature,
and covers transactions relating to both bound, and unbound items. The AB left
no doubt with its interpretation in this respect (Japan—Alcoholic Beverages II,
p 17):
. . . The Article III national treatment obligation is a general prohibition on the use of
internal taxes and other internal regulatory measures so as to afford protection to domestic
production. This obligation clearly extends also to products not bound under Article II.
GATT, Art III does not specify which measures come under its purview. There
are two exceptions to this approach:
first, Art III.5 of the GATT explicitly outlaws local content requirements: under
this provision, all measures whereby a government privileges goods containing a
set amount of domestic added value are outlawed. This prohibition has been fur-
ther underscored through the advent of the TRIMs Agreement which contains
an illustrative list of such measures. No other domestic instruments (other than
local content requirements), are explicitly prohibited by Art III of the GATT;
second, two measures are explicitly exempted from the national treatment
obligation by virtue of Art III.8 of the GATT: subsidies, and government
procurement.

2.2 Subsidies
2.2.1 What is a (domestic) subsidy?
Domestic subsidies are not prohibited. The GATT did not include an elaborate
framework to entertain the subsidies issue: it does not even reflect a definition
of ‘subsidy’. The first definition is reflected in the steel disputes between the

⁵ A note by the GATT Secretariat summarizing the meetings of the Working Party on Border
Adjustments on 18 to 20 June 1968 (GATT Doc L/3039 of 11 July 1968) captures the same point
in the following terms: ‘In the case of Article III, the rules were designed to safeguard tariff conces-
sions and to prevent hidden discrimination’.
196 Domestic Instruments

United States and the European Community.⁶ There, the panel introduced the
idea that, for a subsidy to exist, the measure challenged must be attributed to a
government. The report, however, was never adopted, and, as a result, its merit on
this issue notwithstanding is of limited legal value. GATT, Art XVI was designed
to discipline subsidies, without inquiring at all into their welfare implications;⁷
there is an implicit acknowledgement in Art XVI of the GATT, that subsidies
should be discouraged, but not outlawed.⁸
The Tokyo round SCM was enacted in 1979 and regulated subsidies in more
detail. It was, however, binding only on a sub-set of the GATT membership,
the countries that voluntarily adhered to this agreement, and it still lacked a
definition of ‘subsidy’.
The disciplines on subsidies underwent substantial transformation during
the Uruguay round negotiations. The new SCM Agreement is a multilateral
agreement, that is, it binds all WTO members.
The SCM stipulates that for a measure to constitute a subsidy the following
conditions must be met:
(1) a financial contribution or income support by a government;⁹ which
(2) confers a benefit;
(3) to a specific recipient(s).

2.2.1.1 Financial contribution

A financial contribution can consist of direct transfer of funds, or of omission
to request payments due; it can be performed by the government, or by a private
entity that a government entrusted with the operation (Art 1 of the SCM).
The SCM mentions illustrations of transfer of funds: grants, loans, equity
infusions, as well as loan guarantees are examples of such transfers. A financial
contribution also exists when the government does not collect, or, in general,
foregoes revenue which is otherwise due. But who determines which revenue the
government is entitled to? In other words, when is government revenue ‘other-
wise due’? In US—FSC, the panel and the AB had to deal head on with this issue.
The European Community challenged the US practice to exempt US economic

⁶ See the panel report on US—Lead and Bismouth I.

⁷ See on this issue, Janow and Staiger (2003). On the economics of the subsidies regulation, see
Sykes (2003a).
⁸ The attitude of the GATT on subsidies was benign. Indeed the original Art XIV which was
meant to regulate subsidies (see UN Economic and Social Council Doc E/PC/T/C.6/85 of 15
February 1947) provided for an obligation to consult, when subsidies harmed foreign producers.
As noted in Chapter 2, Bagwell and Staiger (2006) take issue with the change of the attitude
towards subsidies, as a result of the Uruguay round agreements.
⁹ One might still find in literature the term ‘cost to government’ used as synonym to financial con-
tribution. In its report on Canada—Aircraft, the AB confirmed the panel’s understanding regard-
ing the irrelevance of the term ‘cost to government’ for the purposes of interpreting the SCM, In its
view, this term can be misleading since financial contributions are not necessarily a cost to government
and a cost to government does not necessarily confer a benefit. The AB privileged the use of the terms
as they appear in Art 1 of the SCM (§§ 149–61).
 The Coverage 197

operators from payment of taxes for income earned outside the US territory.
In the EC view, such omissions constituted an export subsidy. The European
Community argued that the income was taxable under US tax laws; it is, the EC
delegates submitted, US tax laws that should constitute the benchmark to find
whether a subsidy had been granted. The panel originally, and the AB subse-
quently, agreed with the EC view on this score and held accordingly that (§ 90):
We also agree with the Panel that the basis of comparison must be the tax rules applied by
the Member in question. To accept the argument of the United States that the compara-
tor in determining what is “otherwise due” should be something other than the prevailing
domestic standard of the Member in question would be to imply that WTO obligations
somehow compel Members to choose a particular kind of tax system; this is not so.¹⁰
A third type of financial contribution listed in Art 1 of the SCM was discussed
in US—Softwood Lumber IV, where the AB entertained an argument whether the
Canadian stumpage arrangements constituted a subsidy in accordance with Art
1 of the SCM. In the AB’s view, since stumpage arrangements give tenure holders
a right to enter onto government lands, cut standing timber, and enjoy exclusive
rights over the timber that is harvested, such arrangements represent a situation
in which provincial governments provide standing timber. It thus disagreed with
Canada’s argument that the granting of an intangible right to harvest standing
timber cannot be equated with the act of providing that standing timber, and
held that such measures could very well come under the purview of Art 1 of the
SCM (§ 75).
Case-law seems to suggest that a control criterion is most appropriate in order
to ascertain whether a body granting a subsidy is governmental or not. Such a
criterion can cut in a consistent manner across WTO members with substantial
regulatory diversity on this score. The question to ask in this context is whether a
particular body is effectively (de facto) controlled by a government, irrespective
of idiosyncratic (eg WTO member-specific) legal manipulations. If the response
to this question is affirmative, then the body at hand will be acknowledged to be
a government body.
The SCM also contains an anti-circumvention provision on this score: gov-
ernments that ‘channel’ action through private entities can still be found to be
subsidizing, if it is demonstrated that the entity providing subsidies is directed
(entrusted) to do so by the government. In US—Countervailing Duty Investigation
on DRAMS, the AB reached the following conclusion concerning the meaning of
government entrustment or direction of private bodies (§ 116):

¹⁰ Taxation is a domestic issue. Consequently, the European Community cannot, for example,
accuse the United States of subsidizing its domestic industry by not adhering to the Kyoto Protocol.
It is up to the US government to decide to adhere or not, and a refusal to adhere does not constitute
a subsidy (in favour of US industry) in the sense of Art 1 of the SCM. In this vein, the arguments
advanced by the French Prime Minister Dominique de Villepin should be dismissed (see ‘French
PM calls for EU carbon levy’, The Financial Times, 13 November 2006).
198 Domestic Instruments
In sum, we are of the view that, pursuant to paragraph (iv), “entrustment” occurs where
a government gives responsibility to a private body, and “direction” refers to situations
where the government exercises its authority over a private body. In both instances, the
government uses a private body as proxy to effectuate one of the types of financial con-
tributions listed in paragraphs (i) through (iii). It may be difficult to identify precisely,
in the abstract, the types of government actions that constitute entrustment or direction
and those that do not. The particular label used to describe the governmental action
is not necessarily dispositive. Indeed, as Korea acknowledges, in some circumstances,
“guidance” by a government can constitute direction. In most cases, one would expect
entrustment or direction of a private body to involve some form of threat or inducement,
which could, in turn, serve as evidence of entrustment or direction. The determination of
entrustment or direction will hinge on the particular facts of the case.
A finding of entrustment or direction, therefore, requires that the government
give responsibility to a private body—or exercise its authority over a private
body—in order to effectuate a financial contribution.¹¹
Apart from a financial contribution by the government, any form of income
or price support which confers a benefit may also be considered a subsidy (Art
1.1(a)(2) of the SCM). Income or price support mechanisms play an important
role in agricultural goods, and commodities in general.¹²

2.2.1.2 Benefit to recipient

For a benefit to be demonstrated, a government needs to show that the recipient
obtained an advantage which it could not obtain in the market place. There is,
however, a sequence between financial contribution and benefit. As the AB put it
in its report on US—Countervailing Duty Investigation on DRAMS, if no contri-
bution took place, no benefit can result either (§ 205).
Case-law has used the private investor test¹³ to establish whether a benefit has
been conferred. The intuition there is that a subsidy is whatever the beneficiary
economic operator could not have procured under market conditions.¹⁴ SCM, Art
14 which deals with the calculation of the amount of a subsidy in terms of benefit
to the recipient makes explicit what is implied in SCM, Art 1.1(b): it provides that

¹¹ Circumstantial evidence was enough for the AB to reach this conclusion.

¹² See, for example, the panel report on US—Upland Cotton for an appropriate illustration.
¹³ There is, however, also case-law which uses the cost of production as benchmark to define
whether a benefit has been bestowed. Th is narrower construction has been used in two cases only
(in Canada—Dairy and, more recently, in the EC—Export Subsidies on Sugar case). Both cases
concerned farm subsidies, and were adjudicated under the WTO AG. In the latter case, the AB
accepted a claim by the complainant that the EC regime effectively led to cross-subsidization, even
though the arguments in support of this thesis cover only a sub-set of the possible situations that
can give rise to cross-subsidization.
¹⁴ In practice, this test can prove quite difficult to implement. Assume for example, that com-
pany X issues new shares at a price say $100/share. Assume that on day one, government buys 70%
of the new shares. Is it a subsidy (since, probably, prices of the remaining shares will shoot up as a
result)? What if the government buys 50% of the shares, or say, 50% 2 months after the opening of
the procedure, and assuming that the private sector has not or has with little enthusiasm responded
to the sale?
 The Coverage 199

the term “benefit” refers to benefit to the recipient, and uses the market or private
investor as the benchmark for determining the existence and amount of benefit.
The government provision of equity capital shall not be considered as conferring
a benefit unless the investment decision can be regarded as inconsistent with the
usual investment practice (including the provision of risk capital) of private inves-
tors in the territory of the member concerned. In case of a government loan, no
benefit is conferred unless there is a difference between the amount that the firm
receiving the loan pays on the government loan and the amount the firm would
pay on a comparable commercial loan which the firm could actually obtain on
the market. Similarly, a loan guarantee by the government confers a benefit in
case there is a difference between the amount that the firm receiving the guaran-
tee pays on a loan guaranteed by the government and the amount that the firm
would pay on a comparable commercial loan absent the government guarantee.
Finally, if the government provides goods or services, such financial contribu-
tion shall not be considered as conferring a benefit unless it is made for less than
adequate remuneration, which is to be determined in relation to the prevailing
market conditions for the good or service in question.
In US—Softwood Lumber IV, the question arose whether the list of bench-
marks included in Art 14 of the SCM is exhaustive (as the wording suggests),
or indicative. Recall that this provision requests that a comparison is made
between the market price in a given (exporting) market and the price at which
a good is being procured. In this case, however, it was difficult to establish what
the market price in the exporting market was: the overwhelming majority of
Canadian production of timber was taking place in the context of stumpage
arrangements; the lease, at least according to the US argument, was below a
market price. The United States chose, consequently, to disregard the Canadian
price, and compare the price charged to Canadian producers to the US market
price. Canada challenged this measure. The panel agreed with Canada: in its
view, the United States had to use a benchmark included in the list of Art 14 of
the SCM. The panel recognized the problems with the current drafting of this
provision, considered however, its role to be confined to the application of the
law (Art 3.2 of the DSU), rather than of its own perceptions about what law
should be. The AB disagreed with the panel: while, in its view, private prices in
the market of provision will generally represent an appropriate measure of the
adequacy of remuneration for the provision of goods, this may not always be
the case. Therefore (§ 90):
investigating authorities may use a benchmark other than private prices in the coun-
try of provision under Article 14(d), if it is fi rst established that private prices in that
country are distorted because of the government’s predominant role in providing those
goods.
The AB opened the door to alternative benchmarks, but it did not explain at all
what benchmarks WTO members are allowed to use.
200 Domestic Instruments

2.2.1.3 The specificity requirement

A subsidy is specific in the sense of the SCM Agreement, if the granting authority
or the legislation pursuant to which it operates, explicitly limits access to a sub-
sidy to certain enterprises or industries. Apart from this type of de jure specificity,
a subsidy may also be considered specific if factors, such as the use of the subsidy
by a limited number of certain enterprises, the predominant use by certain enter-
prises, or the granting of disproportionately large amounts of subsidy to certain
enterprises give reasons to believe that the subsidy is in fact specific. The SCM
Agreement explicitly provides that a subsidy shall not be considered specific, if
access to a subsidy has been limited by objective criteria or conditions defined as:
criteria or conditions which are neutral, which do not favour certain enterprises over
others, and which are economic in nature and horizontal in application, such as number
of employees or size of enterprise. (Art 2.1(b) and footnote 2 of the SCM)
These criteria must be clearly spelled out in an official document in order to
facilitate verification. Eligibility under these criteria should be automatic and
the criteria should be strictly adhered to. De facto specificity is also covered.¹⁵
In US—Softwood Lumber IV, the panel dismissed the argument that specifi-
city requires investigating the intent of the WTO member providing the subsidy
(§ 7.116). This should be a legitimate approach with respect to de jure specificity.
Yet, absent such an inquiry, pronouncing on specificity might prove a quixotic
test for WTO adjudicating bodies, in case of de facto specificity. Otherwise, one
might be punishing non-specific interventions which accidentally might yield
specific beneficiaries. Such was presumably not the intent of the framers of the
SCM Agreement.

2.2.2 The classification of subsidies

Unless the three elements discussed above (cost to government, benefit to recipi-
ent, specificity) are cumulatively present, a subsidy does not exist. Subsidies were
originally classified under three headings:
(1) prohibited subsidies (Art 3 of the SCM);
(2) actionable subsidies (Art 5 of the SCM); and
(3) non-actionable subsidies (Art 8 of the SCM).
The first and the third category are defined in the SCM Agreement: a subsidy
is prohibited if it is conditioned upon export performance or upon local content
requirements. A subsidy is non-actionable if it respects the conditions set out in
Art 8 of the SCM (as to its amount) and is granted for purposes of environmen-
tal protection, to make up for regional inequalities within a WTO member, or
in order to promote research and development (R&D). Non-actionable subsidies
were included in the SCM on a provisional basis, subject to review (Art 31 of the

¹⁵ SCM, Art 2.1(c).

 The Coverage 201

SCM). In the absence of an agreement between the WTO membership to extend

the transitional period, the safe harbour for non-actionable subsidies in the SCM
no longer exists.¹⁶

2.2.3 Counteracting subsidies

The SCM offers two possibilities to WTO members wishing to counteract
subsidies:¹⁷
(1) a multilateral avenue, whereby a WTO member may request the establish-
ment of a panel to adjudicate whether a particular subsidy program is WTO
consistent; and
(2) an unilateral avenue, whereby a WTO member which has complied with
the relevant conditions set out in the SCM imposes countervailing duties
(CVDs) against subsidized imports.¹⁸
While the multilateral avenue might lead to an eradication of the subsidy, CVDs
can only neutralize competitive effects of the subsidy in the domestic market; they
cannot address or rather, counter the competitive effect on the country’s exports
to a foreign country’s market. The question arises whether the two forms of relief
(that is, countermeasures and CVDs) can be used simultaneously. Footnote 35 to
the SCM Agreement answers this question in the negative. It reads in pertinent
part:
however, with regard to the effects of a particular subsidy in the domestic market of the
importing Member, only one form of relief (either a countervailing duty, if the require-
ments of Part V are met, or a countermeasure under Articles 4 or 7) shall be available.
Accordingly, this provision requires WTO members to choose one of the two
forms of relief only with respect to injury in their import market; that is, WTO
members can legitimately attack a subsidy by imposing CVDs to address injury
in their domestic market (unilateral avenue), and pursue the multilateral avenue
to address injury suffered in their export markets.

2.2.4 Remedies¹⁹
As noted in Chapter 2, export subsidies must be withdrawn, or the country grant-
ing them will be facing countermeasures up to the level of the subsidy paid. It is a

¹⁶ Howse (2006) has argued in favour of their re-introduction.

¹⁷ Counteracting subsidies is of course a matter of discretion for aff ected WTO members. Recall
our discussion in Chapter 2 about the welfare implications of subsidies. WTO members, especially
those with no competing industry, might not wish to counteract subsidized imports at all.
¹⁸ The conditions under which CVDs can be lawfully imposed are explained in Chapter 4.
¹⁹ In this sub-section, we do not discuss CVDs. They are discussed in Chapter 4. We further
do not discuss remedies in case a prohibited subsidy has been granted, since this issue has been
discussed in Chapter 2. The discussion here is confined to remedies in case an actionable subsidy
has been granted.
202 Domestic Instruments

different case with actionable subsidies: pursuant to Art 5 of the SCM, a WTO
member challenging an actionable subsidy must demonstrate that the subsidy has
adverse effects: by either (a) causing injury to its domestic industry (b) nullification
or impairment of benefits accruing to it, or (c) causing serious prejudice to its inter-
ests. Serious prejudice is explained in Art 6.3 of the SCM and can take one of the
following forms: price undercutting (the price of subsidized commodity is lower
than that of the like product of another WTO member in the same market); price
suppression (prices do not go up, as they should, absent subsidy); price depression
(prices fall); displacement or impediment of sales/exports resulting from subsidized
sales. The subsidizing member will be requested to remove the adverse effects or
withdraw the subsidy; if this has not proved to be the case, then the injured WTO
member could be authorized to take countermeasures commensurate with the degree
and the nature of the adverse effects, and not the level of subsidy paid.
It follows that, with respect to subsidies, as is the case when CVDs, AD duties,
or safeguards are being imposed, the legal texts privilege an injury to compet-
itors—as opposed to an injury to competition—standard. Whereas under the
former, injury will be established any time there are negative welfare implications
for a sub-set of the domestic economy (producers’ welfare), the economy-wide
implications will be evaluated under the latter.

2.2.5 Farm subsidies

2.2.5.1 The regulation of farm trade in the GATT era
During the GATT years, the Contracting Parties to the GATT could not agree
on a substantial reduction in protection for agricultural products. At the end of
the GATT years (1994), protection of agricultural products was considered to be
a ‘tariff peak’, an observation that was confirmed at the end of the tariffication
(see infra) process that occurred during the Uruguay round that was secured for
most other categories of products.
Due originally to US pressure, some quantitative restrictions on agricultural
imports were permitted by virtue of the limited exception set out in Art XI.2 of
the GATT. GATT, Art XI.1 prohibits of course, QRs. GATT, Art XI.2 reflects a
limited exception (for farm products) to this rule.
Of relevance also, is Art XVI.3 of the GATT which regulates export subsidies
for primary (farm) products. Although no clear cut prohibition of such schemes
is spelled out, this provision aims at discouraging the subsidization of primary
goods, and allows subsidization only to the extent that the subsidizing member
does not end up with more than an equitable share of the world market for the
subsidized good at hand.
Beyond these two provisions, there is no other GATT provision that deals
specifically with the trade in agricultural goods. It is of course true that the GATT
applied to all products. It is also true that when the GATT Contracting Parties
 The Coverage 203

wanted to exclude specific products from the GATT disciplines, they signed a
product-specific agreement. Such was for example the nature of the MFA which
excluded trade in textile products from the scope of GATT. Agricultural prod-
ucts had a more ‘limited’ exception, since GATT Contracting Parties did not
conclude a product-specific agreement, as they did for trade in textiles; through
Arts XI.2 and XVI.3 of the GATT, the Contracting Parties did, however, opt for
a product-specific regime of limited scope (as described above). Concessions on
farm goods could, of course, be exchanged during the GATT years during the
various trade rounds (ie Art XXVIII bis of the GATT), through the procedures
enshrined in Art II of the GATT. This was, however, at least with respect to some
members, hardly the case.
The attitude of some key GATT Contracting Parties was to erect additional
barriers to the trade in agricultural products.
As the Executive Secretary of the GATT once observed, though Art XI was:
largely tailor-made to United States requirements [ . . . ] the tailors cut the cloth too
fine.²⁰
What was initially thought of as an instrument to address certain market
disequilibria, changed faces over the years and turned, through a combination of
factors, into a green light for agricultural protectionism.
The two major GATT players capitalized on the unique treatment accorded to
agricultural products in different ways. The United States requested and obtained
a waiver in 1955 of unlimited duration but which was subject to annual review.
Through this waiver, the United States was free to heavily subsidize its farm pro-
duction. As a result of the waiver, US farm policy remained unchallenged dur-
ing the GATT years. Following the US example, some other GATT members
also managed to keep their farm policies immune from legal challenges, either
through waivers (Belgium and Luxembourg), or through special clauses in their
protocol of accession (Switzerland). Besides quantitative restrictions and high
tariffs on imports, public policy in most developed countries included price and
supply management tools, seen as appropriate and necessary to maintain the
viability of rural and agrarian communities and ensure adequate food supply.
For some agricultural products (eg rice in Japan) market access for imports was
effectively non-existent.
The European Community extensively used a variable levies system, whereby a
farm good, when imported into the EC market, would pay not a fixed, but a variable
levy, which would ensure that its price becomes equal to that of the domestic (EC)
goods. As a result, exports rarely, if at all, penetrated into the EC market and farm-
ers for years were completely insulated from the competitive pressures of the world
market. Variable import levies were a key feature of the EC Common Agricultural

²⁰ Eric Wyndham-White quoted in Dam (1970, p 260).

204 Domestic Instruments

Policy (CAP) and were present in many relevant farm product markets. The con-
sistency of this scheme with the GATT was, at the very least, questionable.²¹
Josling et al (1996, pp 71ff ) have argued that the political importance of
European integration and the central role that the CAP played in that process
made the United States (which was also covered by the 1955 waiver), at least
initially, reluctant to request the establishment of a GATT panel to challenge
the legality of the CAP. Such challenges, in the authors’ view, would have jeop-
ardized the goal of European integration itself. The unwillingness of the United
States to destabilize the European integration process thus explains its decision
not to challenge the EC policy in this respect.
When its perception towards the CAP changed, the United States realized that
the GATT dispute settlement procedures were of limited help: what could not be
achieved through negotiations could not be achieved through adjudication either,
as the positive consensus rule of dispute settlement would have proven to be a for-
midable obstacle. Although Hudec (1993a) notes that consensus in the adoption of
panel reports over the GATT years was less of a problem than one might suspect,
and this is definitely true for most cases, most panel reports involving farm-related
issues remain to this date unadopted (and hence of limited legal value).
The protectionist policies of both the European Community and the United
States entangled these GATT super-powers in export subsidy battles in the 1980s
and early 1990s, partly as consequence of thinly veiled attempts to dump surplus
agricultural production onto the world market that had resulted from domes-
tic price and supply management policies. These tensions strained the existing
GATT rules and created a new impetus for liberalization through multilateral
disciplines. But achieving liberalization in this area would clearly require address-
ing the underlying domestic policy regimes, quantifying or estimating their trade
distorting impacts, and instigating their reform—not an easy or simple matter.
There was no political will to do that.
Absence of willingness to reform the domestic market (very much a reality,
especially as far as the EC was concerned) emerged as the key reason for lack of
progress in terms of liberalization of agricultural trade during the GATT years.
Even those practices that could potentially be eliminated through recourse to
dispute settlement, continued to be tolerated since it was perceived that legal
challenges would not bear any results.
In the Uruguay round, a coalition of exporters of agricultural products coun-
tries—the so-called CAIRNS Group²²—provided the necessary stimulus to

²¹ GATT discussions reveal that variable duties that exceeded the level of bound duties would
be deemed illegal, see GATT Doc AG/M/3, p 63 (1984). Anyway, this is a non-issue now following
the advent of the AG Agreement, concluded during the Uruguay round.
²² Borrowed the name of an Australian city where the original founders first met. The Group
was originally composed of Argentina, Australia, Brazil, Canada, Chile, Colombia, Fiji, Hungary,
Indonesia, Malaysia, New Zealand, Philippines, Thailand, and Uruguay and later when Hungary
left the group, Bolivia, Costa Rica, Guatemala, Paraguay and South Africa joined.
 The Coverage 205

push for a worldwide liberalization of agricultural trade. Although it is probably

an overstatement to credit the CAIRNS Group with the successful completion
of the negotiations on agricultural trade, it did nonetheless play a decisive role in
getting the item on the agenda, in making farm liberalization one of the priorities
of the Uruguay round and eventually, making it round-breaker so to speak (in
the sense that it became common knowledge that absent some sort of an agree-
ment on farm trade, the whole round risked being left in limbo).²³
The European Community, on the other hand, slowly moved towards a ration-
alization of its CAP. Although as Messerlin (2001) describes material change
started only after the conclusion of the Uruguay round, the seeds were neverthe-
less to some extent sown during the Uruguay round. By the end of the round, EC
enlargement of Austria, Finland, and Sweden had been completed and enlarge-
ment towards the ex-communist European countries (many of them import-
ant farming countries) was well on track. Messerlin (2001, pp 95ff ) succinctly
discusses the evolution of the EC stance in the field of CAP. The Berlin deci-
sions (1999 Berlin Council) went some way towards reversing the tide inside the
European Community. Importantly, through these decisions:
(1) the risk for over-compensation of EC farmers was reduced;
(2) the CAP budget was stabilized in absolute monetary amounts;
(3) roll-back of existing aid was not excluded; and
(4) a reduction of all direct subsidies by 20% at most was called.
Following these decisions,²⁴ Agenda 2000 was decided and it included an extra
20% cut of the guaranteed prices for cereals, 30% for beef and 15% for dairy
products (without touching upon ‘thorny’ issues such as sugar and olive oil). In
Messerlin’s view:
More important, the whole debate revealed increasingly stronger forces in favour of fur-
ther reforms for domestic reasons—the need to reallocate scarce funds to other purposes
than agriculture, the necessity of preparing for the accession of Central European coun-
tries (some of them having substantial capacities in agriculture in the long run) and last
but not least consumers’ interests in getting cheaper products (. . .). In particular, the
debate began to clarify the merits of “renationalizing” EC farm subsidies (co-financing,
in EC parlance) in comparison with “modulating” (reducing) them. (original emphasis)
The gradual reform of the CAP was the key towards breaking the deadlock
during the Uruguay round.²⁵

²³ On the negotiation of the agreement on agriculture, see Olsen (2005). For an excellent survey
of the WTO AG, see McMahon (2006).
²⁴ Similarly to Messerlin’s reading of the situation, Swinbank and Tanner (1996) argue that
the eventual accession of the 10 new EC member states (and most notably, Poland, a large farming
country) and the resulting implications for the EC budget, proved to be a decisive factor arguing in
favour of rationalization of the CAP.
²⁵ One should not forget that the Brussels meeting of 1991 led to a deadlock essentially because
of disagreements between the EC and its partners on the issue of farm trade.
206 Domestic Instruments

Upon conclusion of the WTO Agreement,²⁶ negotiators wanted to make

it clear that the historical agreement reached on agricultural disciplines was
only the start of the liberalization process and that future negotiations would
be needed to further liberalize agricultural trade. AG, Art 20 addresses this
concern explicitly:
Recognizing that the long-term objective of substantial progressive reductions in support
and protection resulting in fundamental reform is an ongoing process, Members agree
that negotiations for continuing the process will be initiated one year before the end of
the implementation period, taking into account:
(a) the experience to that date from implementing the reduction commitments;
(b) the effects of the reduction commitments on world trade in agriculture;
(c) non-trade concerns, special and differential treatment to developing country
Members, and the objective to establish a fair and market-oriented agricultural trad-
ing system, and the other objectives and concerns mentioned in the preamble to this
Agreement; and
(d) what further commitments are necessary to achieve the above mentioned long-term
objectives.²⁷
It is important to understand the AG Agreement as the first genuine agreement
to move towards liberalization of agricultural trade. In principle, there is no com-
pelling intellectual reason arguing in favour of a separate product-specific agree-
ment: one could very well imagine that agricultural trade (as well as trade in
textiles for that matter) should also abide by the disciplines applicable to all other
products. Subjecting farm products to the existing GATT disciplines (eg start-
ing an exchange of concessions; prohibiting QRs, enforcing the disciplines on
subsidies) could very well do the trick. However, practical political concerns and
domestic pressures often impel Members to favour a sector-specific industry and
devise a special agreement, the parameters of which might well be put into ques-
tion in the future.²⁸

2.2.5.2 The WTO Agreement in a nutshell

The negotiators adopted a rather convoluted liberalization model, which could be
presented as follows:
(1) an obligation imposed on all WTO Members to ‘tariff y’, that is, con-
vert to tariffs, existing non-tariff protection. The so-called tariffication process
aims at converting into tariffs all forms of protection;²⁹ that is, quantitative
restrictions and measures having an equivalent effect as well as other similar

²⁶ For a brief, yet concise, description, see O’Connor (2005).

²⁷ A similar point of view is reflected in the panel report on Canada—Dairy (§§ 7.25 and 7.26).
²⁸ Although no such signs are visible at this juncture.
²⁹ The Illustrative list of measures that must be converted to tariff s includes variable import lev-
ies (a predominantly EC practice in the pre-Uruguay Round years), and minimum import prices
(see Art 4.2 of the AG).
 The Coverage 207

forms of protection (domestic subsidies included). Tariffication becomes thus the

common language among negotiators to use in future rounds aiming at liberal-
ization of agricultural trade. Absent such common language, there was a percep-
tion among some WTO members that reciprocal negotiations could not easily
come to fruition. Additionally, WTO members committed not to reinstate or
introduce other non-tariff forms of trade protection, with the exception of safe-
guards provisions. Following tariffication, WTO members agreed to engage
in negotiations aiming to reduce tariffs progressively. The percentage by which
duties would be reduced was pre-defined.³⁰ Domestic support (domestic subsid-
ies) will be gradually reduced at an aggregate level (and this is where the aggregate
measurement of support (AMS) becomes relevant). The AG required from WTO
members to report their total AMS³¹ in a base period (1986–8), bind it and
then reduce it according to an agreed schedule.³² WTO members enjoy the
flexibility of decreasing the protection in one agricultural commodity, while
increasing domestic subsidies for another commodity, and still be in compliance
with their respective commitment to the continuous reduction of their overall vol-
ume of domestic subsidies. Domestic subsidies were classified into green box, blue
box, and amber box subsidies, each box imposing particular conditions for tolerat-
ing subsidies. Each WTO member must notify its programmes coming under each
one of the three boxes. Although the agreement contains fairly detailed criteria for
notification, disagreements might arise across members as to the designation of a
particular national programme.³³ Green box (Annex 2) includes programmes
that are not at all or are minimally distorting, such as agricultural research; blue
box includes production limited support for either crop or livestock production
(Art 6.5 of the AG). Under the de minimis provisions, WTO members are allowed
to use subsidies up to 5% (10% for developing countries) of the total value of
domestic agricultural production. Because the level of tariffs post-tariffication
was sometimes quite high, WTO members offered a de minimis market access, that
is, they introduced a quota with low tariffs. Amber box includes the products that
do not come under the exempted categories (blue box, green box, and de minimis
exceptions). It is capped and reduced in the subsequent trade liberalizing rounds;

³⁰ Panagariya (2005) observes a great discrepancy between the MFN and the applied rates.
³¹ The AMS was based, but does not correspond to entirely, the OECD Producer Subsidy
Equivalent (PSE), which measures the total increment in revenues earned by domestic producers
over the world price, through any measure affecting the additional earnings.
³² Hoekman and Messerlin (2006) estimate that the European Community accounted for
$ 4.95 billion of export subsidies, Switzerland, for $292 million, and the United States, for $147
million. The same figures for 1999 were $2.6 billion for the European Community and $80
million for the United States. The numbers are substantially smaller for domestic subsidies (see
WTO DocTN/AG/S/1). Moreover, Hart and Beghin (2006) demonstrate that the European
Community for example has moved a lot of its programmes from the amber box (which includes
subsidies that should be eliminated) to the green box (which includes subsidies that are tolerated).
³³ In US—Upland Cotton, for example, Brazil successfully challenged the US designation of
production flexibility contract (PFC) as fully decoupled, and therefore, green box consistent.
208 Domestic Instruments

(2) with respect to export subsidies,³⁴ WTO members undertake product (or
group of products) specific commitments either in the form of budgetary outlay
reduction commitments, or in the form of export quantity reduction commit-
ments; and finally,
(3) WTO members undertake some other, specific obligations (such as due
restraint, export prohibition, transparency requirements and obligations vis-à-vis
the net food importing developing countries).

2.2.5.3 AG over SCM

Subsidies provided to agricultural products are governed by the WTO AG.³⁵ The
AG provisions constitute lex specialis to the SCM provisions and, as a result, to
the extent of difference between the two sets of provisions, they prevail over those
in the SCM. AG, Art 21 provides the response to this question:
1. The provisions of GATT 1994 and of other Multilateral Trade Agreements in Annex
1A to the WTO Agreement shall apply subject to the provisions of this Agreement.
This provision gives precedence to the AG Agreement. The panel in its report on
Canada –Dairy (§§ 7.20ff ), explained that:
(1) the observance of the AG Agreement disciplines does not create a ‘safe har-
bour’ for the measures in question under the SCM Agreement;
(2) a subsidy which is consistent with disciplines of the AG Agreement, but
does violate the SCM Agreement and GATT 1994 is provisionally (that is, until
the expiration of the ‘peace clause’ pursuant to Art 13 of the AG) exempted from
any legal challenge under the WTO Agreement. That is, until 31 December 2003
(or sometime in 2004 depending on when the marketing year cycle of the spe-
cific agricultural product ends)³⁶ no legal challenge can be directed against such
measures under the SCM Agreement or the GATT 1994 that satisfy the condi-
tions of Art 13 of the AG. Today, such legal challenges are perfectly possible;
(3) a subsidy which violates the AG Agreement could be found to be inconsist-
ent with the SCM Agreement and GATT as well.

³⁴ Export subsidies are of course a trade and not a domestic instrument. Its inclusion here is sim-
ply for the sake of completing the discussion on the disciplining of farm trade.
³⁵ Annex 1 contains a list of tariff lines (products) covered by the disciplines of the AG
Agreement.
³⁶ Although the issue is moot today, it is nevertheless interesting to note that ‘implementation
period’ is defined in Art 1(f) of the AG as the 9-year period beginning in 1995 for the purposes of
Art 13 of the AG (ie peace clause), but the definition of ‘year’ set out in Art 1(i) of the AG depends
on whether a member has specified for a particular product the calendar, financial or marketing year
in its Schedule of Commitments. Consequently, if the marketing year for a commodity runs from,
let us say, June to May, then it is possible that the implementation period began in June 1995 and
ends in May 2004, which would make it a 9-year period within the meaning of the definition of
‘implementation period’.
 The Coverage 209

The confirmation of this last point (point (3)) came with the AB report on
US—Upland Cotton. The AB had to face, inter alia, an appeal by the United
States concerning the panel’s findings on the consistency of the so-called Step 2
payments with the WTO. A Step 2 payment is defined as follows (§ 514):
Under the program, marketing certificates or cash payments (collectively referred to by
the Panel as “user marketing (Step 2) payments”) are issued to eligible domestic users and
exporters of eligible upland cotton when certain market conditions exist such that United
States cotton pricing benchmarks are exceeded. “Eligible upland cotton” is defi ned as
“domestically produced baled upland cotton which bale is opened by an eligible domestic
user . . . or exported by an eligible exporter”. An “eligible domestic user” of upland cotton
is defined under the regulations as:
A person regularly engaged in the business of opening bales of eligible upland cotton
for the purpose of manufacturing such cotton into cotton products in the United
States (domestic user), who has entered into an agreement with CCC to participate
in the upland cotton user marketing certificate program.
Brazil did not contest that the United States was in compliance with its obli-
gations under Art 6.3 of the AG (which imposes a cap on the total spending).
It contested, nevertheless, the consistency of the US payments with the WTO,
arguing that such payments, which amounted to local content subsidies, were
inconsistent with Art 3.1(b) of the SCM. The question hence, before the WTO
adjudicating bodies was whether the AG Agreement provides a safe haven for
subsidies otherwise illegal under the SCM Agreement. The panel had responded
in a negative manner, and the AB confirmed the panel’s findings. The key to the
AB’s conclusion was its understanding of Art 21.1 of the AG. The AB agreed with
the panel that Art 21.1 of the AG applies in three situations (§ 532):
. . . where, for example, the domestic support provisions of the Agreement on Agriculture
would prevail in the event that an explicit carve-out or exemption from the disciplines in
Article 3.1(b) of the SCM Agreement existed in the text of the Agreement on Agriculture.
Another situation would be where it would be impossible for a Member to comply with its
domestic support obligations under the Agreement on Agriculture and the Article 3.1(b)
prohibition simultaneously. Another situation might be where there is an explicit author-
ization in the text of the Agreement on Agriculture that would authorize a measure that, in
the absence of such an express authorization, would be prohibited by Article 3.1(b) of the
SCM Agreement. (original emphasis)
The AB held the view that none of these three situations was the case in the trans-
action before it, and then concluded that (§ 550):
In providing such domestic support, however, WTO Members must be mindful of
their other WTO obligations, including the prohibition in Article 3.1(b) of the SCM
Agreement on the provision of subsidies that are contingent on the use of domestic over
imported goods. (original emphasis)
210 Domestic Instruments

If, hence, a WTO member can simultaneously comply with both the AG and the
SCM Agreements, it should do so. If not, the AG provisions take precedence.³⁷
As a result, it outlawed the US Step 2 payments for being illegal under the SCM
Agreement, although in full compliance with the provisions of the AG Agreement
(§ 552).
The peace clause has now expired but it seems that its re-negotiation during the
Doha round should be a priority. Absent a new peace clause, trading nations risk
seeing the outcome of their negotiations on farm products challenged under the
SCM rules. That of course would be odd, since the very existence of the nego-
tiations and the existence of the Agreement on Agriculture are evidence enough
that the intent of negotiators was to shield their negotiated farm policies from
challenges under the SCM. Paradoxically, the conclusion of the original peace
clause is the cause for this mess: absent such a clause, a reasonable panel (in case
a dispute arose) would have probably interpreted the Agreement on Agriculture
in a self-contained manner. Now that a peace clause has been concluded and
has expired, a reasonable panel might infer that its non-renewal evidences the
willingness of negotiators to subsume their farm negotiations to the SCM dis-
ciplines as well. The report on US—Upland Cotton certainly goes in this dir-
ection. To avoid trouble in the future, negotiators will be well-advised to draft
a new peace clause that carefully delineates the carve-out of the Agreement on
Agriculture, and which will remain in place until farm products have been fully
integrated in the GATT disciplines.

2.3 Government procurement

2.3.1 What is government procurement?
The GPA covers cases where governments buy but not with the intention to re-sell.
As such, its obligations are to be distinguished for example, from the obligations
imposed on governments when they have recourse to state trading (Art XVII
of the GATT). The signatories to the GPA commit that certain (governmental,
but also non-governmental) entities, which they have included in their lists of
concessions, will purchase on a non-discriminatory basis from suppliers originat-
ing in other signatories’ sovereignties.³⁸

³⁷ For example, export farm subsidies cannot be challenged now for being inconsistent with Art
3 of the SCM, the passage of the peace clause notwithstanding: the AG explicitly authorizes the use
of export subsidies, provided that the imposed cap has been respected.
³⁸ Although the importance of the government procurement market tends to be over-stated
sometimes, it is not negligible either, as various contributions in Hoekman and Mavroidis (1997)
demonstrate. See, on this score, Francois et al (1997).
 The Coverage 211

2.3.2 The basic discipline: non-discrimination

Why was GPA necessary to introduce non-discrimination in government pur-
chasing? Essentially, because WTO members, by virtue of Art III.8 of the GATT,
are not bound by NT with respect to government purchasing. By the same token,
Art XIII of the GATT excludes national treatment from government procure-
ment in the services sector. Signatories of the GPA, consequently, have brought,
with respect to their list of covered entities, government procurement back within
the four corners of NT.³⁹
The issue has been raised whether WTO members have to respect MFN with
respect to government purchasing. That is, the question is whether, in the absence
of a paragraph in Art I of the GATT equivalent to Art III.8 of the GATT, WTO
members must observe MFN (although not national treatment) with respect to
public purchasing. Assuming of course that an affirmative response to this ques-
tion were warranted, the whole purpose of the GPA could be put into question: any
time a WTO member granted an advantage to a GPA signatory, it would have to
extend it to all WTO members, irrespective of whether they are GPA signatories
or not. The panel report on EC—Commercial Vessels held that this should not
be the case. Looking into the negotiating history of the relevant legal instru-
ments, it concluded that, with respect to government procurement, GATT con-
tracting parties aimed to introduce a caveat from both NT as well as MFN (§§
7.85–7.87).
The GPA, in a nutshell, re-introduces the national treatment obligation only
for the list of entities (and for purchases above an agreed monetary threshold)
committed by each participant.

2.3.3 The past and present of GPA membership

Participation in the GPA is open only to WTO members (Art XXIV.2 of the
GPA). The modalities of the accession procedures are reflected in a WTO docu-
ment (WTO/GPA/1, Annex 2). In short, a working party will be established
where all signatories will participate and discuss the terms of accession for the
acceding country. At the end of negotiations, their report will be submitted to the
GPA committee⁴⁰ which will decide on consensus.
The WTO GPA is not the first agreement in this area; it is a successor agree-
ment to the Tokyo round GPA.⁴¹ The following WTO members are signator-
ies to the GPA: Canada; the European Community (25 members); Hong Kong,
China; Iceland; Israel; Japan; Korea; Liechtenstein; Netherlands with respect to
Aruba; Norway; Singapore; Switzerland; the United States. As is obvious from

³⁹ Arrowsmith (2005) has the most comprehensive account of the GPA, its historic origins and
the issues that nowadays occupy the interest of the membership. See also Clerc (2000).
⁴⁰ Each signatory has a delegate participating in the GPA committee. An indicative time-frame
for accession negotiations has been agreed (WTO Doc GPA/W/109/Rev.2). The GPA committee
has also agreed on a check-list of issues for the provision of information by the applicant govern-
ments (WTO Doc GPA/35).
⁴¹ On the history of GPA, see Blank and Marceau (1997).
212 Domestic Instruments

the list of signatories, the WTO GPA continues the sad pattern established in the
Tokyo round GPA where no developing country had joined its ranks. To encour-
age participation by developing countries in the WTO GPA, Art V.3 of the GPA
reflects the following:
With a view to ensuring that developing countries are able to adhere to this Agreement
on terms consistent with their development, financial and trade needs, the objectives
listed in paragraph 1 shall be duly taken into account in the course of negotiations with
respect to the procurement of developing countries to be covered by the provisions of
this Agreement. Developed countries, in the preparation of their coverage lists under the
provisions of this Agreement, shall endeavour to include entities procuring products and
services of export interest to developing countries.
GPA, Art V.1 contains a best endeavours clause (‘parties shall duly take into
account’) imposed on parties to GPA to account for the need of developing
countries to safeguard their balance of payments position; to promote the estab-
lishment of domestic industries; to support their industrial units; and to encourage
their development through arrangements among developing countries. GPA, Art
V.3 is essentially a flexibility clause, aimed at relaxing the statutory requirements
and facilitate accession of developing countries to the GPA. So far it has had some
success:⁴² Albania, Bulgaria, Chinese Taipei, Georgia, Jordan, Kyrgyz Republic,
Moldova, Oman, and Panama are in the process of acceding to the GPA.

2.4 The coverage of Art III of the GATT—revisited

So we know that two instruments are not covered by the NT obligation. Does
this mean that all other domestic instruments are covered? GATT, Art III.2
refers to taxes and charges. It does not contain a list of measures covered, not even
one of indicative nature. Does this mean that all internal taxes are covered by Art
III.2 of the GATT? What lends to ambiguity is the imprecise language of this
provision: the term ‘directly or indirectly’ can be interpreted in many different
ways and, depending on the interpretation chosen, the coverage of Art III.2 of
the GATT will be larger or narrower.
Instead of theorizing about possible interpretations, one natural place to start
this discussion is the preparatory work of this provision. The words ‘directly or
indirectly’ were added during the drafting of this provision. The initiative was
that of the UK delegate, who replaced the terms ‘in connexion with’ previously
used.⁴³ At Havana, it was recorded that ‘neither income taxes nor import duties

⁴² It is of course, rather presumptuous to claim that many developing countries are in the pro-
cess of accession because of Art V.3 of the GPA. This provision is probably a contributing factor. On
the other hand, there might be some good economic reasons why developing countries might find
the disciplines imposed by the GPA as not working always to their interest, see on this score the
analysis of Mattoo (1997).
⁴³ See § 13 of the annex to the working party report on Border Tax Adjustments, GATT Doc
L/3464, adopted on 2 December 1970, GATT Doc BISD 18S/97ff.
 The Coverage 213

fall within the scope of Art 18 (of the Havana Charter—Art III of the GATT)
which is concerned solely with internal taxes on goods’.⁴⁴ If at all, prepara-
tory work suggests that the intent of the negotiators was to include only taxes
which hit products. Opening the door to taxes which indirectly hit products
is not unproblematic: a series of tax measures might influence decisions to
invest in a particular sector, some of them not intimately linked to the business
activity chosen. Should all of them come under the purview of Art III of the
GATT? And should states be allowed to penalize imported products that
were subjected to lower indirect taxes in their market of origin? Should they
further be obliged to compensate them in case they had been subjected to
higher taxes?
The working party on Border Tax Adjustments addressed some of these ques-
tions, albeit in a rather unsatisfactory manner since, at the end of the day, GATT
CONTRACTING PARTIES could not agree on this score. The final report
issued is a decision that has been adopted by the GATT CONTRACTING
PARTIES. It should, consequently, by virtue of Art XVI of the WTO Agreement,
guide the WTO judge. In this vein, the WTO judge should note that:
. . . the Working Party concluded that there was convergence of views to the effect that
taxes directly levied on products were eligible for tax adjustment. Examples of such
taxes comprised specific excise duties, sales taxes and cascade taxes and the tax on value
added. It was agreed that the TVA, regardless of its technical construction (fractioned
collection), was equivalent in this respect to a tax levied directly—a retail or sales tax.
Furthermore, the Working Party concluded that there was convergence of views to the
effect that certain taxes that were not directly levied on products were not eligible for
tax adjustment. Examples of such taxes comprised social security charges whether on
employers or employees and payroll taxes.⁴⁵
The term ‘border tax adjustment’ is defined in § 4 of the report:
. . . as any fiscal measures which put into effect, in whole or in part, the destination prin-
ciple (i.e. which enable exported products to be relieved of some or all of the tax charged
in the exporting country in respect of similar domestic products sold to consumers on the
home market and which enable imported products sold to consumers to be charged with
some or all of the tax charged in the importing country in respect of similar domestic
products.)⁴⁶
The GATT Contracting Parties took this exercise, that is, the coverage of Art III.2
of the GATT, quite seriously: the Secretariat was asked to compile a document

⁴⁴ See § 12 of the annex to the working party report on Border Tax Adjustments, op cit.
⁴⁵ See the working party report on Border Tax Adjustments, § 14, op cit.
⁴⁶ The destination principle, as explained above was taken over from bilateral agreements nego-
tiated in the 1930s, such as the agreement of 6 May 1936 between the United States and France, see
§ 10 of the annex to the working party report on Border Tax Adjustments, op cit.
214 Domestic Instruments

where it included cases of border tax adjustments.⁴⁷ Nevertheless, beyond the

noted convergence on some taxes, the GATT CONTRACTING PARTIES did
not go any further.⁴⁸
Ambiguity exists with respect to internal measures of non-fiscal nature, as well.
GATT, Art III.4 covers all internal laws, regulations and requirements aff ecting
the life of imported products into a given market. The term ‘aff ecting’ denotes the
legislative will to cast the net, if in doubt, wide rather than narrow.⁴⁹ This is not
paradoxical. Even policies that are the most remotely linked to trade can affect
trade flows: an oppressive regime, for example, will restrict free speech which in
turn, might restrict imports of newsprint.⁵⁰
The next question is thus, what do, in terms of contractual expression, in light
of the awareness that all policies at least potentially and indirectly can affect trade?
One could imagine writing a complete contract whereby, negotiators will reflect
each and every domestic instrument that they use at a point in time. The amount
of work associated with contracting in this manner should not be underesti-
mated: indeed, not only does one need to enumerate each and every policy which
potentially affects trade, one might need to enter into continuous renegotiation;
policies change, as revealed preferences of societies change over time, and, conse-
quently, there will be an unavoidable need to renegotiate the original contract.
Besides, contracting social policies might be deemed politically undesirable in
some quarters. For all these reasons, and probably for other reasons as well, the
GATT founding fathers opted for an incomplete contract, when designing the
disciplines to be applied on domestic policies:⁵¹ in light of the negotiating costs
involved, they rationally decided to avoid negotiating this issue any further than
imposing NT on all instruments affecting trade.⁵² Pushing, however, the deci-
sion to the adjudicating bodies is not risk-free, as we will see infra. Contractual

⁴⁷ See GATT Doc L/3379 of 6 May 1970. See on this score, the excellent analysis in Démaret
and Stewardson (1994).
⁴⁸ So far, GATT/WTO case-law has not dealt with cases of indirect taxes, such as payroll taxes.
⁴⁹ As we will see in more detail infra, case-law has maintained this approach.
⁵⁰ This is not to suggest that NT can equally meaningfully apply to all domestic policies/
measures. Its application on substantive (as opposed to procedural) antitrust laws is, for example,
problematic: a WTO member which accepts a merger between two domestic companies and, sub-
sequently, rejects a merger between a foreign and a domestic company operating all four in the
same relevant product market has not necessarily violated NT; moving from say six to five or from
five to four companies in a market involves different considerations (inter alia, because the degree
of concentration as measured in Herfindhal Hirschmann Index terms will be different). The WTO
case-law is a far cry from the jurisprudential evolution in the corresponding provisions of the Treaty
establishing the European Community, where the court, over the years, narrowed its scope to avoid
false positives (Type I errors), see Weatherill (2002) on this score.
⁵¹ See the analysis along these lines, in Horn and Mavroidis (2004). For a formal explanation
see Horn et al (2006).
⁵² Horn and Mavroidis (2004) characterize the GATT an obligationally incomplete contract.
Maskin and Tirole (1999) have persuasively argued that probably the manner in which we think
about incomplete contracts is not optimal. They point out that instead of discussing contingencies,
contractual parties could be discussing payoff s. There are doubts, however, as to whether their
model can fit the GATT. With respect to some of the policies (potentially) covered by NT, it is at
 National Treatment, and Negative Integration 215

incompleteness by itself is not necessarily a problem. Governance problems are

posed when incomplete contracts are combined with opportunism.⁵³ The GATT
does not eliminate the potential for such behaviour: WTO members will usu-
ally have private information and a strong incentive to cheat (by pretending, for
example, to be internalizing environmental externalities when acting solely, or
predominantly in the interest of their domestic producer, and imposing thus
costs on their trading partners through beggar thy neighbour policies). The GATT
insurance policy against opportunism is non-discrimination (Art III of the
GATT), an instrument of limited use, as we will see in what follows.
We should finally note, that the locus where the measure will be enforced does
not necessarily prejudge its status: domestic measures enforced at the border
come, according to the Interpretative Note ad Art III of the GATT, under the
purview of Art III of the GATT. Thus, a sales ban which is enforced at the border
(and could be mistaken for trade embargo) will still be considered to be a sales
ban (and, consequently, come under the purview of Art III of the GATT).⁵⁴

3 National Treatment, and Negative Integration

NT means that WTO members can unilaterally define their domestic policies
and, to the extent that there is negative trans-boundary spill-over stemming from
this exercise, it will be internalized through the obligation not to discriminate
across domestic and foreign products: there is, thus, no compulsory adherence to
international standards; there is no pre-defined set of policies to which all WTO
members must subscribe.⁵⁵
Non-discrimination does not exhaust the potential for beggar thy neighbour
policies; depending on how this term is defined, it addresses only a sub-set of all
imaginable cases.⁵⁶

least doubtful that governments would be willing to negotiate specific disciplines and (eventually)
payoffs.
⁵³ See Williamson (2005).
⁵⁴ Now, this raises the question whether it is for the GATT (in a centralized manner) or individ-
ual WTO members to decide whether an instrument is trade or domestic. As we noted in Chapter
2, in light of the incentives associated with this choice, this issue is moot.
⁵⁵ With the notable exception of course of TRIPs, adherence to which is compulsory for all
WTO members. As we will see infra, not even the SPS or the TBT Agreements impose the compul-
sory enactment of common policies.
⁵⁶ Country A, for example, might impose an 80% consumption tax on product X that it does
not produce, and a 0% tax on product Y that it produces. If the products X, Y are not directly com-
petitive, there is no violation of Art III of the GATT, as we will see infra. Nevertheless, nothing in
GATT prohibits A from financing its social policies through the proceeds from taxes imposed on
product X. Such policies can be qualified as beggar thy neighbour; they are not, however, outlawed
through recourse to Art III of the GATT, because the latter is concerned with the fight against pro-
tectionism, alas an elusive notion as we will see in what follows.
216 Domestic Instruments

4 National Treatment: The Legal Discipline

4.1 Fiscal measures

GATT, Art III.2 reads:
The products of the territory of any contracting party imported into the territory of any
other contracting party shall not be subject, directly or indirectly, to internal taxes or
other internal charges of any kind in excess of those applied, directly or indirectly, to like
domestic products.
Moreover, no contracting party shall otherwise apply internal taxes or other internal
charges to imported or domestic products in a manner contrary to the principles set forth
in paragraph 1.
GATT, Art III.1 reads:
The Members recognize that internal taxes and other internal charges, and laws, regula-
tions and requirements affecting the internal sale, offering for sale, purchase, transpor-
tation, distribution or use of products, and internal quantitative regulations requiring
the mixture, processing or use of products in specified amounts or proportions, should
not be applied to imported or domestic products so as to afford protection to domestic
production.
The Interpretative Note ad Art III.2 of the GATT, second sentence, reads:
A tax conforming to the requirements of the first sentence of paragraph 2 would be con-
sidered to be inconsistent with the provisions of the second sentence only in cases where
competition was involved between, on the one hand, the taxed product and, on the other
hand, a directly competitive or substitutable product which was not similarly taxed.
Therefore, a complaining party can choose between two possible routes. One is
to argue that:
(1) the domestic and the foreign products are like; and
(2) the latter is taxed in excess of the former.
The other is to claim that:
(1) the two products are DCS;
(2) the two products are not similarly taxed; and that
(3) the dissimilar taxation operates so as to aff ord protection (SATAP) to domestic
production.
Central to the scope of the national treatment provision, is the adjudicating
bodies’ interpretation of the italicized terms.
 National Treatment: The Legal Discipline 217

4.1.1 DCS products

The AB in its report on Korea—Alcoholic Beverages held that like products are a
sub-set of DCS products: if two products are like, they are, by definition, DCS as
well (§ 118):
‘Like’ products are a subset of directly competitive or substitutable products: all like
products are, by definition, directly competitive or substitutable products, whereas not
all ‘directly competitive or substitutable’ products are ‘like’.
It follows that the case-law regarding the definition of DCS is ipso facto relevant
for the interpretation of the term like products. § 17 of the working party report
on Border Tax Adjustments, quoting from the preparatory work of the GATT,
reflects the rationale for the inclusion of the DCS category:
During the drafting of the Havana Charter, and thus the GATT, it was felt that this might
occur where there was no, or negligible, domestic production of the imported product.
Various examples were quoted; it was for instance suggested that a country which did not
produce coffee could not impose tax on coffee, unless it placed a similar tax on chicory, a
competitive product.⁵⁷

4.1.1.1 Defining DCS

The AB in its report on Japan—Alcoholic Beverages II provided its understanding
of DCS. The dispute arose because of a Japanese taxation scheme, on its face neu-
tral, which subjected predominantly Western drinks, to a heavier taxation than
predominantly Japanese drinks: as a result, sochu (an alcoholic beverage predom-
inantly produced in Japan) was subjected to less burdensome taxation than, inter
alia, whisky (an alcoholic beverage predominantly produced in Europe and the
United States). Europe and the United States protested arguing that the products
at hand were at least DCS, if not like. The panel had already accepted that all of
the products concerned (with the exception of vodka that was deemed to be like
product to sochu) were DCS products. The AB upheld the panel’s findings in
this respect. In its view, (a) physical characteristics, (b) common end-uses, and (c)
tariff classification are appropriate elements to take into account when defining
whether two products are DCS. Importantly, upholding the panel’s findings in

⁵⁷ § 18 of the same report offers the following narrative: ‘At Havana it was also agreed that under
the provisions of Art 18 (of the Havana Charter—Art III of the GATT) regulations and taxes
would be permitted which, while perhaps having the effect of assisting the production of a par-
ticular domestic product (say, butter) are directed as much against domestic production of another
product (say, domestic oleomargarine) of which there was a substantial domestic production as
they are against imports (say, imported oleomargarine). At the Review Session the representative
of Sweden said that his Government continued to interpret the provision in this way and this view
was not challenged.’ So, in this understanding of the NT provision, were a WTO member, substan-
tially producing both butter and oleomargarine, to tax the former at say 10% and the latter at 20%
ad valorem, it would not be violating Art III.2 of the GATT. Th is view, as we will see infra, was
implicitly rejected by the AB in Chile—Alcoholic Beverages.
218 Domestic Instruments

this respect, the AB made it clear that the test to define whether two products are
DCS is in the market-place, in the sense that, it is consumers who will ultimately
decide whether two products are indeed in competition with each other. To this
effect, econometric indicators (in the instant case, cross-price elasticity)⁵⁸ are rele-
vant to define whether two products are indeed in competition with each other.⁵⁹
The European Community had submitted some consumer surveys to this effect,
suggesting that Japanese consumers in the absence of discriminatory taxation
would be prepared to substitute sochu for a host of Western drinks (p 25):
(a) ‘Directly Competitive or Substitutable Products’
In this case, the Panel emphasized the need to look not only at such matters as physical
characteristics, common end-uses, and tariff classifications, but also at the ‘market place.’
This seems appropriate. The GATT 1994 is a commercial agreement, and the WTO is
concerned, after all, with markets. It does not seem inappropriate to look at competition
in the relevant markets as one among a number of means of identifying the broader cat-
egory of products that might be described as ‘directly competitive or substitutable.’
Nor does it seem inappropriate to examine elasticity of substitution as one means of
examining those relevant markets. The Panel did not say that cross-price elasticity of
demand is ‘the decisive criterion’ (footnote omitted) for determining whether products
are directly competitive or substitutable. The Panel stated the following:
In the Panel’s view, the decisive criterion in order to determine whether two
products are directly competitive or substitutable is whether they have common
end-uses, inter alia, as shown by elasticity of substitution.
We agree. And, we find the Panel’s legal analysis of whether the products are ‘directly
competitive or substitutable products’ in paragraphs 6. 28–6.32 of the Panel Report to be
correct. (original emphasis)
In Korea—Alcoholic Beverages, the AB moved in to explain itself as to the rele-
vance of econometric indicators when deciding whether two products are DCS.
In this case, the facts were quasi-identical to those in Japan—Alcoholic Beverages
II: beverages predominantly produced in Korea (soju) were hit by a substantially
lower tax burden than their counterparts which were predominantly produced in
the European Community, Canada, and the United States (vodka, whisky, etc).
The European Community, Canada, and the United States complained, arguing
that the Korean regime was GATT inconsistent. Korea, however, argued that
its system could not be held to be discriminatory since, the products concerned

⁵⁸ With this instrument, we measure the additional demand for product B, in case the price for
product A increases by a fi xed amount. The degree of elasticity can vary: hugh degrees suggest that
two products are perfect substitutes.
⁵⁹ Through the use of this measure, we capture an important aspect of the substitution among
products in response to price changes: what is the change in the demanded quantity for product X,
in case of a price change in product Y (formally, it is derived by dividing the percentage change in
quantity of good X by the percentage change in price for good Y). If the ratio is negative, the prod-
ucts concerned are sometimes said to be complements; if positive, substitutes. Neither the panel nor
the AB put a number on this issue stating, for example.
 National Treatment: The Legal Discipline 219

were not DCS in the first place: the price of (diluted) soju⁶⁰ was a small fraction
of the price of the Western drinks at hand. Consequently, following the ana-
lysis in Japan—Alcoholic Beverages, and the relevance of econometric indicators
in deciding whether two products are DCS, Korea argued that, with respect to
(diluted) soju at least, no claim under Art III.2 of the GATT could be sustained.
Complaining parties attempted to relegate the importance of econometric indi-
cators. The panel essentially upheld the complaining parties’ view. Dissociating
econometric indicators from other elements (such as, consumer preferences, end
uses of the product), the panel held the view that the products were indeed in a
DCS relationship. Only a reading whereby cross-price elasticity would be ele-
vated to the decisive criterion conferring DCS status, would lead the panel to rule
otherwise. Such a reading of Art III.2 of the GATT, however, was in the panel’s
eyes unwarranted.
The AB upheld the panel’s findings without any modification in this respect
(§§ 114 ff and especially 133–4, 135–8):
A. “Directly Competitive or Substitutable Products”
1. Potential Competition
. . . In our view, the word ‘substitutable’ indicates that the requisite relationship may exist
between products that are not, at a given moment, considered by consumers to be substi-
tutes but which are, nonetheless, capable of being substituted for one another. . . .
2. Expectations
As we have said, the object and purpose of Article III is the maintenance of equality of
competitive conditions for imported and domestic products. (footnote omitted) . . .
3. ‘Trade Effects’ Test
. . . the Panel stated that if a particular degree of competition had to be shown in quanti-
tative terms, that would be similar to requiring proof that a tax measure has a particular
impact on trade. It considered such an approach akin to a ‘type of trade effects test.
We do not consider the Panel’s reasoning on this point to be flawed.
4. Nature of Competition
The Panel considered that in analyzing whether products are ‘directly competitive or sub-
stitutable,’ the focus should be on the nature of competition and not on its quantity. . . .
For the reasons set above, we share the Panel’s reluctance to rely unduly on quantitative
analyses of the competitive relationship. [footnote omitted] In our view, an approach that
focused solely on the quantitative overlap of competition would, in essence, make cross-
price elasticity the decisive criterion in determining whether products are ‘directly com-
petitive or substitutable.’ We do not, therefore, consider that the Panel’s use of the term
‘nature of competition’ is questionable.
5. Evidence from the Japanese Market
. . . It seems to us that evidence from other markets may be pertinent to the examination
of the market at issue, particularly when demand on that market has been influenced by
regulatory barriers to trade or to competition. Clearly, not every other market will be

⁶⁰ The price of non-diluted soju was substantially higher.

220 Domestic Instruments
relevant to the market at issue. But if another market displays characteristics similar to
the market at issue, then evidence of consumer demand in that other market may have
some relevance to the market at issue. This, however, can only be determined on a case-
by-case basis, taking account of all relevant facts. (original emphasis)
The attitude of the AB in this report is quite problematic. Dissociating cross-price
elasticity from criteria such as end-uses, or consumer preferences is highly con-
troversial; such criteria are inherent in a cross-price elasticity test. Conversely,
when cross-price elasticity is privileged, the relative price of two products is also
considered. The critical mass of consumers operates under scarcity of monet-
ary resources; purchasing decisions are thus usually based on their availability.
Consumers, when responding to a consumer survey, will naturally take into
account price, for money is a scarce commodity and there is an opportunity cost
associated with any purchasing decision: two products might serve the same end
use in the eyes of the consumer, but one of them can simply be unaffordable to
most of them; a small car and a luxury car fit this picture.⁶¹ Consumers do, of
course, think of properties such as end-uses, physical characteristics, etc when
making their choices in the market. It is, however, scarcity of monetary resources
that is the ultimate arbiter. And anway, all these elements (end-uses, physical
characteristics, and price-related concerns as well) are all part of a cross-price elas-
ticity test. The AB turned a deaf ear to this simple reality of life with its decision.
It is, of course, not necessarily the case that econometric indicators always yield
satisfactory responses to questions such as this one asked before the panel. They
are, however, a more robust method to analyze a DCS relationship than impres-
sionistic out-of-context evaluations; they should provide the starting point of
an analysis which, eventually, might need to be corrected through recourse to
other criteria. The AB did not even attempt to explain why this sequence (econo-
metric indicators, and if necessary, use of other criteria) was inappropriate in the
particular instance. It established an absolute parallelism between a scientific
methodology and guess-work. This is highly regrettable.⁶²
This discussion has not continued after the AB report on Korea—Taxes on
Alcoholic Beverages. It is, hence, now accepted that two products can be DCS irre-
spective whether one has had recourse to econometric indicators or not.

⁶¹ Horn and Mavroidis (2004) explain how, by using non-econometric indicators, one might
end up conferring DCS status to a pair of products that, in the eyes of consumers, are not at all dir-
ectly competitive with each other.
⁶² In the case at hand, the argument was made that demand in Korea was latent because of the
regulatory barriers that impeded access for Western drinks. Hence, evidence from third country
markets was necessary to establish whether soju and a series of Western beverages were indeed
DCS products. This should not, however, have led the panel and the AB to import ‘lock, stock,
and barrel’ the Japanese consumers’ reaction into the Korean market and hold that soju and the
Western drinks at hand were DCS products. As Korea pointed out, the price of shochu was higher
than that of soju, and closer to that of the Western drinks. It could of course be the case that absent
price considerations, Korean consumers might have treated soju and, say, whisky as substitutes.
Price considerations do matter, though, and consumers will base their purchasing decisions on
them.
 National Treatment: The Legal Discipline 221

In a nutshell, two products will be DCS if they are viewed as such by consum-
ers; recourse to econometric indicators is not passage obligé: a DCS relationship
can also be established through recourse to criteria such as physical characteris-
tics, end uses, consumer preferences. These three criteria have been consistently
referred to in case-law.⁶³

4.1.1.2 SATAP
As stated above, for taxation between DCS products to be GATT inconsistent, it
should not operate so as to afford protection to domestic production; this is essen-
tially what the reference to Art III.1 of the GATT amounts to. The Interpretative
Note ad Art III of the GATT further explains that, for this to be the case, the pair
of products (domestic, imported) must not be similarly taxed. Two interpretative
issues arise:
(1) what is the relationship between Arts III.1 and III.2 of the GATT? and
(2) how does the Interpretative Note fit into this relationship?
The response to these two questions will provide us with the answer to the question
what does the SATAP-test amount to? The AB addressed the relationship between
Art III.1 and Art III.2 of the GATT in its report on Japan—Alcoholic Beverages II.
It held that, with respect to like products, taxation in excess should be understood
as an instance of a measure operating so as to afford protection: consequently, a
complainant who has established that taxation on imported products is in excess
of that on domestic like products, does not also have to establish that the measure
at hand operates so as to afford protection (pp 18–19):
Article III:1 informs Article III:2, first sentence, by establishing that if imported products
are taxed in excess of like domestic products, then that tax measure is inconsistent with
Article III. Article III:2, first sentence does not refer specifically to Article III:1. There is
no specific invocation in this first sentence of the general principle in Article III:1 that
admonishes Members of the WTO not to apply measures so as to afford protection. This
omission must have some meaning. We believe the meaning is simply that the presence of
a protective application need not be established separately from the specific requirements
that are included in the first sentence in order to show that a tax measure is inconsistent
with the general principle set out in the first sentence. However, this does not mean that
the general principle of Article III:1 does not apply to this sentence. To the contrary, we
believe the first sentence of Article III:2 is, in effect, an application of this general prin-
ciple. The ordinary meaning of the words of Article III:2, first sentence leads inevitably
to this conclusion. Read in their context and in the light of the overall object and purpose
of the WTO Agreement, the words of the first sentence require an examination of the
conformity of an internal tax measure with Article III by determining, first, whether the

⁶³ Sharing the same tariff classification is not a necessary criterion to decide whether two prod-
ucts are DCS. It has been used, however, in order to decide whether two products are like (as we
will see infra). Since like products are a sub-set of DCS products, it is reasonable to assume that this
criterion is relevant (albeit not necessary) to decide whether two products are DCS.
222 Domestic Instruments
taxed imported and domestic products are ‘like’ and, second, whether the taxes applied
to the imported products are ‘in excess of’ those applied to the like domestic products.
If the imported and domestic products are ‘like products,’ and if the taxes applied to the
imported products are ‘in excess of’ those applied to the like domestic products, then the
measure is inconsistent with Article III:2, first sentence. (original emphasis)
With regard to DCS products, nevertheless, a similar reading was rejected
(p 27):
Unlike that of Article III:2, first sentence, the language of Article III:2, second sentence,
specifically invokes Article III:1. The significance of this distinction lies in the fact that
whereas Article III:1 acts implicitly in addressing the two issues that must be considered
in applying the first sentence, it acts explicitly as an entirely separate issue that must be
addressed along with two other issues that are raised in applying the second sentence.
Giving full meaning to the text and to its context, three separate issues must be addressed
to determine whether an internal tax measure is inconsistent with Article III:2, second
sentence. These three issues are whether:
(1) the imported products and the domestic products are ‘directly competitive or substitut-
able product’ which are in competition with each other;
(2) the directly competitive or substitutable imported and domestic products are ‘not
similarly taxed ’; and
(3) the dissimilar taxation of the directly competitive or substitutable imported domes-
tic products is ‘applied . . . so as to aff ord protection to domestic production.’ (original
emphasis)
Hence, in the AB’s view, Art III.1 is GATT relevant for the whole of Art III.2 of
the GATT:
(1) with respect to like products, taxation in excess of the imported like product
ipso facto amounts to a violation of the SATAP requirement;
(2) whereas in the case of DCS products, establishment of taxation in excess of
the imported product is necessary, but not sufficient condition for finding that
a measure operates so as to afford protection.
The next question is, what else should a successful complainant demonstrate, in
order to discharge its burden of proof that two DCS products (imported, domes-
tic) have been treated in violation of NT? Case-law so far, can be summed up as
follows:
(1) there is a threshold issue: the tax differential must be more than de minimis
(we do not know, however, what exactly constitutes de minimis, but we do
know that mere arithmetic difference satisfies the in excess-requirement, but
not the de minimis requirement);⁶⁴

⁶⁴ Infinitesimal tax differentials will satisfy the in excess—but not the SATAP—criterion. It has
to be then, that tax differentials affect the formation of price (the post-tax price).
 National Treatment: The Legal Discipline 223

(2) if the tax differential is substantial, then it will suffice, in and of itself, to
establish a violation of Art III.2 of the GATT;⁶⁵
(3) if not, an inconsistency with Art III.2 of the GATT can be established only
through recourse to other factors as well (that help to establish the protective
intent of the taxing state).
Quoting from a prior GATT panel report, which had dealt with a quasi-identical
issue, the AB, in its report on Japan—Alcoholic Beverages II clarified this point in
the following terms (pp 29–32):
In the 1987 Japan—Alcohol case, the panel subsumed its discussion of the issue of ‘not
similarly taxed’ within its examination of the separate issue of ‘so as to afford protection’:
. . . whereas under the first sentence of Article III:2 the tax on the imported prod-
uct and the tax on the like domestic product had to be equal in effect, Article III:1
and 2, second sentence, prohibited only the application of internal taxes to imported
or domestic products in a manner ‘so as to afford protection to domestic production.’
The Panel was of the view that also small tax differences could influence the com-
petitive relationship between directly competing distilled liquors, but the existence
of protective taxation could be established only in the light of the particular circum-
stances of each case and there could be a de minimis level below which a tax difference
ceased to have the protective effect prohibited by Article III:2, second sentence.
To detect whether the taxation was protective, the panel in the 1987 case examined a
number of factors that it concluded were ‘sufficient evidence of fiscal distortions of the
competitive relationship between imported distilled liquors and domestic shochu afford-
ing protection to the domestic production of shochu.’ These factors included the con-
siderably lower specific tax rates on shochu than on imported directly competitive or
substitutable products; the imposition of high ad valorem taxes on imported alcoholic
beverages and the absence of ad valorem taxes on shochu; the fact that shochu was almost
exclusively produced in Japan and that the lower taxation of shochu did ‘afford protec-
tion to domestic production’; and the mutual substitutability of these distilled liquors.
The panel in the 1987 case concluded that ‘the application of considerably lower internal
taxes by Japan on shochu than on other directly competitive or substitutable distilled liq-
uors had trade-distorting effects affording protection to domestic production of shochu
contrary to Article III:1 and 2, second sentence.
As in that case, we believe that an examination in any case of whether dissimilar
taxation has been applied so as to afford protection requires a comprehensive and
objective analysis of the structure and application of the measure in question on
domestic as compared to imported products. We believe it is possible to examine
objectively the underlying criteria used in a particular tax measure, its structure,
and its overall application to ascertain whether it is applied in a way that affords
protection to domestic products.

⁶⁵ Once again, WTO adjudicating bodies have not provided an arithmetic benchmark. In
Chile—Alcoholic Beverages, nevertheless, a 20% tax differential across two DCS products was
judged substantial (see the discussion infra).
224 Domestic Instruments

Although it is true that the aim of a measure may not be easily ascertained,
nevertheless its protective application can most often be discerned from the
design, the architecture, and the revealing structure of a measure. The very mag-
nitude of the dissimilar taxation in a particular case may be evidence of such a
protective application, as the Panel rightly concluded in this case. Most often,
there will be other factors to be considered as well. In conducting this inquiry,
panels should give full consideration to all the relevant facts and all the relevant
circumstances in any given case.
In this respect, we note and agree with the panel’s acknowledgment in the 1987 Japan—
Alcohol Report:
. . . that Article III:2 does not prescribe the use of any specific method or system of
taxation. . . . there could be objective reasons proper to the tax in question which
could justify or necessitate differences in the system of taxation for imported and
for domestic products. The Panel found that it could also be compatible with Article
III:2 to allow two different methods of calculation of price for tax purposes. Since
Article III:2 prohibited only discriminatory or protective tax burdens on imported
products, what mattered was, in the view of the Panel, whether the application of
the different taxation methods actually had a discriminatory or protective effect
against imported products.
We have reviewed the Panel’s reasoning in this case as well as its conclusions on the issue
of ‘so as to afford protection’ in paragraphs 6.33–6.35 of the Panel Report. We find cause
for thorough examination. The Panel began in paragraph 6.33 by describing its approach
as follows:
. . . if directly competitive or substitutable products are not ‘similarly taxed,’ and
if it were found that the tax favours domestic products, then protection would be
afforded to such products, and Article III:2, second sentence, is violated.
This statement of the reasoning required under Article III:2, second sentence is correct.
However, the Panel went on to note:
. . . for it to conclude that dissimilar taxation afforded protection, it would be
sufficient for it to find that the dissimilarity in taxation is not de minimis. . . . the
Panel took the view that ‘similarly taxed’ is the appropriate benchmark in order
to determine whether a violation of Article III:2, second sentence, has occurred as
opposed to ‘in excess of’ that constitutes the appropriate benchmark to determine
whether a violation of Article III:2, first sentence, has occurred.
In paragraph 6.34, the Panel added:
(i) The benchmark in Article III:2, second sentence, is whether internal taxes operate
as to afford protection to domestic production, a term which has been further inter-
preted in the Interpretative Note ad Article III:2, paragraph 2, to mean dissimilar
taxation of domestic and foreign directly competitive or substitutable products.
And, furthermore, in its conclusions, in paragraph 7.1(ii), the Panel concluded that:
(ii) Shochu, whisky, brandy, rum, gin, genever, and liqueurs are directly competitive or
substitutable products and Japan, by not taxing them similarly, is in violation of its
 National Treatment: The Legal Discipline 225
obligation under Article III:2, second sentence, of the General Agreement on Tariffs
and Trade 1994.
Thus, having stated the correct legal approach to apply with respect to Article III:2,
second sentence, the Panel then equated dissimilar taxation above a de minimis level with
the separate and distinct requirement of demonstrating that the tax measure ‘affords pro-
tection to domestic production.’ As previously stated, a finding that ‘directly competi-
tive or substitutable products’ are ‘not similarly taxed’ is necessary to find a violation of
Article III:2, second sentence. Yet this is not enough. The dissimilar taxation must be
more than de minimis. It may be so much more that it will be clear from that very differ-
ential that the dissimilar taxation was applied ‘so as to afford protection.’ In some cases,
that may be enough to show a violation. In this case, the Panel concluded that it was
enough. Yet in other cases, there may be other factors that will be just as relevant or more
relevant to demonstrating that the dissimilar taxation at issue was applied ‘so as to afford
protection.’ In any case, the three issues that must be addressed in determining whether
there is such a violation must be addressed clearly and separately in each case and on a
case-by-case basis. And, in every case, a careful, objective analysis, must be done of each
and all relevant facts and all the relevant circumstances in order to determine ‘the exist-
ence of protective taxation.’ Although the Panel blurred its legal reasoning in this respect,
nevertheless we conclude that it reasoned correctly that in this case, the Liquor Tax Law
is not in compliance with Article III:2. As the Panel did, we note that:
. . . the combination of customs duties and internal taxation in Japan has the fol-
lowing impact: on the one hand, it makes it difficult for foreign-produced shochu
to penetrate the Japanese market and, on the other, it does not guarantee equality of
competitive conditions between shochu and the rest of <white= and <brown= spir-
its. Thus, through a combination of high import duties and differentiated internal
taxes, Japan manages to ‘isolate’ domestically produced shochu from foreign com-
petition, be it foreign produced shochu or any other of the mentioned white and
brown spirits. (original emphasis)⁶⁶
In Korea —Taxes on Alcoholic Beverages (WTO Doc WT/DS 75. 84/AB/R of 18
January 1999) the AB held the same view as to the evidence required to demon-
strate protective intent⁶⁷ (§ 150):
Although it is true that the aim of a measure may not be easily ascertained, nevertheless
its protective application can most often be discerned from the design, the architecture,
and the revealing structure of a measure. The very magnitude of the dissimilar taxation in

⁶⁶ The AB repeated this case-law almost verbatim in its Korea—Alcoholic Beverages jurispru-
dence (§ 150).
⁶⁷ The AB privileged, however, the use of the term ‘protective application’ over ‘protective intent’.
Were one to take the AB seriously on this score, the inescapable conclusion would be that the AB
is of the view that unintended or ancillary effects, to the extent that they weigh more on foreign
(imported) products, lead to a demonstration of inconsistency with Art III.2 of the GATT. If true,
this is indeed problematic since it can lead to false positives (such as punishing interventions that
genuinely intend to protect public health or environment, and not maximize the domestic produc-
ers’ welfare). The AB tried to address such concerns by interpreting the term ‘like’ products in a very
innovative manner in EC—Asbestos, and by relaxing the evidentiary standards in Art XX of the
GATT. By doing that, however, it created additional distortions, as we will see infra.
226 Domestic Instruments
a particular case may be evidence of such protective application . . . Most often, there will
be other factors to be considered as well.
The shortcomings of this test became apparent through the litigation on Chile—
Alcoholic Beverages, where the AB was asked to pronounce on the consistency of
the Chilean tax system for alcoholic beverages with the GATT; in the view of
some exporters of ‘Western’ alcoholic drinks, the Chilean tax regime favoured
predominantly locally produced alcoholic beverages (some categories of pisco).
The scheme distinguished between two categories of alcoholic beverages, using
alcoholic content as the distinguishing criterion: below 35° and above 39°.⁶⁸
The tax differential across the two categories was more than de minimis, in the
AB’s view (27% and 47%, respectively). The complaining party (European
Community) had argued that many Western products of slightly more than 39°
were DCS products to Chilean products of less than 35°, and, that the tax differ-
ential operated SATAP. Chile responded that in the 39° and above tax category,
the majority of the products hit by high taxation were domestic, and that no pro-
tection could thus result from such a taxation scheme (§ 58 of the report). The
AB condemned the Chilean fiscal scheme (§§ 44–55 of the report), holding the
view that the tax differential was in and of itself a sufficient reason to establish a
violation of Art III.2 of the GATT. The AB agreed that as a matter of fact, most of
the alcoholic drinks hit by the higher taxation were of Chilean origin. However,
it dismissed the relevance of this observation for the interpretation of the SATAP
requirement in the following terms (§ 67):
This fact does not, however, by itself outweigh the other relevant factors, which tend to
reveal the protective application of the New Chilean System. The relative proportion of
domestic versus imported products within a particular fiscal category is not, in and of
itself, decisive of the appropriate characterization of the total impact of the New Chilean
system under Article III:2, second sentence, of the GATT 1994. This provision, as noted
earlier, provides for equality of competitive conditions of all directly competitive or sub-
stitutable imported products, in relation to domestic products, and not simply, as Chile

⁶⁸ Note that all three cases discussed so far are cases of alleged de facto discrimination. It is often
alleged that it is case-law that extended the coverage of Art III of the GATT to cases of de facto
discrimination as well. The second sentence of Art III.2 of the GATT seems to go into the direction
of de facto discrimination: it prohibits any taxation which might operate so as to afford protection,
that is, de facto discriminatory as well. The first sentence, by contrast, sanctions a tax differential
predicated on the origin of the good. On the other hand, the second sentence, and especially its
Interpretative Note, extend the coverage of NT in the following manner: assume strawberries and
bananas are DCS in a given market, and that A produces 1 ton of the former and 99 of the latter. B
produces 99 tons of strawberries and 1 ton of bananas. A imposes a 50% ad valorem consumption
tax on strawberries, whereas bananas are burdened with a 1% tax. A respects its obligations with
respect to the first sentence of Art III.2, first sentence (like products), but not with respect to Art
III.2, second sentence (DCS). The second sentence thus, operates as an anti-circumvention device
and protects trading partners from such beggar thy neighbour behaviour. This illustration rests on
rather extreme facts and, of course, there will be other, less controversial situations. What if A pro-
duces, for example, 40 tons of strawberries and 60 of bananas? Th is is where, unavoidably, an intent
test is warranted. More on this score in section 4.3.
 National Treatment: The Legal Discipline 227
argues, those imported products within a particular fiscal category. The cumulative con-
sequence of the New Chilean System is, as the Panel found, that approximately 75 per-
cent of all domestic production of the distilled alcoholic beverages at issue will be located
in the fiscal category with the lowest tax rate, whereas approximately 95 percent of the
directly competitive or substitutable imported products will be found in the fiscal cat-
egory subject to the highest tax rate. (original emphasis)
Note that the AB does not compare volume-wise, the remaining 25% of all
domestic production to the 95% of imported products. Chile was, essentially,
arguing that domestic producers were carrying the vast majority of the tax burden
in the high tax category. In its view, this was enough to establish that the legisla-
tion at hand did not operate SATAP.
This is a very reasonable argument: most tax regimes concerning alcoholic bev-
erages include a progressive taxation scheme, whereby each marginal degree of
alcohol adds a fi xed percentage to the tax.⁶⁹ It is true that, absent classifications
many Chilean consumers could have found that a drink of say 34° and a drink of
say 40° are DCS products. In the absence of a tax differential, however, Chilean
consumers would have probably consumed more of Chilean 40° drinks as well.
The fact that Chilean consumers find two products to be DCS does not mean
that Chile cannot regulate and afford different tax treatment to the products that
domestic producers treat as DCS; Chile can of course regulate and enact a dif-
ferent tax treatment for two DCS products: Chile did not, by virtue of Art III of
the GATT, promised to de-regulate; it simply promised not to discriminate, that
is, not to protect its domestic producers through means other than import tariffs.
The question hence, is whether the Chilean measure at hand operated SATAP.
On what evidence did the AB move to find against Chile? And is this case-law
consistent with the prior rulings?
We start from the latter question. In Japan—Alcoholic Beverages II, the SATAP
issue was not much of an issue: Japan did not refute that the tax differential
favoured its domestic production, and offered only ex post facto excuses to justify
its tax discriminatory regime (which the panel legitimately rejected not because
they were unfounded, but because they had been untimely); Japan, in short, for
legitimate due process reasons, was not given a real chance to refute the claims by
the complainants to the effect that the Japanese regime was essentially designed
to protect Japanese producers of shochu against competition of Western drinks.
In Korea—Alcoholic Beverages, Korea did not even prepare a defence on SATAP,
arguing, correctly, that the products at hand were not in a DCS relationship.

⁶⁹ All this in the name of protection of public health: this could be however, a highly ineffi-
cient instrument since it might push consumers to larger consumption of low alcoholic drinks
(depending on the marginal taxation). The efficiency of a measure, however, is of no concern in a
discussion concerning consistency with Art III of the GATT: no matter how inefficient a measure
is, all that matters is that it is non-discriminatory. Efficiency might be (indirectly) a concern in a
necessity analysis in the context of Art XX of the GATT, as well as in the context of the TBT, and
SPS Agreements.
228 Domestic Instruments

Chile is thus the first case where the panel had to entertain an argument by the
defendant that the measure at hand was not operating SATAP. Since the question
of consistency with prior case-law does not arise, we move to discuss the merits of
the evidentiary basis for the AB’s finding against Chile.
In Chile — Alcoholic Beverages, we have some information about the trade
impact of the measure: Chile, as stated above, argued (and its argument was not
refuted) that the majority of producers that paid the higher tax premium were
Chilean producers. The AB chose to disregard this evidence: in its view, all DCS
products must be treated in a similar manner, otherwise Art III.2 of the GATT
will be violated. This is an extraordinary statement. It essentially means that
states cannot intervene and distinguish between two products that are in a DCS
relationship in the eyes of the consumer. At the very least, they cannot do that in
the context of Art III of the GATT.
Defenders of the AB view in this respect would argue that nothing stops Chile
from defending itself under Art XX of the GATT: even assuming that Art XX of
the GATT⁷⁰ should be viewed as a defence for measures found to violate Art III
of the GATT, Chile would find it impossible to justify its progressive tax regime
under Art XX of the GATT: it would have to, as per the chapeau of Art XX of the
GATT, ensure that its tax regime does not constitute a means of discrimination
between countries where similar conditions prevail. In short, no discrimination can
be afforded. Chile would thus lose again, this time under Art XX of the GATT.
But how can it be that Chile loses a case like this? The answer probably lies in
the fact that Art III.2 of the GATT has been interpreted in an a-contextual man-
ner. Recall that Art III.1 of the GATT, the immediate context of Art III.2 of the
GATT, essentially outlaws regulatory interventions that operate so as to afford
protection. The objective of this provision is to combat protectionism; the means
is NT. If NT is to serve its purpose, it must be interpreted in light of the stated
objective, that is, to combat protectionism. One cannot decide whether a meas-
ure operates so as to afford protection, unless one asks the question about the gov-
ernment intentions and its trade impact. That is, whereas it is consumers that will
decide whether two products are DCS, the response to the question concerning
whether a regulatory intervention operates so as to afford protection necessitates
an inquiry into the regulating government’s motives and the trade impact of the
challenged measure. This construction is quite consistent with the nature and
the design of Art III of the GATT: it is about non-discrimination, not about de-
regulation. This means, that governments retain the right to intervene and treat
differently two products which, in the eyes of the consumers, are DCS products,
to the extent that their intervention does not operate so as to afford protection.⁷¹

⁷⁰ GATT, Art XX contains an exhaustive list of public order reasons which, if invoked, justify
violations of GATT obligations. GATT, Art XX is, according to standing case-law, an exception to
NT. For the reasons mentioned in Chapter 4, this should not be the case.
⁷¹ This could be the case because, for instance, governments have private information about
a trade hazard. Assuming symmetry of information, governments might still need to intervene, if
 National Treatment: The Legal Discipline 229

Recall that we lack an operational definition of protection, some sort of off the

shelf rule of thumb that we can use across transactions. Under the circumstances,
recourse to proxies is inevitable.⁷² The allocation of burden of proof in this respect
could be crucial. It is reasonable to assume that the complainant should at the very
least prove that there are negative effects stemming from a measure by another trad-
ing partner. In a case like Chile—Alcoholic Beverages, however, it will be difficult
for the complainant to pass this first obstacle: the majority of the tax burden in the
high tax category was shouldered by Chilean producers.⁷³ If the Chileans intended
to protect the welfare of their own producers, they would most likely have designed
legislation that would exonerate their producers from such a burden.
The test for compliance with Art III.2 of the GATT, as applied in Chile—
Alcoholic Beverages, does not ask such questions. Intent, if at all, will be relevant
only in one case: when we deal with DCS products and the tax differential is
more than de minimis but not large enough to establish, in and of itself, an incon-
sistency: in other words, the larger the tax differential the higher the likelihood
that a measure operates so as to afford protection (hence, the less need to have
recourse to proxies outside the tax differential), in the eyes of the AB.⁷⁴ Is this,
however, a reasonable assumption to make? Why cannot governments tax differ-
ently an environmentally friendly and an environmentally unfriendly good and
still be found to be in full consistency with their obligations? What if, in doing
that, the majority of the adjustment cost will fall on domestic producers? Will
the fact that one foreign producer have to shoulder a small portion of the cost be
enough to establish inconsistency with Art III of the GATT?⁷⁵ That would be an
odd definition of protectionism.
What went wrong has probably little to do with law, but rather with case-
law. The term ‘so as to aff ord protection’ suggests a test along the lines suggested
by Hudec (2000), Howse and Regan (2000), and Horn and Mavroidis (2004),
where both regulatory intent and trade effects could have found their place.⁷⁶

consumers do not behave in a manner to the government’s liking. Assuming, of course, that con-
sumers have, through their behaviour, internalized the externality, there will be little reason for
governments to intervene. What matters is that regulatory intervention will occur (in the majority
of the cases) when there are asymmetrical preferences across the government and its citizens.
⁷² See on this score, Horn and Mavroidis (2004).
⁷³ As Ehring (2002) very perceptively notes, it could be the case that one transaction suffices for
the AB to find that a measure violates Art III.2 of the GATT, irrespective whether the evaluation of
all other transactions points to a different conclusion.
⁷⁴ In Chile—Alcoholic Beverages, the AB mentioned the design of the legislation as a factor that
could be relevant in establishing the protective application of a measure. The choice of the term (pro-
tective application) probably suggests that the AB did not want to go against its prior case-law and
now introduce an intent test in Art III of the GATT. For all practical purposes, however, it did.
For DCS products where the tax differential is large and for like products (see infra), no recourse to
other factors is necessary.
⁷⁵ The response would of course be affirmative, were one to follow the Chile—Alcoholic Beverages
case-law. Similar questions are asked in Howse and Regan (2000), Hudec (1998 and 2000), and
Horn and Weiler (2003).
⁷⁶ More on this infra in section 4.3.
230 Domestic Instruments

Panels and the AB, however, saw too much in an earlier GATT case and repro-
duced the test embodied there, without asking too many questions about its val-
idity (intellectual legitimacy) across transactions.
In what follows, we provide a short narrative in an effort to trace the origins
of the exit of an intent test from GATT/WTO case-law. The leading GATT case
regarding the standard of review to be applied in Art III cases is the notorious
GATT panel report on US—Superfund.⁷⁷ The United States had been taxing
imported petroleum products slightly higher than their domestic counterparts.
When the measure was challenged before the GATT dispute settlement process,
the US response was that the difference was so minimal that it could not reason-
ably have an impact on the prices in the US market. The products concerned by
the US taxation scheme were like products, a point conceded by the the United
States, which did not advance any defence other than the one mentioned. In light
of the above, the GATT panel dismissed the US argument that their discrimin-
atory taxation scheme did not constitute a violation of Art III of the GATT, since
the effects on the market were negligible (§ 5.1.9):
. . . Article III:2, first sentence, cannot be interpreted to protect expectations on export
volumes; it protects expectations on the competitive relationship between imported and
domestic products. A change in the competitive relationship contrary to that provision
must consequently be regarded ipso facto as a nullification or impairment of benefits
accruing under the General Agreement. A demonstration that a measure inconsistent
with Article III:2, first sentence, has no or insignificant effects would therefore in the
view of the Panel not be a sufficient demonstration that the benefits accruing under that
provision had not been nullified or impaired even if such a rebuttal were in principle per-
mitted. (original underlining; italics added)
Note that here we are dealing with like products and the only US defence was the
one mentioned. What followed was a wide extension of this case-law. The AB in
its report on Japan—Alcoholic Beverages II reproduced this idea in the following
terms (p 16):
. . . it is irrelevant that ‘the trade effects’ of the tax differential between imported and
domestic products, as reflected in the volumes of imports, are insignificant or even non-
existent; Article III protects expectations not of any particular trade volume but rather of
the equal competitive relationship between imported and domestic products.
So a demonstration of trade effects became obsolete not only for like, but for
DCS products as well, since most of the products that the AB dealt with in this
case were DCS.
The relevance of intent, as we saw above, is a slightly different story. Following a
couple of GATT reports (US—Malt Beverages, and US—Taxes on Automobiles),⁷⁸
where panels held that intent matters in the context of analysis under Art III of

⁷⁷ We briefly discussed this case in Chapter 2, when analyzing the legal discipline on quotas.
⁷⁸ Both cases are discussed infra.
 National Treatment: The Legal Discipline 231

the GATT, WTO panels explicitly overruled the approach adopted there. WTO
panels did not, nevertheless, totally evict intent outside the four corners of ana-
lysis under Art III of the GATT: case-law has time and again stated that it is the
subjective intent of the legislator that does not matter, leaving thus room for the
relevance of objective intent (assuming, of course, that such distinction is war-
ranted): in Japan—Alcoholic Beverages II, the AB distinguishes between subject-
ive intent and the purpose of a regulatory intervention, as disclosed by objective
features of the design of the measure. The former is irrelevant when it comes
to establishing an Art III.2 of the GATT violation; the latter could be relevant
(p 27):
This third inquiry under Article III:2, second sentence, must determine whether ‘dir-
ectly competitive or substitutable products’ are ‘not similarly taxed’ in a way that affords
protection. This is not an issue of intent. It is not necessary for a panel to sort through
the many reasons legislators and regulators often have for what they do and weigh the
relative significance of those reasons to establish legislative or regulatory intent. If the
measure is applied to imported or domestic products so as to afford protection to domes-
tic production, then it does not matter that there may not have been any desire to engage
in protectionism in the minds of the legislators or the regulators who imposed the meas-
ure. It is irrelevant that protectionism was not an intended objective if the particular tax
measure in question is nevertheless, to echo Article III:1, ‘applied to imported or domestic
products so as to afford protection to domestic production.’ This is an issue of how the
measure in question is applied.
...
As in that case, we believe that an examination in any case of whether dissimilar tax-
ation has been applied so as to afford protection requires a comprehensive and objective
analysis of the structure and application of the measure in question on domestic as com-
pared to imported products. We believe it is possible to examine objectively the under-
lying criteria used in a particular tax measure, its structure, and its overall application
to ascertain whether it is applied in a way that affords protection to domestic products.
(original emphasis)
In Chile—Alcoholic Beverages, the AB made this cryptic statement more explicit
by bringing objective regulatory purpose, that is, the purpose as revealed through
the design and architecture of the measure, within the analysis of the SATAP
criterion. At the same time, however, it continued to explicitly reject a subjective
intent test.⁷⁹ Moreover, it did not establish any criteria as to how it will evaluate

⁷⁹ Th is terminology is indeed confusing: any intent is by definition almost, subjective. Indeed

it could not be otherwise. What the AB probably means to say here is that, whereas it does not
care about the motives that drove legislation, it does care about legislative expression. Th is latter
(legislative expression) will often reveal the intent of the intervening state. This approach is not
unproblematic: it might often be counter-productive difficult to distinguish the motives of those
lobbying for legislation from the legislation as such. The AB was probably worried with cases where
various lobbies with distinct preferences press for the same (more or less) legislation. An intent test,
in its view would be ill-equipped to deal with such cases. On the other hand, it continues to believe
that the place for intent test is in Art XX of the GATT. Th is latter view is wrong, for the reasons
discussed in Chapter 3. The former view should not be too much of a worry either: all that matters
232 Domestic Instruments

the objective regulatory purpose, other than referring to the design and the archi-
tecture of the measure at hand. It discussed summarily the four regulatory objec-
tives advanced as justification of the measure by Chile, and it explicitly rejected
the relevance of the necessity criterion when evaluating a claim under Art III of
the GATT (§§ 71–2):
We recall once more that, in Japan—Alcoholic Beverages, we declined to adopt an
approach to the issue of “so as to afford protection” that attempts to examine “the many
reasons legislators and regulators often have for what they do.” We called for examination
of the design, architecture and structure of a tax measure precisely to permit identifica-
tion of a measure’s objectives or purposes as revealed or objectified in the measure itself.
Thus, we consider that a measure’s purposes, objectively manifested in the design, archi-
tecture and structure of the measure, are intensely pertinent to the task of evaluating
whether or not that measure is applied so as to afford protection to domestic production.
In the present appeal, Chile’s explanations concerning the structure of the New Chilean
System—including, in particular, the truncated nature of the line of progression of tax
rates, which effectively consists of two levels (27 per cent ad valorem and 47 per cent ad
valorem) separated by only 4 degrees of alcohol content—might have been helpful in
understanding what prima facie appear to be anomalies in the progression of tax rates.
The conclusion of protective application reached by the Panel becomes very difficult to
resist, in the absence of countervailing explanations by Chile. The mere statement of
the four objectives pursued by Chile does not constitute effective rebuttal on the part of
Chile.
At the same time, we agree with Chile that it would be inappropriate, under Article
III:2, second sentence, of the GATT 1994, to examine whether the tax measure is neces-
sary for achieving its stated objectives or purposes. The Panel did use the word “necessary”
in this part of its reasoning. Nevertheless, we do not read the Panel Report as showing
that the Panel did, in fact, conduct an examination of whether the measure is necessary
to achieve its stated objectives. It appears to us that the Panel did no more than try to
relate the observable structural features of the measure with its declared purposes, a task
that is unavoidable in appraising the application of the measure as protective or not of
domestic production. (original emphasis)
Regulatory intent can thus be an issue in the limited case where two products are
DCS and the tax differential is not large enough. To detect intent, the design of
the measure can be quite telling: in this case, the fact that the tax differential was
substantial (20%) and the difference in alcoholic degree was not (4°), sufficed, in
the eyes of the AB, to detect objective regulatory intent. The underlying theory
justifying this approach must be that the Chilean law was designed with current
production in mind; Chileans knowing what drinks their competitors produce
and export to their market, designed a tax regime that conferred a competitive
advantage to their own producers. If this is indeed the hypothesis, not only intent
but also trade eff ects matter, and the panel and the AB should have taken them

is that there is no intent to afford protection to the domestic producer. It is irrelevant what other
intent might exist or creep in.
 National Treatment: The Legal Discipline 233

into account. They did not. Had they done just that, they would have had to ser-
iously entertain the Chilean argument, that is, that it was Chilean producers that
carried the burden of the higher tax premium.
The attitude of the AB has important repercussions on standing requirements
as well: what if at the moment Chile enacts legislation no one produces the high
alcohol drinks? Since trade effects do not matter, anyone could have brought a
complaint to the WTO. Would it succeed arguing that the sharp tax differential
between the two categories is in violation of Art III.2 of the GATT? A positive
response seems quite likely in light of the manner in which the expectations about
competitive relationship test, is now being understood.⁸⁰
To conclude on this score, the whole idea of de facto discrimination is to
punish camouflaged de jure discrimination. In the case of de jure one can legit-
imately presume intent.⁸¹ In the case of de facto, this is not necessarily the case.
The complainant will need to show that a measure as it stands does not serve its
stated purpose but hides protectionist intent. Trade effects analysis is an imper-
fect proxy for protectionist intent.⁸² Case-law (especially in the Chile—Alcoholic
Beverages dispute) has made a mockery out of such a test by neglecting both
effects and intent, and reducing the analysis under Art III of the GATT to a
mere comparison between two tax regimes, irrespective of considerations such
as how is the tax burden apportioned across foreign and domestic players, and
what is the regulatory intent. But if the judge refuses to check intent and effect,
then on what premises does de facto discrimination stand? The short answer is
on nothing much.⁸³

⁸⁰ Standing is thus, being conferred quite liberally in proceedings concerning legal challenges
against measures under Arts III and XI of the GATT: affected parties do not have to show trade
effects at all.
⁸¹ Th is is not to suggest that all de jure discrimination is protectionist. Assuming worldwide
sea pollution, for example, it could be the case that only the products of land locked states remain
unaffected. Such cases, however, are rather infrequent and anyway legislation can be drafted in
such a manner that it serves its purpose without naming the countries the goods of which will con-
tinue to be traded legitimately (ie all goods polluted by their exposure to x environmental pollution
will not be cleared for importation in this market).
⁸² Assume, for example, that a legislation distinguishes between environmentally friendly and
environmentally unfriendly products, any products. Assume further that all environmentally
unfriendly products are foreign. Such legislation would impair the income of foreign producers
(trade effects). It would hardly qualify, nonetheless, as protectionist legislation, by any reasonable
definition of the term protectionist. Th is is precisely the reason why the legal test to use in order to
decide on the consistency of such legislation with the GATT cannot be limited to an evaluation of
trade effects only.
⁸³ At the end of the report, the AB seems to suggest that what it found problematic was not the
fact that rates were progressing as the alcohol content was growing; it was that rates were progress-
ing too sharply. This opinion can only make sense in the GATT context if, as briefly alluded to
above, the sharp tax differentials are used as proxy for intent (which the AB refutes). Otherwise, it
makes no sense at all, since NT implies that WTO members are free to decide on the tax rates for
alcoholic drinks as long as they do not discriminate; there is nothing like a common tax regime to
which they must adhere. In section 4.3 we discuss this issue further, this time from a more norma-
tive angle.
234 Domestic Instruments

4.1.2 Like products

As stated supra, for a complainant to demonstrate that the national treatment
obligation has been violated with respect to like products (Art III.2, first sentence
of the GATT), it will have to demonstrate that:
(1) the domestic and the foreign products are like; and
(2) the latter is taxed in excess of the former.
The term ‘like products’ defines a class of products (as is the case of DCS products),
which come under the purview of Art III of the GATT.

