Gorton, Pennacchi, Gilbert Eds - The Changing Market in Financial Services

The Changing Market in Financial
Services
The Changing Market in Financial
Services
Proceedings of the Fifteenth Annual Economic Policy
Conference of the Federal Reserve 8ank of St. Louis.
edited by
R. Alton Gilbert
The Federal Reserve Bank of St. Louis
"
~.
Springer Science+Business Media, LLC

Library 01 Congress Cataloging-in-Publication Data
Economic Policy Conference of the Federal Reserve Bank of

SI. Louis (15th: 1992: Federal Reserve Bank of SI. Louis)
The changing market in financial services: proceedings of the
Fifteenth Annual Economic Policy Conference of the Federal
Reserve Bank of SI. Louis/edited by R. Alton Gilberl.
p. cm.
Includes bibliographical relerences.
ISBN 978-94-010-5322-8 ISBN 978-94-011-2976-3 (eBook)
DOI 10.1007/978-94-011-2976-3
1. Financial services industry-United States-Congresses.
1. Gilbert, R. Alton. II. Federal Reserve Bank of SI. Lauis. III. Title.
HG1818.E36 1992
332.1 '0973-dc20 91-40394
CIP
Copyright © 1992 by SpringerScience+Business Media NewYark

Originally published by Kluwer Academic Publishers in 1992
Softcover reprint of the hardcover 1st edition 1992
AII rights reserved. No part of this publication may be reproduced,

stored in a retrieval system or transmitted in any form or by any
means, mechanical, photo-copying, recording, or otherwise, without
the prior written permission of the publisher, Springer Science+Business
Media, LLC.
Printed an acid-tree paper.

Contents
Contributing Authors vii
Preface ix
1
The Opening of New Markets for Bank Assets 3
Gary Gorton and George G. Pennacchi
Commentary by Stuart I. Greenbaum 35
II 39
2
Interstate Banking, Bank Expansion and Valuation 41
Gerald A. Hanweck
Commentary by Peter S. Rose 93
3
The Market for Home Mortgage Credit: Recent Changes and Future
Prospects 99
Patrie H. Hendershott
Commentary by Herbert M. Kaufman 125
4
Equity Underwriting Risk 129
J. Nellie Liang and James M. O'Brien
v
VI THE CHANGING MARKET IN FINANCIAL SERVICES
III 159
5
The Competitive Impact of Foreign Commercial Banks in the United
States 161
Lawrence G. Goldberg
Commentary by Gary C. Zimmerman 201
6
The Competitive Impact of Foreign Underwriters in the United States 211
Robert Nachtmann and Frederick J. Phillips-Patrick
Commentary by Samuel L. Hayes 241
Index 247
Contributing Authors
Lawrence G. Goldberg Patrie H. Hendershott

Department of Finance Academic Faculty of Finance
P.O. Box 248094 Ohio State University
University of Miami 1775 College Road
Coral Gables, Florida 33124 Columbus, Ohio 43210-1309
Gary Gorton
The Wharton School Herbert M. Kaufman
University of Pennsylvania Financial Systems Research
Philadelphia, Pennsylvania 19104 Arizona State University
Bac 519
Stuart I. Greenbaum Tempe, Arizona 85287
J.L. Kellogg Graduate School of
Management
Northwestern University J. Nellie Liang
Leverone Hall Division of Research and Statistics
2001 Sheridan Road Board of Governors of the Federal
Evanston, Illinois 60208 Reserve System
Washington, D.C. 20551
Gerald A. Hanweck
Department of Finance
School of Business Administration Robert Nachtmann
George Mason University KATZ Graduate School of Business
4400 University Drive University of Pittsburgh
Fairfax, Virginia 22030 352 Mervis Hall
Pittsburgh, Pennsylvania 15260
Samuel L. Hayes III
Graduate School of Business
Administration James M. O'Brien
Harvard University Division of Research and Statistics
Baker 333 Board of Governors of the Federal
Soldier Field Reserve System
Boston, Massachusetts 02163 Washington, D.C. 20551
vii
viii THE CHANGING MARKET IN FINANCIAL SERVICES
George G. Pennacchi Peter S. Rose