4.1.2.1 Defining like products

The term ‘like products’ has been interpreted in many GATT panels. GATT case-
law evidences two trends. First, there are a number of cases that follow a market-
place test: likeness is defined by reference to consumers’ reactions.⁸⁴ Among the
many examples in this line of thinking, we can single out the reports on Border
Tax Adjustments and Japan—Alcoholic Beverages I.
The working party report on Border Tax Adjustments ⁸⁵ established four pertin-
ent criteria to define likeness (§ 18):
(1) the properties, nature and quality of the products;
(2) the end-uses of the products;
(3) consumers’ tastes and habits—more comprehensively termed consumers’
perceptions and behaviour—in respect of the products; and
(4) the tariff classification of the products.
The working party report on Border Tax Adjustments did not assign a particu-
lar weight on each of them. Relying on these four criteria, the panel report on
Japan—Alcoholic Beverages I held (§ 5.6) that alcoholic beverages:
should be considered as ‘like products’ in terms of Article III:2 in view of their similar
properties, end-uses and usually uniform classification in tariff nomenclatures.
The second category of cases, reveal a willingness to take into account regula-
tory intent when establishing likeness among domestic and foreign products.
There are two reported cases following this line of thinking: the Panel report on
US — Malt Beverages laid down, to some extent, the groundwork for the notorious
Gas Guzzler report (US—Taxes on Automobiles).⁸⁶
In its US — Malt Beverages report, the panel defines ‘likeness’ in the following
manner (§ 5.25):

⁸⁴ So far, there is not one single case where supply-substitutability has been accounted for when
defining likeness or DCS relationship.
⁸⁵ See GATT Doc L/3464, adopted on 2 December 1970, BISD 18S/97.
⁸⁶ This report remains unadopted and is, hence, of limited legal value.
 National Treatment: The Legal Discipline 235
Consequently, in determining whether two products subject to different treatment are
like products, it is necessary to consider whether such product differentiation is being
made ‘so as to afford protection to domestic production’. While the analysis of ‘like prod-
ucts’ in terms of Article III:2 must take into consideration this objective of Article III, the
Panel wished to emphasize that such an analysis would be without prejudice to the ‘like
product’ concepts in other provisions of the General Agreement, which might have dif-
ferent objectives and which might therefore also require different interpretations.
In its US—Taxes on Automobiles report,⁸⁷ the panel had the opportunity to
elaborate further on this proposition by providing its own legal benchmark to
establish likeness, the so-called aims and eff ects test (§§ 5.7 and 5.10):
In order to determine this issue, the Panel examined the object and purpose of
paragraphs 2 and 4 of Article III in the context of the article as a whole and the General
Agreement.
...
The Panel then proceeded to examine more closely the meaning of the phrase ‘so as to
afford protection.’ The Panel noted that the term ‘so as to’ suggested both aim and effect.
Thus the phrase ‘so as to afford protection’ called for an analysis of elements including
the aim of the measure and the resulting effects. A measure could be said to have the aim
of affording protection if an analysis of the circumstances in which it was adopted, in
particular an analysis of the instruments available to the contracting party to achieve the
declared domestic policy goal, demonstrated that a change in competitive opportunities
in favour of domestic products was a desired outcome and not merely an incidental con-
sequence of the pursuit of a legitimate policy goal. A measure could be said to have the
eff ect of affording protection to domestic production if it accorded greater competitive
opportunities to domestic products than to imported products. The effect of a measure
in terms of trade flows was not relevant for the purposes of Article III, since a change in
the volume or proportion of imports could be due to many factors other than government
measures. (original emphasis)
According to this view, consequently, likeness will not be defined by reference to
prevailing perceptions in the market-place about the pair of products, but, instead,
by reference to the regulatory aims pursued by the intervening government.⁸⁸
As already stated, supra, WTO case-law has made it clear that like products are
a sub-set of the wider category of the DCS products. Since the latter are defined in

⁸⁷ In this case, the European Community challenged the consistency of US tax scheme applic-
able to cars, according to which, the total fleet of a producer would be taken into account in order
to decide on the tax that would be imposed. Producers with a fleet that consisted of large cubism
cars (gas guzzlers) would suffer most, as a result. Many European producers belonged to this cat-
egory. The US regime was apparently enacted at a time when those suffering most were US produc-
ers, in an effort to dissuade consumers eager to buy such cars from buying them.
⁸⁸ The drafters of this report had probably all the good intentions; they realized the short-
comings of the mechanistic approach followed in Japan—Alcoholic Beverages I and Border Tax
Adjustment reports, and were looking for a way out where regulatory intent would be still part of
the test. Introducing, however, the intent test under the likeness analysis created more problems
than it actually solved, as explained in more detail in section 4.3. See, however, the very perceptive
critique of Mattoo and Subramanian (1998) on this score.
236 Domestic Instruments

the market-place, it can be inferred that the US — Malt Beverages and US—Taxes
on Automobiles case-law is no longer good law. The only remaining question is
what, besides DCS status, a pair of products (domestic, imported) must share, in
order to be regarded as like products. In Japan—Alcoholic Beverages II, the AB (pp
19ff ) ruled that the term ‘like products’ invites a narrow reading and that customs
classification is relevant to establish likeness, beyond the criteria traditionally used
to establish DCS relationship. In this often quoted passage, the AB held that:
The concept of ‘likeness’ is a relative one that evokes the image of an accordion. The
accordion of ‘likeness’ stretches and squeezes in different places as different provisions of
the WTO Agreement are applied. The width of the accordion in any one of those places
must be determined by the particular provision in which the term ‘like’ is encountered as
well as by the context and the circumstances that prevail in any given case to which that
provision may apply. We believe that, in Article III:2, first sentence of the GATT 1994,
the accordion of ‘likeness’ is meant to be narrowly squeezed.
However, not just any customs classification can assist in the definition of like-
ness. A necessary condition is that the classification be precise (pp 23–4):
If sufficiently detailed, tariff classification can be a helpful sign of product similarity.
...
It is true that there are numerous tariff bindings which are in fact extremely precise
with regard to product description and which, therefore, can provide significant guid-
ance as to the identification of ‘like products.’
This would usually be the case with respect to six-digit classifications. Four-digit
classifications are usually uninformative, and eight-digit classifications are a mat-
ter of national inscription, and not of worldwide acceptance.⁸⁹

4.1.2.2 Taxation in excess

The AB has stated that even a minimal tax differential suffices to satisfy the in
excess criterion. In Japan—Alcoholic Beverages II, the AB relevantly held (p 23):
Even the smallest amount of ‘excess’ is too much. ‘The prohibition of discriminatory
taxes in Article III:2, first sentence, is not conditional on a ‘trade effects test’ nor is it
qualified by a de minimis standard.’ (original emphasis)
This approach is very much in line with the GATT panel report on
US—Superfund: a mere arithmetic difference suffices for a demonstration of vio-
lation of the in excess requirement.⁹⁰ We will discuss the merits of this approach
in section 4.3 infra.

⁸⁹ Nothing, however, stops trading nations from treating imported products coming under an
eight-digit classification as like to their domestic counterparts. It is just that the refusal to grant this
status to other imported products does not make such other imported products by definition unlike
to products they could be competing against in a given market.
⁹⁰ Contrast this approach with that followed with respect to DCS products under the SATAP
requirement, where a de minimis difference is a threshold issue. Although the AB has not defined
 National Treatment: The Legal Discipline 237

4.2 Non-fiscal measures

GATT, Art III.4 reads:
The products of the territory of any Member imported into the territory of any other
Member shall be accorded treatment no less favourable than that accorded to like prod-
ucts of national origin in respect of all laws, regulations and requirements affecting their
internal sale, offering for sale, purchase, transportation, distribution or use.
Therefore, in order to show a violation of Art III.4 of the GATT, the complainant
will have to demonstrate that:
(1) a measure (law, regulation or requirement);
(2) aff ecting internal sale, offer for sale, purchase, transportation, distribution or
use;
(3) is affording the foreign like product;
(4) less favourable treatment.
The italicized terms hold the key to the understanding of Art III.4 of the GATT.

4.2.1 A measure . . .
GATT/WTO case-law has understood the term ‘all laws, regulations and
requirements’ figuring in Art III.4 of the GATT to be equivalent to the term
‘measure’ featured in Art XXIII.1b of the GATT, and in Art XI of the GATT.
As we have seen in the discussion of Art XI of the GATT, the term ‘measure’
has been interpreted widely. The same holds true in the context of Art III.4 of
the GATT. The next question is, of course, to what extent a measure is attrib-
uted to a government, since the GATT does not regulate private behaviour.
Once again, GATT/WTO case-law addressed the issue of attribution in the
context of Art III.4 of the GATT, in parallelism with the approach followed
in the context of Art XI of the GATT: hence, not only acts of governments,
but also acts of private parties can be challenged under Art III.4 of the GATT,
provided that they are attributed to a government.⁹¹ The panel in its report
on US—FSC accepted this well-established GATT case-law in the following

as yet what precisely constitutes a de minimis difference, it seems reasonable to infer that it should
correspond to whatever is required to aff ect prices. Th is is an a contrario argument to support this
thesis: in US—Superfund, the GATT panel dismissed the relevance of the US argument (that the
tax scheme did not affect prices at all) because, in its view, a trade effects test was unwarranted with
respect to tax differentials across like products. By contrast, the WTO case-law seems to suggest
that, moving away from like products, we need something more than mere arithmetic difference
in order to satisfy that Art III.2 has been violated. Large tax differentials have been considered to
satisfy in and of themselves this criterion, presumably, because of their effect on prices. A lower tax
differential, albeit one that still has an effect on prices, is probably what is required to overcome the
de minimis threshold.
⁹¹ Note that the issue of attribution is a non-issue in the context of Art III.2 of the GATT: fiscal
impositions can only be imposed by governments, or by non-governmental organs, following dele-
gation of authority by governments.
238 Domestic Instruments

terms (§ 10.376):
A literal reading of the words all laws, regulations and requirements in Article III:4 could
suggest that they may have a narrower scope than the word measure in Article XXIII:1(b).
However, whether or not these words should be given as broad a construction as the word
measure, in view of the broad interpretation assigned to them in the cases cited above, we
shall assume for the purposes of our present analysis that they should be interpreted as
encompassing a similarly broad range of government action and action by private parties
that may be assimilated to government action. In this connection, we consider that our
previous discussion of GATT cases on administrative guidance in relation to what may
constitute a ‘measure’ under Article XXIII:1(b), specifically the panel reports on Japan—
Semi-conductors and Japan—Agricultural Products, is equally applicable to the definitional
scope of ‘all laws, regulations and requirements’ in Article III:4. (original emphasis)

4.2.2 . . . affecting sale, offering for sale, . . .

GATT/WTO case-law has consistently opted for wide interpretations of the term
‘aff ecting’. Under this term, measures that not only actually, but also potentially
and/or indirectly affect trade have been subjected to judicial review.
The AB in its report on US—FSC had the opportunity to reaffirm the wide
interpretation of the term ‘aff ecting’ (§§ 208–10):
. . . the word ‘affecting’ assists in defining the types of measure that must conform to the
obligation not to accord ‘less favourable treatment’ to like imported products, which is
set out in Article III:4.
The word ‘affecting’ serves a similar function in Article I:1 of the General Agreement
on Trade in Services (the ‘GATS’), where it also defines the types of measure that are sub-
ject to the disciplines set forth elsewhere in the GATS but does not, in itself, impose any
obligation. . . .
In view of the similar function of the identical word, ‘affecting’, in Article III:4 of the
GATT 1994, we also interpret this word, in this provision, as having a ‘broad scope of
application.’ (original emphasis)
There is not one single reported case where a challenged measure failed to meet
the aff ecting requirement.⁹²

4.2.3 . . . of like products

GATT, Art III.4 does not, in contrast to Art III.2 of the GATT, distinguish
between like and DCS products. The threshold question, hence, arises whether,

⁹² WTO case-law is thus, a far cry from the jurisprudential evolution in the corresponding
provisions of the Treaty establishing the European Community (TEC), where the court, over the
years, narrowed their scope to avoid Type I errors (false positives). The European courts have over
the years excluded from the scope of Art 28 of the TEC (the corresponding provision to Art III.4 of
the GATT) measures which indirectly affect trade (provided that they are non-discriminatory) the
so-called modalities of selling (assuming they are non-discriminatory). They have further excluded
measures which correspond to essential requirements, that is, measures which pursue a public order
objective, such as the protection of environment. See on this score, Weatherill (2002) who compre-
hensively discusses the case-law evolution in this respect.
 National Treatment: The Legal Discipline 239

the term ‘like’ should have the same meaning in the two paragraphs. The AB
held the view that the term ‘like’ in Art III.4 of the GATT should be interpreted
in light of the over-arching purpose of Art III of the GATT: absent some paral-
lelism in the coverage across the two paragraphs (Art III.2 and Art III.4 of the
GATT), WTO members would be in the position to circumvent the prohibition
with respect to fiscal measures, by simply enacting non-fiscal measures to the
same effect (protectionism). Understanding that the two provisions are more or
less co-extensive, was thus very much, in the eyes of the AB, an anti-circumven-
tion device. It, thus, naturally held that the term ‘like’ in Art III.4 of the GATT
cannot have coverage wider than the combined coverage of the terms ‘like’ and
‘DCS products’ in Art III.2 of the GATT. In its report on EC—Asbestos, the AB
had to deal with a French decree which banned the sales of asbestos containing
construction material. The sales ban was non-discriminatory, that is, asbestos
containing construction material was banned, irrespective of its origin. One of
the questions before the panel (and the AB) was whether all construction mater-
ial, irrespective of whether asbestos containing or asbestos free products, were like
products. In order to respond to this question, the AB had to first define the scope
of like products (§§ 98–100):
As we have said, although this ‘general principle’ is not explicitly invoked in Article III:4,
nevertheless, it ‘informs’ that provision. Therefore, the term ‘like product’ in Article
III:4 must be interpreted to give proper scope and meaning to this principle. In short,
there must be consonance between the objective pursued by Article III, as enunciated in
the ‘general principle’ articulated in Article III:1, and the interpretation of the specific
expression of this principle in the text of Article III:4. This interpretation must, there-
fore, reflect that, in endeavouring to ensure ‘equality of competitive conditions,’ the ‘gen-
eral principle’ in Article III seeks to prevent Members from applying internal taxes and
regulations in a manner which affects the competitive relationship, in the marketplace,
between the domestic and imported products involved, ‘so as to afford protection to domes-
tic production.’
As products that are in a competitive relationship in the marketplace could be affected
through treatment of imports ‘less favourable’ than the treatment accorded to domestic
products, it follows that the word ‘like’ in Article III:4 is to be interpreted to apply to
products that are in such a competitive relationship. Thus, a determination of ‘likeness’
under Article III:4 is, fundamentally, a determination about the nature and extent of a
competitive relationship between and among products. In saying this, we are mindful
that there is a spectrum of degrees of substitutability of products in the marketplace, and
that it is difficult, if not impossible, in the abstract, to indicate precisely where on this
spectrum the word ‘like’ in Article III:4 of the GATT 1994 falls. We are not saying that
all products which are in some competitive relationship are ‘like products’ under Article
III:4. In ruling on the measure at issue, we also do not attempt to define the precise scope
of the word ‘like’ in Article III:4. Nor do we wish to decide if the scope of ‘like products’
in Article III:4 is co-extensive with the combined scope of ‘like’ and ‘directly competi-
tive or substitutable’ products in Article III:2. However, we recognize that the relation-
ship between these two provisions is important, because there is no sharp distinction
240 Domestic Instruments
between fiscal regulation, covered by Article III:2, and non-fiscal regulation, covered
by Article III:4. Both forms of regulation can often be used to achieve the same ends. It
would be incongruous if, due to a significant difference in the product scope of these two
provisions, Members were prevented from using one form of regulation—for instance,
fiscal—to protect domestic production of certain products, but were able to use another
form of regulation—for instance, non-fiscal—to achieve those ends. This would frus-
trate a consistent application of the ‘general principle’ in Article III:1. For these reasons,
we conclude that the scope of ‘like’ in Article III:4 is broader than the scope of ‘like’ in
Article III:2, first sentence. Nonetheless, we note, once more, that Article III:2 extends
not only to ‘like products,’ but also to products which are ‘directly competitive or sub-
stitutable,’ and that Article III:4 extends only to ‘like products.’ In view of this different
language, and although we need not rule, and do not rule, on the precise product scope
of Article III:4, we do conclude that the product scope of Article III:4, although broader
than the first sentence of Article III:2, is certainly not broader than the combined product
scope of the two sentences of Article III:2 of the GATT 1994.
We recognize that, by interpreting the term ‘like products’ in Article III:4 in this
way, we give that provision a relatively broad product scope—although no broader
than the product scope of Article III:2. (original emphasis)
Having decided on the scope of the term ‘like’ products, the next question would
naturally be whether the approach followed with respect to fiscal measures would
also be followed with respect to non-fiscal measures. Which should, in other
words, be the comparator for likeness in Art III.4 of the GATT: the marketplace
or government intent?⁹³
Recall that the ruling in the panel report on Border Tax Adjustment, that we
examined supra, has been repeated in virtually all cases which discussed likeness
in the context of Art III.4 of the GATT. Recall further that this report never made
clear whether one of the four criteria was more important relative to the others,
or conversely, whether they all retained equal value in the eyes of the GATT/
WTO adjudicating bodies. In its report on EC—Asbestos, the AB interpreted
one of them (physical characteristics) in some detail. In this case, as stated above,
Canada launched a complaint against France, because the latter had enacted a
statute imposing a sales ban on asbestos containing construction material, irre-
spective of origin. France’s prohibition (administrative decree) was motivated by
scientific evidence to the effect that asbestos containing construction material
contributed to at least one form of cancer (mesothelioma). The asbestos contain-
ing construction material was chrysotile fibres (an input to the final product).
Canada had argued that there was no difference between construction material
containing chrysotile fibres on the one hand, and construction material contain-
ing PCG fibres (which is an asbestos free input to the final product), on the other.

⁹³ On this issue, see various contributions in Bhagwati and Hudec (1996), Horn and Weiler
(2003), Howse (2000), Howse and Regan (2000), Howse and Turk (2001), Hudec (1998 and
2000), Mattoo and Subramanian (1998), Petersmann (2000), Roessler (1996), and Trachtman
and Marceau (2002).
 National Treatment: The Legal Discipline 241

Construction material containing PCG fibres was being legally sold in France.
Canada argued, that, by virtue of the French prohibition to sell material contain-
ing chrysotile fibres, less favourable treatment was afforded to a like product.
The panel decided that the two products were like and, consequently, held that
Art III.4 of the GATT had been violated. The panel paid particular attention to
the end uses of the two products: since they were the same (construction mater-
ial), the panel held that the two products were indeed like. Having established vio-
lation of Art III.4 of the GATT, the panel then moved to examine the EC defence
under Art XX of the GATT. On appeal, the AB reversed the panel’s findings with
respect to likeness. In its view, the panel should have examined all four criteria
mentioned in Border Tax Adjustments, and not just one of them. Had it done
so, the panel would, in the AB’s view, always observed the differences in physical
characteristics between the two products: the composition of a product (production
process) is very much part of the physical characteristics analysis. Chrysotile fibres
and PCG fibres are not the same: the first are carcinogenic, whereas the latter
are not. This, in the AB’s view, most likely would have led consumers to stop
purchasing material containing chrysotile fibres. The likelihood that the different
composition might affect consumers’ choices in this respect, was sufficient reason
to raise a presumption that the two products are unlike. The burden of proof for
Canada, in light of the difference in physical characteristics, would be now, in the
words of the AB, much higher. Canada never managed to discharge its burden
of proof, and consequently, its original legal challenge against the French decree
was rejected. The AB found additional support for its overall finding, in the fact
that chrysotile fibres and PCG fibres do not share the same tariff classification,
and, also, in the fact that scientific evidence was cited in support of the carcino-
genic nature of chrysotile fibres (§§ 101–54).⁹⁴ It bears repetition that the AB did
not possess any consumers’ surveys indicating that they indeed treated the two
products as unlike. It presumed that this would have been the case, had the con-
sumers known about the health risk associated with the consumption of asbestos
containing construction material (§ 122):
In this case especially, we are also persuaded that evidence relating to consumers’ tastes
and habits would establish that the health risks associated with chrysotile asbestos fibres
influence consumers’ behaviour with respect to the different fibres at issue. We observe
that, as regards chrysotile asbestos and PCG fibres, the consumer of the fibres is a manufac-
turer who incorporates the fibres into another product, such as cement-based products or
brake linings. We do not wish to speculate on what the evidence regarding these consum-
ers would have indicated; rather, we wish to highlight that consumers’ tastes and habits
regarding fibres, even in the case of commercial parties, such as manufacturers, are very

⁹⁴ In a separate but concurring opinion, an unnamed member of the AB held the view, that the
scientific proof cited in this case was sufficient to conclude that the two products were unlike. One
way to understand the need for a separate opinion is probably that, in this member’s eyes, the diffe-
rence in physical characteristics does not merely raise a presumption, but amounts to a home run:
Canada could never rebut such evidence.
242 Domestic Instruments
likely to be shaped by the health risks associated with a product which is known to be highly
carcinogenic. A manufacturer cannot, for instance, ignore the preferences of the ultimate
consumer of its products. If the risks posed by a particular product are sufficiently great, the
ultimate consumer may simply cease to buy that product. This would, undoubtedly, affect
a manufacturer’s decisions in the marketplace. Moreover, in the case of products posing
risks to human health, we think it likely that manufacturers’ decisions will be influenced
by other factors, such as the potential civil liability that might flow from marketing prod-
ucts posing a health risk to the ultimate consumer, or the additional costs associated with
safety procedures required to use such products in the manufacturing process.
It follows that the test for likeness remains in the marketplace, in name only,
since there was no empirical evidence submitted to the AB regarding consumers’
preferences. The AB de facto adopted a reasonable consumer test, speculating on
its reactions in case of symmetry of information. This approach is problematic for
a number of reasons:
(1) if manufacturers, as the AB contends, were indeed anticipating consumers’
reactions, then, by backwards induction, they would not be producing asbestos
containing products either; as a result, there would be no need for regulation in
the first place. Manufacturers stopped producing asbestos containing construc-
tion material, not because of consumers’ reactions, but because of the statutory
prohibition to produce;
(2) if all that is needed was information to ensure symmetry of information
across government agencies and private consumers, then, in all likelihood, France
would have financed an information campaign, and it would not have enacted a
statutory prohibition on sales of asbestos containing products. The government
was probably worried that private parties might through their behaviour impose
health externalities on the society, and this what prompted the prohibition;
(3) some manufacturers at least do produce such products and some consum-
ers continue to buy them. They are located in Canada and not in France. Are
they unreasonable? Assuming that the risk associated with the consumption of
such goods is, relatively speaking, low, and the price difference between asbestos
containing and asbestos free products substantial, a country might think it makes
good sense to allow for the production of both products, tax production/con-
sumption of the former, and finance research to combat mesothelioma.
This is not to suggest that France should have lost that case. France should have
won anyway: its victory, however, should have come under the less favourable
treatment analysis, a point to which we return immediately.

4.2.4 Less favourable treatment

The AB in its report on EC—Asbestos held the view that the term ‘less favour-
able treatment’ (LFT) appearing in Art III.4 of the GATT echoes the principle
set forth in Art III.1 of the GATT: WTO members should not use domestic
 National Treatment: The Legal Discipline 243

measures to afford protection to domestic production (§ 100):

The term ‘less favourable treatment’ expresses the general principle, in Article III:1, that
internal regulations ‘should not be applied . . . so as to afford protection to domestic produc-
tion.’ If there is ‘less favourable treatment’ of the group of ‘like’ imported products, there is,
conversely, ‘protection’ of the group of ‘like’ domestic products. However, a Member may
draw distinctions between products which have been found to be ‘like,’ without, for this
reason alone, according to the group of ‘like’ imported products ‘less favourable treatment’
than that accorded to the group of ‘like’ domestic products. (original emphasis)⁹⁵
The AB clarified, in its report on US — FSC, that, in parallel with its case-law
under Art III.2 of the GATT, for a complainant successfully to absolve its burden
of proof and demonstrate that the group of imported products have been afforded
LFT than the group of domestic like products, there is no obligation to show
trade effects (§ 215):
The examination of whether a measure involves ‘less favourable treatment’ of imported
products within the meaning of Article III:4 of the GATT 1994 must be grounded in
close scrutiny of the ‘fundamental thrust and effect of the measure itself.’ This examin-
ation cannot rest on simple assertion, but must be founded on a careful analysis of the
contested measure and of its implications in the marketplace. At the same time, however,
the examination need not be based on the actual eff ects of the contested measure in the
marketplace. (original emphasis)
Following this line of thinking, the AB outlawed an EC scheme which pro-
vided traders with an incentive to trade EC goods at the expense of imported
like products, even in the absence of any tangible evidence suggesting that there
was a quantified trade impact stemming from this measure (§§ 213–14 of the
AB report on EC—Bananas III). By the same token, the panel, in its report on
Canada—Wheat Exports and Grain Imports, outlawed a routine authorization
scheme simply because it was not being imposed on domestic products (§ 6.193).
The fact that the scheme was not onerous at all (§ 6.193), that economic opera-
tors were familiar with it (§ 6.197), that in the past authorization had always
been granted (§ 6.200), that the agency at hand had enough discretion to always
grant authorization (§ 6.203), that there was an institutional possibility to obtain
advance authorization (§ 6.207), that a more attractive than the authorization
scheme alternative was available to foreign products (§ 6. 213) were not reason
enough for the panel to change its mind on this issue.⁹⁶

⁹⁵ Note, however, that earlier, in EC—Bananas III, the AB had held: ‘Article III:4 does not spe-
cifically refer to Article III:1. Therefore, a determination of whether there has been a violation of
Article III:4 does not require a separate consideration of whether a measure “afford[s] protection to
domestic production” ’ (original emphasis)
⁹⁶ WTO members might have to quantify the effects of a trade measure they have success-
fully challenged, at a later stage of the proceedings: assuming no compliance has occurred and the
complainant requests authorization to impose counter-measures, it will have to quantify the trade
damage it suffered, in light of the obligation enshrined in Art 22.4 of the DSU, that the overall level
of counter-measures should not exceed that of the trade damage incurred.
244 Domestic Instruments

This interpretation of the less favourable treatment requirement is not problem

free; as is the case with the SATAP requirement, the standard of review employed
by the WTO adjudicating bodies is a no effects no intent test.⁹⁷ The AB report
on Korea—Various Measures on Beef exemplifies the various problems encoun-
tered when using this test. This dispute, between Korea on the one hand, and
a host of beef exporters on the other, concerned the distribution of beef in the
Korean market. Korea had enacted a law, whereby traders at the retail level could
sell either domestic or imported beef: this was the notorious Korean dual retail
system. Korea also had in place an import quota for beef, the legality of which
with the GATT rules was not an issue during this dispute. The complainants had
argued that, as a result of the dual retail system, there were only 5,000 points of
sale for imported beef, whereas there were over 45,000 points of sale for domestic
beef. Korea responded, arguing that it had absorbed all of the import quota dur-
ing all years when the system was in place, with one exception only: following the
financial crisis in 1997, when overall consumption (of both domestic and foreign
beef) fell dramatically, it did not manage to absorb the quota. Korea also argued
before the panel that the dual retail system was not discriminatory, in light of the
fact that traders were free to choose whether they would sell domestic or foreign
beef, and that there was no legal compulsion obliging them to choose one category
of beef over the other; they could further subsequently switch at no cost, and sell
say imported beef, when they had previously been selling domestic. In Korea’s
view, the dual retail system established equality of competitive conditions. Since no
demonstration of trade effects is required for a violation of Art III of the GATT
to be established, Korea requested the panel reject the complaint; the number of
outlets is immaterial since all that matters is the establishment of equal competi-
tive conditions across domestic and foreign beef, something the Korean law amply
protected.⁹⁸
The panel rejected all of Korea’s claims, and found in favour of the complain-
ants. On appeal, the AB held the view that this system, although formally non-
discriminatory, still modified the conditions of competition to the detriment of the
imported product and found Korea’s practices to be inconsistent with Art III.4
of the GATT. The AB failed to explain, however, why modification of the con-
ditions of competition amounts to a violation of Art III of the GATT. At the
same time, the AB was careful enough to avoid outlawing private initiatives (ie
exclusivity contracts) that might lead to the same effect, such as, separation of
outlets selling domestic and imported like products. This is a rather extraordin-
ary judgment, and to do justice to the AB argument, we privileged a citation of

⁹⁷ As noted above, regulatory intent is relevant only in cases where the foreign DCS product has
been taxed higher than the domestic one, but not substantially higher.
⁹⁸ As we will see in more detail infra (section 4.4), Korea argued that the system was in place in
order to combat tax fraud: traders had the incentive to sell imported beef for domestic, in light of
the very substantial price-differential between imported and domestic beef (the latter being sub-
stantially more expensive than the former).
 National Treatment: The Legal Discipline 245

the relevant lengthy passage (§§ 143–51):

Korean law in effect requires the existence of two distinct retail distribution systems so
far as beef is concerned: one system for the retail sale of domestic beef and another system
for the retail sale of imported beef. A small retailer (that is, a non-supermarket or non-
department store) which is a ‘Specialized Imported Beef Store’ may sell any meat except
domestic beef; any other small retailer may sell any meat except imported beef. A large
retailer (that is, a supermarket or department store) may sell both imported and domes-
tic beef, as long as the imported beef and domestic beef are sold in separate sales areas. A
retailer selling imported beef is required to display a sign reading ‘Specialized Imported
Beef Store.’
Thus, the Korean measure formally separates the selling of imported beef and domes-
tic beef. However, that formal separation, in and of itself, does not necessarily compel
the conclusion that the treatment thus accorded to imported beef is less favourable than
the treatment accorded to domestic beef. To determine whether the treatment given to
imported beef is less favourable than that given to domestic beef, we must, as earlier
indicated, inquire into whether or not the Korean dual retail system for beef modifies the
conditions of competition in the Korean beef market to the disadvantage of the imported
product.
When beef was first imported into Korea in 1988, the new product simply entered into
the pre-existing distribution system that had been handling domestic beef. The beef retail
system was a unitary one, and the conditions of competition affecting the sale of beef were
the same for both the domestic and the imported product. In 1990, Korea promulgated
its dual retail system for beef. Accordingly, the existing small retailers had to choose
between, on the one hand, continuing to sell domestic beef and renouncing the sale of
imported beef or, on the other hand, ceasing to sell domestic beef in order to be allowed
to sell the imported product. Apparently, the vast majority of the small meat retailers
chose the first option. The result was the virtual exclusion of imported beef from the
retail distribution channels through which domestic beef (and until then, imported beef,
too) was distributed to Korean households and other consumers throughout the country.
Accordingly, a new and separate retail system had to be established and gradually built
from the ground up for bringing the imported product to the same households and other
consumers if the imported product was to compete at all with the domestic product.
Put in slightly different terms, the putting into legal effect of the dual retail system for
beef meant, in direct practical effect, so far as imported beef was concerned, the sudden
cutting off of access to the normal, that is, the previously existing, distribution outlets
through which the domestic product continued to flow to consumers in the urban centers
and countryside that make up the Korean national territory. The central consequence of
the dual retail system can only be reasonably construed, in our view, as the imposition of
a drastic reduction of commercial opportunity to reach, and hence to generate sales to,
the same consumers served by the traditional retail channels for domestic beef. In 1998,
when this case began, eight years after the dual retail system was first prescribed, the con-
sequent reduction of commercial opportunity was reflected in the much smaller number
of specialized imported beef shops (approximately 5,000 shops) as compared with the
number of retailers (approximately 45,000 shops) selling domestic beef.
We are aware that the dramatic reduction in number of retail outlets for imported beef
followed from the decisions of individual retailers who could choose freely to sell the
246 Domestic Instruments
domestic product or the imported product. The legal necessity of making a choice was,
however, imposed by the measure itself. The restricted nature of that choice should be
noted. The choice given to the meat retailers was not an option between remaining with
the pre-existing unified distribution set-up or going to a dual retail system. The choice
was limited to selling domestic beef only or imported beef only. Thus, the reduction of
access to normal retail channels is, in legal contemplation, the effect of that measure. In
these circumstances, the intervention of some element of private choice does not relieve
Korea of responsibility under the GATT 1994 for the resulting establishment of competi-
tive conditions less favourable for the imported product than for the domestic product.
We also note that the reduction of competitive opportunity through the restriction of
access to consumers results from the imposition of the dual retail system for beef, not-
withstanding the ‘perfect regulatory symmetry’ of that system, and is not a function of
the limited volume of foreign beef actually imported into Korea. The fact that the WTO-
consistent quota for beef has, save for two years, been fully utilized does not detract from
the lack of equality of competitive conditions entailed by the dual retail system.
We believe, and so hold, that the treatment accorded to imported beef, as a conse-
quence of the dual retail system established for beef by Korean law and regulation, is less
favourable than the treatment given to like domestic beef and is, accordingly, not consist-
ent with the requirements of Article III:4 of the GATT 1994.
It may finally be useful to indicate, however broadly, what we are not saying in reach-
ing our above conclusion. We are not holding that a dual or parallel distribution system
that is not imposed directly or indirectly by law or governmental regulation, but is rather
solely the result of private entrepreneurs acting on their own calculations of comparative
costs and benefits of differentiated distribution systems, is unlawful under Article III:4 of
the GATT 1994. What is addressed by Article III:4 is merely the governmental interven-
tion that affects the conditions under which like goods, domestic and imported, compete
in the market within a Member’s territory.
Finally, we note that Korea requires that imported beef be sold in a store displaying a
sign declaring ‘Specialized Imported Beef Store.’ The Panel found that the sign require-
ment was ‘essentially ancillary to the dual retail system,’ and, since the dual retail system
was inconsistent with Article III:4, the Panel found that the sign requirement was as
well. The Panel found also that, since the sign requirement went beyond an obligation to
indicate origin, it was inconsistent with statements contained in a 1956 GATT Working
Party Report. On appeal, Korea claims that the Panel used ‘erroneous logic’ in its conclu-
sion based on the ancillary character of the sign requirement, and that the statement in
the Working Party Report did not apply to the sign requirement.
Without a system of specialized imported beef stores, the sign requirement would have
no meaning and would not be required. When considered independently from a dual
retail system, a sign requirement might or might not be characterized legally as consistent
with Article III:4 of the GATT 1994. On the other hand, when considered simply as a
detail of the dual retail system, the sign requirement partakes of the legal characterization
given to that system itself. We believe it unnecessary to pass upon separately the con-
sistency of the display sign requirement with Article III:4 of the GATT 1994. (original
emphasis)
Note that the AB did not state that the decision of traders was prejudged by the
Korean legislation. However, in its view, traders would not have made any such
 National Treatment: The Legal Discipline 247

decision, absent the Korean legislation. Hence, that they made such a decision
(although not compelled, and arguably, not even prompted by the legislation,⁹⁹ but
merely while the legislation on dual retail system was in place), and because, as a
result, we ended up with the known market outcome (fewer outlets for imported
beef), was enough in the AB’s view to find the Korean measure in violation of the
GATT. In short, the AB outlawed a non-discriminatory measure which, follow-
ing rational, independent choices by traders, led to an outcome which, in the eyes
of the AB was not consonant with Art III.4 of the GATT. It bears repetition that
the choice by traders was rational, in light of the fact that there was a legal quota
in place and consumers would satisfy their needs in beef essentially through pur-
chases of domestic beef.
It is probably worth noting too, that the AB, when dealing with Korea’s
defence under Art XX of the GATT (which we discuss infra), accepted that
Korea, through this legislation, intended to protect consumers’ interests. It found
against Korea, on the narrow(er) issue that the measure at hand did not meet the
necessity test enshrined in Art XX(d) of the GATT. It did not, however, find that
Korea’s measures were a disguised restriction of trade.
In short, the Korean law at hand did not in and of itself produce any trade
effects negative to Korea’s trading partners, and the intent of the legislation,
as the AB itself admits, was not to protect Korean producers’ welfare. Yet, the
law at hand was struck down as inconsistent with Art III.4 of the GATT. This
is indeed troubling. And it is the direct outcome of the perverse standard of
review employed by WTO adjudicating bodies. Is there a solution to avoid such
outcomes?¹⁰⁰

4.3 A critique of the case-law concerning NT

Case-law in the context of NT has not been helpful: if the role of case-law, beyond
resolving a dispute, is to complete the contract and make it clearer and, crucially,
more predictable, then case-law on NT has failed its basic mission.
Indeed, how can anyone respond to the question what are DCS/like products
in the GATT?
First, as we saw above, price difference is not necessarily a criterion to define
them, so guess work about consumers’ reactions (as opposed to surveys where use
of econometric indicators has been made) becomes the current benchmark to

⁹⁹ Indeed, because of the price difference, Korean traders had an interest in distributing US beef
which would be circulated anyway (as available statistics showed).
¹⁰⁰ So far, there has been no case in the WTO where an omission was found to be a measure
inconsistent with Art III of the GATT. In the EU context, the court has found that France was
in breach of its obligations under the ECT (Art 28) by not taking measures sufficient to deter
French farmers from emptying trucks and destroying Spanish farm products that had been cleared
through customs, see C-265/95 Commission of the European Communities v French Republic
[1997] ECR I-6959.
248 Domestic Instruments

define ‘likeness’. To justify case-law, one would have to accept as true one of the
following two views:
(1) non-econometric indicators, such as consumer preferences, end uses of a
product, are not captured in a cross price elasticity test;
(2) the critical mass of consumers make their purchasing decisions irrespective
of scarcity of pecuniary resources.
Both views are wrong: a consumer survey where cross price elasticity has been
used will typically ask the question ‘which product would consumers be switching
to, if the price of product A rises by x%?’; on the other hand, with the exception of
very very few outliers indeed, the overwhelming majority of consumers purchase
in accordance with their purchasing capacity. Price does matter. Excluding price
considerations from a test concerning DCS products takes away the foundation
stone from the establishment of the relevant product market.
Second, case-law seems to suggest that tax differentials almost always hide pro-
tectionist attitude. It confuses two distinct issues: the allocation of the burden
of proof (production and persuasion), and the sovereignty to pursue any prefer-
ence a society might deem worth pursuing. We explain. Tax differentials might
be necessary to pursue legitimate public order policies, such as public health and/
or environmental protection. They might be hiding protectionist intent, and one
could argue that the regulator, since it is best placed to explain legislative intent,
should carry the burden of proof to justify a tax differential. It should not, how-
ever, following a demonstration of a tax differential, be pushed to do so outside
the confines of Art III of the GATT, and into the exhaustive list of Art XX of
the GATT. As will be explained in more detail infra, this construction of
Art III of the GATT puts into question the foundation of the whole GATT edifice,
negative integration: WTO members can pursue any (and not just those mentioned
in Art XX of the GATT) regulatory objective they deem worth pursuing, to the
extent that, when doing so, they abide by the NT discipline. Wisely, the judges at
the ECJ through the essential requirements and the Keck and Mithouard doctrines,
reduced the scope of Art 28 of the ECT (corresponding to Art III of the GATT
provision) and left it open to reduce it even more in the future. The WTO judges
dealing with a much less ambitious agenda (the GATT today is not about common
policies, it is about trade liberalization) have followed the opposite pattern.
Third, even when some steps have been taken to exit the current deadlock,
case-law has at best established halfway houses. Take, for example, the statement
that less favourable treatment in Art III.4 of the GATT has a function equivalent
to that of the SATAP test. What does it mean? The AB will not inquire into regu-
latory intent, when applying the SATAP test in cases where the tax differential is
large. What are the corresponding cases in the realm of non-fiscal policies? The
AB, conversely, will inquire into regulatory intent (as reflected in the design of
the legislation), when dealing with cases where the tax differential is more than de
 National Treatment: The Legal Discipline 249

minimis but not large. Once again we can ask, what are the corresponding cases
in the realm of non-fiscal policies?
Fourth, there is a more general problem with the current standard of review,
which does not inquire into trade effects at all, and, in limited cases only, if at
all, will inquire into the intent of the regulating state: effectively, the AB will not
inquire into regulatory intent. Why is this case? One can reasonably assume that
the usual case is that, say producing an environmentally friendly good involves
a costlier process of production than producing the corresponding environmen-
tally unfriendly good: the former will be burdened with a fi xed cost, the cost cor-
responding to turning it into an environmentally friendly product. A domestic
measure aiming at discouraging consumption of the environmentally unfriendly
good will have to take care of this cost difference and provide a price advantage
to the environmentally friendly good (in other words, regulation will not simply
try to equate the two prices; it will try to tilt the advantage towards the side of the
environmentally friendly good). Marginal tax differentials will usually not do for
this purpose. Consequently, WTO adjudicating bodies will not be looking into
regulatory intent in most cases. They will not be looking into effects either. By
disregarding effects and intent, WTO adjudicating bodies are confusing means and
ends: tax diff erentials are equated to protectionism. This is, however, absurd. Tax
differentials, as the discussion above shows, can and do occur for reasons other
than protectionism: a carbon tax for example, can be imposed in order to guaran-
tee that polluters pay for the pollution they have created. GATT, Art III does not
outlaw tax diff erentials; it outlaws protectionism. WTO adjudicating bodies have
turned this test on its head.
True, establishing the intent of the regulating state is far from being an easy
exercise. It is not for concerns of this type, however, that we should substitute this
test for a test that essentially looks into two numbers without asking questions
about their rationale.
A basic premise of our discussion here is that Art XX of the GATT should not
be construed as an exception to Art III of the GATT. This line of thinking is at
odds with current case-law, and is explained infra (section 4.4). However, even if
the current case-law were correct, the interpretation of Art III of the GATT is not
problem-free: GATT, Art III covers all domestic instruments, whereas Art XX of
the GATT only a sub-set of them. This means that with respect to, at least some,
domestic instruments, WTO members have no defence under Art XX of the
GATT. This conclusion makes all the more palatable the need for a substantive
discrimination test in Art III of the GATT.
A number of papers in the literature have criticized the current test.¹⁰¹ The
standard of review as it has developed in the context of NT is zealously oriented
towards avoiding Type II errors (false negatives), ignoring the huge institutional
risk that Type I errors (false positives) might represent. Horn and Mavroidis

¹⁰¹ We referred to many of these papers in the previous sections.

250 Domestic Instruments

(2004) essentially propose a two-tier test whereby, it is consumers who will define
whether two goods are DCS-like, whereas it is government intentions¹⁰² that will
ultimately decide whether a matter was designed SATAP. In their scheme, adju-
dicating bodies should employ proxies to establish (protectionist) intent, trade
effects being one of them. Although the authors do not elaborate on that, it is
clearly the case that the allocation of burden of proof holds one of the keys in
unraveling the true intent behind a measure: considerations such as who is best
equipped to provide information, should matter in allocating the burden of pro-
duction of proof. This understanding is in line with the overarching purpose of
the GATT: to eliminate all forms of protection, other than tariffs (in order, it
bears repetition, to provide trading nations with an incentive to continue lib-
eralizing tariffs). The authors acknowledge that we are still some way before we
can celebrate a fit-for-all test. It is not by accident, one should recall, that many
jurisdictions have struggled with national treatment type disciplines: the ECJ has
even explicitly stated that it will distance itself from its prior case-law on Art 28
of the ECT (Weatherill, 2002), whereas there are many accounts showing how
judges in various US jurisdictions have struggled to implement the US Commerce
Clause.¹⁰³ As mentioned above, trade effects are no perfect proxy to demonstrate
behaviour that would satisfy the SATAP test. Moving to an evaluation of the
regulatory intent becomes thus a necessity. But when we move there, we are mov-
ing into an asymmetry of information context: the regulator has private informa-
tion about its preferences, whereas the trading partners, affected by its choices,
do not. Moreover, the regulator has a strong incentive to cheat: arguing that its
interventions, its true intent notwithstanding, aim at protecting public order,
will exonerate it from liability under the GATT rules. That is, the party with
the private information is the party with the incentive to cheat (to behave in an

¹⁰² There is also a lot of textual support for the suggestion that intent matters: we signaled above
the second sentence of Art III.2 of the GATT, along with the term ‘so as to aff ord protection’ which
appears in Art III.1 of the GATT. There is substantial legal literature that suggests that an intent
based test is quite suitable in this context. Fauchald (1998) and Ortino (2003) have expressed such
views in a coherent manner (along of course with the already mentioned work of Hudec (1998),
Howse and Regan (2000), Regan (2002 and 2003), and Horn and Weiler (2003)). A recurring theme
in this literature, however, is that, whereas the relevance of intent is acknowledged, no concrete
method of applying it in particular disputes is advanced. Not that to come up with a step-by-step,
fit-for-all strategy is an easy task; indeed, if this was really the case, efforts in this direction would
probably have already seen the light of the day. Verhoosel (2002) takes a different view. In his view
the starting-point of a review under Art III of the GATT should be consumers’ reactions: products
will be deemed to be like to the extent that consumers perceive them to be so. The use of economet-
ric indicators in this respect is very much favoured by the author. Assuming that goods are like and
that a taxation is discriminatory, the author suggests that a shift in the burden of proof is warranted
so that the informed party (that is, the WTO member the regulatory intervention of which is being
challenged) can defend say the dissimilarity in taxation across products perceived to be like by
consumers. This approach has definitely some undeniable advantages but might lead to unneces-
sary litigation if, for example, the reason for tax differential is observable (that is, symmetrically
known to the regulating state and the affected trading partner). The shift is probably warranted, in
case of private information.
¹⁰³ See, for example, Hudec (1988).
 National Treatment: The Legal Discipline 251

opportunistic manner). Both the affected party and the judge are ignorant about
the true intent of the party intervening through regulatory means. The role of
the judge is, thus, to make the regulator come up with the truth. In this vein, the
judge will have to devise appropriate questions (for example, who suffers from the
regulatory intervention? why is it that foreigners suffer most?), as well as allocate
the burden of proof in a manner that will serve the judge’s function. Here the
judge will face a trade-off: if the burden of proof is allocated in accordance with
considerations of the type who is best positioned to advance information, then
the judge might be inciting a flood of complaints (since affected trading partners
will have to show adverse trade effects only). In this scenario, the probability to
commit Type I errors is probably high. If, conversely, the complainant has to
demonstrate not only effects but intent as well, then it might be dissuaded to
complain in the first place (and, thus, the judicial system might be committing
more Type II errors than otherwise). There are of course variations: one could,
for example, request that the complainant shows why trade effects are indicative
of protectionist intent, or why they are not evenly distributed between domestic
and foreign producers, before shifting the burden of proof to the regulator.¹⁰⁴
Note that even if the burden shifts to the regulating state (assuming, for example,
that all the complainant will have to demonstrate is discriminatory effects, in
the sense that imported products are burdened disproportionately from the chal-
lenged measure), the regulating government will still have an easier burden to
meet under Art III, as compared to Art XX of the GATT: it will not have to meet
the necessity requirement, for example, allocation of burden of proof within Art
III of the GATT, thus, is not the same as allocation of burden of proof within a
combined Art III/XX of the GATT forum. WTO case-law has failed to come up
with a meaningful test so far. Research is only starting to deal with this issue in a
comprehensive manner. Still, even without formal amendment of the GATT, we
could make some progress.¹⁰⁵ With respect to DCS products, there is no issue at
all: consumers only will decide whether two goods are substitutes. With respect to
like products, an issue could arise since Art III.2, first sentence, does not explicitly
refer to the SATAP requirement. The judge would thus have to choose between
two alternatives:
(1) overcome the absence of explicit reference to Art III.1 of the GATT, and
interpret the in excess requirement in a contextual manner (that is, in light of
the object and purpose of Art III of the GATT as embedded in Art III.1 of the
GATT), that is, once an arithmetic difference has been established, he/she will

¹⁰⁴ The next question will, of course, be which proxies are appropriate in this quest. There are
some natural candidates, such as trade effects, necessity. Research is taking its first steps in iden-
tifying additional elements. Proxies will be necessary for the judge to verify the intentions of the
intervening state. On this score, see Dixit (2004, pp 28ff ).
¹⁰⁵ Indeed, the two authors, like all authors writing on this score, have not argued that the cur-
rent wording of Art III of the GATT is the reason for the non-linear case-law that we observe in
this context.
252 Domestic Instruments

ask the additional question whether taxation in excess serves a public policy pur-
pose or not. In the latter case, the WTO member taxing an imported good in
excess of a domestic like good, will be violating its obligations under Art III.2 of
the GATT; or
(2) interpret like products in a manner that controls for the regulatory pur-
pose.¹⁰⁶ In this construction, tariff classification is not the only property, add-
itional to being DCS, that two products must share in order to be like. The pair of
products (domestic, imported) must be like not in the eyes of consumers but also
in the eyes of the government,¹⁰⁷ that is, they must be equally equipped to achieve
the regulatory purpose that the regulating government intends to achieve.¹⁰⁸
Irrespective of the method chosen, the key is that, when entertaining a complaint
under Art III of the GAT, adjudicating bodies will give meaning to the words so
as to aff ord protection, and less favourable treatment in accordance with the over-
arching objective of Art III of the GATT, embedded in its first paragraph.¹⁰⁹
New generation agreements (such as the TBT, and the SPS, which we discuss
infra in this chapter) have added some proxies (such as, the obligation to base
interventions on science, the obligation to adopt coherent policies, the obligation
to adopt the least restrictive for international trade policies) to non-discrimina-
tion in order to address the asymmetry of information between regulator and
affected parties and provide some checks to evaluate whether a given intervention
is protectionist or not. Although WTO members do not have to observe such
obligations under Art III of the GATT, the judge could still have a better feeling
about the motives of the regulator by inquiring into such issues. Assuming that
de facto a regulatory intervention which does not have to be based on scientific
proof has done so (as was the French law in EC—Asbestos), the judge will be on
safer grounds when pronouncing in favour of the regulator.
The GATT panel report on US—Taxes on Automobiles went some way towards
this direction. It fell short, nevertheless, of explaining why recourse to proxies
was necessary and which proxies are appropriate to establish protectionist intent.

¹⁰⁶ Th is interpretation assumes that the absence of explicit reference to Art III.1 of the GATT
was conscious since like products control for such concerns. Similar thoughts have been expressed
in Hudec (1998), Howse and Regan (2000), and Horn and Weiler (2003).
¹⁰⁷ This means that econometric indicators will not do when it comes to defining whether two
products are like, for even the highest number we might obtain running a cross-price elasticity test
does not guarantee that a set of products are like in the eyes of the government: think for example,
of a case where there is asymmetry of information between the government and consumers.
¹⁰⁸ It follows, that, for Art III.4 of the GATT purposes as well, some products will be considered
DCS, and some like.
¹⁰⁹ Lawyers, with an obsession for dictionary inspired textual interpretation, will object to this
analysis, arguing that the wording of Art III.1 of the GATT points to a best-endeavours clause,
not a legally binding obligation. Fair enough. The suggested approach here, nevertheless, is not an
interpretation of Art III.1 of the GATT; it is a contextual interpretation of Art III.2 and III.4 of the
GATT. It is the overarching objective embedded in the first paragraph that provides the compass
to guide the key terms in the second and fourth paragraphs to the safe harbour of understanding
the disciplines with respect to domestic instruments as intended by the founding fathers.
 National Treatment: The Legal Discipline 253

Crucially, it suffers from the following weakness: Art III of the GATT has two
legs, a likeness leg and a SATAP leg. If all regulatory distinctions across like (DCS)
products were illegal, there would be no need for the legislator to narrow down the
scope of illegality to cases where protection has been aff orded only. It must logic-
ally be the case that many regulatory distinctions across like (DCS) products are
perfectly legitimate, and the legislator wished to outlaw only those that afford
protection to domestic producers. The aims and eff ect test overextends the scope
of like products (since regulatory intent becomes part of the likeness test), and
makes the SATAP leg redundant.
The EC—Asbestos report made a mockery out of the consumers’ reactions,
and drove the LFT requirement to redundancy. The two products (construction
material with chrysotile, and with PCG fibres) were, in the eyes of consumers
substitutes. The French government intervened precisely because consumers
had not internalized the health externality. It did not intervene, however, in
all likelihood, in order to protect domestic producers’ welfare. Indeed French
producers as well incurred adjustment costs. France did not afford imported
products treatment less favourable than that afforded to domestic products; it is
under this requirement that the discussion in EC—Asbestos should have taken
place. Moreover, its law was based on sound scientific expertise; science is uni-
versal, not local. France should have then been exonerated because, by making
a regulatory distinction across two (in the eyes of consumers) DCS products, it
did not afford imported products treatment less favourable than that accorded
to domestic products.
Finally, there is one more reason why current case-law is unsatisfactory. It has
failed to discuss in a comprehensive manner the destination principle. As the
quoted discussions at the working party on Border Tax Adjustments show, there
was no agreement across trading partners as to which regulatory regime should
prevail during some trade transactions. The working party dealt with fiscal issues
only. The discussion is even richer when we move to non-fiscal elements: should,
for example, the European Community be allowed to ban imports from a devel-
oping country because the latter practices child labour? And what if the country
at hand had paid the European Community through a concession that it made,
in order to get the right to export footballs at a low tariff rate, that it cannot
export now because it is being accused of child labour? We will revert to this dis-
cussion during our presentation of the case-law under Art XX of the GATT, since
it is there that these issues were presented squarely before the WTO adjudicating
bodies. It is, nevertheless, perfectly possible that such issues could come in the
context of litigation under Art III of the GATT, as well. As we will see infra, the
AB has failed to name some obvious benchmarks for the allocation of jurisdic-
tion across trading nations. To make matters worse, it is probably the case that
existing legal tools are not perfectly well-equipped to deal with trade cases that
we increasingly experience.
254 Domestic Instruments

4.4 General exceptions

4.4.1 Coverage
GATT, Art XX contains a list of measures ranging from public morals to pro-
duction restrictions, which can function as exceptions to various WTO obliga-
tions.¹¹⁰ Case-law has consistently understood the list of measures included in
Art XX of the GATT to be of exhaustive character.¹¹¹ As we will see in more
detail infra, the various measures included in this provision are associated with
different tests for compliance. No matter which sub-paragraph has been invoked,
however, the WTO member concerned must, in accordance with the chapeau of
Art XX of the GATT, make sure that its measures:
are not applied in a manner which would constitute a means of arbitrary or unjustifiable
discrimination between countries where the same conditions prevail.

4.4.2 An exception to which provisions?

GATT, Art XX does not explicitly state the GATT obligations to which it is an
exception. Its title (General Exceptions) leaves, in principle, little room for doubt
that the various paragraphs included in this provision are legal exceptions to the
obligations assumed under the GATT. It should be, consequently, incumbent
upon the WTO member invoking this provision to carry the burden of proof to
demonstrate that its measures can justifiably come under Art XX of the GATT
Case-law has consistently confirmed this point.
There should be no doubt that Art XX of the GATT ought to be an exception
to measures which are discriminatory, in the sense that they apply to say only
foreign, or only domestic products.¹¹² As such, it could serve as justification for
deviations from obligations assumed with respect to trade instruments, such as a
quota (Art XI of the GATT): a WTO member could, for example, legitimately

¹¹⁰ Some of the measures appearing under Art XX of the GATT can be qualified as state contin-
gencies. For example, Art XXa of the GATT deals with public morals, and can hardly be described
as state contingency. On the other hand, Art XXb of the GATT could be a state contingency:
assuming, for example, that by reason of environmental hazard some goods have been infected, a
WTO member can take exceptional measures and, assuming adherence to the test established by
Art XX(b) of the GATT, lawfully impose a QR. The inclusion of all measures covered by Art XX of
the GATT in this chapter is thus, in part, abusive. It is justified only by the fact that the legal test is
at least in part (chapeau) the same across the various sub-paragraphs of Art XX of the GATT.
¹¹¹ See, for example, the AB report on US—Shrimp, § 157.
¹¹² Th is seems to have been the original intent of the negotiators in both the London Conference
and the New York Conference. It seems that what the UK negotiator (one of the architects of this
provision) had in mind was to include a provision which would operate as exception to import and
export restrictions only, and not to internal measures: ‘The undertakings in Chapter IV of this
Charter relating to import and export restrictions shall not be construed to prevent the adoption or
enforcement by any member of measures for the following purposes, provided that they are not
applied in such a manner as to constitute a means of arbitrary discrimination between countries
where the same conditions prevail, or a disguised restriction of international trade’. See E/PC/T/C.
II/50, pp 7ff, and E/PC/T/C.II/54, pp 32ff (emphasis added).
 National Treatment: The Legal Discipline 255

impose a trade embargo on imports of a commodity which is polluting envir-

onment.¹¹³ By the same token, Art XX of the GATT could serve as justifying
grounds for breaching a contractual promise to apply, say, a 10% import tariff on
wheat, if the wheat at hand has been infected by disease.
The problem arises with respect to those GATT provisions that reflect the non-
discrimination obligation, such as Art III of the GATT. Can Art XX of the GATT
be construed as an exception to the GATT provisions espousing non-discrimination?
There is tension between the title of Art XX of the GATT (General Exceptions) on the
one hand, and its chapeau, which espouses itself the non-discrimination principle,
on the other. How, in other words, can a measure found to violate the non-discrimi-
nation obligation be justified through recourse to Art XX of the GATT, which, itself
reflects the same principle? It simply cannot, as a matter of logic, be ever the case.
There is no doubt however, that, as a matter of positive law, this is indeed the
case: the AB, in its report on Korea—Various Measures on Beef proceeded to
examine whether a Korean measure (separation of outlets) found to be inconsist-
ent with Art III.4 of the GATT could still be justified through recourse to Art
XX(d) of the GATT; later, in its report on EC—Asbestos, the AB did not exclude
the possibility that Art XX of the GATT can serve as legal basis to justify meas-
ures found to be inconsistent with Art III of the GATT (§ 115). Consequently, it
is by now settled that the measures included in Art XX of the GATT can operate
as an exception to violations of Art III of the GATT (non-discrimination).
Case-law has proceeded to discuss a measure under Art XX of the GATT, with-
out, however, offering an explanation why the test of consistency under Art XX of
the GATT is any different from that under Arts I or III of the GATT. There is one
exception: the AB report on US—Gasoline. There the AB held that (p 21):
The exceptions listed in Article XX thus relate to all of the obligations under the General
Agreement: the national treatment obligation and the most-favoured-nation obligation,
of course, but others as well. (original emphasis)
It explained its rationale for doing that (pp 20–1):
The enterprise of applying Article XX would clearly be an unprofitable one if it involved
no more than applying the standard used in finding that the baseline establishment rules
were inconsistent with Article III:4. Th at would also be true if the finding were one of
inconsistency with some other substantive rule of the General Agreement. The provisions
of the chapeau cannot logically refer to the same standard(s) by which a violation of a sub-
stantive rule has been determined to have occurred. To proceed down that path would be
both to empty the chapeau of its contents and to deprive the exceptions in paragraphs (a)

¹¹³ It would, of course, have to stop importing all similar imports (like products), otherwise it
would be violating the chapeau of Art XX of the GATT. There are only two discriminatory QRs
which are permissible: those imposed on balance of payments grounds (Art XIV of the GATT),
and those imposed on national security grounds (Art XXI of the GATT), the latter, however, being
a matter of interpretation (it could be argued that the same conditions do not prevail across one
country threatening and one not threatening the country imposing the trade restrictive measure).
256 Domestic Instruments
to (j) of meaning. Such recourse would also confuse the question of whether inconsist-
ency with a substantive rule existed, with the further and separate question arising under
the chapeau of Article XX as to whether that inconsistency was nevertheless justified.
One of the corollaries of the “general rule of interpretation” in the Vienna Convention
is that interpretation must give meaning and effect to all the terms of a treaty. An inter-
preter is not free to adopt a reading that would result in reducing whole clauses or para-
graphs of a treaty to redundancy or inutility. (original emphasis)
So the AB affirms that the test across the two provisions is different, but where
exactly does the difference lie? When the AB moves to discuss the specific
measures that were being challenged, it offered nothing, beyond additional affir-
mations, to justify its view that the test for discrimination under Art XX of the
GATT is any different than that under Art I or Art III of the GATT (p 26):
We have above located two omissions on the part of the United States: to explore
adequately means, including in particular cooperation with the governments of Venezuela
and Brazil, of mitigating the administrative problems relied on as justification by the
United States for rejecting individual baselines for foreign refiners; and to count the costs
for foreign refiners that would result from the imposition of statutory baselines. In our
view, these two omissions go well beyond what was necessary for the Panel to determine
that a violation of Article III:4 had occurred in the first place.
Why could not, these two omissions, be discussed under Art III.4 of the GATT?
Is it, in the AB’s view, the case that WTO members cannot legitimately impose
adjustment costs (here, costs resulting from the imposition of statutory baselines)
under Art III.4 of the GATT, but not so, under Art XX of the GATT? Non-
discrimination, by definition, entails that adjustment costs will be imposed on
trading partners who, unless if they adjust, will not be in a position to export
to the market with the higher say environmental protection. If the US govern-
ment has enacted statutory deadlines that will serve it in reaching its regulatory
objective, then it will be obliged, under Art III.4 of the GATT, to make sure that
it applies the same regime to both domestic and foreign products. This regime
might entail adjustment costs for foreign producers. It entails similar costs for
domestic producers as well, since they cannot produce and sell in the US market
in disrespect of the statutory baselines. On the other hand, the term ‘arbitrary or
unjustifiable’ does not point to a sub-set of all discrimination. It cannot be the
case that some discrimination can be tolerated under Art XX of the GATT, but
not under Art III of the GATT: in this case, it is the same AB which sees less of a
duty not to discriminate for WTO members under Art XX (as compared to Art
III of the GATT), when Art XX of the GATT is supposed to serve as an excep-
tion to an otherwise rule. Subsequent case-law confirms that the AB sees a non-
discrimination obligation under the chapeau which does not look any different
from the obligation embedded in Art III of the GATT. Moreover, subsequent
(to US—Gasoline) case-law takes it for granted that Art XX of the GATT is an
exception to Art III of the GATT, but it offers no explanation why this should
 National Treatment: The Legal Discipline 257

be the case. There are good arguments, nevertheless, that this should not in fact
hold. Recall that, whereas the coverage of Art III of the GATT is not specified,
that of Art XX of the GATT is; the latter is narrower than the former: luxury
taxes, for example, cannot come under Art XX(a) of the GATT, or any other
item included in Art XX of the GATT. Are luxury taxes hence, per se illegal?
Interpreting the GATT like this amounts to understanding it as an instrument
mandating deregulation which, as noted supra, is not consonant with its intended
function. US—Shrimp was the first case after US—Tuna (Mexico) that, essen-
tially, outlawed any form of regulatory diversity, as we will see infra. This probably
explains the cautious attitude of the US lawyers not to challenge the complaint
arguing that the dispute was a genuine Art III of the GATT case. There should be
little doubt, nevertheless, that similar cases will be handled under Art III of the
GATT nowadays, following EC — Asbestos.¹¹⁴
GATT, Art XX has been wrongly construed as an exception to the non-
discrimination obligation: a measure in violation of Art III of the GATT, which
is not included in the exhaustive list of Art XX of the GATT, can never be jus-
tified under Art XX of the GATT, simply because it does not figure in its list.
Conversely, a measure, which is in violation of Art III of the GATT, and has
been included in the exhaustive list of Art XX of the GATT, cannot be justified
through recourse to Art XX of the GATT either: such measure will have been
found to be discriminatory (otherwise no violation of Art III of the GATT can
be established in the first place). A discriminatory measure violates Art XX of the
GATT as well, and more precisely, its chapeau. Consequently, a rational under-
standing of the GATT, as it now stands, would require that:
(1) either, we interpret the two provisions (Arts III and XX) as independent
from each other. In this vein, Art III of the GATT should be understood as a one
stop shop: we increase the burden of persuasion upfront (when the consistency of
a measure is challenged under Art III of the GATT), and request that for a meas-
ure to be judged that it operates so as to afford protection, proof will be required
to the effect that the intent and the eff ect of the measure was to protect domestic
producers at the disadvantage of its foreign competitors;
(2) or, we understand the two provisions to form part of a logical continuum.
In this vein, assuming we have espoused an intent and eff ect test, the plaintiff will
be requested to provide the eff ects-related information, and, upon successful dem-
onstration, the burden of production (and persuasion) will shift to the defendant
(the regulating state) to counteract the submitted evidence by demonstrating that
the intent was to promote public order (ordre public).

¹¹⁴ In US—Shrimp (Art 21.5—Malaysia), for example, the AB exonerated the United States
from any liability when it agreed to offer to all WTO members the same advantage it had previ-
ously offered to Caribbean states only (an opportunity to negotiate).
258 Domestic Instruments

The second approach¹¹⁵ suffers from the exhaustive character of the current list
embedded in Art XX of the GATT. It might act as a strain on regulatory freedom
if the regulating state aims at an objective not included in the current list. On the
other hand, it presupposes that protection has been afforded, otherwise recourse
to Art XX of the GATT is not an issue: but then, protection will be deemed to
have been afforded any time trading partners suffer adverse effects, irrespective
of the rationale for the regulation. This could lead to zealous over-enforcement
(since the burden under Art III of the GATT would be reduced to the minimum),
and, consequently, lead to increased likelihood to commit Type I errors.
The first approach has the merit of reducing administrative costs (by raising
the burden of persuasion, it discourages potential plaintiffs to sue). The downside
might be over-deterrence assuming allocation of the burden of proof to the effect
that the plaintiffs must show both effects and (protectionist) intent on behalf of
the regulating state. True, as argued above, adherence to this approach carries
the risk for increased Type II errors (since the burden of proof bestowed on the
complainant will be substantially higher than it is today). And yes, we express a
preference here to commit Type II rather than Type I errors, since the legislation is, at
least on its face, origin-neutral.¹¹⁶ The next step will be to design mechanisms that
will help the judge reduce such errors. Allocation of the burden of proof will have
to be rethought and, ideally, it should take into account the potential of parties
to the dispute to provide the information required, assuming that the function
of litigation is to arrive at the truth. The current transparency obligation covers
measures of general obligations only, and affected traders risk being uninformed
about other more specific measures (Art X of the GATT). Along these lines, one
would eventually need to institutionalize recourse to proxies: in an asymmetry
of information context, it might be a quixotic test to request from the plaintiff to
show the regulatory intent of the intervening state. Such private information rests
with the regulator, the only revelation mechanism being the transparency obliga-
tion reflected in Art X of the GATT. Upon demonstration of proxies, such as, for
example, a least restrictive alternative, the burden of proof could shift to the other
party, the regulating state to show that its intent was not to protect. The remedies
would have to be designed in accordance with the persuasiveness of the response,
assuming that the complainant has met its burden of proof. A lot of the case-law
under Art XX of the GATT discussed infra, would be of direct relevance in this
context.¹¹⁷

¹¹⁵ Effectively, this approach amounts to adding Art XX of the GATT to all GATT obligations
to which it acts as an exception.
¹¹⁶ Ideally, of course, no errors should be committed. But we do live in a second best world, do
we not?
¹¹⁷ The extreme textualist will observe that there is no necessity requirement included in the
current version of Art III of the GATT. Fair enough, and probably an amendment to this effect
would solve the issue in a more elegant manner. However, what stops the judge from recourse to
such proxies in his/her quest to define whether protection has been afforded? How will the judge be
persuaded that the design, architecture of a domestic legislation is (not) such to afford protection?
 National Treatment: The Legal Discipline 259

As things stand, however, Art XX of the GATT is being routinely construed as

an exception to Art III of the GATT, and we will proceed on this understanding
in order to present the case-law in what follows.

4.4.3 Regulatory diversity

During the GATT years, two panels outlawed legislation aimed at protect-
ing dolphins caught in the nets of tuna fishermen, simply because the meas-
ure was a unilateral one. The United States had enacted legislation which
required from all fishermen (domestic and foreigners alike), wishing to sell tuna
in the US market, to observe a regulation that required them to fish for tuna
using purse seine nets: this form of fishing for tuna guaranteed that the acci-
dental taking of dolphins would be highly reduced. The danger of accidental
dolphin taking when fishing for tuna was particularly high in some areas where
tuna and dolphin swim together. Mexico is one of these areas, and protested
that the measure at hand amounted to an unjustifiable restriction in violation
of Art XI of the GATT. The GATT panel, having established a violation of Art
XI of the GATT, rejected the argument advanced by the United States that the
measure was consistent with Art XX of the GATT: in its view, this could never
be the case when measures are enacted unilaterally and applied so as to pre-
judge behaviour outside of the jursidictional limits of the regulating state. The
panel, in its report on US—Tuna (Mexico) outlawed the US measure because of
its unilateral character (§ 5.27):
The Panel considered that if the broad interpretation of Article XX(B) suggested by the
United States were accepted, each contracting party could unilaterally determine the
life or health protection policies from which other contracting parties could not deviate
without jeopardizing their rights under the General Agreement. The General Agreement
would then no longer constitute a multilateral framework for trade among all contracting
parties but would provide legal security only in respect of trade between a limited num-
ber of contracting parties with identical internal regulations.
The AB, in its landmark case-law on US—Shrimp, reversed this case-law by
making it clear that measures are not, merely because of their unilateral char-
acter, inconsistent with Art XX of the GATT. In the absence of transfer of
sovereignty to the international plane, WTO members remain free unilat-
erally to regulate their market, provided that they respect the relevant GATT

By opening this door (to investigate the design, architecture of a domestic legislation) through the
case-law examined above, the WTO judge opened the door to designing proxies that will help
detect the true intent of a given legislation. What is argued here is that this quest should not be lim-
ited only to cases where a tax differential across DCS products is more than de minimis but not large
enough, so as to constitute by itself proof enough that protection has been afforded. Th is approach
is consistent with the approach developed in section 4.3 and can either apply to all transactions
coming under Art III of the GATT, or be limited to DCS products only (the test for like products,
being designed along the lines argued in that section).
260 Domestic Instruments

disciplines (§ 121):
. . . conditioning access to a Member’s domestic market on whether exporting Members
comply with, or adopt, a policy or policies unilaterally prescribed by the importing Member
may, to some degree, be a common aspect of measures falling within the scope of one or
another of the exceptions (a) to (j) of Article XX. Paragraphs (a) to (j) comprise measures
that are recognized as exceptions to substantive obligations established in the GATT 1994,
because the domestic policies embodied in such measures have been recognized as import-
ant and legitimate in character. It is not necessary to assume that requiring from exporting
countries compliance with, or adoption of, certain policies (although covered in principle
by one or another of the exceptions) prescribed by the importing country, renders a measure
a priori incapable of justification under Article XX. Such an interpretation renders most, if
not all, of the specific exceptions of Article XX inutile, a result abhorrent to the principles of
interpretation we are bound to apply. (original emphasis)
This case-law has been consistently reproduced ever since.¹¹⁸ As a result, it is now
uncontested as a matter of jurisprudential finding that Art XX of the GATT con-
dones and does not suppress regulatory diversity.

4.4.4 A two-tier test for compliance with Art XX of the GATT

The AB has constructed Art XX of the GATT akin to a two-tier test, whereby the
legal benchmark for the substantive conformity of a measure with Art XX of the
GATT is provided by the sub-paragraph invoked, whereas compliance with the
chapeau of Art XX of the GATT ensures that a national measure is applied in a
GATT consistent manner. The AB stated as much in its report on US—Gasoline
(p 22):
. . . In order that the justifying protection of Article XX may be extended to it, the meas-
ure at issue must not only come under one or another of the particular exceptions—para-
graphs (a) to (j)—listed under Article XX; it must also satisfy the requirements imposed
by the opening clauses of Article XX. The analysis is, in other words, two-tiered: first, pro-
visional justification by reason of characterization of the measure under XX(g); second,
further appraisal of the same measure under the introductory clauses of Article XX.
In its report on US—Shrimp, the AB provided the rationale for the two-tier
approach in the following manner (§§ 119 and 120):
The sequence of steps indicated above in the analysis of a claim of justification under
Article XX reflects, not inadvertence or random choice, but rather the fundamental
structure and logic of Article XX. . . .
The task of interpreting the chapeau so as to prevent the abuse or misuse of the specific
exemptions provided for in Article XX is rendered very difficult, if indeed it remains pos-
sible at all, where the interpreter (like the Panel in this case) has not fi rst identified and
examined the specific exception threatened with abuse.

¹¹⁸ See, for example, US—Shrimp (Art 21.5—Malaysia) in §§ 137–8.

 National Treatment: The Legal Discipline 261

The chapeau thus functions as a fixed element that should be added (and must
be complied with) to each and every sub-paragraph included in Art XX of the
GATT. The various sub-paragraphs of Art XX of the GATT on the other hand,
reflect divergent legal tests that serve as benchmarks in order to evaluate the con-
sistency of a particular measure: Art XX(b) of the GATT requires that measures
be necessary to protect human health, whereas Art XX(g) of the GATT requires
that they must relate to the objective pursued. Quoting from the AB report on
US—Gasoline (pp 17–18):
. . . In enumerating the various categories of governmental acts, laws or regulations which
WTO Members may carry out or promulgate in pursuit of differing legitimate state pol-
icies or interests outside the realm of trade liberalization, Article XX uses different terms
in respect of different categories:
“necessary”—in paragraphs (a), (b) and (d); “essential”—in paragraph (j); “relating
to”—in paragraphs (c), (e) and (g); “for the protection of”—in paragraph (f); “in pursu-
ance of”—in paragraph (h); and “involving”—in paragraph (i).
It does not seem reasonable to suppose that the WTO Members intended to require,
in respect of each and every category, the same kind or degree of connection or relation-
ship between the measure under appraisal and the state interest or policy sought to be
promoted or realized.
In what follows, we discuss some of the sub-paragraphs included in Art XX of the
GATT, those that gave rise to concerns about their ambit.

4.4.5 Public morals (Art XX(a) of the GATT)

4.4.5.1 Coverage
WTO members can adopt measures necessary¹¹⁹ to protect their public morals. The
term ‘public morals’ has not been interpreted in case-law. Feddersen (1998) points
to the preparatory work of the GATT, where it is made clear that the negotiating
partners had in mind a narrow construction of the term (as opposed to some form
of generic public order type of exception). Recourse to travaux préparatoires is of
course informative and, probably, warranted on most occasions.¹²⁰ Charnovitz
(1998) alludes to a wider understanding of the term ‘public morals’ and constructs
examples where, through recourse to Art XX(a) of the GATT, WTO members

¹¹⁹ We discuss necessity in section 4.4.6.

¹²⁰ Following the travaux préparatoires is a different issue. Good arguments could be advanced
in order to occasionally avoid an obligation to follow an agreement which took place years ago (and
probably such arguments explain the placing of the travaux préparatoires in the VCLT system): it
could be, for example, that following a historic interpretation leads to obvious mistakes. Scalia
(1997) mentions the dangers stemming from the adoption of this approach. Or, it could be that
only few of the current members participated in the negotiation, and hence the negotiating history
does not correspond to the wish of all contractual partners. Or, even that the travaux préparatoires
do not point to a particular result: negotiators preferred an ‘incomplete’ contract (to be completed
through recourse to adjudication) to no contract at all. Hence, it is not surprising that recourse to
travaux préparatoires can certainly be informative, its value however, for good reasons, needs to be
contained.
262 Domestic Instruments

can promote social values. His approach, although not consonant either with a
textual or with a historic interpretation of the term, is probably the only one which
can resist the de-regulatory pressure resulting from the current (AB) understand-
ing and construction of the legal terms appearing in Ar III of the GATT.¹²¹
GATT, Art XX(a) does not define the permissible territorial scope for action;
presumably, it is through recourse to public international law that disputes on this
score will be resolved, an issue to which we will return later in this chapter.¹²²

4.4.5.2 The necessity requirement

See the discussion infra under sections 4.4.6.2 and 4.4.7.2.

4.4.6 Human, animal and plant life (Art XX(b) of the GATT)
4.4.6.1 Coverage
GATT, Art XX(b) covers measures aimed at protecting human, animal, and
plant life and health. There is thus an overlap with the coverage of the TBT and
SPS Agreements. By virtue of law and case-law, as we will see infra, measures fall-
ing simultaneously under the GATT, the TBT, and the SPS will be scrutinized
under the latter. Measures falling under the GATT and the TBT will be scruti-
nized under the TBT. Hence, Art XX(b) of the GATT is the default provision, to
the extent that the TBT and SPS Agreements are not applicable.

4.4.6.2 The necessity requirement

For a measure to be consistent with Art XX(b) of the GATT, it must be necessary
to achieve the ends pursued. The necessity requirement implies that there is no
other measure (other than the one privileged) which is less restrictive (in terms
of impact on international trade transactions), and which can achieve the same
result. The AB made this point clear in § 172 of its report on EC—Asbestos:
. . . In this case, the objective pursued by the measure is the preservation of human
life and health through the elimination, or reduction, of the well-known, and life-
threatening, health risks posed by asbestos fibres. The value pursued is both vital and
important in the highest degree. The remaining question, then, is whether there is an
alternative measure that would achieve the same end and that is less restrictive of trade
than a prohibition.

¹²¹ Wisely, Art XIV of the GATS, which is the corresponding provision to Art XX of the GATT,
uses the term ‘public order’ in lieu of public morals.
¹²² On the relevance of public international law in general, see Pauwelyn (2003), and Trachtman
(1999). An interesting case concerning the allocation of jurisdiction arose in the European
Community when the United Kingdom refused to grant export licenses to producers interested
in exporting livestock to Spain since, in its view, Spain had not eliminated the potential for cruel
treatment of animals in its slaughterhouses (Case C–5/94 The Queen v Ministry of Agriculture,
Fisheries and food ex p Hedley Lomas (Ireland) Ltd (ECJ, 23 May 1996). The dispute was resolved
on technical grounds, but the interesting question is to what extent public morals can provide an
exception not only to import- but also to export-related measures.
 National Treatment: The Legal Discipline 263

The necessity test thus involves a relationship between two elements: a common
(fi xed) element, which is trade liberalization, and a variable element, which is the
objective pursued by the regulating WTO member. Recall that, with respect to
Art XX of the GATT, the list of objectives that can legitimately be pursued is
exhaustively reflected in this provision.
GATT, Art XX contains two implicit working hypotheses: first, that trade
liberalization yields to any one of the objectives included in this provision. Trade
liberalization is, from a hierarchy perspective, accorded a lower value than any of the
objectives included in Art XX of the GATT; second, the recognition that a WTO
adjudicating body cannot put into question the legitimacy of the ends sought (a nat-
ural consequence of the legitimacy of regulatory diversity condoned by Art XX of
the GATT and recognized by the AB), but can extend its judicial review only with
respect to the means used to achieve the ends: ends are not justiciable, means are.
It stems from the above that, crucial to the fulfillment of the necessity require-
ment, is that the alternative measure is as efficient as the one chosen: adjudicating
bodies cannot impose on a regulating WTO member a less efficient, but probably
less restrictive, measure, since it is up to WTO members unilaterally to reveal
their preferences and formulate the ends sought.¹²³
The AB distinguished, when interpreting the term ‘necessary’ in connection
with another sub-paragraph (Art XXd of the GATT), between a measure that is
indispensable to achieve the member’s objective, and one that is not indispensable,
but that bears a close relationship to the objective. The AB held that a measure
need not always be indispensable in order to be necessary within the meaning of
Art XX of the GATT; a measure with a close relationship to the objective might
also be deemed necessary, provided it is proportional.
Also, when dealing with disputes in the context of Art XX(d) of the GATT,
the AB endorsed the view that, for the necessity requirement to be satisfied, one
should not simply point to another, less restrictive measure which is theoretically
available; the less restrictive measure should be reasonably available to the defend-
ant. These findings are discussed infra.

4.4.6.3 No science required

The AB clarified in § 178 of its report on EC—Asbestos that WTO members
could of course have recourse to scientific expertise to justify a regulatory inter-
vention. In such a case, they did not have to abide by the majority opinion in
the relevant scientific field. Nevertheless, recourse to scientific expertise is not a
necessary requirement for consistency of a national measure with Art XX(b) of
the GATT.

¹²³ Mattoo and Mavroidis (1997) advance some thoughts trying to see how much the necessity
requirement can borrow from the targeting literature to address this issue. On targeting, see the
classic piece by Bhagwati and Ramaswami (1963).
264 Domestic Instruments

4.4.7 Domestic laws and regulations (Art XX(d) of the GATT)

4.4.7.1 Coverage
GATT, Art XX(d) covers measures which aim to secure compliance with laws (or
regulations) which are GATT consistent. It does not define what a GATT con-
sistent law (or regulation) is in an exhaustive manner.¹²⁴ It does, however, provide
an indicative list to this effect: customs enforcement, enforcement of monopolies,
protection of patents, trade marks and copyrights, and prevention of deceptive
practices. Such laws (and regulations) are by legislative fiat GATT consistent.¹²⁵

4.4.7.2 The necessity requirement

The AB has provided a number of clarifications on this score over the years.
First, it made clear that the necessity requirement points to a test where the
judge can pronounce on the consistency of the means used only, leaving it to the
total discretion of WTO members to choose the ends themselves.¹²⁶ In its report
on Korea—Various Measures on Beef, the AB held as much (§ 176):
It is not open to doubt that Members of the WTO have the right to determine for them-
selves the level of enforcement of their WTO-consistent laws and regulations. We note
that this has also been recognized by the panel in United States—Section 337, where it
said: “The Panel wished to make it clear that this [the obligation to choose a reasonably
available GATT consistent or less inconsistent measure] does not mean that a contracting
party could be asked to change its substantive patent law or its desired level of enforcement
of that law . . . .” . (original emphasis)
Second, the AB, in the same report, held that necessary should not be equated to
indispensable, since, in its view, the term ‘necessary’ in Art XX(d) of the GATT
refers to a range of necessary options. The AB explained that it understands neces-
sity to be on one end of a logical continuum, and indispensability, on the other.

¹²⁴ It would have been a very high consuming enterprise to attempt something like that in light
of the existing regulatory diversity, and the need to renegotiate the contract in light of changes both
at the domestic, as well as at the multilateral level.
¹²⁵ All indicative lists serve two very important functions: first, they alert the judge so that he/
she will not commit a Type II error: the transactions covered in the indicative list will anyway come
under the purview of the law at hand; second, they inform the judge as to the class of transactions
that the legislator had in mind when drawing the list (that it did not complete). Any laws resem-
bling the subject matter of those mentioned in Art XX(d) of the GATT should come under the
purview of this provision as well.
¹²⁶ Check, however, the AB report on Korea—Various Measures on Beef, where the AB held
(§ 162) that: ‘… It seems to us that a treaty interpreter assessing a measure claimed to be necessary
to secure compliance of a WTO-consistent law or regulation may, in appropriate cases, take into
account the relative importance of the common interests or values that the law or regulation to be
enforced is intended to protect. The more vital or important those common interests or values are,
the easier it would be to accept as “necessary” a measure designed as an enforcement instrument’.
This implies some sort of hierarchy across the various legitimate objectives, an exercise which is at
odds with the idea that ends are not justiciable. Sykes (2003b) takes the view that this is exactly
what the AB does: it will be more deferential to WTO members if measures aimed at protecting
human life and health are at stake, and less so when other regulatory objectives are being pursued.
 National Treatment: The Legal Discipline 265

The boundaries are hard to define but it seems reasonable to conclude that, what
the AB had in mind, was to provide some leeway to the regulating WTO member
by relaxing the standard of review (§§ 161ff ). In its own words (§ 164):
In sum, determination of whether a measure, which is not “indispensable”, may never-
theless be “necessary” within the contemplation of Article XX(d), involves in every case
a process of weighing and balancing a series of factors which prominently include the
contribution made by the compliance measure to the enforcement of the law or regula-
tion at issue, the importance of the common interests or values protected by that law or
regulation, and the accompanying impact of the law or regulation on imports or exports.
(original emphasis)
It follows that measures which are not, strictly speaking, indispensable to
reach the objective pursued, might still qualify as necessary, depending on the
circumstances.
Third, the AB, in the same report, held for the proposition that, under the
necessity requirement, a WTO adjudicating body should not force WTO mem-
bers to use measures that are not reasonably available to them (§ 166):
In our view, the weighing and balancing process we have outlined is comprehended in
the determination of whether a WTO-consistent alternative measure which the Member
concerned could ‘reasonably be expected to employ’ is available, or whether a less WTO-
inconsistent measure is ‘reasonably available’.
This is probably the most decisive attempt by the AB to make sure that necessity
does not operate as a stranglehold. By endogenizing the necessity requirement,
the AB shows its willingness to respect national idiosyncratic attributes (endow-
ments). The absolutely least restrictive measure might, for example, sometimes
be a subsidy, which could be simply unaffordable to a WTO member with severe
budgetary constraints.¹²⁷ Through reasonable availability thus, the AB under-
stands that the necessity requirements cannot exclude that two different WTO
members must use different measures to comply with Art XX of the GATT while
pursuing the same ends.
Fourth, case-law has also provided clarifications as to the allocation of the bur-
den of proof in order to ensure that a measure is indeed reasonably available.
GATT/WTO case-law has consistently supported the proposition that a WTO
member should privilege a WTO consistent measure over a WTO inconsistent
measure, if both are reasonably available. There was divergent case-law, concern-
ing the issue whether it is for the panel to come up with a reasonably available
alternative, or whether it is up to the party challenging the measure to do so. The
panel on Canada—Wheat Exports and Grain Imports is an example of the former
approach, whereas the AB report on Japan—Agricultural Products II (§§ 123–30)
of the latter. This issue was resolved in the AB report on US—Gambling. There
the AB explained itself clearly as to the allocation of the burden of proof. In this

¹²⁷ Note that Art XX of the GATT does not a priori limit the choice of instruments at all.
266 Domestic Instruments

case, the AB held that the original burden of proof rests with the WTO member
raising the defence. Assuming it has made a prima facie case (that the necessity
requirement has been met), the burden of proof shifts to the other party (the ori-
ginal complainant) who will have to demonstrate that another, less restrictive (but
equally effective) measure should have been privileged. Assuming the existence
of such a measure has been confirmed, the burden of proof will shift again to the
party raising the defence under Art XX of the GATT.¹²⁸ This time, however, it
will have to demonstrate that this measure was not reasonably available to it (§§
309–11).

4.4.8 Exhaustible natural resources (Art XX(g) of the GATT)

4.4.8.1 Coverage
GATT, Art XX(g) deals with exhaustible natural resources. The AB, in its report
on US—Shrimp held that the term ‘exhaustible natural resources’ should not be
confined to non-living resources. At dispute was a US measure which banned
the import of shrimps fished in a manner that led to the accidental taking of
the life of sea turtles. Since sea turtles often swim together with shrimp, the US
legislation at hand requested from all fishermen (domestic and foreign alike) who
wanted to export shrimp to the US market, to guarantee that they have caught
them using TEDs (turtle excluding devices). The US measure banned imports of
shrimps which did not respect this requirement. Through monitoring and report-
ing requirements, the US authorities could satisfy themselves that the exporter
had indeed respected the statutory requirements when catching shrimp. The
United States argued that the measure at hand was in full conformity with the
requirements of Art XX(g) of the GATT. The question naturally arose whether
the US measure should come under the purview of Art XX(b) or Art XX(g) of the
GATT.¹²⁹ The first step in analyzing this claim is, of course, the question whether
sea turtles are indeed an exhaustible natural resource. In the view of the AB, the
term must be given a meaning in accordance with today’s perceptions.¹³⁰ It went
on to state that sea turtles, a living organism, could very well be regarded as an
exhaustible natural resource. International conventions recognizing sea turtles as
exhaustible natural resources were cited in support (without, however, prescribing

¹²⁸ US—Gambling concerned trade in services and was adjudicated under Art XIV GATS,
the provision corresponding to Art XX of the GATT. There should be no doubt, however, that the
same standard should find application in the goods trade as well.
¹²⁹ As we will see shortly, the standard for consistency with Art XXg of the GATT is less
demanding than that under Art XXb of the GATT. Hence, this was not a trivial issue, as the US
attorneys had strong interest in subjecting the measure under the less demanding standard.
¹³⁰ To make the bridge to modern reality, and interpret the term in its contemporary dimen-
sion, the AB invoked the principle of contemporaneity, an interpretative principle not mentioned
anywhere in the VCLT, the instrument that the AB has always had recourse to when interpreting
the WTO, by legal compulsion in its own admission (see Chapter 5). Moreover, this principle has
been used only infrequently by other fora such as the International Court of Justice (ICJ).
 National Treatment: The Legal Discipline 267

in a definitive manner the legal relevance of such instruments in the WTO legal
order). The defining criterion (which distinguishes exhaustible from non-exhaust-
ible natural resources) is, in the AB’s view, the response to the question whether
the item at hand is being depleted faster than it is being reproduced (§§ 128, 130
and 153):
. . . Textually, Article XX(g) is not limited to the conservation of “mineral” or “non-liv-
ing” natural resources. The complainants’ principal argument is rooted in the notion that
“living” natural resources are “renewable” and therefore cannot be “exhaustible” natural
resources. We do not believe that “exhaustible” natural resources and “renewable” nat-
ural resources are mutually exclusive. One lesson that modern biological sciences teach
us is that living species, though in principle, capable of reproduction and, in that sense,
“renewable”, are in certain circumstances indeed susceptible of depletion, exhaustion and
extinction, frequently because of human activities. Living resources are just as “finite” as
petroleum, iron ore and other non-living resources.
...
. . . From the perspective embodied in the preamble of the WTO Agreement, we note
that the generic term “natural resources” in Article XX(g) is not “static” in its content or
reference but is rather “by definition, evolutionary”. It is, therefore, pertinent to note that
modern international conventions and declarations make frequent references to natural
resources as embracing both living and non-living resources.
...
The language [of the Preamble of the WTO Agreement] demonstrates a recognition by
WTO negotiators that optimal use of the world’s resources should be made in accordance
with the objective of sustainable development. As this preambular language reflects the
intentions of negotiators of the WTO Agreement, we believe it must add colour, texture
and shading to our interpretation of the agreements annexed to the WTO Agreement, in
this case, the GATT 1994. We have already observed that Article XX(g) of the GATT
1994 is appropriately read with the perspective embodied in the above preamble. (ori-
ginal emphasis)
A number of arguments can be advanced against this view. It is highly debat-
able whether this understanding of the term ‘exhaustible natural resources’ is con-
sonant with the original understanding of the GATT founding fathers. Indeed,
a look into the travaux préparatoires of the GATT confirms that the founding
fathers had in mind items such as petrol and minerals (ie non-living organisms)
when they drafted Art XX(g) of the GATT.¹³¹ On the other hand, the rest of the
provision does not make much sense were one to understand the term as cover-
ing living organisms as well: indeed, how can a state restrict the domestic repro-
duction of sea turtles, absent recourse to extravagant devices? Additionally, the
AB takes the view that rebalancing the coverage between Art XX(b) and XX(g)
of the GATT is perfectly consistent with the VCLT prescriptions. Is this the
case? GATT, Art XX(g) reflects a consistency threshold which is less demanding
than that included in Art XX(b) of the GATT (which is the other candidate for

¹³¹ See E/PC/T/C.II/50, pp 5ff.

268 Domestic Instruments

adjudicating the present dispute): whereas the former suggests that any means
which is appropriate in order to reach a particular objective is fine, the latter exon-
erates from inconsistency only a sub-set of all appropriate measures, that is, those
that are necessary to achieve the stated ends.¹³² Is the judge respecting the balance
of rights and obligations as struck by the GATT founding fathers by modifying
the legislative standard of consistency for a particular transaction? The obligation
of the agent (judge), by virtue of Art 3.2 of the DSU, is not simply to ensure that
the total policy space as contracted by the principals (GATT founding fathers)
is not modified; it is also to ensure that the expressed wish as to the standard of
consistency of particular transactions with the negotiated edifice will be applied
as negotiated. Otherwise he/she will be undoing the balance of rights and obliga-
tions struck by the founding fathers of the treaty.¹³³

4.4.8.2 Jurisdictional limits

The AB, in its report on US—Shrimp faced, inter alia, the question whether Art XX
(g) of the GATT includes a jurisdictional limit, in the sense that WTO members
can intervene to protect exhaustible natural resources within their jurisdiction, as
the latter is defined by public international law.¹³⁴ Nationality and territoriality
are the two most common bases, in public international law, for exercising juris-
diction: under the former, a state can regulate behaviour of its nationals; under
the latter, a state can regulate transactions taking place in its territory. Public
international law practice gives the edge to the latter, when a conflict between the
two arises. On the other hand, one further distinguishes between objective and
subjective territoriality: the former captures activities occurring within national
frontiers; the latter captures activities which occur in state A, but which affect

¹³² See infra.

¹³³ The defendant in this case should have won anyway, even if the transaction had been sub-
jected under Art XXb of the GATT. Indeed, the TEDs are a very inexpensive device; one can
hardly make the argument that they are not necessary (in the sense least restrictive and reasonably
available) to fishermen around the world. Moreover, by correctly requiring additional flexibility
from the original relatively inflexible US Statute (Section 609), the AB obliged the United States
to accept shrimps fished in any manner which guarantees results comparable to those obtained
through the use of TEDs. The decision, however, to subject the transaction to Art XXg of the
GATT was a positive signal to the civic (and more specifically, the environmental) community.
Placed in its historical context (1998), the decision is a welcome change (a sea change indeed)
and a complete break with past GATT law on this issue. The GATT had been accused of being
a clique of faceless, unaccountable bureaucrats who place trade liberalization above all societal
values. Through this decision, not only does the WTO signal to the world that unilateral trade
obstructing measures can survive in the WTO, it also ensures that many of them (through the
expansive understanding of the term ‘exhaustible natural resources’) will be reviewed favourably.
By this, I do not mean to suggest that political motivations explain the decision of the AB, but
simply mean to highlight the positive externalities of this decision. This case should probably have
been adjudicated under Art III.4 of the GATT where it would have been hard to find fault with the
US measure.
¹³⁴ Since, as already noted, Art XX of the GATT does not contain a jurisdictional clause.
 National Treatment: The Legal Discipline 269

state B. B could legitimately, by virtue of subjective territoriality (effects doctrine)

punish, for example, A’s export cartels which cartelize its own (B’s) market.¹³⁵
States can, of course, extend or reduce the limits of their jurisdictional reach
by international agreement. This, however, raises the question regarding the
legal relationship between the GATT and multilateral environmental agree-
ments (MEAs), a largely unresolved issue, as a matter of WTO law.¹³⁶ In § 133
of its report on US—Shrimp, the AB avoided responding directly to the question
whether the US measure respected the territoriality principle:
. . . We do not pass upon the question of whether there is an implied jurisdictional limi-
tation in Article XX(g), and if so, the nature or extent of that limitation. We note only
that in the specific circumstances of the case before us, there is a sufficient nexus between
the migratory and endangered marine populations involved and the United States for
purposes of Article XX(g).
Implicitly, nevertheless, the AB, by ultimately justifying the US measure under
Art XX(g) of the GATT, it accepted the GATT consistency of a measure based
on subjective territoriality.¹³⁷

4.4.8.3 Measure must be relating to the objective pursued

A measure will be judged consistent with Art XX(g) of the GATT, if it is relat-
ing to the conservation of exhaustible natural resources. The AB, in its report
on US—Gasoline, noted that past case-law had constructed the term ‘relating to’
akin to primarily aimed at. The AB was not at ease with such understanding, even
though the parties to the dispute seemed to endorse it (pp 18 and 19):
. . . All the participants and the third participants in this appeal accept that a measure
must be “primarily aimed at” the conservation of exhaustible natural resources in order
to fall within the scope of Article XX(g). Accordingly, we see no need to examine this
point further, save, perhaps, to note that the phrase “primarily aimed at” is not itself treaty
language and was not designed as a simple litmus test for inclusion or exclusion from Article
XX(g). . . . (emphasis added)
This was just a signaling device. In its report on US—Shrimp, the AB had the
opportunity to express its views on this issue. Taking distance from prior practice,

¹³⁵ Indeed, except for some extreme cases, state A would have little incentive to regulate the
behaviour of an export cartel operating from its market but affecting the rest of the world. Antitrust
advocacy usually attributes such inactivity to the fact that domestic antitrust laws are there to pro-
tect domestic (as opposed to foreign) consumers’ welfare.
¹³⁶ Commentators might have legitimately thought that the door to the legal relevance of
MEAs was slightly opening following the AB US—Shrimp jurisprudence. In EC—Approval and
Marketing of Biotech Products, the panel held nevertheless, that only MEAs to which all WTO
membership is a party are legally relevant for the purposes of adjudicating WTO disputes (§§
7.67—7.72). Th is view is arguably at odds with Art 30 of the VCLT. See on this score, Palmeter and
Mavroidis (1998), Pauwelyn (2003), and Trachtman (1999).
¹³⁷ On the relevance of public international law in the WTO legal order, see the comprehensive
study by Pauwelyn (2003).
270 Domestic Instruments

it held that relating to implied a rational connection between a measure and the
conservation of exhaustible natural resources, and nothing beyond that (§ 141):
In its general design and structure, therefore, Section 609 is not a simple, blanket prohib-
ition of the importation of shrimp imposed without regard to the consequences (or lack
thereof) of the mode of harvesting employed upon the incidental capture and mortality
of sea turtles. Focusing on the design of the measure here at stake, it appears to us that
Section 609, cum implementing guidelines, is not disproportionately wide in its scope
and reach in relation to the policy objective of protection and conservation of sea turtle
species. The means are, in principle, reasonably related to the ends. The means and ends
relationship between Section 609 and the legitimate policy of conserving an exhaustible,
and, in fact, endangered species, is observably a close and real one. (original emphasis)
Arguably, the rational connection standard endorsed by the AB is more deferential
towards the regulating WTO member than the previously employed primarily
aimed standard, since even measures which do not primarily aim at the conser-
vation of natural resources can be justified through recourse to Art XXg of the
GATT, assuming that they have a rational connection with the objective stated
in this provision.

4.4.8.4 Measures restricting domestic production or consumption

GATT, Art XX(g) requires that, when imposing trade restrictions to protect
exhaustible natural resources, WTO members must also adopt measures aimed
at restricting domestic consumption or production (as the case may be). The AB,
in its report on US—Gasoline, explained that the requirement to demonstrate
that import-restricting measures are taken in conjunction with domestic meas-
ures aiming at the conservation of exhaustible natural resources is an even-hand-
edness requirement. It went on to stress that there is no need for an effects test in
order to comply with Art XX(g) of the GATT in this respect (pp 20–1):
. . . the clause “if such measures are made effective in conjunction with restrictions on
domestic product or consumption” is appropriately read as a requirement that the meas-
ures concerned impose restrictions, not just in respect of imported gasoline but also with
respect to domestic gasoline. The clause is a requirement of even-handedness in the impos-
ition of restrictions, in the name of conservation, upon the production or consumption of
exhaustible natural resources.
. . . if no restrictions on domestically-produced like products are imposed at all, and
all limitations are placed upon imported products alone, the measure cannot be accepted
as primarily or even substantially designed for implementing conservationist goals. The
measure would simply be naked discrimination for protecting locally-produced goods.
We do not believe . . . that the clause “if made effective in conjunction with restrictions
on domestic production or consumption” was intended to establish an empirical “effects
test” for the availability of the Article XX(g) exception. (original emphasis)¹³⁸

¹³⁸ An example of an even-handed measure is offered in §§ 144–5 of the AB report on

US—Shrimp.
 National Treatment: The Legal Discipline 271

This construction seems quite rational: absent a non-discrimination obligation,

WTO members could be conferring regulatory subsidies to their domestic pro-
ducers, and thus contravene the quintessential element of the GATT, that is, to
provide protection to domestic producers only through tariffs.

4.4.9 The chapeau of Art XX of the GATT

4.4.9.1 The scope of the chapeau
Recall that the test for consistency of a measure with Art XX of the GATT is a
two-tier test: following a finding on the substantive consistency of the challenged
measure with a particular sub-paragraph, adjudicating bodies will move to dis-
cuss whether a measure is also applied in conformity with the requirements of the
chapeau.¹³⁹ On p 22 of its report on US—Gasoline, the AB held that a WTO
adjudicating body cannot review the substantive consistency of a national meas-
ure with Art XX of the GATT under the chapeau of Art XX of the GATT. When
moving to the chapeau, the only remaining question is to what extent the meas-
ure at hand is applied in a GATT consistent manner:
The chapeau by its express terms addresses, not so much the questioned measure or its
specific contents as such, but rather the manner in which that measure is applied. It is,
accordingly, important to underscore that the purpose and object of the introductory
clauses of Article XX is generally the prevention of “abuse of the exceptions”.

4.4.9.2 Complying with the chapeau

The AB, in its report on US—Shrimp (Art 21.5—Malaysia), held that for a meas-
ure to be chapeau consistent, it should not (§ 118):
(1) amount to an arbitrary discrimination between countries where the same
conditions prevail;
(2) amount to an unjustifiable discrimination; and
(3) be a disguised restriction to trade either.¹⁴⁰

4.4.9.2.1 Arbitrary or unjustifiable discrimination

GATT/WTO case-law has often examined the arbitrary or unjustifiable discrim-
ination requirement in tandem, without distinguishing between its two elements
(AB report on US—Shrimp, § 150).
The AB, in its report on US—Gasoline, discusses the issue whether the term
‘between countries where the same conditions prevail’ should be understood as refer-
ring only to exporting countries or, conversely, whether it should encompass the
importing country as well. Although the AB did not formally rule on this issue
on this occasion, it saw no reason to deviate from the prevailing practice of WTO

¹³⁹ Since the two conditions must be cumulatively met, a finding that a measure has failed the
test for substantive consistency makes recourse to the second part of the test redundant.
¹⁴⁰ See also the AB report on US—Shrimp, § 150.
272 Domestic Instruments

members which privileged the latter interpretation (pp 23–4):

It was asked whether the words incorporated into the first two standards “between coun-
tries where the same conditions prevail” refer to conditions in importing and exporting
countries, or only to conditions in exporting countries. The reply of the United States
was to the effect that it interpreted that phrase as referring to both the exporting coun-
tries and importing countries and as between exporting countries. . . . At no point in the
appeal was that assumption challenged by Venezuela or Brazil.
. . . we see no need to decide the matter of the field of application of the standards set
forth in the chapeau nor to make a ruling at variance with the common understanding of
the participants.
As a result of this interpretation, a regulating state, for the purposes of compli-
ance with the chapeau, cannot treat its products any differently from the way it
treats products originating in other WTO members, provided that the condi-
tions across countries are the same.
The AB, in its report on US—Shrimp added that, actively accounting for differ-
ences among various WTO members is a necessary component of the obligation
imposed on WTO members not to discriminate in an unjustifiable or arbitrarily
manner. Flexibility becomes thus a necessary component of such measures for
compliance to be achieved under circumstances where different conditions pre-
vail among different WTO members (§§ 164–5 and 177):
. . . It may be quite acceptable for a government, in adopting and implementing a domes-
tic policy, to adopt a single standard applicable to all its citizens throughout that country.
However, it is not acceptable, in international trade relations, for one WTO Member to
use an economic embargo to require other Members to adopt essentially the same com-
prehensive regulatory program, to achieve a certain policy goal, as that in force within
that Member’s territory, without taking into consideration different conditions which
may occur in the territories of those other Members.
. . . We believe that discrimination results not only when countries in which the same
conditions prevail are differently treated, but also when the application of the measure at
issue does not allow for any inquiry into the appropriateness of the regulatory program
for the conditions prevailing in those exporting countries.
...
. . . Section 609, in its application, imposes a single, rigid and unbending requirement
that countries applying for certification . . . adopt a comprehensive regulatory program
that is essentially the same as the United States’ program, without inquiring into the
appropriateness of that program for the conditions prevailing in the exporting coun-
tries. Furthermore, there is little or no flexibility in how officials make the determination
for certification pursuant to these provisions. In our view, this rigidity and inflexibility
also constitute “arbitrary discrimination” within the meaning of the chapeau. (original
emphasis)
 National Treatment: The Legal Discipline 273

The AB had the opportunity to further explain itself during the compliance
stage¹⁴¹ of this litigation. The original US measure conditioning access of exports
of shrimps to the US market upon a particular fishing method (the use of TEDs),
was found to be in substantive compliance with Art XX(g) of the GATT; it was
found, however, to be inconsistent with the requirements of the chapeau. In the
AB’s view, the US measure should be modified to allow exports of shrimps fished
through other (than TEDs) fishing methods, of comparable to the TEDs’ effect-
iveness. Subsequent to the original condemnation of its measure, the United
States modified its statute so as to allow for certification of other fishing methods.
The AB held that this amendment brought the US measure into compliance with
its obligations in this respect (§§ 144 and 149–50).
In the same report, and in the same vein, the AB had the opportunity to
address another related issue: the United States was also found to be inconsist-
ent with the requirements of the chapeau, because it had offered international
negotiations to resolve the problems encountered by the enactment of US Section
609 to some (Caribbean countries), but not to all WTO members. Subsequently,
the United States offered the same opportunity to all other (including Malaysia,
the complainant in this case) shrimp-exporters affected by the US measure. The
AB found that, in so doing, the United States had effectively complied with its
obligations in this respect. In its view, the United States could not have been
asked to achieve through international negotiations with Malaysia an outcome
comparable to that achieved with other WTO members (as Malaysia indeed
had argued), which were prepared to accept fishing using a particular technique
which minimizes the incidental taking of sea turtles. The United States had ori-
ginally been held accountable for not offering the same opportunity to all WTO
members; having done that, it could not be held accountable for an event that did
not solely depend on its own will (§§ 122, 123 and 130):
We concluded in United States—Shrimp that, to avoid “arbitrary or unjustifiable dis-
crimination”, the United States had to provide all exporting countries “similar opportun-
ities to negotiate” an international agreement. Given the specific mandate contained in
Section 609, and given the decided preference for multilateral approaches voiced by WTO
Members and others in the international community in various international agreements
for the protection and conservation of endangered sea turtles that were cited in our previ-
ous Report, the United States, in our view, would be expected to make good faith efforts
to reach international agreements that are comparable from one forum of negotiation to
the other. The negotiations need not be identical. Indeed, no two negotiations can ever be
identical, or lead to identical results. Yet the negotiations must be comparable in the sense
that comparable efforts are made, comparable resources are invested, and comparable

¹⁴¹ As we will see in Chapter 5, assuming that a WTO member has been found to violate its
obligations, it will be granted a reasonable period of time during which it should comply with its
multilateral obligations. Disagreements among complainant and defendant whether compliance
has indeed occurred, might lead to the establishment of a compliance panel (Art 21.5 of the DSU),
the report of which can also be appealed.
274 Domestic Instruments
energies are devoted to securing an international agreement. So long as such comparable
efforts are made, it is more likely that “arbitrary or unjustifiable discrimination” will be
avoided between countries where an importing Member concludes an agreement with
one group of countries, but fails to do so with another group of countries.
Under the chapeau of Article XX, an importing Member may not treat its trading part-
ners in a manner that would constitute “arbitrary or unjustifiable discrimination”. With
respect to this measure, the United States could conceivably respect this obligation, and
the conclusion of an international agreement might nevertheless not be possible despite the
serious, good faith efforts of the United States. Requiring that a multilateral agreement be
concluded by the United States in order to avoid “arbitrary or unjustifiable discrimination”
in applying its measure would mean that any country party to the negotiations with the
United States, whether a WTO Member or not, would have, in effect, a veto over whether
the United States could fulfill its WTO obligations. Such a requirement would not be rea-
sonable. For a variety of reasons, it may be possible to conclude an agreement with one
group of countries but not another. The conclusion of a multilateral agreement requires the
cooperation and commitment of many countries. In our view, the United States cannot be
held to have engaged in “arbitrary or unjustifiable discrimination” under Article XX solely
because one international negotiation resulted in an agreement while another did not.
...
At no time in United States—Shrimp did we refer to the Inter-American Convention
as a “benchmark”. The Panel might have chosen another and better word—perhaps, as
suggested by Malaysia, "example". Yet it seems to us that the Panel did all that it should
have done with respect to the Inter-American Convention, and did so consistently with
our approach in United States—Shrimp. The Panel compared the efforts of the United
States to negotiate the Inter-American Convention with one group of exporting WTO
Members with the efforts made by the United States to negotiate a similar agreement with
another group of exporting WTO Members. The Panel rightly used the Inter-American
Convention as a factual reference in this exercise of comparison. It was all the more rele-
vant to do so given that the Inter-American Convention was the only international agree-
ment that the Panel could have used in such a comparison. As we read the Panel Report,
it is clear to us that the Panel attached a relative value to the Inter-American Convention
in making this comparison, but did not view the Inter-American Convention in any way
as an absolute standard. Thus, we disagree with Malaysia’s submission that the Panel
raised the Inter-American Convention to the rank of a “legal standard”. The mere use
by the Panel of the Inter-American Convention as a basis for a comparison did not trans-
form the Inter-American Convention into a “legal standard”. Furthermore, although the
Panel could have chosen a more appropriate word than “benchmark” to express its views,
Malaysia is mistaken in equating the mere use of the word “benchmark”, as it was used by
the Panel, with the establishment of a legal standard. (original emphasis)

4.4.9.2.2 Disguised restriction of trade

On p 25 of its report on US—Gasoline, the AB opined that the term is not limited
to concealed or unannounced restrictions only. In other words, the obligation to
avoid disguised restrictions of trade is not a mere transparency exercise:
 National Treatment: The Legal Discipline 275
“Arbitrary discrimination”, “unjustifiable discrimination” and “disguised restriction” on
international trade may, accordingly, be read side-by-side; they impart meaning to one
another. It is clear to us that “disguised restriction” includes disguised discrimination in
international trade. It is equally clear that concealed or unannounced restriction or dis-
crimination in international trade does not exhaust the meaning of “disguised restric-
tion.” (original emphasis)
That much should be obvious. On p 25 of the same report, the AB went on and
provided the framework for analysis of the term. In its view, through this term,
the founding fathers of GATT wanted to outlaw abusive invocations of Art XX of
the GATT:
. . . It is equally clear that concealed or unannounced restriction or discrimination in inter-
national trade does not exhaust the meaning of “disguised restriction.” We consider that
“disguised restriction”, whatever else it covers, may properly be read as embracing restric-
tions amounting to arbitrary or unjustifiable discrimination in international trade taken
under the guise of a measure formally within the terms of an exception listed in Article
XX. Put in a somewhat different manner, the kinds of considerations pertinent in decid-
ing whether the application of a particular measure amounts to “arbitrary or unjustifiable
discrimination”, may also be taken into account in determining the presence of a “dis-
guised restriction” on international trade. The fundamental theme is to be found in the
purpose and object of avoiding abuse or illegitimate use of the exceptions to substantive
rules available in Article XX. (original emphasis)
This view reminds us of the French doctrine of abus de droit. It has been repro-
duced as such in subsequent case-law. The panel in its report on EC—Asbestos, for
example, held the view that the term should be understood as outlawing inter-
ventions, which, under the guise of protection of one of the grounds mentioned
in the sub-paragraphs of Art XX. of the GATT, aim at promoting other interests
(§ 8.236):
Referring also to the remark made by the Appellate Body in the same case according
to which “the provisions of the chapeau [of Article XX] cannot logically refer to the
same standard(s) by which a violation of the substantive rule has been determined to
have occurred”, we consider that the key to understanding what is covered by “disguised
restriction on international trade” is not so much the word “restriction”, inasmuch as,
in essence, any measure falling within Article XX is a restriction on international trade,
but the word “disguised”. In accordance with the approach defined in Article 31 of
the Vienna Convention, we note that, as ordinarily understood, the verb “to disguise”
implies an intention. Thus, “to disguise” (déguiser) means, in particular, “conceal beneath
deceptive appearances, counterfeit”, “alter so as to deceive”, “misrepresent”, “dissimu-
late”. Accordingly, a restriction which formally meets the requirements of Article XX(b)
will constitute an abuse if such compliance is in fact only a disguise to conceal the pursuit
of trade-restrictive objectives. However, as the Appellate Body acknowledged in Japan—
Alcoholic Beverages, the aim of a measure may not be easily ascertained. Nevertheless, we
note that, in the same case, the Appellate Body suggested that the protective application
276 Domestic Instruments
of a measure can most often be discerned from its design, architecture and revealing
structure. (original emphasis)
This is all we know so far as a matter of case-law. We still lack information as to
the operationalization of this analytical framework. To some extent, this issue
is moot: with respect to measures that have respected the necessity requirement,
the likelihood of abuse is probably small; were WTO members to indeed abuse
their rights, they would, presumably, be wishing to inflict the highest possible
trade cost on their partners. The necessity test internalizes such externalities.
Accordingly, a measure that is both necessary and non-discriminatory will hardly
be found to fail the disguised restriction requirement.¹⁴²

4.4.10 A critique of the case-law

4.4.10.1 The discontents of an exhaustive list
We have already held the view that constructing the list included in Art XX of
the GATT to be exhaustive, risks turning the basic obligation assumed under
the GATT not to discriminate on its head, and equate it to an obligation to
deregulate.¹⁴³ The only provision that could be construed in a rather open-ended
manner, and thus reduce substantially if not eliminate altogether the risk of con-
struing the GATT as a legal instrument mandating deregulation, is Art XX(a)
of the GATT. It all depends of course on the manner in which the term ‘public
morals’ will be construed in future experience (since so far, this term remains
uninterpreted).
On one side of the spectrum, one could see this term being equated to some
sort of ordre public exception. Such a construction would establish a symmetry
between the general exceptions list reflected in the GATT and in the GATS (Art
XIV of which does not refer to public morals, but instead, to public order). And
of course, such an understanding of the term ‘public morals’ would leave WTO
members ample discretion to justify measures taken for an, in principle, unlim-
ited number of regulatory objectives. On the other side of the spectrum, pub-
lic morals could be construed in a substantially narrower manner and refer only
to measures such as a ban on pornography, that one intuitively would subject
to its purview. Although Feddersen (1998) who favours a narrow interpretation,
as cited above, has probably the better arguments, it would seem that following
Charnovitz (1998) on this score is the only way to continue understanding the
GATT as an instrument of negative integration.¹⁴⁴

¹⁴² By this, I do not mean that this cannot be the case. Practice has often defied foresight, and
this is probably the case why legislators have a preference for indicative (as opposed to exhaustive)
lists.
¹⁴³ The remarks here are pertinent only if Art XX of the GATT continues to be construed as an
exception to NT.
¹⁴⁴ It bears repetition that this is a second best construction, the first best being a radical
re-orientation of the NT test under Art III of the GATT.
 National Treatment: The Legal Discipline 277

4.4.10.2 Extra-territoriality
Trade transactions are by definition international transactions where, sometimes,
the right of more than one jurisdiction to regulate a particular transaction will be
implicated. Regulatory diversity might be an obstacle to trade. Assigning juris-
diction to one country for a type of transactions might thus facilitate trade liber-
alization. Law has not addressed this issue in a conclusive manner: recall that the
working group on Border Tax Adjustments endorsed only in part the destination
principle. Case-law, so far, has not had to face a challenge against the legitimacy
of a particular regulatory intervention: questions of the type, ‘has country A the
right to regulate process of production in a manner that prejudges exports from
country B to its market, and if yes, is this an unconditional right?’, have not been
explicitly submitted to the WTO. As a result, there has been no need to respond
to such questions. Implicitly, nevertheless, case-law has adhered to an uncondi-
tional destination principle, whereby the country granting market access has the
right to impose conditions, upon the satisfaction of which, market access will be
granted. The destination principle is at odds with the classic bases for regulat-
ing jurisdictional issues under public international law, in that no balancing of
interests (to regulate) will ever take place. At the same time, it might run counter
a pure contractual construction of the GATT, since, in the name of public order,
the content of negotiations will be prejudged. Case-law, in this respect, rests on
no theory, and it is, probably, because of the self-selection of disputes that we have
so far avoided cases which might rock the boat. In what follows, we advance a
framework that can serve as basis to classify cases where regulatory diversity can
lead to tariff barriers. We conclude that, even if we were to retain the destination
principle, there are good arguments for tempering it.
Case-law, so far, gives the impression that the WTO adjudicating bodies
are sitting on the fence when it comes to deciding the territorial reach of meas-
ures taken under Art XX of the GATT.¹⁴⁵ Take the US—Shrimp litigation,
for example: exhaustible natural resources, assuming that living organisms can
come under the purview of this term as the AB suggested, might ‘travel’ around
the world; their protection hence cannot be guaranteed unless through either
extra-territorial application of national laws, or multilateral agreements, or both.
The AB¹⁴⁶ stopped short of making any sweeping statements as to the territorial
reach of national laws coming under the purview of Art XX of the GATT. In

¹⁴⁵ The same analysis is valid for Art III of the of the GATT as well, assuming a proper discrim-
ination test has been included therein.
¹⁴⁶ Absent the emergence of a worldwide consensus on the protection of exhaustible natural
resources, an issue discussed to which we return infra in this sub-section, Art XX(g) of the GATT
could be construed as allowing WTO members to intervene only with respect to exhaustible nat-
ural resources within their territorial limits. The judge, hence, would have to decide between effect-
iveness of a measure (since purely territorial actions could be rather inefficient) and respect for the
national sovereignty principle. The GATT, as Chang (1995) has pointed out, does not necessarily
take a position in favour of efficient environmental protection. One way out of this construction is
what the AB did in US—Shrimp: without denying that there are transboundary effects, it implicitly
278 Domestic Instruments

the absence of a specific challenge by the complainant to this effect, the AB did
not have to answer the question whether the challenged US measure was intra-
(and hence, legal) or extra-territorial. Implicitly, nevertheless, it acknowledged
the right of the US to regulate conditions for granting access to its market for
shrimps. The US action had an effect on foreign producers’ welfare. The outcome
was that the US government did not have to compensate affected traders for the
damage suffered. Should WTO adjudicating bodies condone such behaviour?
And, if yes, on what basis should they do so?
To start with, there would be nothing to discuss if the US action did not prod-
uce any extra-territorial effects. What is intra- and extra-territorial is not always
an easy question to answer. Law in this area has moved sometimes by intuition,
sometimes by social convention, and rarely based on sound intellectual grounds.
Leaving the jus cogens issue aside,¹⁴⁷ it is generally accepted that the hierarchy
between the two basic bases conferring the right to regulate is territoriality before
nationality: when in Rome, do as the Romans do, in other words.
Territoriality, however, is an elusive concept. The main reason for the difficulty
in this respect has to do with the increasing acceptance of a rather imprecise tool
for conferring jurisdiction: subjective territoriality, or, as is widely known, the
eff ects doctrine. The eff ects doctrine is an extension of the (objective in the geo-
graphic sense) territoriality principle. For a number of reasons, which largely have
to do with the increasingly free movement of production factors, effects of regula-
tory (or private) interventions can be felt across various jurisdictions.
Its origins, however, were much more modest. The notorious Lotus jurispru-
dence, discussed before the Permanent Court of International Justice (PCIJ),
where the effects are quite dramatic and the interest to regulate can be hardly put
into question through an application of a reasonableness (however defined) test,
is usually cited in literature as the origin of the effects doctrine: shooting from
a French boat and wounding people in Turkey was accepted by the PCIJ (and
is widely accepted) as basis for conferring jurisdiction to Turkey, irrespective of
whether the shooter was on French territory when committing the act.
Over the years, this kind of thinking crept into other legal statutes. Antitrust was
one of the first statutes to embrace the effects doctrine. Assuming the anti-competitive
(price fixing) effects of a Swedish cartel are felt among Finnish consumers, Finland
will have good reasons to exercise its jurisdiction and punish the foreign practice.
This much is accepted on the two sides of the Atlantic and increasingly beyond.
But even in the field of antitrust, there are disagreements as to which effects con-
fer jurisdiction: take outward trade, as an example; it is highly debatable whether
moving from protection against inward trade to measures to protect outward

takes the view that effects are present in the US market as well, since depletion of exhaustible nat-
ural resources might affect the US sovereignty as well.
¹⁴⁷ A legitimate omission for the purposes of the present exercise, since the GATT does not
reflect any rules that could be qualified jus cogens.
 National Treatment: The Legal Discipline 279

trade is consistent with international law. Can, for example, Finland take action
against a Swedish import cartel which entails (negative) welfare implications for
Finnish exporters to Sweden? Mavroidis and Neven (1999) take a sceptical view
on this and related issues, but the jury is still out on issues like this.¹⁴⁸
In the GATT context, the issue can be even more complicated. Essentially, as
mentioned above, every trade transaction is by definition international. Take, for
example, the case where a Swedish industry is producing steel without respecting
environmental norms and contributes to the creation of acid rain. Should Finland
await until acid rain falls in Helsinki before intervening, or should it be allowed
under Art XX(g) of the GATT to ban imports of steel from Sweden anyway?
The degree of complication can be heavily increased, once we leave the realm of
human life, animal life or environmental protection, and move into the realm of
public order issues: can a WTO member legitimately stop imports from another
member which does not ban child labour?¹⁴⁹ In such an example, the distortion
is not transboundary, since child labour is not being imported. Its products are
however, and a lato sensu understanding of the eff ects doctrine would be offered
as an argument to support that there are effects in the importing market.¹⁵⁰ Or,
what if a WTO member stops trading with another member because the latter
permits the death penalty?
There is no easy answer to any of these questions. And these are not tomorrow’s
questions; they are very much today’s world. Take the three cases discussed in this
and previous chapters, EC—Asbestos, US—Shrimp, and EC—Tariff Preferences.
These are all cases where the defendant, that is, the regulating state, prevailed.¹⁵¹
Implicitly, at the very least, the exercise of jurisdiction by the defendants was
accepted as lawful by WTO adjudicating bodies.
EC—Asbestos is a case where the regulating state (France) is intervening against
a consumption externality which will produce its effects in France. Although the
sales ban might have effects outside France (Canadians, assuming France is an
important export market, will have to change their production process if they wish
to continue exporting to France), it also has effects in France, since the production
process has been incorporated in the final product. Even if French consumers can-
not distinguish between asbestos containing and asbestos-free materials, exposure
to the former might have very negative health effects to consumers.

¹⁴⁸ For a comprehensive discussion on extra-territoriality and its treatment in public inter-
national law, see the various contributions in Meessen (1996).
¹⁴⁹ Recall the AB report on EC—Tariff Preferences discussed in Chapter 2. Th is report leaves the
door open to conditioning extra preferences upon enactment of policies based on objective criteria.
From preferences to sanctions there is a short walk, if any.
¹⁵⁰ To avoid any misunderstanding, mentioning such examples should not be equated to accept-
ance of their intellectual solidity on our part.
¹⁵¹ Recall that the defendant lost in EC—Tariff Preferences only because the list of beneficiaries
was closed. In other words, the right of the European Community to condition granting of extra
preferences, upon satisfaction of criteria unilaterally advanced by the European Community, was
not put into question by the AB.
280 Domestic Instruments

US—Shrimp is a slightly different case. What is addressed here is a produc-

tion externality: by fishing in a particular way, there is a negative external effect,
namely, the accidental taking of the life of sea turtles. The US measure might have
effects outside the US territory, assuming foreign producers wish to continue to
export to the US market. There will be, however, no consumption externality in
the US market, at least, not an immediate one. Still, the US legislator can claim
that the (negative) effects in the US (like in the world) market will be unavoid-
able, unless action has been taken to protect exhaustible natural resources, and,
thus, protect the world wide ecological equilibrium. In other words, absent regu-
lation, there will be a global inefficiency.
EC—Tariff Preferences is yet again, different. Regulation here is justified
because the production process is deemed to be unjust or unfair. There is neither
a production, nor a consumption externality that is being internalized through
EC laws. EC legislator is not directly regulating the production process in India,
Pakistan, etc either. India, etc nevertheless, might feel that they have to respect
the EC preferences if export income gained in the EC market outweighs the costs
of meeting the EC standards.
WTO adjudicating bodies, as already alluded to above, found nothing wrong
with the exercise of jurisdiction by defendants in any of the three cases. They
accepted the right of the WTO member granting market access to regulate on
this score. In a way, the destination principle, so much debated during the Border
Tax Adjustments case with respect to fiscal measures, has now been uncritically
welcome, it seems. WTO members (like India in EC—Tariff Preferences) already
complain about extra-territorial behaviour and it will not be long before the exer-
cise of jurisdiction as such will be questioned.
Many would view public international law as the appropriate forum to discuss
such issues. The WTO is an international law contract, and, as a result, this is a
natural place to entertain this question. A classic public international law type of
analysis would revolve around the following questionnaire:
(1) which countries have an interest to regulate a particular transaction?
(2) if more than one, which country has the greater interest to regulate this par-
ticular transaction?
Assuming, of course, the countries with an interest to regulate have identical
laws/application of laws, then the issue is moot, since the result will be the same
irrespective who the final regulator is. This is, nevertheless, a highly unlikely
scenario.
Examples like the three cases we discussed above are typical of cases where
more than one WTO member can lawfully assert the right to regulate a par-
ticular transaction. Question (2) above does not exclude that some balancing (of
interests) might be required when more than one jurisdiction has an interest in
regulating one particular transaction. Balancing involves the exercise of discre-
tion. There is often wide dispute across scholars and national legal services about
 National Treatment: The Legal Discipline 281

the outcome of this exercise.¹⁵² WTO case-law does not even enter the discussion
on balancing: it assumes that it is the right of the state granting market access
to regulate access to its market, irrespective of the external effects of such regu-
lations on third (exporting) countries. It is, in a nutshell, destination principle
unlimited. There has been some support for this view, in literature: Bartels (2002)
for example, argues that Art XX of the GATT can be used to stop imports of
products produced through child labour. In the author’s view, however, recourse
to this provision is not legitimate if it aims at enforcing other WTO members’
obligations. By the same token, Charnovitz (1998 and 2002) argues that, since
process based measures are not per se GATT inconsistent, WTO members can
through recourse to the various sub-paragraphs of Art XX of the GATT advance
their own public order agenda.
The problem with this approach is that it neglects the external effects of such
actions. True, the WTO contract clearly states that public order concerns should
take precedence over trade commitments. What about cases, however, where
country A negotiates with country B, only to subsequently block trade from coun-
try B because it conflicts to its public order? Has not A accepted that B is a legitim-
ate trading partner by agreeing to negotiate with it? Or should B go to negotiate
with A, knowing ex ante, that the outcome of negotiations will hinge upon the
eventual acceptance of its public order by A? B, in such cases, would have no
incentive to negotiate at all. The papers by Bartels and Charnovitz do not respond
to these questions. Moreover, this view also neglects another, institutionally
important dimension that the unlimited acceptance of the destination principle
entails: unilateral definition of policies can be a powerful tool when exercised by
some, it is simply not the case when exercised by others. Take the US—Shrimp
dispute, for example, and paraphrase the facts: Malaysia and Thailand
challenged the US measure because it was the United States that was adopting it.
Their exports to the US market represent a substantial percentage of their export
income. In other words, they know that the United States, through its policies,
can inflict adjustment costs on to them. Would they be attacking the same meas-
ure with the same intensity (the dispute dragged on for over 5 years) had St Kitts
and Nevis adopted this measure? By backwards induction, the United States (and
other WTO members similarly positioned) knows that it can affect terms of trade
through such interventions and might have an incentive to do so in the future as
well. Importantly, there is absolutely no guarantee as to the moral standing of
any public order unilaterally defined. The only guarantee that those being in a
position to affect terms of trade will behave in a morally acceptable manner, is

¹⁵² See, for example, the divergent views expressed with respect to the treatment by the US
Supreme Court of the Insurance Antitrust Case (Hartford Insurance) by Lowenfeld (1995), Trimble
(1995), and Kramer (1995). See also the discussion on the problematic, for many, treatment of
outward trade by some US courts in Mattoo and Mavroidis (1995). More generally, see the interest-
ing thoughts expressed in Vagts (2003), and the treatment of balancing in various contributions in
Meessen (1996).
282 Domestic Instruments

their internal decision making process. Public choice literature evidences a series
of contributions showing that the decision making process can be manipulated
by homogeneous lobbies, or, worse, can be manipulated and sold in some form
of socially acceptable package (protectionism under the guise of promotion of
moral values).¹⁵³
Bagwell et al (2002) tried to respond to this question, by endorsing a contrac-
tual approach and argue that trading partners negotiating concessions during a
round should be presumed to morally accept each other, otherwise why negoti-
ate in the first place? The periodicity of rounds (every 6–7 years) makes it highly
unlikely to see societies change their moeurs so drastically as to justifying recourse
to trade-blocking measures based on public order issues. Changes in policy which
are not GATT inconsistent and might affect trade could, in their framework, be
challenged through NVCs.¹⁵⁴ Bagwell et al (2002) do not put into question the
right of the WTO member granting market access to regulate the conditions
under which market access will take place; they simply state that, assuming such
exercise has negative effects on foreign producers, the regulating state will have to
compensate. Consequently, this approach does not put into question the destin-
ation principle at all.
In its defence, one could add that WTO members could, if they do not want to
transact with a particular state, simply invoke Art XIII of the WTO Agreement to
this effect. This possibility does not, of course, take care of all possible contingencies:
first, this provision cannot be invoked post-accession of a particular country to the
WTO; second, negotiators also know that they can block trade (and thus, not hon-
our concessions made) if trade goes against their country’s public order (Arts XX and
XXI of the GATT). Respect of the public order is part of the contractual promise.
The relationship between public order and tariff concessions is quite problematic.
This approach, even if endorsed, is of limited value, since it does not, for example,
address cases where no change has taken place. Assume, for example, that A and B
negotiate a concession (by A) on shrimps. Subsequently, A does not honour its com-
mitment by invoking its pre-existing legislation banning imports of shrimps if fished
in a particular manner. Should B be penalized for not being careful? Or, conversely,
should A be penalized because B might legitimately have thought that A, by negoti-
ating with B, is about to change its laws? And what about cases where no law exists?
Are cases of new laws cases of change which would yield the obligation to compen-
sate? If yes, then we risk adding to the deregulatory pressure, a rather ill-conceived
strategy both as a matter of law and a matter of reason. Finally, should the form of
negotiation matter? Recall that most negotiations take place on a bilateral basis, that

¹⁵³ See the excellent analysis offered by Sykes (1999) on this score. Recall the facts of EC—Tariff
Preferences: the European Community is compensating beneficiaries who will reduce the size of an
EC problem (as well). It does not compensate beneficiaries for, for example, investing in education
in their home market.
¹⁵⁴ Mavroidis (2004b) argues along the same lines but also suggests that some degree of self-
restraint is warranted: any policy affects at least indirectly, and potentially, trade.
 National Treatment: The Legal Discipline 283

is, few countries negotiate directly. The results are being multilaterally extended. If A
has negotiated with C and B has not done so, B can hardly be accused for accepting
C as a legitimate partner. How to respond to all these cases?
There is a need to come up with a benchmark, and there are no obvious can-
didates. Public international law is of limited help. The eff ects doctrine provides
satisfactory responses on issues such as the one treated in the Lotus jurisprudence,
or in (some) antitrust practice. It is a less useful instrument when it comes to
discussing the regulatory interface across countries which trade (and thus have
a channel of continuous communication between them) and at the same time
are (perceived to be) ‘exporting’, along with goods, their cultural, moral values
as well. The Lotus formula works fine, as long as we talk direct, physical effects.
Extending the effects doctrine to cover indirect, non-physical effects as well,
amounts to opening Pandora’s box.
So how do states deal with such issues? Do they prefer to let uncertainty loom
over an area intimately linked with the ever-increasing liberalization of production
factors? Uncertainty is very much the product of unilateral exercise of jurisdiction.
Cooperative solutions that ex ante guarantee legal certainty and political acceptabil-
ity have been designed and are being practiced. Through bilateral or multilateral
agreements, sovereign states solve issues relating to the limits of national jurisdic-
tion, issues that unless solved in a contractual manner, risk imposing important
transaction costs on all interested parties. In the field of environmental protection,
for example, this is the role of MEAs: through an MEA, countries can agree on
the jurisdictional reach of their respective actions and thus, reduce (and hopefully)
eliminate uncertainty resulting from the (eventual) application of the effects doc-
trine. Palmeter and Mavroidis (1998) early on argued that, the VCLT obliges WTO
adjudicating bodies, when adjudicating a dispute, to look into MEAs (to the extent
to which they are relevant of course) even if such MEAs are concluded between a
sub-set of the total WTO membership. In such a case, they will, of course, have to
guarantee the pacta tertiis nec nocent nec prosunt maxim.¹⁵⁵ Trachtman (1999) agrees
that, as a matter of interpretation, such an approach is quite consistent with the legal
nature of the WTO contract. The WTO adjudicating bodies have so far refrained to
open the door, in a clear-cut manner, to this perspective. It is high time they do so.
MEAs as well, alas, are of limited help: typically MEAs will deal with cases
of global inefficiencies, such as the transaction in US—Srhimp. Indeed, unless a
concern is common, there will be no interest to co-regulate. What to do with the
remaining transactions?
The proposal advanced here is meant to spark discussion on this issue, a discus-
sion very much needed today. This proposal rests on two key foundations:
(1) the destination principle should not be questioned, it should, however, be
tempered;

¹⁵⁵ In plain English, they will not be imposing obligations on non-signatories.

284 Domestic Instruments

(2) by tempering, we understand that, depending on the transaction, the regu-

lating state might have to compensate the affected trading partners.
There should be no compensation due in case of:
(1) a consumption externality, such as in EC—Asbestos;
(2) production externalities that affect global commons. In cases where an
MEA has been signed to deal with it, and the parties to the dispute are both
parties to the MEA, behaviour which is inconsistent with the MEA should be
sanctioned. By the same token, one could legitimately block trade assuming the
production externality is the outcome of behaviour inconsistent with an interna-
tional arrangement (say, an agreement promoting public health) to which both
the exporting and the importing country have adhered;¹⁵⁶
(3) violation of jus cogens.¹⁵⁷
Case (1) is a case of direct effect. The effects could also be direct in case (2)
although scientific proof might have a higher mountain to climb in this case;
even if not, there is an agreement to behave in a particular manner that is being
violated. There is definitely a worldwide agreement to behave in a particular man-
ner in case (3), and an agreement to prosecute such acts anywhere, no matter the
nationality of those committing them.
Conversely, compensation should be paid in cases where:
(1) production externalities do not affect global commons;
(2) a behaviour is unilaterally judged unfair.
In none of these cases there will be direct physical effects in the importing mar-
ket. In this line of thinking, it is clear that the outcome in EC—Asbestos is cor-
rect. It is also clear that the outcome in EC—Tariff Preferences is incorrect: the
European Community should, of course, be in position to regulate access to its
market; it should, nevertheless, be required to compensate India, in a setting
like the one discussed in this litigation.¹⁵⁸ Cases like US—Shrimp (assuming

¹⁵⁶ One might be tempted to say that the same solution should be followed also in cases where
the affected state has not signed the MEA, but the MEA codifies customary international law. The
problem with this approach is that, with few exceptions, there is substantial uncertainty as to iden-
tity of customary international law, see Goldsmith and Posner (2005). See, however, also Alvarez
(1994) who makes persuasive arguments to the contrary.
¹⁵⁷ We are not subscribing to the view that trade sanctions is the best means to address such
cases. It should, however, be the privilege of states to decide on the efficiency of their response.
¹⁵⁸ If the European Community genuinely wanted to combat drug trafficking, or, eventu-
ally, child labour, this is the solution it should arrive at anyway: by restricting export money, the
European Community is also reducing the (Indian) budgetary expenditure for social policies
(since taxable income will fall). When faced with this choice, the Indian government will, other
things being equal, have even less money to spend on policies designed to combat child labour.
Conditioning the payment of compensation upon adoption of specific policies (eg building add-
itional schools so as to reduce the distance that children have to travel in India in order to visit their
schools) might be appropriate.
 National Treatment: The Legal Discipline 285

sea turtles are endangered species, and an MEA, say the Convention on the
International Trade of Endangered Species (CITES), has been concluded rec-
ognizing as much, and the measure taken is in accordance with the MEA)
should also fall under the category of cases for which no compensation should
be paid.¹⁵⁹

4.4.10.3 A two-tier test?

One can legitimately ask the question whether the distinction between sub-
stantive conformity and application of a measure scrutinized under Art XX of
the GATT is meaningful at all. Should, for example, the original US measure
challenged in US—Shrimp be provisionally exonerated under Art XX(g) of
the GATT, before it is struck down by virtue of the chapeau? And what if the
measure at hand had yet to find application, but was already affecting market
behaviour? According to standard case-law, WTO members can attack legisla-
tion even before its application, to the extent that it mandates GATT inconsist-
ent behaviour and it leaves no room for discretion on this score. Assume that a
WTO member enacts such legislation which, in accordance with current law,
upsets the competitive condition between two DCS products, and is, conse-
quently, found to be in violation of Art III of the GATT. It then raises an Art
XX of the GATT defence, and argues that its law aims at protecting a legitim-
ate objective. It will find it impossible to prevail if its law has not been applied.
The AB, it seems, has added in this respect an unnecessary complication. Its
reading of Art XX of the GATT would be smoother had it accepted that the
chapeau reflects the obligation to adopt only non-discriminatory regulatory
interventions and is relevant for each and every sub-paragraph of Art XX of the
GATT. It would also lead to faster reviews, since all discriminatory cases would
be outright considered to be WTO inconsistent.

4.4.10.4 The standard of review

In a short paper, Sykes (2003b) makes an insightful empirical observation:
WTO adjudicating bodies have adopted a deferential standard of review when
the protection of human/animal life is being sought. Th is is not the case when
other sub-paragraphs of Art XX of the GATT are being invoked. In other words,

¹⁵⁹ The CITES prohibits trade in endangered species. The US measure does not regulate trade in
sea turtles, assuming the latter are accepted as endangered species. The US measure takes positive
steps to ensure (eventually) that sea turtles are not extinct. In that, it rejoins the spirit of CITES.
Recall that this discussion is normative and not positive. As a matter of positive law, it bears repe-
tition, the US measure should have been found consistent with Art III.4 of the GATT. Note that
the advanced framework does not respond to all questions. What if, for example, A blocks trade
from B for environmental reasons but cannot point to an international agreement to this effect, or
B has not agreed to participate in the MEA that A invokes. Assuming A can show that actions by
B are a threat to global commons, such cases should be assimilated to the first category of cases (no
compensation should be due).
286 Domestic Instruments

there is a correlation between the nature of the risk at stake and the intrusive-
ness of the judicial review. This is probably a wise approach in light of the insti-
tutional risk in case Type I errors are being committed where, for example,
social values such as protection of human life and health are being protected.
Th is does not mean, however, that WTO adjudicating bodies should be tak-
ing light-heartedly the rest of the grounds mentioned in Art XX of the GATT.
In the Korea—Various Measures on Beef case discussed supra, for example, the
panel and the AB outlawed the dual retail system operated by the Koreans for
being overly restrictive, without explaining how another, less restrictive option
would lead to an equally efficient outcome. Th is approach is also at odds with
the very nature of the necessity requirement to prejudge the means but not the
ends sought.

4.5 A specific exception: films

GATT, Art IV regulates screen quotas. In essence, it allows WTO members to
keep such quotas in place in order to protect domestic films. Consequently, this
provision is a deviation from NT, since it allows to impose quotas on foreign films.
Note that the quotas involved here are domestic and not import or export quotas,
hence, they should not come under the purview of Art XI of the GATT. Even if
enforced at the border (with respect to foreign films), they should be considered a
domestic instrument merely enforced at the border, as per the Interpretative Note
ad Art III.4 of the GATT.
It was felt, at the time of the drafting of the NT provision, that NT could cause
difficulties to film screening. Many countries held the view that tariffs were not
an effective means to protect the domestic film industry, and thought quotas
would be a more appropriate instrument.¹⁶⁰
Th is provision should have been superseded by the advent of the GATS. It
is true that few commitments have been made in the audiovisual sector; still,
it is commitments under GATS that now prevail over Art IV of the GATT.
So far, no formal action has been taken to eliminate Art IV of the GATT.¹⁶¹
It is reasonable nevertheless, to assume that this provision has fallen into
desuetudo.

¹⁶⁰ See on this issue the GATT Doc MTN.GNS/AUD/W/1.

¹⁶¹ Interestingly, WTO Doc S/C/W/40 of 15 June 1998, which discusses commitments in the
audiovisual sector, only in passing refers to Art IV of the GATT.
 Dealing with NTBs 287

5 Dealing with NTBs

5.1 The TBT and SPS Agreements

Over the years, the trading partners have added new agreements to the multi-
lateral legal arsenal aiming at addressing NTBs.¹⁶² Where tariffication has been
impossible,¹⁶³ disciplining domestic regulation has emerged as the second best. The
WTO TBT and SPS Agreements deal with a sub-set of domestic instruments that
would otherwise come under the purview of Art III.4 of the GATT. These two agree-
ments innovate, in comparison to Art III of the GATT, in two important ways:
(1) they incorporate a more elaborate (than the GATT) test to detect protec-
tionist intent.¹⁶⁴ Recall our discussion concerning the imperfections we observed
when evaluating the non-discrimination test included in Art III of the GATT.
Such imperfections stem mainly from the incompleteness of the GATT in this
respect, and the incentive for WTO members to behave in an opportunistic man-
ner. The TBT/SPS Agreements include a series of proxies, such as scientific evi-
dence, coherence in the formulation of policies which help distinguish beggar thy
neighbour behaviour (opportunism) from policies genuinely aiming at protecting
health/environment. International standards as well, become proxies indicating
absence of protectionist intent. All these devices are at the service of the judge
who will eventually be called to sanction opportunism;
(2) the second innovation has to do with the kind of beggar thy neighbour pol-
icies that they outlaw. Recall that Art III of the GATT imposes a non-discrimina-
tion principle and nothing beyond that. Let us assume that a WTO member does
not produce product A, and it wants to exclude it from its market, because, in its
view, it is associated with adverse health implications. A trade embargo would
never run foul of Art III of the GATT, since there is no discrimination issue here.
It might, however, be caught by the disciplines included in the SPS Agreement:
failing consistency with the legal requirements imposed by the SPS Agreement,
the measure at hand will be judged WTO inconsistent even if it affords no advan-
tage to the competing domestic production since there is no such production. In
other words, through the SPS Agreement domestic measures which affect solely

¹⁶² Kono (2006) observes that it is democracies that typically create NTBs. In his analysis, by
increasing the transparency of some policies relative to others, democracy induces politicians to
replace transparent trade barriers (such as tariffs) with less transparent ones (such as NTBs). He
tests his hypothesis using a sample of 75 countries and concludes that democracy leads to lower
tariffs, higher core NTBs, and even higher quality NTBs. In his view, democracy promotes opti-
mal obfuscation that allows politicians to protect their markets while maintaining a veneer of lib-
eralization. The various TBT (and SPS) disciplines are meant to tame the tendency for optimal
obfuscation.
¹⁶³ Tariffication has been successfully used in the agriculture negotiations, where a series of
domestic policies have been tariffied (that is, expressed in tariff terms), bound and, eventually, will
be reduced in subsequent negotiating rounds.
¹⁶⁴ See the analysis of this issue in various chapters in Melloni (2005).
288 Domestic Instruments

imported products (without conferring an advantage to the domestic production

as the measures captured by Art III of the GATT) can still be found to be WTO
inconsistent. The SPS Agreement thus outlaws beggar thy neighbour policies in
addition to those outlawed by the GATT.¹⁶⁵

5.2 The legal relationship between GATT, TBT, and SPS

Since the TBT and the SPS Agreements cover a sub-set of domestic instruments,¹⁶⁶
a transaction can, in principle, be subjected to the disciplines of either the GATT,
or the TBT, or the SPS. If the legal discipline across the three agreements was iden-
tical, there would be no need to prolong this discussion any further. As mentioned,
however, the three agreements do not reflect the same test for evaluation of the
consistency of a particular transaction with the WTO. Consequently, it is impera-
tive to first determine which agreement takes precedence.¹⁶⁷ The TBT and the
SPS Agreements are Annex 1A agreements and their relationship with the GATT
is in principle, addressed in the General Interpretative Note to Annex 1A, dis-
cussed in Chapter 1. As we saw in Chapter 1, case-law with respect to the General
Interpretative Note to Annex 1A has not been linear. In EC—Asbestos, nevertheless,
the AB concluded that a measure, which revealed the characteristics of a technical
regulation, and thus, simultaneously fell under the TBT and the GATT should
have been reviewed under the TBT Agreement, and not under the GATT (as the
panel had originally done). The subsequent panel report on EC—Sardines contains
an even more explicit reflection of this approach (§§ 7.14–7.16).¹⁶⁸ The relationship
between TBT and SPS is regulated in Art 1.5 of the TBT, which clearly states that
SPS takes precedence over TBT. In a nutshell, it is by now settled that the legal hier-
archical relationship between the three agreements or, in WTO parlance, the order
of analysis is as follows: SPS over TBT over GATT 1994.

5.3 TBT¹⁶⁹
5.3.1 Coverage
The TBT Agreement imposes multilateral disciplines on two types of legal instru-
ments issued by WTO members: technical regulations and standards. Whereas

¹⁶⁵ Even in this situation, however, a non-discrimination test is still embedded in the SPS
Agreement: the ban will cover also potential national production; the import prohibition for for-
eign products amounts to a production ban for domestic products.
¹⁶⁶ For the time being it is only an affirmation, but as we will see infra this is indeed the case.
¹⁶⁷ On this issue, see Marceau and Trachtman (2002).
¹⁶⁸ TBT, Art 1.5 solves the hierarchy issue between TBT and SPS: measures which conceivably
could come under the disciplines of either the SPS or the TBT Agreement must be addressed under
the disciplines of the former.
¹⁶⁹ As is the case with all Annex 1A agreements other than the GATT, what follows is not
meant to be an exhaustive discussion of all TBT issues. It is rather a sketchy discussion of the main
features of TBT.
 Dealing with NTBs 289

compliance with a technical regulation is mandatory in the sense that the prod-
ucts concerned cannot be guaranteed market access in the regulating WTO
member in its absence, compliance with a standard is optional.¹⁷⁰
A WTO member wishing to enact a technical regulation has two options: either
adopt an international standard (assuming that one exists and can be appropri-
ately used), or proceed unilaterally.

5.3.2 International standards

There is a great deal of discussion taking place in economic theory regarding
whether and under what circumstances harmonization (international standards)
is welfare enhancing. The answer is far from clear.¹⁷¹ It is clear that standard-
ization of production at the international level is a better guarantee for market
access, in regulated sectors. Non-discrimination (NT) is not a guarantee for mar-
ket access; in fact, in case of regulatory diversity, the WTO member with the
higher protection (eg higher environmental standards) can behave in a perfectly
GATT consistent manner by keeping all products originating in countries with
lower environmental standards outside its market. In such case, the products ori-
ginating in a country with lower standards cannot be sold in the market of the
country with the higher standards. On occasion, it might even have the incentive
to enact prohibitively high standards that impose substantial adjustment costs
on its trading partners. Adherence to international standards guarantees that,
in principle, the products concerned can be sold in foreign markets (adhering to
such standards). What is also clear is that, in order to avoid rent-seeking behav-
iour from domestic lobbies, a certain degree of cooperation is warranted.¹⁷² The
remaining part is the tough part of the question: what is the optimal amount of

¹⁷⁰ The term ‘technical regulation’ is defined in Annex 1, §1 of the TBT in the following terms:
‘Document which lays down product characteristics or their related processes and production
methods, including the applicable administrative provisions, with which compliance is manda-
tory. It may also include or deal exclusively with terminology, symbols, packaging, marking or
labelling requirements as they apply to a product, process or production method.’ Reviewing its
prior jurisprudence on the issue (essentially §§ 67–72 from its report on EC—Asbestos), the AB
defined ‘technical regulation’ in its report on EC—Sardines in the following manner (§§ 175 and
176): ‘In doing so, we set out three criteria that a document must meet to fall within the definition
of “technical regulation” in the TBT Agreement. First, the document must apply to an identifiable
product or group of products. The identifiable product or group of products need not, however, be
expressly identified in the document. Second, the document must lay down one or more character-
istics of the product. These product characteristics may be intrinsic, or they may be related to the
product. They may be prescribed or imposed in either a positive or a negative form. Third, com-
pliance with the product characteristics must be mandatory’ (original emphasis) The term ‘stand-
ard’, also defined in Annex A, reads as follows: ‘Document approved by a recognized body, that
provides, for common and repeated use, rules, guidelines or characteristics for products or related
processes and production methods, with which compliance is not mandatory. It may also include
or deal exclusively with terminology, symbols, packaging, marking or labelling requirements as
they apply to a product, process or production method’.
¹⁷¹ See on this issue Drabek (2005). See also the discussion in various parts of the World Trade
Report (2005, pp 29ff ).
¹⁷² See on this score Sykes (2000), and Pelkmans (2003).
290 Domestic Instruments

cooperation? Should one go for mutual recognition by signing an agreement to

this effect (MRA), or should one go all the way so to speak, and harmonize?
Unilateral setting of technical regulations and/or standards can have nega-
tive implications on international trade:¹⁷³ a transaction costs based argument
in favour of harmonization is thus easy to make. On the other hand, harmoniza-
tion reduces (if not eliminates altogether) gains from innovation. Hence, mutual
recognition often emerges as the most appropriate candidate for regulatory
cooperation.¹⁷⁴ Mutual recognition, however, is not unproblematic. Technical
regulations and/or standards will continue to be set unilaterally. There is no guar-
antee at all that tomorrow’s policies will be judged equivalent to yesterday’s by all
parties to the original contract. Breaking from the contract might be costly.¹⁷⁵
Although empirical evidence to this effect across sectors is less persuasive than
before, one cannot exclude a race to the bottom, or, even, as Vogel (1997) shows,
an unnecessary race to the top. Back to square one, and the discussion on harmon-
ization. Harmonization, however, is costly not only in terms of negotiation, but
also because it does not allow participants to profit from gains from innovation
that would have resulted had they been operating in a competition across regimes
context. This is why some economists¹⁷⁶ will agree that harmonization defin-
itely makes sense in the presence of global externalities, such as environmental
pollution, and in the presence of network externalities: Katz and Shapiro (1985)
suggest that, in case of network externalities, enacting compatibility standards
can have positive welfare implications.¹⁷⁷ In this case, however, the market might
reach this outcome anyway. The former case is more problematic: the World Trade
Report (2005) mentions (p 53) that, in such instance, governments tend to under-
regulate. What is, however, the counterfactual? Take the case of the negotiations
on the Kyoto Protocol which deals with an admittedly global environmental

¹⁷³ Although, as Deardorff and Stern (1998) have shown, the measurement of technical bar-
riers to trade is not a day playing in the field even for the most gifted among econometricians.
Some sector-specific empirical research, however, aimed to provide numbers: Wilson and Otsuki
(2004), without pretending that they inquired in a comprehensive manner into the well-founded
of some technical barriers (although they allude to the opposite), conclude that, in their sectors
under examination (289 firms from 25 industries), the negative implications especially for develop-
ing countries from unilateral standard-setting should not be underestimated.
¹⁷⁴ Mutual recognition is a decisive step closer (than the classic non-discrimination recipe
embodied in Arts I and III of the GATT) to ensuring market access: in a non-discrimination scen-
ario, unless a country meets the regulatory standards in its exporting market, it simply can not
export to that market. In an MRA scenario, it is irrelevant if a partner has regulated more or less
or differently; they are all bound by an MRA and cannot be denied market access on regulatory
grounds. Vogel (1997) has shown that cooperation can lead, under circumstances, to a race to
the top. Piermartini (2005), shows that MRAs are associated with significant positive effects on
intra-EC trade.
¹⁷⁵ Which probably explains why MRAs are typically signed between players in frequent inter-
action (repeat players), among which some (quantified?) element of ‘trust’ exists.
¹⁷⁶ See the discussion in the World Trade Report (2005, pp 53ff ).
¹⁷⁷ Literature offers numerous examples, some more ‘classic’ than others (computers and plugs,
hardware and software, etc). The argument is that standard setting is preferable to the use of adap-
tors which come at a cost anyway.
 Dealing with NTBs 291

issue. One reading of the situation is that (some) states, either because they will
not be directly affected by their own pollution, or because there is a problem of
collective action, might prefer not to intervene at all, or intervene in an uncostly
and probably ineffective manner. Another reading, however, suggested by Gollier
et al (2000), is that some states, such as the United States, believe that no inter-
vention is required on an issue like the environment until more about the hazard
is known. Assuming this view is correct, states might be under-regulating for fear
that by doing so they might be creating more harm than good. One should not
necessarily impose a negative value to under-regulation.¹⁷⁸ The proliferation of
international standards is very much the outcome of this type of thinking.
International standards¹⁷⁹ are defined in Art 2.4 of the TBT:
Where technical regulations are required and relevant international standards exist or
their completion is imminent, Members shall use them, or the relevant parts of them, as a
basis for their technical regulations except when such international standards or relevant
parts would be an ineffective or inappropriate means for the fulfilment of the legitimate
objectives pursued, for instance because of fundamental climatic or geographical factors
or fundamental technological problems.
The term ‘international standard’ is not defined, except for the following
reference:
Standards prepared by the international standardization community are based on con-
sensus. This Agreement covers also documents that are not based on consensus.
Unlike the SPS Agreement, the TBT Agreement, except for some oblique refer-
ences to the ISO (International Organization for Standardization, with its head-
quarters in Geneva), does not mention by name the standard-setting institutions,
the standards of which it recognizes; Art 2.4 of the TBT simply calls on coun-
tries to use international standards when they exist, or when their completion is
imminent. Consequently, even an ISO standard would have to be judged on a
case-by-case basis in order to determine whether it is an international standard in
the TBT sense of the term.
International standards, on the other hand, do not necessarily carry the same
legitimacy. For one, they are not always adopted in the same manner. A num-
ber of factors (eg difference in composition/voting across standard setting insti-
tutions) might influence the manner in which standards covered by the TBT
Agreement will be adopted.¹⁸⁰ Legitimately one might ask the question whether
all standards, irrespective of the decision mode should enjoy the same legitimacy

¹⁷⁸ It is not suggested here that action in accordance with the Kyoto Protocol is unnecessary. It
is merely suggested that under-regulation could be a value judgment.
¹⁷⁹ To avoid any confusion one should always distinguish between an international standard
and a (domestic) standard. Whereas compliance with the former is required by virtue of Art 2.4 of
the TBT, compliance with the latter is not necessary for access to a WTO member’s market.
¹⁸⁰ See on this issue, the very interesting account with respect to the hormones standards pro-
vided in Abdel Motaal (2004).
292 Domestic Instruments

and, consequently, the same status under WTO law. The AB rejected an argu-
ment advanced by the European Community that only standards adopted by
consensus are international standards. Noting that the ISO/IEC (International
Electrotechnical Commission) Guide accepts only consensus-based standards as
international standards, the AB went on to say that the deviation included in the
Explanatory Note in Annex 1 of the TBT Agreement, according to which even
standards adopted without consensus are recognized as international standards,
must have been voluntary: it thus held that an international standard adopted
without a unanimity vote still qualifies as standard under Art 2.4 of the TBT
Agreement (EC—Sardines § 225).¹⁸¹
The panel and the AB accepted in the EC—Sardines dispute that a standard
adopted by the Codex Alimentarius Commission (which is, thus, recognized as
an international standardization body) relating to the species that can lawfully
carry the name sardine was an international standard in the TBT sense of the
term, without, however, providing any additional clarificaion as to the criteria
to be used in future transactions to define an international standard. The TBT
Agreement qualifies the term ‘international standard’ by adding the word rele-
vant in front of the term. In EC—Sardines, the European Community argued
that since the product coverage between the international standard and the EC
technical regulation was not identical, the former was not relevant for the latter:
the international standard covered the marketing of 21 fish species while the EC
technical regulation only covered the marketing of one of them. However, since
the EC technical regulation had legal implications for the marketing of the other
20 species covered by the international standard, the AB held the view that the
international standard was relevant for the EC technical regulation (§§ 222ff ).
This seems an eminently reasonable conclusion: if one were to accept the EC
argument, circumvention of all international standards would become a simple
act of acknowledging rights to only a sub-part of the right-holders in an inter-
national standard context.
Pursuant to Art 2.4 of the TBT, international standards or the relevant parts
of them should be used as the basis for the adoption of technical regulations.¹⁸²
The AB in its report on EC—Sardines, explained that the terms used in Art 2.4
of the TBT necessarily entail that a technical regulation should at the very least

¹⁸¹ In the same report, the AB, upholding the panel’s findings in this respect, confirmed that
the definition of standard in § 2 of Annex 1 to the TBT is relevant not only for domestic (that is,
non-compulsory), but also for international standards (§ 220).
¹⁸² Sometimes, in literature, one finds the view that the TBT and SPS Agreements are positive
integration instruments. The supporters of this view point to the provisions on international stand-
ards to make their point. Th is view, however, should be discarded. WTO members do not have to
enact international standards irrespective whether they want to intervene in a particular field. They
will have to base their intervention on such standards, only after they have decided to intervene.
This latter decision, nevertheless, is sovereign, and not a matter of legal compulsion: WTO mem-
bers which do not want, for example, to enact legislation regarding the safety of children’s toys, do
not have to anyway adopt the relevant ISO standards.
 Dealing with NTBs 293

not be contradictory to the relevant international standard. The AB also pointed

out that all relevant parts of an international standard, not only some of them,
must form the basis of a technical regulation (§§ 248 and 250). In the same
report, the AB had the opportunity to provide its understanding of the terms
ineff ective or inappropriate. In the dispute, Peru complained that the European
Community had deviated unjustifiably from an international standard reflect-
ing the denomination of sardines. The facts of the case¹⁸³ are as follows: the dis-
pute concerned the name under which certain species of fish could be marketed
in the European Community. An EC Council Regulation ((EEC) 2136/89 of
21 June 1989 laying down common marketing standards for preserved sardines
[1989] OJ L212/79), setting forth common marketing standards for preserved
sardines, allowed only products conforming to the following four requirements
to carry this name commercially: (i) be covered by CN codes 1604 13 10 and ex
1604 20 50; (ii) be prepared exclusively from fish of the species Sardina pilchar-
dus Walbaum; (iii) be pre-packaged with any appropriate covering medium in
a hermetically sealed container; and (iv) be sterilized by appropriate treatment.
Sardina pilchardus Walbaum (‘Sardina pilchardus’), is found mainly around the
coasts of the eastern North Atlantic Ocean, in the Mediterranean Sea, and in the
Black Sea. In 1978, the Codex Alimentarius Commission of the United Nations
Food and Agriculture Organization (FAO) and the World Health Organization
(WHO) adopted a worldwide standard for preserved sardines and sardine-type
products, which regulates matters such as presentation, essential composition
and quality factors, food additives, hygiene and handling, labelling, sampling,
examination and analyses, defects and lot acceptance. This standard (CODEX
STAN 94–1981, Rev.1–1995), covers preserved sardines or sardine-type products
prepared from 21 fish species, and among them the Sardina pilchardus and the
Sardinops sagax. Section 6.1 of CODEX STAN 94–1981, Rev.1–1995 regulated
the name that the 21 mentioned species could legitimately carry in the following
manner:
6.1.1 (i) “Sardines” (to be reserved exclusively for Sardina pilchardus (Walbaum)); or
(ii) “X sardines” of a country, a geographic area, the species, or the common name
of the species in accordance with the law and custom of the country in which
the product is sold, and in a manner not to mislead the consumer. (original
emphasis)
Peru exported preserved products prepared from Sardinops sagax. This species is
found mainly in the eastern Pacific Ocean, along the coasts of Peru and Chile.
Since, pursuant to the EC regulation, Peru was prohibited from exporting its
product under the name sardine, it requested the establishment of a WTO panel.
There was no dispute among the parties that the European Community had devi-
ated from the international standard. The AB had to first decide the meaning of

¹⁸³ For a more detailed description, see §§ 2–8 of the AB report.

294 Domestic Instruments

the term ‘ineff ective or inappropriate’ (§ 285):

. . . we noted earlier the Panel’s view that the term “ineffective or inappropriate means”
refers to two questions—the question of the eff ectiveness of the measure and the question
of the appropriateness of the measure—and that these two questions, although closely
related, are different in nature. The Panel pointed out that the term “ineffective” “refers
to something which is not ‘having the function of accomplishing’, ‘having a result’, or
‘brought to bear’, whereas [the term] ‘inappropriate’ refers to something which is not
‘specially suitable’, ‘proper’, or ‘fitting’ ”. The Panel also stated that:
Thus, in the context of Article 2.4, an ineffective means is a means which does not
have the function of accomplishing the legitimate objective pursued, whereas an
inappropriate means is a means which is not specially suitable for the fulfillment of
the legitimate objective pursued. . . . The question of effectiveness bears upon the
results of the means employed, whereas the question of appropriateness relates more
to the nature of the means employed. (original emphasis)
We agree with the Panel’s interpretation. (original emphasis)
The AB then had to respond to the question who carries the burden of proof to
show that a deviation from an international standard has been TBT consistent.
The AB reversed in this respect, the panel’s holding on the issue of burden of proof
in EC—Sardines, concluding that, when a WTO member does not use a relevant
international standard as the basis for its technical regulation on the ground that
it believes that the standard is either ineffective or inappropriate, the burden still
rests with the complaining party to prove that the international standard at issue
is appropriate and effective for the defendant to reach its objectives (§ 282). The
AB, referring to its case-law under the SPS Agreement, saw no reason to have a
disparate treatment between SPS and TBT Agreements on the issue of the legal
relevance of international standards.
The AB position is hard to defend: in principle, deviation from an international
standard means extra costs on trading partners; substantial negotiating efforts
might take place for (at least) some international standards. The AB effectively
reduces such concerns to irrelevance.¹⁸⁴ From a pure technical perspective,
it seems highly unlikely¹⁸⁵ that the burden of proof will ever shift back to the
defendant: assuming that Peru had demonstrated that the international standard
at hand was effective and appropriate, how could the European Community ever
show the opposite? Assuming it has not done so, its complaint will be rejected
altogether, since, absent proof to the opposite, there is a presumption that WTO
members have behaved in a WTO consistent manner. It is also quite odd that the
burden of proof in these circumstances is assigned to the uninformed, and not

¹⁸⁴ For a very comprehensive analysis on how international standards are being prepared and
their trade relevance, see Sykes (1995).
¹⁸⁵ It all, of course, depends on the quantum of proof needed (burden of persuasion) associated
with what WTO adjudicating bodies have named a prima facie case: if by this term they aimed to
capture not proof but probability that an argument might hold true, then yes, the burden of proof
will shift back and forth many times during proceedings.
 Dealing with NTBs 295

to the informed party.¹⁸⁶ Note finally, that, as Art 2.5 of the TBT makes clear,
whenever an international standard has served as basis for the enactment of a
technical regulation, there will be ipso facto a presumption that the technical regu-
lation at issue is consistent with the necessity requirement (ie it is not more trade-
restrictive than necessary to achieve the stated legitimate objective):
Whenever a technical regulation is prepared, adopted or applied for one of the legitimate
objectives explicitly mentioned in paragraph 2, and is in accordance with relevant inter-
national standards, it shall be rebuttably presumed not to create an unnecessary obstacle
to international trade.
The AB did not pay (not even) lip-service to such considerations, and, as a result,
we ended up with a rather counter-intuitive allocation of the burden of proof.

5.3.3 Unilateral technical regulations

The TBT Agreement makes it clear that technical regulations should not be
enacted unless necessary and, if enacted, they should not be more trade restrictive
than necessary to achieve the stated regulatory objective. Moreover, they should,
whenever feasible, be drafted in terms of performance requirements, and finally,
WTO members are encouraged to recognize each other’s technical regulations as
equivalent and to sign MRAs on how conformity with technical regulations can
be assessed—in particular, with regard to the recognition of results of conformity
assessment procedures.
The TBT Agreement does not enquire as to the motives why a WTO mem-
ber might want to enact a technical regulation. What WTO members cannot
do, however, is to enact technical regulations in order to confer a regulatory sub-
sidy to their domestic production. The disciplines meant to distinguish wheat
from chaff, by leaving intact regulatory interventions meant to pursue a legit-
imate objective, while outlawing the rest are essentially two: non-discrimina-
tion, and the necessity principle. The two disciplines must be cumulatively met:
every technical regulation must be necessary to achieve a legitimate objective¹⁸⁷
and it must be applied in an NT consistent manner (Arts 2.1 and 2.2 of the
TBT). Consequently, the NT obligation in TBT goes one step beyond the
NT obligation in GATT, by adding that all instruments must be necessary, that
is, the least trade restrictive instruments reasonably available to a WTO member

¹⁸⁶ Paradoxically, Peru did not add one iota to its original claim when called to carry the bur-
den of proof under Art 2.4 of the TBT. It still, however, prevailed before the AB. Under the cir-
cumstances, one may wonder why the shift of the burden of proof was necessary? What was its
functionality?
¹⁸⁷ TBT, Art 2.2 contains an indicative list of legitimate objectives. Indicative lists help avoid
Type II errors (false negatives): the WTO judge cannot, for example, argue that national security
(which figures prominently in the list of Art 2.2 of the TBT) is not a legitimate objective. They also
help to identify what other objectives should be regarded as legitimate since, arguably, objectives
which share properties with those included in the list are strong candidates to be recognized as
legitimate.
296 Domestic Instruments

to achieve a unilaterally defined legitimate objective.¹⁸⁸ All international stand-

ards are, by virtue of Art 2.5 of the TBT, presumed necessary to achieve a legit-
imate objective. It bears repetition that necessity also serves as proxy to detect
whether a WTO member indeed pursues a legitimate objective, and not, under
its guise, aims to provide a regulatory subsidy to its domestic producer: having
to pick the least restrictive instruments means, that the intervening state, will
impose the least possible cost on its trading partners, while pursuing a legitim-
ate objective (which, by common agreement, the trading partners decided not
to upset through trade concessions).

5.3.4 Domestic standards

TBT, Art 4.1 states that all WTO members are bound by the disciplines
included in the Code of Good Practice which is included in Annex 3 of the TBT
Agreement. Compliance with the Code of Good Practice amounts ipso facto to
compliance with the principles of the TBT Agreement. TBT, Art 4.2 is pertinent
in this respect:
Standardizing bodies that have accepted and are complying with the Code of Good
Practice shall be acknowledged by the Members as complying with the principles of this
Agreement.
Code of Good Practice, § D reads:
In respect of standards, the standardizing body shall accord treatment to products originat-
ing in the territory of any other Member of the WTO no less favourable than that accorded
to like products of national origin and to like products originating in any other country.
Except for § D, the Code of Good Practice imposes, in essence, the following
obligations:
§ E: standards must respect the necessity principle;
§ F: WTO members must use international standards when appropriate (this
provision corresponds to Art 2.4 of the TBT);
§ G: to further harmonization among nationally defined standards, standard-
izing authorities are encouraged to participate in the work of standardizing
bodies;
§ H: national standardizing bodies should avoid duplication of the work done at
the international level;
§ I: standards should be drafted in terms of performance requirements, when
feasible;
§ J: standardizing authorities should publish at least once every 6 months their
work programme;

¹⁸⁸ We will not be discussing necessity any further here. A detailed discussion of this issue is pre-
sented in Chapter 4, where we discuss public order exceptions (Art XX of the GATT).
 Dealing with NTBs 297

§ K: standardizing authorities should join the ISONET;

§ L: WTO members should allow 60 days, if possible, to lapse between adoption
and entry into force of the standard;
§ M: ex ante transparency is required in the sense that draft standards must be
made available to interested parties to comment upon;
§ N: comments received on draft standards should be, if appropriate, taken into
account when drafting the final text;
§ O: all standards must be published.
§ P: standardizing authorities should provide, if requested, information about
their work programme.

5.4 The WTO SPS Agreement

5.4.1 Coverage
The SPS Agreement (SPS)¹⁸⁹ applies to sanitary and phyto-sanitary measures
(Art 1 of the SPS). Annex A contains the definition of an SPS measure:
1. Sanitary or phytosanitary measure—Any measure applied:
(a) to protect animal or plant life or health within the territory of the Member from
risks arising from the entry, establishment or spread of pests, diseases, disease-
carrying organisms or disease-causing organisms;
(b to protect human or animal life or health within the territory of the Member from
risks arising from additives, contaminants, toxins or disease-causing organisms in
foods, beverages or feedstuffs;
(c) to protect human life or health within the territory of the Member from risks aris-
ing from diseases carried by animals, plants or products thereof, or from the entry,
establishment or spread of pests; or
(d) to prevent or limit other damage within the territory of the Member from the
entry, establishment or spread of pests.

Sanitary or phytosanitary measures include all relevant laws, decrees, regulations,

requirements and procedures including, inter alia, end product criteria; processes and
production methods; testing, inspection, certification and approval procedures; quaran-
tine treatments including relevant requirements associated with the transport of animals
or plants, or with the materials necessary for their survival during transport; provisions
on relevant statistical methods, sampling procedures and methods of risk assessment; and
packaging and labelling requirements directly related to food safety. (original emphasis)

¹⁸⁹ For a comprehensive discussion of the SPS, see Iynedjan (2002).

298 Domestic Instruments

5.4.2 The SPS Agreement in a nutshell

As is the case under the TBT Agreement, WTO members can either use inter-
national standards, or proceed unilaterally and enact their own SPS measures.
Unlike the TBT Agreement, no distinction exists between technical regulations
and standards: all measures coming under the purview of the SPS Agreement are
legally binding on their addressees. The SPS Agreement reflects merely an indica-
tive list of measures coming under its purview (Annex A). The agreement further
makes provisions for the use of expertise, and contains special and differential
treatment provisions aiming at promoting the interests of developing countries.
In parallel with the TBT Agreement, SPS measures can be enacted irrespect-
ive of whether or not domestic production exists. The term ‘ including’, which
appears in Art 2.3 of the SPS, makes this point clear:¹⁹⁰ assuming there is domes-
tic production, the regulating member is required to subject the measure to the
disciplines discussed immediately below (NT), otherwise, the disciplines dis-
cussed are applicable on an MFN basis.
As is the case with the GATT, only measures attributable to a government can
come under the purview of the SPS Agreement. In recent years, however, we have
seen a proliferation of private standards. Besides anecdotal evidence to this effect,
the WTO SPS committee dedicated a session¹⁹¹ to this issue to discuss, inter
alia, a European private standard concerning the commercialization of bananas
which, in the view of the exporters complaining about it, is more stringent than
relevant governmental standards. Discussions so far have not been fruitful, in part
because the SPS Agreement does not include provisions that could adequately
address this issue. WTO members are required, by virtue of Art 13 of the SPS, to
take reasonable measures, in order to ensure that, bodies other than governmental
will observe the SPS Agreement. This provision is not associated with any mean-
ingful procedural requirement to this effect. Consequently, assuming a standard
has its origins in a body which is considered to be a non-governmental body,
there is little an affected party can do.¹⁹²
In contrast to the TBT Agreement, the SPS Agreement identifies several
standard-setting institutions; standards prepared by these institutions (Codex
Alimentarius, International Office of Epozootics, International Plant Protection
Convention) will be accepted as international standards. The list, however, is
indicative: international standards can be recognized as such, even if they do not

¹⁹⁰ Our discussion of this issue under the TBT Agreement is relevant here as well.
¹⁹¹ See WTO Doc G/SPS/ meeting of 29–30 June 2005.
¹⁹² GATT, Art XXIV.12 which regulates the obligations of a central governmental vis à vis
its sub-central entities, offers one possible source of inspiration to interpret the term ‘reasonable’
appearing in Art 13 of the SPS would be. Case-law in that context suggests a rather cautious, non-
intrusive attitude by panels. One would imagine, that at least the same attitude should be observed
when dealing with issues coming under the purview of Art 13 of the SPS: if a panel does not
wish to light-heartedly disturb the relationship between a central government and its sub-central
entities, all the more so it might not want to disturb a relationship between a government and pri-
vate behaviour.
 Dealing with NTBs 299

originate in one of the institutions mentioned in Art 3.4 of the SPS; the SPS com-
mittee (a WTO organ where all WTO members participate) can identify which
other standards by which institutions are relevant as well.¹⁹³

5.4.3 International standards

There is nothing much to add to our discussion of this issue under the TBT
Agreement. In EC—Hormones, the AB held that WTO members that choose not
to avail themselves of the institutional possibility to use an international standard
should not be penalized for their decision to do so (§ 102).
We can thus turn to the study of unilateral SPS measures which, in principle,
must be based on scientific evidence, must be necessary to achieve their stated
objective, must be part and parcel of a wider internally consistent policy, and
must be applied on a non-discriminatory basis. We will take each item in turn in
what follows.

5.4.4 Unilateral measures based on scientific evidence

SPS, Art 2.2 reads:
Members shall ensure that any sanitary or phytosanitary measure is applied only to the
extent necessary to protect human, animal or plant life or health, is based on scientific
principles and is not maintained without sufficient scientific evidence, except as provided
for in paragraph 7 of Article 5.¹⁹⁴
This obligation requires members to ensure that SPS measures are based on
an assessment of the risk involved (Arts 5.1 and 5.2 of the SPS). The AB in
EC—Hormones (§ 250) accepted as much. Following this line of logic, the AB
clarified in Australia—Salmon that a violation of Art 2.2 of the SPS amounted
ipso facto to a violation of Art 5.1 and vice-versa (§ 138). The principle underlying
Arts 2.2 and 5.1 of the SPS is, according to the AB, to strike the appropriate
balance between the interest to promote world trade, and the interest to pro-
tect life and the health of humans (§§ 177 and 180). The AB, in its report on
Japan—Agricultural Products II held that, as per Art 5.2 SPS, a legal requirement
is imposed on WTO members to provide a rational relationship¹⁹⁵ between the
SPS measure enacted, and the available scientific evidence that exists (§ 84).
It is difficult to define what science is, as a matter of WTO law. In
EC—Hormones, the AB held the view that the scientific evidence that should be
evaluated in the risk assessment, need not be understood as a requirement to base

¹⁹³ At the time of writing, no other institution has been identified by the SPS Committee. The
door is, of course, always open. Eventually, one could see for example, the International Health
Regulations of the WHO, or the Cartagena Protocol coming under the purview of Annex A § 3(d).
Besides, the question is still open as to whether panels as well, in the absence of action by the WTO
SPS committee, have the inherent power to identify institutions in this context.
¹⁹⁴ This paragraph reflects the precautionary principle, which is discussed infra.
¹⁹⁵ See the excellent analysis in Dunoff (1999 and 2005) on this score.
300 Domestic Instruments

SPS measures solely on the prevailing opinion in the relevant scientific field. In
its view, SPS measures based on minority scientific opinions could very well be
deemed to be WTO consistent (§ 194). Beyond this statement, there is nothing
much in WTO case-law so far. The AB dismissed, in the EC—Hormones dispute,
evidence submitted orally and not substantiated by scientific experiments, with-
out defining what the minimum requirements to define are.¹⁹⁶
Scientific evidence will need to point to a risk for an SPS measure to be legit-
imately enacted. In EC—Hormones, the AB made two important clarifications
regarding the definition of risk. First, it explained that the risk must be identifi-
able, as opposed to a mere hypothetical possibility (§ 186):
. . . In one part of its Reports, the Panel opposes a requirement of an “identifiable risk” to
the uncertainty that theoretically always remains since science can never provide absolute
certainty that a given substance will not ever have adverse health effects. We agree with
the Panel that this theoretical uncertainty is not the kind of risk which, under Article 5.1,
is to be assessed. (original emphasis)
Second, the AB held the view that the risk envisaged in the SPS Agreement is not
just a laboratory risk but a real life risk that takes into account behavioural factors
(§ 187):
. . . It is essential to bear in mind that the risk that is to be evaluated in a risk assessment
under Article 5.1 is not only risk ascertainable in a science laboratory operating under
strictly controlled conditions, but also risk in human societies as they actually exist, in
other words, the actual potential for adverse effects on human health in the real world
where people live and work and die.¹⁹⁷
The first statement is not an example of coherence: when reference is made to
risk assessment, one usually knows the distribution of probabilities that the risk
might occur; there is, thus, certainty as to the probability that a risk might occur.
Societies wanting to take measures when there is uncertainty as to whether a risk

¹⁹⁶ One can, of course, enter into interminal debates about what is and what is not scientific
discovery, see, inter alia, Kuhn (1962), but also Popper (1992) and his verification/falsification
criteria. Various courts have had to deal with this issue, and, in other jurisdictions, case-law has
come up with a test to distinguish science from non-science. For instance, in Daubert (Daubert v.
Merrill Dow Pharmaceuticals, Inc. 509 US 579 (1993)), the US Supreme Court (SCt) stood for the
proposition that, for an opinion to be accepted as scientific opinion, the methodological specificities
prevailing in a particular scientific field must be established. Absent respect of the relevant meth-
odology, evidence submitted would not be held scientific, in the eyes of the US SCt. In the WTO
context, beyond banal statements, we are still in the dark as to what distinguishes science from non-
science. See the excellent analysis by Trebilcock and Soloway (2002), where the authors reflect their
disappointment with the manner in which the AB has treated such an important issue. They con-
clude that we simply do not know what science is in the SPS sense of the term, although implicitly
at least, the AB has accepted that expertise provided to it by various experts appearing before it by
virtue of Art 11.2 of the SPS Agreement is scientific. Similar thoughts have been expressed by Horn
and Mavroidis (2003). Pauwelyn (1999) as well, in his comprehensive account of the first three SPS
cases submitted to the WTO, discusses at length this issue and ends up with similar conclusions.
¹⁹⁷ The panel in its report on Japan—Apples (Art 21.5—US) dismissed the relevance of two
studies, because they did not correspond to natural conditions, see §§ 8.65 and 8.140ff.
 Dealing with NTBs 301

might occur are not risk-averse, but ambiguity-averse societies. The two concepts
are related but are not identical.¹⁹⁸ The second statement is difficult to decipher
since scientific evidence normally occurs in laboratories aimed at producing
results useful to the real world where normal people live and die. Probably, the
AB wanted to state that, when transposing the outcome of a scientific experiment
in regulatory language, there is no need to transpose it in an absolutely faithful (to
the science) manner; some deviation will be tolerated and, eventually, forgiven.
There are two kinds of risk assessment: all risks arising from the presence of
additives, contaminants, toxins or disease-causing organisms in food, beverages or
feedstuffs, must be assessed and their potential effects on human or animal life
evaluated (risk assessment I); on the other hand, as far as pests or diseases are con-
cerned, the SPS Agreement provides for assessment of the likelihood of a pest
or disease entering, establishing and spreading and the associated potential bio-
logical and economic consequences (risk assessment II).
In Japan—Apples, the issue was whether the SPS Agreement prejudges
the methodology to be used in the context of risk assessment. The AB found
that the agreement does not impose a particular methodology to this effect
(§ 204).¹⁹⁹ In the same report, the AB, confirming prior case-law on this issue (ie
EC—Hormones), held that the obligation to base measures on risk assessment as
expressed in § 4 of Annex A to the SPS Agreement, entails that the WTO mem-
ber wishing to enact an SPS measure, cannot carry out a risk assessment in a
manner that precludes phytosanitary measures other than the scheme already in
place to be considered. The AB found (§ 209) that Japan violated its obligations
under the SPS Agreement by conducting risk assessment justifying the measure
it had in place, without, however, inquiring into the possibility of other, poten-
tially applicable, measures. In this regard, the AB concluded that the SPS does
not request that a particular methodology be applied, but that any methodology
chosen must be specific to (in close connection with) the factual situation inves-
tigated. Moreover, WTO members must not perform a risk assessment aiming at
justifying the measure already in place; rather, in the AB’s view, when choosing
their methodology they must make sure that they make room for investigating
whether measures, other than those in place, could appropriately address the fac-
tual situation at hand.
The AB explained in its report on EC—Hormones that a WTO member can
base its measures on risk assessment performed either by another WTO member,
or by an international organization (§ 190). In the same report, the AB reversed a

¹⁹⁸ Ambiguity-aversion is probably what is captured by the precautionary principle, which we

discuss infra. On this issue, see Gollier et al (2001) and also Sunstein (2002).
¹⁹⁹ However, depending on the interpretation of Art 5.1, one could legitimately hold the view
that the SPS Agreement does impose a certain methodology, that is, the methodology privileged
by the relevant international organizations. The AB seems to suggest that risk assessment techniques
and methodology used are two distinct issues. There is not much support for this view, however, in
scientific discourse.
302 Domestic Instruments

finding by the panel to the effect that Arts 2.2 and 5.1 of the SPS impose a mini-
mum procedural requirement, in the sense that a WTO member adopting an SPS
measure must provide evidence that it did base its measure on scientific evidence.
In the case at hand, the panel found no evidence in the body of the EC regulation
(reflecting the SPS measure), or in its preamble, that the European Community
had indeed based its measure on science (§§ 188–90).²⁰⁰
Through risk assessment, WTO members aim at determining the probabil-
ity that a certain risk materializes. Assuming existence of risk,²⁰¹ it is for WTO
members to decide whether to intervene or not. Throughout the GATT years,
and until Japan—Apples, it was thought that a WTO adjudicating body could
not prejudge the level of risk aversion that a given society unilaterally sets. If at
all, what was justiciable was not the ends sought, but the means serving ends.²⁰²
This approach was in line with the negative integration character of the GATT
and was repeated in all cases dealing with Art XX of the GATT. In the WTO era,
the AB reaffirmed this point in the most unambiguous terms in EC—Hormones
(§ 186):
To the extent that the Panel purported to require a risk assessment to establish a minimum
magnitude of risk, we must note that imposition of such a quantitative requirement finds
no basis in the SPS Agreement. A panel is authorized only to determine whether a given
SPS measure is ‘based on’ risk assessment.
Then came Japan—Apples.²⁰³ Japan imposed a series of measures to block trade
of apples originating in the United States, fearing that some of them might suffer
from fire blight. Fire blight affects apples and not human life.²⁰⁴ There was no
evidence that apples infected with fire blight had been exported to Japan,
although there was evidence of a shipment of infected apples to the separate

²⁰⁰ It is true that a textual reading of the provisions does not impose such a requirement, how-
ever, from an evidentiary perspective it is almost impossible for uninformed parties to discern
the basis of an SPS measure unless some evidence is provided. In fact, pursuant to § 1 of Annex
B, WTO members must ensure, ‘… that all sanitary and phytosanitary regulations which have
been adopted are published promptly in such a manner as to enable interested members to become
acquainted with them’. Annex B, § 3 further requires members to introduce enquiry points whereby
interested parties can request (and obtain) responses to reasonable queries that they might have.
All the panel did was to read in context (since transparency is provided for in the SPS Agreement)
the articles concerning evidence that there is scientific basis for the risk assessment performed.
Following the AB’s ruling the door to ex post facto rationalizations is not shut.
²⁰¹ There is of course, always a certain degree of uncertainty surrounding such experiments, due,
inter alia, to the ever evolving nature of science. Although history suggests that paradigm shifts
might be infrequent, still, measurement undergoes a steady evolution, see Kuhn (1962 and 1996).
²⁰² Th is attitude, theoretically, can of course lead to perverse outcomes: since there is wide-
spread uncertainty as to a number of issues, a very ambiguity-averse society could, essentially, stop
trade altogether. However, such a danger is not realistic: in light of the national treatment obliga-
tion it would not be stopping only trade, but domestic production as well. Moreover, governments
are not necessarily irrational.
²⁰³ Neven and Weiler (2006) have provided an excellent, critical account of this dispute.
²⁰⁴ At least according to today’s scientific knowledge.
 Dealing with NTBs 303

customs territory of Taiwan, Penghu, Kinmen, and Matsu.²⁰⁵ However, the

apples shipped to Taiwan, Penghu, Kinmen, and Matsu were not infected with
fire blight, but rather from another disease (codling moth). Japan claimed that
it possessed sufficient scientific evidence that risk existed. The record shows that
apples infected with fire blight had been exported to New Zealand in the early
twentieth century, to the United Kingdom in the 1950s, and to Egypt a little
later. However, scientific evidence showed (with a considerable degree of confi-
dence) that the disease had been transmitted because of the nature of the trade
involved (trade in root stocks, that is, in apple trees and not in apples). Since the
trade involved in Japan—Apples dispute concerned apples and not apple trees, the
two situations were not comparable.
The expertise provided to the panel suggested that the risk of completing the
pathway (and thus, transmitting the disease) was negligible. The panel rejected
the argument of the United States that it was required to confine its review only
to mature, symptom-less apples, since only such apples were being exported to
Japan. The panel nevertheless went ahead with its assessment and held the risk
(by human error) that immature, symptom-full apples be exported to Japan
should also be taken into account. In the experts’ view, however, even if this were
the case the risk was still negligible because the disease could only be transmitted
through birds flying from infected apples to uninfected apple trees. The panel,
however, relying on the guidance of the experts on this point, did not hold that
there was no risk at all. In fact, the Japan—Apples panel explicitly accepted that
there was negligible risk. It still went ahead, nevertheless, and found, that there
was no rational or objective relationship between the measure and the relevant
scientific evidence. In light of the negligible risk identified on the basis of the sci-
entific evidence and the nature of elements composing the measure, the panel
concluded that the Japanese measure was clearly disproportionate to the risk
(§ 8.198 and 8.199 of the panel report). In other words, according to the panel,
the absence of a rational or objective relationship equals absence of proportionality.
The challenged measure was in fact nine prohibitions or requirements (§§ 8.5ff of
the panel report). Among these measures was the prohibition of importing apples
from US states other than Oregon and Washington, certification requirements
and the prohibition of exports if fire blight had been detected in a neighbouring
geographical area. According to the Japanese regulator, in the absence of a total
embargo on imports, the risk that the disease would spread was very much alive.
Critically, the panel did not address this issue under Art 5.6 of the SPS (the
necessity requirement). It decided to exercise judicial economy in this respect. It
held that the measure was inconsistent with Art 2.2 of the SPS, since there was no
scientific evidence to support it, although it accepted that risk, albeit negligible,
did exist.

²⁰⁵ See § 160 of the AB report and footnote 289.

304 Domestic Instruments

The AB, approved the rational or objective relationship as one of the appro-
priate methodologies to establish whether an SPS measure is being maintained
without scientific evidence (§ 164). It thus endorsed the idea that in light of the
negligible risk, the challenged measure was disproportionate (§§ 160 and 163ff ).
Although the necessity requirement has an impact on the analysis as well, the AB
followed the panel’s lead by not conducting an analysis either: rather, the AB held
that the challenged measure was disproportionate as it was maintained without
any scientific evidence supporting it. The foundation of the AB’s decision here is
not unproblematic. A number of questions are raised: where do we draw the line
between negligible and non-negligible risk? Should a judicial body such as the AB
be entrusted with the responsibility to define the level of risk aversion for a given
society? Do these bodies carry such legitimacy? Even more disturbing is the AB’s
finding that when scientific evidence is insufficient (as opposed to uncertainty, see
infra), a WTO member cannot have recourse to precautionary measures (§ 188).
Hence, in the case at hand, where science has not categorically pronounced on
the absence of risk, but accepts some, albeit negligible risk, Japan cannot protect
itself (since it can neither base its SPS measure on science, nor on precaution).²⁰⁶
Basing a measure on scientific evidence is only the first step; WTO members
must also ensure that their measures are necessary and consistent. The neces-
sity requirement is reflected in Art 5.6 of the SPS,²⁰⁷ whereas the consistency
requirement is discussed in Art 5.5 of the SPS, which requires consistency in the

²⁰⁶ Had the panel and the AB addressed the issue under Art 5.6 of the SPS (as opposed to Art
2.2 of the SPS) then, probably, a different result would have been reached. Actually, there is a
logical connection between the two provisions. SPS, Art 2.2 in a sense is a reflection of the neces-
sity requirement: absent scientific justification, it is unnecessary to keep in place SPS measures. It
thus, has both an ex ante and an ex post function: it is a threshold issue for any SPS measure to be
enacted; it also prejudges the lifetime of a measure since, assuming new scientific evidence con-
cludes that, say, no risk exists at all from GMOs, a measure adopted to address such issues must
be repealed. SPS, Art 5.6 places the necessity requirement in its trade context: assuming there is
evidence that a risk exists, a measure taken to address such risk should not be more trade restrictive
than necessary to address it (the risk at hand). SPS, Art 5.6 therefore balances the positive effects
of addressing the risk with the negative effects that implementation of such a measure entails for
the trading partners of the intervening state. There is hence a sequence between the two provisions:
if it is demonstrated that a certain risk is not purely theoretical (in the AB’s terminology), but real,
albeit negligible, it means that there is at least some scientific evidence to the effect that such a risk
exists. This is what the panel and the AB understood to be the case in Japan—Apples. In such a
scenario, they should have logically left the room occupied by Art 2.2 and gone on to address the
trade necessity of the measure under Art 5.6. There the analysis should have been whether, in light of
the negligibility of the risk, the measures applied by Japan satisfied Art 5.6. Assuming that a differ-
ent yet equally effective and reasonable measure was available to Japan, it would have failed on that
score. Regrettably, this is not what the AB did. Instead, it held the proposition that, in the presence
of an undefined threshold (which both the panel and the AB conveniently called negligible risk), the
measures chosen were unsupported by scientific evidence. In going down this route, both WTO
adjudication bodies left themselves vulnerable to criticism. See on this score, the excellent analysis
in Neven and Weiler (2006).
²⁰⁷ We have already discussed the necessity requirement in this chapter. There is no need to add
anything to the discussion above.
 Dealing with NTBs 305

appropriate level of protection across situations.²⁰⁸ WTO case-law (AB report on

EC—Hormones in §§ 212 and 238) has established that this provision should
be read together with Art 2.3 of the SPS, which requires non-discriminatory
behaviour.
Recall that Art 2.3 of the SPS reads:
Members shall ensure that their sanitary and phytosanitary measures do not arbitrar-
ily or unjustifiably discriminate between members where identical or similar conditions
prevail, including between their own territory and that of other members. Sanitary and
phytosanitary measures shall not be applied in a manner which would constitute a dis-
guised restriction on international trade.
In Australia—Salmon, the AB went one step further and explained that a vio-
lation of Art 5.5 of the SPS ipso facto entails a violation of Art 2.3 of the SPS
(§ 252). The natural conclusion, when reading these two provisions together, is
that Art 5.5 of the SPS requires consistency across situations across WTO mem-
bers.²⁰⁹ Whereas the latter part (non-discrimination across WTO members)
seems easier to grasp, the former part (consistency across situations) is far from
being a walk in the park for the adjudicator.
The WTO members adopted on 22 June 2000²¹⁰ Guidelines, which pro-
vide some clarification on the scope of the obligation assumed under Art 5.5 of
the SPS.
WTO members must, according to the Guidelines, in order to observe their
consistency (coherence) obligation, indicate the level of protection which they
consider appropriate, and also indicate if there is a difference in the level of pro-
tection under consideration and levels already determined by the member in dif-
ferent situations: WTO members must further compare the level of protection
now being sought with that already considered in previous situations which con-
tain sufficient common elements so as to render the two situations comparable.²¹¹
It follows that the Guidelines attach substantial importance on the compar-
ability across transactions. Comparability across transactions is of course the key
input to decide whether discrimination has occurred under Art 2.3 of the SPS as
well. Various factors can, of course, affect comparability. Case-law has contrib-
uted some clarifications on this score:
(a) First, in EC—Hormones, the AB held that, for a violation of Art 5.5 of
the SPS to be established, a complaining party must satisfy the elements of a

²⁰⁸ On the economics of Art 5.5 of the SPS, see the excellent analysis in Pienaar (2003).
²⁰⁹ SPS, Art 6 states that SPS measures are not required to be directed against all exports when
a pest or a disease has surfaced, which necessitated the adoption of the SPS measure. Pest-free and
disease-free areas (among WTO members, of course) can legitimately be excluded. The terms ‘pest-
free or disease-free area’ and ‘area of low pest or disease prevalence’ are further detailed in §§ 6 and
7 of Annex A to the SPS Agreement.
²¹⁰ See WTO Doc G/SPS/15 of 18 July 2000.
²¹¹ See p 3 of the Guidelines, op cit, under A4. The italicized terms appear as such in the body of
the Guidelines.
306 Domestic Instruments

three-prong test (§§ 214–15):

Close inspection of Article 5.5 indicates that a complaint of violation of this Article must
show the presence of three distinct elements. The first element is that the Member impos-
ing the measure complained of has adopted its own appropriate levels of sanitary pro-
tection against risks to human life or health in several different situations. The second
element to be shown is that those levels of protection exhibit arbitrary or unjustifiable
differences (“distinctions” in the language of Article 5.5) in their treatment of different
situations. The last element requires that the arbitrary or unjustifiable differences result
in discrimination or a disguised restriction of international trade. We understand the last
element to be referring to the measure embodying or implementing a particular level of
protection as resulting, in its application, in discrimination or a disguised restriction on
international trade.
We consider the above three elements of Article 5.5 to be cumulative in nature; all
of them must be demonstrated to be present if violation of Article 5.5 is to be found. In
particular, both the second and third elements must be found. The second element alone
would not suffice. The third element must also be demonstrably present: the implement-
ing measure must be shown to be applied in such a manner as to result in discrimination
or a disguised restriction on international trade. The presence of the second element—
the arbitrary or unjustifiable character of differences in levels of protection considered
by a Member as appropriate in differing situations—may in practical effect operate as a
“warning” signal that the implementing measure in its application might be a discrimin-
atory measure or might be a restriction on international trade disguised as an SPS meas-
ure for the protection of human life or health. Nevertheless, the measure itself needs
to be examined and appraised and, in the context of the differing levels of protection,
shown to result in discrimination or a disguised restriction on international trade. (ori-
ginal emphasis)
To perform this test, however, the complaining party needs to first establish com-
parability across situations (§ 217 of the AB report on EC—Hormones).
(b) Second, the AB explained in EC—Hormones that the letter and the spirit
of Art 5.5 of the SPS does not require WTO members to guarantee an absolute
uniformity across the various appropriate levels of protection that they pursue. In
its view, Art 5.5 of the SPS should be properly understood as a legal prohibition of
arbitrary or unjustifiable discrimination (§ 213).
So far, so good. However, the uphill battle is still ahead since the real ques-
tion is: what is the benchmark for comparability? In Australia—Salmon, the AB
offered such a benchmark (§§ 146 and 152):
. . . the panel was correct in stating that situations can be compared under Article 5.5
if these situations involve either a risk of entry, establishment or spread of the same or a
similar disease, or a risk of the same or similar “associated potential biological and eco-
nomic consequences”.
...
. . . we believe that for situations to be comparable under Article 5.5, it is sufficient
for these situations to have in common a risk of entry, establishment or spread of one
 Dealing with NTBs 307
disease of concern. There is no need for these situations to have in common a risk of entry,
establishment or spread of all diseases of concern.” (original emphasis)
Therefore, similarity of the disease or of the associated risk provides comparability.
Following this line of reasoning, the AB introduced in Australia—Salmon the
term ‘warning signals’, which refers to elements or properties of particular SPS
measures that could be relevant in establishing a violation of Art 5.5 of the SPS.
The quantity and quality of such warning signals will ultimately prove to be the
decisive factor in determining whether Art 5.5 of the SPS has been violated: sub-
stantial difference in the level of protection (§ 164), and/or violation of Art 5.1 of
the SPS (§ 166), could serve as warning signals:
. . . in this case the degree of difference in the levels of protection (prohibition versus tol-
erance) is indeed, as the panel stated, “rather substantial”. We, therefore, consider it legit-
imate to treat this difference as a separate warning signal.
...
. . . We note that a finding that an SPS measure is not based on an assessment of the
risks to human, animal or plant life or health—either because there was no risk assess-
ment at all or because there is an insufficient risk assessment—is a strong indication that
this measure is not really concerned with the protection of human, animal or plant life or
health but is instead a trade-restrictive measure taken in the guise of an SPS measure, i.e.,
a “disguised restriction on international trade”. We, therefore, consider that the finding
of inconsistency with Article 5.1 is an appropriate warning signal for a “disguised restric-
tion on international trade”.²¹²
The AB seems to have cast too wide a net: comparability across risks might cover
dozens of situations. For instance, assuming the AB had been requested by the
complainant to resolve the issue in the EC—Asbestos dispute under the SPS
Agreement, should it have impugned France on the basis that, on the one hand, it
took measures banning the sale of asbestos containing construction material but,
on the other, did nothing to stop the sale of cigarettes? Both items (asbestos con-
taining construction material and cigarettes) are health-impairing. Both might
also contribute to the same disease. It is possible that the risk distribution in
either case is more or less comparable. Moreover, it could further be the case that
France produces cigarettes but no asbestos containing construction material. But
is this enough to lay blame on France for being inconsistent (in the sense of Art
5.5 of the SPS)? The SPS Agreement is part and parcel of a wider trade agreement,
the WTO. SPS measures, in the absence of the SPS Agreement, would have been
adjudicated under Arts III/XX of the GATT. The purpose of a trade agreement,
we recall, is to address external effects stemming from unilateral definition of
trade policies. Domestic instruments (such as SPS measures), are not enumerated
one by one in the contract. The incompleteness of the contract coupled with the
potential for opportunism make the need for a test to distinguish wheat from

²¹² Evidence of this attitude can also be traced in § 240 of the AB’s report on EC—Hormones.
308 Domestic Instruments

chaff imperative. A test based on non-discrimination only is highly imperfect,

as we saw in our discussion of Art III of the GATT. The SPS adds some proxies,
such as scientific evidence, necessity, coherence in an effort to (eventually) help
the judge in his/her judgment to outlaw beggar thy neighbour policies, while not
disturbing measures that genuinely pursue a health/environmental objective. A
presumption that protectionism has been sought is much stronger when the test
of Art 5.5 of the SPS is confined within a relevant product market.²¹³
Moreover, casting the net too wide has one important consequence: WTO pan-
els and the AB become the arbiters of consistency in the formulation of national
health policies. The risk for false positives by the WTO adjudicating bodies in the
field of health related matters is not comparable, both from a human, and from
an institutional perspective, to the risk of false positives say in the field of AD.
It is submitted that restricting the test, for the purposes of Art 5.5 of the SPS
analysis, to an examination of national (health, environmental) policies within
the relevant product market under consideration is consistent with different
criteria, some of them explicitly and some of them implicitly mentioned in the
Guidelines as well: comparability of the risk distribution is such an element; so is
comparability of the disease (similar disease in the Guidelines lingo).
At least these factors should be evaluated cumulatively. Were one to isolate, for
example, comparability of risk distribution, one might end up with results that
would hardly make sense within a trade agreement context. Assume, for example,
that a consumer smoking a fi xed number of cigarettes and a fi xed number of
hormone-treated beef incurs the same probability to suffer from the same disease.
Assume further that a country bans hormone-treated beef, but simply taxes sales
of cigarettes. Has this country violated Art 5.5 of the SPS? Yes, if comparability
of risk distribution and/or disease are viewed as the only relevant criteria to decide
on the consistency requirement. No, were one to control for the purpose of sign-
ing the SPS: it is highly unlikely, if not implausible altogether, that consumers
of hormone-treated beef will start smoking more as a result of the ban on beef.
Moreover, the tax on cigarettes will be borne by both domestic and foreign pro-
ducers. By the same token adjustment costs have been imposed to both domestic
and foreign beef producers. This is hardly a case of discrimination, if by this term,
in a trade context, we understand a ban on protectionism.
In light of the fact that Art 5.5 of the SPS deals with more detail (than Art 2.3 of
the SPS) with discrimination, it seems, in principle at least, plausible to argue that
this provision reflects a different facet of the non-discrimination obligation. This
addition is necessary in light of the idiosyncratic character that we must acknow-
ledge to the non-discrimination obligation as embedded in the SPS Agreement.
An illustration is definitely helpful in this context. Take the EC—Hormones case,
for example, and the manner in which the AB treated the fact that the defend-
ant had not imposed a similar measure on hormone-treated pork (as it did with

²¹³ More on this score, later. See nevertheless, the analysis in Horn and Mavroidis (2003).
 Dealing with NTBs 309

respect to hormone-treated beef). The European Community must, by virtue of

Art 2.3 of the SPS, ban sales of hormone-treated beef irrespective of its origin: Art
2.3 of the SPS, in other words, covers only the identical product market and not
the relevant product market.²¹⁴ SPS, Art 5.5 extends this obligation to the rele-
vant product market: the European Community must, by virtue of this provision,
examine whether a ban on hormone-treated products, which are substitutable to
beef, and, provided that at least the comparability factors mentioned above have
been satisfactorily met, is warranted.²¹⁵
In this vein, the obligation reflected in Art 5.5 of the SPS cements the obliga-
tion reflected in Art 2.3 of the SPS; it serves as an anti-circumvention provision:
WTO members will not arbitrarily choose levels of protection when faced with
comparable risks stemming from comparable diseases within a relevant product
market.