University of Illinois Finance Department
304G David Kinley Hall Blocker Building, Room 340
1206 South Sixth Street Texas A&M University
Champaign, Illinois 61801 College Station, Texas 77843
Gary C. Zimmerman
Frederick J. Phillips-Patrick Economic Research Department
Office of Economic Research Federal Reserve Bank of San
Office of Thrift Supervision Francisco
1700 G Street NW. P.O. Box 7702
Washington, D.C. 20552 San Francisco, California 94105
Preface
The articles and commentaries included in The Changing Market in

Financial Services were presented at the Fifteenth Annual Economic Policy
Conference of the Federal Reserve Bank of St. Louis. The conference
focused on the effects of a variety of recent changes in the market for
financial services in the United States. This market has been changing
rapidly in recent years: Business loans have become more liquid, as the
market for loan sales grows; banks have been permitted to participate in
a limited form of interstate banking; commercial banks have been given
permission to offer additional underwriting services; the market for resi-
dential mortgage credit has been transformed through securitization and
the declining role of savings and loan associations; and foreign financial
firms have taken a rising share of the market financial services. The art-
icles describe these changes and examine their implications for financial
institutions and their customers.
The article by Gary Gorton and George Pennacchi examines why loans
sold by commercial banks in recent years have grown rapidly. In 1989,
about $240 million of commercial and industrial loans were sold, compared
to insignificant amounts five years earlier. This observation raises ques-
tions about the nature of relationships among banks, their customers, and
those who buy the loans. The theory of financial intermediation devel-
oped in recent years is based on an information asymmetry between banks
and other investors: By providing payments services, banks gain valuable
information about the financial condition of their depositors that is not
available to other investors. If banks have special information, why would
others buy loans from them, since banks are likely to sell only the poorest
quality loans? Gorton and Pennacchi explore the empirical evidence on
several competing hypotheses about why investors would buy loans from
ix
x THE CHANGING MARKET IN FINANCIAL SERVICES
commercial banks. The authors conclude that technological progress has