5.4.5 Measures adopted on the basis of precaution

WTO members can, in accordance with Art 5.7 of the SPS, adopt provisionally,
even in the absence of scientific backing, SPS measures on the basis of available
information. In this case, however, members are under an obligation to collect
information that will enable them to perform a risk assessment within a reasonable
period of time. Although Art 5.7 of the SPS does not explicitly refer to the precau-
tionary principle, WTO adjudicating bodies have held that this provision reflects
indeed the precautionary principle. In EC—Hormones, the AB summarized the
relevance of the precautionary principle in the WTO, in the following terms (§§
123–5):
(1) the precautionary principle is reflected in Art 5.7 of the SPS, but is also reflected
in other SPS provisions such as the preamble and Art 3.3 of the SPS. Hence,
the precautionary principle is not exhaustively reflected in Art 5.7 of the SPS;
(2) the status of precautionary principle under customary international law is
unclear;²¹⁶
(3) WTO panels should keep precautionary principle in mind when interpreting
the SPS; but

²¹⁴ The argument here is that a very narrow construction of the obligation included in Art 3.2 of
the SPS is warranted. Th is is probably not always the correct level of disaggregation: if, for example,
growth hormones might have a differential effect on the animal (and thus present a different risk
for human health) depending on the age of the animal when they are injected (the growth hor-
mone), then a WTO member should legitimately control for such differences without risking a
panel finding that it has violated its obligations under Art 3.2 of the SPS.
²¹⁵ The appropriate level of protection in the two cases must not be identical; it suffices that it is
comparable.
²¹⁶ In EC—Approval and Marketing of Biotech Products, the panel, when dealing with the EC
régime for approval of GMOs even cited the International Tribunal of the Law of the Sea (ITLOS)
as support for this finding that the precautionary principle had an uncertain status under custom-
ary international law (§ 7.89).
310 Domestic Instruments

(4) the precautionary principle does not override the explicit wording of specific
SPS provisions.
The AB in Japan—Agricultural Products II ²¹⁷ explained the relationship between
the possibility to invoke precautionary principle, and the obligation to base meas-
ures on scientific evidence in the following terms (§ 80):
. . . Article 5.7 allows Members to adopt provisional SPS measures “[I]n cases where
relevant scientific evidence is insufficient” and certain other requirements are fulfilled.
Article 5.7 operates as a qualified exemption from the obligation under Article 2.2 not
to maintain SPS measures without sufficient scientific evidence. An overly broad and
flexible interpretation of that obligation would render Article 5.7 meaningless. (original
emphasis)
The AB, hence, held that Art 5.7 of the SPS is an exception from the obligation
to use scientific evidence when enacting SPS measures. In Japan—Apples, the AB
went one step further and clarified its understanding of the relationship between
the two provisions. In its view, if science is well settled on an issue, recourse to
precaution is unwarranted. In a sense, the AB sees a firewall between scientific
evidence and precaution (§ 184):
The application of Article 5.7 is triggered not by the existence of scientific uncertainty,
but rather by the insufficiency of scientific evidence. The text of Article 5.7 is clear: it
refers to ‘cases where relevant scientific evidence is insufficient’, not to ‘scientific uncer-
tainty’. The two concepts are not interchangeable. Therefore, we are unable to endorse
Japan’s approach of interpreting Article 5.7 through the prism of ‘scientific uncertainty’.
Now, that is a bold statement. Where does one draw the line between scientific
uncertainty and scientific insufficiency? Although the text of Art 5.7 of the SPS
does in fact mention insufficiency and not uncertainty, is this enough of a basis to
proceed in such a categorical fashion as the AB did? Admittedly, the precaution-
ary principle is not an easy concept. Many analysts have trouble trying to ‘pin
it down’ to a measurable dimension.²¹⁸ Irrespective of the AB’s qualification on
this issue, which is the product of an exaggerated textual approach to interpret-
ation rather than logical reasoning (since the AB nowhere explains the dichotomy
between insufficiency of evidence, and scientific uncertainty), science can some-
times persuasively point in a certain direction and sometimes not.²¹⁹ Assuming
the probability that a risk might occur is known (certainty), a member will choose
whether to intervene or not, depending on its preferred level of risk aversion.
Assuming, however, that this is not the case (ie assume that there is ambiguity
as to the realization of a risk), an ambiguity-averse society might still prefer to

²¹⁷ See Dunoff (2005) for an excellent account of this dispute.

²¹⁸ See for example the diverging views of Sunstein (2002 and 2003), and Scott (2003) on the
others, on this issue.
²¹⁹ See the excellent and highly informative analysis in Gollier et al (2000). See also the relevant
passage in Neven and Weiler (2006).
 Dealing with NTBs 311

intervene, rather than stay idle. The significance of the precautionary principle
should lie in the ability of WTO members to intervene when certainty about risk
distribution is lacking.²²⁰ This is the first interpretative error in Japan—Apples.
The second error (and probably, the more dramatic)²²¹ is the firewall estab-
lished between precaution and scientific evidence; between uncertainty and cer-
tainty. Gollier et al (2000) explain that it is quite a daring move to establish a
firewall in this context: the two concepts are better viewed as being in a state of
osmosis; uncertainty might lead to certainty, and it is not improbable that the
latter might lead to uncertainty again. This description captures to a wide extent,
the history of scientific evolution which judicial pronouncements should avoid
undermining.²²² We will return to this issue shortly.
It follows, that, in the eyes of the AB, a measure can either be based on science
or on Art 5.7 of the SPS, and that recourse to the latter is appropriate in cases of
scientific insufficiency, but not of scientific uncertainty. WTO case-law has also
dealt with the mechanics of the compliance (with Art 5.7 of the SPS) test. The
AB, in its report on Japan—Agricultural Products II established a four-prong test
that must be met for a measure to be deemed consistent with Art 5.7 of the SPS.
As enunciated by the AB, all four elements of the test must be cumulatively met
(§ 89):
Article 5.7 of the SPS Agreement sets out four requirements which must be met in order to
adopt and maintain a provisional SPS measure. Pursuant to the first sentence of Article
5.7, a Member may provisionally adopt an SPS measure if this measure is:
(1) imposed in respect of a situation where “relevant scientific information is insuffi-
cient”; and
(2) adopted “on the basis of available pertinent information”.
Pursuant to the second sentence of Article 5.7, such a provisional measure may not be
maintained unless the Member which adopted the measure:
(1) “seek[s] to obtain the additional information necessary for a more objective assess-
ment of risk”; and
(2) “review[s] the . . . measure accordingly within a reasonable period of time”.
These four requirements are clearly cumulative in nature and are equally important for
the purpose of determining consistency with this provision. Whenever one of these four
requirements is not met, the measure at issue is inconsistent with Article 5.7. (original
emphasis)

²²⁰ It is of, course, often very difficult to talk of certainty in science. Certainty is intimately
linked to the predictive power.
²²¹ On this issue, we follow Neven and Weiler (2006).
²²² Gollier et al (2001) represent here the mainstream scientific view, see Stirling (2001). Viewed
in this light, the AB report on EC—Hormones was probably moving in the right direction when
it held that the precautionary principle is not confined to Art 5.7 of the SPS. The firewall was the
invention of Japan—Apples.
312 Domestic Instruments

Case-law has contributed some additional clarifications. The term ‘reasonable

period of time’, for example, was discussed in the AB report on Japan—Agricultural
Products II. There, the AB held that this term will have to be interpreted on a case
by case basis. In this particular case, 4 years of inaction by Japan subsequent to
the adoption of a measure under Art 5.7 of the SPS was deemed to be unrea-
sonable (§ 93). In the same dispute, the AB held that the additional information
sought during the reasonable period of time must be germane in conducting risk
assessment (§ 92).
Case-law, thus, has refrained from adding anything meaningful regarding the
interpretation of the substantive content of the precautionary principle. It has
provided some procedural clarifications and nothing beyond that. Could we risk
that Art 5.7 of the SPS should be understood to encompass the non-discrimi-
nation obligation, at least as viewed in Art 3.2 of the SPS? And what about the
consistency and necessity obligations embedded in Arts 5.5 and 5.6 of the SPS
respectively? Does Art 5.7 of the SPS encompass them as well?
To respond to these questions, one needs to be clear about the function of Art
5.7 of the SPS. One natural place to start this discussion is the travaux prépara-
toires. The negotiating history of this provision²²³ suggests that it was drafted to
deal with emergency situations, such as an outbreak of a disease. It is the AB in its
report on EC—Hormones that extended Art 5.7 of the SPS to cover precaution-
ary measures as well. Consequently, were Art 5.7 of the SPS to be limited to emer-
gency situations, then, absent extraordinary circumstances, it would be difficult
to envisage application of the non-discrimination obligation, let alone necessity,
or consistency.
Extending Art 5.7 of the SPS to cover precautionary measures as well prob-
ably changes the picture: precautionary measures do not necessarily have to be
taken under urgency. The European Community circulated a proposal inviting
the WTO membership to think further about the content of the precaution-
ary principle:²²⁴ in its view, measures based on precaution must be proportional
to the chosen level of protection, non-discriminatory in their application, and
also consistent with past practice (assuming, of course, that comparability across
transactions has been established).²²⁵ To appreciate the EC proposal, it is prob-
ably warranted to see how the EC institutions understand precaution.
One of the leading cases in EC-law in this field is Commission v. Netherlands,²²⁶
where the court outlawed a Dutch measure banning marketing of food-
stuffs to which nutrients had been added. The court made it first clear that the

²²³ See, inter alia, GATT Docs MTN.GNG/NG5/WGSP/7 of 20 November 1990, and MTN.
GNG/NG5/WGSP/17 of 30 April 1990.
²²⁴ See WTO Doc G/SPS/GEN/168 of 14 March 2000.
²²⁵ See also the European Council Resolution on the Precautionary Principle, WTO Doc
G/SPS/GEN/225 of 2 February 2001.
²²⁶ C-41/02 Commission of the European Communities v Kingdom of the Netherlands [2004]
ECR I-11375.
 Dealing with NTBs 313

proportionality requirement (adopt the least restrictive for trade measure) should
be observed in cases where the precautionary principle has been invoked as well
(§ 46). It then moved on to provide its understanding that the precautionary
principle should not be dissociated completely from scientific evidence. The court
adopts a, correct in our view, understanding of precaution as one line in a con-
tinuum that might lead to scientific proof (§§ 51–2):
A proper application of the precautionary principle requires, in the first place, the iden-
tification of the potentially negative consequences for health of the proposed addition of
nutrients, and, secondly, a comprehensive assessment of the risk for health based on the
most reliable scientific data available and the most recent results of international research
(see Commission v Denmark, paragraph 51).
Where it proves to be impossible to determine with certainty the existence or extent of
the alleged risk because of the insufficiency, inconclusiveness or imprecision of the results
of studies conducted, but the likelihood of real harm to public health persists should the
risk materialise, the precautionary principle justifies the adoption of restrictive measures
(see Commission v Denmark, paragraphs 52 and 53).
It should come thus as no surprise that proportionality must be respected even
when recourse to precaution has been made. This view (that is, to dismantle a fire-
wall between uncertainty, to which precaution could be an antidote, and science)
has strong underpinnings in the epistemic community. In a very elaborate docu-
ment prepared for the European Scientific Technology Observatory (ESTO),²²⁷
the view is held that precaution should be a response to ignorance. Ignorance is
but one element in a wider context. The starting-point is a distinction between
knowledge about likelihoods and knowledge about outcomes. The former is further
distinguished between firm, shaky, and no basis for probabilities; the latter, in a
continuum of outcomes, set of discrete outcomes, and outcomes poorly defined.
Depending on the combination between the various elements of the two variables,
we can end up with a state of uncertainty, ambiguity, or ignorance. Uncertainty is
the state where there is a continuum of outcomes, but we have no basis for prob-
abilities; ambiguity is the state where outcomes are poorly defined but we have
a firm or at the very least, a shaky basis of probabilities; finally, ignorance is the
state where we have no basis for probabilities and outcomes are poorly defined.
In epistemic terms, scenario analysis is recommended in cases of uncertainty,
sensitivity analysis in cases of ambiguity, and precaution in cases of ignorance.
Precaution emerges as the scientific approach to address incommensurability, that
is, the situation where one compares apples to pears. Should precaution in WTO
parlance be confined to cases of ignorance and incommensurability only? Or
should we understand precaution as an antidote for uncertainty, and ambiguity as
well? It seems reasonable to assume that the latter should be the case. Recall that

²²⁷ See Stirling (2001). It is made clear that the views expressed in this document are personal.
The EC document, nevertheless, (G/SPS/GEN/168), explicitly refers to this (upcoming at the time)
study as the source of its understanding of precaution.
314 Domestic Instruments

sensitivity and scenario analyses are scientific, not regulatory responses. The view
to understand precaution as an antidote for uncertainty and ambiguity as well, is
very much in line with the idea that there is no firewall between science and no
science; that science is an ever-evolving construct; and that regulatory responses
might be warranted irrespective of the degree of certainty about scientific evi-
dence. Ultimately, regulation is the privilege of the statesman. One should be
able to distinguish between scientifically informed and scientifically dependent
regulation: in the very famous and oft quoted words of Churchill, ‘science should
be on tap, not on top.’
Where does all this lead us? There is probably no horizontal response to the
question whether Art 5.7 of the SPS should cover non-discrimination, neces-
sity, and coherence. If its coverage extends from emergency situations, to cases
of ignorance, uncertainty, and ambiguity, then one should first acknowledge that
its coverage is so wide that it cannot accommodate horizontal solutions in this
respect. The response will depend on the availability of information. Assuming,
for example, that there is sufficient information, one cannot a priori exclude that
respect of non-discrimination, consistency, and necessity might be warranted.
Non-discrimination, necessity, and consistency are hardly compatible with emer-
gency situations. They might be compatible with uncertainty, and ambiguity, and
less so with cases of ignorance.
If the key criterion is the manner in which WTO case-law has been evolving,
then it is probably safe to conclude that a negative response is likelier: recall that
case-law has not clarified what is the minimal threshold that the precautionary
principle must respect. Does public anxiety, for example, suffice? Or, should the
intervening state be able to show at least some observations that science cannot
(at the moment of regulatory intervention) explain? Is, for example, fear that
GMOs might impair environment in and of itself a basis for precaution? Or, con-
versely, should a state invoking precaution be obliged to demonstrate some dam-
age (probably unrelated to a GMO) which, however, time-wise at least coincides
with the use of GMOs in the area at hand? In the absence of such information, it
is very difficult to evaluate whether the statutory requirements included in Arts
5.5 and 5.6 of the SPS must be respected, when recourse to precaution is being
made. If the AB’s silence on this issue is meant to convey the message that pre-
caution can be invoked any time scientific proof is deficient, then, even assuming
relevance of Arts. 5.5 and 5.6 of the SPS, it will be hard, at the very least as a mat-
ter of evidence, to show non-compliance with these two provisions.
In a nutshell, the response to the question regarding the substantive content of
precautionary principle largely depends on whether future case-law will be will-
ing to maintain or destruct the firewall it has erected between science and precau-
tion. What is being advocated here is the latter option.
 State Trading Enterprises (STEs) 315

6 State Trading Enterprises (STEs)

6.1 STEs: definitional issues

A natural consequence of the negative integration-character of the GATT is that
the status of property in individual WTO members is not prejudged; WTO
members remain free to decide on the extent of public sector in their jurisdic-
tion. As a result, it could very well be the case that they allow for state-trading
enterprises (STEs) to operate and conduct business.²²⁸ If this is the case, WTO
members will have to ensure that STEs do not frustrate the cornerstone of the
GATT, the principle of non-discrimination.
STEs are domestic institutions (instruments) often entrusted with the import-
ation, and, sometimes, the distribution of goods.²²⁹ GATT, Art XVII does not
define STEs in any meaningful manner. It uses side by side the terms ‘STEs’, and
‘enterprises’ which have been granted de jure or de facto exclusive or special privi-
leges. Some legislative documents provide illustrations:
(1) the Interpretative Note ad Art XVII of the GATT makes it clear that market-
ing boards are covered anyway (§ 1);
(2) the Understanding on the Interpretation of Art XVII of the GATT adopted
during the Uruguay round, reflects the following definition of the term,
which makes it clear that marketing boards are but one of the entities that
could come under the purview of Art XVII of the GATT:
governmental and non-governmental enterprises, including marketing boards, which
have been granted exclusive or special rights or privileges, including statutory or constitu-
tional powers, in the exercise of which they influence through their purchases or sales the
level or direction of imports or exports.
Practice, as well, has offered illustrations. To this effect, the GATT panel report
on Korea—Beef (US) held that a producer controlled import monopoly for beef
should be subjected to the discipline of Art XVII of the GATT. By the same
token, the Canadian Wheat Board, entrusted with the exclusive right to purchase
and sell western Canadian wheat for export and human consumption, was con-
sidered to be an STE (Canada—Wheat Exports and Grain Imports).

²²⁸ In Miller’s (2000) account, it is the British delegation that, during the London Conference,
insisted on the inclusion of STEs in the GATT.
²²⁹ The rationale for STEs is not prejudged by Art XVII of the GATT. A marketing board, for
example, could operate as some sort of quality-control mechanism. It could also, however, be the
case, that through a marketing board, WTO members pursue beggar thy neighbour policies, costly
to their trading partners: it could be, for example, that a marketing board, entrusted with the
distribution of imported goods, intentionally inflicts harm on imported products by distributing
them inadequately. Unless disciplines are imposed on STEs, WTO members might see the value of
their negotiated concessions be reduced substantially. This is the role that Art XVII of the GATT
is called to play.
316 Domestic Instruments

GATT, Art XVII.2 reflects a caveat which is critical for the coverage of Art
XVII of the GATT: the legal obligations assumed, by virtue of Art XVII of the
GATT, do not concern government procurement. Hence, STEs should be under-
stood to be entities which buy and resell, and not as entities which buy for gov-
ernmental use.²³⁰

6.2 The legal obligation in a nutshell

STEs must respect the non-discrimination principle. This point needs some fur-
ther elaboration. To the extent that an STE is acting as an importer only, respect-
ing the non-discrimination principle is tantamount to respecting MFN. To the
extent, however, that an STE acts as a distributor as well, it must not only observe
MFN, but also NT: it must make sure that it behaves in accordance with commer-
cial considerations, and that it affords companies (with respect to their purchases
or sales) adequate opportunities to compete.
GATT, Art XVII.1(a) enshrines the non-discrimination obligation. The AB,
in its report on Canada—Wheat Exports and Grain Imports held that this pro-
vision is an anti-circumvention device: WTO members cannot, by establishing
STEs, circumvent their non-discrimination obligation (§ 85):
Subparagraph (a) seeks to ensure that a Member cannot, through the creation or main-
tenance of a State enterprise or the grant of exclusive or special privileges to any enter-
prise, engage in or facilitate conduct that would be condemned as discriminatory under
the GATT 1994 if such conduct were undertaken directly by the Member itself. In other
words, subparagraph (a) is an “anti-circumvention” provision.
Earlier case-law suggests that the non-discrimination obligation did not
extend to cover NT as well (Belgian Tax Allowances, § 4). There is also case-law
(Canada—FIRA) where the panel refused to pronounce on the issue whether
non-discrimination in Art XVII of the GATT extended to cover NT as well
(§ 5.16).²³¹ This issue was re-discussed in the context of the dispute between a series
of beef exporting countries (Australia, New Zealand, and the United States) and
Korea. The facts of the case are worth recounting: Korea had a legitimate quota on
beef in place (for BoP reasons), and also established a producer-controlled import
monopoly, entrusted with the importation and distribution of beef. The import
monopoly was the sole importer and distributor of beef, in times when the price
of imported beef was substantially lower than that of domestic beef. The Korean
measure was under attack for various reasons, one of them being for its incon-
sistency with Art XVII of the GATT. The panel in its report on Korea—Various
Measures on Beef held that the legal obligation not to discriminate enshrined in

²³⁰ See on this issue, Blank and Marceau (1997).

²³¹ A similar attitude is observed in the panel report on Canada—Provincial Liquor Boards
(EEC) where, once again, the panel refused to pronounce on the issue, in light of its findings that
the measure at hand contravened Art XI of the GATT (§ 4.27).
 State Trading Enterprises (STEs) 317

Art XVII.1 of the GATT covered at least both MFN and NT, without specify-
ing, however, what else should come under its purview (§ 7.53).²³²
WTO members must ensure that their STEs make their purchases in accord-
ance with commercial considerations, and that they afford all suppliers involved
adequate opportunities to compete (Art XVII.1(b) of the GATT). Case-law has
by now established that the two obligations are legally distinct.²³³ The question
arose in dispute settlement practice as to whether these distinct obligations are
mere illustrations of the non-discrimination obligation, or, conversely, whether
they should be understood as obligations additional to the obligation not to dis-
criminate. The AB, in its report on Canada—Wheat Exports and Grain Imports,
held that these two obligations are a mere illustration of the obligation not to dis-
criminate (§§ 89–106).
The term ‘commercial considerations’ has been interpreted in the panel report
on Canada—Wheat Exports and Grain Imports. There, the panel held that, were
an STE to make purchases or sales on considerations such as the nationality of
potential buyers or sellers, the policies pursued, or the national (economic or pol-
itical) interest of the member maintaining the STE, it would not be acting solely
in accordance with commercial considerations (§ 6.88). We can infer that such
considerations are not per se inconsistent with commercial considerations: indeed
if, for example, such considerations are ancillary to welfare-maximizing concerns,
they could have been considered in accordance with Art XVII of the GATT. The
same panel went on to make some affirmative statements as to what constitutes
behaviour in accordance with commercial considerations: the United States claimed
that to act in accordance with commercial considerations required the Canadian
STE at hand (the Canadian Wheat Board) not to sell with the objective to maxi-
mize sales, but with the objective to maximize profits. Although the first type of
behaviour might be viewed as rational, in the US point of view, it was not per se
a commercial consideration. Continuing this line of thinking, the United States
claimed that Canada should not be allowing its STE to use its privileges in order
to maximize sales. The privileges of the Canadian Wheat Board were:
(1) the exclusive right to purchase and sell western Canadian wheat for export
and domestic human consumption;
(2) the right to set, subject to government approval, the initial price pay-
able for western Canadian wheat destined for export or domestic human
consumption;
(3) the government guarantee of the initial payment to producers of western
Canadian wheat;
(4) the government guarantee of its borrowing; and
(5) government guarantees of certain of its credit sales to foreign buyers.

²³² This panel finding was not appealed.

²³³ See the panel report on Canada—Wheat Exports and Grain Imports (§ 6.60).
318 Domestic Instruments

The US view was that the Canadian Wheat Board was acting inconsistently
with Art XVII of the GATT, because when selling it should not be discrimin-
ating across markets, and it should not be selling below market rates anyway.
Additionally, the United States found that the Canadian STE’s behaviour was
GATT inconsistent, since it was seeking to maximize revenue and not profit, and
was, hence, not acting like a private grain trader (a profit maximizing firm). The
panel rejected all US claims and arguments in this respect. It took the view that
the STE at hand could legitimately use its privilege to the disadvantage of com-
mercial actors (§ 6.106), that selling below market prices was perfectly legitimate
as well (§ 6.129), and that not selling for its own profit should not be equated to
acting without respecting commercial considerations (§ 6.133). The United States
appealed the first finding. The AB, first, in § 140 of the report, explained in clear
terms the panel’s overall understanding of the term ‘commercial considerations’:
The panel began its analysis by considering the meaning of the term “commercial consid-
erations” in subparagraph (b) and found that this term should be understood as mean-
ing “considerations pertaining to commerce and trade, or considerations which involve
regarding purchases or sales ‘as mere matters of business’.” The panel also determined
that the requirement that STEs act solely in accordance with such considerations “must
imply that they should seek to purchase or sell on terms which are economically advan-
tageous for themselves and/or their owners, members, beneficiaries, etc.” Thus, the panel
interpreted the term “commercial considerations” as encompassing a range of different
considerations that are defined in any given case by the type of “business” involved (pur-
chases or sales), and by the economic considerations that motivate actors engaged in busi-
ness in the relevant market(s).
Having agreed with the panel’s understanding of the term, the AB went on to
examine the merits of the US claim. In its view, the request by the United States
was unreasonable. STEs have to act in accordance with commercial consider-
ations; they do not, however, in the AB’s view, have to undo themselves, and
behave as if they did not have any privileges. In the AB’s view, Art XVII of the
GATT does not impose competition law-type of disciplines.²³⁴ Consequently, as
long as they do not discriminate, STEs can be deemed to have acted in accord-
ance with commercial considerations and thus, can make use of their privileges
(§§ 146–51).²³⁵
In the same dispute, the United States also complained about the practices of
the Canadian STE vis-à-vis other enterprises, arguing that it did not afford other
companies adequate opportunities to compete and was, thus, in violation of the
second clause embedded in Art XVII.1(b) of the GATT. The United States was
arguing that the obligation to afford adequate opportunities to other enterprises

²³⁴ On the EC experience on this issue, see Mavroidis and Messerlin (1998).
²³⁵ Th is is a problematic statement. Acting with commercial considerations might entail dis-
criminatory behaviour. It is difficult to understand what the AB had in mind here. Earlier case-law
which understood this term as imposing an obligation to, for example, not privilege suppliers based
on nationality only was probably wiser.
 State Trading Enterprises (STEs) 319

to compete, extended to cover enterprises in competition with an STE, that is,

enterprises which could substitute the STE at hand as a purchaser or a seller. The
panel had already rejected this claim, arguing that in cases where the STE was a
seller, the obligation to afford adequate opportunities to compete should regard
buyers only and not other sellers as well (that is, it should not extent to cover com-
panies which could substitute the STE). The AB confirmed the panel’s reading of
the second clause of Art XVII.1(b) of the GATT (§ 157).
All STEs must be notified to the WTO. The Understanding on the
Interpretation of Art XVII of the GATT allows for the possibility to cross-notify
STEs, which have not been duly notified by the WTO member the nationality
of which they carry. The CTG has adopted a series of decisions relating to the
periodic notifications requirements with respect to STEs, that all WTO mem-
bers must observe.²³⁶

²³⁶ See WTO Docs G/C/M/1, G/STR/N/1 and all subsequent numbers in this series.
 4
State Contingencies

1 Introduction 321
2 National Security (Art XXI of the GATT) 322
2.1 The legal discipline 322
2.2 Practice in the GATT years 322
2.3 Practice in the WTO era 330
3 BoP 331
4 Exchange Restrictions 335
5 Antidumping (Art VI of the GATT) 338
5.1 An unfair (?) price discrimination 338
5.2 Finding dumping 343
5.3 Establishing injury 349
5.4 Sunset reviews 352
5.5 Price differentiation in antitrust statutes 358
5.6 What does the ‘anti’ in AD refer to? 361
5.6.1 Discriminatory pricing 361
5.6.2 Predation 362
5.6.3 Strategic dumping 363
5.6.4 AD as a safeguard mechanism 364
6 Countervailing Duties (CVD) 365
7 Safeguards 365
7.1 The regulatory framework 365
7.2 The typology of safeguards 366
7.3 Safeguards imposed by PTAs 367
7.4 Conditions for lawful application of safeguards 368
7.4.1 Initiation of investigation 368
7.4.2 Unforeseen developments 370
7.4.3 Increased imports 371
7.4.4 Injury 371
7.4.5 Are safeguards truly non-discriminatory? 375
7.4.6 The issue of compensation 376
7.4.7 The duration of safeguards 378
7.5 Safeguards and voluntary export restraints (VERs) 380
7.6 Safeguards: a safe harbour too far? 381
8 The Notion of Market Access Revisited 385
 Introduction 321

Overview

In Chapter 1, we described the GATT as a mechanism whereby members com-

mit default protection, and state contingent protection. Economic theory sug-
gests that there are good arguments to support contract flexibility for a contract
like the GATT: countries will have more of an incentive to commit when pre-
sented with inexpensive exit options, since the pace of adjustment will not be too
onerous for them. There is, thus, a very solid argument in favour of the introduc-
tion of a safeguard in the GATT contract. As we will see, however, a number of
other contingent protection instruments have been included as well. And while
it is difficult to argue against the insertion of some of them (national security),
good arguments can be made against the introduction of others. For example, the
BOP restriction was probably more to the point when we were living in a world
of fixed exchange rates. In a world where this is not the case anymore, why not
simply have recourse to devaluation? Recourse to AD duties is hard to justify
(from an economic perspective). By introducing the loose concept of ‘unfair
trade’, GATT/WTO members agreed to punish behaviour, which they find quite
normal and acceptable business behaviour when practised at home: what is unfair
internationally is quite fair at the national level, in other words. Following a brief
introduction to all contingent protection instruments (section 1), we move to
discuss national security (section 2), BOP (section 3), exchange restrictions (sec-
tion 4), AD (section 5), countervailing duties (section 6), and, finally, safeguards
(section 7).

1 Introduction

In this chapter, we discuss the state contingencies, the occurrence of which allows
WTO members to deviate from their default promise, and either use a tariff
higher than the bound rate, or have (legitimate) recourse to a QR, or a TRQ, that
is, a combination of the two instruments, whereby a WTO member imposes a
low duty for a limited quantity and a higher duty beyond that.
We will be discussing:
(1) the national security exception (Art XXI of the GATT);
(2) restrictions imposed for BOP reasons (Arts XXII and XVIII of the GATT);
(3) exchange restrictions (Art XV of the GATT);
(4) AD (Art VI of the GATT, and the WTO AD Agreement);
(5) countervailing duties (Art XVI of the GATT, and the WTO SCM
Agreement);
(6) safeguards (Art XIX GATT, and the WTO SG Agreement).
322 State Contingencies

2 National Security (Art XXI of the GATT)

2.1 The legal discipline

GATT, Art XXI reads:
Nothing in this Agreement shall be construed
(a) to require any contracting party to furnish any information the disclosure of which it
considers contrary to its essential security interests; or
(b) to prevent any contracting party from taking any action which it considers necessary
for the protection of its essential security interests
(i) relating to fissionable materials or the materials from which they are derived;
(ii) relating to the traffic in arms, ammunition and implements of war and to such
traffic in other goods and materials as is carried on directly or indirectly for the
purpose of supplying a military establishment;
(iii) taken in time of war or other emergency in international relations; or
(c) to prevent any contracting party from taking any action in pursuance of its obliga-
tions under the United Nations Charter for the maintenance of international peace
and security.
The title of this provision (Security Exception) leaves no room for doubt that
it is intended to function as an exception to the obligations assumed under the
GATT. Hence, it is for the WTO member invoking it to carry the associated bur-
den of proof. The contingency justifying deviation from the obligations assumed
is the threat to national security.

2.2 Practice in the GATT years

GATT, Art XXI has been invoked only once during the GATT years.¹ Nicaragua
was facing a two-way embargo from the United States² which practically ended
any trade relations between the two countries. Nicaragua complained that the
measure at hand was in violation of Art XI of the GATT. Even before the panel
had been established, the United States had argued that the measure was neces-
sary to protect the national security interests of the United States. The United
States also claimed that an invocation of Art XXI of the GATT is not justiciable.
Before the panel, the United States repeated the same arguments.
The panel report on US—Nicaraguan Trade which dealt with the complaint
by Nicaragua was never adopted, and, hence, remains of limited legal value. It
reveals, however, the opinio juris of some influential WTO members, with respect

¹ On this issue, see the excellent analysis offered in Hahn (1991) who discusses the case-law and
pays particular attention to the standard of review employed in national security cases. The author
further advances his own framework for analysis of such issues.
² The United States had stopped exports to and imports from Nicaragua as a reaction to the rise
of the Sandinista regime.
 National Security (Art XXI of the GATT) 323

to Art XXI of the GATT. The motivation for the trade embargo is laid out in
§ 3.1 of the report, and § 3.3 of the report reflects what the US embargo meant
for trade between the two countries:
3.1 On 1 May 1985 the President of the United States issued an Executive Order
which reads:
. . . I, RONALD REAGAN, President of the United States of America, find that
the policies and actions of the Government of Nicaragua constitute an unusual and
extraordinary threat to the national security and foreign policy of the United States
and hereby declare a national emergency to deal with that threat.
I hereby prohibit all imports into the United States of goods and services of
Nicaraguan origin; all exports from the United States of goods to or destined for
Nicaragua, except those destined for the organized democratic resistance, and
transactions relating thereto.
I hereby prohibit Nicaraguan air carriers from engaging in air transportation to
or from points in the United States, and transactions relating thereto.
In addition, I hereby prohibit vessels of Nicaraguan registry from entering into
United States ports, and transactions relating thereto.
The Secretary of the Treasury is delegated and authorized to employ all powers
granted to me by the International Emergency Economic Powers Act to carry out
the purposes of this Order.
The prohibition set forth in this Order shall be effective as of 12:01 a.m., Eastern
Daylight Time, May 7, 1985 and shall be transmitted to the Congress and published
in the Federal Register.
3.3 According to calculations made by the GATT Secretariat almost all imports
(more than 99 per cent) from Nicaragua into the United States are items for which the
duties are bound under the General Agreement.
It is noteworthy that the parties agreed on special terms of reference for this panel,
which are reproduced in §§ 1.4–1.5:
1.4 At the meeting of the Council on 12 March 1986, the Chairman announced that
the following terms of reference of the Panel had been agreed:
To examine, in the light of the relevant GATT provisions, of the understanding
reached at the Council on 10 October 1985 that the Panel cannot examine or judge
the validity of or motivation for the invocation of Article XXI:(b)(iii) by the United
States, of the relevant provisions of the Understanding Regarding Notification,
Consultation, Dispute Settlement and Surveillance (BISD 26S/211-218), and
of the agreed Dispute Settlement Procedures contained in the 1982 Ministerial
Declaration (BISD 29S/13-16), the measures taken by the United States on 7 May
1985 and their trade effects in order to establish to what extent benefits accruing to
Nicaragua under the General Agreement have been nullified or impaired, and to
make such findings as will assist the CONTRACTING PARTIES in further action
in this matter (C/M/196, page 7).
1.5 Following this announcement, the representative of the United States said the
terms of reference had been drafted specifically for this case and would govern the Panel
324 State Contingencies
in this particular dispute. However, this should not imply that panels in other cases would
not have to determine whether nullification or impairment existed. Only in this case did
the United States not dispute the effects of a two-way trade embargo. Furthermore, the
above terms of reference should not be interpreted to mean that any further action by
the CONTRACTING PARTIES in this matter was necessary or appropriate. The rep-
resentative of Nicaragua replied that, in his view, this Panel was not an exception; its
functions would be those described in the 1979 Under-standing (BISD 26S/211-218).
Consequently, the CONTRACTING PARTIES would have to take appropriate action
on the Panel’s report (C/M/196, page 8).
This panel was, consequently, lame from birth. The main arguments of the par-
ties focused on the standard of review to be applied by the panel. The parties
presented radically opposing views—Nicaragua argued for a substantive review
of the case, whereas the United States argued for the opposite:
4.5 Nicaragua stated that the United States could not properly rely on Article XXI:
(b)(iii) in this case. This provision could be invoked only if two conditions were met:
first, the measure adopted had to be necessary for the protection of essential security
interest and, second, the measure had to be taken in time of war or other emergency in
international relations. Neither of these conditions were fulfilled in this present case.
Obviously, a small developing country such as Nicaragua could not constitute a threat
to the security of the United States. The embargo was therefore not necessary to protect
any essential security interest of that country. Nor was there any “emergency” in the sense
of Article XXI. Nicaragua and the United States were not at war and maintained full
diplomatic relations. If there was tension between the two countries, it was due entirely
to actions by the United States in violation of international law. A country could not be
allowed to base itself on the existence of an “emergency” which it had itself created. In
that respect, Article XXI was analogous to the right of self-defence in international law.
This provision could be invoked only by a party subjected to direct aggression or armed
attack and not by the aggressor or by parties indirectly at risk. Nicaragua added that it
must be borne in mind that GATT did not exist in a vacuum but was an integral part
of the wider structure of international law, and that the General Agreement must not
be interpreted in a way inconsistent with international law. The International Court of
Justice had found that the embargo was one element of a whole series of economic and
military actions taken against Nicaragua in violation of international law and that it was
not necessary for the protection of any essential security interest of the United States, and
it had declared that the United States must make reparation for the damage caused. The
Security Council (Resolution 562) and the General Assembly (Resolution 40/188) of the
United Nations had also condemned the embargo for infringing the principles of free
trade and had explicitly demanded its rescinding. Consequently, Nicaragua held that the
United States could not base itself on Article XXI in the particular case, and that the trade
measures under consideration constituted coercive measures applied for political reasons
in contravention of paragraph 7(iii) of the Ministerial Declaration of November 1982,
which obliged contracting parties to “abstain from taking restrictive trade measures, for
reasons of a non-economic character, not consistent with the General Agreement.”
4.6 The United States said that Article XXI applied to any action which the contract-
ing party taking it considered necessary for the protection of its essential security interest.
 National Security (Art XXI of the GATT) 325
This provision, by its clear terms, left the validity of the security justification to the exclu-
sive judgement of the contracting party taking the action. The United States could there-
fore not be found to act in violation of Article XXI. In any case, the Panel’s terms of
reference made it clear that it could examine neither the validity of, nor the motivation
for, the United States’ invocation of Article XXI:(b)(iii). The United States’ compliance
with its obligations under the General Agreement was therefore not an issue before the
Panel. The United States added that it disagreed with Nicaragua’s assessment of the secur-
ity situation but it did not wish to be drawn into a debate on a matter that fell outside the
competence of the GATT in general and the Panel in particular.
4.7 Nicaragua, while recognizing that it was not within the competence of the Panel
to examine or judge the validity of or motivation for the invocation of Article XXI:
(b)(iii), nevertheless felt that the Panel had sufficient legal material and other information
before it to arrive at a conclusion on the consistency of the embargo with the provisions of
the General Agreement.
The panel’s reaction to these arguments is reproduced in what immediately
follows. The panel was precluded, by its terms of reference, from addressing the
justification for the embargo. Moreover, the panel realized that a recommenda-
tion to the effect that the United States withdraw the embargo would not mean
much in practice, in light of the resolve of the United States (repeated a number
of times before and during the panel proceedings) not to comply with such a
recommendation. Finally, even if the panel had opted for a recommendation
which would be favourable to Nicaragua and the latter would, as a result, sub-
sequently request authorization to impose counter-measures, such a recommen-
dation would be totally counter-productive: the two-way embargo by the United
States essentially insulated one country from the other; presence or absence of
counter-measures by Nicaragua, would have no effect under the circumstances.
The full panel report is reproduced hereinafter:³
5.1 The Panel first considered the question of whether any benefits accruing to
Nicaragua under the General Agreement had been nullified or impaired as the result of
a failure of the United States to carry out its obligations under the General Agreement
(Article XXIII:1(a)). The Panel noted that, while both parties to the dispute agreed that
the United States, by imposing the embargo, had acted contrary to certain trade-
facilitating provisions of the General Agreement, they disagreed on the question of
whether the non-observance of these provisions was justified by Article XXI(b)(iii), the
relevant part of which reads:
Nothing in this Agreement shall be construed . . . to prevent any contracting party
from taking . . . in time of war or other emergency in international relations . . . any
action which it considers necessary for the protection of its essential security interests.
5.2 The Panel further noted that, in the view of Nicaragua, this provision should be
interpreted in the light of the basic principles of international law and in harmony with

³ The reader should keep in mind that this is the only GATT/WTO panel that has discussed Art
XXI of the GATT. It might provide guidance to subsequent cases dealing with this issue, assuming
there are any.
326 State Contingencies
the decisions of the United Nations and of the International Court of Justice and should
therefore be regarded as merely providing contracting parties subjected to an aggression
with a right to self-defence. The Panel also noted that, in the view of the United States,
Article XXI applied to any action which the contracting party taking it considered neces-
sary for the protection of its essential security interests and that the Panel, both by the
terms of Article XXI and by its mandate, was precluded from examining the validity of
the United States’ invocation of Article XXI.
5.3 The Panel did not consider the question of whether the terms of Article XXI pre-
cluded it from examining the validity of the United States’ invocation of that Article as
this examination was precluded by its mandate. It recalled that its terms of reference put
strict limits on its activities because they stipulated that the Panel could not examine or
judge the validity of or the motivation for the invocation of Article XXI:(b)(iii) by the
United States (cf. paragraph 1.4 above). The Panel concluded that, as it was not author-
ized to examine the justification for the United States’ invocation of a general exception
to the obligations under the General Agreement, it could find the United States neither to
be complying with its obligations under the General Agreement nor to be failing to carry
out its obligations under that Agreement.
5.4 Being precluded from examining the embargo in light of paragraph (a) of Article
XXIII:1, the Panel proceeded to examine it in the light of paragraph (b) of Article XXIII:1.
Consequently, it considered the question of whether benefits accruing to Nicaragua
under the General Agreement had been nullified or impaired by the embargo whether or
not it conflicted with the provisions of the General Agreement.
5.5 The Panel noted that the previous cases under paragraph (b) of Article XXIII:1
(BISD Vol. II/192-193 and BISD 1S/58-59) involved measures that had been found to be
consistent with the General Agreement while in the present case it could not be determined
whether or not the measure was consistent with the General Agreement. The Panel never-
theless considered the principles established in the previous cases to be applicable in the
present case because a contracting party has to be treated as if it is observing the General
Agreement until it is found to be acting inconsistently with it.
5.6 The Panel noted that the embargo had virtually eliminated all opportunities for
trade between the two contracting parties and that it had consequently seriously upset
the competitive relationship between the embargoed products and other directly com-
petitive products. The Panel considered the question of whether the nullification or
impairment of the trade opportunities of Nicaragua through the embargo constituted
a nullification or impairment of benefits accruing to Nicaragua within the meaning of
Article XXIII:1(b). The Panel noted that this question raised basic interpretative issues
relating to the concept of non–violation nullification and impairment which had nei-
ther been addressed by the drafters of the GATT nor decided by the CONTRACTING
PARTIES. Against this background the Panel felt that it would only be appropriate for it
to propose a ruling on these issues if such a ruling would enable the CONTRACTING
PARTIES to draw practical conclusions from it in the case at hand.
5.7 The Panel then noted that Article XXIII:2 would give the CONTRACTING
PARTIES essentially two options in the present case if the embargo were found to have
nullified or impaired benefits accruing to Nicaragua under the General Agreement
independent of whether or not it was justified under Article XXI. They could either (a)
recommend that the United States withdraw the embargo (or, which would amount in
 National Security (Art XXI of the GATT) 327
the present case to the same, that the United States offer compensation) or (b) author-
ize Nicaragua to suspend the application of obligations under the General Agreement
towards the United States.
5.8 As to the first of the above options the Panel noted the following: It is clear from
the drafting history that in case of recommendations on measures not found to be incon-
sistent with the General Agreement, the contracting parties “are under no specific and
contractual obligations to accept those recommendations” (EPCT/A/PV/5, p.16). The
report of the Sixth Committee during the Havana Conference notes with respect to
the power of the Executive Board to make recommendations to member States in any
matter arising under Article 93:1(b) or (c) of the Havana Charter (which corresponds to
Article XXIII:1(b) and (c) of the General Agreement): “It was agreed that sub-paragraph
2(e) of Article 94 does not empower the Executive Board or the Conference to require a
Member to suspend or withdraw a measure not in conflict with the Charter”. The 1950
Working Party on the Australian Subsidy on Ammonium Sulphate took the same view
as to the powers of the CONTRACTING PARTIES (BISD Vol. II/195). In their 1982
Ministerial Declaration, the CONTRACTING PARTIES stated that the dispute settle-
ment process could not “add to or diminish the rights and obligations provided in the
General Agreement” (BISD 26S/16).
5.9 In the light of the above drafting history and decisions of the CONTRACTING
PARTIES the Panel found that the United States, as long as the embargo was not found
to be inconsistent with the General Agreement, was under no obligation to follow a rec-
ommendation by the CONTRACTING PARTIES to remove the embargo.
5.10 The Panel noted that in the past cases under paragraph (b) of Article XXIII:1, the
CONTRACTING PARTIES had recommended that the contracting party complained
against consider ways and means to restore the competitive relationship that existed when
the tariff concession was made (BISD Vol. II/195 and BISD 1S/31). However, the Panel
also noted that these recommendations had been made only because they were consid-
ered to offer the best prospect of a mutually agreed settlement of the dispute. It noted in
particular the following statement in the report of the Working Party on the Australian
Subsidy on Ammonium Sulphate:
The sole reason why the [withdrawal of a measure not found to be inconsistent with
the General Agreement] is recommended is that, in this particular case, it happens
that such action appears to afford the best prospect of an adjustment of the matter
satisfactory to both parties (BISD Vol. II/195).
The Panel noted that the United States had declared from the outset that it would not
remove the embargo without a solution to the underlying political problem (para-
graph 4.9 above). It also noted that Nicaragua had recognized that “it seemed unfor-
tunately unlikely that the United States would accept a recommendation to lift the
embargo” (paragraph 4.10 above). The Panel therefore considered that a decision of the
CONTRACTING PARTIES under Article XXIII:2 recommending the withdrawal of
the embargo would not seem to offer the best prospect of an adjustment of the matter
satisfactory to both parties and that, in these circumstances, it would not appear to be
appropriate for the CONTRACTING PARTIES to take such a decision unless they had
found the embargo to be inconsistent with the General Agreement.
5.11 The Panel then turned to the second option available to the CONTRACTING
PARTIES under Article XXIII:2 in the present case, namely a decision to authorize
328 State Contingencies
Nicaragua to suspend the application of obligations to the United States. The Panel
noted that, under the embargo imposed by the United States, not only imports from
Nicaragua into the United States were prohibited but also exports from the United States
to Nicaragua. In these circumstances, a suspension of obligations by Nicaragua towards
the United States could not alter the balance of advantages accruing to the two con-
tracting parties under the General Agreement in Nicaragua’s favour. The Panel noted
that the United States had stated that an authorization permitting Nicaragua to suspend
obligations towards the United States “would be of no consequence in the present case
because the embargo had already cut off all trade relations between the United States and
Nicaragua” (paragraph 4.9 above) and that Nicaragua had agreed that “a recommenda-
tion by the Panel that Nicaragua be authorized to withdraw its concessions in respect of
the United States would indeed be a meaningless step because of the two-way embargo”
(paragraph 4.10 above). The Panel therefore had to conclude that, even if it were found
that the embargo nullified or impaired benefits accruing to Nicaragua independent of
whether or not it was justified under Article XXI, the CONTRACTING PARTIES
could, in the circumstances of the present case, take no decision under Article XXIII:2
that would re-establish the balance of advantages which had accrued to Nicaragua under
the General Agreement prior to the embargo. In the light of the foregoing considerations
the Panel decided not to propose a ruling in this case on the basic question of whether
actions under Article XXI could nullify or impair GATT benefits of the adversely affected
contracting party.
5.12 The Panel proceeded to consider the request by Nicaragua that the Panel recom-
mend that the CONTRACTING PARTIES grant, in accordance with Article XXV:5
and footnote 2 to paragraph 2 of the Enabling Clause (BISD 26S/203), a general waiver
which would permit the contracting parties which so desire to compensate the effects of
the embargo by giving, notwithstanding their obligations under Article I, differential
and more favourable treatment to products of Nicaraguan origin.
5.13 The Panel examined whether it was appropriate for a panel established under
Article XXIII to make recommendations on requests for waivers under Article XXV. It
noted the following GATT practices and procedures on this question: Only once in the
history of the GATT, in 1971, has a panel established under Article XXIII recommended
a waiver pursuant to Article XXV. This waiver released the party complained against
from an obligation which it had failed to observe (BISD 18S/33, 183-188). All other
panels have proposed recommendations and rulings of the CONTRACTING PARTIES
under Article XXIII:2 and not decisions under Article XXV. This practice is reflected
in the 1979 Understanding on dispute settlement which states that “the function of
panels is to assist the CONTRACTING PARTIES in discharging their responsibil-
ities under Article XXIII:2” (BISD 26S/213). The procedures for waivers adopted by the
CONTRACTING PARTIES in 1956 (BISD 5S/25) provide that requests for waivers are
in principle to be submitted with a thirty-day notice, must be preceded by consultations
between the applicant contracting party and other contracting parties having made rep-
resentations and should be granted only if the CONTRACTING PARTIES are satisfied
that the legitimate interests of all contracting parties are adequately safeguarded. Th is
procedure ensures that the CONTRACTING PARTIES do not grant waivers without
first considering the views of the contracting parties that would be directly affected by
the waiver.
 National Security (Art XXI of the GATT) 329
5.14 The Panel recognized that its mandate was to “. . . make such findings as will assist
the CONTRACTING PARTIES in further action in this matter” (paragraph 1.4 above)
while panels were normally asked “to assist the CONTRACTING PARTIES in mak-
ing recommendations or rulings, as provided for in Article XXIII:2” (cf. for instance
BISD 31S/68, 76 and 94 and BISD 32S/56) and that a recommendation on the waiver
proposed by Nicaragua would therefore not be excluded by the Panel’s terms of reference.
However, the Panel concluded that it would be acting contrary to the GATT practices
and procedures described in the preceding paragraph if it were to recommend a change
in the obligations of third contracting parties that had no part in the Panel’s proceed-
ings and whose views it could therefore not consider. The Panel wishes to emphasize,
however, that Nicaragua has the right to submit a proposal for a waiver directly to the
CONTRACTING PARTIES and that the Panel’s decision not to make a recommen-
dation on the waiver is based on purely procedural grounds and should therefore in no
way be interpreted as prejudging a decision by the CONTRACTING PARTIES on such
a request. In this respect, the Panel also recalls that the consequences of the embargo on
Nicaragua’s trade and economy were severe and that, as noted in paragraph 5.6 above, the
embargo had seriously upset the competitive relationship between the embargoed prod-
ucts and other directly competitive products.
5.15 The Panel wishes to note that in the course of the Panel proceedings Nicaragua
had maintained that GATT could not operate in a vacuum and that the GATT provi-
sions must be interpreted within the context of the general principles of international law
taking into account inter alia the judgement by the International Court of Justice and
United Nations resolutions. While not refuting such argumentation, the Panel neverthe-
less considered it to be outside its mandate to take up these questions because the Panel’s
task was to examine the case before it “in the light of the relevant GATT provisions”,
although they might be inadequate and incomplete for the purpose.
5.16 The Panel, noting that it had been given not only the mandate to prepare a deci-
sion of the CONTRACTING PARTIES under Article XXIII:2 but the wider task of
assisting the CONTRACTING PARTIES in further action in this matter, examined
the effects of the embargo on Nicaragua’s economy and on the international trading sys-
tem. The Panel noted that the embargo had brought the trade between two contracting
parties to a standstill and that it had a severe impact on the economy of a less-developed
contracting party. The Panel further noted that embargoes imposed for security reasons
create uncertainty in trade relations and, as a consequence, reduce the willingness of
governments to engage in open trade policies and of enterprises to make trade-related
investments. The Panel therefore concluded that embargoes such as the one imposed by
the United States, independent of whether or not they were justified under Article XXI,
ran counter to basic aims of the GATT, namely to foster non-discriminatory and open
trade policies, to further the development of the less-developed contracting parties and to
reduce uncertainty in trade relations. The Panel recognized that the General Agreement
protected each contracting party’s essential security interests through Article XXI and
that the General Agreement’s purpose was therefore not to make contracting parties
forego their essential security interests for the sake of these aims. However, the Panel
considered that the GATT could not achieve its basic aims unless each contracting party,
whenever it made use of its rights under Article XXI, carefully weighed its security needs
against the need to maintain stable trade relations.
330 State Contingencies
5.17 The above considerations and the conclusions to which the Panel had to arrive,
given its limited terms of reference and taking into account the existing rules and proced-
ures of the GATT, raise in the view of the Panel the following more general questions:
If it were accepted that the interpretation of Article XXI was reserved entirely to the
contracting party invoking it, how could the CONTRACTING PARTIES ensure that
this general exception to all obligations under the General Agreement is not invoked exces-
sively or for purposes other than those set out in this provision? If the CONTRACTING
PARTIES give a panel the task of examining a case involving an Article XXI invocation
without authorizing it to examine the justification of that invocation, do they limit the
adversely affected contracting party’s right to have its complaint investigated in accord-
ance with Article XXIII:2? Are the powers of the CONTRACTING PARTIES under
Article XXIII:2 sufficient to provide redress to contracting parties subjected to a two-way
embargo?
5.18 The Panel noted that in 1982 the CONTRACTING PARTIES took a “Decision
Concerning Article XXI of the General Agreement” which refers to the possibility of
a formal interpretation of Article XXI and to a further consideration by the Council
of this matter (BISD 29S/23-24). The Panel recommends that the CONTRACTING
PARTIES, in any further consideration of this matter in accordance with that Decision,
take into account the questions raised by the Panel above. (original emphasis )
The quintessence of the Panel’s approach is captured in § 5.17: although not expli-
citly stating it, the panel seems to be opposed to the view that the invocation of
Art XXI of the GATT is not justiciable. However, in light of its mandate (special
terms of reference), the panel did not have the opportunity to explain itself as to
the appropriate standard of review in cases like this.

2.3 Practice in the WTO era

The notorious US Helms/Burton Act (Cuban Liberty and Democratic Solidarity
Act, or LIBERTAD) gave rise to the only dispute in the WTO era thus far.
Through this Act, the United States imposed trade sanctions on products of
Cuban origin, but also on products of other (than Cuban) origin, a percentage of
the value added of which was Cuban (primary embargo). Eventually, the United
States, through a secondary embargo, stopped trading with nations trading with
Cuba. The European Community complained that the US measure at hand was
in violation of a series of GATT provisions. The official description of the EC
complaint is reproduced here:
The European Community and its Member States wish to convey to you a request for con-
sultations with the United States of America pursuant to Article 4 of the Understanding
on Rules and Procedures Governing the Settlement of Disputes (DSU), Article XXIII:1
of the General Agreement on Tariffs and Trade 1994 (GATT 1994) and Article XXIII:1
of the General Agreement on Trade in Services (GATS) concerning the Cuban Liberty
and Democratic Solidarity (LIBERTAD) Act of 1996, other legislative provisions con-
solidated therein, and any implementing measures taken thereunder.
 BoP 331
The European Community and its Member States wish to express their profound
concern about the apparent lack of conformity of certain aspects of this Act, including
other legislative provisions consolidated therein and any implementing measures taken
thereunder, to the international obligations of the United States under GATT 1994 and
GATS. This concern relates in particular, but not necessarily exclusively, to the following
aspects:
The Cuban Democracy Act and its companion the Cuban Liberty and Democratic
Solidarity Act contain a number of provisions which have the intent and effect to restrain
the liberty of the EC to export to Cuba or to trade in Cuban origin goods, as well as to
restrict the freedom of EC registered vessels and their cargo to transit through US ports.
In addition, there are provisions which require the provisions of certificates in respect
of trade in Cuban sugar. If such certificates are not provided, access to the US sugar quota
is denied.
Finally, there are measures which may lead to the refusal of visas and the exclusion of
non-US nationals from US territory in a way which may contravene US commitments
under GATS.
The European Community and its Member States are of the view that these and com-
parable measures taken under the two laws mentioned above may not be in conformity
with at least the following provisions: Articles I, III, V, XI and XIII of GATT 1994 and
Articles I, III, VI, XVI and XVII of GATS and in particular in relation to the Annex on
Movement of Natural Persons Supplying Services under the Agreement.⁴
Following inconclusive consultations, a panel was established to adjudicate the
EC complaint. Subsequent to the establishment of the panel, the European
Community requested, in accordance with Art 12.12 of the DSU, the suspension
of the panel’s work, in order to allow it to reach a mutually acceptable solution
with the United States.⁵ The panel suspended its work on 21 April 1997, and since
no request was tabled to reconvene within one year, the authority of the panel,
as per Art 12.12 of the DSU, lapsed on 22 April 1998.⁶ The WTO was never
notified of any mutually acceptable solution. The European Community has not
resurrected the original panel either.
It is difficult to draw any conclusions from two cases. Indeed, the inaction is
probably more telling in this context: WTO members, in general, prefer not to
submit such disputes to the WTO dispute settlement system.⁷

3 BoP

GATT, Art XII (for use by for developed countries), and Art XVIIIb of the GATT
(for use by developing countries) permit the imposition of trade restrictions to
safeguard a country’s external financial position. The contingency is the threat to

⁴ See WTO DocWT/DS38/1 of 13 May 1996.

⁵ See WTO DocWT/DS38/5 of 25 April 1997.
⁶ See WTO Doc.WT/DS38/6 of 24 April 1998.
⁷ See the chapter dedicated to this issue in Matsushita et al (2006).
332 State Contingencies

the financial stability of the WTO member concerned. The inclusion of such pro-
visions was deemed necessary in light of the inflexibilities associated with the sys-
tem of fi xed (but adjustable) exchange rates that prevailed when the GATT/ITO
was originally negotiated: under fi xed exchange rates, a country with a payments
deficit cannot devalue easily. As import restrictions (in conjunction with export
subsidies) are equivalent to a nominal devaluation, allowing (temporary) import
barriers to deal with a balance of payments can make some sense. Depending, of
course, on the satisfaction of the Marshall/Lerner condition.
By virtue of Art XV of the GATT, the IMF had an important role to play in
fi xing the exchange arrangements for members of the GATT, newly acceding
countries, as well as for countries departing from the Fund. We quote from Art
XV.6 of the GATT:
Any contracting party which is not a member of the Fund shall, within a time to be deter-
mined by the CONTRACTING PARTIES after consultation with the Fund, become
a member of the Fund, or, failing that, enter into a special exchange agreement with the
CONTRACTING PARTIES. A contracting party which ceases to be a member of the
Fund shall forthwith enter into a special exchange agreement with the CONTRACTING
PARTIES. Any special exchange agreement entered into by a contracting party under
this paragraph shall thereupon become part of its obligations under this Agreement.
A specific committee was established to help administer the two GATT provisions
(Art II and Art XVIII), the WTO Committee on Balance of Payments Restrictions
(BoP committee). The WTO BoP committee carries out consultations in order to
review all restrictive import measures taken for BoP purposes. The committee
follows the procedures for consultations on BoP restrictions approved on 28 April
1970.⁸ A WTO member willing to apply a new restriction, or to raise the gen-
eral level of its existing restrictions, shall enter into consultations with the WTO
BoP committee within 4 months of the adoption of such measures. The member
adopting such measures may request that a consultation be held (Art XII.4a of the
GATT, and Art XVIII.12a of the GATT). A consultation can also be held upon
invitation by the chairman of the WTO BoP committee. All restrictions applied
for BoP purposes shall be reviewed periodically by the WTO BoP committee.
A simplified procedure for consultations is also available: it was approved on 19
December 1972.⁹ This procedure is available to LLDCs and developing countries
alike. The latter, however, can have recourse to it, provided that they are pursuing
trade liberalization efforts in conformity with the schedule presented to the com-
mittee in previous consultations.
There are notification requirements whenever recourse to BoP restrictions is
being made: a WTO member must notify to the WTO General Council the intro-
duction of, or any change in the application of trade-restricting import measures

⁸ See GATT Doc BISD 18S/45.

⁹ See GATT Doc BISD 20S/47. The WTO understanding on BOP restrictions refers to this
document as ‘simplified consultation procedures’.
 BoP 333

taken for BoP purposes. It must also notify any modification concerning the tim-
ing of withdrawal of such measures. Finally, every year, the WTO member con-
cerned shall make a comprehensive notification of all trade-restricting measures
taken for BoP reasons. Notifications shall include information at the tariff-line
level, the product coverage and trade flows affected.
At the request of any member, notifications may be reviewed by the WTO BoP
committee. Recourse to dispute settlement procedures is also available.
Most countries have shifted by now to flexible exchange rates. Given that
the exchange rate is a more appropriate instrument to deal with balance of pay-
ments disequilibria—as part of a comprehensive macroeconomic adjustment
program—the GATT provisions on BoP have become largely redundant: other
things equal, WTO members would rather devaluate and profit from increase in
export income, rather than impose a QR and keep their exchange rate intact.
Developing countries, however, made frequent use of Art XVIII(b) of the GATT
as cover for the use of QRs, subject to the largely ineffective surveillance by the
WTO GATT BoP committee (see Table 4.1):

Table 4.1 Invocations of Art XVIIIb of the GATT

Argentina 1972–78 1986–91

Bangladesh 1974–2001¹⁰
Brazil 1962–71 1976–91
Chile 1961–80
Colombia 1981–92
Egypt 1963–95
Ghana 1959–89
India 1960–97
Indonesia 1960–79
Korea 1969–89
Nigeria 1985–98
Pakistan 1961–2001
Philippines 1980–95
Peru 1968–91
Sri Lanka 1960–98
Tunisia 1967–97¹¹

During the Uruguay round surveillance of restrictions was strengthened. In

principle, surcharges or similar measures must be applied on an across-the-board
basis—as that is what is needed from a BoP perspective.

¹⁰ Bangladesh notified its intention to phase out its remaining restrictions by 1 January 2005 (see
WTO Doc WT/BOP/N/54 of 15 December 2000, and also WT/BOP/N/62 of 18 February 2004).
Subsequent to this notification, Bangladesh imposed import restrictions under Art XVIIIc of the
GATT, invoking infant industry protection, see the discussion in Chapter 2 and WTO Doc G/C/7
of 16 January 2002.
¹¹ Source: notifications to the BoP committee.
334 State Contingencies

Disputes have been brought against developing countries that invoked Art XVIII
of the GATT. A high profile case was brought by the United States in 1997 against
India. India justified QRs on over 2,700 agricultural and industrial product tariff
lines, by invoking Art XVIII(b) of the GATT. The panel found India’s measures
to be inconsistent with Art XI.1 of the GATT, and further that they could not
be justified through recourse to Art XVIII.11 of the GATT.¹² Most importantly,
the panel found (and the AB upheld) that, in contrast to the argument advanced
by India, BoP restrictions can be the subject-matter of judicial review. The AB, in
its report on India—Quantitative Restrictions, specifically noted the relevance of
institutional balance within the framework of the WTO Agreement, and rejected
India’s argument that only the WTO BoP committee could review the consistency
of its measures with the GATT, noting that (§ 105):
such a requirement would be inconsistent with Article XXIII of the GATT 1994, as
elaborated and applied by the DSU, and footnote 1 to the BOP Understanding which, as
discussed above, clearly provides for the availability of the WTO dispute settlement pro-
cedures with respect to any matters relating to balance-of-payments restrictions.
The panel originally, and the AB subsequently, found against India, because it
could not honour its burden of proof and explain why the removal of QRs would
lead to a change in its development policy, and ultimately to a deterioration of its
overall situation (§ 150):
In that respect, we note that, on the issue of whether India’s balance-of-payments restric-
tions were justified under Article XVIII:9, we note that in paragraphs 5.170 to 5.184
of its Report, the Panel took into account evidence adduced by both the United States
and India, including information from the report of the Reserve Bank of India. As to
whether India’s balance-of-payments restrictions were justified under Article XVIII:11
and the Note Ad Article XVIII:11, the Panel reached its conclusion after weighing all
the evidence and considering the arguments of the parties. With respect to the proviso
to Article XVIII:11, the Panel based its conclusion primarily on the fact that India failed
to supply convincing evidence that the removal of its balance-of-payments restrictions
would entail a change in its development policy. Moreover, it is clear from the Panel’s
analysis in paragraphs 5.211 and 5.220 of its Report that it critically assessed the views of
the IMF on this issue.

¹² The United States carried the burden of proof with respect to Art XI of the GATT, and India
then carried the burden of proof with respect to Art XVIII of the GATT. Oddly, however, the panel
held (and the AB upheld in § 138 of its Report) that the United States carried the burden of proof
with respect to the Interpretative Note ad Art XVIII.11 of the GATT. Th is note concerns the con-
ditions under which a WTO member should progressively relax import restrictions that have been
in place. It is highly unlikely that a WTO member affected by such restrictions adequately knows
the facts mandating such behaviour. The panel probably relied on the solemnity accompanying
the IMF involvement on this score. Still, it would have been on safer grounds had it imposed the
burden of proof on India.
 Exchange Restrictions 335

4 Exchange Restrictions

GATT, Art XV.9 allows WTO members to impose exchange restrictions:

Nothing in this Agreement shall preclude:

(a) the use by a contracting party of exchange controls or exchange restrictions in accord-
ance with the Articles of Agreement of the International Monetary Fund or with
that contracting party’s special exchange agreement with the CONTRACTING
PARTIES, or
(b) the use by a contracting party of restrictions or controls in imports or exports, the
sole effect of which, additional to the effects permitted under Articles XI, XII, XIII
and XIV, is to make effective such exchange controls or exchange restrictions.
The contingency here is again the threat to the financial stability of WTO members.¹³
The relationship between this provision and Art XII of the GATT is underlined in
§ 2 above. The IMF has, of course, a natural role to play in this context, and Art XV
of the GATT explicitly recognizes its role.¹⁴ In recent years, cooperation between the
WTO and the IMF on BoP related issues has been strengthened.¹⁵ An agreement
was signed¹⁶ that provides for the establishment of a steady channel of information
between the two institutions: the WTO has invited the IMF to participate as observer
in a series of meetings relating to its subject-matter, and vice-versa.
In general, whether a measure is an exchange restriction has to do, according to
the IMF, with the question as to whether it involves a direct governmental limita-
tion on the availability or use of foreign currencies.¹⁷
The panel, in its report on Dominican Republic— Import and Sale of Cigarettes
dealt with an exchange restriction imposed by the Dominican Republic. The facts
of the case are reflected in §§ 7.135–7.137 of the report in the following terms:
A reading of the 1991 Resolution reveals the nature of foreign exchange fee as it was
applied in 1991. Paragraph 12 provides: “the Central Bank shall charge users of official

¹³ The panel, in its report on Dominican Republic—Import and Sale of Cigarettes, held that the
party invoking Art XV of the GATT carries the burden of proof associated with this provision
(§ 7.131).
¹⁴ Hence, deference to the IMF is warranted in this context, whereas this is not necessarily
the case in the context of BoP provisions, as we saw in the AB report on India—Quantitative
Restrictions.
¹⁵ The BoP and the exchange restrictions are related: it could be the case that a BoP restriction
is complemented by an exchange restriction, in case, for example, that there is increased specula-
tion about the exchange rate of the country imposing the restriction. It could also be that a WTO
member imposes only an exchange restriction (and thus opens the door to barter trade); assuming
a WTO member does not wish to open the door to barter trade, it will of course impose a BoP
restriction.
¹⁶ See WTO Doc WT/L/195 of 18 November 1996.
¹⁷ See § 7.132 of the panel report on Dominican Republic—Import and Sale of Cigarettes.
Exchange restrictions do not have to be recorded in the schedules of concession, precisely because
they are a contingency which could be unanticipated when concessions had been scheduled.
336 State Contingencies
foreign exchange transactions and commercial banks, for the delegation of foreign
exchange operations, the equivalent in Dominican Pesos (RD$) of two and a half per cent
(2–1/2%) of the selling exchange rate applied to each transaction”. This means that the
fee was originally imposed on foreign exchange transactions conducted in both official
and private foreign exchange markets through either the Central Bank or commercial
banks. This Resolution actually required that the foreign exchange fee be paid for all
kinds of foreign currency transactions, including both selling and purchasing transac-
tions. This measure actually increased the cost for the use of foreign currency, for all kinds
of transactions regardless of whether the transaction was related to importation. The Panel
considers that the foreign exchange fee as it was applied in 1991 could be characterized
as an exchange measure, as it is possible that the IMF would deem the increase of cost for
the availability of foreign exchange as a means of “restriction” on the availability of the
foreign currency. However, it is not necessary for the Panel to decide on the measure as
it was applied in 1991. The measure to be examined by the Panel for the resolution of the
dispute is the currently applied foreign exchange fee measure.
With regard to the currently applied foreign exchange fee measure, the Panel notes
that the relevant provisions in the 1991 Resolution of the Monetary Board have been
amended by later resolutions. On the issue of whether the fee is imposed on all types of
transactions or solely on import-related transactions, the applicable rules are those under
the Resolution of 20 August 2002. Paragraph 2 of this Resolution provides “exchange
commissions shall henceforth no longer be applied to external debt service payments,
capital repatriation, remittances of profits, technology transfers, sales for travel expenses
and medical services, credit cards and other services”. The Dominican Republic confirms
in its replies to a Panel question that “[t]he foreign exchange fee applies only to import-
ation of goods. It does not apply to foreign exchange payments of non-import related
services, nor to foreign currency payments made by Dominican Republic residents, nor
to remittance of dividends from companies located in the Dominican Republic.
The Panel considers that the ordinary meaning of the “direct limitation on availability
or use of exchange . . . as such” means a limitation directly on the use of exchange itself,
which means the use of exchange for all purposes. It cannot be interpreted in a way so as
to permit the restriction on the use of exchanges that only affects importation. To con-
clude otherwise would logically lead to the situation whereby any WTO Member could
easily circumvent obligations under Article II:1(b) by imposing a foreign currency fee or
charge on imports at the customs and then conveniently characterize it as an “exchange
restriction”. Such types of measures would seriously discriminate against imports while
not necessarily being effective in achieving the legitimate goals under the Articles of
Agreement of the IMF. Therefore, the Panel finds that because the fee as currently applied
is imposed only on foreign exchange transactions that relate to the importation of goods,
and not on other types of transactions, it is not “a direct limitation on the availability or
use of exchange as such”. (original emphasis)

Essentially, the Dominican Republic had modified its original measure which
covered all transactions, and put in its place a measure which covered only
imports. It was thus discriminating against imports, and was not genuinely
addressing the issue that, in name, it was purporting to address. The panel con-
cluded as much. In order, however, to cement its opinion on this issue, the panel
 Exchange Restrictions 337

decided to consult with the IMF. In the panel’s view, in case the IMF authorities
held that the measures imposed by the Dominican Republic were in accordance
with the IMF Articles of Agreement, these measures would ipso facto be deemed
to be GATT consistent as well (§ 7.139). This result is the only one which, in the
Panel’s view, is consistent with a textual reading of Art XV.2 of the GATT:
In all cases in which the CONTRACTING PARTIES are called upon to consider
or deal with problems concerning monetary reserves, balances of payments or for-
eign exchange arrangements, they shall consult fully with the International Monetary
Fund. In such consultations, the CONTRACTING PARTIES shall accept all fi nd-
ings of statistical and other facts presented by the Fund relating to foreign exchange,
monetary reserves and balances of payments, and shall accept the determination of
the Fund as to whether action by a contracting party in exchange matters is in accord-
ance with the Articles of Agreement of the International Monetary Fund, or with
the terms of a special exchange arrangement between that contracting party and the
CONTRACTING PARTIES . . .
So, the panel decided to submit two questions to the IMF:
(1) how the restriction was being implemented by the Dominican Republic, and
(2) whether, as applied, it should be accepted as a restriction in accordance with
the Articles of the Agreement of the IMF.
In light of the IMF’s response that the measure at hand was not a multiple cur-
rency practice (since it was only targeting imports, and, as such, could not qualify
as an exchange restriction in accordance with Art XV of the GATT), and, hence,
was not approved by the IMF anymore, the panel concluded that the Dominican
Republic could not justify its measures through recourse to Art XV.9 of the
GATT (§§ 7.143–7.145):
On the first issue, the IMF General Counsel replied that:
(a) The ‘exchange commission’ is levied under the legal authority of the Banco Central
de la República Dominicana (BCRD). Since its introduction in January 1991, the
commission has undergone a number of changes in the way that it is levied. Initially,
the commission was payable on sales of foreign exchange and was calculated as a
percentage of the selling rate.
(b) Since August 2002, however, pursuant to the Agreement between the BCRD and
the Directorate General for Customs (DGC) of August 22, 2002, the commission
has been collected in its entirety by the DGC. Moreover, although the commission
is still referred to as an “exchange commission” (because it is levied on the basis of
the legal authority vested in the BCRD to charge a commission on sales of foreign
exchange), the commission is no longer payable on sales of foreign exchange. Rather,
it is payable as a condition for the importation of goods, and the amount of the com-
mission is now calculated exclusively on the CIF valuation of the imported goods as
determined by the DGC (Article 1 of the Agreement between the BCRD and the
DGC). By Notice of Resolution No. 1 of the Monetary Board of October 22, 2003,
the rate of the commission was increased to ten per cent in October 2003.
338 State Contingencies
On the second issue that the Panel requested information on, the reply states:
As applied since August 2002, the exchange commission is no longer a measure
subject to Fund approval. As noted above, the commission is no longer payable on
sales of foreign exchange. It is payable as a condition for the importation of goods
and the amount to be paid is based on the CIF value of the imported goods (rather
than the amount of foreign exchange sold to an importer for the payment of goods).
As such, it does not constitute a multiple currency practice or an exchange restriction
notwithstanding its label or the fact that the commission is charged on the basis of
the legal authority vested in the BCRD to charge an exchange commission on sales
of foreign exchange. For the same reasons, it is not an exchange control measure.
(emphasis added)
The Panel fully agrees with the opinion of the IMF. For the reasons set out above by
the Panel and considering the opinion expressed by the IMF, the Panel finds that the for-
eign exchange fee measure as it is currently applied by the Dominican Republic does not
constitute an “exchange restriction” within the meaning of Article XV: 9(a) of the GATT
1994. ( original emphasis)¹⁸
This finding was not appealed. It is instructive in two respects: first, as to the
allocation of the burden of proof: it is for the WTO member invoking this pro-
vision to carry the burden of proof; second, as to the deference to the IMF. Such
deference is warranted in light of the wording of Art XV of the GATT which,
wisely, acknowledges that the IMF is best equipped (in terms of institutional
expertise) to deal with exchange restrictions.

5 Antidumping (Art VI of the GATT)

5.1 An unfair (?) price discrimination

GATT, Art VI was designed to discipline dumping. Dumping is defined as a
price differential between the price in the home market (normal value) and a
comparable price at the export market (export price). The difference between the
two prices is the dumping margin. To the extent that dumping causes injury¹⁹ to
the state importing dumped imports, it can lead to the imposition of AD duties.
AD duties are a duty super-imposed on the bound duty, and equal to the dump-
ing margin (the difference between normal value and export price).
GATT, Art VI condemns but does not outlaw dumping. Dumping has been
recognized as an unfair practice by the AB (US—Line Pipe, § 80). Unfairness

¹⁸ See also § 7.150 of the panel report which states that ‘In fact, the reply of the IMF General
Counsel concludes that since the foreign exchange commission does not constitute an exchange
restriction, “the issue of its consistency or inconsistency with the Funds Articles for purpose of
paragraph 8 of the Co-operation Agreement does not arise”. The Panel fully agrees with this
statement’.
¹⁹ As will be shown in more detail infra, the AD Agreement adopts an injury to competitors, as
opposed to injury to competition standard.
 Antidumping (Art VI of the GATT) 339

does not amount to illegality, however.²⁰ As a result, the GATT panel report on
Canada—FIRA naturally recognized that governments are under no obligation
to prevent dumping, a private decision (§ 5.18):
the General Agreement does not impose on contracting parties the obligation to prevent
enterprises from dumping.²¹
Consequently, qualifying dumping as unfair practice means that, although it is
not illegal to dump, WTO members have a right to react to it. Indeed, all the
AD Agreement does is to regulate the conditions under which the right to react
will be exercised. On the other hand, AD duties are the only permissible way to
address dumping (Art 18.1 of the AD). The AB, in its report on US—Off set Act
(Byrd Amendment), ruled as much, and upheld the panel ruling on this score. The
Continued Dumping and Subsides Offset Act (CDSOA)²² was being challenged.
Through this measure, the United States was distributing proceedings from the
imposition of AD duties only to the US economic operators that had supported
the initiation of an investigation. According to the complainants this measure
dissuaded exporters from dumping: since it acted against dumping and it had not
taken the form of AD duties, the AB found it to be inconsistent with Art. 18.1 of
the AD (§§ 255–6 and 265).²³
GATT, Art VI is not explicit about the procedures that WTO members should
follow in order to impose AD duties. The procedural requirements have been
developed over the years in side agreements dealing specifically with AD. Until
the Uruguay round AD Agreement, participation in the agreement was optional:
all GATT Contracting Parties merely had to respect was Art VI of the GATT.
Those who signed the Tokyo round AD Agreement, on the other hand, were
bound by its disciplines as well. The disciplines on dumping underwent substan-
tial transformation during the Uruguay round negotiations. The new AD agree-
ment²⁴ is a multilateral construct, that is, participation is not optional anymore.
The AD Agreement regulates the investigation process that WTO members
wishing to impose AD duties must follow. Throughout the investigation, WTO

²⁰ On the historical underpinnings of the notion fair trade, see the excellent analysis of Beviglia-
Zampetti (2006).
²¹ By the same token, the panel report on Japan—Semi-conductors held in § 120: ‘The provi-
sion was silent on actions by exporting countries. The Panel therefore found that Article VI did
not provide a justification for measures restricting the exportation or sale for export of a product
inconsistently with Article XI.1’. Dumping is a private decision. There are only few instances where
the WTO Agreement requests from WTO members to discipline private behaviour. One such rare
instance is the GATS Telecoms Reference Paper.
²² The official acronym of the US Byrd Amendment.
²³ This is not an unproblematic decision. Horn and Mavroidis (2005a) point out that it is
equally plausible that, as a result of Byrd payments, exporters continue to dump and possibly, more
aggressively so. On this score, see also Bhagwati and Mavroidis (2004). Th is ruling suggests a wide
understanding of the obligation included in Art 18.1 of the AD.
²⁴ See, in general, Palmeter (1995) who discusses in a compelling manner all innovations con-
cerning the AD Agreement that were agreed during the Uruguay round. Moore (2004) notes that,
for many US Congressmen, preserving AD was a top priority during the Uruguay round. Horlick
(1993) reflets a similar discussion.
340 State Contingencies

members must respect certain due process clauses. They must, in a nutshell, per-
form an objective examination and ensure that interested parties:
(1) are essentially given a chance to adequately present their views (right of
defence);
(2) have access to all information having a bearing on the case (right to access all
relevant information).²⁵
The term ‘interested parties’ is defined in Art 6.11 of the AD²⁶ as including
exporters, their government(s), as well as representatives of the competing domes-
tic industry. WTO members might wish to add to that list other parties, such as
consumers’ organizations.²⁷ The AB, in its report on EC—Bed Linen (Art 21.5—
India), understood the duty to perform an objective examination entails a duty of
even-handedness (§ 114):
In short, an ‘objective examination’ requires that the domestic industry, and the effects
of dumped imports, be investigated in an unbiased manner, without favouring the interests
of any interested party, or group of interested parties, in the investigation. The duty of the
investigating authorities to conduct an ‘objective examination’ recognizes that the deter-
mination will be influenced by the objectivity, or any lack thereof, of the investigative
process. (original emphasis)
The AD Agreement contains numerous provisions that reflect this obligation
(Art 6 of the AD), such as Art 6.1 of the AD (right to information, opportunity

²⁵ Stewart et al (2006) conclude that transparency and due process are the key areas in the evo-
lution of GATT/WTO provisions dealing with AD.
²⁶ From now on, references are made to provisions of the Uruguay round AD Agreement, unless
otherwise specified. This does not mean, however, that Art VI of the GATT is completely obso-
lete. GATT, Art VI ‘to the extent that there is no conflict with the provisions included in the AD
Agreement, remains, by virtue of the General Interpretative Note regulating the legal relationship
between the GATT and the other Annex 1A agreements, legally relevant. Case-law has consist-
ently acknowledged as much, see, for example, US—Off set Act (Byrd Amendment). There is one very
important consequence of the continuing legal relevance of Art VI of the GATT: the AD Agreement
is silent on non market economies (NMEs), but the Interpretative Note ad Art VI. 2 of the GATT
refers to them and allows for the possibility to neglect normal values in such countries. It reads: ‘It is
recognized that, in the case of imports from a country which has a complete or substantially complete
monopoly of its trade and where all domestic prices are fixed by the State, special difficulties may exist
in determining price comparability for the purposes of paragraph 1, and in such cases importing
contracting parties may find it necessary to take into account the possibility that a strict comparison
with domestic prices in such a country may not always be appropriate’ (emphasis added). In practice,
WTO members have accorded NME status to countries which do not fix all prices; they have also
used prices in other markets as surrogate for normal values in NMEs. For example, the European
Community used Mexico and Brazil as surrogate countries for Vietnam when imposing AD duties
against imports of bicycles and leather shoes respectively from Vietnam. The United States imposed
AD duties ranging from 37%–64% on imports of Vietnamese catfish. They calculated the dumping
margin by assuming the price that would have been charged had the fish farm existed in Bangladesh,
using water bought in India and transported by Bangladeshi truckers, with labour purchased at a price
the US department thought should prevail in Vietnam (see The Economist, print edition, 2 November
2006). China still has NME-status in US practice, see Stewart et al (2006).
²⁷ These are the so-called public interest clauses.
 Antidumping (Art VI of the GATT) 341

to present in writing all evidence); Art 6.2 of the AD (right of defence through-
out the AD investigation); Art 6.3 of the AD (right to provide oral statements);
Art 6.4 of the AD (right to see all relevant information); Art 6.5 of the AD (treat-
ment of confidential information); Art 6.8 of the AD (conditional only recourse
to best information available); Art 6.9 of the AD (right to be informed of the final
determination before it is made public); Art 6.13 of the AD (obligation of author-
ities to account for difficulties in providing information).²⁸
Many WTO members do not transpose as such the AD Agreement into their
domestic legal order, but instead add language that is not necessarily compatible
with the letter and the spirit of their WTO obligations. This phenomenon has led
over the years to extensive challenges not only of specific applications, but also of
national AD legislation as such²⁹ before the WTO. A series of papers appearing
in Boltuck and Litan (1991) for example, make the point that the US implemen-
tation of the WTO AD Agreement has not been in line with the international
obligations of the United States.³⁰
The AD is by now a widely used instrument; dozens of WTO Members
avail themselves of this possibility.³¹ A number of reasons have contributed to
the proliferation of AD duties:³² first, AD allows for targeted protection, in the
sense that the WTO member imposing AD duties can hit one specific source of
supply;³³ second, the provisions for calculating dumping are quite favourable to
the investigating authority, since they are allowed to deviate from actual prices
and construct them, subject to little discipline, as we will see infra; third, the
special standard of review included in Art 17.6 of the AD is arguably more defer-
ential to the investigating authority (than the general standard included in Art 11
of the DSU, which requests panels to make an objective assessment of the matter

²⁸ The political economy of domestic investigating authorities suggests that there is true need
to ensure that an even-handedness requirement will be observed. One cannot help irresponsibly
thinking what the situation, in terms of AD duties imposition, would be like were they to be
imposed by one World AD Authority.
²⁹ When legislation is challenged as such, the complainant is claiming that the legislation even
before (or in absence of) application is GATT inconsistent.
³⁰ Stewart et al (2006) go so far to state that the mandate of the US negotiators negotiating the
original GATT was to ensure that there would be no need to modify the US laws on AD in force
at that time.
³¹ See Human et al (2003). Many scholars have studied the proliferation of AD duties and the
reasons behind this phenomenon. Prusa (2001 and 2005) offers a very comprehensive look on
this issue. Finger and Nogues (2006) have studied practice in seven Latin American countries
(Argentina, Brazil, Chile, Colombia, Costa Rica, Mexico, and Peru). They confirm that the use of
AD has increased substantially over the years.
³² Kucik and Reinhardt (2007) have argued that countries are more likely to form domestic
AD mechanism when they join the WTO: correcting for that, however, they have estimated that,
having an AD system makes a country 4.4 times as likely to join the WTO, and induces it to set its
average tariff bindings a full 17.9 percentage points than otherwise.
³³ In contrast, safeguards, as will be shown infra, have to be applied on a non-discriminatory
basis.
342 State Contingencies

before them).³⁴ The AD standard of review requests that WTO adjudicating bod-
ies accept any permissible interpretation of the AD Agreement reached by an investi-
gating authority. Where exactly the difference lies between Art 17.6 of the AD and
Art 11 of the DSU is still unknown. Actually, many panels have adopted an attitude
of reconciliation between the two standards:³⁵ the panel report on US—Hot-Rolled
Steel took the view that, in order to evaluate whether the interpretation reached is
a permissible one, the starting point of the panel’s analysis should be the VCLT
(§§ 7.26 and 7.27). On appeal, the AB confirmed the panel’s position. In §§ 57–62
of its report, it took the position that, in principle, the two standards should not
be viewed as being mutually exclusive (inclusio unius, exclusio altrius). To reach a
conclusion whether more than one interpretation could be permissible, exhaustion
of the interpretative elements reflected in the VCLT was the necessary first step.
What is clear is that there is no room for de novo review in Art 17.6 of the AD:
the panel, in its report on EC—Pipe Fittings, held that panels cannot substitute
their own judgment to that of the investigating authority that had previously
dealt with the case (§ 76):
In light of this standard of review, in examining the matter referred to us, we must evalu-
ate whether the determination made by the European Communities is consistent with
relevant provisions of the Anti-Dumping Agreement. We may and must find that it is con-
sistent if we find that the European Communities investigating authority has properly
established the facts and evaluated the facts in an unbiased and objective manner, and
that the determination rests upon a “permissible” interpretation of the relevant provi-
sions. Our task is not to perform a de novo review of the information and evidence on
the record of the underlying anti-dumping investigation, nor to substitute our judgment
for that of the EC investigating authority even though we may have arrived at a different
determination were we examining the record ourselves. (original emphasis)
Such pronouncements confirm the intuition that probably, the most appropri-
ate way to describe the standard of review embodied in Art 17.6 of the AD is a

³⁴ Stewart et al (2006) point out, that this standard has only infrequently, if ever, been used in
WTO litigation. It is also true, however, that most commentators have expressed their disappoint-
ment with the manner in which WTO panels have acted in this context. A WTO panel went so far
as to defy an AB ruling which had outlawed zeroing, a practice used by some authorities according
to which positive dumping margins are zeroed and the overall dumping margin is inflated since
only few (favourable to the authority) transactions matter for purposes of establishing the dumping
margin (US—Zeroing (Japan), WT/DS322/R). On appeal, the AB, in its report on US—Zeroing
(Japan), clarified that all forms of zeroing are illegal, irrespective of the methodology or the stage
of the proceedings at which zeroing has been used. On this case, and a survey of the literature,
see Mavroidis et al (2007). A similar pro-deferential attitude is also observable when it comes to
domestic courts discussing antidumping impositions. The European Court of First Instance (CFI)
for example, show a rather deferential attitude in its judgment on Case T-413/03 Shandong Reipu
Biochemicals Co Ltd v Council of the European Union [2006] ECR II-02243. Compare its standard
of review with that employed in Case C-13/03 Commission of the European Communities v Tetra
Laval BV [2005] OJ C106/4, where the ECJ found that the Commission had not applied a proper
standard of review when refusing to clear a merger.
³⁵ There are, however, examples to the opposite, such as the panel report on US—Softwood
Lumber V (§ 7.22).
 Antidumping (Art VI of the GATT) 343

compulsory inclusion of the in dubio mitius maxim. This maxim which is not
enshrined in the VCLT but is a direct reflection of the sovereignty principle, sug-
gests that when in doubt, a court should presume absence of transfer of sover-
eignty to the international plane, rather than the opposite. The panel report on
Argentina—Poultry Antidumping Duties, for example, was requested to judge
whether 46% of all domestic producers should be considered as a major proportion
of the total domestic production, in accordance with Art 4.1 of the AD. Without
delving too much into a thorough discussion of this issue, the panel accepted that
this is indeed a permissible interpretation of the term. In other words, following
exhaustion of the elements enshrined in the VCLT, and absent a clear case of a
unique interpretation as the only tenable interpretation, panels, when adjudicat-
ing disputes in the AD context, will arguably accept as lawful any interpretation
which satisfies a reasonableness test.

5.2 Finding dumping

AD investigations can be launched upon request, or ex officio, the latter option
being rather exceptional. AD, Art 5.4 regulates the standing requirements for
initiations upon request: the supporting companies must represent 50% of the
producers that have expressed an opinion and at least 25% of the total production
(ie those that have and those that have not expressed an opinion), otherwise no
investigation will be launched.³⁶ An investigating authority receiving a petition
must ensure that it is not only accurate, but also sufficient. Sufficiency takes us
to a discussion on the quantum of proof (burden of persuasion) necessary for an

³⁶ Liberalization of investment, and free movement of production factors more generally, has
given food for thought to investigating authorities in charge of AD investigations. The reasons
for off shoring and outsourcing have been explained in on original manner in Friedman (2005).
Because of offshoring and outsourcing, the question arises for investigating authorities: who are
they protecting? Increasingly, because of outsourcing, it is the case that say, EC companies request
imposition of AD duties against inputs of EC goods (since, many EC companies, established in
the European Community, outsource part of their production abroad). In October 2006, the
European Community imposed AD duties of 16.5% and 10% on certain leather shoes imported
into the EC market. Although many EC companies produce leather footwear inside the European
Community, a significant number of EC companies have outsourced the production of footwear to
third countries, while keeping other parts of their operations in the European Community. Those
EC companies that produce leather shoes in the third countries were subjected to AD duties. Such
incidents have caused concern in Brussels, to the point that the Trade Commissioner has issued a
Green Paper asking, inter alia, whether the EC investigating authority should somehow take into
account the interests of companies which have retained significant operations and employment in
Europe, even though they have moved some part of their production outside the EC market. See
pp 6 and 7 in Commission of the European Communities, Global Europe, Europe’s trade defense instru-
ments in a changing global economy, A Green Paper for public consultation, Brussels, 6 December
2006, COM (2006) Final. Off shoring (trade in tasks) has been hailed by Richard Baldwin as the
second great unbundling (the first being the distancing of consumption from production through
trade liberalization), see <http://www.tinyurl.com/2ol2n8>. Note that there is by now at least
some empirical evidence that offshoring can have positive implications for national welfare, see
Grossman and Rossi-Hansberg (2006).
344 State Contingencies

investigation to be launched. Assuming it accepts the petition, it will launch an

investigation. The total duration of an investigation should not, in accordance
with Art 5.10 of the AD, exceed 18 months from the date of initiation.³⁷
The AD Agreement does not discuss in extenso the period of investigation
(POI) during which the data with respect to dumping and injury should be
collected. The WTO Antidumping Committee (AD Committee) has adopted a
recommendation which regulates the POI: it³⁸ distinguishes between a period of
data collection for the dumping investigation and a period of data collection for
the injury investigations and provides, inter alia:
(1) that the period of data collection for dumping investigations should normally
be twelve months, and in any case no less than 6 months;
(2) that this period should end as close to the date of initiation as is practicable;
(3) that the period of data collection for injury investigations should normally be
at least 3 years, unless a party from whom data is being gathered has existed
for a lesser period; and
(4) that the period of data collection for injury investigations should include the
entirety of the period of data collection for the dumping investigation.³⁹
An investigating authority must perform a fair comparison between two prices; it
will carry the bulk of the burden of proof; interested parties, however, are under
the duty to cooperate (Art 2.4 of the AD). This duty is detailed in various provi-
sions of the AD Agreement but essentially requests from an investigating author-
ity⁴⁰ to construct normal value and export price in a WTO consistent manner
and compare them, after it has made due allowances, in accordance with the AD
Agreement.⁴¹

³⁷ Occasionally, requests for investigation are withdrawn. The argument has been made that
withdrawals of requests by the domestic industry might be signalling collusion with exporters.
Rutkowski (2007) examines 45 such withdrawals in EC practice (between 1992 and 2004) and
tests this hypothesis.
³⁸ See WTO Doc G/ADP/6 (adopted by the Committee on 5 May 2000), and entitled
Recommendation Concerning the Periods of Data Collection for Anti-Dumping Investigations.
³⁹ The AB in its report on Mexico—Anti-dumping Measures on Beef and Rice upheld a panel
finding to the effect that Mexico had violated its obligations by using 15-month-old data without
explaining why more recent data could not have been reasonably used (§ 167).
⁴⁰ The AB in its report on US—Hot Rolled made it amply clear that the duty to perform a fair
comparison rests with the investigating authority, interested parties having to cooperate but being
under no duty to perform a fair comparison (§ 178).
⁴¹ It could, however, be the case that parties might choose not to cooperate. The AD Agreement
moves the process out of this potential deadlock by introducing the recourse to best information avail-
able (BIA), an institutional possibility to conduct an investigation when facing non-cooperation.
Assuming that the exporter refuses access to information or otherwise significantly impedes the
process, an investigating authority can move on and complete the record on the basis of BIA, provided
that the conditions of Art 6.8 of the AD and Annex II of the AD have been met. These conditions
can be summed up as follows:
(1) a reasonable period to provide the information has been set (Art. 6.8 of the AD);
(2) the investigating authority has specified the information requested (Annex II, § 1);
 Antidumping (Art VI of the GATT) 345

The investigating authority must, in principle, use actual prices; it can, never-
theless, assuming the conditions of Art 2.2 of the AD have been met, use an alter-
native basis for deriving the normal value if:
(1) there are no sales in the exporting country of the like product in the ordinary
course of trade; or
(2) sales in the exporting country’s market do not ‘permit a proper comparison’
because of ‘the particular market situation’; or
(3) sales in the exporting country’s market do not ‘permit a proper comparison’
because of their low volume (low volume sales).
The first alternative is further detailed in Art 2.2.1 of the AD, which states that
sales below (fi xed and variable) cost of per unit production plus administra-
tive, selling and general expenses (SG&A), can be regarded as sales not in the
ordinary course of trade.⁴² The second alternative is nowhere detailed in the AD
Agreement, and there is no relevant practice either.⁴³ As to the third alternative,
footnote 2 to the AD Agreements states that:
Sales of the like product destined for consumption in the domestic market of the export-
ing country shall normally be considered a sufficient quantity for the determination of
the normal value if such sales constitute 5 per cent or more of the sales of the product
under consideration to the importing Member, provided that a lower ratio should be
acceptable where the evidence demonstrates that domestic sales at such lower ratio are
nonetheless of sufficient magnitude to provide for a proper comparison.
Assuming that it is presented with one of the three scenarios described above,
the investigating authority can set aside the price in the exporter’s home market,
and pick one of the two alternative bases for the calculation of NV, as per Art 2.2
of the AD:
(1) third country sales, that is, the investigating authority will use data from sales
of the like product in an ‘appropriate’ third country, provided the price is
‘representative’;⁴⁴ or

(3) the information has not arrived in a timely fashion (Annex II, § 3)
(4) the information has not been appropriately transmitted (Annex II, § 2);
(5) the information falls short of what has been requested and the party concerned has not
acted to the best of its ability (Annex II, § 5);
(6) the party concerned has been informed of the reasons why recourse to BIA has been
decided, and it has not reacted (Annex II, § 6);
(7) when using information from secondary sources, investigating authorities should do so
with special circumspection (Annex II, § 7).
⁴² The AB, in its report on US—Hot-Rolled Steel, has suggested that the list of Art 2.2.1 of the
AD is not exhaustive; in its view, there could be other cases, where sales should not be considered to
be in the ordinary course of trade (§ 142).
⁴³ A proposal has been tabled during the Doha round to scrap this possibility altogether.
⁴⁴ That is, assuming that the United States investigates dumping practices by a Korean company,
it can use sales by the company at hand to say Sweden and use the price obtained in the Swedish
346 State Contingencies

(2) constructed price, that is, the investigating authority can calculate de novo the
NV, which will be the arithmetic addition of the following elements:
(a) the cost of production in the country of origin;
(b) a reasonable amount for SG&A expenses; and
(c) a reasonable amount for profits.
The AD Agreement does not allow full discretion to investigating authorities
when constructing the normal value. In short, they must:
(1) with respect to cost of production, follow the disciplines embedded in Art
2.2.1.1 of the AD;
(2) with respect to SG&A and profits, follow the disciplines included in Art
2.2.2 of the AD.⁴⁵
Once the normal value has been established (either using market or constructed
price), it will be compared to the export price.⁴⁶ The investigating authority can
proceed to make due allowances for factors that affect price comparability (Art 2.4
of the AD).⁴⁷ At the end of the day, it must compare two prices at the same level
of trade, usually the ex-factory level. The AD Agreement provides for two nor-
mal methods at the disposal of investigating authorities in order to perform the
comparison:
(1) weighted average to weighted average of transactions as close as possible time-
wise;
(2) transaction to transaction, again respecting the time-proximity
requirement.⁴⁸

market as normal value. The letter of law leaves substantial discretion to investigating authorities
as to the choice of third country (provided of course that the price used is representative). Practice
is scarce, most investigating authorities preferring to use the other option (constructed price). The
United States, for example, which occasionally uses third country sales, will usually select one coun-
try and, if there is more than one candidate, its final choice will be based on considerations such
as: similarity of the products exported (to the United States and the third country at hand); the
country to which the largest volume of sales is directed; and other appropriate factors (see WTO
Doc G/ADP/AHG/W/166 of 26 November 2004).
⁴⁵ The disciplines of these provisions require from the investigating authority to base their inves-
tigation on information provided by the books of the company under investigation, to the extent
that they conform with the generally accepted accountancy principles (GAAP).
⁴⁶ This is usually a market price, unless, for example, the exporter has vertically integrated with
the importer and there are good reasons to believe that the export price is not the outcome of an
arm’s length transaction.
⁴⁷ Allowances, however, do not include some obvious candidates, such as start up costs, where
the export price could be much higher than the price of the same good later in time, see p 9 in
Commission of the European Communities, Global Europe, Europe’s trade defense instruments in a
changing global economy, A Green Paper for public consultation, Brussels, 6 December 2006, COM
(2006) Final.
⁴⁸ Note, however, that Art. 2.4.2 of the AD allows, in specific circumstances, a comparison
between a weighted average and a specific transaction.
 Antidumping (Art VI of the GATT) 347

The last hurdle is that the dumping margin must be above the de minimis thresh-
old established in Art 5.8 of the AD (it must be more than 2%); unless this is the
case, the investigation must be immediately terminated.⁴⁹
In practice, dumping investigations have a ‘national’ character: country A
might end up imposing duties on all exporters from country B (Art 9.2 of the
AD). Indeed, Art 2 of the AD defines dumping in terms of countries with no
reference to individual firms. Mavroidis and Sapir (2007) have made the argu-
ment that the AD Agreement is incoherent in this respect: it simply cannot be
the case that dumping is at the same time a private business practice (as the AD
Agreement itself recognizes in Art 5 of the AD when it talks of the identity of each
known exporter, and Art 6 of the AD where it requests that the authorities deter-
mine an individual margin of dumping for each known exporter or producer),
and that duties can be imposed on a nationwide basis. Yet, this is precisely what
the AD Agreement indeed condones. Importantly, the AD Agreement does not
exclude that different dumping margins are calculated for different exporters: the
working hypothesis thus, is that all producers in a given country are presumed to
dump, if one is dumping, and that there is nothing wrong with all of them dump-
ing at different degrees. This cannot be.
Under perfect competition, the law of one price would hold, at least within
the same geographical market. This is, however, a rather hypothetical scenario.
The law of one price does not hold because most products competing in a rele-
vant product market are not identical. But even if two products are homogeneous
products, it is by now commonplace in economic analysis that price disper-
sion is likely to be the rule because of imperfect information and search costs
for consumers. Product differentiation and imperfect information imply that
companies might be in a position to exercise some market power and thus price
discriminate.
The operationalization of the possibility to impose duties on a countrywide
basis is equally paradoxical. The AD Agreement requests from investigating

⁴⁹ We should note here that terminating an investigation is not without consequences. Prusa
(1992) looks at a wider sample of cases that also includes cases where a petition has been with-
drawn. His data shows that these withdrawn cases have at least as great an effect on trade as cases
which resulted in duties. What he terms nuisance suits, that is, petitions with low probability of
success, actually confer large gains to both domestic and foreign firms. Actually, an effect on the
market might exist from the moment an announcement of a petition is made public. On the other
hand, some national investigating authorities are more prone to find dumping (and eventually
establish injury as well) than others. Knetter and Prusa (2003) note that of 800 cases investigated
between 1980–98, the US Department of Commerce (DOC), in charge of establishing dumping
margin, issued a negative dumping decision only in 28 cases (3.5% of the total); on the other hand,
the US International Trade Commission (ITCom), in charge for establishing injury, made negative
injury decisions in 37.5% of all cases submitted to it during the same period. Finally, the timing of
the fi ling might have something to say about the chances to succeed: Feinberg (2005), and Knetter
and Prusa (2003) discuss how business-cycle effects might affect the chance to prevail when fi ling
for antidumping duties. Feinberg (2005), using financial health of the domestic industry as proxy
for (lack of) injury, showed why in a booming market it will be hard to demonstrate that injury has
occurred as a result of dumping.
348 State Contingencies

authorities to calculate dumping margins for all known exporters. So if there are
known, there must also be unknown exporters. Who are they? The only cat-
egory of unknown exporters that we know is the so-called new shipments, that
is, producers who were not exporting during the period of investigation (Art 9.5
of the AD). They are presumed guilty, and, as a consequence, will pay a duty
following a fast-track review. New exporters (new shipments) originating in the
same country will pay an undefined duty until their review starts (it should start
promptly), no duty while being reviewed, and the individually calculated dump-
ing margin eventually (Art 9.5 of the AD). Now do new shipments exhaust the
realm of unknown exporters? No, says practice.
The issue was in the background of the litigation on Mexico—Antidumping
Measures on Beef and Rice. There the Mexican authority calculated dumping
margins for two identified exporters, for two more that voluntarily presented
themselves to the authority and imposed a default duty on all other US producer
of beef and rice. The US disagreed with the level of duty that remaining produc-
ers had to pay,⁵⁰ and challenged the Mexican practice. The panel found a decent
way to minimize the problem: it felt that it was incumbent upon the investigat-
ing authority to look for information and identify exporters. The AB disagreed
(§§ 250ff ). In its view, the Mexican authority was under no duty to investigate
any further. In its reading, hence, known are those exporters identified by the
petitioner and those that have voluntarily presented themselves to the authority.
But what about those exporters who do not belong to one of the two classes men-
tioned above, and are not new shipments either? The AB did not squarely address
this issue. One (probably extreme) reading between the lines of the AB Report, is
that such exporters pay no duties since an investigating authority cannot use best
information available against them.⁵¹ One other, less dramatic reading, suggests
that authorities cannot use information that has particularly adverse conse-
quences for exporters. As things stand, we simply do not know what the duty is
that should be paid by unknown exporters. A legislative intervention to clarify
things seems quite appropriate here.⁵²
The AD Agreement contains specific provisions which allow an investigat-
ing authority, under specific circumstances, to investigate only a sample of all
known exporters, and impose duties on the non-investigated known exporters of

⁵⁰ Actually, the AD Agreement does not prescribe what duty unknown exporters should be pay-
ing, although paying the residual duty seems like a reasonable option. In US practice, they will pay
the residual duty. In EC practice, unknown exporters pay the highest and not the residual duty. So
far, no panel has had to deal with the consistency of the EC practice. For discussion of state practice
on this score, see Mavroidis et al (2007).
⁵¹ As the Mexican authority did, and was found in violation of its obligations.
⁵² Both the exporters and the investigating authority can, in theory, act in bad faith: a produ-
cer who dumps more than the average of its compatriots, could hide hoping to get (at least in the
United States) the residual duty only. Bad faith behaviour is unlikely when it is the EC authorities
investigating, though. An authority, on the other hand, might choose to investigate only those who
dump a lot and thus raise the level of the residual duty. As we will see in what immediately follows,
there is no obligation to investigate a statistically viable sample.
 Antidumping (Art VI of the GATT) 349

the same country of origin along with those investigated (Art 6.10 of the AD).
The next question is what level of duty they should be paying. Assuming produc-
ers have been sampled,⁵³ the AD Agreement suggests that only two rates of AD
duties are permissible:
(1) an individually calculated duty for all sampled exporters and all other export-
ers who have come forward and requested to be investigated;
(2) the weighted average of the individually calculated duties will be applied
to all other sources (originating in the same country), the so-called residual
duty.

5.3 Establishing injury

Establishing dumping is only the first step towards the imposition of AD duties.
An investigating authority will also have to establish injury (of the domestic
industry producing the like product, Art 3.4 of the AD), and also show that dump-
ing actually caused injury (Art 3.5 of the AD). Injury will result from increased
(dumped) imports. Not every change, however, in the volume of dumped imports
will do. According to Art 5.8 of the AD, negligible imports, that is, imports repre-
senting less than 3% of imports of the like product in the investigating member,
even if dumped, cannot lead to a lawful imposition of AD duties.⁵⁴
AD, Art 3.4 comprises a list of factors that show injury, such as loss of market
share, sales, etc. All factors reflected in Art 3.4 of the AD must be examined.
Investigating authorities must, however, also examine all other known factors:
factors raised by an interested party qualify as known factors under Art 3.5 of the
AD (AB report on EC— Pipe Fittings (§ 178)).
The law imposes a dual obligation on WTO members: their investigating
authorities must ensure that:
(1) injury is attributed to dumped imports; and
(2) that injury is not attributed to factors other than dumped imports.
The requirement under (2) is often referred to as non-attribution. Under this
requirement, an investigating authority will be required to distinguish the effects
of dumped imports, from the effects of any other relevant factors, on the domestic

⁵³ Sampling can lawfully take place if the number of exporters is too large. In this case, the
investigating authority can either investigate a statistically viable sample, or the largest number of
exporters that it can reasonably investigate. The second alternative will have to control for the
administrative capacity of any given investigating authority.
⁵⁴ AD, Art 5.8 allows an investigating authority to cumulate imports from different sources,
and lawfully establish injury, assuming they pass the statutory threshold (imports for example
from three countries must represent at least 7% of the imports of the like product). The AB, in its
report on E—Pipe Fittings upheld the panel’s opinion that cumulation does not have to include a
country-specific analysis of injurious effects (§§ 107 and 117).
350 State Contingencies

industry producing the like product. Satisfying these two requirements amounts
to a demonstration of causality: it is dumped imports that caused injury.⁵⁵
The agreement does not contain a definition of injury: it does state, however,
in Art 3.2 of the AD that the effect of dumped imports must be significant price
undercutting, or significant price depression. WTO adjudicating bodies have
not interpreted these terms in any meaningful manner so far: they have limited
themselves to banal statements such as: significant does not mean insignificant.
Moreover, the agreement does not state how much of the injury should be attrib-
uted to dumping. Does it suffice, for example, that even a small part of the injury
is caused by dumping? And how small is small? WTO adjudicating bodies have
not come up with anything meaningful on this score either. This is less of an
issue in the context of the AD Agreement than it is for the purposes of causation-
analysis under the SG Agreement: causation-analysis is a mere procedural require-
ment in the context of the AD Agreement; following a demonstration of injury,
a WTO member will be imposing duties up to the amount of the dumping mar-
gin. It is more of an issue in the SG Agreement, since there, WTO members
can legally impose safeguards only to counteract the part of injury caused by
increased imports.
Next comes the question of the methodology to use in order to separate the
effects from various factors. An investigating authority is facing a situation where
more than one factors are simultaneously (potentially) influencing one out-
come.⁵⁶ A correlation between two variables simply indicates that these variables
tend to move together. It does not say anything about whether one variable has
a causal influence on the other.⁵⁷ Indeed, the fact that they move together may
be the result of the fact that a change in the first variable led to a change in the
second variable and vice-versa, or that their move is simply the result of a third
factor (multi-collinearity). Statisticians, thus, widely caution, and the AB itself
has on many an occasion accepted that correlation does not imply causality.
Classic legal anlysis will ask the question whether dumped imports are a suf-
ficient cause (causa adequata), or whether absent dumped imports there would
be no injury (conditio sine qua non). The former test involves some value judg-
ment; the latter is reminiscent of the but for approach, used in economics. A small
detour to economics is thus warranted in this respect.

⁵⁵ The causation-analysis is the same across the AD, SCM, and SG Agreements, so there is no
need to reproduce this discussion infra, where we examine the SCM and SG Agreements. The SG
Agreement, nevertheless, differs since safeguards can address only the injury caused by increased
imports, and not injury caused by other factors. Conversely, AD and CVD duties will be imposed
up to the level of the dumping margin and the subsidy paid respectively. Thus, in theory at least,
it could be the case that through AD and/or CVD duties a WTO member goes beyond what is
strictly necessary in order to address the injury caused by dumping and/or subsidization.
⁵⁶ This is a rather common story in many situations.
⁵⁷ As Prusa notes in various of his writings, however, correlations, when in the hands of an
investigating authority, often prove to be irresistible.
 Antidumping (Art VI of the GATT) 351

Economists would proceed to establish causality, either through direct estima-

tion, or through establishing the counterfactual (but for approach).⁵⁸ The most
common method used for direct estimation purposes is regression analysis, where
one estimates the relationship between a dependent and a series of independent
(explanatory) variables. In contingent protection cases, economists will usually
use Granger causality, whereby they will test whether one variable consistently
precedes another. This method requires many observations of both the depend-
ent and the independent variables, otherwise it might not be statistically reliable.
It could thus be used in the context of AD, where, usually,⁵⁹ many transactions
are observed, but not in the context of subsidization.
As Sapir and Trachtman (2007) have noted, the intuition behind the counter-
factual approach (but for):
. . . is the close association which is being made between a cause of an event and a sine qua
non condition of its occurrence. Here a cause is the condition ‘but for’ which the effect
would not have occurred.⁶⁰
So, instead of asking the question did A cause B, we try to show that if A had
not occurred, B would not have occurred either. The key issue here is how to
construct the counterfactual. Economists will either use trend analysis, where it
is assumed that today’s world is a continuation of yesterday’s world, or use empir-
ical models. There are two categories of such models: econometric models, which
use statistical techniques to test their validity, and calibration models (simula-
tion) which take them for granted. Their differences, nevertheless, should not be
exaggerated since (some of) the assumptions of calibration models are based on
econometric estimations.⁶¹
Grossman (1986), for example, uses an econometric model, the multivariate
analysis, in order to analyze the true causes of injury in the US steel industry. His
purpose is to control for a third factor, ie to see whether the variables still move
together when the influence of other variables has been removed.⁶²
Although one would expect that some or all of these methods should be given
some prominence in the WTO,⁶³ the panel in its report on EC—Pipe Fittings,
held the view that the non-attribution requirement does not prescribe a particular
methodology that all investigating authorities must use to this effect (§ 7.366).⁶⁴

⁵⁸ For a very succinct explanation, see Sapir and Trachtman (2007).

⁵⁹ Its relevance in the context of targeted dumping, that is, in cases for example, of dumped sea-
sonal goods, might be questionable. In such cases, we might be observing one or very few export
transactions.
⁶⁰ See Sapir and Trachtman (2007, p 27, emphasis original).
⁶¹ See, on this score, Kelly (1988), and Irwin (2003).
⁶² See Greene (1993, pp 486–507) and for a practical application, see Grossman (1986). The
manner in which the model is constructed is the key, however.
⁶³ Some domestic investigating authorities use this method.
⁶⁴ The AB, as well, has in a number of cases acknowledged the difficulties associated with non-
attribution, but has not privileged the use of a particular method in order to show compliance with
352 State Contingencies

The AD Agreement adopts an injury to competitors standard (Art 3.4 of the

AD): what matters is whether the producer welfare (that is, the welfare of the
the domestic industry producing the like product)⁶⁵ has been negatively affected
because of dumping by foreign exporters.⁶⁶ Once injury has been shown to have
been, at least partially, caused by dumped imports, the whole injury analysis
becomes moot for the remaining part of the process: AD duties will be imposed
to counteract the dumping margin and not the resulting injury for the domestic
industry producing the like product. That is, demonstrating that injury has been
caused by dumped imports is a mere procedural obligation and nothing beyond
that. Assume for example, that the dumping margin established through inves-
tigation is 20%. Assume further that the resulting injury for domestic producers
is $10 million. Assume that AD duties of 5% suffice for the domestic industry
to be compensated (since they guarantee the lost market share). A WTO mem-
ber can lawfully impose the 20% mark-up; the lesser (in this case 5%) duty rule,
according to which a WTO member should impose a lower than the established
margins duties if such lower duties suffice to counteract the injury caused, is an
option, but not an obligation.⁶⁷
WTO members can, instead of imposing AD duties, request or accept price
undertakings from willing exporters (Art 8 of the AD). Through a price undertak-
ing, an exporter agrees to raise prices up to the level that no injury results for the
domestic industry of the like product. The maximum price increase is up to the
level of the dumping margin, a lesser duty rule being an option.⁶⁸

5.4 Sunset reviews

AD duties will remain in place, in principle, as long as necessary to counteract
injurious dumping. Five years after their imposition, a sunset review must take
place and duties should be removed, unless if, during the review, it is established
that removal of duties will lead to recurrence of dumping and injury (Art 11.3 of
the AD). During the 5 years, an administrative review can be requested ex officio

the non-attribution requirement. It is difficult, however, to imagine how causality can be shown
without recourse to econometric analysis.
⁶⁵ The term ‘like product’ is interpreted very narrowly in the AD Agreement: the use of a product
coming under the same tariff classification (and sometimes, even a sub-part of it) will be privileged
and, in its absence, a product competing with the imported dumped product.
⁶⁶ See the excellent overview in Human et al (2003). On EC AD practice, see Vermulst (1996).
⁶⁷ The possibility for a lesser duty rule (that is, a duty which would be less than the full dump-
ing margin, but enough to counteract the resulting injury, exists in the AD Agreement (Art 9.1 of
the AD). It is, however, a best endeavours clause, although a matter of legal compulsion in some
jurisdictions (European Community), see Vermulst and Waer (1997). A proposal has been tabled
in the Doha round in favour of mandatory application of the lesser duty rule, see WTO Doc TN/
RL/GEN/99 of 3 March 2006.
⁶⁸ Vermulst and Waer (1997) shows, that this practice is actively pursued by the EC authority.
In practice, this works like a voluntary export restraint.
 Antidumping (Art VI of the GATT) 353

or upon request.⁶⁹ The administrative review can deal with the question whether
the continued imposition of duties is necessary to offset dumping, whether injury
would be likely to recur if duties were removed, or both (Art 11.2 of the AD).
The AD Agreement does not impose an obligation to start ex officio sunset
reviews on a specific date. As a result, WTO members retain some discretion on
this score. Since the imposed duties will remain in place while the review is taking
place, there is a risk that WTO members might keep duties in place by starting a
review at as late a stage as possible. This risk is somewhat addressed through the
discipline enshrined in Art 11.4 of the AD which requests that a review should
normally be completed within 12 months.⁷⁰ ⁷¹
The 5-year period (for sunset reviews) counts:
(1) either from the date of the original imposition; or
(2) from the date of the most recent administrative review under Art 11.2 of the
AD, if the review at hand covered both dumping and injury; or
(3) from the date of the most recent sunset review.
Hence, AD duties can, in theory, remain in place for a period longer than 5 years.
The time-frame involved in this context depends on the outcome of successive
sunset reviews. Since duties can only stay in place after the 5-year period follow-
ing a review, it is inferred that absent such a review any AD duties imposed will
have to be eliminated.
Cadot et al (2006) examine data from 1979–2005, and conclude that, since the
advent of sunset clauses, the majority of AD duties imposed do not stay in place
beyond the 5-year period. The authors, nevertheless, conclude that one should
not readily attribute this success story to the WTO: the agreement itself may be
endogenous to autonomous political forces in member countries, precisely those
forces that made the agreement possible. They observe that the 5-year deadline
did not change much with respect to EC practice: duties are rarely imposed for
more than 5 years running whereas, when the European Community imposes

⁶⁹ An administrative review can focus on dumping, or injury, or both.

⁷⁰ The term ‘normally’ suggests that a certain degree of flexibility is very much on the cards.
⁷¹ The US laws provide for an automatic initiation of sunset reviews. In other words, it is never
the case that duties will lapse absent sunset review in the United States, since a review will always be
initiated. The AB, in its report on US—Carbon Steel, faced a complaint regarding review of coun-
tervailing duties. It described the US law in the following manner (§ 101):
Section 751(c)(2) of the Tariff Act directs USDOC to publish a notice of initiation
of a sunset review no later than 30 days before, inter alia, the fifth anniversary of
the date of publication of a countervailing duty order. Section 351.218(b) of Title
19 of the Regulations confirms that USDOC will conduct a sunset review of each
countervailing duty order. Both the Sunset Policy Bulletin and the SAA describe the
initiation of sunset reviews by USDOC as “automatic.” (original emphasis)
The AB held the view that this law was not inconsistent with the requirements of the SCM
Agreement (§ 118). Confirmation that this interpretation is good law in the AD context as well,
came with the panel report on US—Corrosion-Resistant Steel Sunset Review.
354 State Contingencies

price undertakings, they run out after 5 years anyway. The United States is,
according to the results in Cadot et al (2006), the only outlier: it continues to
keep most duties in place after reviewing them at the sunset stage.⁷² Bloningen
and Park (2004) show why reverse causation might also explain why some duties
stay in place in the United States after the review stage: under incomplete pass-
through, the exporter will have to reduce the normal value in order to absorb
part of the imposed AD duties. Under the US system of annual reviews, this
practice will raise the dumping margin measured by the DOC and will lead to
an upward revision of duties. On the opposite side of the spectrum, Vermulst and
Waer (1997, pp 133–8), discussing the EC practice, show that in case the normal
value is being constructed, an exporter will have to raise its duties by more than
double the dumping margin in order to obtain a refund.
In many cases, duties stay in place following a sketchy review, usually sup-
ported by data used already in the initial investigation. The same data can logic-
ally lead to divergent results at the initial and the review stage: increased imports
during the original investigation could lead to a finding of injurious dumping;
increased imports during the review stage could lead to the conclusion that
dumping was not the cause of injury. Investigating authorities nevertheless, can
hide behind the permissive standard of review (Art 17.6 of the AD) applied in
AD cases, and not entertain all possible explanations why, for example, imports
continue to increase at the review stage.⁷³ The basic problem is that the WTO
adjudicating bodies have not so far to come up with a meaningful test that should
be applied in review cases. Clearly, the test should be forward-looking. It would
further be unreasonable to demand a finding of certainty from the investigating
authority that dumping and injury will recur absent duties. It would be natural,
however, to request from investigating authorities to look into, for example, the
contractual relationship between the domestic industry and its customers and
other relevant factors. It would further be natural to request from investigating
authorities to show the counter-factual, that is, what would have happened had
duties been removed. In this vein, Howse and Staiger (2006) have advanced some
preliminary ideas that would help WTO adjudicating bodies to construct a test.
Their test would request that investigating authorities respond to two questions:
(1) what conditions led to dumping in the first place? And
(2) whether those conditions have changed so as to remove the original reason
for dumping?
Absent a response to question (1), it is simply impossible to evaluate whether a
removal of duties will lead to dumping and injury anew. Dumping, for example,
could occur because of excess capacity in the business-cycle sense of the term, or

⁷² Thus, Cadot et al (2006, p 6) do not question the results obtained in Moore (2004) on this
score.
⁷³ See on this score Moore (2002), and Prusa (2005).
 Antidumping (Art VI of the GATT) 355

because of exchange rate fluctuations, or for many other reasons. Unless we know
what caused dumping, it is impossible to know whether dumping will vanish or
recur in case the duties in place are revoked. Then, an investigating authority
would also need to have an idea about the conditions of competition in the par-
ticular industry: is there perfect, imperfect competition, etc? How many com-
panies operate in the same relevant product market? Absent such knowledge, it
will be difficult to predict the pricing behaviour of the companies concerned in
case duties are revoked.
Unfortunately, we are still very far from seeing a test along these lines. As
things stand, absence of certainty has been more or less equated to carte blanche
to keep duties in place.⁷⁴
Sunset reviews are performed at the end of the 5-year period, and some-
times extend much beyond the 5-year period: for example, in the European
Community, it is quite common that a sunset review extends 15 months beyond
the 5-year period.⁷⁵ Case-law has not had to deal with the issue whether AD
duties perceived during these 15 months should be reimbursed (although good
arguments can support the thesis that reimbursement should indeed be the case).
The question whether standards applied during the original investigation are
relevant at the review stage has also been discussed in case-law. Panels and the AB
have repeatedly stated that, since reviews and original investigation are distinct
processes with different purposes,⁷⁶ the disciplines applicable to original investi-
gations cannot be automatically imported into review processes.⁷⁷
Based on this premise, the panel on US—Corrosion-Resistant Steel Sunset
Review considered that the de minimis thresholds applicable during the original
investigation in the absence of explicit language or cross-referencing to this effect
are not applicable in the context of a review.⁷⁸ The AB had reached a similar

⁷⁴ For some egregious cases, see Mavroidis et al (2007). Moore (2002) points to a large number
of cases (47% of the total between 1980 and 2000) where the US DOC relied on data that had been
used in the original investigation. Dordi (2006) on the other hand, refers to some promising EC
cases, where duties have been imposed for 2 years (Council Reg (EC) 1472/2006 of 5 October 2006
imposing a definitive anti-dumping duty and collecting definitely the provisional duty imposed on
imports of certain footwear with uppers of leather originating in the People’s Republic of China
and Vietnam [2006] OJ L275/1), and cases where, following a sunset review, duties have been
re-imposed for 2 years only (Council Reg (EC) 1583/2006 of 23 October 2006 imposing a defini-
tive anti-dumping duty on imports of ethanolamines originating in the United States of America
[2006] OJ L294/2).
⁷⁵ See p 11 in Commission of the European Communities, Global Europe, Europe’s trade defense
instruments in a changing global economy, A Green Paper for public consultation, Brussels, 6 December
2006, COM (2006) Final.
⁷⁶ AB report, US—Corrosion-Resistant Steel Sunset Review, paras 106–7; AB report,
US—Carbon Steel, para 87.
⁷⁷ AB report on US–Oil Country Tubular Goods Sunset Reviews, para 359.
⁷⁸ Panel report, US—Corrosion-Resistant Steel Sunset Review, §§ 7.70–7.71. The panel thus
concluded as follows:
On the basis of this textual analysis of the relevant provisions of the Anti-dumping
Agreement, we conclude that the 2 per cent de minimis standard of Article 5.8 does
not apply in the context of sunset reviews. In this context, we again observe that, in
356 State Contingencies

conclusion in the countervailing duty context in the US—Carbon Steel case.⁷⁹

The AB, in its report on US—OCTG Sunset Reviews, continued along the same
line, and held that no injury examination, in the sense of Art 3 of the AD, is
required in a sunset review. It thus upheld the panel’s finding that the obligations
set out in Art 3 of the AD do not apply to likelihood of injury determinations in
sunset reviews (§ 280). The AB added, however, that an investigating authority
may, without being obliged to do so, ‘borrow’ from its analysis under Art 3 of
the AD (the original investigation), when conducting its review analysis (§ 284).
In a similar vein, the AB expressed the view that Art 11.3 of the AD requires
an authority to make a determination concerning likelihood of dumping and
injury, but not of a causal link between the two. In US—AD Measures on OCTG,
the AB first confirmed that a causal link between dumping and injury to the
domestic industry is fundamental to the imposition and maintenance of an AD
duty under the AD Agreement. It added nevertheless that, because the review
contemplated in Art 11.3 is a distinct process different from the original
investigation purpose, it was of the view that a causal link between dumping and
injury is not required to be established anew in a review conducted under Art
11.3 of the AD (§§ 117–18 and 123–4). The AB underlined that the nexus to be
demonstrated under Art 11.3 of the AD is ‘between the expiry of the duty’ on
the one hand, and the likelihood of ‘continuation or recurrence of dumping and
injury’, on the other.
The difference between nexus and causal link is nowhere explained and is hard
to understand as well. It seems that the AB is playing with words here.
Next, came the issue whether cumulation was permissible at the review stage.
During the original investigation, no dumping duties can be lawfully imposed,
unless dumped imports from a particular source (country) are over 3% of the
total imports in the market of the investigating member. A WTO member can,
however, cumulate imports and, assuming, imports from three different sources
constitute more than 7% of the total imports in its market, it can still go ahead
and impose duties, assuming of course, that such imports have caused injury
(Art 5.8 of the AD).⁸⁰ The panel on US—OCTG Sunset Reviews faced, inter alia,
a claim that, in the absence of specific language to this effect, cumulation was
not permissible in the context of sunset reviews; consequently, the United States,
by cumulating imports from various sources, was acting inconsistently with its

light of the qualitative differences between sunset reviews and investigations, it is

unsurprising that the obligations applying to these two distinct processes are not
identical. (§ 7.85)
⁷⁹ AB Report, US—Carbon Steel, §§ 81–4.
⁸⁰ Hansen and Prusa (1996) have shown that, assuming constant market share, the probabil-
ity of a positive dumping finding is higher the larger the number of exporters. Hence, investigat-
ing authorities might have an incentive to cumulate. This is what they termed super-additivity.
Gupta and Panagariya (2006) provided a theoretical explanation for super-additivity: the larger
the number of exporters involved, the higher the risk of free-riding (producers will free ride on
other producers’ recourse to legal experts, etc).
 Antidumping (Art VI of the GATT) 357

obligations under the AD Agreement. The panel, based on textual and context-
ual arguments, held the opinion that various provisions in the AD Agreement
make it clear that cumulation is permissible throughout the investigation and
the review processes, but that the standards regarding cumulation during the
original investigation reflected in Art 3.3 of the AD were not applicable in
the context of reviews.⁸¹ The AB confirmed this view,⁸² rejecting the suggestion
that this would imply a carte blanche for investigating authorities when cumu-
lating (§ 302).
In US —AD Measures on OCTG, the AB, on the one hand, confirmed its view
that Art 3.3 of the AD does not apply to sunset reviews, but emphasised, on the
other hand, that this does not imply that it is never necessary for an authority to
determine whether such a cumulative assessment is appropriate in light of the
conditions of competition in the market place. In other words, while there does
not exist a legal requirement to do so, contrary to what is the case in an original
investigation because of Art 3.3 of the AD, the specific facts of the case may
nevertheless require such an analysis of the appropriateness of cumulation (§ 171).
With respect to the calculation of dumping duties at the review stage, the obliga-
tions imposed on an investigating authority are also less stringent than the cor-
responding obligations during the original investigation. The panel, in its report
on US—Corrosion-Resistant Steel Sunset Review, held that, during the review,
an investigating authority need not calculate in a precise manner the dumping
margins which will result in case it removes the duties in place. Rather, because
uncertainty is inherent in any forward-looking study, some reasonableness stand-
ard is most warranted in this context, and consequently, investigating authorities
should not be requested to make a determination of dumping in the sense of Art 2
of the AD, or provide a precise amount of dumping margins. This does not mean,
according to the panel, that evidence of dumping may not be relevant for a likeli-
hood of recurrence or continuation of dumping determination (§§ 7.162–7.180).
On appeal, the AB confirmed this view (§§ 123–4). However, the AB did make
one important clarification, and overturned the panel’s ruling in this respect:
it stated that, in case a WTO member goes ahead and does calculate dumping

⁸¹ Panel report, US—OCTG Sunset Reviews , §§ 7.323–7. 336. The panel was of the view that,
‘the Agreement generally allows the use of cumulation and that Article 3.3 is not an authorization
for cumulation. Rather, it sets out the conditions that must be fulfi lled when cumulation is used in
investigations.’
⁸² AB report, US—OCTG Sunset Reviews , §§ 300–1:
‘Given the express intention of Members to permit cumulation in injury determin-
ations in original investigations, and given the rationale behind cumulation in
injury determinations, we do not read the Anti-Dumping Agreement as prohibiting
cumulation in sunset reviews.
Turning to Argentina’s argument that the prerequisites specified in Article 3.3(a)
and (b) should be satisfied by investigating authorities when performing cumulative
analyses in sunset reviews, we note that Argentina offers no textual support for its
claim. Indeed, as we observed above, the opening text of Article 3.3 plainly limits its
applicability to original investigations’ (footnote omitted).
358 State Contingencies

margins—although no such requirement exists in the AD Agreement—it should

do so only in accordance with Art 2 of the AD (§§ 127–8).
A closely connected issue occupied the panel in its report on US—Corrosion-
Resistant Steel Sunset Review: the question was asked, inter alia, whether an inves-
tigating authority would be required, by analogy with Art 6.10 of the AD, to
calculate an individual margin of dumping for each exporter or producer exam-
ined, and to make a determination of likelihood of recurrence or continuation of
dumping and injury for each exporter or producer under review. The panel con-
sidered that no such ‘company-specific’ likelihood determination is required, and
a determination could thus be made on an ‘order-wide’ basis (§§ 7.207–7.208).
The AB confirmed this view. It acknowledged that Art 11.4 of the AD contains
an explicit cross-reference to Art 6 of the AD, regarding evidence, and proced-
ure. Consequently, in its view, these rules were made applicable to review situ-
ations. However, Art. 6.10 AD, requiring the authority to calculate individual
margins of dumping, cannot apply in a review because, according to the AB, in a
review, an authority is not required under Art 11.3 of the AD to calculate dump-
ing margins in the first place. Hence, the requirement to make an individual
company-specific determination as set forth in Art 6.10 of the AD cannot apply
in a review situation (§ 155).
A more stringent sunset review is being requested by some exporting states
during the ongoing Doha round. It should be noted, nevertheless, that some
exporters might find it profitable to keep duties in place. Bown and Wauters
(2007) contemplate such a scenario where the most efficient exporter is hit with
the highest dumping margin. In this case, at least some trade will be diverted to
less efficient dumpers who have been hit by a lower margin. Removal of duties
will tilt the balance once again in favour of the most efficient producer. In such
cases, collusion between some exporters and the investigating authority is not
unthinkable.

5.5 Price differentiation in antitrust statutes

Dumping is a price discrimination scheme. It can, of course, have positive wel-
fare implications for consumers abroad, and indeed it does, in the overwhelming
majority of tested cases.⁸³ AD, however, proceeds without enquiring at all into
the overall welfare implications of dumping; injury to domestic producers com-
peting with dumped imports suffices for duties to be lawfully imposed. This is
what the injury to competitors test amounts to.⁸⁴
Before we move to discuss the basic economics of AD, it is probably war-
ranted to make a short detour into representative antitrust statutes to see how

⁸³ See Finger (1993), and Messerlin (2001).

⁸⁴ See on this issue, Bown et al (2004), Prusa (2005 and 2001), Petersmann (1993), and
Hoekman and Mavroidis (1996).
 Antidumping (Art VI of the GATT) 359

they deal with price discrimination.⁸⁵ In substance, AD is an exception to Art III

of the GATT: a different statute is applied when price discrimination is domes-
tic (antitrust), than that applied when the price discrimination is international
(AD). Technically, however, since AD is a border instrument and NT covers only
domestic instruments, it would be difficult to sustain such an argument before
a WTO adjudicating body. Still, trade liberalization is meant to be an exten-
sion of the relevant geographic market (in the antitrust sense of the term): by
removing tariff barriers and NTBS, conditions of competition become (more)
homogeneous across national jurisdictions. It is with this in mind that we turn to
antitrust practice on the two sides of the Atlantic.
In antitrust practice, which is not necessarily by definition the best reflection
of cutting-edge economics⁸⁶ but has been substantially more open to economic
analysis, a price differential is not treated in the same way it is treated in the AD
Agreement. Of course, there is nothing like a world antitrust statute and hence,
ipso facto, no harmonized practice either. A look into two quite well-known sys-
tems, the European Community and the United States, is appropriate.
The ECJ had the opportunity to pronounce on this issue in its Akzo and
Tetrapak case-law.⁸⁷ In the EC system, price differentiation will be treated as an
abuse of (single or collective) dominance. As a result, for price differentiation
to enter the picture of antitrust enforcement, a prior finding of dominance is
required. We quote from the Tetrapak decision (§ 41), which is the most recent
one and can be considered the authentic expression of the test for predatory pri-
cing established by the ECJ:
In AKZO this Court did indeed sanction the existence of two different methods of ana-
lysis for determining whether an undertaking has practiced predatory pricing. First,
prices below average variable costs must always be considered abusive. In such a case,
there is no conceivable economic purpose other than the elimination of a competitor,
since each item produced and sold entails a loss for the undertaking. Secondly, prices
below average total costs but above average variable costs are only to be considered abu-
sive if an intention to eliminate can be shown.
With respect to some of the sales at hand, the ECJ found that the prices charged
were between average variable cost and average total cost. According to its own test,
therefore, it would have to show intent to eliminate the competition. The ECJ in
§ 44 ruled as follows on this point:
. . . it would not be appropriate, in the circumstances of the present case, to require in
addition proof that Tetra Pak had a realistic chance of recouping its losses. It must be

⁸⁵ See Kovacic (2006) for an excellent discussion of the manner in which price discrimination is
treated in antitrust and trade statutes.
⁸⁶ See Hovenkamp (2005, pp 12ff ).
⁸⁷ See Case C-62/86 Akzo Chemie BV v Commission [1991] ECR I-3359 (3 July 1991); Case
T-83/91 Tetrapak International SA v Commission [1995] ECR II-762 (CFI) and Case C-333/94P
[1996] ECR I-5951.
360 State Contingencies
possible to penalize predatory pricing whenever there is a risk that competitors will be
eliminated. . . . The aim pursued, which is to maintain undistorted competition, rules out
waiting until such a strategy leads to the actual elimination of competitors.
A dominant company that prices below a certain threshold is thus found to be
practising predatory prices, in EC practice. There is no need to further show that
the company at hand would realistically recoup its original investment. In the
eyes of the court, a company behaving in this manner can only intend to monop-
olize the market, and it is predatory intent that is ultimately being punished.⁸⁸
The US Supreme Court also had the opportunity to pronounce on this issue.
In the US statute, monopolization plays a comparable role to abuse of dominance
in the EC regime, and it is under its aegis that the Supreme Court discussed this
issue. The leading case is Brooke Group:⁸⁹
A plaintiff must prove (1) that the prices complained of are below an appropriate measure
of its rival’s costs and (2) that the competitor had a reasonable prospect of recouping its
investment in below cost prices. . . . The plaintiff must demonstrate that there is a likeli-
hood that the scheme alleged would cause a rise in prices above the competitive level suffi-
cient to compensate for the amounts expended on the predation, included the time value
of the money invested in it. Evidence of below cost pricing is not alone sufficient to per-
mit an inference of probable recoupment and injury to competition. The determination
requires an estimate of the alleged predation’s cost and a close analysis of both the scheme
alleged and the relevant market’s structure and conditions. Although not easy to estab-
lish, these prerequisites are essential components of real market injury. . . . Predatory pri-
cing schemes, in general, are implausible . . . and even more improbable when they require
coordinated action among several firms . . . They are least likely to occur where . . . the
cooperation among firms is tacit, since effective tacit coordination is difficult to achieve;
since there is a high likelihood that any attempt by one oligopolist to discipline a rival
by cutting prices will produce an outbreak of competition; and since a predator’s present
losses fall on it alone, while the latter supracompetitive profits must be shared with every
other oligopolist in proportion to its market share, including the intended victim.
As can be seen, the evidentiary standard in US law is higher and makes the pos-
sibility of successfully challenging predatory schemes quite unlikely. Still, both
regimes punish only a sub-set of price discrimination: predatory pricing: assum-
ing that such a scheme has been successfully implemented, the predator will be
in a position to recoup the original investment, having driven the competitors
out of the market and having ensured that there is no risk, at least for some time,
that they re-enter the market once the predator raises its prices. In other words,
what antitrust statutes punish is behaviour which causes injury to competition, as
opposed to injury to competitors.

⁸⁸ At the time of writing, the EC Commission had already issued the Discussion Paper on
the application of Art 82 of the Treaty to exclusionary abuses (EC Commission, DG Competition,
December 2005), and it had organized a public hearing to entertain reactions. It had not, however,
published its Guidelines which, presumably, will modify the current regime.
⁸⁹ Brooke Group Ltd v Brown & Williamson Tobacco Corp, 509 US 209 (1993).
 Antidumping (Art VI of the GATT) 361

The US test is more reasonable since it avoids punishing unsuccessful preda-

tory pricing: in such cases, consumers will profit from lower prices, as the preda-
tor will not be in a position to recoup the original investment. Rational economic
behaviour by governments would suggest that, before imposition of duties, the
welfare implications of dumping on the whole of the society, and not simply a
sub-set of it (the injured producers), should have been undertaken. This is what
the two antitrust statutes aim to do, and this is where the AD Agreement fails.

5.6 What does the ‘anti’ in AD refer to?

The regulation of AD has emerged as probably the most contentious issue among
policy-makers and trade economists. Policy-makers advance four categories of
arguments, that is, four concerns that AD aims to address:⁹⁰
(1) discriminatory pricing is wrong;
(2) AD addresses predatory prices;
(3) strategic dumping can reduce the incentives for participating in the WTO;
(4) AD is a necessary safeguard.

5.6.1 Discriminatory pricing

There is an argument that discriminatory pricing is unfair. Price discrimination
will occur when price arbitrage is impossible. This could be due to various reasons.
The absence of arbitrage may flow from intrinsic factors outside the control of the
exporting firm, (such as different technologies in the two markets; designs seg-
menting markets for more economically sound reasons—the home country con-
sumers could be much more sensitive to the latest technologies than the foreign
country consumers; or other technical barriers to trade) or might be generated
from the firm itself in its effort to protect the home market. Whatever the rea-
son, the fact remains that dumping occurs because there is no price arbitrage. If
discriminatory pricing is the true concern of the proponents of AD duties, why
not address discriminatory pricing as such? Why have recourse to a second best
instrument such as AD (which leads to increasing prices to the level of the higher
price), and not try to move prices to the level of the lower price? Do we really need
to burn the village to save it?
More practically, the first question to ask would be why importers cannot sim-
ply re-export the dumped products to the exporting country charging the higher
price? Is it because of (state) trade barriers, product differentiation, or any other
reason? The former could be decomposed into tariffs, QRs, export subsidies, and

⁹⁰ This section borrows from Mavroidis et al (2007). The welfare implications of AD are well
documented in economics literature and there is no need to reinvent the wheel here. The reader
could, inter alia, check Boltuck and Litan (1991), Finger (1993), various contributions by Messerlin
culminating in Messerlin (2001), and Prusa (2001 and 2005).
362 State Contingencies

domestic instruments: QRs and export subsidies are illegal per se, and domestic
instruments have to be applied in a non-discriminatory manner. Hence, the only
remaining issue would be whether the reason for price arbitrage relates to import
tariffs. Assuming this is indeed the case, it would be more sensible to try to reduce
the level of tariffs in the context of multilateral rounds, rather than increase
the prices by adding the dumping margin to the bound tariff. Cases of product
differentiation can also be discarded, since, assuming this is indeed the case, AD
duties cannot be imposed: the investigating authority will not be in a position
to overcome the like product requirement.⁹¹ Price arbitrage can occur because of
legitimate prohibition of parallel imports: the TRIPs Agreement allows for regu-
latory diversity in this field.⁹²
Inquiring into the reasons for lack of arbitrage, assuming one exists, will allow
trading partners to use the first best instruments to address distortions (assuming
they exist).⁹³ It will, in all likelihood, also lead to fewer AD investigations. This
is not to say, of course, that one needs to address any price discrimination. This is
simply to suggest that we do not know whether price discrimination is a genuine
concern for those proposing the regulation of AD. It most likely is not.

5.6.2 Predation
Complaining firms in AD cases often argue that the foreign exporters are pricing
at a level below their home price, or at a price lower than their cost of production
(but generally without specifying whether they talk about total, fi xed or vari-
able costs). They also suggest that such a pricing behaviour is driven by the will
of foreign firms to eliminate domestic competitors in their own market in order
to increase prices when they will ultimately be in a monopoly position in the
import-competing market. Does this argument make sense?
No, is the short answer. First, if this was indeed true, one could simply reduce
AD actions to cases of predatory pricing. Actually, for some jurisdictions at
least, this would mean that there would be no need to rely on an AD Agreement
drafted to this effect, since their own competition laws would lead to a compar-
able outcome.⁹⁴

⁹¹ Admittedly, things can be much more complicated. But developing this line of thought
for argument’s sake will allow us to understand how much of a concern price discrimination
really is. Like product defi nitions are, in the overwhelming majority of cases, quite narrow in AD
proceedings.
⁹² See Li and Maskus (2006) who also discuss the cost of such policies.
⁹³ A similar pattern is described by Goldberg and Verboven (2005) who discuss the cars’ price
differentiation across EC member states, and the manner in which the EC Commission addressed it.
⁹⁴ Messerlin (2001) asked the question what would have been the outcome of AD investiga-
tions, had the investigating authorities employed an antitrust-type standard. His analysis is specific
to the EU. The period of his investigation runs from 1980 to 1997 during which he examines 461
cases of initiation of investigation for which adequate information exists. To answer this question
he proposes a five screen test. The five screens include: first a screen to assess the capacity of the
foreign firms involved in AD cases to behave as predators. He proposes a screen which eliminates
from investigation all foreign companies with forecasted aggregate market share of 40% or less.
 Antidumping (Art VI of the GATT) 363

Second, as we saw, US antitrust practice has adjusted to the idea that preda-
tion is unlikely and, consequently, not much time and effort should be invested
in thinking about ways to address it. This is so because a predator’s pricing is a
strategy in two phases: eliminating competitors necessarily generates costs for
predator, and recouping these costs during the second phase, which almost neces-
sarily creates incentives for old competitors to come back or new competitors to
enter a market that they ignored before the price war. A predator should follow a
predatory pricing strategy only if the second phase brings higher profits than the
losses generated by the first phase. If phase 1 is long, the likelihood of recoup-
ing in phase 2 thereby becomes slimmer. At the end of the day only a prospect-
ive evaluation of the barriers to entry in the market where the predator operates
can provide responses as to the likelihood of recouping the original investment.
The unlikelihood of this occurrence is supported by an additional argument. If
the goal of predation is monopolization of a market, there is a better, less costly
manner to do it: merger. A predator can buy the other companies in the relevant
market where it operates, at a price which would reflect their stock value in the
expected monopolistic (not in a competitive) market.⁹⁵

5.6.3 Strategic dumping

Strategic dumping crucially relies on cost conditions: an exporting firm will
benefit from a closed home market (a safe harbour), where it can charge the full
(fi xed and variable) costs of production on the home consumers, allowing the sale
of the product in the export market at a price only inclusive of the variable cost
(its competitors in the export market are assumed not to operate in a safe har-
bour). Consequently, the competitors of the strategic dumper cannot emulate its
behaviour: they thus must be disadvantaged enough, in terms of the relative size
of accessible markets and scale economies, with respect to the strategic dumper.

He thus screens out 311 out of 461 cases. The second screen screens out 14 additional cases for
which a negative outcome (no injury) resulted from the investigation. The third screen requests
elimination of all cases where four or more countries were involved. The rationale for this screen
as explained by the author (p 358) is that ‘if four or more countries are involved in simultaneous
antidumping actions on the same product, the possibility of joint predatory behaviour seems rather
low’. Seventy-five more cases are thus screened out. The fourth screen eliminates all cases where
eight or more economic operators are involved (for the same reason as per screen three, the possibil-
ity for predatory behaviour in case eight or more companies are involved, is quite low). Seventeen
more cases are screened out. The remaining 44 cases (ie less than 10% of the original total) are sub-
jected to a test whereby the author examines the market share of the economic operators involved
and/or the Herfindahl-Hirschmann Index (HHI) and concludes (having eliminated cases where
companies have either small market share or in application of the HHI no reasonable competition
authority would have intervened) that only 12 cases were worth investigating, that is less than 2%
of the original total.
⁹⁵ A predator’s strategy can be further complicated, depending on the facts of the case: if for example,
the other companies are small compared to predator, they can try to convince the demand side of the
market (the consumers) to help them to survive to the first phase of predator’s dumping. They have a
good argument: it is that if predator wins, consumers will pay the eventual monopoly price.
364 State Contingencies

It follows that, when the safe harbour is small relative to the export market, it is
highly unlikely that strategic dumping will occur. In this vein, a high percent-
age of AD investigations (of goods originating in say Singapore, or Hong Kong,
China) cannot be justified on strategic dumping grounds.⁹⁶

5.6.4 AD as a safeguard mechanism

The argument has also been often made that AD introduces a flexibility elem-
ent in the WTO contract: trading partners will be induced to make more
commitments than otherwise, because they know that they can always subse-
quently increase their protection by the dumping margin. Recently, Finger and
Nogues (2006) provided some support for this argument by researching the
MERCOSUR⁹⁷ experience: in their account, national administrations in the
MERCOSUR countries used the potential for AD as a carrot to persuade domes-
tic lobbies to support trade liberalizing commitments.
This argument rests on the premise that the current safeguard clause does not
operate to the satisfaction of the trading partners. This is probably true.⁹⁸ But the
natural question to ask is, why do negotiating efforts not concentrate on fi xing the
current safeguard clause? Using AD as a safeguard is not a rational choice: filing
fees in AD largely exceed those for safeguards; the former takes place predomin-
antly upon request of the domestic industry, the latter is predominantly a matter

⁹⁶ Independently from AD measures, there is evidence that AD measures have a pro-cartel

effect at the two stages of the complaint and of the decision. Messerlin (2001) persuasively argues
that exchanging information for lodging a complaint requires a minimum exchange of informa-
tion from the complainant; even if handled through lawyers, plaintiffs could draw common con-
clusions from the complaint. On the other hand, nascent collusion between plaintiffs can be made
sustainable by AD measures. It is not by accident that, at least, in his count, one quarter of EU
AD cases of the 1980s were ‘twin’ AD and competition cases dealing with similar products and
EU firms. A more systematic indication of this pro-cartel dimension flows from the fact that the
combined market shares of the plaintiff s and defendants are often extremely high—on average,
around 80%–85% in the United States and the European Union. Defendants and complainants
have a combined market share of less than 90% in only 55% of the US and EC AD cases examined
in the 1980s and mid-1990s. Foreign cartels (that is, outside the country enforcing AD measures),
as well, have been created by AD cases, as best illustrated by the quasi-official Canadian potash car-
tel triggered by US AD. Another interesting observation that Messerlin (2001) reports is echoing in
AD investigations: it has been shown that EU and US AD actions against China targeted the same
Chinese exports in 75% of AD cases initiated against Chinese exports by the United States and in
68% of the cases initiated by the European Union. Most of these cases echoed each other within a
year or less, and all but three of these cases (cycles, hammers, and pocket lighters) resulted in AD
measures of some kind. Such a large proportion of echoing cases and the similarity of their out-
comes are signs that AD is a protectionist instrument that petitioners are using in a strategic way to
segment the two largest world markets. World cartels are likely to be fostered by a series of AD cases
echoing each other all over the world. Large firms almost simultaneously lodge AD complaints in
Brussels, Washington and elsewhere in order to segment the world market as they wish (AD duties
and measures increase the transaction costs, hence the firms’ ability to charge different prices
(discriminate) in different locations).
⁹⁷ MERCOSUR is the preferential trade agreement signed by Argentina, Brazil, Paraguay, and
Uruguay.
⁹⁸ See the discussion in section 7.
 Safeguards 365

of governmental discretion.⁹⁹ It is probably true that, in practice, AD has been

used as a surrogate for safeguard measures. This was not, however, its intended
use: indeed the AB has time and again insisted on the distinction between fair
and unfair trade, upon which, in its view, the differentiated legal requirements in
the two agreements are predicated. Consequently, it is not that it is irrational to
use AD as a safeguard, it is contra legem.

6 Countervailing Duties (CVD)

CVDs have a function similar to that of AD duties. The only exception is that
CVDs address government practice (subsidization), whereas AD duties address
private practice (dumping). Our discussion above is, consequently, directly rele-
vant for CVDs as well.
There is one aspect of CVDs regulation that we probably need to elabor-
ate on: sunset reviews. The test is the same. CVDs will be reviewed, unless the
investigating authority establishes that, in all likelihood, removal of CVDs will
lead to recurrence of subsidization that might lead to injury. The problem is
that such a forward looking analysis is probably feasible when private practices
such as dumping are being addressed: the analyst will examine the competi-
tive conditions in the product market, as discussed in section 5 above. It is less
obvious, however, to perform the same test in the context of CVDs: what kind
of information should an investigating authority possess in order to establish
that subsidies would recur if duties were removed? The investigating authority
of country A must know the future political importance of a lobby in country
B. This is probably asking too much, or being content with pure speculation:
the test is either too demanding or simply unworkable and, hence, adjudicating
bodies will adopt a pro-deference attitude in this respect. Practice so far, tends to
support the latter view.

7 Safeguards

7.1 The regulatory framework

GATT, Art XIX allows WTO members to take safeguard action, in order to
react to a sudden, unforeseen surge in imports, as a result of tariff concessions
concluded during a trade round. A multilateral agreement on safeguards (SG)
was concluded during the Uruguay round, which adds detail to Art XIX of

⁹⁹ Safeguards, as we will see, can also be imposed upon request. In this case, nevertheless, there
are no standing requirements: an informal petition by one operator suffices.
366 State Contingencies

the GATT.¹⁰⁰ Based on the General Interpretative Note regulating the legal rela-
tionship between the GATT and the other Annex 1A agreements, the AB, in its
report on Argentina —Footwear, held that WTO members wishing to impose
safeguards must, besides the conditions mentioned in Art 2 of the SG (import
surge which causes injury to the domestic industry producing the like product),
also satisfy the requirement included in Art XIX of the GATT: they must show
that the import surge is the result of unforeseen developments.¹⁰¹

7.2 The typology of safeguards

The SG Agreement only indirectly discusses the type of measures that WTO
members can use. A look into prevailing practice is thus, necessary. Based on the
notifications by WTO members to the WTO safeguards committee, it appears
that ad valorem tariff increases are the most widely used safeguards instrument.
Almost as popular are tariff quotas (TRQs) whereby members reserve a favour-
able tax rate for a small initial quantity imported. Specific tariff increases are third
on the ranking. Quantitative restrictions with some quota system allocation
are a distant fourth.¹⁰² In other words, although one often associates safeguard
measures with quantitative restrictions, a quantitative restriction clearly does not
exhaust the realm of possible safeguard measures that can be lawfully imposed
under the SG Agreement, quite the contrary in fact.

¹⁰⁰ For an excellent overview of the WTO Agreement, see Sykes (2006). Ostry (1997) correctly
underscores the crucial place that safeguards held in the original GATT contract.
¹⁰¹ This is an incorrect interpretation. The AB reading is based on Art 1 of the SG, which refers
to the measures that WTO members can use as safeguards. SG Art 2, entitled Conditions, does not
mention unforeseen developments among the requirements that must be met for safeguards to be
lawfully imposed. This was probably a conscious departure from the Art XIX test. The AB over-
looked this point. Moreover, the relevance of the GATT depends on the understanding of the term
‘conflict’ appearing in the General Interpretative Note regulating the legal relationship between the
GATT and the other Annex 1A agreements. By narrowing down its meaning to an either/or type of
situation, where the GATT and another Annex 1A agreement cannot simultaneously apply, the AB
opened wide the door to the reintroduction of unforeseen developments among the requirements for
lawful imposition of safeguards. Th is led to confusion and discontent, as the very comprehensive
study of Sykes (2003c) elegantly shows, but see contra Lee (2005). GATT, Art XIX also requests
that increased imports are the result of obligations assumed under the GATT, a condition that has
not been explicitly included in the SG Agreement. Th is is, however, a rather innocuous addition:
assuming duties are unbound, a WTO member could simply raise its duties rather than go through
the investigation process that might lead to the imposition of safeguards. Crowley (2006) finds that
there is indeed a correlation between tariff cuts and recourse to safeguards. In her model, she shows
that (for a sample of 41 countries) a tariff cut by 10% would lead to six times as many safeguard
actions, whereas a tariff cut of 5.5% would lead to 2.7 times as many safeguard actions. Indeed,
assuming substantial variance between applied and bound duties, a WTO member will probably
raise duties to the bound level, instead of initiating an investigation.
¹⁰² Based on the notifications until 7 November 2005, we counted 24 ad valorem tariffs, 21
TRQs, 18 specific tariff increases, and seven QR/quota measures. Two variable tariff increases
complete the picture.
 Safeguards 367

7.3 Safeguards imposed by PTAs

Individual WTO members may, according to Art 2 of the SG, impose safe-
guards. A CU may, according to footnote 1 to Art 2 of the SG, impose safe-
guard measures, either as a single unit, or on behalf of one of its members.¹⁰³
The AB distinguished between a measure imposed by the CU on behalf of one
of its members, and a measure imposed by a WTO member, which also hap-
pens to be a member of a, CU for itself.¹⁰⁴
A different issue is whether a member of a PTA should include or exclude
preferential imports when imposing safeguards. The AB applied an obli-
gation of parallelism in such cases: Argentina (member of MERCOSUR)
has had recourse to safeguards and its measures were challenged, because
it had included imports from its preferential partners in the examination
of increased imports and consequent serious injury. The AB held, in its
Argentina—Footwear (EC) report, that, in so doing, Argentina was not
allowed subsequently to exclude from the application of the measure its pref-
erential partners (§ 113):
On the basis of this reasoning, and on the facts of this case, we find that Argentina’s
investigation, which evaluated whether serious injury or the threat thereof was
caused by imports from all sources, could only lead to the imposition of safeguard
measures on imports from all sources. Therefore, we conclude that Argentina’s
investigation, in this case, cannot serve as a basis for excluding imports from other
MERCOSUR member States from the application of the safeguard measures. (ori-
ginal emphasis)
In US—Wheat Gluten, the AB confirmed that a necessary parallelism existed
between the imports examined and the imports covered by the measure. An
authority which examines imports from all sources is precluded from applying
the measure only to a sub-set of such imports examined. In other words, such an

¹⁰³ Up to the end of 2005, case-law evidences no case where a CU imposed safeguards on behalf
of one of its members. It has always been individual members of a CU imposing safeguards on
their own.
¹⁰⁴ AB report, Argentina—Footwear (EC), § 108:
Therefore, at the time the safeguard measures at issue in this case were imposed by
the Government of Argentina, these measures were not applied by MERCOSUR
“on behalf of ” Argentina, but rather, they were applied by Argentina. It is Argentina
that is a Member of the WTO for the purposes of Article 2 of the Agreement on
Safeguards, and it is Argentina that applied the safeguard measures after conducting
an investigation of products being imported into its territory and the effects of those
imports on its domestic industry. For these reasons, we do not believe that footnote
1 to Article 2.1 applies to the safeguard measures imposed by Argentina in this case.
As a result, we find that the Panel erred in assuming that footnote 1 applied, and we,
therefore, reverse the legal reasoning and findings of the Panel relating to footnote 1
to Article 2.1 of the Agreement on Safeguards.
368 State Contingencies

authority puts itself in any case off-side.¹⁰⁵ It is doubtful whether this case-law is
consistent with Art XXIV of the GATT.¹⁰⁶

7.4 Conditions for lawful application of safeguards

7.4.1 Initiation of investigation
There is nothing in the SG Agreement relating to the initiation phase. There is
also nothing which subjects the decision to initiate an investigation to certain
procedural or substantive conditions. So, unlike the AD or the SCM Agreements,
there is no distinction between self-initiated (ex officio), and upon request investi-
gations. This does not mean that an investigation cannot be requested by a pri-
vate party. The SG Agreement does not prejudge this issue. In fact, the legislative
requirements are the same irrespective of whether an investigation has been initi-
ated by a private party or ex officio. There are no standing requirements in the SG
Agreement either: it could be, for example, the case that one economic operator
representing a very minor proportion of the domestic industry requests the initi-
ation of an investigation; an investigation could be launched, assuming the inves-
tigating authority agrees. Nor are there any preliminary evidence requirements:
neither the private party, nor the investigating authority needs to show prelim-
inary evidence that increased imports, resulting from unforeseen developments,
are causing injury to the domestic industry.¹⁰⁷

¹⁰⁵ AB Report, US—Wheat Gluten, § 96:

The same phrase—“product . . . being imported”—appears in both these paragraphs
of Article 2. In view of the identity of the language in the two provisions, and in the
absence of any contrary indication in the context, we believe that it is appropriate
to ascribe the same meaning to this phrase in both Articles 2.1 and 2.2. To include
imports from all sources in the determination that increased imports are causing
serious injury, and then to exclude imports from one source from the application of
the measure, would be to give the phrase “product being imported” a diff erent mean-
ing in Articles 2.1 and 2.2 of the Agreement on Safeguards. In Article 2.1, the phrase
would embrace imports from all sources whereas, in Article 2.2, it would exclude
imports from certain sources. This would be incongruous and unwarranted. In the
usual course, therefore, the imports included in the determinations made under Articles
2.1 and 4.2 should correspond to the imports included in the application of the measure,
under Article 2.2 (emphasis added).
¹⁰⁶ Summing up prior case-law, the AB, in its report on US—Line Pipe, shied away from clari-
fying in more general terms the relationship between Art XXIV of the GATT and Art 2.2 of the
SG and avoided addressing the issue head on (§ 198). Implicitly, however, it did, since parallelism
can lead to a situation where preferential partners are also hit by a safeguard. GATT, Art XIX is
not mentioned in the list of measures excluded from the obligation to liberalize substantially all
trade, which binds members of a PTA (Art XXIV.8 of the GATT). Assuming the list is exclusive
(a hypothesis not yet tested in case-law), parallelism can run counter to Art XXIV.8 of the GATT.
¹⁰⁷ Standing and preliminary evidence requirements play an essential role in maintaining the
balance between the authorities’ right to investigate and impose measures and the exporters’ right
to be able to trade without undue interference or harassment. Their absence from the SG Agreement
means that harassment resulting from nuisance suits is not squarely addressed. On the other hand,
there are two counter forces: as we will see, safeguards are less attractive than AD duties, since
 Safeguards 369

The SG Agreement does not prejudge the duration of the investigation. In

practice, it takes approximately 12 months (across WTO members), and the data
used extends to the 3 years prior to the initiation of the investigation.¹⁰⁸ The
investigating authority is under a due process obligation: Art 3.1 of the SG requires
from a WTO member:
(1) to provide reasonable public notice to all interested parties;
(2) to provide all such parties with the opportunity to present evidence and their
views and to respond to the presentations of other parties; and
(3) to publish a report setting forth their findings and reasoned conclusions
reached on all pertinent issues of fact and law.¹⁰⁹
Recall that WTO members, in accordance with Art 2 of the SG, can impose
safeguards only upon prior demonstration that four conditions have been cumu-
latively met:
(1) a product is being imported in increased quantities;
(2) so as to cause;
(3) serious injury to the domestic industry producing the like or directly competi-
tive product;
(4) increased imports have to be the result of unforeseen developments.
Assuming that the four conditions have been met, a WTO member can lawfully
impose safeguards. The AB, in its report on US—Line Pipe, advanced a distinc-
tion between (§§ 83–4):
(1) the right to impose a safeguard; and
(2) the lawful application of a safeguard.
For a right to exist, a WTO member must ensure that it has met all of the require-
ments enshrined in Art 2.1 of the SG and Art XIX of the GATT, namely, it must
show that its domestic industry has suffered serious injury as a result of unfore-
seen developments that have led to increased imports; for an application to be
lawful, the safeguard measure may be applied only to the extent necessary to
counteract the resulting damage (Art 5 of the SG).

they cannot exclusively target individual exporters. Moreover, the statutory requirements and the
standard of review favour the use of AD over safeguards.
¹⁰⁸ It can, however, be longer. The panel in its report on US—Line Pipe found that a 5-year
period was consistent with the SG Agreement (§ 7.201). The same panel found that the same period
for examining data on imports and injury must be used (§ 7.209).
¹⁰⁹ The only other obligation is that members conduct such safeguards investigations in accord-
ance with procedures previously established and made public.
370 State Contingencies

7.4.2 Unforeseen developments

Evidently, assuming a concession has been made, other things equal, one would
expect (foresee) increased imports. Such normal expansion of imports is not cap-
tured by the term ‘unforeseen developments’. We need something more, but where
should we draw the line? It is of course not an easy exercise to determine precisely
ex ante by how much trade will expand as a result of tariff reductions. Hence, this
test cannot be applied with mathematic precision.
Case-law has contributed some, albeit not far-reaching clarifications in this
respect: first, in its report on Korea —Dairy, the AB held that, in its view, unfore-
seen should be read as synonymous to unexpected, as opposed to unpredictable
which would be synonymous to unforeseeable (§ 84). This is not a particularly
helpful distinction. Second, case-law addressed the question as to when devel-
opments should be regarded as unforeseen. The panel report on Argentina—
Preserved Peaches, reflecting prior case-law by the AB, records the view that
developments should be unforeseen at the time when concessions were made
(§§ 7.26–7.28). Third, in US—Steel Safeguards, the AB stated that it does not suf-
fice that a WTO member considered data which could be relevant to an unfore-
seen developments evaluation. It must also explain how such data satisfies the
unforeseen developments requirement (§ 329).¹¹⁰
In a nutshell, we know that the WTO member imposing safeguards must
demonstrate that it took account of the unforeseen developments, that the term
should not be equated to unforeseeable, and that developments should be unfore-
seen when concessions had been made. Is this helpful? Not particularly. Sykes
(2003c) explains that in light of the variables that might influence a potential
trade outcome, it is simply impossible to reasonably request a WTO member
to foresee events that will occur in the not immediate future. To make matters
worse, it could, theoretically at least, be the case that the time-span is quite long:
what if a WTO member did not make any bindings after the Kennedy or the
Tokyo round?¹¹¹ Should it still be held liable for not having foreseen events occur-
ring 30 years later? Grossman and Mavroidis (2004) advanced a proposal to the
effect that the last round of trade negotiations should be the point in time that
counts for the purposes of this exercise.¹¹²

¹¹⁰ The AB was upholding a panel finding to the effect that a mere reflection of a plausible
scenario does not suffice. The panel in its report on US—Steel Safeguards found that the follow-
ing passage from an ITCom decision failed short of meeting the requirement to show unforeseen
developments (§ 10.122):
the weakness of the USITC Report is that, although it describes a plausible set of
unforeseen developments that may have resulted in increased imports to the United
States from various sources, it falls short of demonstrating that such developments
actually resulted in increased imports into the United States causing serious injury
to the relevant domestic producers.
¹¹¹ The first round took place in the sixties, whereas the second in the seventies.
¹¹² Still, the authors accept that even such delimitation cannot by itself take care of all problems
that the satisfaction of the unforeseen developments requirement might give rise to.
 Safeguards 371

So far, the GATT working party in US—Fur Felt Hats emerges as the only case where
the unforeseen developments requirement was interpreted in a meaningful manner:
“that the fact that hat styles had changed did not constitute an ‘unforeseen development’
within the meaning of Article XIX”, but that the effects of the special circumstances of
this case, and “particularly the degree to which the change in fashion aff ected the competitive
situation, could not reasonably be expected to have been foreseen by the United States
authorities in 1947, and that the condition of Article XIX that the increase in imports
must be due to unforeseen developments and to the effect of the tariff concessions can
therefore be considered to have been fulfilled.”¹¹³(emphasis added)

7.4.3 Increased imports

The term ‘increased quantities’ (imports) was interpreted by the AB in its report on
Argentina—Footwear (EC), where it ruled that panels should be looking at trends
instead of isolated transactions or absolute numbers based on an end-point to end-
point comparison (§ 129). In its report on Argentina—Footwear (EC), the AB
emphasized that not just any increase over a period of time will do, rather it is neces-
sary for the competent authorities to examine recent imports.¹¹⁴ In sum, the AB
held the much repeated view that trends of imports should be recent, sudden, sharp,
and significant enough, both quantitatively and qualitatively, so as to cause serious
injury. The AB in its report on US—Steel Safeguards explained that what matters is
a reasoned explanation that the trend supports a finding that there is a trend to the
effect that imports are increasing. Even a decrease at the end of the period does not
necessarily detract from a finding that imports have increased, assuming a reason-
able explanation has been provided supporting such a conclusion (§ 367):
We agree with the United States that Article 2.1 does not require that imports need to
be increasing at the time of the determination. Rather, the plain meaning of the phrase
“is being imported in such increased quantities” suggests merely that imports must have
increased, and that the relevant products continue “being imported” in (such) increased
quantities. We also do not believe that a decrease in imports at the end of the period of inves-
tigation would necessarily prevent an investigating authority from finding that, neverthe-
less, products continue to be imported “in such increased quantities.” (original emphasis)

7.4.4 Injury
Safeguards can be shown upon demonstration of injury, or threat to injury. The AB, in
its report on US—Lamb supplied its understanding of the two terms (§§ 124–5):
The standard of ‘serious injury’ set forth in Article 4.1(a) is, on its face, very high.
Indeed, in United States—Wheat Gluten Safeguard, we referred to this standard as ‘exact-
ing’. Further, in this respect, we note that the word ‘injury’ is qualified by the adjective

¹¹³ Report of the Inter-sessional Working Party on the Complaint of Czechoslovakia Concerning the
Withdrawal by the United States of a Tariff Concession under Article XIX of the GATT (US—Fur Felt
Hats), GATT/CP/106, adopted 22 October 1951, § 12, as quoted in panel report, US—Lamb, § 7.23.
¹¹⁴ AB report, Argentina—Footwear (EC), § 130.
372 State Contingencies
‘serious’, which, in our view, underscores the extent and degree of ‘significant overall
impairment’ that the domestic industry must be suffering, or must be about to suffer,
for the standard to be met. We are fortified in our view that the standard of ‘serious
injury’ in the Agreement on Safeguards is a very high one when we contrast this stand-
ard with the standard of ‘material injury’ envisaged under the Anti-Dumping Agreement,
the Agreement on Subsidies and Countervailing Measures (the ‘SCM Agreement’) and the
GATT 1994. We believe that the word ‘serious’ connotes a much higher standard of
injury than the word ‘material’.
Moreover, we submit that it accords with the object and purpose of the Agreement on
Safeguards that the injury standard for the application of a safeguard measure should be
higher than the injury standard for anti-dumping or countervailing measures, since, as
we have observed previously:
[t]he application of a safeguard measure does not depend upon ‘unfair’ trade actions, as
is the case with anti-dumping or countervailing measures. Thus, the import restrictions
that are imposed on products of exporting Members when a safeguard action is taken
must be seen, as we have said, as extraordinary. And, when construing the prerequisites
for taking such actions, their extraordinary nature must be taken into account.
Returning now to the term ‘threat of serious injury’, we note that this term is con-
cerned with ‘serious injury’ which has not yet occurred, but remains a future event whose
actual materialization cannot, in fact, be assured with certainty. We note, too, that Article
4.1(b) builds on the definition of ‘serious injury’ by providing that, in order to constitute
a ‘threat’, the serious injury must be ‘clearly imminent’. The word ‘imminent’ relates to
the moment in time when the ‘threat’ is likely to materialize. The use of this word implies
that the anticipated ‘serious injury’ must be on the very verge of occurring. Moreover, we
see the word ‘clearly’, which qualifies the word ‘imminent’, as an indication that there
must be a high degree of likelihood that the anticipated serious injury will materialize
in the very near future. We also note that Article 4.1(b) provides that any determination
of a threat of serious injury ‘shall be based on facts and not merely on allegation, conjecture
or remote possibility.’ (emphasis added) To us, the word ‘clearly’ relates also to the fac-
tual demonstration of the existence of the ‘threat’. Thus, the phrase ‘clearly imminent’
indicates that, as a matter of fact, it must be manifest that the domestic industry is on the
brink of suffering serious injury. (original emphasis)
So, serious is a higher standard than material, but what is the material injury stand-
ard in the first place? If at all, this passage suggests that the AB is probably seeing
a parallelism in the definition of injury across the AD and the SG Agreements.
What is significant or serious is of course very subjective. Panels have not addressed
this concept in any of the cases dealt with so far. It remains to be seen whether in
practice there is any meaningful difference between material injury and serious
injury. It is important to recall the standard of review applied by panels in review-
ing an injury determination in the safeguards context: whether an adequate,
reasoned and reasonable explanation has been provided by the investigating
authority of how the facts support the determination made. In other words, a
panel may neither substitute its conclusions for those of the authorities (de novo
review), but it should not show total deference to the investigating authority’s
determination either, if not the review would not be meaningful. In US—Lamb,
 Safeguards 373

the AB translated this general standard in practical terms in the following way,
highlighting the need to examine the adequacy of the authority’s explanation and
reasoning in light of some alternative plausible explanation of the facts (§ 106):
Panels must, therefore, review whether the competent authorities’ explanation fully
addresses the nature, and, especially, the complexities, of the data, and responds to other
plausible interpretations of that data. A panel must find, in particular, that an explan-
ation is not reasoned, or is not adequate, if some alternative explanation of the facts is
plausible, and if the competent authorities’ explanation does not seem adequate in the
light of that alternative explanation. Thus, in making an ‘objective assessment’ of a claim
under Article 4.2(a), panels must be open to the possibility that the explanation given by
the competent authorities is not reasoned or adequate.¹¹⁵
Investigating authorities must examine all factors mentioned in Art 4.2 of the
SG. The AB, in its report on US—Line Pipe had the opportunity to address the
question as to whether an investigating authority can, using the same set of facts,
show serious injury or threat of serious injury. It had to respond to address a claim
to the effect that a discrete finding of either injury or threat of injury is required,
for a lawful imposition of safeguard measures to be the case. The panel had taken
the view that a discrete finding is necessary, since, in its view, the same set of
facts cannot simultaneously support a finding of injury and a finding of threat of
injury. Adopting a textual interpretation, the AB reversed the panel in this respect
and held the view that, a discrete finding is not required by the SG Agreement. In
doing so, the AB took the view that the threat standard is a lower threshold than
the serious injury standard. What is important for the right to impose a safeguard
measure to exist is that at least a threat of injury be established (§§ 169–72).
This view pays no attention at all to one crucial difference between the two tests:
whereas an injury standard is by definition a backward looking review (injury
must have been established as a result of increased imports), a threat of injury

¹¹⁵ The panel in Argentina—Preserved Peaches considered ‘the temporal focus of the competent
authorities’ evaluation of the data in making their determination of a threat of serious injury, and
whether their explanation was adequate in the light of any plausible alternative explanation of the
facts’, § 7.104. It reached the following conclusion (§ 7.104):
The directors who voted in favour of the preserved peaches measure viewed the
data for the most recent period in isolation and did not acknowledge the alternative
plausible explanation. The considerable increase in imports in 2000 and deterior-
ation in certain injury factors—viewed in isolation—led them to a potentially very
different conclusion from an evaluation in the light of all data before the competent
authorities. They explained their finding on the basis of the most recent period and
did not offer any explanation of that data in light of the longer term data which was
before them. They did not seek to deal with the alternative plausible explanation,
even though it was disclosed in the technical report.
The Panel is not substituting its own opinion for that of the competent author-
ities. In fact, the Panel has not formed its own opinion on either the situation of the
domestic industry or the capacity of imports to cause serious injury in 2001. Rather,
the Panel finds that for the reasons given above, the explanation of the determin-
ation of a threat of serious injury was not reasoned or adequate as required by Article
4.2(a) (§§ 7.116–7.117).
374 State Contingencies

standard is by definition the opposite: a forward looking exercise (injury might

occur as a result of the explanatory variables employed in the SG Agreement).
There is thus certainty that something happened, as opposed to uncertainty that
something might happen. If the same facts would yield the same results always,
there would be no need to introduce a threat of injury standard. It is precisely
because of the uncertainty that this will always be the case, that the threat of
injury standard has been introduced. At the very least, an investigating authority
that confines itself to facts that have already established injury must ‘enrich’ its
analysis with some predictive power: a set of facts, under assumptions, will always
lead to injury.
While the AD and SCM Agreements contain a definition of the term ‘like
product’, the SG Agreement does not. It seems justified to assume, however, that
the same definition of like product applies here as well: a product which is identical
to the imported product, ie alike in all respects, or in the absence of such a prod-
uct, a product which has characteristics closely resembling those of the imported
product. The AB, in its report on US—Cotton Yarn explained that, the term
‘domestic industry of the like or directly competitive product’ should be interpreted
in a product-oriented (as opposed to producer-oriented) manner. It thus expressed
its preference for demand-side criteria and relegated the relevance of supply-side
considerations (§ 86). In US—Lamb, the AB, like the panel before it,¹¹⁶ rejected
this argument in no uncertain terms, concluding that it is not permitted under
Art 4.1(c) of the SG to include input producers as part of the industry producing
the ‘like’ end-product (§ 92). The AB thus concluded that by expanding the term
‘domestic industry’ to include producers of other products, namely, live lambs,
the ITC defined the domestic industry inconsistently with Art 4.1(c) of the SG; it
should have limited its reach to packers and breakers of lamb meat only.¹¹⁷
An investigating authority must ensure that it attributes injury to increased
imports and it does not attribute injury to factors other than increased imports
that might have contributed to the injury (Art 4.2 of the SG). To do that, it
must:
(1) examine all relevant (submitted and known) factors that might be causing
injury;

¹¹⁶ Panel report, US—Lamb, §§ 7.71–7.77 and 7.118. The panel considered that past panel
reports concerning industry definition in the context of the SCM and AD Agreements were rele-
vant to its interpretation and application of the industry definition under the SG Agreement and
found support for its analysis in GATT panel reports in United States—Wine and Grape Products,
Canada—Manufacturing Beef CVD, and New Zealand—Finnish Transformers. See panel report
on US—Lamb, §§ 7.75 and 7.78–7.100.
¹¹⁷ AB report on US—Lamb, §§ 95–6. For the sake of completeness, it should be added that
the ITCom had not made a finding that live lamb was a product directly competitive with lamb
meat. In particular, the ITCom defined lamb meat as the like product, and identified the growers,
feeders, packers and breakers as producers of that like product. See panel report on US—Lamb,
footnote 154.
 Safeguards 375

(2) separate the effects caused by increased imports from those caused by other
factors; and
(3) still find that increased imports have caused injury.¹¹⁸
Then, and only then, can it lawfully impose safeguards. The extent of permissible
safeguards depends on compliance not with the causality requirement (Art 4.2
of the SG), but with Art 5.1 of the SG which was probably intended to serve as
mere procedural requirement. WTO members must, according to Art 5.1 of the
SG, impose safeguards necessary to counteract the amount of injury caused by
increased imports only, and not the total amount of injury suffered by the domes-
tic industry. Thus, in contrast, to the AD and the SCM Agreements, the amount
of duties eventually imposed is not calculated on using the contingency (dump-
ing, subsidy) as benchmark, but using the extent of injury as benchmark (AB,
US—Line Pipe, § 216).
Finally, there is a procedural requirement that must be met: an investigating
authority must provide a reasoned and adequate explanation that injury caused by
factors other than increased imports was not attributed to increased imports, and
that it is increased imports that caused injury (AB, US—Lamb, § 103).

7.4.5 Are safeguards truly non-discriminatory?

When safeguard measures are imposed in the form of QRs, quotas, by virtue of
Art 5.2(a) of the SG, shall be allocated to supplying members by reference to their
share in the import market during a previous representative period.¹¹⁹ SG, Art.
5.2(a) does not specify any further the length of the representative period. Any
representative period will do, so it seems.¹²⁰ Consequently, although the idea is
often expressed that safeguards must be imposed on a non-discriminatory man-
ner, what happens through the imposition of a safeguard in the form of a quota is
that historic market shares are being maintained throughout the period when the
safeguard is in place.
On the other hand, WTO members can depart from the obligation to respect
historic market shares in their import market, and target the relatively more effi-
cient sources of supply by allocating them quotas which are less than their market
share as observed during the investigation period. This will be the case if certain
WTO members have increased their market share in the market of the member

¹¹⁸ See the extensive analysis of the causality requirement in Grossman and Mavroidis (2004).
¹¹⁹ Actually, Art 5.2(a) of the SG first provides that the member imposing the safeguard measure
may seek agreement with respect to quota allocation with all other Members having a substantial
interest in supplying the product concerned. It is, when this method is not reasonably practicable
that the Member imposing the safeguard measure ‘shall allot . . . shares based upon the proportions
supplied by such members during a previous representative period . . . due account being taken of
any special factors which may have affected . . . the trade in the product’.
¹²⁰ SG, Art 5.2(a) does specify that an authority is to take ‘due account . . . of any special factors
which may have affected ... the trade in the product’. Th is could be interpreted to use a ‘representa-
tive’ period which does not include the most recent period of increased imports.
376 State Contingencies

imposing a safeguard in disproportionate quantities. This is what quota modula-

tion under Art 5.2(b) of the SG amounts to.
SG, Art 5.2(b) allows such quota modulation, in case a clear demonstration has
been provided to the WTO committee on safeguards to the effect that:
(1) imports from certain members have increased in disproportionate percentage
in relation to the total increase of imports of the product concerned in the
representative period;
(2) the reasons for the departure from the historic patterns are justified; and
(3) the conditions of such departure are equitable to all suppliers of the product
concerned.¹²¹
It is noteworthy that the agreement does not provide that the WTO committee
on safeguards has to authorize such a departure from historical patterns. In case
of disagreements as to the persuasiveness of the provided demonstration, it seems
that recourse to panel procedure is the only available option to decide on the
merits of the case.
Allowing a certain minimum amount of imports to enter the country and
respecting historical patterns only apply in cases where the safeguard measure
takes the form of a QR. There is no room for differentiated tariffs for example, in
case a WTO member prefers to use tariffs instead of QRs.¹²² The only obligation
in such cases is to ensure that the level of the tariff increase is not higher that the
amount of injury caused by increased imports alone.¹²³

7.4.6 The issue of compensation

SG, Art 8 provides that before imposing safeguard measures, the WTO member
wishing to impose it will enter into negotiations with the affected members, the
object of which is to compensate through concessions in another product market
for loss of market shares in the product market where a safeguard is being taken
(Art 8.1 of the SG).¹²⁴ In other words, a member imposing safeguard measures has
to pay¹²⁵ for offering this protection to its domestic industry. Such an obligation

¹²¹ Such targeted safeguard measures may only be used in case of a finding of current serious
injury and not in the case of a threat of serious injury.
¹²² Collusion thus with exporters would strongly suggest in favour of using a QR rather than a
tariff.
¹²³ In US—Line Pipe, the panel rejected the argument by Korea that tariff quotas are a form of
quotas/QR (§ 7.69).
¹²⁴ Hence the importance of the notification requirement under Art 12.3 of the SG which
will form the basis for any meaningful consultations under Art 8.1. See panel report, US—Wheat
Gluten, § 8.206.
¹²⁵ It, of course, redistributes wealth across its domestic producers, by over-exposing some to
foreign competition while shielding off others. Rosendorf and Milner (2001, pp 853ff ) take the
view that escape clauses are useful only when they impose some kind of cost for their use, because
paying signals an intention to comply with the agreement in future time.
 Safeguards 377

to compensate does not exist in other contingent protection instruments (neither

in AD, nor in countervailing).
The obligation to compensate probably corresponds to the idea that safeguards
might improve national welfare, but have detrimental effects on global welfare:
compensation must thus be paid to affected parties in order to ensure that the
effect of beggar thy neighbour policies is minimized.
Assuming that there is no agreement within 30 days between the affected
WTO member(s) and the member proposing to impose the safeguard, the affected
member(s) can withdraw substantially equivalent concessions or other obligations
under GATT 1994, unless the WTO Council for Trade in Goods disapproves of
such action.¹²⁶ It is highly unlikely, however, that the WTO Council for Trade in
Goods will disallow such action since the affected members will, in all likelihood,
be blocking a consensus to this effect.¹²⁷
The term ‘affected parties’ is not specified any further in the SG Agreement.
There is, however, a link between Art 8.1 of the SG and Art 12.3 of the SG which
refers to consultations
with those Members having a substantial interests as exporters of the product concerned.
The panel, and the AB report on US—Wheat Gluten, pointed explicitly to the link
between the two provisions.¹²⁸ In light of their approach on this issue, it seems safe
to conclude that the term ‘affected parties’, appearing in Art 8.1 of the SG, should be
understood as equivalent to the term ‘members having a substantial interest as export-
ers of the product’, appearing in Art 12.3 of the SG. Consequently, only a sub-set of
the WTO membership will be entitled to suspend concessions in case of disagree-
ment as to the adequate means of trade compensation to be paid.¹²⁹
In fact, the obligation to compensate can be avoided, if the WTO member
concerned proposes a safeguard action the maximum duration of which will not
exceed three years. This is so because Art 8.3 of the SG provides that the right of
suspension of equivalent level of concessions by the affected members shall not be
exercised for the first 3 years that a safeguard measure is in effect. The condition
is that the measure:

¹²⁶ SG, Art 8.2. Th is provision further states that a member should do this at the latest 90
days following imposition of the measure. It must give the WTO Council on Trade for Goods 30
days to disapprove the proposed suspension. It seems therefore that a member has to announce its
suspension at the latest 60 days following application of the safeguard measure.
¹²⁷ This might give the aff ected members the incentive to over-shoot their injury. The tables will
now be turned and this time it is the member imposing the safeguard who will feel affected by the
amount of countermeasures imposed against it. In such a case, it can only initiate dispute settle-
ment proceedings against the member(s) imposing countermeasures. Assuming that over-shooting
takes place, it might provide the member imposing the safeguard with a disincentive to do so. SG,
Art 8.1 seems to have been drafted in a sloppy manner since it can be abused in both directions.
¹²⁸ See AB report, US—Wheat Gluten, § 146.
¹²⁹ The term ‘substantial interest’ is not defined any further. One could, of course, seek inspir-
ation in the term ‘principal supplying interest’ appearing in Art XXVIII of the GATT. Such a con-
struction of the term ‘substantial interest’, however, has not as yet been condoned by the AB.
378 State Contingencies

(1) has been taken as a result of an absolute increase in imports, and

(2) it was taken in conformity with the provisions of the agreement (Art 8.3 of
the SG).
The obvious question that follows is, of course, who will determine whether the
measure conforms with the provisions of the SG Agreement? In other words, can
an affected member suspend an equivalent level of concessions effective from the
moment of imposition of the measure, unilaterally determining that the measure
does not conform with the provisions of the SG Agreement? This would, in effect,
force the member taking the safeguard action to establish before a WTO panel
that its measure conforms with the SG Agreement and thus that the suspension
of concessions during the first three years was illegitimate. The other possibility
is what happened for example in the US—Steel Safeguards case.¹³⁰ In this case,
however, the United States repealed the steel safeguard shortly after the AB issued
its ruling, and even before the DSB had had a chance to adopt the report.¹³¹ The
announced countermeasures by affected members were abandoned shortly there-
after. The potential for abuse is high in light of the duration of the panel process,
and the fact that, as we will see in Chapter 5, WTO remedies for violation are de
facto always prospective.¹³²

7.4.7 The duration of safeguards

A safeguard measure can be imposed for an initial period of up to 4 years (Art 7.1
of the SG). It can be extended for a maximum 4 years more, in case it has been
determined that:
(1) the safeguard measure continues to be necessary to prevent or remedy serious
injury; and
(2) that there is evidence that the industry is adjusting.

¹³⁰ Actually, the European Union first published a list of products on which additional duties
were going to be levied as of the third birthday of the US Safeguard Measure, or the fifth day fol-
lowing the date of a decision by the WTO DSB that the measure is incompatible with the WTO
Agreement, if that is earlier. Council Regulation (EC) 131/2002 of 13 June 2002.
¹³¹ The measures were repealed on 4 December 2003, while the reports were adopted by the
DSB on 10 December 2003 only. The EC repealed its planned countermeasures on 12 December
2003. See Council Regulation (EC) 2168/2003 of 12 December 2003 repealing Regulation (EC)
No 1031/2002 establishing additional customs duties on imports of certain products originating
in the United States of America [2003] OJ L326/1.
¹³² The potential for abuse is from both sides: assume A imposes safeguards and announces that
its measure will stay in place for less than 3 years, and that B disagrees with the consistency of the
safeguard imposed. If B acts in good faith and takes A to a panel, it will eventually prevail after the
3-year period has expired and, in light of the prospective nature of the remedies, B will satisfy itself
that A stopped its illegal practice. If B reacts unilaterally (illegally) and imposes countermeasures
against A, it is A that will take B to the court and will have to await the final outcome of the process
before it can react to B’s unilateral measures.
 Safeguards 379

Such an extended measure may not be more restrictive than it was at the end of
the initial period.¹³³ Eight years is the maximum period for a safeguard measure
(Art 7.3 of the SG), an exception is possible for developing countries only. A safe-
guard measure imposed by a developing country is allowed to stay in place for a
maximum period of 10 years.¹³⁴
Measures imposed must be progressively liberalized at regular intervals.¹³⁵
The SG Agreement provides for a mandatory review of any safeguard measure of
more than 3 years, at the latest before the mid-term of this measure. If appropri-
ate, the measure is to be withdrawn or the pace of liberalization increased. Hence,
for any measure of more than 3 years there is a mandatory mid-term review, as
well as a sort of sunset review after 4 years.¹³⁶
Following the imposition of a safeguard measure for any amount of years, the
WTO member concerned, immediately after the expiry of the said period, can-
not impose a safeguard measure on the same product for the equal number of
years (Art 7.5 of the SG): for example, if country A takes safeguard action in the
area of cars for 8 years, it has to desist from a safeguard action with respect to cars
for a period of 8 years following the expiry of the original safeguard measure.
There is thus a dynamic use constraint instituted in the SG Agreement.¹³⁷ This
obligation not to impose safeguards on the same product for a period equivalent
to the period of imposition is often referred to as peace clause or grace period.¹³⁸
This grace period shall in any case not be shorter than 2 years, even if the measure
itself was applied for a shorter period of time.¹³⁹
One might find it counter-intuitive that safeguard measures will be imposed
for a period longer than 3 years: in practice, no compensation is due when safe-
guards do not extend 3 years, and there is no need for a review of the measure
either. This contrasts with the situation in case of a safeguard measure of more
than 3 years. Members imposing a maximum 3-year safeguard do not impose

¹³³ SG, Art 7.4 in fine.

¹³⁴ SG, Art 9.2.
¹³⁵ SG, Art 7.4. The exception is a safeguard measure of less than one year, which does not have
to be liberalized.
¹³⁶ There is nothing like a sunset review à la AD or SCM Agreement; the investigating authority
will still have to decide, however, whether to keep the safeguard measure in place.
¹³⁷ Bagwell and Staiger (2005) find that this feature raises the opportunity cost of using the
escape clause in the first place and it might raise the welfare of negotiating governments.
¹³⁸ When developing countries impose safeguard measures, this grace period is only half of the
time of application of the measure. In our example, a developing country member would have to
wait only 4 years before being allowed to impose safeguards on cars again. SG, Art 9.2.
¹³⁹ SG, Art 7.6 allows for a limited exception for very short safeguard measures of less than
6 months (up to 180 days) to this rule, provided that its substantive conditions have been met.
Bagwell and Staiger (2005) note that the peace clause provides WTO members with the incentive
to choose strategically the sectors where they will take protective action. Their point is well-taken
and would have its maximum value if safeguards provided the only means of contingent protec-
tion. In practice, however, it could be (and has been) the case that a WTO member, during the
peace clause, imposes AD duties on the products that were previously protected by safeguards.
380 State Contingencies

costs on other producers’ interests,¹⁴⁰ since no compensation will be paid. This is

all the more relevant given the fact that the WTO member imposing a safeguard
cannot choose the area where it will be paying compensation: affected WTO
members might suspend concessions in fields of their interest; all they have to
ensure is substantial equivalence between damage and concessions withdrawn.¹⁴¹
Still, based on the notifications of safeguard measures to the WTO committee
on safeguards, we notice that in fact, quite a number of safeguard measures are
imposed for a period exceeding 3 years.¹⁴²

7.5 Safeguards and voluntary export restraints (VERs)

GATT, Art XIX and SG, Art 2.2 do not allow interested states to target particu-
lar exporters. In the seventies and the eighties, some GATT contracting parties,
upon request, would agree to limit their export towards particular destinations.
This is how the notorious voluntary export restraints (VERs) saw the light of
day. So, a VER is not a formal treaty; it is, in practice, a unilateral reduction of
exports.¹⁴³
The legality of VERs was not formally tested by a GATT panel. There were
admittedly little, if any, incentives to mount a legal challenge: the requesting state
would not normally be attacking a practice it requested; and the country limiting
its exports was adjusting itself in a comfortable second-best, as the work of Smith
and Venables (1991), inter alia, has demonstrated: exporters would see their vol-
ume of exports reduced, but would be charging higher prices (as close as possible
to that of their competitors), thus pocketing monopoly rents. They would have
little incentive to challenge such measures. Those who would be facing a higher
flux of imports would have a mountain to climb in terms of evidence: if at all,
they could raise an NVC provided of course that they did not try to address such
practices through AD.¹⁴⁴ The trade regime was thus ill equipped to deal with
such cases, and so was probably the antitrust regime as well.¹⁴⁵ The absence of
a formal condemnation notwithstanding, a series of good arguments could be

¹⁴⁰ Consumer welfare will of course be negatively affected, since consumers will now have to
pay a higher price for the good on which a safeguard has been in place. The relative importance of
consumers’ interests, however, has been weighed before the decision to take safeguards was taken
(and obviously, set aside).
¹⁴¹ DSU, Art 22, is legally irrelevant here, since there is no dispute between the parties. The
aff ected members are free to choose the sectors where they will impose counter-measures.
¹⁴² See, for example, WTO Docs G/SG/N/8/EEC/2, G/SG/N/10/EEC/2, and G/SG/N/11/
EEC/2/Suppl. 1 of 16 March 2004.
¹⁴³ The reduction, of course, would not take place absent a request to this effect.
¹⁴⁴ The GATT panel report Japan—Trade in Semi-conductors faced facts which closely resemble
a VER.
¹⁴⁵ Mattoo and Mavroidis (1995) discuss the possibility to invoke the EC competition laws
to challenge the consistency of the VER on cars concluded between Japan and the European
Community. They conclude that, in all likelihood, Japanese producers could hide behind a ‘sover-
eign compulsion’ defence with good chances of thwarting a legal challenge against them.
 Safeguards 381

advanced in support of the opinion that they violate both Art XI of the GATT
(since they effectively amount to a QR), and Art XIX of the GATT (in light of
their targeting nature).
SG, Art 11.2 put an end in this discussion, as far as its legal dimension is con-
cerned, by outlawing recourse to VERs.¹⁴⁶ However, a footnote to Art 11.2 of the
SG reads:
An import quota applied as a safeguard measure in conformity with the relevant provi-
sions of GATT 1994 and this Agreement may, by mutual agreement, be administered by
the exporting Member.
The wording of this provision does not exclude that VERs have been introduced
through the back door: indeed, a QR administered by the exporter has all the
ingredients of a VER. A closer look points, nevertheless, to a different interpret-
ation: first, only QRs in conformity with this agreement will, if at all, be admin-
istered by the exporter. This means that the procedural requirements reflected in
the SG Agreement will have to be respected (quod non in a case of a VER); second,
the term ‘administered’ appearing in the footnote probably means that the export-
ing country is simply entrusted with the duty to ensure that no more than the
quantities unilaterally defined by the member taking the safeguard action will be
exported. It does not necessarily mean that the rents will stay with the exporter.
The rents could, theoretically at least, stay with the importer.¹⁴⁷

7.6 Safeguards: a safe harbour too far?

In his comprehensive and well-documented study, Sykes (2003c) criticizes the
manner in which case-law has approached the interpretation of the SG Agreement:
up until 2005, no national imposition has been judged GATT consistent by the
WTO adjudicating bodies. Some of these are impositions by the most sophisti-
cated national administrations. One could imagine that, assuming case-law con-
tinues to progress in a linear fashion in this context, WTO members might have
an incentive to have recourse to other forms of protection: AD emerges as prime
candidate; temporary violation of the agreement is another option, since remed-
ies in the WTO, as we will see in Chapter 5, are prospective.¹⁴⁸
It is true that nothing, in theory, stops a WTO member from imposing AD
duties during the constraint period, that is, after an original imposition of safe-
guards. On the other hand, nothing stops WTO members from having recourse

¹⁴⁶ This does not mean, however, that the incentives to conclude such arrangements have been
abolished as well. On the other hand, in the absence of an ex officio complaint in the WTO legal
system, it is difficult to see how, if concluded, they will be challenged before the WTO.
¹⁴⁷ This point is credited to lengthy discussions with Jagdish Bhagwati and Anastasios Tomazos.
It is probably true, however, that it is highly unlikely that the rents will be passed to the importer.
¹⁴⁸ Renegotiation of duties (Art XXVIII of the GATT) is not necessarily a close substitute, in
light of the requirements for lawful recourse to it (see Chapter 2).
382 State Contingencies

to AD duties instead of safeguards in the first place. At first glance, the two
instruments are not perfect substitutes: an AD investigation is more demanding
with respect to standing requirements, and provides protection against targeted
exporters only. It does not shield domestic producers from all foreign competi-
tion. A safeguard investigation is quite easy to kick off (no standing requirements)
and provides protection erga omnes. It is, however, temporary (dynamic use con-
straint): Art XIX of the GATT safeguards are meant to slow temporarily the pace
of adjustment to changes in the external economic environment; AD duties de
facto are often not temporary: they can be in place for as along as the dumping or
subsidization continues.¹⁴⁹ The differences notwithstanding, it seems reasonable
to assume that AD duties are used de facto as a form of safeguard, also because
case-law has been quite deferential to the investigating authority imposing AD
duties.¹⁵⁰ Does it make sense to privilege the use of AD duties over safeguards?
Recall that we found no compelling economics rationale justifying the use of
AD duties. Recall nevertheless, that the inclusion of a safeguard mechanism was
not discussed in the section on AD above. There are good arguments in favour
of introducing safeguards in a trade agreement: provision for safeguards may
ultimately increase the size of the cake that its members share, but their prac-
tical implementation is beset with the risk of abuse.¹⁵¹ Theory has identified two
positive aspects for including safeguards in a trade agreement like the GATT:
they provide trading partners with an extra incentive to make tariff reductions
(the argument being that absent safeguards, WTO members would be reluctant
to make generous tariff concessions), and they correspond to insurance policy
against shocks.
Dam (1970, p 80) described in the following manner the former (safeguards,
because of their flexibility, induce trade liberalization):
The GATT has a special interest in seeing that as many agreements for the reduction of
tariffs as possible are made. Enforcement of bindings is important in the GATT insofar
as such enforcement gives contracting parties the confidence necessary to rely upon tariff
concessions offered by other contracting parties. But because of the economic nature of

¹⁴⁹ Suppose that a WTO member imposes a 4-year safeguard measure on steel. At the end of
this period, it has two options: either to extend the measure up to 4 (additional) years (Art 7.2 of
the SG), or alternatively not to extend it. In the first case, it has to wait for another 8 years before
imposing a safeguard measure on steel anew. In the second case, 4 years. No similar rule applies in
the case of AD, where after the sunset of a measure 5 years after its imposition (Art 11.3 of the AD),
the importing country can effectively extend the AD measure.
¹⁵⁰ Crowley (2006) finds that there is an increase in the imposition of safeguards, the total num-
ber of which, however, is nowhere near the corresponding number of AD impositions. According
to the official WTO website (<http:// www.wto.org>), until the end of November 2006, 2938 AD
investigations and 155 safeguards investigations had been launched by all WTO members (starting
from 1 January 1995).
¹⁵¹ See Sykes (1990 and 1991) for analysis of the role of safeguards in trade agreements, in
particular as viewed from the perspective of public choice theory. See also Deardorff ’s (1987) treat-
ment of the role of tariff and non-tariff safeguards when social preferences are represented by a
Corden ‘conservative social welfare function’. See also Bagwell and Staiger (1990 and 2005).
 Safeguards 383
tariff concessions and the domestic political sensitivity inherently involved in tariff issues,
a system that made withdrawals of concessions impossible would tend to discourage the
making of concessions in the first place. It is better, for example, that 100 commitments
should be made and that 10 should be withdrawn than that only 50 commitments should
be made and that all of them should be kept.
Ethier (2002) provides good arguments in favour of this argument.¹⁵² Another
version of this argument, more closely modelled on a public choice approach
where governments are driven at least partly by other motives than social welfare
maximization, is discussed by Sykes (1991). The argument here is that govern-
ments may, after trade negotiations, face strong pressure for protection in certain
industries. A safeguard mechanism makes it possible to give in to such pressures,
and thus to avoid political setbacks if participating in liberalization. As a result,
governments are more prone to liberalize ex ante.
Horn and Mavroidis (2003) discuss the latter (insurance policy) in the follow-
ing terms: the economic environment is constantly changing; new products and
production technologies are discovered, consumer tastes change, governments
come and go, wars are fought, investments are made, new firms see the light of
day, etc. For a trade agreement to be fully efficient, it would need to adapt to these
changes. This adaptation could be fully achieved only under very special circum-
stances. If the parties could perfectly foresee the future path of events, then they
could specify a contract at the outset that would specify how the contract terms
would change in accordance with this path. In the absence of such information,
the parties may renegotiate the contract any time a change occurs. Alternatively,
the parties could write a fully state contingent contract, which specified commit-
ments for each possible outcome of the underlying economic environment.¹⁵³ It
would, under either of these circumstances, be possible to specify a trade contract
that ex post ensures the desirable levels of trade. If desirable, this contract could
ensure for a gradual adjustment to the changed environment. Hence, in neither
case would there be a role for any provision that allowed for an ex post change in
tariff bindings.¹⁵⁴ Tariff bindings in actual trade agreements are typically not
conditioned on external events, however. There is therefore a need for instru-
ments that allow ex post adjustment of effective levels of bindings—that is, for

¹⁵² Staiger and Tabellini (1987 and 1999), however, have advanced arguments casting doubt on
this proposition.
¹⁵³ The difference between the two scenarios is that in the former, the contract would specify
the commitments of the members at each date. These would then vary over time in response to
changes in the external environment. In the latter case, since the realization of these external events
would be unknown at the contracting date, the contract would specify for each date commit-
ments for each possible realization. Th is would thus be a significantly larger contract, but would
in principle achieve the same thing as the first contract. Similar arguments have been advanced by
Brainard and Verdier (1994).
¹⁵⁴ Tariffs are the only permissible form of protection in the WTO regime which cannot be
countered; quotas are illegal, domestic measures have to observe the non-discrimination principle,
and subsidies can be challenged either through countervailing duties or through other legal mech-
anisms. As of 2001, there are no non-actionable subsidies.
384 State Contingencies

escape clauses—and the GATT includes several provisions to this effect, such as
Arts XII, XV and XXVIII of the GATT. What, then, is the role of Art XIX of the
GATT, and the SG Agreement¹⁵⁵ in this arsenal of escape clauses? They can be
unilaterally imposed, and might for this reason be a quicker response to changes
in the economic environment in particular industries (depending on the admin-
istrative requirements imposed on safeguard investigations), than for instance an
Art XXVIII of the GATT renegotiation.¹⁵⁶ Safeguards thus, not only increase
the efficiency of the contract after external shocks, but they also have a separate
role to play: to reduce temporarily the rate of adjustment.
The analysis above points to the conclusion, that there is need for a safeguard
clause in the GATT. Is, however, the design of the GATT safeguard clause up to
the task? Sykes (2003c) points to a series of shortcomings relating to both the law
as such, as well as the case-law. In his view, largely shared by this author, case-law
has made a rather unworkable test out of a problematic drafting. For a start, as
Sykes (2003c) explains, it simply cannot be that imports, in and of themselves,
injure the domestic production of the like product. Imports constitute a mere
arithmetical difference between total consumption and domestic production at
a given price. It is the causes that lead to import surcharge that might be causing
injury. GATT, Art XIX does not impose on WTO members the obligation to
address these causes at all, and case-law, following a textual interpretation, never
did.¹⁵⁷
Reintroducing the unforeseen developments requirement was not a clever idea
either. The wording implies that some uncertainty is required.¹⁵⁸ So in stage 1,
two countries exchange concessions, and say country A believes that its tariff cut
will lead to extra imports not going beyond quantity Q. In stage 2, imports into
A go beyond Q. Now A will be called to show that it could not foresee this rise.
Assuming we reduce this requirement to emergency situations such as the domes-
tic industry has been burned down, or to technological shocks, then, meeting the
unforeseen requirement might be a relatively easy exercise. The problem is that
there is no indication that the legislator intended such use. Indeed, if this were

¹⁵⁵ As will be shown in detail infra, the safeguards regime of the WTO is reflected in Art XIX of
the GATT, and the WTO SG Agreement.
¹⁵⁶ It is implicitly assumed that there are costs associated with contracting, so that the parties do
not negotiate a new contract each time the environment changes.
¹⁵⁷ Grossman and Mavroidis (2004), while agreeing with Sykes, take the view that an imagina-
tive judge could by-pass this hurdle and still inquire into the causes that drive imports. They point
to another shortcoming, which however, is legislative, and not the result of case-law: GATT, Art
XIX nowhere defines the objective function of the safeguard action. Unless we know for sure the
purpose for introducing safeguards, it is simply impossible to judge the legitimacy of actions under-
taken by WTO members. A number of explanations ranging from structural adjustment assistance
to simply responding to import surge can be offered as valid excuses for safeguard action. So even if
their imaginative judge inquires into the causes that drove imports, he/she will have a hard time to
appreciate the legitimacy of actions taken to address such concerns.
¹⁵⁸ Compare Dam’s (1970, p 80) cited passage: his story is not based on the assumption that
imports have increased as a result of unforeseen developments.
 The Notion of Market Access Revisited 385

the case, Art XIX of the GATT would probably include (at least) an illustrative
list to this effect. Unless we link the demonstration of this requirement to some
form of quantitative estimation as to the quantity of additional imports that a
tariff cut would reasonably induce, the test risks becoming unworkable. The prob-
lem, of course, with this approach is what is a reasonable amount of imports, how
much evidence is required to show the expectations when concessions had been
exchanged, how easy it is to gather such data years after concessions had been
exchanged: the passage of time makes it even harder to meet this requirement.
The AB has nowhere so far provided its understanding of this term. Beyond the
banal interpretations that we discussed above, some of which by-pass the issue
and some of which are of a procedural nature, there is absolutely no indication as
to how the AB understands this term. It is probably a good idea, as Sykes (2003c)
argues, simply to drop it from the list of requirements that must be met for safe-
guards to be lawfully imposed.
Finally, we are still in no man’s land when it comes to calculating the amount
of injury that is required for a safeguard action to be GATT consistent. We know
that serious injury must be the result of increased imports, but what exactly ser-
ious injury is (after one has deducted the amount of injury caused by factors
other than imports) is still an open issue. It is thus not a paradox that even the
most sophisticated national administrations cannot get it right, in the eyes of the
AB. This is quite problematic. A very important weapon in the arsenal of WTO
members which, both in theory and empirically, can help maximize the welfare
of negotiating parties, has suffered important blows. The main effect so far has
probably been the switch to AD duties, a very distorting instrument as our ana-
lysis in section 5 above has shown.

8 The Notion of Market Access Revisited

Our discussion of the substantive obligations regulating trade in goods could be

summarized as follows:
(1) WTO members negotiate and agree to apply two forms of protection: a
default protection, and a state-contingent protection. The former can take
the former of tariffs only, while the latter can also take the form of a quota.
Default protection is further distinguished between MFN and preferential
tariffs. Preferential tariffs will apply to a sub-set of the WTO membership,
developing countries, or PTA partners;
(2) MFN tariff protection will be gradually reduced, principally, in the context
of multilateral negotiating rounds. As things stand, tariff protection, with
the exception of two tariff peaks (agriculture, and textiles) is at an all time
low and does not severely hamper international trade;
386 State Contingencies

(3) MFN tariff liberalization has led to preference erosion for beneficiaries and
the ensuing discussion on how to address it. At the time of writing, no con-
crete decision on this score has seen the light of day;
(4) The main impediment to international trade now is NTBs. There are many
problems associated with the regulation of NTBs. For a start, it is almost
impossible to write a complete contract, since initial negotiating costs would
be quite high. Moreover, social preferences change over time and the contract
would need to be renegotiated quite frequently, adding thus to the initial
negotiating costs;
(5) Contract incompleteness is not per se a problem. Assuming parties always
have the incentive to reveal the truth about their actions, the contract will be
naturally completed each time information has been revealed. The problem
is that trading partners often have the incentive to cheat: they might wish to
reveal that they have been pursuing certain policies not in order to achieve
an undesirable (by all Contracting Parties) outcome, but a desirable one.
It is, thus, the incentive for opportunism that exacerbates the problems of
contract incompleteness;
(6) The WTO judge will be asked to decide whether a policy is genuine or not,
operating in an asymmetry of information context. He/she can only rely on
the legislative proxies that the various agreements regulating trade in goods
have included;
(7) The default GATT recipe to address NTBs is through the non-discrimi-
nation principle. Non-discrimination is far from being a perfect proxy for
understanding whether protectionist behaviour has taken place. It simply
makes some choices costlier for the regulator, does not, however, eliminate
the potential for beggar thy neighbour behaviour. At the same time, non-
discrimination illustrates the negative integration that is being privileged by
the GATT;
(8) Case-law has added to the problems. It has not progressed in a linear man-
ner and has failed so far to come up with a comprehensive test that would
bring non-discrimination back within the four corners of the intended (by
the framing fathers) test: non-protectionism. Absent a genuine intent test,
case-law will fall short of meeting this standard;
(9) There have been attempts by the WTO judiciary to reverse the original
deregulatory bias, a by-product of case-law. They have failed to produce a
reliable framework of analysis. Notably, even these attempts have explicitly
distanced themselves from introducing an intent test in the context of Art III
of the GATT;
(10)The new generation agreements, such as TBT and SPS, which deal with a sub-
set of all NTBs, are better equipped to deal with constraining opportunism.
They have added proxies such as international standards, necessity, scientific
 The Notion of Market Access Revisited 387

evidence, and consistency, which can provide the judge with a better idea
about the true intent of the regulating state;
(11) This is far from being a won battle though. At the end of the day, WTO
judges will have to decide on which side (Type I, or Type II) they would most
likely to err, when uncertain about the outcome of a particular dispute;
(12) State-contingent protection is a necessity in a contract like the GATT. There
are very sound arguments in favour of including one operational safeguard
in this contract;
(13) The current SG Agreement, as interpreted by the AB, is both poorly equipped
to deal with the needs of WTO members which might legitimately wish to
avail themselves of this possibility. A redraft is probably necessary since, it is
quite unlikely at this stage that the AB will be willing to perform a spectacu-
lar revirement de jurisprudence;
(14) The SCM Agreement is also problematic. It is true, however, that the opti-
mal regulation of subsidies (or, at least, some of them) is not an easy exer-
cise. Case-law has added to the problems, especially in the field of remedies
against prohibited subsidies;
(15) It is very hard to come up with persuasive arguments in favour of retaining
the AD Agreement in the WTO edifice. Even the argument that it is an
appropriate surrogate for the current unworkable safeguards clause is far
from being convincing. It should be scrapped altogether;
(16) As things stand now, the main issues facing the world community are the
regulation of farm trade, the erosion of preferences, the PTAs, and the
ongoing struggle to address NTBs. The Doha round promises incremen-
tal changes in the first, and probably the second item, and nothing on the
remaining two items;
(17) It is questionable whether PTAs and NTBs can be addressed in the context
of a negotiating round, where concessions are being exchanged. It is ques-
tionable whcether they could be addressed off rounds as well. The GATT
legislative process, as we will see in Chapter 5 is far from being operational.
At the same time, these issues are horizontal, policy issues, and should not
be left to the discretion of the judge who is best equipped to make transac-
tion-specific judgments.
 5
Institutional Provisions

1 Introductory Remarks 389

2 Participation 390
2.1 Contracting Parties to the GATT, WTO members 390
2.2 Accession 391
2.3 Non-application 394
3 Decision Making 395
3.1 Consensus, the default rule 395
3.2 Special procedures 396
3.2.1 Interpretations 396
3.2.2 Amendment of the GATT 397
4 Dispute Settlement 398
4.1 The early GATT years 398
4.2 The late GATT years 400
4.3 The WTO DSU 402
4.4 The WTO: exclusive forum for adjudication of
trade disputes 404
4.5 A closer look into procedures 406
4.5.1 Request for establishment of a panel 406
4.5.2 Burden of proof 407
4.5.3 Discovery powers 408
4.5.4 Standard of review 410
4.5.5 The process before the AB 411
4.5.6 The types of complaints before panels and the AB 411
4.5.6.1 Violation complaints 412
4.5.6.2 NVCs 412
4.5.6.3 Situation complaints 415
4.5.7 Recommendations and suggestions by panels
and the AB 416
4.6 Enforcing WTO obligations 417
4.6.1 The process in a nutshell 417
4.6.2 The types of complaint matter for enforcement
purposes 418
4.6.3 Compliance within a reasonable period of time 418
4.6.4 Compliance panels 419
4.6.5 Compensation 420
4.6.6 Suspension of concessions 421
4.6.6.1 Procedural issues 421
 Introductory Remarks 389
4.6.6.2 Quantifying suspension of concessions 422
4.7 Monitoring compliance 424
4.8 The value of precedent in the WTO 425
4.9 Does WTO enforcement work? 426
5 Waivers 435
6 Transparency 436
6.1 Introductory remarks 436
6.2 Trade Policy Review Mechanism (TPRM) 437
6.3 Art X of the GATT 440
6.3.1 The discipline in a nutshell 440
6.3.2 The coverage 441
6.3.3 The standard of review 442
6.3.4 Laws of general application 443
6.3.5 Uniform, reasonable, and impartial administration 443
6.3.5.1 Uniform administration 444
6.3.5.2 Reasonable administration 444
6.3.5.3 Impartial administration 445
6.3.6 The obligation to maintain independent tribunals 445
7 Relation to the Havana Charter 446

Overview

The institutional provisions of the GATT have been overtaken by the provision
included in the Agreement Establishing the WTO. The GATT, as we saw in Chapter
1, was not designed to be an institution; it gradually became one when it lost its insti-
tutional umbrella, the ITO. The Agreement Establishing the WTO on the other
hand, largely replicated the GATT approach to institutional life: consensus remains
the default rule for adopting decisions, except for dispute settlement. Consensus is
the most telling reflection of the GATT pragmatism: it might have delayed the lib-
eralization process, but, critically, provided it with a much needed legitimacy during
difficult times. The various GATT working parties and committees, that saw the
light of day over the years, have either been consolidated (Art XXIV working par-
ties became the committee on regional trade arrangements (CRTA), as we saw in
Chapter 2), or been absorbed by WTO committees. In this chapter, following a brief
presentation of the issues we entertain (section 1), we first discuss participation in
the GATT/WTO (section 2), then decision making (section 3), dispute settlement
(section 4), waivers (section 5), transparency (section 6), and finally, the relationship
of the GATT to the Havana Charter following the advent of the WTO (section 7).

1 Introductory Remarks

With the advent of the WTO, the whole institutional arrangement of the world
trading system has been reshaped. The GATT provisions, however, have not been
390 Institutional Provisions

abolished. In all probability, for reasons of administrative inertia, or even because

some might have felt that a renegotiation of the GATT would be tantamount to
opening up Pandora’s box, the GATT remained as such, the advent of the WTO
notwithstanding. Recall that the GATT 1947 has few institutional provisions.
On the other hand, the WTO institutional provisions were drafted in a manner
compatible with the corresponding provisions in the GATT.
WTO institutional provisions reflect not only the letter, but also the prac-
tice of GATT provisions. A prominent example is offered by Art XXV.4 of the
GATT which indicates that majority of the votes cast is the default rule for
decision making in the GATT. In practice, however, the GATT Contracting
Parties rarely, if ever, voted: decisions were taken by consensus.¹ Art IX is the
corresponding provision in the Agreement Establishing the WTO (decision
making); it calls for a continuation of GATT practice (rather than GATT law)
and consensus, which was the de facto default rule in the GATT era, and now
becomes the de jure default rule in the WTO era. This is probably the most expli-
cit acknowledgement of the positive influence that GATT pragmatism has exer-
cised over the years on the shaping of world trade relations: WTO members thus,
by formalizing their links through the advent of the WTO, did not want to break
with GATT practice.
In what follows, we examine the GATT and WTO provisions relating to:
(1) Participation;
(2) Decision making process;
(3) Dispute settlement;
(4) Waivers;
(5) Transparency;
(6) Relationship with the Havana Charter.

2 Participation

2.1 Contracting Parties to the GATT, WTO members

GATT, Art XXVI makes it clear that states, as well as customs territories can
become Contracting Parties to the GATT. GATT, Art XXVI.5(c) reads in this
respect:
If any of the customs territories, in respect of which a contracting party has accepted this
Agreement, possesses or acquires full autonomy in the conduct of its external commercial
relations and of the other matters provided for in this Agreement, such territory shall,

¹ See on this score Hoekman and Kostecki (2001). As noted before, decisions gain in legitimacy
when consensus-based.
 Participation 391
upon sponsorship through a declaration by the responsible contracting party establishing
the above-mentioned fact, be deemed to be a contracting party.
Hong Kong (now Hong Kong, China) is an appropriate illustration to this effect:²
it joined the GATT in is own right on 23 April 1986. Before that, Hong Kong’s
participation in the GATT was governed by a declaration made on 28 June 1948
by the United Kingdom, concerning the application of the GATT to Hong Kong.
By the same token, Macau became a Contracting Party to the GATT in its own
right, in January 1991.
The term ‘Contracting Parties’ was used in GATT because the GATT was ori-
ginally designed to be an agreement (a contract) and not an institution: the term,
thus, appropriately reflects the contractual nature of the GATT. The GATT was
transformed de facto into an institution when the ITO failed to materialize,
but the term ‘Contracting Parties’ persisted. The term ‘Contracting Parties’ is
defined in Art XXXII of the GATT, and covers original Contracting Parties and
all acceding countries. Nevertheless, in practical terms, Contracting Parties were
behaving like members of an international institution. The Contracting Parties
to the GATT have, by virtue of Art XI.1 of the WTO Agreement, now become
original members of the WTO. With the accession of Vietnam (January 2007),
the WTO numbers 150 members.

2.2 Accession
Accession to the GATT is regulated in Art XXXIII. Typically in the GATT era,
a working party for the accession of the candidate country (or customs territory)
would be established, which would negotiate the terms of accession. Following
the advent of the WTO, the GATT reverted back to just an agreement. As a
result, new members accede to the WTO, and Art XII of the WTO Agreement
regulates the accession process.³
WTO Agreement, Art XII does not contain any procedural requirements.
Nonetheless, the procedural aspects of the accession process have been developed
in a relatively uniform manner. A country or separate customs territory begins
its accession process by submitting a formal communication to the Director-
General of the WTO indicating its willingness to become a member of the WTO
and requests the establishment of a working party (WP). The communication is
circulated to all WTO members.

² Hong Kong, China continues to be a WTO member, distinct from China (the People’s
Republic).
³ The reason why the term ‘absorbed’ is preferred to the term ‘superseded’, which is technic-
ally the one used in other contexts, has to do with the fact that accession to the WTO includes but
does not exhaust adherence to the GATT: WTO members will also have to respect a series of other
obligations.
392 Institutional Provisions

The WTO General Council will then consider the application and the estab-
lishment of a WP. The terms of reference of a WP, assuming its establishment has
been agreed upon, are typically:
to examine the application for accession to the WTO under Article XII and to submit
to the General Council/Ministerial Conference recommendations which may include a
draft Protocol of Accession.
A chairman of the WP will be appointed following consultations between the
Chairman of the WTO General Council, the applicant state or customs territory
and members of the WP. The secretarial functions of the WP are assumed by one or
more officials of the WTO Accessions Division, a Division of the WTO Secretariat.
Upon formation of the WP, the WTO Secretariat will inform the applicant country
of the procedures to be followed, and especially about the details of the requirement
to submit its Memorandum on Foreign Trade Regime which covers issues ranging
from the design and the operation of the trade instruments employed by the acceding
country, to questions of a more general nature, such as those relating to overall eco-
nomic policy (with specific questions on the macroeconomic policies). The coverage
of the Memorandum in the WTO era covers trade in goods and services as well as the
protection of intellectual property rights covered by TRIPs (but also, occasionally,
examines intellectual property protection not covered by TRIPs).⁴
Upon submission of the Memorandum, the WTO Secretariat will check the
overall consistency of the submitted text with the outline format (that is, the for-
mat reproduced in WTO Doc WT/ACC/1). Although some differences between
submissions and the outline format are tolerated, major deviations are not. At the
same time, copies of the tariffs that are currently applicable by the applicant (in
the HS nomenclature) will be distributed. The WTO Secretariat will circulate the
Memorandum to the members of the WP. The members of the WP will then be
invited to submit questions to the applicant. Depending on the responses and the
trade interests involved, there could be more than one round of question and answer
sessions organized between the members of the WP and the applicant.
When the examination of the Memorandum is sufficiently advanced, members
of the WP will initiate bilateral market access negotiations with the applicant, ie
negotiations between the applicant and particular WTO members. The list of con-
cessions of the acceding country (or customs territory) will, upon conclusion of the
negotiations, be annexed to its Draft Protocol of Accession.⁵ A summary of the

⁴ The WTO Doc WT/ACC/1 of 24 March 1995 contains an outline format for the Memorandum
on Foreign Trade Regime. Although occasionally differences across applicants are observable, the
vast majority of its provisions are reproduced verbatim across national Memoranda.
⁵ For a very comprehensive description of the WTO relevant practice, from the moment when
a WP is established until the conclusion of the ensuing negotiations, see WTO Doc WT/ACC/1
of 24 March 1995. The amount of technical capacity required in order to complete the acces-
sion process should not be underestimated. As Gay (2005) explains, Vanuatu had to suspend its
accession process because of lack of technical expertise on issues under negotiation. Vanuatu came
back to complete the negotiation later.
 Participation 393

discussions between the WP and the applicant is reflected in the report of the WP
to the Ministerial Conference together with a Draft Decision and the Protocol of
Accession. The latter contains the terms of accession agreed by the applicant and
members of the WP. However, it is important to emphasize that in many instances
the Protocol of Accession contains cross-references to the report of the WP that
elaborate on issues of a substantive nature. As stated in Art XII.2 of the Agreement
Establishing the WTO, the report of the WP and the Draft Decision will need to
be approved by a two thirds majority for the Protocol of Accession to enter into
force (30 days after acceptance by the applicant).⁶ There is no time set ex ante for
the acceding country to implement the negotiated market access commitments. In
practice, in light of its own applicable legal procedures and constraints, the appli-
cant makes a proposal as to the period of time it envisages implementation will take
(the proposed periods of implementation vary between 6–9 months). There are no
reported cases where the deadline proposed has not been respected.
On 10 December 2002, the WTO General Council adopted a decision⁷ regard-
ing procedures for the accession of LLDCs to the WTO. During the GATT era
Contracting Parties could withdraw from the GATT. To do so, they had to respect a
formal requirement and send a written notice to this effect to the Secretary-General
of the United Nations (Art XXXI of the GATT). Their withdrawal took effect
6 months after the receipt of this written notice. Of course, GATT Contracting
Parties would, by virtue of Art 70 of the VCLT, be responsible for acts they com-
mitted while they were Contracting Parties to the GATT, even after they had with-
drawn. The full list of countries that withdrew from the GATT, as well as the date
when their withdrawal was made effective, is reproduced here (see Table 5.1):

Table 5.1 GATT Withdrawals

Republic of China 5 May 1950⁸

Lebanon 25 February 1951⁹
Liberia 13 June 1953¹⁰
Syria 6 August 1951¹¹

⁶ However, as noted above, in practice, accessions so far have taken place following consensus-
based decisions.
⁷ See WTO Doc WT/L/508 of 20 January 2003.
⁸ 55 UNTS 196. The Republic of China was an original CONTRACTING PARTY to the
GATT (21 May 1948). After its breakaway from the People’s Republic of China (‘China’), the
Republic of China, in a letter addressed to the GATT on 5 May 1950, denounced its status as a
CONTRACTING PARTY to the GATT, and formally withdrew from the GATT. The People’s
Republic of China never acknowledged the legitimacy of this action and its own (reduced)
participation in the GATT was left in limbo as a result. Th is is precisely why, when negotiations
to clarify the status of China under the GATT began again in the eighties, the WP established
to this effect (1987) carried the title People’s Republic China’s Status as a Contracting Party.
⁹ 77 UNTS 367.
¹⁰ 163 UNTS 375.
¹¹ 90 UNTS 324.
394 Institutional Provisions

During the late eighties, the Yugoslav Federation slowly began to disintegrate
and by the early nineties collapsed as the former ‘breakaway’ republics of the
Yugoslav Federation were widely being recognized as sovereign states on the
international scene. One of them, the Federal Republic of Yugoslavia, claimed
to be the successor of the Yugoslav Federation and requested to be acknowledged
as such and preserve its status as a Contracting Party to the GATT. The other
GATT Contracting Parties, however, ‘froze’ the participation of the Federal
Republic of Yugoslavia in the GATT through collective action:¹² the delegate of
the Federal Republic of Yugoslavia in effect did not participate in the meetings
of the GATT General Council. Later, with the advent of the WTO, the Federal
Republic of Yugoslavia was reconstituted (Serbia and Montenegro participated)
and it requested its accession anew to the WTO and a WP is currently in place to
this effect.¹³ Other ex-Yugoslav republics, such as Croatia, the Former Yugoslav
Republic of Macedonia (FYROM), and Slovenia, have since joined the WTO
following their breakaway from ex-Yugoslavia.
It is Art XIV of the WTO Agreement that now regulates the legal right of
WTO members to withdraw from the WTO. This provision recognizes the right
of a WTO member to unilaterally sever its contractual ties with the WTO; thus
far, no WTO member has done so.

2.3 Non-application
In order to facilitate accession to the GATT, Art XXXV of the GATT provided
flexibility for acceding countries to the GATT in allowing them not to enter
into contractual arrangements at all with some incumbent GATT Contracting
Parties:
1. This Agreement, or alternatively Article II of this Agreement, shall not apply as
between any contracting party and any other contracting party if:
(a) the two contracting parties have not entered into tariff negotiations with each other,
and
(b) either of the contracting parties, at the time either becomes a contracting party, does
not consent to such application.
2. The CONTRACTING PARTIES may review the operation of this Article in particu-
lar cases at the request of any contracting party and make appropriate recommendations.
At first glance, this provision strikes the reader as a legal curiosum: how could
it be that two states enter the same contractual arrangement and have no legal

¹² See GATT Doc C/M/257 of 10 July 1992.

¹³ Initially, only one WP was supposed to work on the accession of Serbia and Montenegro. At
the time, both had separate trade regimes and following a slow start, the WTO moved to establish
two separate WPs to discuss the accession of Serbia and Montenegro. Eventually, Montenegro
seceded (following a referendum in 2006) and is now a different state. Another WP was established
to discuss the accession of Serbia.
 Decision making 395

relationship between them? There is, nevertheless, a persuasive explanation for

this legal curiosum: the GATT was negotiated and concluded during the years
immediately following the Second World War. Many states emerged as enemies
from one of the saddest moments in world history. For many reasons, all states
were not prepared immediately to turn the page, and enter into contractual
arrangements with the rest of the world. GATT, Art XXXV was the means to,
on the one hand, induce participation in the GATT and, on the other, avoid
imposing on incumbents and acceding countries the ‘inconvenience’ of entering
into contractual arrangements with countries they preferred to ignore.
WTO Agreement, Art XIII now regulates the non-application of the WTO
agreements by and among members. The list of members that have invoked the non-
application clause as of the entry into force of the WTO is as follows (see Table 5.2):

Table 5.2 Invocations of the Non-application Clause

Accession Country Invoking Document Revocation

1 Armenia US WT/L/505 WT/L/601
2 Armenia Turkey WT/L/501
3 China El Salvador WT/L/429
4 Moldova US WT/L/395
5 Georgia US WT/L/318 WT/L/385
6 Kyrgyz Republic US WT/L/275 WT/L/363
7 Mongolia US WT/L/159 WT/L/306

3 Decision making

3.1 Consensus, the default rule

Decision making in the GATT was regulated in Art XXV. In two paragraphs (3
and 4), Art XXV of the GATT states that decisions will be taken on the basis of
a majority vote and that every Contracting Party had one vote at all meetings.
GATT, Art XXV.1 explains that the term ‘CONTRACTING PARTIES’ (in
capital letters) refers to the deciding organ in the GATT:
3. Each contracting party shall be entitled to have one vote at all meetings of the
CONTRACTING PARTIES.
4. Except as otherwise provided for in this Agreement, decisions of the CONTRACTING
PARTIES shall be taken by a majority of the votes cast.
In practice, however, GATT CONTRACTING PARTIES always took their
decisions by consensus.¹⁴ Ordinarily decisions taken by consensus on a matter

¹⁴ Hoekman and Kostecki (2001) offer some anecdotal evidence as to why requests to vote in
the General Council have been thwarted in general. It is probably the case that WTO members
fear that if they request voting in one area, other WTO members might start requesting voting in
another area in the future. Consensus is thus cemented in practice because of mutual deterrence.
396 Institutional Provisions

in a particular meeting would be presumed by the chair in the absence of any

party present objecting or calling for a vote.¹⁵ This is what is often referred to as
GATT pragmatism: without engaging in a detailed discussion of this issue,¹⁶ the
downside of GATT pragmatism has been the slowness in reacting to emerging
challenges, the almost obsessive reliance on incremental changes, and sometimes,
even the impotence to act at all; the upshot however, has been the enhanced
legitimacy of each and every GATT decision.
WTO Agreement, Art IX, which reflects the default decision-making process
used in the WTO,¹⁷ explicitly states that decisions will be taken following the trad-
itional GATT consensus-based approach. The WTO has not deviated from GATT
practice. Accordingly, recourse to voting procedures will occur only when consen-
sus has not been reached. A footnote to this provision makes it clear that consensus
will cease to exist only when a WTO member formally objects to a proposal that is
under consideration. On the other hand, Art IX of the WTO Agreement does not
even mention the term ‘quorum’. As a result, unless a specific provision to this effect
exists, there is no need for quorum-based decisions in the WTO.¹⁸
In the first 12 years of the WTO, there has been no deviation from the consen-
sus rule.¹⁹

3.2 Special procedures

3.2.1 Interpretations
WTO Agreement, Art IX.2 provides for the possibility to adopt interpretations
of the WTO Agreement. Since the WTO Agreement is largely an incomplete
contract,²⁰ WTO members might consider it necessary or useful in certain instances
to fill in the gaps of a provision that they deem is unclear. Gaps of course can be
filled through recourse to adjudication; this may not be an equally satisfactory

¹⁵ Hoekman and Kostecki (2001, pp 56ff ) anecdotally refer to a particular instance whereby a
request for a vote by one GATT Contracting Party was followed by a pause and then a return to
consensus procedures. On the powers of the chair more generally, see Odell (2005).
¹⁶ In light of the fact that there are a number of very instructive publications on this topic, see,
inter alia, Barfield (2001), and also the pioneering work of Hudec (1972).
¹⁷ That is, in the absence of specific rules governing decision making in a particular context.
¹⁸ A contrario, of course, whenever a quorum-based decision is provided for in a covered agree-
ment, it will also provide the parameters for quorum.
¹⁹ Informally, WTO members have voted (intermediary vote) during the election of Mike
Moore (the third Director-General (DG) of the WTO). Even on this occasion, however, the final
approval of his nomination was by consensus.
²⁰ The term ‘incomplete contract’ is used here as an acknowledgement that the ‘gaps’ of the con-
tract left unfi lled purposely or inadvertently by negotiators will be completed or ‘fi lled’ through
subsequent practice (and/or adjudication). There is no firewall between the two functions: it could
be the case that practice is inspired by adjudication and vice-versa. There are many examples from
practice in other legal regimes (eg the European Community case-law in the field of free movement
of goods rising as a very appropriate illustration) where the legislator has ‘crystallized’ in regula-
tory language jurisprudence which occurred between the original enactment of a statute and the
moment when it was amended/modified/completed.
 Decision making 397

alternative, however, since the outcome of the adjudication is transaction-specific.

On the other hand, recourse to an authentic interpretation may be appropriate if
WTO members wish to pre-empt the discretion of WTO panels (or the AB), or
even if they wish to undo a panel (or AB) finding. A three-quarters majority of
WTO members is required, according to Art IX.2 of the WTO Agreement, in
order to adopt an interpretation. This provision imposes a quorum, since it requires
the presence of at least three-quarters of the WTO membership at the meeting in
order for an interpretation to be adopted:
The Ministerial Conference and the General Council shall have the exclusive authority
to adopt interpretations of this Agreement and of the Multilateral Trade Agreements. In
the case of an interpretation of a Multilateral Trade Agreement in Annex 1, they shall
exercise their authority on the basis of a recommendation by the Council overseeing the
functioning of that Agreement. The decision to adopt an interpretation shall be taken by
a three-fourths majority of the Members. This paragraph shall not be used in a manner
that would undermine the amendment provisions in Article X.
As of August 2006, no interpretation had been adopted.

3.2.2 Amendment of the GATT

The GATT does not distinguish between amendments and modifications, as the
VCLT does, it regulates amendments only. This should mean that formal modi-
fications are not an institutional possibility in the GATT.²¹ GATT, Art XXX
states that for an amendment to enter into force, all GATT Contracting Parties
will have to accept it. WTO Agreement, Art X regulates amendments to the
GATT (and to other covered agreements as well), and has substituted Art XXX
of the GATT.²²
WTO Agreement, Art X explains the specific procedure to be followed when-
ever a proposal to amend the WTO Agreement is tabled. In essence, if consensus
is not possible, a qualified majority of the Ministerial Conference (two-thirds
of the WTO members²³) is required for a proposal to be submitted by the
Ministerial Conference to WTO members for their approval. Amendments to
five specific provisions set forth in Art X.2 require unanimous acceptance of the
WTO membership:
2. Amendments to the provisions of this Article and to the provisions of the following
Articles shall take effect only upon acceptance by all Members:

²¹ This means that, say, country A cannot modify a provision of the GATT vis-à-vis coun-
try B only. It will have to request an amendment that will, in principle, bind all WTO members.
Modification thus is a means to enhance contract flexibility, whereas amendments are a rather inflex-
ible mechanism, since their advent depends by the will of at least two-thirds of the WTO members.
In a way, the distinction between plurilateral and multilateral agreements has occupied this place.
²² Although Art XXX of the GATT is technically still in force, it is expected, if at all, to serve as
the basis, additional to Art X of the Agreement Establishing the WTO), to amend the GATT.
²³ That is, not of the members present for the vote, but of the total membership. Hence, once
again we are in the presence of a quorum-requirement.
398 Institutional Provisions
(a) Article IX of this Agreement;
(b) Articles I and II of GATT 1994;
(c) Article II:1 of GATS;
(d) Article 4 of the Agreement on TRIPS.
On 6 December 2005, the WTO General Council adopted the first, and only so
far, amendment of the WTO Agreement. The amendment reads as follows:
The Agreement on Trade-Related Aspects of Intellectual Property Rights (the “TRIPS
Agreement”) shall, upon the entry into force of the Protocol pursuant to paragraph 4, be
amended as set out in the Annex to this Protocol, by inserting Article 31bis after Article
31 and by inserting the Annex to the TRIPS Agreement after Article 73.
Reservations may not be entered in respect of any of the provisions of this Protocol
without the consent of the other Members.
This Protocol shall be open for acceptance by Members until 1 December 2007 or
such later date as may be decided by the Ministerial Conference.
This Protocol shall enter into force in accordance with paragraph 3 of Article X of the
WTO Agreement.
This Protocol shall be deposited with the Director-General of the World Trade
Organization who shall promptly furnish to each Member a certified copy thereof and a
notification of each acceptance thereof pursuant to paragraph 3.
This Protocol shall be registered in accordance with the provisions of Article 102 of the
Charter of the United Nations.²⁴

4 Dispute Settlement

4.1 The early GATT years

GATT, Arts XXII and XXIII establish a two-phase system, whereby Contracting
Parties would first go through a bilateral consultation phase aimed at resolving
their dispute, and if this was not possible, the complainant could refer the matter
to the CONTRACTING PARTIES. There was nothing like a panel procedure
explicitly established in the two aforementioned GATT provisions. The panel
procedure is thus an invention of the GATT CONTRACTING PARTIES.
It all began with a complaint in the summer of 1948 by the Netherlands against
Cuba, which presented the question as to whether the MFN obligation (Art I of the
GATT) also applied to consular taxes.²⁵ The matter was referred to the chairman,²⁶

²⁴ See WTO Doc WT/L/641 of 8 December 2005. Through the new Art 31bis, WTO members
could outsource production of goods coming under compulsory licensing.
²⁵ This narrative about the GATT years borrows heavily from Palmeter and Mavroidis (2004).
²⁶ Hart (1998) mentions that ‘the Havana Conference in its closing days established the Interim
Commission for International Trade Organization (ICITO), composed of most members of the
conference, to prepare a program for the ITO when it came into being, and to carry out related
functions. Canada’s Dana Wilgress was elected its first Chairman. The Commission appointed as
its executive secretary, Eric Wyndham-White’ (pp 51–2).
 Dispute Settlement 399

who ruled that Art I of the GATT effectively applied to consular taxes. Rulings by
the chairman were subsequently substituted by a WP procedure, whereby the com-
plaining party, the defendant, and any other Contracting Party that might have an
interest in the dispute, would participate. The GATT procedural rules were very
limited, largely because it was anticipated that the ITO rules would soon come into
force. The ITO Charter contained detailed dispute settlement rules that included a
provision for the referral of questions to the ICJ.
The next mutation of the procedure was to eliminate the complainant and
defendant from the adjudicating panel and instead assign three or five ‘neu-
tral’ panelists to prepare a report that would eventually be submitted to the
CONTRACTING PARTIES for their consideration and adoption. At its early
stages, dispute settlement in GATT reflected its diplomatic roots: the majority of
the panelists were diplomats. No formal requirement to have legal training, let
alone GATT-related legal training, was imposed. Both Hudec (1993a) and Davey
(1987) have observed that the goal of the process was to end up with a mutually
acceptable solution, rather than to judge on the legal merits of the case—the
overall tenor of the process was highly conciliatory.
GATT, Arts XXII.1 and XXIII.1 have a parallel function:²⁷ the former requires
each Contracting Party to afford other parties an adequate opportunity for con-
sultation with respect to any matter affecting the operation of the agreement; the
latter provides that, if any Contracting Party considers that a benefit, accruing to
it directly or indirectly under the agreement, was being nullified or impaired by
another party, it could make written representations or proposals to that other
party. While both provisions regulate the consultations phase, the former has a
wider coverage than the latter.
If consultations could not lead to a satisfactory adjustment, the complain-
ant, assuming the defendant was in agreement, could refer the matter to a panel.
The panel would not decide the issue; it would simply submit a report which,
unless adopted by the CONTRACTING PARTIES, would be of limited legal
value. Assuming again that the defendant consented to the adoption of the report,
the defendant would be requested to bring its measures into compliance with its
GATT obligations. If implementation did not occur, the complainant could be
authorized to impose countermeasures, provided:
(1) that it made a request to the CONTRACTING PARTIES to this effect,
and
(2) the CONTRACTING PARTIES by consensus (eg with the defendant’s
agreement) would acquiesce.

²⁷ An early decision by the CONTRACTING PARTIES held that consultations under Art
XXII.1 of the GATT would be considered as fulfi lling the consultation requirements of Art
XXIII.1 of the GATT, see GATT Doc BISD 9S/20.
400 Institutional Provisions

Countermeasures in the GATT context amounted to an authorization to the

complainant to suspend tariff concessions or other GATT obligations to the
party found acting inconsistent with its obligations under the agreement.²⁸

4.2 The late GATT years

In 1979, during the Tokyo round, the trading partners decided, inter alia, to
take stock of the evolution in dispute settlement procedures. The adoption of
the Understanding on Notification, Consultation, Dispute Settlement and
Surveillance of 28 November 1979, which included an annex setting out an
Agreed Description of the Customary Practice of the GATT in the Field of
Dispute Settlement, signals the first step towards providing some institutional
recognition in this field. The relevant part of the Description reads:
Panels set up their own working procedures. The practice for the panels has been to hold
two or three formal meetings with the parties concerned. The panel invited the parties
to present their views either in writing and/or orally in the presence of each other. The
panel can question both parties on any matter which it considers relevant to the dispute.
Panels have also heard the views of any contracting party having a substantial interest
in the matter, which is not directly party to the dispute, but which has expressed in the
Council a desire to present its views. Written memoranda submitted to the panel have
been considered confidential, but are made available to the parties to the dispute. Panels
often consult with and seek information from any relevant source they deem appropriate
and they sometimes consult experts to obtain their technical opinion on certain aspects
of the matter. Panels may seek advice or assistance from the secretariat in its capacity as
guardian of the General Agreement, especially on historical or procedural aspects. The
secretariat provides the secretary and technical services for panels.²⁹
From there onwards, we observe incremental steps towards greater formalism: it
was decided in 1982 that:
The contracting party to which such a recommendation [ie to bring a challenged measure
into conformity with GATT] has been addressed, shall report within a reasonable speci-
fied period on action taken or on its reasons for not implementing the recommendation
or ruling by the Contracting Parties.³⁰
The overall culmination of the preceding developments comes with the so-called
Montreal Rules that grew out of a Ministerial meeting in Montreal (December
1988), held to review the progress of the Uruguay round and to reap an ‘early har-
vest’ of any results accomplished during the round (in GATT parlance, a ‘mid-term
review’). Following further consideration in early 1989 in Geneva, the Montreal

²⁸ In the only instance where such an authorization occurred in the GATT era, The Netherlands
was authorized to suspend tariff commitments vis-à-vis the United States. The Netherlands never
acted upon its right to do so.
²⁹ See GAT Doc BISD 26S/215.
³⁰ See GATT Doc BISD 29S/13.
 Dispute Settlement 401

Rules were adopted by the CONTRACTING PARTIES in April 1989.³¹ To

a great extent, the Montreal Rules formed the basis of what eventually became
the DSU.
The CONTRACTING PARTIES agreed to apply the Montreal Rules:
on a trial basis from 1 May 1989 to the end of the Uruguay Round in respect of com-
plaints brought during that period under Article XXII or XXIII.³²
Time-limits were imposed on consultations and the right to initiate panel pro-
ceedings was established (the defendant could no longer ‘block’ such a request
when it appeared on the agenda for the second time). Parties receiving a request
for consultations would be required to reply to that request within 10 days, and
agree to enter into consultations in good faith in no less than 30 days. In the
absence of an agreement to consult, the complainant could proceed directly to
request the establishment of a panel. If consultations failed to settle the dispute
within 60 days of the request, the complainant could request the establishment
of a panel.³³
The system performed admirably well when compared to other international
regimes. Although the institutional possibility to block the establishment of a panel
existed for over 40 years (1947–89), little use was made of it. Dispute settlement in
the GATT is a truly unique paradigm in international relations. (Hudec, 1993a)
shows that Contracting Parties to the GATT de facto behaved as if compulsory
third party adjudication had been agreed among them.³⁴ His study demonstrates
that from 1947 to 1992, with one exception, panels were always established, and
the losing party eventually accepted the results and conceded the adoption of an
adverse panel report in approximately 90% of all cases.³⁵ The record would be
even more impressive, absent some highly political disputes in the late eighties
early nineties (essentially between the ‘big two’, the European Community and
the United States, which occasionally reverted to dispute adjudication to obtain
institutional recognition of positions that they could not successfully advance at
the negotiating table). We have little information on the actual implementation
of GATT rulings. For a number of reasons having to do with the lack of adequate
notification procedures in the GATT system, we simply cannot accurately tell
what the end-result was in each and every dispute. Monitoring procedures were, to

³¹ Improvements to the GATT Dispute Settlement Rules and Procedures, Decision of 12 April
1989, GATT Doc BISD 36S/61.
³² Idem.
³³ Benedek (1991) has taken the view that by that time a customary right to a panel had emerged.
Even if this view is probably over-optimistic, it does reflect the outcome of the empirical analysis
conducted by Hudec (1993a) discussed in what immediately follows.
³⁴ Hudec’s study did not answer the question of how many cases were not submitted because of
an anticipated negative response by the (eventual) defendant. Such a study is difficult to perform,
of course. Backwards induction might (at least, intuitively) suggest that, expecting a negative
response, potential complainants might have felt unwilling to start the process altogether.
³⁵ Adoption of the report does not necessarily mean implementation, as noted below.
402 Institutional Provisions

a large extent, dependent on the diligence and the incentive of interested parties.
Recourse to proxies is thus inevitable under the circumstances. This is, neverthe-
less, a rather imperfect exercise. For example, one could reasonably assume that
the absence of legal challenges on the same factual issues indicates satisfaction
with the implementation that was undertaken; the lack of challenge could, how-
ever, hide strategic interaction between the players involved (eg A does not chal-
lenge B expecting that B will do the same), or, crudely, non-transparent bilateral
deals.³⁶ Hudec (1993a) reveals some data to this effect.
Within these limits, the GATT dispute settlement system performed well by
any reasonable standard. To an extent, the GATT’s conciliatory and incremental
approach to legalization paved the way to the DSU as we know it today. Hudec’s
study brought home the fact that an empirical analysis of dispute settlement did
not concur with the oft-heard voices of discontent as to the effectiveness of the
system. The DSU—the crown jewel of the WTO system—was largely shaped
between 1947 and 1995 and did not emerge out of the blue the night before the
eventual agreement on the WTO’s inception.

4.3 The WTO DSU³⁷

The DSU is one of many annexes to the agreement establishing the WTO. The
administration of disputes is entrusted to the DSB, where representatives of all
WTO members participate. The system is highly decentralized: disputes cannot
be initiated ex officio; there is no authority assigned to a supra-national entity
(a watchdog) to initiate complaints against WTO members; and disputes are
launched at the initiative of a WTO member.
Adjudication in the WTO system, as in GATT, has two phases: one, in prin-
ciple, bilateral, and one multilateral. The bilateral phase consists of consultations
between the complainant and the defendant. Few disputes are of a purely bilateral
nature however, and even if this is the case, other WTO members might have an
interest in the interpretation of the rules pertinent to the particular transaction at
hand, since, arguably, such interpretation might be influential in interpreting (or
construing) their own commitments in the future. To this effect, when request-
ing consultations, the complainant has to notify the WTO as to the subject-mat-
ter of the dispute. Other WTO members wishing to join as co-complainants can
do so, provided that the defendant accepts their request (Art 4.11 of the DSU).³⁸

³⁶ One could, of course, adopt a very wide understanding of the term ‘implementation’, wide
enough to encompass bilateral deals to mutually avoid prosecution of GATT inconsistent practices.
This is not, nevertheless, the manner in which the term ‘implementation’ is used in legal literature.
³⁷ This section is based on Horn and Mavroidis (2007).
³⁸ Commentators often disregard this phase. It is, however, quite important in many ways. For
lawyers this could be an invaluable source of WTO practice. While it is true that settlements are
not always notified in a user-friendly manner, this is not reason enough though to stop research
in this area. Busch and Reinhardt (2001) are two of the few authors with substantial work in this
research area.
 Dispute Settlement 403

The subject-matter of a particular dispute can range from disagreement over

a particular transaction and its consistency with the relevant WTO law (eg
A believes that B imposed AD duties without having demonstrated any injury
resulting from dumped imports), to disagreements over the consistency of a piece
of legislation with the WTO rules (eg A believes that B, by enacting legislation
which precludes its investigating authorities from conducting injury analysis in
the context of an AD investigation, is violating its obligations under the WTO).
The standard of review, however, is more demanding in the latter case.
Assuming that the parties reach no solution during the consultation stage, the
complainant can request the establishment of a panel to adjudicate the dispute.
Such a request leads to the second (multilateral) phase, which consists of two
parts: the first involves the panel procedure, the panel being analogous to a first
instance court; the second part is the procedure before the AB, the court of final
instance. Whereas the three-person panels are ad hoc adjudicating bodies, the
composition of which depends, in principle, on the discretion of the parties to the
dispute (and, in case of disagreement between them, panelists will be appointed
by the Director-General of the WTO),³⁹ the AB is composed of seven judges
appointed on 4-year terms, renewable once. Panels have competence to review
the factual record and the legal issues before them, whereas the AB’s remit/
purview is limited to the latter.⁴⁰
Following a request for establishment of a panel, other WTO members have
a limited time within which they can request to appear before the panel as third
parties and present their views on a particular case. Panels are assisted by mem-
bers of the WTO Secretariat (usually lawyers, and occasionally economists as
well). The role of the Secretariat should not be underestimated: it normally drafts
the reports, even though panelists, of course, have the last word. Importantly,
with the exception of Secretariat members working for the AB, members of the
Secretariat, besides advising the members of the panel they are assigned to, also
advise WTO members on all issues of WTO law. They are thus not assigned to
exclusively service panels.
Assuming that the AB accepts the original complaint,⁴¹ the defendant will
be requested to implement the judgment. Implementation should, if possible,
occur immediately, although this is hardly ever the case. Instead, defendants are
granted an implementation period, that is either agreed bilaterally, or through
recourse to binding arbitration (Art 21 of the DSU).

³⁹ Horn and Mavroidis (2006) have estimated that in 78/139 panels (up to February 2005),
the Director-General had appointed at least one member of the panel. Selection of panelists has
become problematic and this is why voices are increasing in favour of appointing permanent pan-
elists, see on this score Cottier (2004), and Ehlermann (2004).
⁴⁰ However, the distinction between a factual and a legal issue is often overstated in legal litera-
ture, since many disagreements about facts can be formulated in terms of a legal issue.
⁴¹ If the AB rejects the original complaint, the case will, of course, be closed.
404 Institutional Provisions

Defendants do not always have sufficient guidance to implement a judgment

against them. The lack of guidance is a direct consequence of adjudicating body
reports that merely recommend that the losing respondent to bring its measures
into compliance with its obligations. In the unusual instance where adjudicating
body reports suggest ways to implement the final judgment, the addressees of
suggestions are nevertheless free to choose their way of complying, since sugges-
tions are not legally binding. As a result, disagreements as to whether implemen-
tation has truly occurred can easily arise. In such a case, the dispute is referred to a
compliance panel whose task it is to determine whether implementation occurred
(Art 21.5 of the DSU); the panel’s report can be appealed.
If the complainant believes that no implementation has occurred, they can
request authorization to suspend concessions. If granted, the complainant will
have the right to raise its bound duties to the level necessary to inflict on the
defendant damage equal in value to the damage the complainant suffered as a
result of the practice that was found to be illegal.⁴²
The entire adjudication process can, in practice, take up to 3 years or even
longer. Nothing, however, prohibits WTO members from reaching a MAS at
any stage throughout the process. A MAS must be notified to the DSB, where
any WTO member can raise questions as to its consistency with the WTO rules.

4.4 The WTO: exclusive forum for adjudication of trade disputes

DSU, Art 23.2 (a) reads:⁴³
1. When Members seek the redress of a violation of obligations or other nullification or
impairment of benefits under the covered agreements or an impediment to the attain-
ment of any objective of the covered agreements, they shall have recourse to, and abide
by, the rules and procedures of this Understanding.
2. In such cases, Members shall:
(a) not make a determination to the effect that a violation has occurred, that ben-
efits have been nullified or impaired or that the attainment of any objective of
the covered agreements has been impeded, except through recourse to dispute set-
tlement in accordance with the rules and procedures of this Understanding, and shall
make any such determination consistent with the findings contained in the panel
or AB report adopted by the DSB or an arbitration award rendered under this
Understanding. (emphasis added)
The AB provided its understanding of this obligation in US—Certain EC
Products (§ 111):
Article 23.1 of the DSU imposes a general obligation on Members to redress a viola-
tion of obligations or other nullification or impairment of benefits under the covered
agreements only by recourse to the rules and procedures of the DSU, and not through
unilateral action. Subparagraphs (a), (b) and (c) of Article 23.2 articulate specific and

⁴² See on this score, Lawrence (2003).

⁴³ On this issue, see Bourgeois (2005).
 Dispute Settlement 405
clearly-defined forms of prohibited unilateral action contrary to Article 23.1 of the DSU.
There is a close relationship between the obligations set out in paragraphs 1 and 2 of
Article 23. They all concern the obligation of Members of the WTO not to have recourse
to unilateral action. (. . .)
The panel report on EC—Commercial Vessels added that Art 23.2(a) of the DSU
serves a double purpose:
to ensure that WTO Members will seek to redress violations of the WTO Agreement
through recourse to the DSU procedures and will thus avoid unilateral action: and
that WTO Members will use exclusively the DSU procedures when trying to redress
violations of the WTO Agreement and will not have recourse to the dispute settlement
procedures of other fora. Importantly, the panel, through this second element, adds to
prior case-law dealing with the interpretation of Art. 23 DSU. (§§ 7.194–7.196)
De facto, the GATT was an exclusive forum as well, and, from this angle Art
23.2(a) of the DSU should not be viewed as an innovation: as Jackson (1969) notes,
although the possibility to submit trade-related disputes to the ICJ was in prin-
ciple open to GATT Contracting Parties, no use of it was ever made. DSU, Art
23.2(a) thus eliminated this theoretical possibility. On the other hand, one may,
in light of recent practice, legitimately ask the question whether Art 23.2(a) of the
DSU should continue to be understood in the manner elaborated above. The DSU
provides for three different adjudication avenues: the normal panel/AB proceed-
ings; mediation, under Art 5 of the DSU; arbitration, under Art 25 of the DSU.⁴⁴
On two occasions, however, WTO members opted for resolution of their disputes
under an ad hoc procedure different to the three formal procedures described
above.⁴⁵ Although, officially, these two procedures constitute a deviation from Art
23.2(a) of the DSU, they are nevertheless in line with the spirit of this provision,
since parties to the dispute enlisted third party adjudication (albeit of an ad hoc
nature).
The interaction between Art 23.2(a) of the DSU and choice of forum rules in
preferential trade agreements such as NAFTA is complex: for example, NAFTA
provides that where certain kinds of disputes (for instance, environmental disputes)

⁴⁴ Mediation is a confidential process, the WTO Director-General being competent to act as

mediator. Arbitration, under Art 25 of the DSU, is closer to the usual proceedings: parties will
choose the Arbitrator, and the latter’s report will be published. It is supposed to be a rather exped-
itious process to be used when quick solutions are sought.
⁴⁵ The Tuna dispute between Philippines, Thailand and the European Community and the
Bananas dispute between a host of banana-exporting countries and the European Community.
The only official evidence for the Tuna dispute is in the request for mediation between Philippines,
Thailand, and the European Community (Communication from the Director-General, WTO
Docs WT/GC/66 of 16 October 2002 and WT/GC/71 of 1 August 2003). The final report was
not made public. The Arbitrator was a WTO Deputy Director-General (DDG). The story of the
Bananas arbitration is different. The parties to the dispute made it clear that the arbitration sought,
to examine whether the EC proposals with respect to the import treatment of bananas would be
WTO consistent, was not an Art 5 of the DSU mediation. Still the award of the Arbitrator was
circulated as a WTO Doc, albeit not in the DS series, as is usually the case with dispute settlement
documents, see WTO Doc WT/L/616 of 1 August 2005.
406 Institutional Provisions

between NAFTA parties could be subject to both NAFTA and WTO proceedings,
the complainant is required to use NAFTA facilities exclusively. So far, WTO case-
law has not had to address this issue on head on: in one dispute so far, a WTO panel
rejected the argument that a WTO member (which was also a MERCOSUR⁴⁶
member) was estopped from raising a legal challenge against a preferential partner,
simply because a MERCOSUR court was dealing with this issue.

4.5 A closer look into procedures⁴⁷

4.5.1 Request for establishment of a panel
The process before the panel will kick off with the submission of a request for
establishment of a panel (Art 6.2 of the DSU). The complainant⁴⁸ must include
all its claims in the request; claims cannot be added at a subsequent stage (out of
due process considerations), arguments supporting claims however, can be raised
even at the second meeting before the panel. Panels⁴⁹ will typically organize two
meetings with the parties before drafting their interim, and eventually the final,
report. WTO Members can participate as third parties (Art 10 of the DSU)⁵⁰
before a panel and, through their submissions, side with either the complainant
or the defendant. However, there is no obligation bestowed upon panels to enter-
tain their submissions.⁵¹

⁴⁶ MERCOSUR stands for the CU concluded between Argentina, Brazil, Paraguay, and
Uruguay.
⁴⁷ On this score, see the various contributions in Weiss (2000).
⁴⁸ Only WTO members can act as complainants. Private parties’ interests will be represented
before the WTO under the conditions established in domestic constitutional law. The United States
and the European Community have enacted domestic statutes which clarify the conditions under
which they will accept to represent private claims (Section 301, and the Trade Barriers Regulation,
respectively). Levy and Srinivasan (1996) make a good argument in favour of assigning standing to
governments only and persuasively dismiss the arguments presented by advocates of direct effect.
On Section 301, see Abbott (1996). Shaffer (2003) offers a very comprehensive study regarding the
defence of private interests before the WTO.
⁴⁹ In the GATT years there are reported cases of panels comprising five panelists. Since the
advent of the WTO a panel has always constituted of three panelists (on the nationality of panelists
and how many of them are repeat players, see Horn and Mavroidis (2006)). Panels are assisted by
members of the WTO Secretariat; the role of the latter is quite influential, although formally, by
virtue of the DSU, it should be limited to assisting parties in the selection process of panelists, see
Nordtström (2005b).
⁵⁰ Busch and Reinhardt (2005) advance the argument and provide empirical support to the
effect that third parties’ participation reduces the probability of reaching a settlement.
⁵¹ WTO members, following the AB report on EC—Sardines, and non-state entities (non-gov-
ernmental organizations (NGOs), but also individuals) can also participate as amici curiae. In that
case, however, WTO panels do not have to accept their briefs. Whereas briefs by third parties will
be reflected in panel reports (panels, however, retain discretion as to whether to entertain them),
submissions by amici might be totally neglected. By the same token, whereas third parties have the
right to participate at the first panel meeting (and exceptionally, at the second as well), amici have
no such right; their participation depends entirely on the discretion of panels, see on this issue
Mavroidis (2004a).
 Dispute Settlement 407

4.5.2 Burden of proof

The procedures before panels and the AB are based on two legal maxims: actori
incumbit probatio (the party making a claim carries the burden to prove it) and
jura novit curia (parties can presume that the court knows the applicable law,
the consequence being that they can limit their efforts to proving facts).⁵² The
jura novit curia maxim⁵³ should not be taken too seriously: parties to a dispute
will, in practice, not only present evidence with respect to the facts the truth
of which they assert, but also with respect to what is, in their view, the proper,
understanding of the legal regime.⁵⁴ The party carrying the burden of proof
has to make a prima facie case that its claim holds. The requirement imposed
on the party carrying the burden of proof to make a prima facie case is not
reflected in the DSU. It is judge-made law, and was first introduced by the AB
in its report on US—Wool Shirts and Blouses. It has been cited in practically all
disputes ever since (§ 14):
It is a generally accepted canon of evidence in civil law, common law, and, in fact, most
jurisdictions, that the burden of proof rests upon the party, whether complaining or
defending, who asserts the affirmative of a particular claim or defence. If that party
adduces sufficient evidence to raise a presumption that what is claimed is true, the burden
then shifts to the other party, who will fail unless it adduces sufficient evidence to rebut
the presumption.
In addition to its standpoint on the burden of proof, this passage provides an
indication as to the burden of persuasion:⁵⁵ the party making a claim must prove
it; assuming it has made a prima facie case the burden of proof will shift to the
other party. Making a prima facie case is shorthand for the burden of persuasion.
Unfortunately, we do not know exactly what this term means. In subsequent
(to US—Wool Shirts and Blouses) cases, panels have used the term ‘to make a
prima facie case’ as equivalent to the obligation to ‘raise a presumption’ that what
is claimed is true (US—Stainless Steel, § 6.2).⁵⁶ Raising a presumption is not a
self-interpreting term either and, as a result, the question of the quantum of proof

⁵² This does not mean that they avoid references to the appropriate legal provisions. Indeed,
as the AB case-law has made it amply clear, claims are distinguished on the basis of their factual
as well as their legal components (the legal component clarifies which provision of the WTO a
particular factual situation runs counter to). On this issue, see the excellent analysis by Pauwelyn
(1998).
⁵³ As such, this maxim is not explicitly stated in the DSU. In an oblique manner it is reflected in
Art 8 of the DSU which describes the function of panels.
⁵⁴ Sometimes the distinction between a factual and a legal issue is not crystal clear, see Bronckers
and McNelis (2000).
⁵⁵ See on this issue, the excellent analysis in Sanchirico (1997).
⁵⁶ The panel on Mexico—Taxes on Soft Drinks ruled that the duty of the complainant to make
a prima facie case is not affected by the defendant’s decision not to challenge the claims and argu-
ments made. In the case at hand, Mexico had chosen not to raise any defence in some of the claims
advanced by the United States. The panel implicitly held that Mexico’s inaction did not amount to
admission that the United States had made a prima facie case (§§ 8.16ff ).
408 Institutional Provisions

needed to discharge the obligation of burden of persuasion can only be evaluated

on a case-by-case basis in WTO jurisprudence.⁵⁷

4.5.3 Discovery powers

DSU, Art 13 allows panels to seek information from any source they deem
appropriate:⁵⁸
Each panel shall have the right to seek information and technical advice from any indi-
vidual or body which it deems appropriate. However, before a panel seeks such information
or advice from any individual or body within the jurisdiction of a Member it shall inform
the authorities of that Member. A Member should respond promptly and fully to any
request by a panel for such information as the panel considers necessary and appropriate.
Confidential information which is provided shall not be revealed without formal author-
ization from the individual, body, or authorities of the Member providing the informa-
tion. (emphasis added)
DSU, Art 13.2 specifies that this right includes the possibility for panels to
request information from expert review groups, established in accordance with
the procedures enshrined in Annex 4. So far, no recourse to this procedure has
been recorded. However, recourse to individual experts has occurred, especially
when panels are confronted with conflicting scientific expertise submitted by the
parties to a dispute. This has been especially the case when panels are called to
adjudicate disputes in the SPS-context.
Appendix 3 of the DSU in § 8 makes it clear that panels can address questions
to the parties at any point in time. In practice, a questionnaire will customarily be
addressed to the disputing parties both during and following the first and second
meetings. Although oral responses are furnished, panels also insist on receiving
written responses. Responses to the questions constitute part of the record.⁵⁹ Panels
lack the legal powers to compel disclosure of information from the parties; they

⁵⁷ In private conversations, Horn uses a tennis story to explain burden of persuasion: a party
(player) has assumed its burden of persuasion any time the ball goes over the net. In the real world
we only know whether this indeed has been the case ex post facto. WTO panels will let panelists
argue irrespective of whether they have been persuaded by the arguments made. They will review
the record in toto once arguments have been exhausted.
⁵⁸ Note that Art 13 of the DSU is drafted in a manner that places the initiative to call upon
experts on the panel itself. It could, however, be the case that a party to a dispute requests from a
panel to exercise its right. Th is was indeed the case in EC—Selected Customs Matters. The panel
refused to do so since the United States stated that calling upon experts was not a necessary step
for the panel to decide on the case before it. The panel, hence, declined to exercise its discretion in
this regard (§§ 7.81–7.83). SPS, Art. 11.2 contains a similar SPS-specific provision. Such a pro-
vision is clearly unnecessary in light of Art 13 of the DSU (which is applicable across all covered
agreements). The inclusion of Art 11.2 of the SPS probably serves as an encouragement to SPS
panels to search for outside expertise, deemed necessary in a predominantly technical field. In
EC—Approval and Marketing of Biotech Products, the panel called upon experts that had not met
the prior approval of the parties to the dispute (§§ 7.23–7.27).
⁵⁹ In practice, parties to a dispute can address questions to each other through the chair. The
responses will form part of the record as well. However, this is a right acknowledged to parties and
not part of a panel’s discovery powers, see Palmeter and Mavroidis (2004).
 Dispute Settlement 409

can, however, draw negative inferences where a party has refused to provide certain
information. This issue arose in the dispute between Brazil and Canada relating to
export subsidies paid to the aircraft industry (Canada—Aircraft). Canada refused
to provide information sought by the panel, and the panel drew inferences from the
lack of cooperation. The AB was called to decide whether in such a case a panel can
legitimately draw inferences. The relevant passage reads (§ 203):
Clearly, in our view, the Panel had the legal authority and the discretion to draw infer-
ences from the facts before it—including the fact that Canada had refused to provide
information sought by the Panel.
The panel justified its decision pointing to the duty to cooperate that all WTO
members incur by virtue of Art 13.1 of the DSU. The duty to cooperate imposes,
inter alia, the duty to respond in good faith to questions asked. In the panel’s
view, a WTO member which does not respect its duty to cooperate can legitim-
ately face the consequence of the panel drawing inferences based on the lack of
cooperation (Canada—Aircraft, § 187). There are limits, imposed by case-law,
on a panel’s discovery powers. The AB, in its report on India—Patents (US)
explained that, when having recourse to its discovery powers, a panel must
ensure that it does so within the context of the claims presented to a WTO
adjudicating body. A panel cannot use its discovery powers to alter the claims
submitted to it (§ 94).
The next question is to what extent panels, by respecting the ambit of claims as
submitted to them by the complaining party, can, through use of their discovery
powers, form a view on the claim using arguments not advanced by the parties.
The AB, in its report on Japan—Agricultural Products II addressed, inter alia,
this issue: the panel faced a claim by the United States that the Japanese legisla-
tion at hand was unnecessarily restrictive, and thus in violation of Art 5.6 of the
SPS. To support its claim, the United States advanced the argument that Japan
could have used one particular method, which was arguably less restrictive than
that chosen, and still reach its stated objective. The panel was not persuaded by
the US argument in this respect and requested expert advice. Experts demon-
strated that there was another method (not suggested by the United States) which
was clearly less restrictive than, and as effective as, the method used by Japan. The
panel, on these grounds, moved on to find that Japan had violated its obligations
under Art 5.6 of the SPS. The AB reversed the panel’s findings in this respect,
since in its view, the panel in casu could not have based its decision on the method
suggested by the experts, precisely because the United States had not suggested
this method in its submission (§§ 123–6 and 129–30). The AB ‘straightened the
record’ in its report on EC–Hormones (US) where it took the view that confin-
ing an adjudicating body to what has been argued (in terms of arguments) by the
parties to the dispute, might put into question its ability to respect its legal duty
to perform an objective assessment of the matter before it, as required by Art 11
of the DSU. In this line of thinking, panels should be free to develop their own
410 Institutional Provisions

reasoning, independently of the arguments advanced by the parties but within

the ambit of the claims submitted to them (§ 156).

4.5.4 Standard of review

Panels are required by virtue of Art 11 of the DSU⁶⁰ to perform an objective exam-
ination of the matter before them:
The function of panels is to assist the DSB in discharging its responsibilities under this
Understanding and the covered agreements. Accordingly, a panel should make an object-
ive assessment of the matter before it, including an objective assessment of the facts of
the case and the applicability of and conformity with the relevant covered agreements,
and make such other findings as will assist the DSB in making the recommendations
or in giving the rulings provided for in the covered agreements. Panels should consult
regularly with the parties to the dispute and give them adequate opportunity to develop a
mutually satisfactory solution.
What exactly the duty to perform an objective examination entails has been the
subject-matter of numerous cases submitted to the WTO.⁶¹ In a nutshell, panels
should not be performing a de novo review, but should abstain from total deference
as well.⁶² The term de novo review has traditionally been used in GATT/WTO
practice to indicate the limits of judicial review by panels. The term was first used in
the context of contingent protection instruments (AD, countervailing, safeguards)
to indicate that panels should abstain from performing de novo the whole investi-
gation, but rather, should check whether, based on the facts before it, an investigat-
ing authority reached a reasonable conclusion. With its report on EC—Hormones,
the AB extends the ‘no de novo review—but no total deference either’ standard⁶³
to disputes outside the realm of contingent protection. In the case at hand, the
question revolved around the appropriate standard of review to be applied in a dis-
pute involving a determination by the competent EC authority to block imports
of hormone-treated beef on, allegedly, health grounds (§ 117). This does not mean
that the panels’ role as trier of facts is reduced to redundancy. On the contrary: not
performing a de novo review means that panels, assuming they are satisfied with the
factual basis on which a decision by a domestic investigating authority rests, should
not re-open the whole investigation and should limit themselves to determining
whether the ultimate finding by the domestic authority is a reasonable one. Panels,
however, cannot use facts which occur subsequent to the investigation (being initi-
ated). They should place themselves in the shoes of the national authority at the
moment in time it came to its decision and ask whether the authority:

⁶⁰ On the standard of review issue, in general, see the very comprehensive analysis in Oesch
(2003).
⁶¹ Claims that Art 11 of the DSU has been violated are quite frequent in WTO practice.
⁶² Ehlermann and Lockhart (2004) defend the view that the differences in the standard of review
applied by WTO panels and the AB largely reflect differences across the covered agreements.
⁶³ On this score, see the analysis in Desmedt (1998).
 Dispute Settlement 411

(1) diligently looked for all relevant facts; and

(2) whether the final decision is reasonably supported by the facts before it.⁶⁴
The AB said as much in its report on US—Lamb. We quote from § 103:
Thus, an “objective assessment” of a claim under Article 4.2(a) of the Agreement on
Safeguards has, in principle, two elements. First, a panel must review whether competent
authorities have evaluated all relevant factors, and, second, a panel must review whether
the authorities have provided a reasoned and adequate explanation of how the facts support
their determination. Thus, the panel’s objective assessment involves a formal aspect and a
substantive aspect. The formal aspect is whether the competent authorities have evaluated
“all relevant factors”. The substantive aspect is whether the competent authorities have
given a reasoned and adequate explanation for their determination. (original emphasis)
The AB, in subsequent case-law, clarified that panels still retain a substantial
amount of discretion, even though they should not engage in de novo reviews.

4.5.5 The process before the AB

The process before the AB will start with a notice of appeal. The notice of appeal
serves the same purpose as the request for establishment of a panel. Only three mem-
bers of the AB (AB division) will adjudicate a dispute, although in practice, informal
discussions among all seven members take place. The AB will hold one meeting with
the parties. Its decisions are final. Assuming the AB has condemned the defendant,
it will have to bring its measures into compliance, an issue we discuss infra.
Panels must wrap up the process within 6 (and exceptionally, in 9) months. The
corresponding statutory deadline for the AB is 3 months. Horn and Mavroidis
(2006) show that, whereas the AB has, on average, respected the statutory dead-
line, panels take 386.89 days on average to complete the process.⁶⁵ This observa-
tion is of particular importance when viewed against the background of de facto
prospective remedies, the sine qua non in the WTO context, an issue to which we
will return infra.

4.5.6 The types of complaints before panels and the AB

The WTO DSU reproduces the GATT regime with respect to types of legal
complaints that can be lawfully launched in the context of dispute settlement.
The GATT distinguishes between three types of complaints:
(1) Violation complaints (Art XXIII.1a of the GATT);
(2) Non-violation complaints (Art XXIII.1b of the GATT);
(3) Situation complaints (Art XXIII.1c of the GATT).

⁶⁴ Expressed in this way, the standard of review resembles what is known in some domestic
administrative systems as the due diligence standard.
⁶⁵ Their data, which is being periodically renewed, runs until February 2005.
412 Institutional Provisions

The DSU has not altered this classification; it has, however, added some extra detail
with respect to non-violation and situation complaints (Art 26 of the DSU).

4.5.6.1 Violation complaints

A WTO member alleges, through a violation complaint, that the defendant has
violated the WTO Agreement. This category comprises, in practice, the vast
majority of complaints. On its face, it would seem that a complaining party
would have to prove, beyond violation of a GATT provision, that a benefit accru-
ing to it had been nullified or impaired. Standing GATT case-law however has
made it clear that any time the complaining party has managed to prove violation
of the GATT there was ipso facto a prima facie case that a benefit accruing to it
had indeed been nullified or impaired.⁶⁶ In fact, thus far, there has never been a
case where this presumption has been effectively rebutted.

4.5.6.2 NVCs
NVCs are described in Art XXIII.1b of the GATT:
If any contracting party should consider that any benefit accruing to it directly or indir-
ectly under this Agreement is being nullified or impaired or that the attainment of any
objective of the Agreement is being impeded as the result of
...
(b) the application by another contracting party of any measure, whether or not it con-
flicts with the provisions of this Agreement . . .
In Art 26 of the DSU, the founding fathers reflected some of the jurisprudential
evolution in the context of non-violation complaints: the defendant does not have
to withdraw the challenged measure; instead, a mutually agreed solution should be
encouraged. On the other hand, suggestions by an Arbitrator entrusted with the
task to determine the level of nullified benefits are non-binding on their addressee.
The rationale for non-violation complaints has to do with the explicit acknow-
ledgement (in Art XXIII of the GATT) that benefits (from trade liberalization)
accruing to the various WTO members can be nullified as a result of either a
GATT inconsistent or GATT consistent behaviour. In a way, non-violation com-
plaints are an application of the good faith (bona fides) principle, in the sense that,
by adhering to the WTO, a member should not only abstain from GATT incon-
sistent, but also from GATT consistent behaviour, to the extent that the latter
might nullify benefits from trade liberalization.
The 1990 panel report on EEC—Oilseeds I provided the panel’s understanding
of the rationale for a non-violation complaint (§§ 144 and 148):
The idea underlying [the provisions of Article XXIII:1(b)] is that the improved competi-
tive opportunities that can legitimately be expected from a tariff concession can be frus-
trated not only by measures proscribed by the General Agreement but also by measures

⁶⁶ See Jackson (1969) on this score.

 Dispute Settlement 413
consistent with that Agreement. In order to encourage contracting parties to make tariff
concessions they must therefore be given a right of redress when a reciprocal concession
is impaired by another contracting party as a result of the application of any measure,
whether or not it conflicts with the General Agreement.
...
The Panel considered that the main value of a tariff concession is that it provides an
assurance of better market access through improved price competition. Contracting par-
ties negotiate tariff concessions primarily to obtain that advantage. They must therefore
be assumed to base their tariff negotiations on the expectation that the price effect of the
tariff concessions will not be systematically offset. If no right of redress were given to
them in such a case they would be reluctant to make tariff concessions and the General
Agreement would no longer be useful as a legal framework for incorporating the results
of trade negotiations.
The aforementioned panel dealt with a (legal) subsidy scheme granted by the
European Community. Subsidies have a direct impact on prices since a benefit-
ing producer could use the proceedings to lower its price. Assume now that two
countries exchange concessions. Reciprocity obliges that A has to promise some-
thing to extract a promise from B and vice-versa; A and B are both ‘promisors’
and ‘promisees’. Subsequent to the exchange of concessions, A grants a subsidy
to its producers of the commodity for which it promised B a tariff treatment. B
finds itself thus in a situation where it paid to extract a concession from A and at
the same time cannot make any use of it since its exports are now probably over-
priced when compared to the subsidized domestic production of A. B might thus
lose the incentive to continue negotiating in the future and the whole liberaliza-
tion game might suffer as a result. This is where non-violation complaints kick
in: B cannot request from A to stop subsidizing, since A is not practicing GATT
inconsistent activity. B, however, can request compensation or arrive at any other
mutually agreed solution. In all non-violation complaints launched during the
GATT era, the complainant claimed that a subsidy paid subsequent to a negoti-
ated concession was nullifying benefits accruing to it.⁶⁷ The question, of course,
is where should one draw the line? Could, for example, a complainant attack
GATT consistent behaviour (other than subsidies), which might be nullifying
its benefits under the GATT? And if yes, could a list be drawn (ex ante or ex post)
of measures that could come under the purview of a non-violation complaint?
Certainly, WTO members could, through an imaginative application of non-
violation complaints, challenge a large amount of legal practices that might be
impairing their benefits.⁶⁸ This type of question brings us squarely into a discus-
sion on the scope of non-violation complaints.

⁶⁷ See Petersmann (1991) for an exhaustive discussion on this score.

⁶⁸ In one of the early papers on this issue, Hoekman and Mavroidis (1994) suggested that WTO
members could conceivably attack restrictive business practices through a non-violation complaint
(assuming that they act as regulatory subsidies and thus nullify benefits to trading partners). The
Japan—Film (Kodak—Fuji) litigation provides evidence that this view was not impracticable.
414 Institutional Provisions

The law (Art XXIII.1b of the GATT), as we saw above, was not modified over
the years. The (implicit) response came through case-law: the AB, in its report
on EC—Asbestos had to, inter alia, face an argument advanced by the European
Community to the effect that a non-violation complaint could not be used against
health-based trade-obstructing measures, justified under Art XX of the GATT;
in the EC view a measure could not on the one hand, be explicitly permitted
under a GATT provision while the regulating state remains liable for any related
damage to its trading partners interests. The AB dismissed this argument (§§ 188
and 189). Remarkably, this case-law did not limit the realm of possible applica-
tions of a non-violation complaint to say those health-based measures that have a
direct impact on the value of concessions; it is based on the open-ended language
employed in Art XXIII.1b of the GATT. Following this jurisprudence, it seems
plausible to argue that non-violation complaints can be raised against practically
each and every government measure that might have an impact on the value of
negotiated concessions.
Standing GATT/WTO case-law has clarified that, for a WTO member to
successfully launch a non-violation complaint, it must demonstrate that:
(1) as a result of some action taken by a WTO member, the consistency of which
with the WTO Agreement is not in dispute;
(2) provided that such action occurred subsequently to the conclusion of a tariff
concession; and
(3) provided that such action could not have been reasonably anticipated by the
complaining;
(4) the action at hand reduced the value of the concession negotiated between
the complainant and the defendant (impaired benefits accruing to the com-
plaining party).⁶⁹
Whenever these four conditions have been cumulatively met, panels have always
accepted the non-violation complaint submitted by the complaining party. Points (1)
and (2) above are factual issues. Point (4) would typically require a causality analysis.
So far GATT/WTO case-law has not contributed much on this score: since most of it
dealt with subsidies, panels took it for granted that beneficiaries would use the subsidy
proceedings to lower their prices and thus make it harder for imports to penetrate the
market. The economic rationality of such an approach is, of course, highly questionable.⁷⁰
With respect to point (3), the panel, in its report on Japan—Film established a bench-
mark in order to evaluate whether a measure could have been reasonably anticipated
or not. In the panel’s view, the timing of the occurrence of the challenged measure acts

Bagwell and Staiger (2002) take the view that non-violation complaints are a very attractive means
to ensure that reciprocal commitments will be observed, in a wide variety of contexts.
⁶⁹ The first systematic analysis of non-violation complaints is in Petersmann (1991).
⁷⁰ If for example, they could enjoy higher returns by playing in the stock exchange market, why
not simply invest there instead of investing in their own company?
 Dispute Settlement 415

as the benchmark for the allocation of the burden of proof: if the challenged measure
occurred after the exchange concession, the affected party would be presumed not to
reasonably expect it; the opposite is the case, if the measure pre-exists the exchange of
concessions (§ 10.79–10.81).
It follows that current case-law suggests that WTO members might have to
pay compensation for any, in principle, GATT consistent action that might have
a negative trade impact on their trading partners. It bears repetition that, so far,
compensation has been successfully requested only in cases where, subsequent to
the conclusion of a concession, a subsidy was paid by the country offering the con-
cession for the product on which the concession had been offered. In Japan—Film
the challenge against restrictive business practices failed, mainly because most of
the practices challenged pre-dated the relevant concession. A line, nevertheless,
needs to be drawn. It is probably not by accident that trading partners attacked
subsidization through the NVC mechanism. Subsidizing a product the tariff of
which has been lowered during negotiations might indicate intent to undo what
had been committed. Such intent is, in all likelihood, less probable when a coun-
try issues, for example, a health measure that negatively affects its trading part-
ners: trade instruments only affect one of the two (domestic, foreign) competing
products, whereas domestic instruments affect both. Introducing an intent-test
here, the difficulties surrounding such a test notwithstanding,⁷¹ is probably an
appropriate way to draw the line between cases that could and cases that should
not come under the purview of Art XXIII.1b of the GATT.

4.5.6.3 Situation complaints

The ambit of situation complaints is defined in Art XXIII.1c of the GATT:
If any contracting party should consider that any benefit accruing to it directly or indir-
ectly under this Agreement is being nullified or impaired or that the attainment of any
objective of the Agreement is being impeded as the result of

(c) the existence of any other situation . . .

It is hard to imagine what could come under Art XXIII.1c of the GATT which
is not already covered by Art XXIII.1a, or by Art XXIII.1b of the GATT. Hence,
it is not surprising that throughout the GATT years there is no case-law in this
respect (with the exception of a 1982 request for consultations by the European
Community against Japan on the so-called ‘Japanese way of life’ case,⁷² which,
however, never led to the establishment of a panel; moreover, claims of a compar-
able nature made in the context of the Japan—Film dispute were not discussed

⁷¹ See the discussion in Chapter 3.

⁷² See on this score the excellent analysis in Bronckers (1987, pp 111–19). Bronckers discusses
the political economy of the dispute as well as its eventual legal repercussions, in case the com-
plainant (European Community) would be willing to go the full 9 yards and formally submit the
dispute to a GATT panel.
416 Institutional Provisions

in any comprehensive manner). DSU, Art 26.2 deals with situation complaints.
This provision makes it clear that the DSU is applicable to situation complaints
up to the stage of the proceedings when the report has been circulated to WTO
members. The practical significance of this provision is that there is no negative
consensus with respect to the adoption of the report, and with respect to a request
to authorize countermeasures, as far as situation complaints are concerned.

4.5.7 Recommendations and suggestions by panels and the AB

WTO adjudicating bodies must, according to Art 19 of the DSU, issue a recom-
mendation, when they find that the challenged measure is inconsistent with the
WTO. The content of recommendations is identical across reports: the guilty party
must bring its measures into compliance with its obligations. In addition, WTO
adjudicating bodies may suggest ways to implement recommendations. The con-
tent of suggestions depends on what is being requested. Recommendations and
suggestions are meant to help their addressee to implement the various findings
in a report. Case-law has clarified that recommendations:
(1) must be issued by panels, assuming that an inconsistency has been estab-
lished, with panels enjoying no discretion on this score;
(2) are binding on their addressees;
(3) have an inflexible content: panels must reproduce the language included in
Art 19 of the DSU;
(4) amount to an obligation of result, in the sense that they allow WTO members
to choose the means (specific conduct) in order to reach the requested result
(consistency with the WTO).
Case-law has also clarified that suggestions:
(1) will be issued depending on panels’ discretion. Panels enjoy discretion even if
they have been requested to issue suggestion;
(2) are not binding on their addressees who might choose to ignore them;
(3) have flexible content, and panels have discretion to shape it, provided that
they observe the non ultra petita principle;
(4) amount to an obligation of specific conduct since they prejudge not only the
result (consistency with the WTO) but also the means that must be employed
to this effect.⁷³

⁷³ What matters nevertheless, is that in case a suggestion has been followed, there should
be no doubt that implementation has occurred (irrebutable presumption, at least if the AB has
issued the suggestion, or if it agrees with the appealed panel’s suggestion). Conversely, in case a
suggestion has not been followed, it is not excluded that, following a challenge, the implement-
ing measure falls short of the implementation-requirements, Mavroidis (2000). Panels should be
encouraged to allocate the burden of proof depending on whether a suggestion has been followed
or not (if yes, the complainant should carry it, if not, the defendant). Complainants currently
always carry the burden of proof.
 Dispute Settlement 417

4.6 Enforcing WTO obligations

4.6.1 The process in a nutshell
Assuming an un-appealed panel (or the AB) has found that a WTO member has
violated its obligations,⁷⁴ they will recommend and might suggest ways to imple-
ment the findings. The WTO member whose measures are found to be inconsist-
ent with the WTO will have to abide by the decision/ruling within a reasonable
period of time. If implementation has not occurred, the injured party might be
authorized to suspend concessions. Assuming disagreement between the parties
as to the level of proposed countermeasures, their dispute will be submitted to an
Arbitrator (usually the members of the original panel) who will decide on first and
last resort on the level of countermeasures to be imposed. As a transitional measure,
compensation might be voluntarily agreed among the interested parties. In good
faith, imposed countermeasures must be lifted if implementation occurs at a later
stage. In the case of disagreement regarding whether implementation occurred,
however, the interested party might be forced to challenge the countermeasures by
bringing a new dispute to this effect. The request by the interested party suffices for
suspension of concessions to be authorized (assuming, of course, that the legisla-
tive prescription for equality between damage suffered and requested suspension of
concessions has been observed).
A distinction is usually made between exogenously, and self-enforced contracts,
depending on the identity of the enforcer. The WTO is a self-enforcing contract:
assuming non-compliance with the rulings by a WTO adjudicating body, the
injured WTO member can request and impose countermeasures,⁷⁵ that is, raise
the level of its bound duties vis-à-vis the author of the illegal act. WTO counter-
measures are by law (Art 22.1 of the DSU) transitional: the DSU explicitly reveals a
preference for specific performance of the obligations assumed; it follows that, from a
systemic perspective, countermeasures are a means to induce compliance.
Viewed from the injured WTO member’s perspective, the situation could be
presented as follows: suspending concessions is tantamount to imposing a cost on
the country adopting countermeasures, in the sense that the price of imports will,
as a result of countermeasures, become higher than before (with the ensuing wel-
fare losses for domestic consumers). Adoption of countermeasures will thus largely
depend on a cost-benefit analysis, whereby on the one side of the equation the self-
imposed loss will be calculated, and on the other, the likelihood that the author of
the illegal act might change behaviour as a result of the countermeasures imposed
(and the resulting benefits from such change will be computed). As a result, one
should not assume that, when faced with non-compliance, WTO members will
mechanically move and suspend concessions. WTO members have not often had

⁷⁴ As will be shown infra, the type of complaint matters when it comes to enforcement. I assume
that a violation complaint has been submitted, since this is the most frequent type of complaint.
⁷⁵ I use this term as equivalent to the DSU term ‘suspension of concessions’.
418 Institutional Provisions

recourse to countermeasures.⁷⁶ Increasingly, there is a growing concern with respect

to the effectiveness of the system to take care of deviations from what has been
agreed. It is certainly not a coincidence that developing countries especially have,
for the first time during the ongoing Doha round, tabled wide-ranging proposals
aiming at reshaping the existing regime. We will return to this issue later.

4.6.2 The types of complaint matter for enforcement purposes

Assume that a WTO has raised a non-violation complaint and succeeded: the WTO
adjudicating body will recommend a mutually satisfactory adjustment. In the absence
of practice, we can only speculate as to the scope of this term. To facilitate the reso-
lution of the dispute, an Arbitrator can, upon request, determine the level of benefits
which has been impaired, but such suggestions are not binding on the parties to the
dispute (Art 26.1(c) of the DSU). Recourse to compensation (itself a voluntary option)
can be part of a mutually satisfactory adjustment (Art 26.1(d) of the DSU).
Assume now that a WTO member has raised a situation complaint and suc-
ceeded: Art 26.2 of the DSU makes it clear that the rules of the DSU apply:
only up to and including the point in the proceedings where the panel report has been
circulated to the Members.
As of this stage, the old GATT positive consensus rule, as already discussed supra,
applies: adoption of the panel report and any eventually following action can be
formally decided only if both parties to the dispute agree on the issue.
In what follows, we concentrate on enforcement in case a violation complaint
has been successfully lodged. Violation complaints constitute, anyway, the vast
majority of litigation before the WTO.

4.6.3 Compliance within a reasonable period of time

Following a recommendation and/or suggestion by a WTO adjudicating body, the
WTO Member who is the author of the illegal act will have to bring its measures
into compliance with the WTO within a reasonable period of time (RPT). The
RPT can be defined either by agreement between the parties to the dispute
(bilateral definition), or, in case an agreement is not possible, through recourse
to arbitration (multilateral definition). There is by now rich case-law on arbitral
awards calculating the RPT (Art 21.3(c) of the DSU): the identity of the WTO
member (developed, developing), and the domestic law constraints regarding
implementation emerge as two of the most frequent attendant circumstances that
Arbitrators will control for when calculating the RPT.⁷⁷

⁷⁶ We are of course aware of the fact that it is difficult to establish a benchmark to measure
‘often’ in this respect. Bagwell et al (2005) have counted few cases where countermeasures have
been imposed (adding cases where a request has been tabled). The same authors count a much lar-
ger number of persisting violations, which suggests that quite a few ongoing violations of the WTO
remain unpunished.
⁷⁷ For a lengthy exposition of this issue, see Palmeter and Mavroidis (2004).
 Dispute Settlement 419

4.6.4 Compliance panels

If during the RPT, however defined, no implementing activities at all take place,
then the complaining party can request authorization to suspend concessions.
If, on the other hand, implementing activities do take place, one can distinguish
between two situations:
(1) the complaining party agrees that the author of the illegal act has adequately
implemented its WTO obligations; or
(2) the complaining party does not agree that this has indeed been the case.⁷⁸
The first case marks the end of the matter. In the second case, the original
complainant (the party that now disagrees with the adequacy of the implemen-
tation efforts) can request the establishment of a compliance panel as per Art
21.5 of the DSU. A panel will be composed by the members of the original panel,
if possible. If not, the parties to the dispute and, ultimately in case of disagree-
ment, the Director-General of the WTO will compose the panel. DSU, Art 21.5
requests that the panel issue its reports within 90 days. In practice, empirical ana-
lysis suggests that deviations from this statutory deadline are quite common.⁷⁹
The question as to whether consultations are necessary for recourse to a com-
pliance panel to be lawful is still open. The AB, when confronted with this issue in
its Mexico—Corn Syrup report, decided not to address it (§ 65). In this case, both
the panel and the AB proceeded to the review of the case before them, regardless
of the absence of consultations.⁸⁰ The applicability, mutatis mutandis, of Art 6.2
of the DSU has by now been endorsed by the AB. In its report on US—FSC (Art
21.5—EC II), the AB took the view that Art 6.2 of the DSU is generally applic-
able in Art 21.5 of the DSU proceedings (§ 62).
The mandate of compliance panels⁸¹ has been clarified in the AB jurisprudence
in its report on Canada—Aircraft (Art 21.5—Brazil). Following condemnation
by a panel and the AB, Canada revised its original TPC programme (the TPC
programme was the measure which had been found to be inconsistent with the
WTO) so that it did not amount to an export subsidy anymore (and thus, be con-
sistent with Art 3.1(a) of the SCM). The complainant, however, was not in agree-
ment with Canada as to the adequacy of its implementation efforts. The dispute
was submitted to a compliance panel, the report of which was then appealed. On

⁷⁸ There is a third (rather unlikely) situation where the parties to the dispute agree that imple-
menting activities that took place were inadequate. In this case as well, the injured party can request
authorization to suspend concessions.
⁷⁹ Horn and Mavroidis (2006) calculate the average length of compliance panels to be 208.62
days. An example of non-identical composition across the original and the compliance panels is
offered by the panel on US—Softwood Lumber V (Art 21.5—Canada), where a member of the ori-
ginal panel subsequently became DDG of the WTO and had to withdraw from the panel and be
replaced.
⁸⁰ Another open question is whether third parties in a compliance panel must have been third
parties before the original panel as well. DSU, Art 10 does not address this issue head on.
⁸¹ See on this score, but also, in general on the role of panels, Davey (2004).
420 Institutional Provisions

appeal, the AB had, inter alia, the opportunity to provide its understanding of
the mandate of a compliance panel:
. . . in our view, the obligation of the Article 21.5 Panel, in reviewing ‘consistency under
Article 21.5 of the DSU’, was to examine whether the new measure—the revised TPC
programme—was ‘in conformity with’, ‘adhering to the same principles of’ or ‘compatible
with’ Article 3.1(a) of the SCM Agreement. In short, both the DSU and the Article 21.5
Panel’s terms of reference required the Article 21.5 Panel to determine whether the revised
TPC programme involved prohibited export subsidies within the meaning of Article
3.1(a) of the SCM Agreement (§ 37, original emphasis)⁸²
Compliance panels are not required to issue recommendations: assuming they
find that no adequate implementation has occurred, the complainant will have
to decide whether to pursue a request for suspension of concessions. Conversely,
in case the compliance panel finds that compliance indeed occurred, the case
is, of course, regarded as having been brought to an end (assuming the AB has
endorsed the panel’s view, since all compliance panel reports can be appealed).⁸³

4.6.5 Compensation
Compensation and suspension of concessions or other obligations are temporary
means that can be used alternatively until specific performance (and thus, reso-
lution of the dispute) has been secured. Compensation is voluntary. Its form is
not prejudged in the DSU. Recourse to it has been very infrequent: in fact, com-
pensation has only been agreed once; following the condemnation of US copy-
right practices in US—Section 110(5) Copyright Act, the European Community
(complainant) and the United States (defendant) agreed to submit to an Art 25 of
the DSU arbitration, since they could not agree on the amount of compensation
to be paid.⁸⁴ It has been sometimes argued that the reason why recourse to com-
pensation is not more frequent has to do with the fact that it must respect MFN.
The legal underpinnings for this opinion are in Art 22.1 of the DSU which calls
for WTO consistent compensation. Compensation does not necessarily have to
respect MFN, however; it all depends on the nature of the violation: if A has been
found to have wrongfully imposed AD duties against imports from B, and agrees
to provisionally compensate B until full implementation has occurred, A cannot

⁸² The AB confirmed its understanding of this issue in § 88 of its report on US—Shrimp (Art
21.5 – Malaysia).
⁸³ Practice reveals that a second compliance panel can be effectively established in the context
of the same original dispute. Brazil – Aircraft (Art 21.5 – Canada, Second recourse) is an appropriate
illustration. However, practice on this score has probably developed contra legem. The purpose of
compliance panels is to evaluate whether compliance occurred during the RPT. There is, however,
only one RPT. Hence, there can be only one compliance panel to evaluate implementation in the
context of a particular dispute.
⁸⁴ See the discussion of the case in Grossman and Mavroidis (2003). The authors note, inter
alia, that it is questionable whether Art 25 of the DSU was meant to serve this purpose. It seems
that the Arbitrator here assumed a role normally entrusted to an Art 22.6 of the DSU arbitration.
At any rate, the compensation in this case was monetary.
 Dispute Settlement 421

be requested to also compensate C and D, which have suffered no damage at all

from A’s illegality.
It is true that monetary compensation is a very attractive proposition, especially
when compared to the alternative, suspension of concessions, discussed infra.⁸⁵
This is the case because compensation does not create the kind of distortions that
suspension of concessions does.

4.6.6 Suspension of concessions

4.6.6.1 Procedural issues
Through suspension of concessions, the injured WTO member(s) will raise its
duties above the bound level, in an effort to inflict trade damage equal to that
it suffered as a result of the commission of an illegal act. WTO members might
agree on the level of suspension of concessions. If they do not, the interested mem-
ber will request from an Arbitrator (usually, the members of the original panel) to
decide on the level of suspension. To do this, it will have to come up with a num-
ber and a list of products where suspension of concessions is sought.⁸⁶
The Arbitrators’ report is the final word as far as the permissible level of coun-
termeasures is concerned. DSU, Art 22.7 pertinently provides that:
The parties shall accept the arbitrator’s decision as final.⁸⁷

⁸⁵ Good arguments in favour of more frequent recourse to compensation have been advanced
by Bronckers and van den Broek (2005), and Pauwelyn (2000), calling for compulsory use of this
mode of implementation. The problem is that there is no response (in these proposals at least)
to the question as to what happens in the case when the WTO member concerned refuses to
compensate?
⁸⁶ Importantly, cross-retaliation can occur and the injured WTO member might retaliate in
GATS against a violation occurring under GATT. The Arbitrators’ report on EC—Bananas III
(Art 22.6—EC) imposed some limits on cross-retaliation; they did not accept the list presented by
Ecuador as such. In their view, Ecuador’s proposal does not bind the Arbitrators, who retain the
discretion to review it and amend it, as it did, in this case (§ 52, § 101, §§ 125–6 and §§ 167–70. In
§ 173 of their report, the Arbitrators indicated the sectors where suspension of concessions should
take place). In a nutshell, the Arbitrators in this case held the view that their understanding of what
is practicable and eff ective prevails. Would they be prepared to take the blame in case their predic-
tion was to be proved wrong? This is problematic in a number of respects. It is not the task of the
Arbitrators to second-guess evaluations of efficiency and practicability. Wisely, the text of Art 22.3
of the DSU prefaces every sentence to this effect with the words if that party considers (that action is
impracticable or ineffective). This suggests that the legislator’s intent was to leave WTO members
with the maximum of discretion as to the sectors where action should be taken, disciplining only
the amount of permissible action (by virtue of Art 22.4 of the DSU). Luckily, this view has not been
followed in subsequent practice. It seems fair to argue that, nowadays, the discretion of Arbitrators
is limited to a review of whether the requesting party has provided an explanation why it was not
practicable or effective to suspend concessions within the same sector and does not extend beyond
this point. In subsequent practice, the Arbitrators have refrained from reviewing the nature of the
concessions to be suspended, limiting their review on the amount of proposed suspension of con-
cessions. Since practice so far, however, is quite limited, it is probably too early in the day to draw
any definitive conclusions as to the content WTO practice on this score.
⁸⁷ The Arbitrators in EC—Bananas III (Ecuador) (Art 22.6—EC) confirmed this view (§ 2.12).
422 Institutional Provisions

4.6.6.2 Quantifying suspension of concessions

DSU, Art 22.4 calls for equivalence between the proposed level of suspension of con-
cessions and the level of nullification and impairment suffered by the injured party:
The level of the suspension of concessions or other obligations authorized by the DSB
shall be equivalent to the level of the nullification or impairment.
Crucially, equivalence is established at the moment when a request for authoriza-
tion to adopt countermeasures is being submitted. It is, however, highly unlikely
that, say $200 million dollars/year which is authorized in September 2005
because it corresponds to the damage suffered in the preceding year, will corres-
pond to the damage suffered each subsequent year that the illegality persists. A
number of variables might influence the amount of damages over time, and there
is substantial uncertainty as to the eventual outcome. The DSU, as it stands, does
not contain any institutional guarantees to ensure fast relief in case of deviation
from the discipline of Art 22.4 of the DSU: all an adversely affected WTO mem-
ber can do is request a new panel. This provides interested parties (as well as those
willing to go ahead with this option) with yet another incentive to overshoot the
quantification of countermeasures.
With this caveat in mind, we turn to quantification issues clarified through
case-law. There are four such elements:
(1) remedies, in practice, are calculated prospectively;
(2) indirect benefits are not recouped;
(3) only value added matters;
(4) legal fees are not recouped.
Prospective remedies: in the GATT era, there are five reported cases where GATT
Contracting Parties recommended that GATT parties which illegally imposed
AD or countervailing duties should reimburse all duties illegally collected from
the date of the first initiation of such duties.⁸⁸ There is only one such case in the
WTO era (the panel report on Australia—Automotive leather II). For the rest, all
Arbitral awards went for prospective remedies, whereby the injury is calculated not
from the time when the illegality had been committed, but from the end of the
RPT. DSU, Art 19 does not prejudge this issue. In principle, panels are free to
suggest retroactive remedies. Customary international law⁸⁹ certainly sides with
retroactive remedies. However, international law does not privilege either cus-
tomary law or treaty provisions; hence, nothing precludes parties from adopting
contractual terms at variance with customary international law. In light of the
GATT past practice, it is doubtful whether the founding fathers of the WTO,
by enacting Art 19 of the DSU, did indeed intend to deviate from customary

⁸⁸ All GATT cases are reported in Petersmann (1993) and Mavroidis (2000).
⁸⁹ See Pauwelyn (2003), and Mavroidis (2000).
 Dispute Settlement 423

international law. Practice seems to go this way, nonetheless. Prospective rem-

edies imply a smaller monetary burden/cost for the non-compliant party than
that under retroactive remedies in the same case. Viewed from this perspective,
they might prove to be a less efficient instrument in inducing compliance. For
instance, what is the incentive for a WTO member to respect the AD disciplines
if it knows that, at worst, it risks seeing itself obliged to stop some 4 years down
the road imposing duties that were illegally rendered in the first place? It will de
facto have provided itself with a safeguard for its producers in the meantime,
without having to comply with the safeguards (or the AD) provisions.⁹⁰ On the
other hand, legislative initiatives aiming at clarifying this issue have not been
translated into law thus far.⁹¹ As things stand, one can only conclude that, as a
matter of WTO law, the prospective calculation of countermeasures (as of the
end of the RPT) represents the norm at present.
Indirect benefits: the Arbitrators in their report on EC—Bananas III (Ecuador)
(Art 22.6—EC) dealt, inter alia, with the following issue: the United States claimed
that it should be compensated for lost profits resulting from the EC bananas import
regime. In its view, the European Community, by blocking imports into its market
of bananas originating in Mexico, was ipso facto blocking exports to the Mexican
market of fertilizers originating in the United States. In other words, in the US
view, Mexico would have less need for US fertilizers in light of the reduced export
opportunities of bananas to the EC market. The Arbitrators decided against the US
claim in this respect. In their view, the European Community could be held liable
for trade in bananas lost by Mexican exporters, but not for trade in fertilizers lost by
US exporters as a result of Mexico’s decision to reduce imports of the said commod-
ity. DSU, Art 22, consequently, must be construed so as to disallow the inclusion of
indirect benefits, when calculating the amount of countermeasures (§§ 6.12–6.14).
This attitude is probably meritorious. Trade is so intertwined across countries that
opening the door to indirect benefits amounts to a quasi-impossibility to draw the
line somewhere. What if, for example, fertilizers have some added value originat-
ing in Japan, which, in turn has added value originating in India and so on and so
forth. This is very much today’s world with the emergence of trade in tasks. On the
other hand, this approach does not fully address the injury suffered. This is prob-
ably one case where, for reasons of administrative ease, sub-optimal mechanisms
are acceptable.
Value added: a direct outcome of the discussion on indirect benefits is the decision
by the Arbitrators in the same case to compute only value added when calculating

⁹⁰ This argument is, of course, short. It definitely makes good sense if violations of the contract
are bad faith violations. There are, nevertheless, cases of genuine ambiguity, and we pointed to a
number of them during the preceding chapters. Should losing WTO members be punished equally
severely in such cases? The problem is that it is quite difficult to draw a clear dividing line between
good- and bad-faith cases and, faute de mieux, a unique solution in case of violation is probably
the most appropriate. The better arguments (inducement to comply) argue in favour of retroactive
remedies, but see also some valid counter-arguments advanced by Lawrence (2003).
⁹¹ See Bagwell et al (2005).
424 Institutional Provisions

the level of nullification and impairment. In other words, assuming that a good
costs €10 and €4 of the total value is from imported goods. In this case, the exporter
of the said good will lose €10/unit in revenue when an illegal trade barrier has been
erected against its exports. However, as a result of the trade measure it is by now
facing, it will probably stop importing the input costing €4. As a result, the actual
nullification and impairment will not be €10, but €6/unit. The Arbitrators in their
report on EC –Bananas III (Ecuador) (Art 22.6—EC) endorsed the view that only
value added can be part of the calculation (§§ 6.18ff ).
Legal fees: the Arbitrators in their report on US—1916 Act (EC) (Art 22.6—US)
made it clear that legal fees paid cannot form part of the calculation of nullifica-
tion and impairment (§ 5.76).

4.7 Monitoring compliance

WTO members found to be in violation of their obligations must periodically
report their implementation efforts to the DSB. The obligation to report kicks in
6 months after the establishment of the RPT. As a matter of practice,⁹² however,
assuming that a compliance panel has been established, there is no need to report,
or, better, the obligation to report is suspended. It will resume only if the compli-
ance panel (and/or the AB, as the case may be) finds that implementation did not
occur during the RPT. The WTO member concerned might fail to implement
its obligations under Art 21.6 of the DSU. In such cases, practice evidences the
tendency, on occasion, of compliance panels to provide their own status reports on
the actual state of implementation.⁹³
Assuming a WTO member reports implementation, one might have to look into
various facts, including the reactions of other WTO members in order to form
an opinion as to whether implementation has indeed occurred. Implementation
must be WTO consistent (Art 3.5 of the DSU), and WTO members can con-
test the WTO consistency of implemented measures during the DSB meeting (Art
3.6 of the DSU). Some scholars take the view that absent a request to establish a
compliance panel, implementation can be inferred. This is probably a reasonable
assumption, but problematic nonetheless: it could be the case that complainant and
defendant have colluded and agreed on a deal which suits them both, but which is
not legal. There are other, even more problematic cases: the complainant might raise
questions or request time to reflect upon the information provided. In such cases,
the defendant does not have to provide a new implementation report. If the com-
plainant provides no status report, then again, it could very well be the case that we
are left in the dark as to whether compliance has occurred in the case concerned.

⁹² Practice is quite important here since the DSU rules are anything but detailed on this score.
Practice is evidenced in the DSB/M-series, the WTO documents that reflect discussions before the
DSB.
⁹³ Very helpful discussions with Bill Davey and Joost Pauwelyn on this score are acknowledged
here.
 Dispute Settlement 425

Remarkably, the DSU does not contain any specific provisions dealing with the
situation where, post-adoption of countermeasures, the WTO member concerned
has implemented its WTO obligations. All Arts 21.6 and 22.8 of the DSU state is
that the DSB will keep all issues regarding implementation under surveillance. All an
interested WTO member can do is request the establishment of a panel, which will
examine the legality of the countermeasures imposed against it when it has already
complied with its obligations.⁹⁴ This of course entails lengthy procedures. This is one
area where a legislative addition would be quite welcome: an additional (à la Art 21.5
of the DSU) fast track procedure would perform a useful function in this context.

4.8 The value of precedent in the WTO

During the GATT years, it was widely accepted that panel reports exerted per-
suasive power on subsequent GATT panels dealing with the same issue; there
was nevertheless, no formal obligation to adhere to the rulings of a panel report
that had dealt with the same issue (no stare decisis, in other words).⁹⁵ The advent
of the AB, it was assumed by some, would tilt the balance more in favour of the
precedent: the AB is confined to an examination of legal issues only, that is, its
pronouncements are, in principle, valid, irrespective of the transaction that gave
birth to the AB process.
This view is probably in line with what the AB had in mind when, in US
Softwood Lumber V, stated that adopted panel reports create legitimate expecta-
tions among WTO members (§§ 109–12), only to add that:
. . . following the Appellate Body’s conclusions in earlier disputes is not only appropri-
ate, but is what would be expected from panels, especially where the issues are the same.
(US—OCTG Sunset Reviews, § 188)
Hence, although there is nothing like formal stare decisis, an institution to which
common law students are accustomed, it is still the case that panels, with few notable
exceptions, will adhere to the AB’s rulings. They will do that, sometimes mechan-
ically, and sometimes only if they find the AB’s analysis persuasive. An appropriate
illustration of the latter category is offered by the panel report on US—Shrimp
(Ecuador): although the AB in US—Zeroing (Japan) had confirmed yet again its
case-law that zeroing (in AD investigations) is illegal under any circumstances, the
panel still had to find the AB’s reasoning persuasive before it decided to adhere to it
(§§ 7.34—7.41). It follows that in case of identity of factual issues across cases, it is
highly likely that a subsequent panel will follow the AB’s rulings.⁹⁶
What about the AB when it is confronted with its prior case-law? Does it have
to abide by its own case-law? One would assume that, in absence of distinguishing

⁹⁴ For reasons mentioned above, recourse to compliance panel should be judged contra legem.
⁹⁵ See Jackson (1969, pp 22ff ).
⁹⁶ As we noted in Chapter 4, however, when discussing the case-law on zeroing, this has not
always been the case.
426 Institutional Provisions

factors, or a new, better theory to deal with a particular transaction, this should
be the case. And indeed, the AB has to some extent followed its prior case-law. It
is not nevertheless, legally bound to do so. Take, for example, the AB reports on
US—Lead and Bismouth II, and US—Countervailing Measures on Certain EC
Products. In both cases, the AB is dealing with non-recurring subsidies. The issue
is to what extent there is pass-through of subsidies in case a state entity has been
privatized at arm’s length. The AB held the first time that this is necessarily the
case, whereas it stated the opposite the second time, without, however, approach-
ing the change of heart as a genuine revirement de jurisprudence.⁹⁷
It is thus, very much the case that the AB strives to pass the message that its case-
law progresses in linear manner, even when this has not necessarily been the case.⁹⁸
This issue is linked to the overall drafting attitude of the WTO adjudicating bod-
ies. The AB especially, has been quite erratic: there are cases where the AB has adopted
a maximalist attitude,⁹⁹ and tried to prejudge a number of transactions. Sweeping
statements, such as arm’s length transactions always exhaust benefits to subsidized
companies, are both unnecessary and wrong. WTO adjudicating bodies, and the AB
especially, will be well-advised to adopt a minimalist approach and judge as much
as necessary, as little as possible (incrementalism).¹⁰⁰ Incrementalism should not be
equated to absence of theory: WTO adjudicating bodies must always explain the
wider theory supporting their ratio decidendi. In an obiter dictum, they are welcome
to go one step further and provide their reading of a class of similar to the adjudicated
transactions. Their ratio decidendi, nevertheless, should be transaction-specific. The
wider issues should be left to the legislator, that is, the WTO members.

4.9 Does WTO enforcement work?

The issue of enforcement of WTO obligations has provoked a lot of interest in
the DSU review. Empirical data confirms that the current enforcement is privi-
leged in the context of asymmetric bargaining power between the parties in
dispute: for example, there is not even one reported case where a ‘smaller’ coun-
try faced with non-compliance by the ‘bigger’ defendant, has had recourse to
countermeasures.¹⁰¹

⁹⁷ See on this score, Grossman and Mavroidis (2005).

⁹⁸ On the attitude of the WTO adjudicating bodies in general, when it comes to interpreting
the WTO, see Qureshi (2006).
⁹⁹ Maximalist attitude should not be confused with judicial economy. The former has to do
with the drafting style, the latter with the number of claims that a court judges necessary to adju-
dicate in order to resolve a dispute. Panels have occasionally adopted judicial economy, and this
attitude is problematic in the WTO setting: assuming reversal of their findings by the AB, in the
absence of remand authority, a complainant has to restart all over again the litigation process, and,
this time, litigate the claims that the panel did not rule on, on judicial economy grounds.
¹⁰⁰ See the US Supreme Court experience on this score, as discussed in Sunstein (1999). On
incrementalism and how economic theory can help us understand it, see Bown and Sykes (2007).
¹⁰¹ See Bagwell et al (2005).
 Dispute Settlement 427

It should come as no surprise then that, during the DSU review, a series of
developing countries tabled proposals requesting changes in the current DSU to
address their interests.¹⁰² Their discontent can be summarized in the belief that
suspension of concessions is not a weapon equally effective of all WTO members;
its efficacy depends on the identity of the players involved.¹⁰³ Their proposals typic-
ally take a developing country view and argue for some sort of special and differ-
ential remedies when developing countries are implicated in the process.
But the existing system is problematic not only because it does not sufficiently
address the interests of ‘smaller, poorer’ markets. Assuming no recourse is made
to extra-WTO measures, a ‘small, poor’ WTO member can also legally abuse (not
an oxymoron) the WTO dispute settlement procedures:¹⁰⁴ it can, for example,
go ahead and impose patently illegal AD measures and essentially provide its
domestic producers with a safeguard (for which, in the absence of retroactive
remedies, no compensation will be paid) until a panel and the AB eventually
ask it to stop such activity. Assuming extra willingness to procrastinate, it can
find refuge in the Art 21.5 of the DSU procedures by doing something during
the implementation period and contest, before a compliance panel and the AB,
a legal challenge to the effect that it had not done enough. Moreover, the DSU
knows of no injunction relief either.
The choice for weak, rather than strong remedies is not paradoxical though.
Ethier (2001) explains why: in his view, a link between enforcement and trade
liberalization is warranted. An optimal amount of liberalization will occur in
a system based on commensurate rebalancing, when, ex ante, trading partners
are unsure whether they might have to resort to violations in the future. The
common incentive of all trading partners is, when in doubt, to introduce weaker
rather than stronger remedies. Th is is so for a number of reasons: uncertainty as
to whether recourse to illegal measures will occur, who will commit illegalities,
whether a dispute will be lodged, how the judge will interpret the contract, the
institutional risks from wrong interpretations, etc. This is probably what explains
the current state of affairs in the DSU and its interpretation. One could probably
add further good arguments to those already advanced by Ethier (2001). The
trading game is a coordination game and countries are well aware of the merits of
trade liberalization. Deviations will be temporary and will not detract from the
common goal that they all have an interest (and thus an incentive) in achieving.
This is not to reduce enforcement to redundancy: enforcement is one of the fac-
tors contributing to implementation of the obligations. Dispute settlement serves

¹⁰² For an excellent discussion of the most important proposals, see Kessie (2004).
¹⁰³ The timing can of course be an issue here as well: the United States was admittedly quite vul-
nerable during the 2004 election year as the European Community was targeting Florida (a ‘swing’
state) with its countermeasures (citrus products) in the aftermath of the US—Steel Safeguard litiga-
tion. Such strategic thinking definitely takes place within national administrations when design-
ing countermeasures.
¹⁰⁴ As a matter of Realpolitik this might not occur for other reasons, see on this Busch and
Reinhardt (2002 and 2003), and Guzman and Simmons (2005).
428 Institutional Provisions

of course other functions as well, and chief among them is to complete the con-
tract: through its case-law, WTO adjudicating bodies have provided additional
information that did not exist in the original contract.¹⁰⁵ This does not mean that
WTO remedies are always weak. The identity of the players matters. Bagwell et al
(2005) find no case where a developing country faced with non-compliance has
had recourse to countermeasures. They, further, find no case where a developed
country challenging a developing country’s practices has had to have recourse
to countermeasures: the threat of countermeasures sufficed for implementation
to occur, an observation consistent with Schelling’s (1960) understanding of the
term ‘threat’. What follows is a list of all arbitration thus far in the context of
countermeasures (see Table 5.3).
Notice that countermeasures have been imposed only between more or less
symmetric countries. There is no evidence that one of the countries named above
changed its behaviour because of the countermeasures imposed. This observation
would lend further support to the argument that WTO remedies are particularly
effective in case of asymmetry across players. The discussion about remedies (not
only in the WTO context) reveals the diverging views between those who see the
DSU espousing property rules, and those who understand it as an expression of
liability rules: the former term captures the idea that the parties to a contract must
always perform the contract; the latter provides Contracting Parties with the pos-
sibility to compensate in lieu of specifically performing their contractual obliga-
tions. Undeniably, the GATT is about rebalancing: Art XXVIII of the GATT
provides parties with a negotiating forum any time they wish to revisit their con-
tractual promises and re-negotiate them. Does it also, however, allow rebalancing
following the commission of an illegal act? The DSU is a puzzle in this respect:
Art 22.1 of the DSU takes a prima facie stance in favour of property rules, when
it states:
... neither compensation nor the suspension of concessions or other obligations is pre-
ferred to full implementation of a recommendation to bring a measure into conformity
with the covered agreements.
On the other hand the DSU does nothing to ensure that the legislative preference
will be honoured. Actually, the means to honour this preference make us wonder
whether it is property rules that negotiators really had in mind, or whether Art
22.1 of the DSU is simply clumsy shorthand for a different preference.
First, the DSU allows suspension of concessions to, in principle, stay in place
forever. Suspension of concessions (a liability rule) is the interim measure until
compliance (the property rule) occurs. Since it can stay in place, in principle, for-
ever, the liability rule becomes not only de facto but also de jure some sort of per-
manent solution of a conflict;

¹⁰⁵ See on this score, Battigalli and Maggi (2002).

Table 5.3 Requests for Authorization to Impose Countermeasures

Date RPT expired Date DSB agreed Date of circulation of Date DSB authorized
to arbitration Arbitration Award countermeasures
DS18—Canada Australia—Salmon 6 November 1998 28 July 1999 Suspended
DS26—US EC—Hormones 13 May 1999 3 June 1999 12 July 1999 26 July 1999
DS27—Ecuador EC—Bananas III 1 January 1999 19 November 1999 24 March 2000 18 May 2000
DS27—US EC—Bananas III 1 January 1999 29 January 1999 9 April 1999 19 April 1999
DS46—Canada Brazil—Aircraft 18 November 1999 22 May 2000 28 August 2000 12 December 2000
DS48—Canada EC—Hormones 13 May 1999 3 June 1999 12 July 1999 26 July 1999
DS103—US Canada—Dairy 31 December 2000 1 March 2001 Suspended
DS108—EC US—FSC 1 November 2000 28 November 2000 30 August 2002 7 May 2003
DS113— Canada—Dairy 31 December 2000 1 March 2001 Suspended

Dispute Settlement
New Zealand
DS136—EC US—1916 Act 31 December 2001 18 January 2002 24 February 2004
DS160—EC US—Section 110(5) 31 December 2001 18 January 2002 Suspended Mutually
Copyright Act satisfactory temporary
arrangement notified
on 23 June 2003
DS162—Japan US—1916 Act 31 December 2001 18 January 2002 Suspended
DS217, DS234 US—Off set Act 27 December 2003 26 January 2004 31 August 2004 26 November 2004 (x 7)
(Byrd Amendment) 17 December 2004 (x 1)
(eight complainants)
DS222—Brazil Canada—Aircraft 20 May 2002 24 June 2002 17 February 2003 18 March 2003
Credits and Guarantees
DS245—US Japan—Apples 30 June 2004 30 July 2004 Suspended
DS257—Canada US—Softwood 17 December 2004 14 January 2005 Suspended
Lumber IV
DS264—Canada US—Softwood 2 May 2005 1 June 2005 Suspended
Lumber V
DS277—Canada US—Softwood 26 January 2005 25 February 2005 Suspended

429
Lumber VI
430 Institutional Provisions

Second, the liability rule has not been credited with enough persuasive force to
ensure that it will indeed be temporary and will, in all likelihood, lead to com-
pliance: the level of suspension of concessions must, by legislative fiat (Art 22.4
of the DSU), be substantially equivalent to the damage done. Consequently, the
author of the illegal act knows that the worst case scenario is that it will be called
to pay back the amount of damage done some time in the future.
Case-law has, of course, further weakened the persuasiveness of this remedy
by, for example, denying it a retroactive force. One cannot of course, blame nego-
tiators for the case-law. In short, the DSU on the one hand declares a property
rule, but on the other does nothing much to impose the stated preference.
All this of course is the positive law. In the normative discussion, the better
arguments rest with the proponents of the liability rule for the reasons already
exposed in Schwarz and Sykes (2002).
The current situation, as stated above, has raised concerns among WTO
Members. Besides the Mexican proposal on tradable remedies (WTO Docs
TN/DS/W/23 of 4 November 2002, and TN/DS/W/40 of 27 January 2003)
discussed infra, WTO members have failed to advance proposals that would
drastically change the current scene. While a substantive legislative amendment
does not, thus, seem to be on the cards, the literature has provided a number of
proposals aiming to address actual (or, even, perceived) distortions. A forgotten
dimension of the discussion, nevertheless, is that we still lack a theory of disputes;
we might know why some disputes have started, but we do not know why dis-
putes in general occur. Absent a clear understanding why disputes occur, it is an
awkward exercise to design enforcement mechanisms. For example, should we
have the same remedy in the case of a blunt violation of the contract as in the case
of a disagreement, in good faith, among WTO members as to the coverage of a
particular provision? And how do we distinguish between the two situations?¹⁰⁶
Moreover, even if the objective function of countermeasures has been agreed,
problems might continue to persist. Take the DSU, for example: as things stand,
there is no connection between the legislative postulate (to induce compliance)
and what happens in practice; the identity of the party taking the countermeas-
ures, rather than the countermeasures as such, may be the best indicator as to
whether compliance will be engendered/induced.
With this in mind, we can turn to some of the proposals advanced in litera-
ture: we should keep in mind, however, that designing the appropriate remedy
is not exactly a walk in the park, and literature has no off-the-shelf solutions that
could be imported lock stock and barrel into the WTO legal regime. Some early
papers, such as Becker (1968) link enforcement to the welfare implications in
case of enforcement (or non-enforcement): assuming certainty that violations
will always be detected and pursued, tough punishments (eg ‘I will shoot you

¹⁰⁶ On this score, see Horn and Mavroidis (2007), where the authors advance some thoughts
while presenting a survey of the law and economics literature on dispute settlement.
 Dispute Settlement 431

if you steal a chewing gum’) come very close to guaranteeing enforcement of

contractual obligations. However, such punishments go against some basic legal
principles, such as the proportionality principle.¹⁰⁷ More importantly, it is simply
not the case that all violations of the WTO Agreement are abusive in nature: the
WTO Agreement is a sufficiently incomplete contract to make room for legitim-
ate disagreements as to its precise scope. One should add that it is not always the
case that a solution provided by a WTO adjudicating body is the correct one: we
run the risk of enforcing Type I errors.
Barfield (2001) has argued that the basic flaw in the WTO edifice is between
the WTO consensus-plagued inefficient rule-making procedures and its highly
efficient dispute settlement system—an imbalance that creates pressure to ‘legis-
late’ new rules through adjudication. Barfield goes on to suggest two ways out of
this flaw: the Director-General should be given the prerogative to remove con-
troversial or ‘political’ disputes from the docket; additionally, a blocking minor-
ity of WTO members should be entitled to overrule a decision by the WTO
adjudicating bodies. The author thus supports strong enforcement of some, but
critically not all, WTO obligations. One of the problems with this approach is
the difficulty in distinguishing between cases that should and should not go to
WTO adjudication. One can possibly reach the same outcome without having
to operate such (quasi-) impossible tasks: Schwarz and Sykes (2002) claim that
the WTO remedies system should provide victorious plaintiff s with ‘expectation
damages’. By this, they mean to illustrate a situation which places the promisee
in as good a position as it would have been if the promisor had delivered on their
commitments. If the promisor is obliged to make the promisee no worse off than
it would have been had the contract been fulfilled, and can fulfil this obligation
even while breaking the contract, then the resulting breach is efficient. In their
view, the more money the victorious plaintiff receives, the better off he will be.
But an increase in the tariff binding beyond a certain level could actually make
the victorious plaintiff worse off. There is, hence, an inherent limit to the amount
of compensation that can be provided through higher tariffs alone. As a result,
were the WTO to institute a system based on ‘expectation damages’, it would
have to introduce additional payments that would compensate victorious plain-
tiffs for damages suffered when they introduce higher tariffs. The authors are thus
seeking to encourage the formal introduction of the efficient breach of contract
concept within the DSU.

¹⁰⁷ The proportionality principle presupposes a benchmark, and agreement on the benchmark
should not be taken for granted. Voicing their disagreement with the final ruling in the US—FSC
dispute, Esserman and Howse (2004) have strongly recommended caution when calculating the
amount of countermeasures. They suggest that Arbitrators should be well aware of the spill-overs
of their decisions and opening up the door to FSC-type of countermeasures (the worth of which
was more or less $4 billion) is, in their view, in disregard of the proportionality principle. Th at is, in
their view, using the amount of subsidy as benchmark to calculate the amount of countermeasures
does not respect the proportionality principle.
432 Institutional Provisions

Bronckers and van den Broek (2005), and Pauwelyn (2000) have forcefully
argued in favour of the introduction of monetary compensation. Ethier (2001)
and Schropp (2005) have argued in favour of tariff compensation, whereby the
author of the illegal act compensates affected parties by reducing the level of
bound duties in other tariff lines. What if, however, the WTO member requested
to provide it simply refuses to do so? We are back to square one. Mexico proposed
that the winning party should have the right to auction off its right to impose
countermeasures and potentially, receive some compensation from parties inter-
ested in purchasing such rights. Bagwell et al (2006), discussing this proposal,
add, in their extended model, that the author of the illegal act should be allowed
to bid. Such an extension, in their view, reduces the risk for auction failure.¹⁰⁸
The Mexican proposal, however, did not capture the imagination of the negotia-
tors. Limão and Shaggi (2005) added an interesting twist to this idea: in their
model, the WTO member who is the author of an illegal act will be requested to
pay monetary compensation. Assuming they refuse to do so, the WTO member
affected by their behaviour will be entitled to auction its right to impose counter-
measures. Recalcitrant states will thus suffer reputation costs twice: once when
they do not comply, and once when they refuse to pay monetary compensation.
Lawrence (2003) has proposed the introduction of contingent liberalization com-
mitments (CLCs) as a means to secure enforcement, and which in part responds
to the question as to what happens if the recalcitrant state refuses to pay.¹⁰⁹
What is probably an often forgotten dimension of WTO dispute settlement,
is that the DSU process matters, probably as much as its outcomes. Yes, it is true
that de facto prospective remedies only might give an incentive to WTO mem-
bers to buy a cheap exit from their obligations: this was probably the case with
US—Steel Safeguards. In this line of thinking, violations of the GATT contract
de facto become a substitute for lawful use of safeguards, any safeguards. This

¹⁰⁸ Bown and Hoekman (2005) correctly point out that developing countries might not be in
a position to impose countermeasures also because they have not liberalized enough: raising their
bound duties might not capture trade equivalent to the damage suffered. Th is is definitely a very
important observation. It loses, however, its significance in the Bagwell et al (2005) model: by auc-
tioning its right to take countermeasures, the retaliatory capacity of the auctioning state becomes
a non-issue (assuming successful auction). This is yet another reason why developing countries
might find it useful to support this initiative.
¹⁰⁹ Quoting from p 87: ‘In this approach, WTO Members would be given the option of offer-
ing a preauthorizing compensation mechanism during the Doha round negotiations. These offers
would be included in the multilateral negotiations. If a country’s offer is accepted, in the event it is
later found to have violated the agreement and failed to come into compliance, winning plaintiff s
would be authorized to select an equivalent package of concessions from the defendant’s commit-
ments. Countries could choose from several options in making their CLCs. They could indicate
a willingness to provide (selective) financial compensation, they could agree to provide across-
the-board (most favoured nation) tariff cuts to generate additional trade equal to the value of the
infraction, or they could agree to liberalize certain sectors on an MFN basis. Since the sectors to
be covered would be negotiated, in the multilateral setting, countries specializing in particular
exports (for example, textiles) could form alliances to ensure that products of interest to them
would be included in the commitments of important trading partners’.
Table 5.4 Waivers Currently in Force
Waivers Granted Expiry Decision
United States—Former Trust 14 October 1996 31 December 2006 WT/L/183
Territory of the Pacific Islands
Canada—CARIBCAN 14 October 1996 31 December 2006 WT/L/185
Preferential Tariff Treatment for 15 June 1999 30 June 2009 WT/L/304
Least-Developed Countries
EC—Autonomous Preferential Treatment to 8 December 2000 31 December 2006 WT/L/380
the Countries of the Western Balkans
EC—The ACP-EC Partnership Agreement 14 November 2001 31 December 2007 WT/L/436
Cuba—Article XV:6 of GATT 1994 20 December 2001 31 December 2006 WT/L/440
LLDCs—Article 70.9 of the TRIPS 8 July 2002 1 January 2016 WT/L/478

Dispute Settlement
Agreement with respect to
pharmaceutical products
Australia, Brazil, Canada, Israel, Japan, Korea, 15 May 2003 31 December 2006 WT/L/518
Philippines, Sierra Leone, Switzerland, Thailand,
United Arab Emirates and the United
States—Kimberley Process Certification
Scheme for rough diamonds
(Additional Members covered by the
waiver pursuant to procedures under
Paragraph 3 of the Decision: Botswana,
Bulgaria, Croatia, Czech Republic, European
Communities, Hungary, India, Malaysia, Mauritius,
Mexico, Norway, Romania; Separate Customs
Territory of Taiwan, Penghu, Kinmen and Matsu;
Slovenia, Switzerland, Venezuela)
Albania—Implementation of Specific
Concessions—Extension of the staging period
of implementation for a number of products 26 May 2005 • 1 January 2007 or 1 January WT/L/610
2009 for Part A of the Annex

433
(Continued )
 434
Table 5.4 (Continued)
Waivers Granted Expiry Decision
• January 2005 or 1 January 2007
for Part B of the Annex
Israel—Introduction of Harmonized 1 December 2005 31 October 2006 WT/L/639
System 1996 Changes into WTO
Schedules of Tariff Concessions
Argentina; Australia; Brazil; Bulgaria; Canada; 1 December 2005 31 December 2006 WT/L/638
China; Costa Rica; Croatia; El Salvador;
European Communities; Hong Kong, China;
Iceland; India; Republic of Korea; Macao,

Institutional Provisions
China; Mexico; New Zealand; Nicaragua;
Norway; Romania; Singapore; Switzerland;
Chinese Taipei; Thailand; United States;
Uruguay—Introduction of Harmonized
System 2002 Changes into WTO
Schedules of Tariff Concessions
Panama—Introduction of Harmonized 28 July 2006 30 April 2007 WT/L/652
System 1996 Changes into WTO
Schedules of Tariff Concessions
Argentina—Introduction of Harmonized 28 July 2006 30 April 2007 WT/L/653
System 1996 Changes into WTO
Schedules of Tariff Concessions
Senegal—Waiver on minimum values 28 July 2006 30 June 2007 WT/L/655
in regard to the Agreement on the
Implementation of Article VII of
General Agreement on Tariffs
and Trade 1994
European Communities—European 28 July 2006 31 December 2011 WT/L/654
Communities’ preferences for Albania,
Bosnia and Herzegovina, Croatia,
Serbia and Montenegro, and the
Former Yugoslav Republic of Macedonia
 Waivers 435

is a problem that deserves the attention of the negotiators. To distinguish, how-

ever, between bad faith recourses to dispute settlement (cases, that is, where the
defendant knows that it is acting wrongfully and will not address its behaviour
prior to a definitive ruling by the WTO) on the one hand, and cases of genuine
ambiguity is a monumental task. Once again one will need to rely on proxies,
such as the identity/very close smiliarity of factual aspects across old and new
cases, the absence of new, better theory to address such transactions, etc. At the
end of the day though, such cases should not detract from the overall contribu-
tion of the WTO dispute settlement process: even when dealing with bad faith
recourses to dispute settlement, WTO adjudicating bodies will have the oppor-
tunity to revisit past case-law, rethink its merits and confirm it if nothing has
changed in between cases. Case-law will thus be progressing in a linear manner.
At the same time, trading partners will continue to submit their disputes, even
the abusive cases, to third party adjudication. This is what Art 23.2(a) of the DSU
essentially amounts to. The process promotes a culture for peaceful resolution, a
unique paradigm in international relations. Moreover, it is quite hard to measure
the positive external effects of the WTO process: trading partners, influenced by
adjudication before the WTO, might be willing to emulate their WTO experi-
ence in other areas of international relations. While criticism of specific instances,
such as lack of retroactive remedies, or lack of transparency as to the criteria for
selection of panelists by the Director-General is welcome, one should always be
mindful of the broader picture as well.

5 Waivers

A WTO member may, in exceptional circumstances, request that it be exempted

from its obligations under the WTO. To this effect, a member must submit a
request for a waiver to the WTO membership. WTO Agreement, Art IX regu-
lates the procedure to this effect. Waiver decisions are to be taken by consensus,
and if this proves not to be feasible, by a three-quarters majority of all WTO
members.
The panel on EC—Bananas III was called upon to address the argument
advanced by the European Community that waivers are not justiciable. In this
case the European Community had requested a waiver provisionally exempting
it from having to bring the Lomé Convention¹¹⁰ in consistency with its obliga-
tions. The panel held the view that it had the requisite jurisdiction to interpret the
contents of a waiver granted by the WTO members (§ 7.97). On appeal, the AB
upheld this view (§ 167).

¹¹⁰ Under this arrangement, the European Community was treating imports from some devel-
oping countries better than it treated imports from others.
436 Institutional Provisions

The same panel was of the view that a waiver should be construed narrowly.
It went on to hold, nevertheless, that, despite the fact Art XIII of the GATT
was not explicitly mentioned in the waiver granted to the European Community,
that provision should nevertheless be applicable by the terms of the waiver,
which were indicative of its eff et utile (§§ 7.105–7.108). Hence, in the panel’s
view, what matters is the rationale for granting a waiver, and it was prepared to
set aside textual arguments that would put into question whether the ambit of
the waiver was consistent with its rationale. The panel took the view that the ten-
sion between interpreting exceptions such as waivers narrowly, and the eff et utile
of the exception itself, should not render the exception unworkable. On appeal,
the AB reversed the panel’s findings and held the view that only what is explicitly
reflected to in a waiver should be understood to be covered by the terms of the
waiver (§§ 182–8).
Table 5.4 provides the list of waivers currently in force.

6 Transparency

6.1 Introductory remarks

The GATT imposes a general transparency obligation: by virtue of Art X of the
GATT, WTO members are required to publish all laws affecting trade of general
application, and to administer them in an unbiased (vis-à-vis foreign products)
manner. Such laws can of course relate to either trade or domestic instruments, or
indeed both together.¹¹¹ GATT, Art X reflects bottom up transparency, since the
obligation to notify lies with the WTO Members. As of 1989, the WTO added
another weapon to its transparency arsenal: the TPRM, whereby the WTO
Secretariat prepares reports regarding the conduct of trade policies by WTO
members. The TPRM thus involves top down transparency, since this time it is
the institution providing the information.
The importance of transparency requirements should not be under estimated.
Information about some trade instruments (tariffs) is easily made available: it suf-
fices to ask a question to the WTO Secretariat. The other two trade instruments
(QRs, and export subsidies) are illegal; it is hence, legally impossible to imagine
transparency requirements in this context. Domestic instruments, however, do
not have to be notified to the WTO. Interested traders could be severely ham-
pered in their efforts to penetrate an export market absent a transparency require-
ment: Art. X of the GATT aims to address the information asymmetry between,
on the one hand, domestic traders who are presumed to know their own law, and,
on the other, their foreign counterparts.

¹¹¹ As such, the classification of the discussion on transparency in Chapter 5 of this volume is
abusive. The counterfactual, however, is not promising either since the same discussion would have
to appear either twice, or once partly improperly classified.
 Transparency 437

6.2 Trade Policy Review Mechanism (TPRM)

The TPRM was established in 1989 on a provisional basis and was continued pursu-
ant to Annex 3 of the Agreement Establishing the WTO. This Annex required that
the operation of the TPRM was to be evaluated by the Trade Policy Review Body
(the WTO body administering the TPRM) in 1999. The evaluation concluded that
the TPRM functions effectively, and, as a consequence, the TPRM has become a
permanent feature of the WTO legal edifice.¹¹² In accordance with Annex 3 of the
Agreement Establishing the WTO, the objectives of the TPRM are:
to contribute to improved adherence by all Members to rules, disciplines and com-
mitments made under the Multilateral Trade Agreements and, where applicable, the
Plurilateral Trade Agreements, and hence to the smoother functioning of the multilateral
trading system, by achieving greater transparency in, and understanding of, the trade
policies and practices of Members.¹¹³
To this effect, the TPRM is required to periodically review the trade policies and
practices of all members. The periodicity of the reviews depends on the shares of world
trade of the countries being reviewed: WTO members such as the United States and
the European Community will be reviewed every 2 years, whereas members with
less significant shares will be reviewed every 4 or 6 years, or even less frequently so.¹¹⁴
Members of the WTO Secretariat visit WTO members to conduct the review (from
a special unit called the WTO Trade Policy Review Mechanism Division). Since the
WTO Secretariat does not have the authority to interpret the covered agreements, a
TPRM report on the trade policies and practices of a member does not provide an
assessment as to the legal consistency of particular policies with the multilateral agree-
ments. Occasionally, the WTO Secretariat reports will be prepared with the help of
outside consultants (this has been the case for the reviews of Maldives, Niger, Senegal,

¹¹² See WTO Doc WT/MIN(99)/2 of 8 October 1999.

¹¹³ There is a de facto asymmetry across WTO members as far as procurement of information
about other members’ trade practices are concerned. Large WTO members with export diversifi-
cation will learn about practices of their partners from their private sector. It is also expected that
diplomatic missions will be established where private interests are prevalent. Hence, the European
Community can learn about others’ practices through its private sector (exporting all over the
world), its many EC missions as well as the missions of its member states. This is not the case for a
typical sub-Saharan state. In this sense, the TPRM provides some information that is more valu-
able to the sub-Saharan state rather than to the European Community. But then, the typical sub-
Saharan state will know enough about EC and US practices, since these will be the two preferred
export markets. If at all, they will need to know about other countries’ markets (which as we will
see, are not reviewed as frequently as the EC and the US markets).
¹¹⁴ Hoekman and Mavroidis (2000) have criticized the criterion for periodicity. In their view,
the order should be reversed and Members at the early development stages with low shares of world
trade should be reviewed more frequently than say the European Community or the United States.
Their argument is that for many members of the WTO, the WTO is policy irrelevant and absent
some periodic monitoring, they will never fully benefit from the benefits associated with partic-
ipation in the world trading system. Transparency is, the argument goes, the first step towards
realization of such benefits. In their view, the TPRM exercise should not be perceived only as an
instrument to provide information about other countries’ practices, but critically, as an instrument
intended to accustom the reviewed entity with the WTO disciplines.
438 Institutional Provisions

and the Southern African CU). The WTO Secretariat’s reports are written in close
consultation with the authorities of the member under review.
The increased importance given to the reviews of LLDCs has led to 14 reviews
of LLDCs since 1998. The LLDCs that have been reviewed since the establish-
ment of the mechanism are: Bangladesh (twice), Benin, Burkina Faso, Burundi,
Guinea, Lesotho (twice), Madagascar, Malawi, The Maldives, Mali, Mauritania,
Mozambique, Niger, Senegal (twice), the Solomon Islands, Tanzania, Togo,
Uganda (twice), and Zambia (twice).¹¹⁵
A WTO document¹¹⁶ reflecting the record of the TPRM as of mid-2003 con-
cluded that:
while each review highlights the specific issues and measures concerning the individual
Member, certain common themes emerged during the course of the reviews conducted in
the period January–July 2003. These included:
transparency in policy-making and implementation; economic environment and trade
liberalization; implementation of the WTO Agreements; regional trade agreements and
their relationship with the multilateral trading system; tariff issues, including peaks, escal-
ation, preferences, rationalization and the difference between applied and bound rates; cus-
toms clearance procedures; import and export restrictions and licensing procedures; the use
of contingency measures such as anti-dumping and countervailing duties; technical and
sanitary measures and market access; standards and their equivalence with international
norms; intellectual property rights legislation and enforcement; government procurement
policies and practices; state involvement in the economy and privatization programmes;
trade-related competition and investment policy issues; incentive measures such as subsid-
ies and tax forgone; sectoral trade-policy issues, particularly liberalization in agriculture and
certain services sectors; GATS commitments; special and differential treatment, including
market access and implementation, particularly for customs valuation, TRIPS and TRIMs;
small-island and small land-locked Members; and technical assistance in implementing the
WTO Agreements and the experience with the Integrated Framework.

Table 5.5 List of Countries Reviewed After the WTO Came into Force

1 Costa Rica 2 Côte d’Ivoire

3 European Union (3rd review) 4 Uganda
5 Mauritius 6 Sri Lanka
7 Slovak Republic 8 Morocco (2nd review)
9 Thailand (2nd review) 10 Venezuela
11 Dominican Republic 12 Czech Republic
13 Switzerland (2nd review) 14 Singapore (2nd review)
15 Norway (2nd review) 16 United States (4th review)
17 Zambia 18 Colombia (2nd review)
19 Republic of Korea (2nd review) 20 New Zealand (2nd review)
21 Brazil (2nd review) 22 Canada (4th review)
23 El Salvador 24 Fiji
25 Cyprus 26 Paraguay

¹¹⁵ See WTO Doc WT/TPR/134 of 27 June 2003.

¹¹⁶ See WTO Doc WT/TPR/134 of 27 June 2003.
 Transparency 439
Table 5.5 (contd.)

27 Benin 28 Chile (2nd review)

29 Mexico (2nd review) 30 European Union (4th review)
31 Malaysia (2nd review) 32 Japan 4th review
33 India (2nd review) 34 South Africa (2nd review)
35 Botswana 36 Lesotho
37 Namibia 38 Swaziland
39 Nigeria (2nd review) 40 Hungary (2nd review)
41 Australia (3rd review) 42 Jamaica
43 Mali 44 Turkey (2nd review)
45 Solomon Islands 46 Burkina Faso
47 Argentina (2nd review) 48 Togo
49 Trinidad & Tobago 50 Uruguay (2nd review)
51 Indonesia (3rd review) 52 Hong Kong, China (3rd review)
53 Canada (5th review) 54 Rep of Guinea
55 Egypt (2nd review) 56 United States (5th review)
57 Bolivia (2nd review) 58 Israel (2nd review)
59 Philippines (2nd review) 60 Romania (2nd review)
61 Nicaragua 62 Papua New Guinea
63 Thailand (3rd review) 64 Kenya (2nd review)
65 Iceland (2nd review) 66 Tanzania
67 Singapore (3rd review) 68 Bangladesh (2nd review)
69 Peru (2nd review) 70 Norway (3rd review)
71 Poland (2nd review) 72 European Union (5th review)
73 Rep of Korea (3rd review) 74 Bahrain
75 Brazil 76 Japan (5th review)
77 Switzerland and Liechtenstein 78 Canada (6th review)
79 Mozambique 80 Madagascar
81 Ghana (2nd review) 82 Macau, China (2nd review)
83 Costa Rica (2nd review) 84 Brunei Darussalam
85 OECS-WTO members 86 Gabon
87 Cameroon (2nd review) 88 United States (6th review)
89 Czech Republic (2nd review) 90 Mauritius (2nd review)
91 Slovak Republic (2nd review) 92 Malaysia (3rd review)
93 Uganda (2nd review) 94 Guatemala
95 Pakistan (2nd review) 96 Malawi
97 Mexico (3rd review) 98 Slovenia
99 Haïti 100 India (3rd review)
101 Barbados 102 European Union (6th review)
103 Mauritania 104 Australia (4th review)
105 Dominican Republic (2nd review) 106 Zambia (2nd review)
107 Japan (6th review) 108 Venezuela (2nd review)
109 Hong Kong, China (4th review) 110 Maldives
111 El Salvador (2nd review) 112 Canada (7th review)
113 Burundi 114 SACU¹¹⁷
115 New Zealand (3rd review) 116 Morocco (3rd review)
117 Indonesia (4th review) 118 Niger
119 Senegal (2nd review) 120 Honduras
121 Bulgaria 122 Guyana
123 Thailand (4th review) 124 Chile (3rd review)

¹¹⁷ SACU stands for South African Customs Union.

440 Institutional Provisions
125 Turkey (3rd review) 126 United States (7th review)
127 The Gambia 128 Sri Lanka (2nd review)
129 Rwanda 130 Singapore (4th review)
131 Benin (2nd review) 132 Burkina Faso (2nd review)
133 Mali (2nd review) 134 Belize
135 Suriname 136 European Communities (7th review)
137 Rep of Korea (4th review) 138 Norway (4th review)
139 Jamaica (2nd review) 140 Brazil (4th review)
141 Switzerland and 142 Japan (7th review)
Liechtenstein (2nd review)
143 Sierra Leone 144 Qatar
145 Mongolia 146 Paraguay (2nd review)
147 Nigeria (3rd review) 148 Ecuador
149 The Philippines (3rd review) 150 Egypt (3rd review)
151 Trinidad and Tobago 152 Tunisia (2nd review)
(2nd review)
153 Guinea (2nd review) 154 Bolivia (3rd review)
155 Romania (3rd review) 156 Malaysia (4th review)
157 Israel (3rd review) 158 Angola
159 Djibouti 160 United States (8th review)
161 People’s Rep of China 162 United Arab Emirates
163 Uruguay (3rd review) 164 Iceland (3rd review)
165 Chinese Taipei 166 Togo (2nd review)
167 Nicaragua (2nd review) 168 Bangladesh (3rd review)
169 Republic of Congo 170 Kyrgyz Republic
171 EAC (Kenya (3), Tanzania 172 Colombia (3rd review)
(2),Uganda (3))

6.3 Art X of the GATT

6.3.1 The discipline in a nutshell
GATT, Art X incorporates/details the general transparency obligation featured in
the GATT. The GATT, and the other Annex 1A agreements, contains various spe-
cific transparency obligations, such as the obligation to notify safeguard measures,
etc. GATT, Art X entails a broader set of responsibilities going beyond these spe-
cific requirements: in § 1, it imposes an obligation to publish all laws, adminis-
trative determinations and judicial decisions of general application affecting trade.
The coverage of Art X of the GATT hinges on the interpretation of the term ‘general
application’. A standstill obligation is imposed by § 2: state acts covered by the first
paragraph will not be enforced before their publication if they represent a new or
more burdensome requirement on imports. Finally, § 3 imposes the obligation that
state acts covered by the first paragraph be administered in a uniform, reasonable,
and impartial manner. The panel report on Argentina—Hides and Leather acknowl-
edged that the satisfaction of the first paragraph is a threshold issue for a review of
a claim under Art X.3 of the GATT. Importantly, the same report recognizes that
Art X of the GATT underscores that the obligation to publish all state acts of gen-
eral application is assumed vis-à-vis individual traders. It is one of the very rare
 Transparency 441

obligations which, by legislative imperative, pertains directly to private parties, that

is foreign producers interested in exporting to the market in question.

6.3.2 The coverage

GATT, Art X covers any state act of general application. Case-law has contributed
two important clarifications.
First, that Art X of the GATT is not concerned with the substantive consist-
ency of a particular act with the GATT rules. This is the subject-matter of other
GATT provisions. GATT, Art X is concerned with the administration of laws.
The AB report on EC—Bananas III held as much in § 200:
The context of Article X:3(a) within Article X, which is entitled “Publication and
Administration of Trade Regulations”, and a reading of the other paragraphs of Article
X, make it clear that Article X applies to the administration of laws, regulations, decisions
and rulings. To the extent that the laws, regulations, decisions and rulings themselves are
discriminatory, they can be examined for their consistency with the relevant provisions
of the GATT 1994.
By the same token, the AB report on EC—Poultry dismissed a claim by Brazil
which in part concerned the substantive consistency of an EC measure. In this
case, the AB entertained a claim by Brazil to the effect that an EC measure relat-
ing to imports of frozen poultry meat did not allow Brazilian traders to know
whether a particular shipment would be subject to the rules governing in-quota
trade or to rules relating to out-of-quota trade. Brazil maintained that this was a
violation of Art X of the GATT. The AB held that this claim concerned in part
the substantive consistency of the EC measure with the GATT rules; this part of
the claim lay outside the coverage of Art X of the GATT (§ 115):
Thus, to the extent that Brazil’s appeal relates to the substantive content of the EC rules
themselves, and not to their publication or administration, that appeal falls outside the
scope of Article X of the GATT 1994. The WTO-consistency of such substantive con-
tent must be determined by reference to provisions of the covered agreements other than
Article X of the GATT 1994. (original emphasis)

In EC—Selected Customs Matters, the AB provided a noteworthy precision: the

substantive content of the instrument that regulates the administration can of course
be challenged. It is the substantive content of instruments being administered that
cannot be challenged. This distinction has of course implications with respect to
the type of evidence that the complainant must submit (§§ 200–1):
The statements of the Appellate Body in EC—Bananas III and EC—Poultry do not
exclude, however, the possibility of challenging under Article X:3(a) the substantive con-
tent of a legal instrument that regulates the administration of a legal instrument of the
kind described in Article X:1. Under Article X:3(a), a distinction must be made between
the legal instrument being administered and the legal instrument that regulates the
application or implementation of that instrument. While the substantive content of the
442 Institutional Provisions
legal instrument being administered is not challengeable under Article X:3(a), we see no
reason why a legal instrument that regulates the application or implementation of that
instrument cannot be examined under Article X:3(a) if it is alleged to lead to a lack of
uniform, impartial, or reasonable administration of that legal instrument.
This distinction has implications for the type of evidence required to support a claim of a
violation of Article X:3(a). If a WTO Member challenges under Article X:3(a) the substan-
tive content of a legal instrument that regulates the administration of a legal instrument
of the kind described in Article X:1, it will have to prove that this instrument necessarily
leads to a lack of uniform, impartial, or reasonable administration. It is not sufficient for the
complainant merely to cite the provisions of that legal instrument. The complainant must
discharge the burden of substantiating how and why those provisions necessarily lead to
impermissible administration of the legal instrument of the kind described in Article X:1.¹¹⁸
Second, it held that the obligation included in Art X of the GATT covers all acts
of general application, irrespective of whether they have been published or not.
Argentina argued before the panel on Argentina—Hides and Leather, that, since
Art X of the GATT imposes a transparency obligation, it should apply only to
unpublished acts and sanction government behaviour to this effect. The panel
rejected this argument (§ 11.71). In its view, such an interpretation would run
counter to the explicit wording of Art X.1 of the GATT.

6.3.3 The standard of review

The standard of review was discussed in a comprehensive manner before the panel
on Argentina—Hides and Leather. Argentina enacted Resolution 2235/96 which,
inter alia, made it possible for representatives of the Argentine tanning indus-
try (Association of Industrial Producers of Leather, Leather Manufactures and
Related Products (ADICMA)), to be present during customs clearing procedures
concerning exports of bovine hides. In the view of the European Community, the
presence of ADICMA representatives amounted to a de facto export restriction
prohibited by Art XI of the GATT (since, its argument goes, ADICMA represent-
atives would have an interest in keeping bovine hides in the Argentine market, in
order to keep the price of the input low; knowledge by exporters on the other hand,
that ADICMA representatives were present at counting, would provide them with
an incentive not to export since, had they done so, they risked being excluded from
the Argentine downstream market. This would be the case if ADICMA blacklisted

¹¹⁸ By attempting to place the procedural and substantive aspects of the law into hermetically
sealed boxes so to speak, the relevant WTO body may end up with a problematic interpretation:
for instance, the manner in which a law is administered should pay due regard to its substantive
nature; in essence, it is surely impractical to strike down administrative procedures without at least
inquiring into the rationale behind the law in question. I do not mean to support the thesis that a
WTO panel should, when entertaining a complaint that Art X of the GATT has been violated, it
should take the liberty to pronounce on the substantive consistency of a law with another GATT
provision. An inquiry into the rationale of the law nevertheless, might occasionally prove a useful
stepping stone in constructing its findings with respect to the consistency of the challenged meas-
ure with Art X of the GATT.
 Transparency 443

all upstream producers exporting hides, rather than selling them to the Argentina
downstream industry). The bulk of the complaint by the European Community
focused on the inconsistency of this practice with Art XI of the GATT. The claim
under Art X of the GATT focused on the administration of Resolution 2235/96.
More specifically, the European Community claimed that, in administering
Resolution 2235/96 in an unreasonable, impartial and not uniform manner,
Argentina was violating its obligations under Art X.3 of the GATT. In this con-
text, the question before the panel was whether, for the European Community to
meet its burden of proof, it must show actual effects (trade damage), as a result of
the administration of the contested legislation. The panel dismissed this thesis. In
its view, it sufficed for the European Community to demonstrate the likelihood
that the interests of traders will be negatively affected from the administration of
Resolution 2235/96, and no actual trade damage had to be shown (§ 11.77):
Thus, it can be seen that Article X:3(a) requires an examination of the real effect that a
measure might have on traders operating in the commercial world. This, of course, does
not require a showing of trade damage, as that is generally not a requirement with respect
to violations of the GATT 1994. But it can involve an examination of whether there is a
possible impact on the competitive situation due to alleged partiality, unreasonableness
or lack of uniformity in the application of customs rules, regulations, decisions, etc.

6.3.4 Laws of general application

A number of WTO reports have dealt with this issue. The panel report on
US—Underwear held that a safeguard imposed by the United States on imports
of cotton was a measure of general application. The relevant criterion, in the pan-
el’s view, was that the measure at hand applied, in principle, to an unidentified
number of economic operators and not to specific shipments (§ 7.65). On appeal,
the AB upheld this point of view (p 21).
A contrario, measures applicable to specific shipments only should not come
under Art X of the GATT. The AB, in its report on EC—Poultry, provided an
explicit confirmation of this point. The AB was facing an appeal against a panel
finding to the effect that import licenses issued by the European Community to
specific companies interested in importing poultry product, or import licenses
applied to specific shipments (poultry products) to the European Community,
were not measures of general application, and thus, not justiciable under Art X of
the GATT. The AB upheld the panel’s findings in this respect (§ 113).

6.3.5 Uniform, reasonable and impartial administration

GATT, Art X.3 imposes on WTO members the obligation to ensure that laws
covered by Art X of the GATT are to be administered in a uniform, reasonable
and impartial manner. The panel, in its report on Argentina—Hides and Leather
held the view that these are three distinct requirements (§ 11.86). The same panel
had an opportunity to explain its views on each one of the three requirements. We
444 Institutional Provisions

recall that the European Community complained that the presence of represent-
atives of the tanning industry in customs clearance procedures was in violation of
Art X.3 of the GATT because such a presence did not guarantee a uniform, rea-
sonable and impartial administration of customs clearance procedures, in light of
the fact that the Argentine tanning industry had a vested interest in reducing (if
not eliminating altogether) exports of bovine hides.

6.3.5.1 Uniform administration

The European Community argued that the Argentine measure was not uni-
formly administered, and thus inconsistent with Art X.3 of the GATT, because it
was applied to bovine hides only and not to other products. The panel dismissed
this claim. In its view, what mattered was not whether the measure was discrim-
inatory across products or not, but whether it provided traders with predictabil-
ity as to future transactions (§§ 11.83–11.85). The panel here dismissed the EC
claim essentially on grounds that it concerned the substantive consistency of the
measure with the GATT, an issue which escapes the standard of review employed
by panels on claims under Art X of the GATT.
The panel report on EC—Selected Customs Matters held that the obligation for uni-
formity covers, inter alia, geographic uniformity. This panel held that the European
Community was in violation of Art X of the GATT because identical products,
within its sovereignty, were subjected to different customs treatment (§§ 7.135 ff ).¹¹⁹

6.3.5.2 Reasonable administration

The European Community claimed before the panel (Argentina—Hides and
Leather), that the Argentine measure was in breach of Art X.3 of the GATT, since it
provided representatives of ADICMA with possibility of having access to business
confidential information: they could namely learn the names of the exporters, and
their pricing schemes. The panel upheld the claim since, in its view, such informa-
tion could then be used to confer an advantage to the Argentine tanning industry,
when negotiating with the upstream segment of this market (§§ 11.92–11.94):
To provide some specific examples, ADICMA representatives should not be able to see the
pricing information of the suppliers to ADICMA’s members. This is information which
ADICMA members could use to their commercial advantage in negotiations with the frig-
oríficos. We should note in this regard that Argentina bases its export duties on prices of
hides quoted in the United States. Thus, even if we were to consider it reasonable for the

¹¹⁹ By the same token, the same panel found that differential customs administration of LCD
monitors within the European Community was tantamount to a violation of Art X of the GATT
(§ 7.305), the AB upholding this finding (§ 260). In this panel’s view, uniformity should not be
confused with identity. Moreover, the degree of uniformity should be determined on a case by case
basis: the narrower the challenge (that is, the more specific the aspects of the administrative action
complained about), the higher the degree of uniformity required (and, consequently, the lower the
burden for the complainant to make a prima facie case). In a rather cryptic passage the precedence
value of which can legitimately be put into question, the panel held that Art XXIV.12 of the GATT
is not an exception to Art X.3(a) of the GATT.
 Transparency 445
tanners to be involved in the export clearance process, there would be no reason whatever
for them to see the prices as these would be irrelevant to the assessment of export duties. We
also see no need for them to be made aware of the destination or quantities involved as these
data are irrelevant to the tasks ADICMA representatives are involved in.
We think it is particularly important for the reasonable administration of Argentina’s
export laws that the tanners not be provided the name of exporters. Argentina claims that
this is no longer possible. However, as it was part of the European Communities’ claims
and was unarguably possible as recently as May of 1999 that such written information was
supplied to ADICMA, we consider it necessary to specifically find that it is unreasonable
for such information to be provided to ADICMA or its members. However, this question
goes beyond just supply of the name in writing. Argentina has stressed in its arguments
under all three conditions of Article X:3(a) that the process is balanced because the export-
ers may be present during the Customs process. However, it necessarily follows that exer-
cising this right would reveal the identity of the exporter. While it could be argued that the
exporter could send a representative or agent and may thereby conceal his identity, impos-
ing such a burden with respect to an exporter’s own products would be unreasonable.
Therefore, we must conclude that a process aimed at assuring the proper classification of
products, but which inherently contains the possibility of revealing confidential business
information, is an unreasonable manner of administering the laws, regulations and rules
identified in Article X:1 and therefore is inconsistent with Article X:3(a). (original emphasis)

6.3.5.3 Impartial administration

Before the panel on Argentina—Hides and Leather, the European Community
argued that the Argentine measure was in breach of Art X.3 of the GATT: in its
view, the law, which allowed for the representation of the Argentine industry during
customs clearance, was not being administered in an impartial manner. In its view,
there was absolutely no legitimate reason why ADICMA representatives should be
present during customs clearance of bovine hides. Moreover, information obtained
by their mere presence, could potentially be used to their own advantage (as already
discussed supra). Hence, there was at least potentially a case of conflict of interest.
The panel upheld this claim, since it agreed that the potential that the Argentine
industry present in customs clearance could use confidential information to its
own advantage, amounted ipso facto to a violation of the impartial administration
requirement embedded in Art X of the GATT (§§ 11.99–11.101).

6.3.6 The obligation to maintain independent tribunals

WTO members must, by virtue of Art X.3(b) of the GATT, establish (judi-
cial, arbitral, administrative) tribunals (or procedures) that will be independ-
ent from the agency the acts of which they will be reviewing. Traders should
have recourse to such tribunals or procedures. The panel report on EC—Selected
Customs Matters underscored this requirement for impartiality of tribunals (§
7. 519). In the same report, the panel also dealt with the question of the legal
effect of a reviewing tribunal’s decision. The question was to what extent such a
decision should govern the practice of all agencies entrusted with administrative
446 Institutional Provisions

enforcement throughout the territory of the WTO member concerned. The panel
held that such a reading of Art X.3(b) of the GATT was unwarranted. It based its
view on its understanding that the tribunals mentioned in Art X.3(b) were in all
likelihood first instance tribunals and requiring such an effect of their decisions
could raise public order type of concerns.¹²⁰
The United States appealed this finding. The AB had, thus, the opportunity to
clarify its understanding of the obligation included in Art X.3(b) of the GATT.
The AB upheld the panel’s finding to the effect that the tribunals envisaged in the
first sentence of this provision are first instance tribunals (§ 294). The AB, based
on textual and contextual arguments (first instance courts do not extend their
jurisdiction throughout the territory of a WTO member; the existence of second
instance courts supports this thesis, §§ 298–9), upheld the panel’s findings on
this score in the following terms (§ 303):
For these reasons, we are of the view that Article X:3(b) of the GATT 1994 requires a WTO
Member to establish and maintain independent mechanisms for prompt review and cor-
rection of administrative action in the area of customs administration. However, neither
text nor context nor the object and purpose of this Article require that the decisions eman-
ating from such first instance review must govern the practice of all agencies entrusted with
administrative enforcement throughout the territory of a particular WTO Member.

7 Relation to the Havana Charter

Recall from our discussion in Chapter 1, that Art XXIX of the GATT which
explicitly regulates the legal relationship between the GATT and the ITO. In § 1
it states:
The contracting parties undertake to observe to the fullest extent of their executive author-
ity the general principles of Chapters I to VI inclusive and of Chapter IX of the Havana
Charter pending their acceptance of it in accordance with their constitutional procedures.
The ITO never entered into force. Its entry into force is no longer pending, since
the WTO now takes the place that the ITO was originally supposed to occupy. By
saying no to the ITO and yes to the WTO, WTO members have endorsed a regime
with no competition type disciplines.¹²¹ It follows that the best endeavours clause
to observe the general principles of Chapters I to VI of the Havana Charter should
not be considered active any more: as a result, WTO members should not be held
liable if they do not observe the disciplines concerning RBPs included in Chapter V
of the Havana Charter, an area which has been left unregulated in the WTO.

¹²⁰ See the analysis leading to the final ruling in § 7.539. The panel, however, did not exclude
that such an effect could de facto be the case. The panel only ruled that there was no WTO obliga-
tion to this effect.
¹²¹ A working group aimed to discuss the interaction between trade and competition policies was
instituted in the aftermath of the Singapore Ministerial Conference (1996). The WTO members
decided, however, to discontinue its operation during the Cancun Ministerial Conference (2003).
 6
The GATT, Then and Now

1 The GATT 1947: A Solid Text (with some

compromises) 447
2 The Evolution of the GATT: No Keeping up with
the Joneses 448
3 Rule of Reason 451

1 The GATT 1947: A Solid Text (with some compromises)

The original GATT text is predicated on the following propositions:

(1) legal disciplines are imposed on both trade and domestic instruments
alike. Since both trade and domestic instruments affect trade, absent dis-
ciplines on the latter, the disciplines on trade instruments might be (easily)
circumvented;
(2) harder disciplines are imposed on trade instruments, since these instruments
affect trade directly: they apply on either domestic or imported goods. By
contrast, domestic instruments frequently pursue non-trade objectives, and
apply to both domestic and imported products;
(3) there is a preference for use of tariffs over quantitative restrictions (the latter
being outlawed), which is justified by the fact that tariffs are more trans-
parent, easier to administer and apply on an MFN-basis and invite less rent
seeking behaviour;
(4) negotiations are based on reciprocal commitments, which is both a means
to reduce (remove) terms of trade externalities, and a politically acceptable
vehicle for trade liberalization;
(5) non-discrimination is the insurance policy against concession erosion
which, if present, will negatively impact on the incentive to liberalize. Non-
discrimination, when applied to domestic instruments, is also the only
feasible way to integrate markets across vastly heterogeneous participants:
448 The GATT, Then and Now

heterogeneity makes other forms of cooperation (such as harmonization, or

mutual recognition) a remote, if at all, possibility;
(6) flexibility mechanisms, such as the possibility to renegotiate bound duties,
and the safeguards provision, provide countries with more of an incentive to
liberalize trade.
The basis of the GATT rests thus, on solid grounds. Some authors tend to over-
idealize the GATT, and see nothing wrong with its trade liberalization model.
This is an exaggeration. The GATT is also the product of some important polit-
ical compromises:
(1) the imperial preferences, so dear to the heart of some original members, were
maintained until the end of the Uruguay round. It is the UK government
that, according to sources cited in Chapter 1, insisted on their survival in
the GATT era. There is now substantial evidence that, the noble intentions
behind them notwithstanding, they hardly contributed to the aims sought
through them;
(2) AD became a GATT institution, largely because the US negotiators were
instructed not to agree to anything that would require substantial change
of pre-existing domestic US laws. AD in 1947 was very much part of the US
trade arsenal.¹ It is difficult to reconcile AD with economic rationale;
(3) trade in farm goods remained for years outside the purview of the GATT
disciplines. The US Agricultural Adjustment Act of 1933 introduced price
supports in the 1930s during the New Deal and Great Depression, and this
is probably the reason why the disciplines on farm goods were weaker than
those imposed on other goods (Art XI.2 of the GATT). No serious argument
has been mounted justifying an idiosyncratic treatment for farm goods.
One should, further, not neglect that some of the GATT provisions, their intellec-
tual coherence notwithstanding, presented a challenge to the interpreter (judge).
The non-discrimination clauses (Arts I and III of the GATT) are, for the reasons
mentioned in Chapters 2 and 3, among the first to spring to mind.

2 The Evolution of the GATT: No Keeping up with the Joneses

A question one could legitimately ask is whether, over the years, we managed to
address the distortions in the original GATT, correct them, and move on to a
new, more sensible agreement. The short answer is no.

¹ In the pre-GATT era, the US government did not have extensive recourse to AD. It seems thus
fair to conclude that the United States introduced a Trojan horse in 1947 which eventually became
a big problem in international trade relations.
 The Evolution of the GATT 449

For a start, one cannot fail to notice that even the simplest of legislative amend-
ments have not been inserted: what is, for example, the place for Art XXIX of the
GATT in today’s world? Does any one seriously believe that, in presence of the
WTO, the ITO project might be resurrected? But in an increasingly politicized
environment, some negotiators might have felt that opening up even the most
innocuous issue for amendment, they would, effectively, be opening up Pandora’s
box: negotiators cannot be accused for being unimaginative when designing quid
pro quos. It is not, hence, an accident that the only legislative amendment reported
so far is that of August 2005, where the TRIPs Agreement was amended so as
to make room for a change in the provision on compulsory licensing, an issue
very much requested by developing countries, the losing party since the advent of
TRIPs. There is thus, prima facie, an observed legislative inertia.
To some extent, nevertheless, this attitude is probably justified by the fact
that a number of GATT provisions, such as Art XXIX of the GATT, have been
superseded by new provisions contracted under the Agreement Establishing the
WTO.² Negotiators might, thus, rationally prefer not to initiate the painstaking
amendment procedures of the Agreement Establishing the WTO (Art X of the
GATT), since the very existence of the WTO amounts to definitively abandon-
ing the ITO endeavour. Desuetudo, in other words, has functioned as substitute
for amendment in this case.
The original GATT included Art VI. If the rationale for it was that we needed
an instrument to address international predation, the rationale vanished with the
wholehearted acceptance of the eff ects doctrine, first in antitrust practice in the
two sides of the Atlantic, and, eventually, with its spreading around the world.
Increasingly it is the case that the explanation offered for AD is that the WTO
regime lacks a user-friendly safeguard mechanism that will provide trading part-
ners with enough ‘breathing space’, when need be, to perform (painful) adjust-
ments at home. But if this is the case, then the source of the distortion is the
current safeguard clause which does not satisfy the needs of the trading partners.
Why not simply address the distortion at its source and fi x it, instead of adding
other safeguard clauses, and creating additional distortions? If what is required
is a safeguard clause that can target specific sources of supply only, then probably
one should be working on a reinforcement of the quota modulation provision in
the SG Agreement. If what is required is some sort of simplistic rule of thumb
(instead of the current convoluted regime where four conditions must be cumula-
tively met) that would allow use of safeguards, then probably this is where negoti-
ating efforts should concentrate.³ It seems, however, senseless to spend substantial

² GATT, Art XXXV for example, has been superseded by Art XIII of the Agreement
Establishing the WTO.
³ Had they decided to go down this road, they would find that there are proposals out there
aiming at addressing this issue: Bhagwati (1998, pp 119ff ), for example, proposed safeguard with
protective tariff to be explicitly set on a declining time-bound schedule, a drastically more user-
friendly option than the current regime.
450 The GATT, Then and Now

negotiating capital in every single trading round since the Kennedy round (not to
mention lobbying efforts in between rounds) trying to reduce the scope of discre-
tion that AD investigating authorities enjoy. And of course this is only the tip of
the iceberg: AD, as empirical analysis cited in Chapter 4 makes amply clear, has
led to substantial welfare losses around the world (stemming directly, or indir-
ectly from AD: VERs, non-transparent collusion are often supported by threats
of AD).
The first steps have been taken to bring farm goods under the aegis of the
multilateral rules. It has so far proved to be a painful negotiation, but at least a
framework is now in place and, eventually, the same rules will apply to both farm
and non-farm goods. As things stand, nevertheless, tariff protection on farm
goods constitutes one of the notorious tariff peaks.
And, probably for reasons extraneous to the GATT itself, imperial preferences
and all sorts of grandfathering have been abolished. This does not mean, however,
as we will see infra, that non-discrimination is de facto the dominant discipline
in international trade.
So, when we look at the failings in the original GATT, one has been addressed
with a caveat that we will discuss infra (grandfathering), one is on its way to being
addressed (trade in farm goods), and one has not been addressed at all; to the con-
trary it has become a mainstream GATT institution (AD).
And, of course, GATT practice unveiled new challenges, to which the system
was called to respond. The list is quite long. The original GATT had a few out-
liers, over the years, however, developing countries have come to represent the
majority in the world trading system. The term ‘developing countries’ is not very
helpful and, as it is now being used in the WTO, brings quite heterogeneous
countries under its aegis. The GATT/WTO has not discarded (even) abusive
invocations of developing country status. What is worse, empirical evidence
mentioned in Chapter 2 suggests that the special and differential treatment, the
GATT response to the requests by developing countries, has done very little, if
anything, to help those in need. Fifty years later, the world trading system is still
discussing this issue and still tenaciously refusing to accept the shortcomings of
the original recipe. To the contrary: the initiatives that see the light of the day
continue to compensate the poorer members of the world community by effect-
ively pushing them to non-participation, when what is required is exactly the
opposite: participation in the round, and painful adjustments.
The decline in the relative importance of tariff protection has naturally moved
the discussion to the realm of NTBs. The ensuing legislative initiatives (TBT,
and SPS) are thoughtful texts but once again very challenging for the interpreter,
an issue to which we return shortly.
The trading community has also decided to take action to minimize dead-
weight costs: the various agreements aimed at providing transparency to customs
and customs-related procedures (CV, PSI, and ILA), and the ongoing negotiation
on trade facilitation are evidence of this will. It is probably too early to draw any
 Rule of Reason 451

definitive conclusions on this score: on the one hand, the lack of disputes could
be interpreted as signaling that concerns about this issue are exaggerated. On the
other hand, the negotiation on trade facilitation was considered necessary to fill
gaps left by the previous legislative initiatives.
As stated above, the imperial preferences were eliminated over the years. The
two main exceptions to non-discrimination, nevertheless, PTAs and special and
differential treatment, proved a different story. PTAs have proliferated over the
years, for a variety of reasons. The lack of GATT surveillance of PTAs is a contrib-
uting factor to their proliferation. It has proved ineffective. As things stand, we
are not even in the beginning of the discussion on what should be done, assuming
trading nations do agree that something should be done in the first place. It has
been clear for some time now that the wording of Art XXIV of the GATT is not
helpful at all. Some legislative change is needed. However, if at all, the multiplica-
tion of PTAs is ample evidence that the revealed preference of trading nations is
for lax rather than meaningful review.
In general, during trading rounds, when quid pro quos are easier to identify,
the trading nations have managed to produce trade law; off trade rounds, how-
ever, legislative initiatives have been rare, in fact only one since 1995 (the TRIPs
amendment). This has prompted scholars, such as Barfield (2001) to speak of
an imbalance across the judiciary and the legislative functions of the GATT.
Although Barfield probably exaggerated the function of dispute settlement, he
is right to point out that the WTO bodies (the various committees, the coun-
cils, etc) have only rarely managed to do anything beyond household activities.
Developing countries, on the other hand, are being sub-divided following all sorts
of unilaterally defined criteria. They are in a race among themselves to attract the
extra preference, which might mean nothing much economically, but quite a lot
in terms of political capital for the ruling classes. The true issues have failed to
be addressed so far, and it seems quite unlikely that this will be the case during
the Doha round. Replacing special and differential treatment with something
else should not be part of a quid pro quo. It deserves the undivided attention of the
WTO membership, and it requests building bridges with other, better equipped
to deal with development issues, institutions.

3 Rule of Reason

The discussion above points to an ongoing inertia, an unwillingness to address

the issues: yes, new agreements are being agreed, but the problems remain; non-
discrimination is all but a smokescreen in today’s trading system, issues relating
to developing countries’ interests are a constant topic of discussion which, in and
of itself provides the proof that the previous talk on the same subject did not do
the job, and so on and so forth. Yet, it is difficult to find many institutions and
452 The GATT, Then and Now

many processes as present in international relations as the WTO. The WTO has
managed to persuade nations to commit to act collectively, and not unilaterally,
and this is its biggest success. It might take time to reach the wanted result, but
eventually it gets there.
We live in different times than our forefathers in the forties. There was a polit-
ical will there to re-do the world, and people rose up to that challenge. Visionaries
like Cordell Hull pushed for international cooperation through a multilateral
trade agreement, and able thinkers found their way in the key national delega-
tions. Recall that Keynes, Meade and Robbins participated, in one way or the
other, in the design of the GATT. The original GATT benefited from the think-
ing of some of the best economics minds of the past century. Recall further, our
discussion concerning the shift of competence over trade issues from Congress to
the US President: it was meant to immunize the process from undue influence by
particular segments of the society.
The identity of the participants has changed over time, and the political pres-
sure keeps mounting. These two ingredients are important contributing factors in
the changing face of the GATT. Vested interests become all the more prevalent,
and largely explain the flourishing of intellectually untenable institutions, such
as AD. They also largely explain the inertia to address the true concerns. It is true
that the GATT, and the trade liberalization process, gain in legitimacy when it
can rely on the support of the societies the life of which it will be affecting. Public
discussion about the winners and losers of trade liberalization is very much a
necessary ingredient of this support. The society should be aware that ongoing
trade liberalization should be viewed in terms of Kaldor Hicks efficiency, it is
not necessarily a Pareto superior outcome: some will lose, but the overall gains
will be enough to compensate the losers. In this vein, a change in the WTO is
largely predicated on a change in domestic societies. Modern democracies have
seen the rise of the administrative state: it was a democratic choice to keep central
banks and antitrust authorities away from mainstream political control, precisely
because they wanted to protect such institutions from important vested interests
(especially since it is simply not the case that all affected by the decisions of the
central bank or the antitrust authority can lobby equally effectively). Trade insti-
tutions have not enjoyed the same privilege.
The world trading system has entered a phase quite comparable to what anti-
trust authorities had to face with the emergence of rule of reason. The institu-
tional design of antitrust authorities had to undergo important changes as a
result: anyone, or almost so, can apply a clear rule of thumb (per se prohibitions);
it takes a substantial amount of training to perform a rule of reason analysis to
the intellectual satisfaction of the test itself. In today’s world, beggar thy neigh-
bour policies are couched in domestic instruments (the gradual elimination of
tariffs, and the total prohibition of import and export quotas are two important
contributing factors) which, in principle, affect both domestic and foreign prod-
ucts. There are two very crucial to the analysis questions: what is protectionist
 Rule of Reason 453

regulatory intervention? What is the permissible extent of domestic jurisdiction?

Economic thinking is struggling to come up with operational definitions of the
former. The current legal tools to discuss the latter (the eff ects doctrine) are highly
imperfect to address non-pecuniary external effects. In the meantime, cases like
EC—Asbestos, EC—Tariff Preferences, and US—Shrimp are being routinely sub-
mitted to WTO adjudication. The reasoning behind such cases might provide
the input to eventual legislative initiatives. So far, this has not been the case,
precisely because the reasoning (irrespective of the political acceptability of the
outcome) has failed to rise to the standard. WTO members, in the short run,
would be well-advised to spend some time thinking about the composition of
the adjudicating bodies, judges, and Secretariat alike. We are not arguing for
judicial activism, far from it. We have, hopefully, made it clear in this book that
the judge, in our view, is an agent, not a principal. It is, nevertheless, astonishing
that adjudication of essentially economic issues is performed by individuals not
trained in economics. Dispute settlement is the only genuine WTO voice. This
voice should be coherent, a pattern, and not unconnected dots on a canvas. Until
the legislative machinery becomes more efficient, the WTO judges must ensure
that they have not failed this remarkable endeavour, the GATT.
This page intentionally left blank
 Appendices

1 Table of GATT Panels 456

2 Table of WTO Panels 464
456 Appendices

Table of GATT Panels

Number BISD/DS Full title Short title

Report/Other
Ref No.

1 BISD II/12 Ruling by the Chairman: The Phrase Cuba—Consular

‘Charges of any Kind’ in Article I:1 in Taxes
Relation to Consular Taxes
2 BISD II/12 Ruling by the Chairman: Application of India—Tax
Article I:1 to Rebates on Internal Taxes Rebates
3 GATT/CP.2/43 Report of Working Party 7 on the Cuban Cuba—Textiles I
Schedule
4 BISD II/28 United States Export Restrictions US—Export
Restrictions
(Czechoslovakia)
5 BISD II/181 & Brazilian Internal Taxes Brazil—Internal
186 Taxes
6 BISD II/11 Margins of Preference US—Margins of
Preference
7 GATT/CP.3/82 Report of Working Party 8 on Cuban Cuba—Textiles II
Textiles
8 BISD II/188 The Australian Subsidy on Ammonium Australia—
Sulphate Ammonium
Sulphate
9 GATT/CP/106 Report of the Intersessional Working US—Fur Felt
Party on the Complaint of Czechoslovakia Hats
concerning the Withdrawal by the United
States of a Concession under the terms of
Article XIX
10 BISD 1S/28 Article XIX—Increase in the United US—Dried Figs
States Duty on Dried Figs
11 BISD 1S/59 Belgian Family Allowances (allocations Belgium—Family
familiales) Allowances
12 BISD 1S/51 Increase of Import Duties on Products Greece—Import
included in Schedule XXV (Greece) Duties
13 BISD 1S/48, Special Import Taxes Instituted by Greece Greece—Import
7S/69 Taxes
14 BISD 1S/53, Treatment by Germany of Imports of Germany—
1S/30 & 7S/69 Sardines Sardines
15 BISD 1S/62 Netherlands Action Under Article US—Suspension
XXIII:2 to Suspend Obligations to the of Obligations
United States
16 BISD 3S/81, Swedish Anti-Dumping Duties Sweden—AD
7S/69 Duties
 Appendices 457

Number BISD/DS Full title Short title

Report/Other
Ref No.
17 BISD 3S/26, French Special Temporary Compensation France—
4S/20, 5S/27, Tax on Imports Compensation
7S/68 Tax
18 BISD 3S/77 German Import Duties on Starch and Germany—
Potato Flour Starch Duties
19 L/575, L/580 Greece—Increase in Bound Duty Greece—
Phonograph
Records
20 BISD 7S/23, Italian Discrimination Against Imported Italy—
7S/60, 7S/22 Agricultural Machinery Agricultural
Machinery
21 BISD 7S/46, French Assistance to Exports of Wheat France—Wheat
7S/22 and Wheat Flour Exports
22 SECRET/105 Article I—United Kingdom Waiver UK—Ornamental
(Ornamental Pottery) Pottery
23 L/1749 United Kingdom Waivers—Application UK—Bananas
in Respect of Customs Duties on Bananas
24 BISD 11S/94 & French Import Restrictions France—Import
11S/55 Restrictions
25 BISD 11S/88 & Exports of Potatoes to Canada Canada—
11S/55 Potatoes
26 BISD 11S/95 Uruguayan Recourse to Article XXIII Uruguay—
Recourse to
Article XXIII
27 L/2088 US/EEC—Panel on Poultry US/EEC– Poultry
28 BISD 18S/183 Jamaica—Margins of Preference Jamaica—
Margins of
Preference
29 BISD 18S/179 Greece—Preferential Tariff Quotas to the Greece—USSR
USSR Tariff Quotas
30 BISD 20S/237 United Kingdom Import Restrictions on UK—Cotton
Cotton Textiles Textiles
31 BISD 20S/230 & United Kingdom—Dollar Area Quotas UK—Dollar
20S/236 Quotas
32 BISD 23S/91 Canadian Import Quotas on Eggs Canada—Eggs
33 BISD 25S/49 EEC—Measures on Animal Feed Proteins EEC—Animal
Feed Proteins
34 BISD 25S/68 EEC—Programme of Minimum Import EEC—Minimum
Prices, Licences and Surety Deposits for Import Prices
Certain Processed Fruits and Vegetables
35 BISD 25S/42 Canada—Withdrawal of Tariff Canada—Lead
Concessions (Lead and Zinc) and Zinc
458 Appendices

Number BISD/DS Full title Short title

Report/Other
Ref No.
36 BISD 25S/107 Japan Measures on Imports of Thrown Japan—Silk Yarn
Silk Yarn
37 BISD 26S/320 Japanese Measures on Imports of Leather Japan—Leather
I (US)
38 BISD 26S/290, European Communities—Refunds on EC—Sugar
28S/80, 29S/82, Exports of Sugar Exports
27S/69 (Australia)
39 BISD 27S/69 & European Communities—Refunds on EC—Sugar
26S/290 Exports of Sugar—Complaint by Brazil Exports (Brazil)
40 BISD 27S/98 EEC Restrictions on Imports of Apples EEC—Apples I
from Chile (Chile)
41 BISD 27S/118 Japan’s Measures on Imports of Leather Japan—Leather
(Canada)
42 BISD 27S/119 Norway—Restrictions on Imports of Norway—Textiles
Certain Textile Products
43 BISD 28S/100 Japanese Restraints on Imports of Japan—Tobacco
Manufactured Tobacco from the United
States
44 BISD 28S/92 European Economic Community— EEC—Imports
Imports of Beef from Canada of Beef
45 BISD 28S/102 Spain—Tariff Treatment of Unroasted Spain—
Coffee Unroasted Coffee
46 BISD 28S/113 Panel on United States Countervailing US—CVD
Duties (India)
47 BISD 28S/90 EEC—United Kingdom Application of EEC—Poultry
EEC Directives to Imports of Poultry from (US)
the United States
48 L/5142 Spain—Measures Concerning the Spain—Soyabean
Domestic Sale of Soyabean Oil—Recourse Oil
to Article XXIII:2 by the United States
49 BISD 23S/127 & Income Tax Practices Maintained by Belgium—
28S/114 Belgium Income Tax
50 BISD 23S/98 & United States Tax Legislation (DISC) US—DISC
28S/114
51 BISD 23S/114 & Income Tax Practices Maintained by France—Income
28S/114 France Tax
52 BISD 23S/137 & Income Tax Practices Maintained by the Netherlands—
28S/114 Netherlands Income Tax
53 BISD 29S/91 United States—Prohibition of Imports of US—Canadian
Tuna and Tuna Products from Canada Tuna
54 BISD 30S/107 United States—Imports of Certain US—Spring
Automotive Spring Assemblies Assemblies
 Appendices 459

Number BISD/DS Full title Short title

Report/Other
Ref No.
55 BISD 29S/110 Panel on Vitamins US—Vitamin
B12
56 SCM/42 European Economic Community— EEC—Wheat
Subsidies on Export of Wheat Flour Flour Subsidies
57 L/5778 European Economic Community— EEC—Canned
Production Aids Granted on Canned Fruit
Peaches, Canned Pears, Canned Fruit
Cocktail and Dried Grapes
58 L/5776 European Community—Tariff Treatment EC—Citrus
on Imports of Citrus Products from
Certain Countries in the Mediterranean
Region
59 BISD 30S/140 Canada—Administration of the Foreign Canada—FIRA
Investment Review Act
60 BISD 30S/129 EEC—Quantitative Restrictions Against EEC—Import
Imports of Certain Products from Hong Restrictions
Kong
61 SCM/43 European Economic Community— EEC—Pasta
Subsidies on Export of Pasta Products Subsidies
62 BISD 31S/74 United States Manufacturing Clause US—
Manufacturing
Clause
63 BISD 31S/94 Panel on Japanese Measures on Imports Japan—Leather
of Leather II (US)
64 BISD 31S/67 United States—Imports of Sugar from US—Sugar
Nicaragua Quota
65 BISD 31S/247 Panel on Value-Added Tax and Threshold EEC - VAT and
Threshold
66 L/5863 Canada—Measures Affecting the Sale of Canada—Gold
Gold Coins Coins
67 BISD 31S/114 Panel on Newsprint EEC—Newsprint
68 BISD 32S/55 New Zealand—Imports of Electrical New Zealand—
Transformers from Finland Finnish
Transformers
69 L/6053 United States—Trade Measures Affecting US—Nicaraguan
Nicaragua Trade

70 BISD 35S/37 Canada—Import, Distribution and Canada—

Sale of Alcoholic Drinks by Canadian Provincial Liquor
Provincial Marketing Agencies Boards (EEC)
460 Appendices

Number BISD/DS Full title Short title

Report/Other
Ref No.
71 BISD 39S/436 Panel on United States Definition of US—Wine and
Industry Concerning Wine and Grape Grape Products
Products
72 BISD 35S/163 Japan—Restrictions on Imports of Japan—
Certain Agricultural Products Agricultural
Products I
73 SCM/85 Canada—Imposition of Countervailing Canada—
Duties on Imports of Manufacturing Beef Manufacturing
from the EEC Beef CVD
74 BISD 34S/194 Panel on United States Initiation of a US—Softwood
Countervailing Duty Investigation into Lumber I
Softwood Lumber Products from Canada
75 BISD 34S/83 Japan—Customs Duties, Taxes and Japan—Alcoholic
Labelling Practices on Imported Wines Beverages I
and Alcoholic Beverages
76 BISD 35S/98 Canada—Measures Affecting Exports of Canada—Herring
Unprocessed Herring and Salmon and Salmon
77 BISD 35S/116 Japan—Trade in Semi-Conductors Japan—Semi-
conductors
78 BISD 34S/136 United States—Taxes on Petroleum and US—Superfund
Certain Imported Substances
79 BISD 36S/345 United States Section 337 of the Tariff Act US—Section 337
of 1930
80 BISD 35S/245 United States—Customs User Fee US—Customs
User Fee
81 BISD 36S/93 European Economic Community— EEC—Dessert
Restrictions on Imports of Dessert Apples
Apples—Complaint by Chile
82 BISD 36S/135 European Economic Community— EEC—Apples
Restrictions on Imports of Apples— (US)
Complaint by the United States
83 BISD 36S/306 Norway—Restrictions on Imports of Norway—Apples
Apples and Pears and Pears
84 BISD 36S/202 Republic of Korea—Restrictions Korea—Beef
on Imports of Beef—Complaint by (Australia)
Australia
85 BISD 36S/234 Republic of Korea—Restrictions on Korea—Beef
Imports of Beef—Complaint by New (NZ)
Zealand
86 BISD 36S/268 Republic of Korea—Restrictions on Korea—Beef (US)
Imports of Beef—Complaint by the
United States
 Appendices 461

Number BISD/DS Full title Short title

Report/Other
Ref No.
87 BISD 36S/68 Canada—Import Restrictions on Ice Canada—Ice
Cream and Yoghurt Cream and
Yoghurt
88 BISD 36S/167 Canada/Japan—Tariff on Imports of Japan—SPF
Spruce, Pine, Fir (SPF) Dimension Dimension
Lumber Lumber
89 BISD 42S/208 United States—Countervailing Duties on US—Non-
Non-Rubber Footwear from Brazil Rubber Footwear
90 BISD 37S/86 European Economic Community— EEC—Oilseeds I
Payments and Subsidies Paid to Processors
and Producers of Oilseeds and Related
Animal-Feed Proteins
91 BISD 37S/132 European Economic Community— EEC—Parts and
Regulation on Imports of Parts and Components
Components
92 BISD 36S/331 United States—Restrictions on Imports US—Sugar
of Sugar
93 BISD 37S/200 European Economic Community— EEC—Copper
Restrictions on Exports of Copper Scrap Scrap
94 BISD 38S/30 United States—Countervailing Duties US—Canadian
on Fresh, Chilled and Frozen Pork from Pork
Canada
95 ADP/47 United States—Imposition of Anti- US—Swedish
Dumping Duties on Imports of Seamless Steel
Stainless Steel Hollow Products from
Sweden
96 BISD 37S/228 & United States—Restrictions on US—Sugar
36S/331 the Importation of Sugar and Waiver
Sugar-Containing Products Applied under
the 1955 Waiver and under the Headnote
to the Schedule of Tariff Concessions
97 BISD 37S/200 Thailand—Restrictions on Importation of Thailand—
and Internal Taxes on Cigarettes Cigarettes
98 BISD 37S/80 Canada/European Communities—Article EC—Article
XXVIII Rights XXVIII
99 SCM/142 German Exchange Rate Scheme for EEC—Airbus
Deutsche Airbus
100 BISD 39S/206 United States—Measures Affecting US—Malt
Alcoholic and Malt Beverages Beverages
101 BISD 39S/27 Canada—Import, Distribution and Sale Canada—
of Certain Alcoholic Drinks by Provincial Provincial Liquor
Marketing Agencies Boards (US)
462 Appendices

Number BISD/DS Full title Short title

Report/Other
Ref No.
102 ADP/82 United States—Anti-Dumping Duties US—Cement
on Gray Portland Cement and Cement
Clinker from Mexico
103 BISD 39S/411 Panel on Canadian Countervailing Duties Canada –Grain
on Grain Corn from the United States Corn
104 BISD 40S/358 Panel on United States—Measures US—Softwood
Affecting Imports of Softwood Lumber Lumber II
from Canada
105 BISD 39S/128 United States—Denial of Most-Favoured- US—MFN
Nation Treatment as to Non-Rubber Footwear
Footwear from Brazil
106 BISD39S/91 European Economic Community— EEC—Oilseeds II
Follow-Up on the Panel Report ‘Payments
and Subsidies Paid to Processors and
Producers of Oilseeds and Related
Animal-Feed Proteins’
107 BISD 41S/I/229 Imposition of Anti-Dumping Duties on US—Norwegian
Imports of Fresh and Chilled Atlantic Salmon AD
Salmon from Norway
108 BISD 41S/II/576 Imposition of Countervailing Duties on US—Norwegian
Imports of Fresh and Chilled Atlantic Salmon CVD
Salmon from Norway
109 GPR/DS1/R United States—Procurement of a Sonar US—Sonar
Mapping System Mapping
110 BISD 40S/319 Panel Report on Norwegian Procurement Norway—
of Toll Collection Equipment for the City Trondheim Toll
of Trondheim Ring
111 BISD 39S/155 United States—Restrictions on Imports US—Tuna
of Tuna (Mexico)
112 BISD 40S/205 Panel Report on Korea—Anti-Dumping Korea—Resins
Duties on Imports of Polyacetal Resins
from the United States
113 ADP/136 EC—Anti-Dumping Duties on Audio EC—Audio
Tapes in Cassettes Originating in Japan Cassettes
114 SCM/178 Australia—Imposition of Countervailing Australia—Glacé
Duties on Imports of Glacé Cherries from Cherries
France and Italy in Application of the
Australian Customs Amendment Act 1991
115 SCM/174 United States—Measures Affecting the US—Magnesium
Export of Pure and Alloy Magnesium
from Canada
 Appendices 463

Number BISD/DS Full title Short title

Report/Other
Ref No.
116 ADP/117 + United States—Anti-Dumping Duties US—Swedish
Corr.1 on Imports of Stainless Steel Plate from Steel Plate
Sweden
117 DS29/R United States—Restrictions on Imports US—Tuna (EEC)
of Tuna
118 DS39/R EEC—Restrictions on Imports of Apples EEC—Apples II
(Chile)
119 DS32/R EEC—Member States’ Import Regimes EEC (Member
for Bananas States)—Bananas
I
120 BISD 41S/II/467 Imposition of Provisional and Definitive Brazil—EEC
Countervailing Duties on Milk Powder Milk
and Certain Types of Milk from the
European Economic Community
121 SCM/185 United States—Imposition of US—Lead and
Countervailing Duties on Certain Hot- Bismuth I
Rolled Lead and Bismuth Carbon Steel
Products Originating in France, Germany
and the United Kingdom
122 DS38/R EEC—Import Regime for Bananas EEC—Bananas II
123 BISD 42S/17 European Economic Community— EEC—Cotton
Imposition of Anti-Dumping Duties on Yarn
Imports of Cotton Yarn from Brazil
124 BISD 41S/I/131 United States Measures Affecting the US—Tobacco
Importation, Internal Sale and Use of
Tobacco
125 DS31/R United States—Taxes on Automobiles US—Taxes on
Automobiles
464 Appendices

Table of WTO Panels

Number WT/DS No. Full title Short title

1 WT/DS2— United States— US—Gasoline

Venezuela WT/ Standards for
DS4—Brazil Reformulated and
Conventional Gasoline
2 WT/DS7—Canada European EC—Scallops (Canada)
Communities—Trade
Description of Scallops
3 WT/DS8—EC Japan—Taxes on Japan—Alcoholic Beverages II
WT/DS10— Alcoholic Beverages
Canada WT/
DS11—US
4 WT/DS9—Canada European EC—Cereals
Communities—Duties
on Imports of Cereals
5 WT/DS12—Peru European EC—Scallops (Peru and Chile)
WT/DS14—Chile Communities—Trade
Description of Scallops
6 WT/DS18— Australia—Measures Australia—Salmon
Canada Affecting Importation
of Salmon
7 WT/DS21—US Australia—Measures Australia—Salmonids
Affecting the
Importation of
Salmonids
8 WT/DS22— Brazil—Measures Brazil—Desiccated Coconut
Philippines Affecting Desiccated
Coconut
9 WT/DS24—Costa United States— US—Underwear
Rica Restrictions on
Imports of Cotton
and Man-Made Fibre
Underwear
10 WT/DS26—US European EC—Hormones (US)
11 WT/DS48— Communities— EC—Hormones (Canada)
Canada Measures Concerning
Meat and Meat
Products (Hormones)
12 WT/DS27— European EC—Bananas III (Ecuador)
Ecuador, Communities— EC—Bananas III (Guatemala
Guatemala, Regime for the and Honduras) EC—Bananas
Honduras, Mexico, Importation, Sale III (Mexico) EC—Bananas
US and Distribution of III (US)
Bananas
 Appendices 465

Number WT/DS No. Full title Short title

13 WT/DS31—US Canada—Certain Canada—Periodicals

Measures Concerning
Periodicals
14 WT/DS32—India United States— US—Wool Coats
Measures Affecting
Imports of Women’s
and Girls’ Wool Coats
15 WT/DS33—India United States— US—Wool Shirts and Blouses
Measures Affecting
Imports of Woven
Wool Shirts and
Blouses from India
16 WT/DS34—India Turkey—Restrictions Turkey—Textiles
on Imports of Textile
and Clothing Products
17 WT/DS35— Hungary—Export Hungary—Agricultural
Argentina Subsides in Respect of Products
Agricultural Products
18 WT/DS38—EC United States—The US—Helms Burton
Cuban Liberty and
Democratic Solidarity
Act
19 WT/DS44—US Japan—Measures Japan—Film
Affecting Consumer
Photographic Film and
Paper
20 WT/DS46— Brazil—Export Brazil—Aircraft
Canada Financing Programme
for Aircraft
21 WT/DS50—US India—Patent India—Patents (US)
Protection for
Pharmaceutical and
Agricultural Chemical
Products
22 WT/DS54—EC Indonesia—Certain Indonesia—Autos
WT/DS55—Japan Measures Affecting the
WT/DS59—US Automobile Industry
WT/DS64—Japan
23 WT/DS56—US Argentina –Measures Argentina—Textiles and
Affecting Imports of Apparel
Footwear, Textiles,
Apparel and Other
Items
24 WT/DS58—India, United States—Import US—Shrimp
Malaysia, Pakistan, Prohibition of Certain
Thailand Shrimp and Shrimp
Products
466 Appendices

Number WT/DS No. Full title Short title

25 WT/DS60— Guatemala—Anti- Guatemala—Cement I

Mexico Dumping Investigation
Regarding Portland
Cement from Mexico
26 WT/DS62—US European EC—Computer Equipment
WT/DS67—US Communities—
WT/DS68—US Customs Classification
of Certain Computer
Equipment
27 WT/DS69—Brazil European EC—Poultry
Communities—
Measures Affecting the
Importation of Certain
Poultry Products
28 WT/DS70—Brazil Canada—Measures Canada—Aircraft
Affecting the Export of
Civilian Aircraft
29 WT/DS72—NZ European EC—Butter
Communities—
Measures Affecting
Butter Products
30 WT/DS75—EC Korea—Taxes on Korea—Alcoholic Beverages
WT/DS84—US Alcoholic Beverages
31 WT/DS76—US Japan—Measures Japan—Agricultural Products
Affecting Agricultural II
Products
32 WT/DS77—EC Argentina—Measures Argentina—Textiles and
Affecting Textiles and Clothing
Clothing
33 WT/DS79—EC India—Patent India—Patents (EC)
Protection for
Pharmaceutical and
Agricultural Chemical
Products
34 WT/DS87—EC Chile—Taxes on Chile—Alcoholic Beverages
WT/DS110—EC Alcoholic Beverages
35 WT/DS88—EC United States— US—Procurement
WT/DS95—Japan Measure Affecting
Government
Procurement
36 WT/DS90—US India—Quantitative India—Quantitative
Restrictions on Imports Restrictions
of Agricultural, Textile
and Industrial
Products
 Appendices 467

Number WT/DS No. Full title Short title

37 WT/DS98—EC Korea—Definitive Korea—Dairy

Safeguard Measure
on Imports of Certain
Dairy Products
38 WT/DS99—Korea United States—Anti- US—DRAMS
Dumping duty on
Dynamic Random
Access Memory
Semiconductors
(DRAMS) of one
Megabit or Above from
Korea
39 Canada—Measures Canada—Dairy
WT/DS103—US Affecting the
WT/DS113—NZ Importation of Milk
and the Exportation of
Dairy Products
40 WT/DS106—US Australia—Subsidies Australia—Automotive
Provided to Producers Leather I
and Exporters of
Automotive Leather
41 WT/DS108—EC United States - Tax US—FSC
Treatment for ‘Foreign
Sales Corporations’
42 WT/DS114—EC Canada—Patent Canada—Pharmaceutical
Protection of Patents
Pharmaceutical
Products
43 WT/DS121—EC Argentina—Safeguard Argentina—Footwear (EC)
Measures on Imports of
Footwear
44 WT/DS122— Thailand—Anti- Thailand—H-Beams
Poland Dumping Duties on
Angles, Shapes and
Sections of Iron or
Non-Alloy Steel and
H-Beams from Poland
45 WT/DS126—US Australia—Subsidies Australia—Automotive
Provided to Producers Leather II
and Exporters of
Automotive Leather
46 WT/DS132—US Mexico—Anti- Mexico—Corn Syrup
Dumping Investigation
of High Fructose Corn
Syrup (HFCS) from
the United States
468 Appendices

Number WT/DS No. Full title Short title

47 WT/DS135— European EC—Asbestos
Canada Communities—
Measures Affecting
Asbestos and
Asbestos-Containing
Products
48 WT/DS136—EC United States–Anti- US—1916 Act (EC)
49 WT/DS162—Japan Dumping Act of 1916 US—1916 Act (Japan)
50 WT/DS138—EC United States— US—Lead and Bismuth II
Imposition of
Countervailing
Duties on Certain Hot-
Rolled Lead
and Bismuth Carbon
Steel Products
Originating in the
United Kingdom
51 WT/DS139—Japan Canada—Certain Canada—Autos
WT/DS142—EC Measures Affecting
the Automotive
Industry
52 WT/DS141—India European EC—Bed Linen
Communities—Anti-
Dumping Duties on
Imports of Cotton-
Type Bed Linen from
India
53 WT/DS146—EC India—Measures India—Autos
WT/DS175—US Affecting the
Automotive Sector
54 WT/DS152—EC United States - US - Section 301 Trade Act
Sections 301–310
of the Trade Act of
1974
55 WT/DS155—EC Argentina—Measures Argentina –Hides and Leather
Affecting the Export
of Bovine Hides and
the Import of Finished
Leather
56 WT/DS156— Guatemala— Guatemala—Cement II
Mexico Definitive Anti-
Dumping Measures on
Grey Portland Cement
from Mexico
57 WT/DS160 - EC United States—Section US—Section 110(5) Copyright
110(5) of the US Act
Copyright Act
 Appendices 469

Number WT/DS No. Full title Short title

58 WT/DS161—US Korea—Measures Korea—Various Measures on
WT/DS169— Affecting Imports of Beef
Australia Fresh, Chilled and
Frozen Beef
59 WT/DS163—US Korea—Measures Korea—Procurement
Affecting Government
Procurement
60 WT/DS164—US Argentina—Measures Argentina—Footwear (US)
Affecting Imports of
Footwear
61 WT/DS165—EC United States— US—Certain EC Products
Import Measures on
Certain Products
from the European
Communities
62 WT/DS166—EC United States— US—Wheat Gluten
Definitive Safeguard
Measures on Imports
of Wheat Gluten
from the European
Communities
63 WT/DS170—US Canada—Term of Canada—Patent Term
Patent Protection
64 WT/DS174—US European EC—Trademarks and
WT/DS290— Communities— Geographical Indications
Australia Protection of
Trademarks and
Geographical
Indications for
Agricultural Products
and Foodstuff s
65 WT/DS176—EC United States—Section US—Section 211
211 Omnibus Appropriations Act
Appropriations Act of
1998
66 WT/DS177—NZ United States— US—Lamb
WT/DS178— Safeguard Measures
Australia on Imports of Fresh,
Chilled or Frozen
Lamb Meat from New
Zealand and Australia
67 WT/DS179—Korea United States—Anti- US—Stainless Steel
Dumping Measures
on Stainless Steel Plate
in Coils and Stainless
Steel Sheet and Strip
from Korea
470 Appendices

Number WT/DS No. Full title Short title

68 WT/DS184—Japan United States— US—Hot-Rolled Steel
Anti-Dumping
Measures on
Certain Hot-Rolled
Steel Products from
Japan
69 WT/DS188— Nicaragua—Measures Nicaragua—Imports from
Colombia Affecting Imports Honduras and Colombia
from Honduras and
Colombia
70 WT/DS189—EC Argentina—Definitive Argentina—Ceramic Tiles
Anti-Dumping
Measures on Imports
of Ceramic Floor Tiles
from Italy
71 WT/DS190—Brazil Argentina— Argentina—Cotton
Transitional Safeguard
Measures on Certain
Imports of Woven
Fabrics of Cotton
and Cotton Mixtures
Originating in Brazil
72 WT/DS192— United States— US—Cotton Yarn
Pakistan Transitional Safeguard
Measure on Combed
Cotton Yarn from
Pakistan
73 WT/DS193—EC Chile—Measures Chile—Swordfish
Affecting the Transit
and Importation of
Swordfish
74 WT/DS194— United States— US—Export Restraints
Canada Measures Treating
Export Restraints as
Subsidies
75 WT/DS195—US Philippines—Measures Philippines—Motor Vehicles
Affecting Trade and
Investment in the
Motor Vehicle Sector
76 WT/DS199—US Brazil—Measures Brazil—Patent Protection
Affecting Patent
Protection
77 WT/DS202— United States— US—Line Pipe
Korea Definitive Safeguard
Measures on Imports
of Circular Welded
Carbon Quality Line
Pipe from Korea
 Appendices 471

Number WT/DS No. Full title Short title

78 WT/DS204—US Mexico—Measures Mexico—Telecoms
Affecting
Telecommunications
Services
79 WT/DS206—India United States—Anti- US—Steel Plate
Dumping and
Countervailing
Measures on Steel Plate
from India
80 WT/DS207— Chile—Price Band Chile—Price Band System
Argentina System and Safeguard
Measures Relating to
Certain Agricultural
Products
81 WT/DS210—US Belgium— Belgium—Rice
Administration of
Measures Establishing
Customs Duties for
Rice
82 WT/DS211— Egypt—Definitive Egypt—Steel Rebar
Turkey Anti-Dumping
Measures on Steel
Rebar from Turkey
83 WT/DS212—EC United States— US—Countervailing Measures
Countervailing on Certain EC Products
Measures Concerning
Certain Products
from the European
Communities
84 WT/DS213—EC United States— US—Carbon Steel
Countervailing
Duties on Certain
Corrosion-Resistant
Carbon Steel Flat
Products from Germany
85 WT/DS214—EC United States— US—Wire Rod and Line Pipe
Definitive Safeguard
Measures on Imports
of Steel Wire Rod
and Circular Welded
Quality Line Pipe
86 WT/DS217— United States— US—Offset Act (Byrd
Australia, Brazil, Continued Dumping Amendment)
Chile, EC, India, and Subsidy Offset Act
Indonesia, Japan, of 2000
Korea, Thailand
WT/DS234—
Canada, Mexico
472 Appendices

Number WT/DS No. Full title Short title

87 WT/DS219—Brazil European EC—Tube or Pipe Fittings
Communities—Anti-
Dumping Duties on
Malleable Cast Iron
Tube or Pipe Fittings
from Brazil
88 WT/DS221— United States—Section US—Section 129(c)(1) URAA
Canada 129(c)(1) of the
Uruguay Round
Agreements Act
89 WT/DS222— Canada—Export Canada—Aircraft Credits and
Brazil Credits and Loan Guarantees
Guarantees for
Regional Aircraft
90 WT/DS227—Chile Peru—Taxes on Peru—Taxes on Cigarettes
Cigarettes
91 WT/DS231—Peru European EC—Sardines
Communities—Trade
Description of Sardines
92 WT/DS236— United States— US—Softwood Lumber III
Canada Preliminary
Determinations with
Respect to Certain
Softwood Lumber
from Canada
93 WT/DS237— Turkey—Certain Turkey—Fresh Fruit Import
Ecuador Import Procedures of Procedures
Fresh Fruit
94 WT/DS238—Chile Argentina—Definitive Argentina—Preserved Peaches
Safeguard Measure on
Imports of Preserved
Peaches
95 WT/DS241—Brazil Argentina—Definitive Argentina—Poultry Anti-
Anti-Dumping Duties Dumping Duties
on Poultry from Brazil
96 WT/DS243—India United States—Rules US—Textiles Rules of Origin
of Origin for Textiles
and Apparel Products
97 WT/DS244—Japan United States—Sunset US—Corrosion-Resistant Steel
Review of Anti- Sunset Review
Dumping Duties on
Corrosion-Resistant
Carbon Steel Flat
Products from Japan
 Appendices 473

Number WT/DS No. Full title Short title

98 WT/DS245—US Japan—Measures Japan—Apples
Affecting the
Importation of Apples
99 WT/DS246—India European EC—Tariff Preferences
Communities—
Conditions for the
Granting of Tariff
Preferences to
Developing Countries
100 WT/DS248—EC United States— US—Steel Safeguards
WT/DS249—Japan Definitive Safeguard
WT/DS251—Korea Measures on Imports of
WT/DS252—China Certain Steel Products
WT/DS253—
Switzerland
WT/DS254—
Norway
WT/DS258—New
Zealand
WT/DS259—Brazil
101 WT/DS250— United States— US—Florida Excise Tax
Brazil Equalizing Excise Tax
Imposed by Florida on
Processed Orange and
Grapefruit Products
102 WT/DS257— United States—Final US—Softwood Lumber IV
Canada Countervailing Duty
Determination with
respect to certain
Softwood Lumber
from Canada
103 WT/DS260—US European EC—Provisional Steel
Communities— Safeguards
Provisional Safeguard
Measures on Imports
of Certain Steel
Products
104 WT/DS261—Chile Uruguay—Tax Uruguay—Tax Treatment
Treatment on Certain
Products
105 WT/DS264— United States— US—Softwood Lumber V
Canada Final Dumping
Determination on
Softwood Lumber
from Canada
474 Appendices

Number WT/DS No. Full title Short title

106 WT/DS265— European EC—Export Subsidies on
Australia Communities—Export Sugar
WT/DS266—Brazil Subsidies on Sugar
WT/DS283—
Thailand
107 WT/DS267—Brazil United States— US—Upland Cotton
Subsidies on Upland
Cotton
108 WT/DS268— United States—Sunset US—Oil Country Tubular
Argentina Reviews of Anti- Goods Sunset Reviews
Dumping Measures on
Oil Country Tubular
Goods from Argentina
109 WT/DS269— European EC—Chicken Cuts
Brazil Communities—
WT/DS286— Customs Classification
Thailand of Frozen Boneless
Chicken Cuts
110 WT/DS270— Australia—Certain Australia—Fresh Fruit and
Philippines Measures Affecting the Vegetables
Importation of Fresh
Fruit and Vegetables
111 WT/DS273—EC Korea—Measures Korea—Commercial Vessels
Affecting Trade in
Commercial Vessels
112 WT/DS276—US Canada—Measures Canada—Wheat Exports and
Relating to Exports of Grain Imports
Wheat and Treatment
of Imported Grain
113 WT/DS277— US—Investigation US—Softwood Lumber VI
Canada of the International
Trade Commission
in Softwood Lumber
from Canada
114 WT/DS280— United States— US—Countervailing Duties
Mexico Countervailing Duties on Steel Plate
on Steel Plate from
Mexico
115 WT/DS281— United States—Anti- US—Anti-Dumping Measures
Mexico Dumping Measures on on Cement
Cement from Mexico
116 WT/DS282— United States—Anti- US—Anti-Dumping Measures
Mexico Dumping Measures on on Oil Country Tubular Goods
Oil Country Tubular
Goods (OCTG) from
Mexico
 Appendices 475

Number WT/DS No. Full title Short title

117 WT/DS285— United States— US—Gambling
Antigua and Measures Affecting the
Barbuda Cross-Border Supply of
Gambling and Betting
Services
118 WT/DS287—EC Australia—Quarantine Australia—Quarantine
Regime for Imports Regime
119 WT/DS291—US European EC—Approval and Marketing
WT/DS292— Communities— of Biotech Products
Canada Measures Affecting
WT/DS293— the Approval and
Argentina Marketing of Biotech
Products
120 WT/DS294—EC United States—Laws, US—Zeroing (EC)
Regulations and
Methodology for
Calculating Dumping
Margins (‘Zeroing’)
121 WT/DS295—US Mexico—Definitive Mexico—Anti-Dumping
Anti-Dumping Measures on Rice
Measures on Beef and
Rice
122 WT/DS296— United States— US—Countervailing Duty
Korea Countervailing Duty Investigation on DRAMs
Investigation on
Dynamic Random
Access Memory
Semiconductors
(DRAMs) from Korea
123 WT//DS299— European EC—Countervailing Measures
Korea Communities— on DRAM Chips
Countervailing
Measures on Dynamic
Random Access
Memory Chips from
Korea
124 WT/DS301—Korea European EC—Commercial Vessels
Communities—
Measures Affecting
Trade in Commercial
Vessels
125 WT/DS302— Dominican Republic— Dominican Republic—Import
Honduras Measures Affecting and Sale of Cigarettes
the Importation
and Internal Sale of
Cigarettes
476 Appendices

Number WT/DS No. Full title Short title

126 WT/DS308—US Mexico—Tax Measures Mexico—Taxes on Soft Drinks
on Soft Drinks and
Other Beverages
127 WT/DS312— Korea—Anti- Korea—Certain Paper
Indonesia Dumping Duties on
Imports of Certain
Paper from Indonesia
128 WT/DS315—US European EC—Selected Customs
Communities— Matters
Selected Customs
Matters
129 WT/DS316—US European EC and certain member
Communities and States—Large Civil Aircraft
Certain Member
States—Measures
Affecting Trade in
Large Civil Aircraft
130 WT/DS317—EC United States— US—Large Civil Aircraft
Measures Affecting
Trade in Large Civil
Aircraft
131 WT/DS320—EC United States— US—Continued Suspension
Continued Suspension
of Obligations in
the EC—Hormones
Dispute
132 WT/DS321—EC Canada—Continued Canada—Continued
Suspension of Suspension
Obligations in the
EC—Hormones
Dispute
133 WT/DS322—Japan United States— US—Zeroing (Japan)
Measures Relating to
Zeroing and Sunset
Reviews
134 WT/DS323— Japan—Import Quotas Japan—Quotas on Laver
Korea on Dried Laver and
Seasoned Laver
135 WT/DS327— Egypt—Anti- Egypt—Matches
Pakistan Dumping Duties on
Matches from Pakistan
136 WT/DS331— Mexico—Anti- Mexico—Steel Pipes and
Guatemala Dumping Duties on Tubes
Steel Pipes and Tubes
from Guatemala
 Appendices 477

Number WT/DS No. Full title Short title

137 WT/DS332—EC Brazil—Measures Brazil—Retreaded Tyres
Affecting Imports of
Retreaded Tyres
138 WT/DS334—US Turkey—Measures Turkey—Rice
Affecting the
Importation of Rice
139 WT/DS335— United States—Anti- US—Shrimp (Ecuador)
Ecuador Dumping Measures on
Shrimp from Ecuador
140 WT/DS336— Japan— Japan—DRAMs (Korea)
Korea Countervailing
Measures on Dynamic
Random Access
Memories from Korea
141 WT/DS337— European EC—Salmon (Norway)
Norway Communities—Anti-
Dumping Measure on
Farmed Salmon from
Norway
142 WT/DS339—EC China—Measures China—Auto Parts
WT/DS340—US Affecting Imports of
WT/DS342— Automobile Parts
Canada
143 WT/DS343— United States— US—Shrimp (Thailand)
Thailand Measures Relating to
Shrimp from Thailand
144 WT/DS344— United States—Final US—Stainless Steel (Mexico)
Mexico Anti-Dumping
Measures on Stainless
Steel from Mexico
145 WT/DS345—India United States— US—Customs Bond Directive
Customs Bond
Directive for
Merchandise subject
to Anti-Dumping/
Countervailing Duties
146 WT/DS347—US European EC and certain member
Communities and States—Large Civil Aircraft
Certain Member (2nd complaint)
States—Measures
Affecting Trade in
Large Civil Aircraft
(Second Complaint)
 References
Aaronson, Susan Ariel. 1999. Who decides? Congress and the debate over trade policy in 1934
and 1974, Council of Foreign Relations: New York, NY.
Abbott, Fred. 2000. ‘The NAFTA and the legalization of world politics’, International
Organzation, 54: 519–50.
Abbott, Kenneth W. 1996. ‘Defensive Unfairness: The Normative Structure of Section
301’ in Jagdish Bhagwati and Robert Hudec (eds), Fair Trade And Harmonization,
MIT Press: Cambridge, MA, pp 415–71.
Abdel Motaal, Dooa. 2004. ‘The “multilateral scientific consensus” and the World Trade
Organization’, Journal of World Trade, 38: 855–76.
Alvarez, Jose. 1994. ‘Positivism regained, nihilism postponed’, Michigan Journal of
International Law, 15: 747–81.
Anderson, James E. and Eric van Wincoop. 2003. ‘Gravity with gravitas: a solution to the
border puzzle’, American Economic Review, 93: 170–92.
Arrowsmith, Sue. 2005. The law of public and utlities procurement, Sweet and Maxwell:
London.
Axelrod, Robert. 1984. The evolution of cooperation, New York Basic Books: New York, NY.
Bagwell, Kyle. 2006. Remedies in the WTO: an economics perspective. Mimeo.
Bagwell, Kyle, Petros C. Mavroidis and Robert W. Staiger. 2002. ‘It is all about market
access’, American Journal of International Law, 96: 56–76.
Bagwell, Kyle, Petros C. Mavroidis and Robert W. Staiger. 2005. ‘The case for tradable
remedies in the WTO’ in Simon Evenett and Bernard Hoekman (eds), Economic devel-
opment and multilateral trade cooperation, Palgrave Macmillan and the World Bank:
Washington, DC, pp 395–414.
Bagwell, Kyle and Robert W. Staiger. 1990. ‘A theory of managed trade’, American
Economic Review, 80: 779–95.
Bagwell, Kyle and Robert W. Staiger. 1995. ‘Protection and the business cycle’, NBER
Discussion Paper 5168.
Bagwell, Kyle and Robert W. Staiger. 2002. The Economics of the World Trading System,
MIT Press: Cambridge, MA.
Bagwell, Kyle and Robert W. Staiger. 2005. ‘Enforcement, private political pressures and
the GATT/WTO escape clause’, Journal of Legal Studies, 34: 471–513.
Bagwell, Kyle and Robert W. Staiger. 2006. ‘Will international rules on subsidies disrupt
the world trading system?’, American Economic Review, 96: 877–95.
Baldwin, Richard E. 1997. ‘The causes of regionalism’, World Economy, 20: 865–88.
Baldwin, Richard E. 2006. ‘Multilateralising regionalism, spaghetti bowls as building
blocks on the path to global free trade’, World Economy, 29: 1451–518.
Baldwin, Robert. 1970. Non-tariff distortions in international trade, Brookings Institution:
Washington, DC.
Barfield, Claude. 2001. Free Trade, Sovereignty, Democracy: The Future of the World Trade
Organization, American Enterprise Institute: Washington, DC.
 References 479
Bartels, Lorand. 2002. ‘Article XX of the GATT and the problem of extraterritorial juris-
diction’, Journal of World Trade, 36: 353–403.
Bartels, Lorand and Federico Ortino. (eds) 2006. Regional trade agreements and the WTO
legal system, Oxford University Press: Oxford.
Basdevant, Suzanne. 1929. La clause de la nation la plus favorisée, Paris.
Battigalli, Pierpaolo and Giovanni Maggi. 2002. ‘Rigidity, Discretion and the Costs of
Writing Contracts’, American Economic Review, 92(4): 798–817.
Becker, Gary S. 1968. ‘Crime and punishment’, Journal of Political Economy, 76:
169–217.
Benedek, Wolfgang. 1991. Das GATT aus völkerrecthlicher Sicht, Springer Verlag:
Heidelberg.
Beviglia-Zampetti, Americo. 2006. Fairness in the World Economy: US Perspectives on
International Trade Relations, Elgar Publishing: Cheltenham.
Bhagwati, Jagdish. 1988. Protectionism. MIT Press: Cambridge, MA.
Bhagwati, Jagdish. 1998. ‘The damages of selective safeguards’ in Jagdish Bhagwati,
A stream of windows, MIT Press: Cambridge, MA, pp 109–12.
Bhagwati, Jagdish. 2002. ‘The unilateral freeing of trade versus reciprocity’ in Jagdish
Bhagwati (ed.), Going alone: the case for relaxed reciprocity in freeing trade, MIT Press:
Cambridge, MA, pp 1–30.
Bhagwati, Jagdish and Robert E. Hudec (eds) 1996. Fair Trade And Harmonization,
Cambridge University Press: Cambridge, MA.
Bhagwati, Jagdish and Petros C. Mavroidis. 2004. ‘Killing the Byrd Amendment with
the right stone’, World Trade Review, 3: 1–9.
Bhagwati, Jagdish and Arvind Panagariya. 1996. ‘Preferential trading areas and multi-
lateralism: strangers, friends or foes?’ in Jagdish Bhagwati and Arvind Panagariya
(eds), Free trade areas or free trade? The Economics of preferential trading agreements, AEI
Press: Washington, DC.
Bhagwati, Jagdish and Arvid Panagariya. 1999. ‘Preferential trading areas and multilat-
eralism—strangers, friends, or foes’ in Jagdish Bhagwati, Pravin Krishna and Arvind
Panagariya, Trading blocks, MIT Press: Cambridge, MA, pp 33–100.
Bhagwati, Jagdish and V.K. Ramaswami. 1963. ‘Domestic distortions, tariffs, and the
theory of optimum subsidy’, Journal of Political Economy, 71: 44–50.
Bickerdike, Charles F. 1907. ‘Review of AC Pigou’s protective and preferential import
duties’, Economic Journal, 17: 98–108.
Blank, Annet and Gabrielle Marceau. 1997. ‘A history of multilateral negotiations on
procurement: from ITO to WTO’, in Bernard M. Hoekman and Petros C. Mavroidis
(eds), Law and Policy in Public Purchasing, University of Michigan Press: Ann Arbor,
MI, pp 31–56.
Bloningen, Bruce and Jee Hyeong Park. 2004. ‘Dynamic pricing in the presence of anti-
dumping policy: theory and evidence’, American Economic Review, 94: 134–54.
Bogdandy von, Armin. 2003. ‘Legitimacy of international economic governance: inter-
pretative approaches to WTO law and the prospects of its proceduralization’, in
S. Griller (ed.), International economic governance and non-economic concerns, Springer
Verlag: Wien and New York, pp 103–48.
Boltuck, Richard and Robert E. Litan. 1991. Down in the dumps, Administration of the
unfair trade laws, Brookings: Washington, DC.
480 References
Bourgeois, Jacques. 2005. ‘Some reflections on the WTO dispute settlement system from
a practitioner’s perspective’, in J.H.J. Bourgeois, Trade Law Experienced: Pottering about
in the GATT and the WTO, Cameron May: London, pp 39–50.
Bown, Chad P. 2002. ‘Why are safeguards under the WTO so unpopular?’, World Trade
Review, 1: 47–62.
Bown, Chad P. 2005. Global Antidumping Database, World Bank Policy Research Paper
No 3737 (available at <http://www.brandeis/edu/-cbown/global_ad/>).
Bown, Chad P. and Bernard Hoekman. 2005. ‘WTO dispute settlement and the miss-
ing developing country cases: engaging the private sector’, Journal of International
Economic Law, 8: 861–90.
Bown, Chad P., B. Hoekman and C. Ozden. 2004. ‘Developing Countries and U.S.
Antidumping: The Path From Initial Filing to WTO Dispute Settlement’, World
Trade Review, 2: 349–71.
Bown, Chad P. and Rachell McCulloch. 2003. ‘Nondiscrimination and the WTO
Agreement on Safeguards’, World Trade Review, 2: 327–48.
Bown, Chad P. and Alan O. Sykes. 2007. US—Softwood Lumber V. Mimeo.
Bown, Chad P. and Jasper M. Wauters. 2007. US—AD Measures on OCTG. Mimeo.
Brainard S. Lael and Thierry Verdier. 1994. ‘Lobbying and adjustment in declining
industries’, European Economic Review, 38: 586–95.
Brander James and Barbara Spencer. 1985. ‘Export subsidies and international market
share rivalry’, Journal of International Economics, 18: 83–100.
Brenton, Paul. 2003. ‘Integrating the Least Developed Countries into the World Trading
System: The Current Impact of EU Preferences under Everything But Arms’, Journal
of World Trade 37(3): 623–46.
Brenton, Paul and Miriam Manchin. 2003. ‘Making EU trade agreements work: the role
of rules of origin’, The World Economy, 26: 755–69.
Broda,Christian,NunoLimãoandDavidE.Weinstein.2006.‘Optimaltariff sandevidence’,
NBERWorking Paper Series12033 (availableat<http://www.nber.org/papers/w12033>).
Bronckers, Marco. 2004. ‘The special safeguard clause in respect of China: (how) will it
work?’ in M. Matsushita and H. Ahn (eds), WTO and Asia: New perspectives, Cameron
May: London, pp 39–50.
Bronckers, Marco C.E.J. 1987. ‘A legal analysis of protectionist measures affecting
Japanese imports into the European Community—revisited’ in E.L.M. Völker (ed.),
Protectionism and the European Community, Kluwer: Deventer, pp 57–120.
Bronckers, Marco C.E.J. and Natalie McNelis. 2000. ‘Fact and law in pleadings before
the WTO Appellate Body’ in Friedl Weiss (ed.), Improving WTO dispute settlement
procedures: issues and lessons from the practice of other international courts and tribunals,
Cameron May: London, pp 321–33.
Bronckers, Marco and Naboth van den Broek. 2005. ‘Financial compensation in the
WTO: improving remedies in WTO dispute settlement’, Journal of International
Economic Law, 8: 101–26.
Brown, Drusilla K. 1987. ‘General equilibrium effects of the US GSP’, Southern Economic
Journal, 54: 27–47.
Brown, Drusilla K. 1989. ‘Trade and welfare effects of the European schemes of the GSP’,
Economic Development and Cultural Change, 40: 532–44.
Brown, Winthrop. 1950. The United States and the restoration of world trade: an an-
alysis and appraisal of the ITO Charter and the General Agreement on Tariff s and Trade,
Brookings Institution: Washington, DC.
 References 481
Busch, Marc L. and Eric Reinhardt. 2001. ‘Bargaining in the Shadow of the Law: Early
Settlement in GATT/WTO Disputes’, Fordham International Law Journal, 24 :
158–72.
Busch, Marc L. and Eric Reinhardt. 2002. ‘Testing International Trade Law:
Empirical Studies of GATT/WTO Dispute Settlement’ in Daniel M. Kennedy
and James D. Southwick (eds), The Political Economy of International Trade Law:
Essays in Honor of Robert Hudec, Cambridge University Press: Cambridge, MA,
pp 457–81.
Busch, Marc L. and Eric Reinhardt. 2003. ‘Developing Countries and GATT/WTO
Dispute Settlement’, Journal of World Trade, 37(4): 719–35.
Busch, Marc L. and Eric Reinhardt. 2005. Three’s a Crowd: The Influence of Third
Parties on GATT/WTO Dispute Settlement. Mimeo. Georgetown University.
Cadot, Olivier, Jaime de Melo and Bolormaa Tumurchudur. 2006. AD sunset reviews:
did the WTO agreement make any difference? Mimeo.
Capling, Ann. 2001. Australia and the global trade system: from Havana to Seattle,
Cambridge University Press: New York, NY.
Chang, Howard. 1995. ‘An economic analysis of trade measures to protect the global
environment’, Georgetown Law Journal, 83: 2131–60.
Charnovitz, Steve. 1995. ‘Promoting higher labour standards’, The Washington Quarterly,
18: 167–90.
Charnovitz, Steve. 1998. ‘The moral exception in trade policy’, Virginia Journal of
International Law, 38: 689–732.
Charnovitz, Steve. 2002. ‘The law of environmental PPMs in the WTO: debunking the
myth of illegality’, Yale Journal of International Law, 27: 59–109.
Charnovitz, Steve. 2006. Exploring the law of WTO accession. Mimeo.
Clerc, Evelyne. 2000. ‘La mondialisation et les marchés publics: bilan et perspective de
l'accord de l'OMC Bruylant, sur les marchés publics’, in D. Batselé et al. (ed.), Les
marchés publics à L’aube du XXI siècle: Brussels, pp 141–94.
Cottier, Thomas. 2004. ‘Proposals for Moving From ad hoc Panels to Permanent WTO
Panelists’ in F. Ortino and E.-U. Petersmann (eds), The WTO Dispute Settlement
System 1995–2003, Kluwer: The Hague, pp 31–40.
Crandall, Samuel B. 1913. ‘The American construction of the most-favored-nation
clause’, American Journal of International Law, 7: 709–10.
Crowley, Meredith. 2005. ‘Do antidumping duties and safeguard tariffs open or close
technology gaps?’, Journal of International Economics, 68: 469–84.
Crowley, Meredith. 2006. Why are safeguards needed in a trade agreement? Mimeo.
Dam, Kenneth. 1970. The GATT: Law and International Economic Organization,
University of Chicago Press: Chicago, ILL.
Dam, Kenneth. 1977. ‘The American Fiscal Constitution’, University of Chicago Law
Review, 44: 271–96.
Dam, Kenneth. 2004. Cordell Hull, The Reciprocal Trade Agreement Act, and the
WTO, Mimeo.
Davey, William, J. 1987. ‘Dispute settlement in GATT’, Fordham International Law
Journal, 11: 51–99.
Davey, William J. 2004. ‘Proposals for improving the working procedures of WTO dis-
pute settlement panels’ in Federico Ortino and Ernst-Ulrich Petersmann (eds), The
WTO dispute settlement system 1995–2003: Kluwer, The Hague, pp 19–30.
482 References
Davey William J. and Joost Pauwelyn. 2000. ‘MFN Unconditionality: a legal analysis of
the concept in view of its evolution in the GATT/WTO jurisprudence with particular
reference to the issue of “Like Products” in Thomas Cottier and Petros C. Mavroidis,
Regulatory barriers and the principle of non-discrimination in world trade law, University
of Michigan Press: Ann Arbor, MI, pp 13–50.
de Jong, David N. and Maria Ripoli. 2006. ‘Tariffs and growth: an empirical exploration
of contingent relationships’, The Review of Economic Statistics, 88: 625–40.
Deardorff, Alan. 1987. ‘Safeguards policy and the conservative social welfare function’,
in Henryk Kierzkowski, Protection and competition in international trade: essays in
honor of W.M. Corden, Blackwell: Oxford.
Deardorff, Alan V. and Robert M. Stern. 1998. Measurement of non tariff barriers,
University of Michigan Press: Ann Arbor, MI.
Démaret, Paul and R. Stewardson. 1994. ‘Border tax adjustments under GATT and EC
law and general implications for environmental taxes’, Journal of World Trade, 28:
5–65.
Desmedt, Axel G. 1998. ‘Hormones: objective assessment and (or as) standard of review’,
Journal of International Economic Law 1: 695–8.
Dewatripont, Mathias, André Sapir and Khalid Sekkat. 1999. ‘Labour market effects
of trade with LLDCs in Europe’ in Mathias Dewatripont, André Sapir and Khalid
Sekkat (eds), Trade and Jobs in Europe: much ado about nothing?, Oxford University
Press: Oxford, pp 60–78.
Diebold, William Jr. 1952. The end of the ITO, Essays in International Finance, No 16,
Princeton International Finance Section, Princeton University: Princeton, NJ.
Dixit, Avinash. 2004. Lawlessness and Economics, Princeton University Press: Princeton, NJ.
Dordi, Claudio. 2006. The Appellate Body case-law on AD and CVD sunset reviews.
Mimeo.
Downs, George, and David Rocke. 1997. Optimal imperfection? Domestic uncertainty and
institutions in international relations, Princeton University Press: Princeton, NJ.
Drabek, Zdenek. 2005. Limits to harmonization of domestic regulation. Mimeo.
Drache, Daniel. 2003. ‘ITO: when labour and investment standards almost mat-
tered: a putative history lesson in trade politics that ought not be forgotten’ in James
Busumtwi-Sam et al. (eds), Global instability: uncertainty and new visions in political
economy, Kluwer: Amsterdam, pp 9–28.
Dunoff, Jeffrey.1999. ‘The Death of the Trade Regime’, European Journal of International
Law, 10: 733–60.
Dunoff, Jeffrey. 2005. ‘Lotus Eaters: The Varietals Dispute, the SPS Agreement, and
WTO Dispute Resolution’, in George Bermann and Petros C. Mavroidis (eds), Trade
and Health in the WTO, Cambridge University Press: Cambridge, MA.
Edgeworth, F.Y. 1894. ‘The theory of international values’, Economic Journal, 4: 35–50.
Ehlermann, Claus-Dieter. 2004. ‘Reflections on the process of clarification and improve-
ment of the DSU’ in F. Ortino and E.-U. Petersmann (eds), The WTO dispute settlement
system 1995–2003, Kluwer: The Hague, pp 105–14.
Ehlermann, Claus-Dieter and Nicolas Lockhart. 2004. ‘Standard of review in WTO
law’, Journal of International Economic Law, 7: 491–521.
Ehring, Lothar. 2002. ‘De facto discrimination in world trade law: national and most
favoured nation treatment—or equal treatment’, Journal of World Trade 36: 921–77.
 References 483
Eichengreen, Barry. 2006. The European economy since 1945: coordinated capitalism and
beyond, Princeton University Press: Princeton, NJ.
Esserman, Susan and Robert Howse. 2004. ‘Trade disputes require fairer arbitration’,
The Financial Times, 12 September 2004 (available at <http://www.ft.com>).
Estevadeordal, Antoni and L.J. Garay. 1996. ‘Protection, preferential tariff s, and rules of
origin in America’, Integration and Trade: 2–25.
Ethier, Wilfred J. 1998. ‘Regionalism in a multilateral world’, Journal of Political Economy,
106: 1214–45.
Ethier, Wilfred J. 2001. Punishments and dispute settlement in trade agreements,
Copenhagen: EPRU.
Ethier, Wilfred J. 2002. ‘Unilateralism in a multilateral world’, Economic Journal, 112:
266–92.
Ethier, Wilfred J. 2004. ‘Political externalities, non-discrimination and a multilateral
world’, Review of International Economics, 12: 303–20.
Fauchald, Ole Christian. 1998. Environmental taxes and trade discrimination, Kluwer:
Amsterdam.
Feddersen, Christoph T. 1998. ‘Focusing on substantive law in international economic
relations,: the public morals of GATT’s Article XX(a) and “conventional” rules of
interpretation’, Minnesota Journal of Global Trade, 7: 75–101.
Feinberg, Robert. 2005. ‘US antidumping enforcement and macroeconomic indicators
revisited: do petitioners learn?’, Review of World Economics, 141: 612–22.
Fenstra, Robert C. 2004. Advanced international trade, Princeton University Press:
Princeton, NJ.
Finger, Michael (ed.) 1993. Antidumping: How It Works And Who Gets Hurt, University
of Michigan Press: Ann Arbor, MI.
Finger, Michael J. and Julio J. Nogues. 2006. Safeguards and antidumping in Latin
American trading liberalization: fighting fire with fire, World Bank: Washington, DC.
Fontagne, Lionel, Thierry Mayer and Soledad Zignago. 2005. A re-evaluation of the
impact of regional agreements on trade patterns. Mimeo.
Fox, Eleanor. 1997. ‘Toward world antitrust and market access’, American Journal of
International Law, 91: 1–25.
Francois, Joseph, Douglas Nelson and David Palmeter. 1997. ‘Public procurement in the
United States: a post-Uruguay round perspective’ in Bernard M. Hoekman and Petros
C. Mavroidis (eds), Law and Policy in Public Purchasing, University of Michigan Press:
Ann Arbor, MI, pp 105–24.
Friedman, Thomas. 2005. The world is flat: a brief history of the twenty-first century, Farrar,
Straus and Giroux: New York, NY.
Fudenberg, Drew and Jean Tirole. 1991. Game Theory, MIT Press: Cambridge, MA.
Gardner, Richard N. 1980. Sterling-Dollar diplomacy in current perspective: the origins
and prospects of our international economic order, Columbia University Press: New
York, NY.
Gay, Daniel. 2005. ‘Vanuatu’s suspended accession bid: second thoughts?’ in Peter
Gallagher, Patrick A. Low and Andrew L. Stoller (eds), Managing the challenges of WTO
participation, 45 case studies, Cambridge University Press: Cambridge, pp 590–606.
Goh, Gavin and Andreas R. Ziegler. 2003. ‘Retrospective remedies in the TO after
Automotive Leather’, Journal of International Economic Law, 6: 545–64.
484 References
Goldberg, Pinelopi K. and Frank Verboven. 2005. ‘Market integration and conver-
gence on the law of one price: evidence from the European car market’, Journal of
International Economics, 65: 49–73.
Goldsmith, Jack L. and Eric Posner. 2005. The limits of international law, Oxford
University Press: New York, NY.
Goldstein, Judith, Doug Rivers and Michael Tomz. 2007. ‘Institutions in inter-
national relations: understanding the effects of GATT and the WTO on world trade’,
International Organization, forthcoming.
Gollier, C., B. Julien and N. Treich. 2000. ‘Scientific progress and irreversibility: an eco-
nomic interpretation of the precautionary principle’, Journal of Public Economics, 75:
229–53.
Grandmont Jean-Michel and Daniel McFadden. 1972. ‘A technical note on classical
gains from trade’, Journal of International Economics, 2: 109–25.
Greene, William H. 1993. Econometric Analysis (2nd edn), Macmillan: New York, NY.
Grinols, Earl L. and Roberto Perrelli. 2006. ‘The WTO impact on international trade
disputes: an event history analysis’, The Review of Economics and Statistics, 88: 613–24.
Grossman, Gene M. 1986. ‘Imports as a cause of injury: the case of the US steel industry’,
Journal of International Economics, 20: 201–23.
Grossman, Gene M. and Elhanan Helpman. 2001. Special interest politics, MIT Press:
Cambridge, MA and London.
Grossman, Gene M. and Elhanan Helpman. 2002. Interest Groups and Trade Policy,
Princeton University Press: Princeton, NJ.
Grossman, Gene M. and Petros C. Mavroidis. 2003. ‘Would’ve or should’ve? Impaired
benefits due to copyright infringement’ in Henrik Horn and Petros C. Mavroidis (eds),
The WTO Case-Law of 2001, The American Law Institute Reporters’ Studies, Cambridge
University Press: Cambridge, pp 281–99.
Grossman, Gene M. and Petros C. Mavroidis. 2004. ‘Not for Attribution: United
States—Definitive Safeguard Measures on Imports of Circular Welded Carbon
Quality Line Pipe From Korea (WTO Doc WT/DS212/AB/R)’ in Henrik Horn and
Petros C. Mavroidis (eds), The WTO Case-Law of 2002: The American Law Institute
Reporters’ Studies, Cambridge University Press: Cambridge, pp 99–132.
Grossman, Gene M. and Petros C. Mavroidis. 2005. ‘Recurring misunderstanding of
non-recurring subsidies’ in Henrik Horn and Petros C. Mavroidis (eds), The WTO
Case-Law of 2002: The American Law Institute Reporters’ Studies, Cambridge University
Press: Cambridge, pp 78–87.
Grossman, Gene and Esteban Rossi-Hansberg. 2006. The rise of offshoring: it’s not wine
for cloth anymore, Mimeo.
Grossman, Gene M. and Alan Sykes. 2005. ‘A preference for development: the law and
economics of GSP’, World Trade Review, 4: 41–68.
Gupta, Poonam and Arvind Panagariya. 2006. ‘Injury investigations in antidumping
and the super-additivity effect: a theoretical explanation’, Weltwirtschaftliches Archiv/
Review of World economics, 142: 151–64.
Guzman, Andrew and Beth Simmons. 2005. ‘Power plays and capacity constraints: the
selection of defendants in WTO disputes’, Journal of Legal Studies, 37: 557–80.
Hahn, Michael J. 1991. ‘Vital interests in the law of the GATT, an analysis of the GATT’s
security exception’, Michigan Journal of International Law, 12: 558–84.
 References 485
Hansen, Wendy L. and Thomas J. Prusa. 1996. ‘Cumulation and ITC decision-making:
the sum of the parts is greater than the whole’, Economic Inquiry, 34: 746–69.
Hart, Michael. 1998. ‘Fifty years of Canadian statecraft: Canada in the GATT 1947–
1997’, Centre for Trade Policy and Law: Ottawa.
Hart, Michael. 2002. A trading nation, Canadian trade policy from colonialism to global-
ization, University of British Columbia Press: Vancouver.
Hart, Chad E. and John C. Beghin. 2006. ‘Rethinking agricultural domestic support
under the WTO’ in Kym Anderson and Will Martin (eds), Agricultural Trade and
Doha Development Agenda, World Bank: Washington, DC, Chapter 8.
Hawkins, Harry C. 1951. Commercial Treaties and Agreement Principles and Practice,
Rinehart & Co: New York, NY.
Hoda, Anwarul. 2001. Tariff Negotiations and Renegotiations Under the GATT and the
WTO, Cambridge University Press: Cambridge.
Hoekman, Bernard. 2004. Operationalizing the Concept of Policy Space in the WTO:
Beyond Special and Differential Treatment. Mimeo.
Hoekman, Bernard M. and Michel Kostecki. 2001. The Political Economy of the World
Trading System, Oxford University Press: Oxford.
Hoekman, Bernard M. and Petros C. Mavroidis. 1994. ‘Competition, competition pol-
icy and the GATT’, World Economy, 17: 121–50.
Hoekman, Bernard M. and Petros C. Mavroidis. 1996. ‘Dumping, Antidumping and
Antitrust’, Journal of World Trade 30: 27–42.
Hoekman Bernard M. and Petros C. Mavroidis (eds) 1997. Law and Policy in Public
Purchasing, University of Michigan Press: Ann Arbor, MI.
Hoekman, Bernard M. and Petros C. Mavroidis. 2000. ‘WTO Dispute Settlement,
Transparency and Surveillance’, World Economy, 23: 527–42.
Hoekman, Bernard M. and Patrick Messerlin. 2006. ‘Removing the exception of agricul-
tural export subsidies’ in Kym Anderson and Will Martin (eds), Agricultural Trade and
Doha Development Agenda, World Bank: Washington, DC, Chapter 7.
Hoekman, B., C. Michalopoulos and L. Alan Winters. 2004. ‘More Favorable and
Differential Treatment of Developing Countries: Towards a New Approach in the
WTO’, World Economy, 27(4): 481–506.
Hoekman, Bernard M., Francis Ng and Marcelo Olarreaga. 2001. ‘Eliminating excessive
tariffs on exports of least developed countries’, World Bank Policy Research Working
Paper 2604, The World Bank Group: Washington DC.
Horlick, Gary. 1993. ‘How the GATT became protectionist: an analysis of the Uruguay
round draft final antidumping code’, Journal of World Trade, 27: 5–17.
Horlick, Gary. 2002. ‘Problems with the compliance structure of the WTO dispute reso-
lution process’ in Dan Kennedy and James Southwick (eds), The Political Economy of the
International Trade Law, Essays in Honour of Robert E. Hudec, Cambridge University
Press: Cambridge, pp 636–45.
Horlick, Gary and Peggy Clark. 1994. ‘The 1994 WTO Subsidies Agreement’, World
Competition, 17: 41–54.
Horn Henrik. 2006. ‘National treatment in trade agreements’, American Economic
Review, 96: 394–404.
Horn, Henrik, Giovanni Maggi and Robert W. Staiger. 2006. Trade Agreements as
Endogenously Incomplete Contracts. Mimeo.
486 References
Horn, Henrik and Petros C. Mavroidis. 2001. ‘Legal and economic aspects of MFN,
European Journal of Political Economy’, 17: 233–79.
Horn, Henrik and Petros C. Mavroidis. 2003. ‘US—Lamb’ in Henrik Horn and Petros
C. Mavroidis (eds), The WTO Case-Law of 2001: The American Law Institute Reporters’
Studies, Cambridge University Press: Cambridge, pp 72–114.
Horn, Henrik and Petros C. Mavroidis. 2004. ‘Still Hazy After All These Years: The
Interpretation of National Treatment in the GATT/WTO Case-Law on Tax
Discrimination’, European Journal of International Law, 15: 39–69.
Horn, Henrik and Petros C. Mavroidis. 2005a. ‘A comment on US—Offset Act (Byrd
Amendment)’ in Henrik Horn and Petros C. Mavroidis (eds), The WTO Case-Law
of 2003: The American Law Institute Reports’ Studies, Cambridge University Press:
Cambridge.
Horn, Henrik, and Petros C. Mavroidis. 2005b. ‘What is a subsidy?’ in Henrik Horn and
Petros C. Mavroidis (eds), The WTO Case-Law of 2002, The American Law Institute
Reporters’ Studies, Cambridge University Press: Cambridge, pp 220–47.
Horn, Henrik and Petros C. Mavroidis. 2006. ‘WTO Dispute Settlement Database’
(available at http://www.worldbank.org/trade/wtodisputes>).
Horn, Henrik and Petros C. Mavroidis. 2007. ‘International Trade: Dispute Settlement’,
forthcoming in Andrew Guzman and Alan O. Sykes (eds), Handbook on International
Trade, Elgar Publishing: Cheltenham, pp 177–210.
Horn, Henrik and Joseph H.H. Weiler. 2003. ‘European Communities—Measures
Affecting Asbestos and Asbestos-Containing Products’ in Henrik Horn and Petros
C. Mavroidis (eds), The WTO Case-Law of 2001: The American Law Institute Reports’
Studies, Cambridge University Press: Cambridge, pp 14–40.
Hovenkamp, Herbert. 2005. Federal Antitrust policy, Thomson, West: St Paul, MN.
Howse, Robert. 2000. ‘Democracy, Science, and Free Trade: Risk Regulation on Trial at
the World Trade Organization’, Michigan Law Review 98, 7: 2329–57.
Howse, Robert. 2002. ‘The Appellate Body rulings in the Shrimp/Turtle case: a new legal
baseline for the trade and environmental debate (Symposium: Trade, Sustainability
and Global Governance)’, Columbia Journal of Environmental Law 27: 491–521.
Howse, Robert. 2003. ‘India’s WTO challenge to Drug Enforcement conditions in the
European Community GSP: a little known case with major repercussions for “pol-
itical” conditionality in US trade policy’, Chicago Journal of International Law, 4:
385–405.
Howse, Robert. 2006. Should all subsidies get punished? The case for the re-introduction
of non-actionable subsidies. Mimeo.
Howse Robert and Don Regan. 2000. ‘The Product/Process Distinction—An
Illusory Basis for Disciplining “Unilateralism” in Trade Policy’, European Journal of
International Law 11: 249–89.
Howse, Robert and Robert W. Staiger. 2006. ‘United States—Anti-Dumping Act of
1916 (Original Complaint by the European Communities)—Recourse to arbitration
by the United States under 22.6 of the DSU, WT/DS136/ARB, 24 February 2004:
A legal and Economic Analysis’, World Trade Review 4(2): 295–316.
Howse, Robert and E. Turk. 2001. ‘The WTO Impact on Internal Regulations—A Case
Study of the Canada-EC Asbestos Dispute’ in G. de Búrca and J. Scott, The EU and the
WTO: Legal and Constitutional Issues, Hart Publishing: Oxford, pp 283–328.
 References 487
Hudec, Robert E. 1972. ‘GATT or GABB? The Future Design of the General Agreement
on Tariffs and Trade’, The Yale Law Journal, 80: 1299–386.
Hudec, Robert E. 1975. The GATT legal system and the world trade diplomacy, Praeger:
New York, NY.
Hudec, Robert E. 1988. ‘Tiger, tiger in the house: a critical evaluation of the case against
discriminatory trade measures’ in Ernst-Ulrich Petersmann and Meinhard Hilf,
The new GATT round of multilateral trade negotiations: legal and economic problems,
Kluwer: Deventer, pp 165ff.
Hudec, Robert E. 1993a. Enforcing International Trade Law, Butterworth: London.
Hudec, Robert E. 1993b. ‘GATT’s influence on regional agreements: a comment’ in
Jaime de Melo and Arvind Panagariya (eds), New dimensions in regional integration,
CEPR, Cambridge University Press: Cambridge, MA, pp 151–5.
Hudec, Robert E. 1998. ‘GATT/WTO Constraints on National Regulation: Requiem
for an “Aims and Effect” Test’, International Lawyer,32: 619.
Hudec Robert E. 2000. ‘ “Like Product”: the differences in meaning in GATT articles
I and III’ in Thomas Cottier and Petros C. Mavroidis, Regulatory barriers and the
principle of non-discrimination in world trade law, University of Michigan Press: Ann
Arbor, MI, pp 101–23.
Hull, Cordell. 1948. The memoirs of Cordell Hull, volumes I and II, Macmillan: New
York, NY.
Human, Johan, Judith Czako and Jorge Miranda. 2003. A Handbook on Anti-Dumping
Investigations, Cambridge University Press: Cambridge.
Inama, Stefano. 2003. ‘Trade preferences and the WTO negotiations on market access:
battling for compensation of erosion of GSP, ACP and other trade preferences or
assessing and improving their utilization and value by addressing rules of origin and
graduation’, Journal of World Trade, 37: 959–76.
Irwin, Douglas. 1995. ‘The GATT in historical perspective’, American Economic Review,
85: 323–8.
Irwin, Douglas. 1996a. Against the Tide, Princeton University Press: Princeton, NJ.
Irwin, Douglas. 1996b. ‘Changes in US tariffs: prices or policies’, NBER Working Paper
5665.
Irwin, Douglas. 1998a. ‘Changes in US Tariffs: the role of import prices and commercial
policies’, American Economic Review, 88: 1015–26.
Irwin, Douglas. 1998b. ‘The Smoot-Hawley Tariff: A quantitative assessment’, Review of
Economics and Statistics, 80: 326–34.
Irwin, Douglas. 2002. ‘Reciprocity and the origins of US trade liberalization’ in Jagdish
Bhagwati (ed), Going alone: the case for relaxed reciprocity in freeing trade, MIT Press:
Cambridge, MA, pp 61–84.
Irwin, Douglas A. 2003. ‘Causing problems? The WTO review of causation and injury
attribution in U.S. Section 201 cases’, World Trade Review, 2: 297–325.
Irwin, Douglas. 2005. Free Trade Under Fire (2nd edn), Princeton University Press:
Princeton, NJ.
Iynedjian, Marc. 2002. L’Accord de l’Organisation mondiale du commerce sur l’application
des mesures sanitaires et phytosanitaires, Une analyse juridique, L.G.D.J. : Paris.
Jackson, John H. 1969. World trade and the law of the GATT, Bobbs-Merril: Indianapolis,
IA.
488 References
Jackson, John H. 1997. The world trading system: law and policy of international economic
relations (2nd edn), Cambridge University Press: Cambridge.
Jackson, John H, William J. Davey and Alan O. Sykes. 2002. Legal problems of inter-
national economic relations, WEST Group: St. Paul, MN.
Janow Merit and Robert W. Staiger. 2003. ‘Canada-Measures Affecting the Importation
of Dairy Products and the Exportation of Milk’ in Henrik Horn and Petros C.
Mavroidis (eds), The WTO Case-Law of 2001: The American Law Institute Reporters’
Studies, Cambridge University Press: Cambridge, pp 236–80.
Jansen, Marion and Eddie Lee. 2007. Trade and Employment: challenges for policy research,
ILO and WTO joint publication: Geneva.
Johnson, H. G. 1953–4. ‘Optimum tariffs and retaliation’, Review of Economic Studies,
1: 142–53.
Josling, Timothy E., Stefan Tangerman and T.K. Warley. 1996, Agriculture in the GATT,
Macmillan: Basingstoke.
Karsenty, Guy and Sam Laird. 1986. ‘The GSP: a quantitative assessment of the dir-
ect trade effects and of policy options’, UNCTAD Discussion Paper 18, UNCTAD:
Geneva.
Katz, M.L. and C. Shapiro. 1985. ‘Network externalities, competition and compatibil-
ity’, American Economic Review, 75: 424–40.
Keck, Alexander and Patrick Low. 2003. Special and Differential Treatment in the
WTO: Why, When and How? Mimeo.
Kelly, Kenneth. 1988. ‘The analysis of causality in escape clause cases’, Journal of
Industrial Economics, 37: 187–207.
Kessie, Edwini. 2004. ‘The early harvest negotiations in 2003’ in Federico Ortino
and Ernst-Ulrich Petersmann (eds), The WTO dispute settlement system 1995–2003,
Kluwer: The Hague, pp 115–50.
Knetter, Michael and Tom Prusa. 2003. ‘Macroeconomic factors and antidumping fil-
ings: evidence from four countries’, Journal of International Economics, 61: 1–17.
Kono, Daniel Y. 2006. ‘Optimal obfuscation: democracy and trade policy transparency’,
American Political Science Review, 100: 369–84.
Kovacic, William E. 2006. Price differentiation in antitrust and trade instruments.
Mimeo.
Kramer, Larry. 1995. ‘Extra-territorial application of antitrust law after the Insurance
Antitrust Case: a reply to Professors Lowenfeld and Trimble’, American Journal of
International Law, 89: 750–8.
Krishna, Kala. 2005. ‘Understanding Rules of Origin’, NBER Working Paper No 11150.
Krishna, Pravin. 1998. ‘Regionalism and multilateralism: a political economy approach’,
Quarterly Journal of Economics, 113: 227–51.
Krishna, Pravin. 2005. Trade Blocks, Economics and Politics, Cambridge University Press:
New York, NY.
Krueger, Anne O. 1997. ‘Free trade areas vs. customs unions’, Journal of Development
Economics, 54: 169–87.
Krugman, Paul. 1983. ‘Targeted industrial policies: theory and evidence’, in Industrial
Change and Public Policy, Kansas City: Federal Reserve Bank: Kansas, KS, pp 123–55.
Krugman, Paul. 1991. ‘The Move Toward Free Trade Zones’, Economic Review,
November–December 1991: 1–24.
 References 489
Krugman, Paul. 1995. ‘Growing world trade: causes and consequences’, Brooking Papers
on Economic Activity, 1: 327–62.
Krugman, Paul and Maurice Obstfeld. 1997. International Economics (4th edn), Addison-
Wessley: New York, NY.
Krugman, Paul and Robin Wells. 2005. Microeconomics, Worth Publishers: New York, NY.
Kucik, Jeffrey and Eric Reinhardt. 2007. ‘Does flexibility promote cooperation? “Efficient
breach” in the global trade regime’. Mimeo, available at <http://user-www.service.
emory.edu/~erein/>.
Kuhn, Thomas. 1962. The structure of scientific revolutions, Harvard University Press:
Cambridge, MA.
Kuhn, Thomas. 1996. The road since structure, Harvard University Press: Cambridge, MA.
Lal, Deepak. 2000. The poverty of development economics, MIT Press: Cambridge, MA.
Lawrence, Robert. 2003. Crimes and Punishment. Retaliation Under the WTO, Institute
for International Economics: Washington, DC.
Leddy, John. 1958. ‘GATT: a cohesive influence in the free world’, Journal of Farm
Economics, 40: 228–37.
Lee, Yong-Shik. 2005. Safeguard measures in world trade: the legal analysis (2nd edn),
Kluwer: The Hague.
Levy, Philip and T.N. Srinivasan. 1996. ‘Regionalism and the (Dis)advantage of Dispute-
Settlement Access’, American Economic Association Papers and Proceedings, 86(2):
93–8.
Levy, Philip I. and T.N. Srinivasan. 1999. ‘Regionalism and the (dis-)advantage of dis-
pute settlement access’ in Jagdish Bhagwati, Pravin Krishna and Arvind Panagariya,
Trading blocks, MIT Press: Cambridge, MA, pp 531–40.
Li Changying and Keith E. Maskus. 2006. ‘The Impact of Parallel Imports on Cost-
Reducing Research and Development’, Journal of International Economics, 68:
443–55.
Limão, Nuno. 2006a. ‘Preferential Trade Agreements as Stumbling Blocks for Multilateral
Trade Liberalization: Evidence for the U.S.’, American Economic Review,96: 896–914.
Limão, Nuno. 2006b. ‘Preferential vs. Multilateral Trade Liberalization: Evidence and
Open Questions’, World Trade Review, 5: 155–76.
Limão, Nuno. 2006c. ‘Trade Preferences to Small Developing Countries and the Welfare
Costs of Lost Multilateral Liberalization’, World Bank Economic Review, 20: 217–40.
Limão, Nuno, Christian Broda and David Weinstein. 2006. ‘Optimal Tariffs: the evi-
dence’, NBER Working Paper 12033.
Limão, Nuno and Kamal Shaggi. 2005. Tariff Retaliation versus Financial Compensation
in the Enforcement of International Trade Agreements. Mimeo.
Lloyd, Peter. 1993. ‘A tariff substitute for rules of origin in free trade areas’, The World
Economy, 16: 691–712.
Low, Patrick A. 1995. ‘Pre-shipment Inspection Services’, World Bank Discussion Paper
No 278, World Bank: Washington, DC.
Lowenfeld, Andreas F. 1995. ‘Conflict, balancing of interests, and the exercise of juris-
diction to prescribe: reflections on the Insurance Antitrust Case’, American Journal of
International Law, 89: 42–53.
Lucas, Robert, E. 1988. ‘On the mechanism of economic development’, Journal of
Monetary Economics, 22: 3–42.
490 References
MacPhee, Craig R. and Victor Iwuagwu Oguledo. 1991. ‘The trade effects of the US
GSP’, Atlantic Economic Journal, 19: 19–26.
Maggi, Giovanni. 1999. ‘The value of commitment with imperfect observability and
private information’, The Rand Journal of Economics, 30: 555–74.
Maggi, Giovanni and Andres Rodriguez-Clare. 1998. ‘The value of trade agreements in
the presence of political pressures’, Journal of Political Economy, 106: 574–601.
Maggi, Giovanni and Andres Rodriquez-Clare. 2006. A political economy theory of
trade agreements. Mimeo.
Marceau, Gabrielle and Joel Trachtman. 2002. ‘The TBT, the SPS and the GATT: a map
of the WTO law of domestic regulation of good’, Journal of World Trade, 36: 811–82.
Martin, Alberto and Wouter Vergote. 2004. Antidumping: Welfare enhancing retali-
ation? Mimeo (available at <http://www.columbia.edu/~wrv13/research.htm>).
Maskin, Eric and Jean Tirole. 1999. ‘Unforeseen developments and incomplete con-
tracts’, Review of Economic Studies, 66: 83–114.
Mathis, James H. 2002. Regional trade agreements in the GATT/WTO, Article XXIV and
the internal trade requirement, TMC Asser Press: The Hague.
Mathis, James H. 2006. ‘Regional trade agreements and domestic regulation: what reach
for “other restrictive regulations of commerce”?’ in Lorand Bartels and Federico Ortino
(eds), Regional trade agreements and the WTO legal system, Oxford University Press:
Oxford, pp 79–108.
Matsushita, Mitsuo, Thomas J. Schonbaum and Petros C. Mavroidis. 2006. The World
Trade Organization, Law Practice, and Policy (2nd edn), Oxford University Press:
Oxford.
Mattli, Walter. 1999. The logic of regional integration: Europe and beyond, Cambridge
University Press: Cambridge.
Mattoo, Aaditya. 1997. ‘Economic theory and the procurement agreement’ in Bernard
M. Hoekman and Petros C. Mavroidis (eds), Law and Policy in Public Purchasing,
University of Michigan Press: Ann Arbor, MI, pp 57–72.
Mattoo, Aaditya and Petros C. Mavroidis. 1995. ‘The EEC/Japan consensus on cars:
interaction between trade and competition policies’, World Economy, 18: 347–66.
Mattoo Aaditya and Petros C. Mavroidis. 1997. ‘Trade, Environment and the WTO:
The Dispute Settlement Practice Relating to Art. XX of the GATT’ in Ernst Ulrich
Petersmann (ed.), International Trade Law and the GATT/WTO Dispute Settlement
System, Kluwer: Amsterdam, pp 325–44.
Mattoo Aaditya, Devesh Roy and Arvind Subramanian. 2003. ‘The Africa Growth and
Opportunity Act and its rules of origin: generosity undermined?’, The World Economy,
26: 829–51.
Mattoo, Aaditya and Arvind Subramanian. 1998. ‘Regulatory Autonomy and Multilateral
Disciplines: The Dilemma and Possible Resolution’, Journal of International Economic
Law, 1: 303.
Mavroidis, Petros C. 1993. ‘Government Procurement Agreement; the Trondheim case:
the remedies issue’, Aussenwirtschaft, 48: 77–94.
Mavroidis, Petros C. 2000. ‘Remedies in the WTO Legal System: Between a Rock and a
Hard Place’, European Journal of International Law, 11: 763–813.
Mavroidis, Petros C. 2002. ‘Judicial supremacy, judicial restraint and the issue of consist-
ency of preferential trade agreements with the WTO: the apple in the picture’ in Dan
 References 491
Kennedy and James Southwick (eds), The Political Economy of the International Trade
Law, Essays in Honour of Robert E. Hudec, Cambridge University Press: Cambridge, pp
583–601.
Mavroidis, Petros C. 2004a. ‘Amicus curiae briefs before the WTO: Much ado about
nothing’ in Armin von Bogdandy, Petros C. Mavroidis and Yves Meny (eds), In honour
of Claus-Dieter Ehlermann, Kluwer: Amsterdam, pp 317–30.
Mavroidis, Petros C. 2004b. ‘The need to micro-manage regulatory diversity’ in K. Basu,
H. Horn, L. Roman, and J. Shapiro (eds), International Labour Standards, Blackwell
Publishing: Oxford, pp 314–25.
Mavroidis, Petros C. 2005. ‘Do not ask too many questions: the institutional arrange-
ments for accommodating regional integration within the WTO’ in E. Kwan Choi
and James C. Hartigan (eds), Handbook of International Trade, volume II, Economic
and Legal Analysis of Trade Policy Institutions, Blackwell Publishing: Malden, MA, pp
239–78.
Mavroidis, Petros C. 2006. ‘If I don’t do it somebody else will (or won’t)’, Journal of World
Trade, 40: 187–214.
Mavroidis, Petros C. and Patrick Messerlin. 1998. ‘Has Art. 90 ECT prejudiced the sta-
tus of property ownership?’ in Thomas Cottier and Petros C. Mavroidis (eds), State-
trading in the 21st Century, The University of Michigan Press: Ann Arbor, MI, pp
345–60.
Mavroidis, Petros C., Patrick Messerlin and Jasper M. Wauters. 2007. The Law and
Economics of Contingent Protection, Elgar: Cheltenham.
Mavroidis, Petros C. and Damien J. Neven. 1999. ‘Some reflections on extraterritoriality
in international economic law: a law and economic analysis’ in Mélanges en homage de
Michel Waelbroeck, Bruylant: Bruxelles, pp 1297–325.
Mavroidis, Petros C. and André Sapir. 2007. Mexico—AD Measures on Rice: Don’t ask
me no questions and I won’t tell you no lies. Mimeo.
McMahon, Joseph. 2006. The WTO Agreement on Agriculture, A Commentary, Oxford
University Press: Oxford.
Meade, James. 1942. The economic basis of durable peace, G. Allen and Unwin Ltd:
London.
Meessen, Karl M. 1996. Extraterritorial jurisdiction in theory and practice, Kluwer:
London.
Melloni, Mattia. 2005. The principle of national treatment in the GATT, Bruylant:
Brussels.
Messerlin, Patrick. 2001. Measuring the Cost of Protection in Europe, Institute Of
International Economics: Washington, DC.
Michalopoulos, Constantine. 2001. Developing Countries in the WTO, Palgrave:
Basingstoke.
Miller, James N. 2000. ‘Origins of the GATT: British resistance to American multi-
lateralism’, Cambridge University, Jerome Levy Economics Institute at Bard College,
Working Paper no 318.
Moore, Michael O. 2002. ‘Commerce Department Antidumping Sunset reviews: a first
assessment’, Journal of World Trade, 36: 675–98.
Moore, Michael O. 2004. Can the US dump antidumping? Evidence from past reforms.
Mimeo.
492 References
Neven, Damien Jules. 2001. ‘How Should Protection be Evaluated in Article III GATT
Disputes?’, European Journal of Political Economy, 17: 421–44.
Neven, Damien J. and Joseph H.H. Weiler. 2006.’ Japan—Measures Affecting the
Importation of Apples: One Bad Apple?’ in Henrik Horn and Petros C. Mavroidis
(eds), The WTO Case Law of 2003, The ALI Reporters’ Studies, Cambridge University
Press: Cambridge, pp 280–310.
Nolde, Boris Baron. 1932. ‘La clause de la nation la plus favorisée et les tariffs préféren-
tiels’, 39 Receuil des cours 35, Hague Academy of International Law: The Hague.
Nordström, Häkan. 2005a. Cost of WTO litigation, legal aid and small claims proce-
dures, Mimeo.
Nordström, Häkan. 2005b. ‘The WTO secretariat in a changing world’, Journal of World
Trade, 39: 819–53.
O’Connor, Bernard. 2005. ‘The structure of the agreement on agriculture’ in Bernard
O’Connor (ed.), Agriculture in WTO Law, Cameron May: London, pp. 83–90.
Odell, John. 2005. ‘Chairing a WTO negotiation’, Journal of International Economic
Law, 8: 425–48.
Odell, John and Barry Eichengreen. 1998. ‘The United States, the ITO and the WTO:
exit options, agent slack, and presidential leadership’ in Anne O. Krueger (ed.), The
WTO as an international organization, University of Chicago Press: Chicago, ILL, pp
181–209.
Oesch, Matthias. 2003. Standards of Review in WTO Dispute Resolution, Oxford
University Press: Oxford and New York, NY.
Olsen, Sven. 2005. ‘The negotiation of the agreement on agriculture’ in Bernard
O’Connor (ed.), Agriculture in WTO Law, Cameron May: London, pp 43–82.
Ortino, Federico. 2003. Basic Legal Instruments for the Liberalisation of Trade: a compara-
tive analysis of EC and WTO Law, Hart Publishing: Oxford.
Ostry, Sylvia. 1997. The post-cold war trading system, who’s on first?, University of Chicago
Press: Chicago, ILL.
Özden, Czaglar and Eric Reinhardt. 2003. ‘The perversity of preferences: GSP and
developing country trade policies, 1976–2000’, World Bank Working Paper 2955, The
World Bank: Washington, DC.
Palmeter, David N. 1995. ‘US Implementation of the Uruguay Round Antidumping
Code’, Journal of World Trade, 29: 39–82
Palmeter, David N. and Petros C. Mavroidis. 1998. ‘The WTO legal system: sources of
law’, American Journal of International Law, 92: 398–413.
Palmeter, David N. and Petros C. Mavroidis. 2004. Dispute Settlement in the WTO,
Practice and Procedure (2nd edn), Cambridge University Press: Cambridge.
Panagariya, Arvind. 2005. ‘Liberalizing agriculture’, Foreign Aff airs, 84.
Pauwelyn Joost. 1998. ‘Evidence, Proof and Persuasion in WTO Dispute Settlement:
Who Bears the Burden’, Journal of International Economic Law, 1: 227–58.
Pauwelyn, Joost. 1999. ‘The WTO agreement on Sanitary and Phytosanitary (SPS) meas-
ures as applied in the first three SPS disputes: EC—Hormones, Australia—Salmon,
and Japan—Varietals’, Journal of International Economic Law, 2: 641–64.
Pauwelyn, Joost. 2000. ‘Enforcement and countermeasures in the WTO: rules are
rules—toward a more collective approach’, American Journal of International Law, 94:
335–47.
 References 493
Pauwelyn, Joost. 2003. Conflicts of norms in public international law. How WTO relates to
other rules of international law, Cambridge University Press: Cambridge.
Pelkmans, Jacques. 2003. ‘Mutual recognition in goods and services: an economic
perspective’, Working Paper no 16, European Network of Economic Policy Research
Institutes.
Penrose, Ernest F. 1953. Economic planning for peace, Princeton University Press:
Princeton, New Jersey, NJ.
Petersmann, Ernst-Ulrich. 1991. ‘Non violation complaints in public international trade
law’, German Yearbook of International Law, 34: 175–231
Petersmann, Ernst-Ulrich. 1993. ‘International competition rules for the GATT—MTO
world trade and legal system’, Journal of World Trade, 27: 35–86.
Petersmann, Ernst-Ulrich. 2000. ‘International Trade Law and International
Environmental Law: Environmental Taxes and Border Tax Adjustment in WTO
Law and EC Law’ in R. Revesz, P. Sands and T. Stewart (eds), Environmental Law,
The Economy and Sustainable Development, Cambridge University Press: Cambridge,
MA, pp 127ff.
Pienaar, Natalie. 2003. Economic aspects of the consistency requirement in the WTO agree-
ment on the application of SPS measures, The Institute for International Economic
Studies: Stockholm.
Piermartini, Roberta. 2005. ‘Harmonization and mutual recognition of product stand-
ards: the effect on intra-EU trade’, WTO Discussion Paper, WTO: Geneva.
Popper, Karl. 1992. The logic of scientific discovery, Routledge: New York, NY.
Prusa, Thomas J. 1992. ‘Why are so many antidumping petitions withdrawn?’, Journal of
International Economics, 33: 1–20.
Prusa, Thomas J. 2001. ‘On the spread and impact of antidumping’, Canadian Journal of
Economics, 34: 591—611.
Prusa, Thomas J. 2005. ‘Anti-dumping: a growing problem in international trade’, The
World Economy 28(5): 683–700.
Qureshi, Asif H. 2006. Interpreting WTO Agreements, Cambridge University Press:
Cambridge.
Raustiala, Kal. 2006. ‘The evolution of territoriality: international relations and
American law’, in Miles Khaler and Barbara Walter (eds), Territoriality and Conflict in
the age of Globalization, Cambridge University Press: Cambridge, MA.
Regan, Donald H. 2002. ‘Regulatory Purpose and “Like Products” in Article III:4 of
the GATT (With Additional Remarks on Article II:2)’, Journal of World Trade 36:
443–78.
Regan, Donald H. 2003. ‘The Dormant Commerce Clause and the Hormones Problem’
in Thomas Cottier and Petros C. Mavroidis (eds), The Role of the Judge in International
Trade Regulation: Experience and Lessons for the WTO, University of Michigan Press:
Ann Arbor, MI, pp 91–117.
Regan, Donald H. 2006. ‘What are trade agreements for? Two conflicting stories told
by economists, with a lesson for lawyers’, Journal of International Economic Law,
forthcoming.
Ricardo, David. 1817. ‘On the principles of political economy and taxation’ in P. Sraffa
(ed.), The works and correspondence of David Ricardo, 1951, Cambridge University Press:
Cambridge.
494 References
Rodriguez, Francisco and Dani Rodrik. 2001. ‘Trade policy and economic growth: a
skeptic’s guide to the cross-national evidence’, in Ben S. Bernanke and Kenneth S.
Rogoff (eds), NBER Macroeconomics Annual 2000, MIT Press: Cambridge, MA.
Rodrik, Dani. 1999. Making Openness Work: the new global economy and the developing
countries, Overseas Development Council, Washington, DC.
Rodrik, Dani. 2002. What is wrong with the (Augmented) Washington Consensus?
Mimeo (available at <http://www.sopde.org/discussion.htm>).
Roessler, Frieder. 1993. ‘The relationship between regional integration agreements and
the multilateral trade order’ in Kym Anderson and Richard Blackhurst (eds), Regional
integration and the global trading system, Harvester Wheatsheaf: Exeter, pp 311–25.
Roessler, Frieder. 1996. ‘Diverging Domestic Policies and Multilateral Trade Integration’
in Jagdish Bhagwati, and Robert E. Hudec (eds), Fair Trade and Harmonization, vol.
2: Legal Analysis, Cambridge University Press: Cambridge, MA, pp 21ff.
Roessler, Frieder. 2000. ‘The institutional balance between the judicial and the polit-
ical organs of the WTO’ in Marco Bronckers and Reinhard Quick (eds), New direc-
tions in international economic law—Essays in honour of John H. Jackson, Kluwer Law
International: The Hague, pp 325–47.
Rose, Andrew K. 2004. ‘Do we really know that the WTO increases trade?’, American
Economic Review, 94: 98–114.
Rosendorf, Peter B. and Helen V. Milner. 2001. ‘The optimal design of international
trade institutions: uncertainty and escape’, International Organization, 55: 829–57.
Rutkowski, Aleksander. 2007. ‘Withdrawals of anti-dumping complaints in the EU: a
sign of collusion’, The World Economy, 30: 470–503.
Sachs, Jeffrey D. and Andrew M. Warner. 1995. ‘Economic reform and the process of
global integration’, Brookings Papers on Economic Activity, 1: 1–95.
Saggi, Kamal and Nese Sara. 2006. National treatment at the WTO: the roles of product
and country heterogeneity. Mimeo.
Sanchirico, Chris. 1997. ‘The burden of proof in civil litigation: a simple model of mech-
anism design’, International Review of Law and Economics, 17: 431–65.
Sapir, André. 1998. ‘The Political Economy of EC Regionalism’, European Economic
Review, 42: 717–32.
Sapir, André. 2000a. ‘EC regionalism at the turn of the Millennium: Towards a new
paradigm?’, The World Economy, 23: 1135–48.
Sapir, André. 2000b. ‘Trade Regionalism in Europe: towards an Integrated Approach’,
Journal of Common Market Studies, 38: 151–63.
Sapir, André and Lars Lundberg. 1984. ‘The US GSP and its impacts in R.E. Baldwin and
A.O. Krueger (eds), The structure and evolution of recent US trade policy, The University
of Chicago Press: Chicago, ILL, pp 195–231.
Sapir, André and Joel Trachtman. 2007. Subsidization, price suppression and expertise:
causation and precision in US—Upland Cotton. Mimeo.
Scalia, Antonin. 1997. A matter of interpretation, Princeton University Press: Princeton, NJ.
Schelling, Thomas. 1960. The strategy of conflict, Harvard University Press: Cambridge, MA.
Schiff, Maurice and L. Alan Winters. 2003. Regional integration and development, Oxford
University Press: Oxford.
Schott, Jeffrey. 1989. ‘More Free Trade Areas?’ in Jeffrey Schott (ed.), Free Trade Areas and
US Trade Policy, Institute of International Economics: Washington, DC, pp 1–58.
 References 495
Schropp, Simon. 2005. ‘The case for tariff compensation in WTO dispute settlement’,
Aussenwirtschaft, 60: 485–528.
Schwartz, Warren F. and Alan O. Sykes. 1998. ‘The positive economics of the most-
favored nation obligation and its exceptions in the WTO/GATT system’in J. Bhandari
and A. Sykes (eds), Economic Dimensions in International Law, Cambridge University
Press: Cambridge, pp 43–75.
Schwarz, Warren F. and Alan O. Sykes. 2002. ‘The economic structure of renegotiation
and dispute resolution in the WTO/GATT system’, Journal of Legal Studies XXXI:
179–204, reprinted in Petros C. Mavroidis and Alan Sykes (eds) 2005. The WTO and
International Trade Law/Dispute Settlement, Edward Elgar Publishing: Cheltenham,
pp 52–77.
Scott, Joanne. 2003. ‘European Regulation of GMOs and the WTO’, Columbia Journal
of European Law, 9: 213–40.
Scott, Robert E. and Paul B. Stephan. 2006. The limits of Leviathan, Cambridge
University Press: New York, NY.
Serra Jaime, Guillermo Aguilar, Jose Cordoba, Gene Grossman, Carla Hills, John Jackson,
Julius Katz, Pedro Noyola and Michael Wilson. 1997. Reflections on regionalism; report
of the study group on international trade, Carnegie Endowment for International Peace,
The Brookings Institution Press: Washington, DC.
Shaffer, Greg. 2003. Defending interests, the public-private partnership in WTO litigation,
Brookings: Washington, DC.
Smith Alasdair and Antony Venables. 1991. ‘Counting the cost of voluntary export
restraints in the European market’ in E. Helpman and A. Razin (eds), International
Trade and Trade Policy. MIT Press: Cambridge, MA, pp 187–220.
Snyder, Richard Carlton. 1948. The most-favored-nation clause: analysis with particular
reference to recent treaty practice and tariff s, Columbia University Press: New York,
NY.
Spadi, Fabio. 2000. ‘Discriminatory Safeguards in the Light of the Admission of the
People’s Republic of China to the World Trade Organization’, Journal of International
Economic Law, 5: 421–43.
Srinivasan, T.N. 1998. Developing Countries and the Multilateral Trading System: GATT
1947 to Uruguay Round and Beyond, Westview Press: Boulder, CO.
Srininivasan, T.N. 2005. ‘Non-discrimination in GATT/WTO: was there anything to
being with and is there anything left?’, World Trade Review, 4: 69–95.
Staiger, Robert W. and Guido Tabellini. 1987. ‘Discretionary trade policy and excessive
protection’, American Economic Review, 77: 823–37.
Staiger, Robert W. and Guido Tabellini. 1999. ‘Do GATT rules help governments make
domestic commitments?’, Economics and Politics, 11: 109–44.
Stevens, Christopher. 2002. The Future of SDT for Developing Countries in the WTO,
Institute for Development Studies, Sussex, Mimeo.
Stewart, Terence P., Amy S. Dyer and J. Daniel Stirk. 2006. Antidumping: Overview of
the Agreement. Mimeo.
Stirling, Andrew (ed.) 2001. ‘On science and precaution in the management of techno-
logical risk’, (available at http://www.esto.jrc.es/detailshort:cfm?ID_report=809>).
Subramanian, Arvind and Shang Jin Wei. 2007. ‘The WTO promotes trade strongly but
unevenly’, Journal of International Economics, forthcoming.
496 References
Summers, Larry. 1991. ‘Regionalism and the world trading system’ in Policy Implications
of Trade and Currency Zones, The Federal Reserve Bank of Kansas City, pp 42–65.
Sunstein, Cass. 1999. One case at a time, Harvard University Press: Cambridge, MA.
Sunstein, Cass. 2002. Risk and reason, Harvard University Press: Cambridge, MA.
Sunstein, Cass. 2003. ‘Beyond the precautionary principle’, University of Pennsylvania
Law Review, 151: 1003–76.
Swinbank, Alan and Carolyn Tanner. 1996. Farm Policy and Trade Conflict, University
of Michigan Press: Ann Arbor, MI.
Sykes, Alan. 1990. ‘GATT safeguards reform: the injury test’, in Michael Trebilcock and
R.York, Fair Exchange: Reforming Trade Remedy Laws, Policy Study 11, C.D. Howe
Institute: Toronto.
Sykes, Alan. 1991. ‘Protectionism as a “safeguard”: a positive analysis of GATT Article XIX
with normative speculations’, University of Chicago Law Review, Winter, 58: 255–82.
Sykes, Alan O. 1995. Product standards for internationally integrated goods markets,
Brookings Institution: Washington, DC.
Sykes, Alan O. 1999. ‘Regulatory protectionism and the law of international trade’,
University of Chicago Law Review, 66: 1–45.
Sykes, Alan O. 2000. ‘Regulatory competition or regulatory harmonization? A silly ques-
tion?’, Journal of International Economic Law, 2: 257–64.
Sykes, Alan O. 2003a. ‘The economics of the WTO rules on subsidies and countervail-
ing measures’, John M. Olin program in Law & Economics working papers, No 186,
University of Chicago.
Sykes, Alan O. 2003b. ‘The Least Restrictive Means’, University of Chicago Law Review,
70: 403–16.
Sykes, Alan O. 2003c. ‘The Safeguards Mess: A Critique of WTO Jurisprudence’, World
Trade Review, 3: 216–96
Sykes, Alan O. 2006. The WTO Agreement on Safeguards, A Commentary, Oxford
University Press: Oxford.
Trachtman, Joel. 1999. ‘The domain of the WTO dispute resolution’, Harvard Journal of
International Law, 40: 333–63.
Trachtman Joel and Gabrielle Mareceau. 2002. ‘TBT, SPS, and GATT: A Map of the
WTO Law of Domestic Regulation’, Journal of World Trade, 36: 811–82.
Trebilcock, Michael and Robert Howse. 2005. International Trade Regulation (3rd edn),
Routledge: London.
Trebilcock, Michael and Julie Soloway. 2002. ‘International trade policy and domes-
tic food safety regulation: the case for substantial deference by the WTO Dispute
Settlement Body under the SPS agreement’ in Daniel M. Kennedy and James
D. Southwick (eds), The political economy of international trade law, Essays in honor of
Robert E. Hudec, Cambridge University Press: Cambridge, MA, pp 537–74.
Trimble, Phillip, R. 1995. ‘The Supreme Court and International Law: the demise of
Restatement Section 403’, American Journal of International Law, 89: 53–7.
Tupy, Marian. 2005. ‘Africa’s war on poverty begins at home’, The Financial Times,
19 December 2005 (available at <http://www.ft.com>).
Tyson D’ Andrea, Laura. 1992. Who’s bashing whom? Trade conflict in high-technology
industries, Institute for International Economics: Washington, DC.
Vagts, Detlev. 2003. ‘Extra-territoriality and the corporate governance law’, American
Journal of International Law, 97: 289–94.
 References 497
Van Damme, Isabelle. 2007. ‘The interpretation of schedules of commitments’, Journal of
World Trade, 41: 1–52.
Van den Bossche, Pieter. 2005. The WTO, Texts, cases and materials, Cambridge
University Press: Cambridge.
Verhoosel, Gaetan. 2002. National treatment and WTO dispute settlement, Hart
Publishing: Oxford.
Vermulst, Edwin. 1992. ‘Rules of origin as commercial policy instruments—revisited’,
Journal of World Trade, 26: 61–102.
Vermulst, Edwin. 1995. ‘Rules of Origin in the Future: Selected Issues’ in Jacques
Bourgeois, Frederique Perrod and Eric Gippini Fournier (eds), The Uruguay Round
Results , Brussels Inter-university Press: Brussels, pp 353–60.
Vermulst, Edwin. 1996. EC Antidumping Law and Practice, Sweet & Maxwell: London.
Vermulst, Edwin and Hiroshi Imagawa. 2005. ‘The Agreement on Rules of Origin’, in
Patrick Macrory, Arthur Appleton, and Michael Plummer (eds), The World Trade
Organization: Legal, Economic and Political Analysis vol. I, Springer: New York, NY, pp
601–78.
Vermulst, Edwin and Paul Waer. 1990. ‘European Community rules of origin as com-
mercial policy instruments’, Journal of World Trade, 24: 55–100.
Vermulst, Edwin and Paul Waer. 1997. EC Antidumping Law and Practice, Sweet &
Maxwell: London.
Viner, Jacob. 1924. ‘The most-favored-nation clause in American commercial treaties’,
Journal of Political Economy, 32: 126–50.
Viner, Jacob. 1950. The customs union issue, Carnegie Endowment for International Peace:
New York, NY.
Vogel, David. 1997. Barriers or benefits? Regulation in transatlantic trade, Brookings
Institution: Washington, DC.
Wang, Z.K. and L.A. Winters. 2000. ‘Putting “Humpty” Together Again: Including
Developing Countries in a Consensus for the WTO’, CEPR Policy Paper No 4.
Weatherill, Stephen. 2002. ‘Preemption, Harmonisation and the Distribution of
Competence to Regulate the Internal Market’ in Catherine Barnard and Joanne Scott
(eds), The Law of The Single European Market, Hart Publishing: Oxford, pp 41–73.
Weiss, Friedl. (ed.) 2000. Improving WTO dispute settlement procedures: issues and lessons
from the practice of other international courts and tribunals, Cameron May: London.
Whalley, John. (ed.) 1989. Developing Countries and the Global Trading System, University
of Michigan Press: Ann Arbor, MI.
Wilcox, Clair. 1949. A charter for world trade, Macmillan: New York, NY.
Williamson, Oliver E. 2005. ‘Why law, economics and organization?’, Annual Review of
Law and Social Science, 1: 369–96.
Wilson, J.S. and T. Otsuki. 2004. Standard and technical regulations and firms in devel-
oping countries: new evidence from a World Bank technical barriers to trade survey, The
World Bank: Washington, DC.
Winters, Alan L. 1996. ‘Regionalism versus multilateralism’, Discussion Paper No 1525,
CEPR: London.
Zanardi, Maurizio. 2004. ‘Antidumping: what are the numbers to discuss?’, The World
Economy: 403–33.
Zeiler, Thomas W. 1999. Free trade, free world: the advent of the GATT, The University of
North Carolina Press: Chapel Hill, NC.
This page intentionally left blank
 Index

abus de droit 275 Brander, James 180

access to markets 385–7 Brenton, Paul 131
accession Bretton Woods conference 3
General Agreement on Tariffs and Trade Bronckers, Marco 432
(GATT) 391
World Trade Organization (WTO) 391–4 Cadot, Olivier 353, 354
ad valorem duties 79 Cairns Group 79, 204–5
administration of quantitative restrictions/ Canada
quotas (QRs) 63–6 commitments 81–2
aggregate measurement of support (AMS) 207 negotiations on GATT and 8
agricultural subsidies see farm subsidies Canada–US Free Trade Agreement 149
alternative duties 79 Capling, Ann 8
amendment of GATT 397–8 cartels 9, 18
animal life, exception to national export 54
treatment 262–3 certificates of origin 159
Antidumping Committee 344 certification of schedules 86
antidumping measures 18, 142, 321 Charnovitz, Steve 261–2, 276, 281
discriminatory pricing 338–43, 361–2 commitment theory 13–14
establishing injury 349–52 Committee on Balance of Payments
finding dumping 343–9 Restrictions 332
predation 362–3 Committee on Regional Trade Arrangements
price differentiation in antitrust (CRTA) 154–5, 156–8, 168,
statutes 358–61 169–70
strategic dumping 363–4 Committee on Trade and Development
sunset reviews 352–8 (CTD) 142
anti-globalization movement 2 Commonwealth preference 9
antitrust statutes, price differentiation comparative advantage 12–13
in 358–61 compensation 420–1
Appellate Body 411 safeguards 376–8
applied duties 71 compliance monitoring 424–5
arbitrary discrimination 271–4 compliance panels 419–20
Australia, negotiations on GATT and 8 compound duties 79
authentic interpretations 396–7 concessions
erosion 120–1, 194
Bagwell, Kyle 15, 16, 17, 18, 180, 282, 428, scheduling of 88–92
432 suspension of 421–4
balance of payments (BoP) 9, 60, 321, 331–4 withdrawing concessions from members
quantitative restrictions/quotas (QRs) leaving WTO 95–7
and 67 consensus 395–6
see also infant industry protection contingencies see state contingencies
Barfield, Claude 431, 451 Contracting Parties 390–1
Becker, Gary S. 430 Council for Trade in Goods (CTG) 100, 101
beggar thy neighbour policies 2, 36, 53, 287 countervailing duties 201, 321, 365
Bhagwati, Jagdish 152 Cuba 330–1
Bloningen, Bruce 354 customs unions (CUs) 148, 159, 162–4, 167
Bogdandy, Armin von 26 customs valuation (CV) 118–19
Boltuck, Richard 341
border tax adjustment 213–14 Dam, Kenneth 382
Bown, Chad P. 174, 358 Davey, William 399
500
decision making
amendment of GATT 397–8
authentic interpretations 396–7
consensus 395–6
developing countries 450
special treatment for 41–2, 137–48
criticism of enabling clause 145–8
enabling clause 137–8, 142, 143–4
historical features 138–43
test for compliance with enabling
clause 144–5
directly competitive or substitutable (DCS)
products 217–33
definition 217–21
so as to afford protection (SATAP) 221–33
directly competitive or substitutable products
(DCS) 61
discriminatory pricing 338–43, 361–2
disguised restriction of trade 274–6
dispute settlement 21–2
early GATT years 398–400
World Trade Organization (WTO)
Appellate Body 411
burden of proof 407–8
discovery powers 408–10
DSU 402–4
exclusive forum for settling trade
disputes 404–6
recommendations and suggestions 416
request for establishment of panel 406
standard of review 410–11
types of complaints 411–16
Dixit, Avinash 251
Doha Round 23, 79–80
domestic laws, exception to national
treatment 264–6
domestic standards 296–7
dumping see antidumping measures
duties see tariffs
economic integration 148
economic theory, rationale for GATT
and 12–18
commitment theory 13–14
terms of trade theory 14–18, 19
effects doctrine 278–9, 449, 453
Ehlermann, Claus-Dieter 403, 410
Eichengreen, Barry 19
elasticity 220
pessimism 139
enforcement of WTO obligations 417–24
compensation 420–1
compliance panels 419–20
compliance within reasonable time
period 418
effectiveness 426–35
summary of process 417–18
suspension of concessions 421–4
Index
types of complaints and 418
equivalence 71–2
Ethier, Wilfred J. 16, 17, 121, 383, 427, 432
European Union 161, 176
farm subsidies 203–4, 205
preferences and
preferential trade agreements (PTAs) 149,
158, 172
special treatment for developing
countries 138–9, 146
tariffs 86
exchange restrictions 321, 335–8
exhaustible natural resources, exception to
national treatment 266–71
coverage 266–8
jurisdictional limits 268–9
measure relating to objective pursued 269–70
measures restricting domestic production or
consumption 270–1
expert review groups 408
export cartels 54
export subsidies 35, 41
counteracting prohibited subsidies 184–9
from discouraging to outlawing 181–4
rigidity required 179–81
export taxes 16–17, 84–5
externalities 14, 16, 17, 19
subsidies 180
extraterritoriality 277–85
farm subsidies 202–10
AG over SCM 208–10
regulation of farm trade in GATT
era 202–6
WTO Agriculture Agreement (AG) 206–8
Feddersen, Christoph 261, 276
fees and charges for services rendered 114–18
films, exception to national treatment 286
financial contribution 196–8
Finger, Michael 364
fiscal measures see taxation
foreign exchange contracts 119
Free Trade Area of the Americas (FTAA) 149
free trade areas (FTAs) 148, 159–62, 164
Friedman, Thomas 19
G 20 group 79, 80
G 90 group 80
General Agreement on Tariffs and Trade
(GATT) 1–2, 389–90
accession 391
acquis 27, 30
Contracting Parties 390–1
decision making
amendment of GATT 397–8
consensus 395–6
dispute settlement 21–2
early GATT years 398–400
 Index 501
late GATT years 400–2 international standards 289–95
genesis of 2–6 International Standards Organization
historical development 20–3, 25–30, (ISO) 291
448–51 international trade
legal relation of GATT with WTO commitment theory 13–14
Agreement 31–4 comparative advantage 12–13
negotiations 7–12 terms of trade theory 14–18, 19
non-application 394–5 International Trade Organization (ITO) 2–3,
reasons for existence 6–20 5, 6, 11, 446
negotiations 7–12 Preparatory Committee 3, 4
reason to ask the question 6–7 Irwin, Douglas 8, 9, 86
suggestions from economic theory 12–18
unexplored avenues 18–20 Jackson, John H. 5, 405
rounds of trade liberalization 23, 79–84 Josling, Timothy 204
rule of reason 451–3
as solid text with some compromises 447–8 Katz, M. L. 290
substantive content 34–8 Kennedy Round 24
WTO context 30–1 Keynes, John Maynard 9, 10
generalized system of preferences (GSP) 138, Krugman, Paul 9, 12, 15, 16, 150, 151, 163,
146 180
genetically modified organisms (GMOs) 61–2,
314 Lawrence, Robert 432
globalization 148–50 Lerner theorem 84
Gollier, C. 291, 311 licensing schemes, import licences 65–6
government procurement 210–12 like products
basic discipline 211 fiscal measures 234–6
definition 210 definition 234–6
past and present of GPA membership taxation in excess 236
211–12 non-fiscal measures 238–42
government trading see state trading enterprises Limão, Nuno 145–6, 432
(STEs) Litan, Robert E. 341
grandfathering 89, 123, 450 litigation, problems with 171–7
Grossman, Gene 146, 147, 351, 370 Lucas, Robert E. 63

Haberler Report 138, 139, 140 Manchin, Miriam 131

Harmonized System (HS) 72–4, 87–8 market access 385–7
Hart, Michael 8 Martin, Alberto 176
Hoekman, Bernard 146, 174 Mattoo, Aaditya 131
Horn, Henrik 229, 249, 383, 411 Mavroidis, Petros 229, 249, 279, 283, 347,
Howse, Robert 229, 354 370, 383, 411
Hudec, Robert 14, 120, 167, 204, 229, 399, MERCOSUR 364, 406
401, 402 Messerlin, Patrick 205
Hull, Cordell 8–9 Mexico, NAFTA and 150, 151
human life, exception to national Miller, James N. 9
treatment 262–3 mixed duties 79
monitoring compliance 424–5
import licences 65–6 most-favoured-nation (MFN) 8, 10, 16, 35,
import substitution 139 70, 80, 102–3
industrial policy 179–80 obligation 120–37
infant industry protection, quantitative extension to like products 126–9
restrictions/quotas (QRs) and 63 immediate and unconditional
information technology 82–4 extension 132–5
initial negotiating rights (INRs) 99–100 measures under purview of MFN
inspection, Pre-shipment Inspection 124–6
Agreement 119–20 no rebalancing permitted 135–6
internal measures, MFN and 125 rules of origin 129–32
International Monetary Fund (IMF) 2, 3, 118, standard of review 136–7
332, 335–8 summary of MFN discipline 123
502 Index
multilateral environmental agreements multilateral trade negotiations (rounds) 23,
(MEAs) 283–5 79–84
Mutual Recognition Agreements (MRAs) 178, renegotiating tariff protection 97–111
290 initial negotiating rights (INRs) 99–100
mechanics of negotiation 103–11
national security 37–8, 321 MFN trade as basis for defining PSIs and
legal discipline 322 SIs 102–3
practice in GATT years 322–30 new products 103
practice in WTO era 330–1 principal supplier interest (PSI)
quantitative restrictions/quotas (QRs) countries 100–2
and 63, 67 substantial interest (SI) countries 102
national treatment 192, 193–5 Neven, Damien 279
coverage of GATT Art III 212–15 Nogues, Julio 364
critique of case law 247–53, 276–86 non-fiscal measures, national treatment
exceptions 254–86 237–47
chapeau of Art XX GATT 271–6 affecting/offering for sale 238
coverage 254 less favourable treatment 242–7
critique of case law 276–86 like products 238–42
domestic laws and regulations 264–6 measures 237–8
exhaustible natural resources 266–71 non-tariff barriers (NTBs) 24, 386, 450
films 286 sanitary and phytosanitary measures 287–8,
human, animal and plant life 262–3 297–314
provisions 254–9 coverage of SPS Agreement 297
public morals 261–2 international standards 299
regulatory diversity 259–60 legal relationship between GATT, TBT
two-tier test for compliance with Art XX and SPS 288
GATT 260–1 measures adopted on basis of
extraterritoriality 277–85 precaution 309–14
fiscal measures 216–36 summary of SPS Agreement 298–9
directly competitive or substitutable unilateral measures based on scientific
(DCS) products 217–33 evidence 299–309
like products 234–6 technical barriers 287–97
so as to afford protection (SATAP) coverage of TBT Agreement 288–9
221–33 domestic standards 296–7
government procurement 210–12 legal relationship between GATT, TBT
basic discipline 211 and SPS 288
definition 210 unilateral technical regulations 295–6
past and present of GPA non-violation complaints (NVCs) 411,
membership 211–12 412–15
negative integration and 215 Nordström, Häkan 174–5
non-fiscal measures 237–47 North American Free Trade Agreement
affecting/offering for sale 238 (NAFTA) 149, 150, 151, 157–8, 170,
less favourable treatment 242–7 171, 405–6
like products 238–42
measures 237–8 opportunism 215, 287
subsidies 195–210 ordinary customs duties 74–5
classification 200–1 typology 78–9
counteracting 201 Organization for Economic Co-operation and
definition 195–200 Development (OECD) 79
farm subsidies 202–10 other duties and charges (ODC) 74, 75–8
remedies 201–2 Özden, Czaglar 147
wide coverage 195
natural resources see exhaustible natural pacta sunt servanda principle 80
resources Palmeter, David 283
negative integration 215 Panagariya, Arvid 152
negotiations Park, Jee Hyeong 354
on GATT 7–12 Pauwelyn, Joost 432
 Index 503
peace clause 210, 379 non-segmentation nature of Art XI GATT 68
Permanent Group of Experts (PGE) 184 textiles 67
Petersmann, Ernst-Ulrich 240, 358, 413, 414,
422 reason, rule of 451–3
plant life, exception to national reciprocity 11, 12, 16, 71, 98
treatment 262–3 Regan, Donald H. 14, 16, 18, 229
Prebisch, Raul 139 regulations
precautionary principle 309–14 exception to national treatment 264–6
precedent 29–30, 425–6 regulatory diversity 259–60
predation 362–3 technical see technical barriers to trade
preferences 9, 35 see also technical barriers to trade
developing countries 41–2 Reinhardt, Eric 147
erosion 80, 145–6 resources see exhaustible natural resources
generalized system of preferences Ricardo, David 13
(GSP) 138, 146 Roessler, Frieder 153, 168
preferential trade agreements (PTAs) 41, rule of reason 451–3
148–79, 451 rules and formalities, MFN and 124–5
bilateral track 168–71 rules of origin 129–32, 159–60
consistency with GATT 152–3 Russia/Soviet Union 19
design and approval 153–6
globalization and 148–50 Sachs, Jeffrey D. 147
important matters when discussing safeguards 321
PTAs 150–2 AD as safeguard measure 364–5
lack of litigation 171–7 compensation 376–8
multilateral track 156–68 criticisms 381–5
permissible deviation from MFN 177–9 duration 378–80
safeguards imposed by 367–8 imposed by PTAs 367–8
rules of origin 131–2, 159–60 increased imports 371
special and differential treatment for initiation of investigation 368–9
developing countries 41–2, 137–48 injury 371–5
criticism of enabling clause 145–8 non-discriminatory 375–6
enabling clause 137–8, 142, 143–4 regulatory framework 365–6
historical features 138–43 typology 366
test for compliance with enabling unforeseen developments 370–1
clause 144–5 VERs and 380–1
principal supplier interest (PSI) 100–2 sanitary and phytosanitary measures 287–8,
principal supplier rule 12 297–314
prisoner’s dilemma 15 coverage of SPS Agreement 297
protectionism 14, 36, 61 international standards 299
see also infant industry protection legal relationship between GATT, TBT and
public morals exception to national SPS 288
treatment 261–2 summary of SPS Agreement 298–9
public sector see government procurement; Sapir, André 148, 347, 351
state trading enterprises (STEs) schedules
certification of 86
quantitative restrictions/quotas (QRs) 9, 35, 41 concessions 88–92
definition and nature 43–66 rectifications and modifications 112–14
attribution of practices to Schelling, Thomas 428
governments 45–50 Schott, Jeffrey 167, 174
measures covered 50–5 Schropp, Simon 432
relationship between Art XI and other Schwarz, Warren F. 120, 430, 431
GATT provisions 58–66 Serra, Jaime 160
standard of review 55–8 Shaggi, Kamal 432432
difference in treatment compared with Shapiro, C. 290
tariffs 71–2 Singer, Hans 139
discriminatory legal QRs 67 single undertaking approach 24
legal text on 42–3 situation complaints 411, 415–16
504
Smith, Adam 12
social policies 214–15
specific duties 79
Spencer, Barbara 180
spill-overs 19
Srinivasan, T. N. 13
Staiger, Robert W. 15, 16, 17, 18, 180, 354
standard of review
MFN obligation 136–7
national treatment 285–6
transparency obligation 442–3
WTO panels 410–11
standards see technical barriers to trade
stare decisis (precedent) 29–30, 425–6
state contingencies
antidumping measures 18, 142, 321
AD as safeguard measure 364–5
discriminatory pricing 338–43, 361–2
establishing injury 349–52
finding dumping 343–9
predation 362–3
price differentiation in antitrust
statutes 358–61
strategic dumping 363–4
sunset reviews 352–8
balance of payments (BoP) 9, 60, 321,
331–4
quantitative restrictions/quotas (QRs)
and 67
countervailing duties 201, 321, 365
exchange restrictions 321, 335–8
market access 385–7
national security 37–8, 321
legal discipline 322
practice in GATT years 322–30
practice in WTO era 330–1
quantitative restrictions/quotas (QRs)
and 63, 67
safeguards 321
AD as safeguard measure 364–5
compensation 376–8
criticisms 381–5
duration 378–80
imposed by PTAs 367–8
increased imports 371
initiation of investigation 368–9
injury 371–5
non-discriminatory 375–6
regulatory framework 365–6
typology 366
unforeseen developments 370–1
VERs and 380–1
state trading enterprises (STEs) 9
definitions 315–16
legal obligations 316–19
Stevenson, Adlai 2
strategic dumping 363–4
Index
subsidies 195–210
classification 200–1
counteracting 201
definition 195–200
benefit to recipient 198–9
financial contribution 196–8
specificity requirement 200
export see export subsidies
farm subsidies 202–10
AG over SCM 208–10
regulation of farm trade in GATT
era 202–6
WTO Agriculture Agreement
(AG) 206–8
remedies 201–2
substantial interest (SI) 102
sunset reviews 352–8
suspension of concessions 421–4
Sykes, Alan O. 120, 146, 147, 180, 285, 381,
384, 430, 431
tariffs 8, 9, 10, 34, 36, 37, 41, 385–6, 450
basic discipline with regard to tariff
protection 70–1
changes in tariff protection after
consolidation 93–114
rectifications and modifications of
schedules 112–14
renegotiating tariff protection 97–111
switching between different types of
duties 94–5
withdrawing concessions from members
leaving WTO 95–7
customs valuation (CV) 118–19
difference in treatment compared with
quantitative restrictions 71–2
fees and charges for services rendered
114–18
Harmonized System (HS) 72–4, 87–8
interpreting tariff commitments 87–93
legal relevance of HS 87–8
(non) impact of legitimate
expectations 93
scheduling of concessions must be WTO
consistent 88–92
legal text 68–71
mechanics of binding 72–86
certification of schedules 86
Harmonized System (HS) 72–4
multilateral trade negotiations
(rounds) 79–84
tariff promise 74–8
treatment of export taxes 84–5
typology of ordinary customs duties
78–9
MFN obligation 120–37
extension to like products 126–9
 Index
immediate and unconditional
extension 132–5
measures under purview of MFN 124–6
no rebalancing permitted 135–6
rules of origin 129–32
standard of review 136–7
summary of MFN discipline 123
preferential trade agreements (PTAs)
148–79
bilateral track 168–71
consistency with GATT 152–3
design and approval 153–6
globalization and 148–50
important matters when discussing
PTAs 150–2
lack of litigation 171–7
multilateral track 156–68
permissible deviation from MFN 177–9
Pre-shipment Inspection Agreement
119–20
special and differential treatment for
developing countries 137–48
criticism of enabling clause 145–8
enabling clause 137–8, 142, 143–4
historical features 138–43
test for compliance with enabling
clause 144–5
terms of trade theory and 15
taxation 36, 60
border tax adjustment 213–14
export taxes 16–17, 84–5
national treatment 216–36
directly competitive or substitutable
(DCS) products 217–33
like products 234–6
so as to afford protection (SATAP)
221–33
see also tariffs
technical barriers to trade 287–97
coverage of TBT Agreement 288–9
domestic standards 296–7
international standards 289–95
legal relationship between GATT, TBT and
SPS 288
unilateral technical regulations 295–6
technical duties 79
terms of trade theory 14–18, 19
textiles, quantitative restrictions/quotas
(QRs) 67
Tokyo Round 21
Trachtman, Joel 283, 351
trade see international trade
Trade Policy Review Mechanism
(TPRM) 436, 437–8
Trade Promotion Authority (TPA) 80
transaction costs 72
transparency obligation 436–46
505
coverage 441–2
laws of general application 443
obligation to maintain independent
tribunals 445–6
standard of review 442–3
summary 440–1
Trade Policy Review Mechanism
(TPRM) 436, 437–8
uniform, reasonable and impartial
administration 443–5
unemployment 9, 10, 11
unilateral measures
sanitary and phytosanitary measures
299–309
technical regulations 295–6
United Kingdom, negotiations on GATT
and 8, 9–10
United Nations 2, 3, 4
United States of America
farm subsidies 203
fast track procedure 80
negotiations on GATT and 8–9, 10
preferences and, preferential trade
agreements (PTAs) 149
special and differential treatment for
developing countries 146, 147
tariffs 86
Uruguay Round 24
valuation, customs valuation (CV) 118–19
van den Broek, Naboth 432
Vergotte, Wouter 176
Vermulst, Edwin 131, 354
Viner, Jacob 120, 150
violation complaints 411, 412
Vogel, David 290
voluntary export restraints (VERs), safeguards
and 380–1
Waer, Paul 354
waiver of obligations 433–4, 435–6
Warner, Andrew 147
Wauters, Jasper 358
welfare, preferential trade agreements (PTAs)
and 150–1
World Bank 2, 3
World Customs Organization (WCO) 73
World Trade Organization (WTO) 2, 25–34,
389, 390
accession 391–4
decision making
authentic interpretations 396–7
consensus 396
dispute settlement
Appellate Body 411
burden of proof 407–8
506
World Trade Organization (cont.)
discovery powers 408–10
DSU 402–4
exclusive forum for settling trade
disputes 404–6
recommendations and suggestions 416
request for establishment of panel 406
standard of review 410–11
types of complaints 411–16
enforcement of WTO obligations
417–24
compensation 420–1
compliance panels 419–20
compliance within reasonable time
period 418
effectiveness 426–35
Index
summary of process 417–18
suspension of concessions 421–4
types of complaints and 418
GATT in WTO context 30–1
legal relation of GATT with WTO
Agreement 31–4
monitoring compliance 424–5
precedent 29–30, 425–6
Textiles Monitoring Body (TMB) 67
waiver of obligations 433–4, 435–6
withdrawing concessions from members
leaving WTO 95–7
Yugoslavia 97, 394
Zanardi, Maurizio 176