drastically reduced the information asymmetry between banks and many
other investors, which is the basis for much of the theory of financial
intermediation and for bank regulation.
Stuart I. Greenbaum suggests in his commentary that Gorton and
Pennacchi may be exaggerating the role of technological change in ex-
plaining the growth of loan sales. Greenbaum points out that other fac-
tors reduce the capacity of major U.S. banks to attract deposits in relation
to their capacity to originate loans, including the downgrading of bank
debt by rating agencies. He suggests that U.S. banks, faced with such
changes, made loan purchases more attractive to potential investors,
compensating them for the moral hazard risk which continues to be
important.
Gerald A. Hanweck analyzes the impact of regional interstate banking
on banking competition and bank performance. Hanweck finds little evi-
dence that regional interstate banking has stimulated competition among
banks or increased their operating efficiency, profitability, or safety and
soundness. The primary impact has been a substantial consolidation
of banking resources within the largest fifty or one hundred bank-
ing companies.
One important reason for the increase in banking concentration at the
national level has been that states currently do not allow banks in other
states to enter by establishing new offices; instead, they must enter by
acquiring existing banks. Interstate banking would induce greater com-
petition if banks could expand their operations across state lines by estab-
lishing new offices.
Peter S. Rose shares Gerald Hanweck's concern with rising banking
concentration at the national level, but raises questions about the nature
of the concentration data and the reasons for the rise in concentration.
Hanweck's data are for domestic commercial banking firms only; they do
not reflect the impact of the growth in U.S. banking assets of foreign
banks or nonbank financial firms. The rise in concentration also reflects,
to a large extent, changes in state laws to permit intrastate branching.
Rose also cites the results of his own work as providing one explanation
for the lack of positive stock price response to the announcements of
plans for interstate mergers reported in several studies. Rose finds that
both the acquiring firms, and those to be acquired in interstate mergers,
tend to underperform other banks.
Patric H. Hendershott takes a look at changes in the market for home
mortgage credit in the 1980s. One of the major developments in this
market has been the securitization of fixed-rate mortgages. Hendershott
PREFACE xi
reports evidence that securitization has affected both the level and pat-
tern of fixed-rate mortgage interest rates, making these rates move more
closely with other long-term rates. Another major change has been the
growth of adjustable-rate mortgages. The continued growth of these
mortgages is likely to be limited by the extent to which these loans are
securitized. A third major development in the 1980s was the rapidly
declining role of savings and loan associations as providers of residential
mortgage credit. Hendershott concludes that the past shrinkage in the
share of mortgage credit provided by savings and loan associations does
not appear to have raised home mortgage rates, nor does he believe that
the future shrinkage will affect mortgage rates.
Herbert M. Kaufman focuses on some of the policy issues raised by
Hendershott's article. One such issue is the status of government agencies
that have developed the market for mortgage-backed securities. Kaufman
concludes that the integration of the mortgage market with other capital
markets is now so well developed that it is time to eliminate the subsidies
to these government-sponsored agencies. The other policy issue that
Kaufman addresses is the rationale for specialized mortgage lenders.
Kaufman argues that because of the integration of the mortgage and
capital markets, reductions in the share of mortgage credit have not raised
mortgage interest rates. This integration is so complete that the rapid
reduction in credit provided by a sector that was important in the past
does not affect market rates.
J. Nellie Liang and James M. O'Brien estimate the distribution on
returns to underwriters on individual issues of corporate equity securities.
Their estimates indicate that the frequency of losses is small, with the
underwriting spread more than covering declines in share prices around
the time new issues come to market. The variability of equity underwriting
returns primarily reflects a variability in the number of offerings, not in
the returns on individual issues.
Liang and O'Brien conclude that large commercial banks are relatively
well positioned to gain entry into corporate securities underwriting, and
that, once established, the variability of returns would primarily reflect
the variability in the frequency of underwriting.
Lawrence G. Goldberg describes the growth of foreign banks in the
United States and summarizes the results of studies of the motivations for
this growth. Goldberg concludes that several of the determinants of foreign
bank presence indicate continuing future growth of foreign banks in this
country. He finds no support for arguments that foreign banks have com-
petitive advantages over domestic banks.
In his comment on Goldberg's paper, Gary C. Zimmerman points out
xii THE CHANGING MARKET IN FINANCIAL SERVICES
that all of the recent foreign bank expansion in the United States has
been by Japanese owned banks. He argues that any analysis of growth
in foreign banks in the United States should focus on the growth of
Japanese banks and the reasons behind that growth.
Zimmerman provides additional perspective on the issue of competi-
tive advantage of foreign banks. Some have argued that foreign banks
have lower cost of funds than domestic banks and that foreign banks are
able to exploit this advantage by lending in the United States at rates
below those offered by domestic banks. Zimmerman's analysis indicates
that, rather than using relatively low cost funds raised in Japan to lend in
the United States, the Japanese banks, on net, are raising funds in the
United States and lending them in Japan.
Robert Nachtmann and Frederick J. Phillips-Patrick study the market
shares of domestic and international securities firms in underwriting
corporate securities. They characterize the domestic underwriting business
as very competitive, with falling underwriting spreads throughout most of
the decade and entry by foreign securities firms providing competitive
pressure. The share of securities underwritten in the United States by
foreign securities firms had risen in the 1980s, but the share of securities
underwritten in other countries by firms headquartered in the United
States had also risen. On net, the global market share of firms headquartered
in the United States had remained relatively constant in the 1980s. These
changes in shares reflect the integration of world financial markets with
no evidence of net disadvantage for domestic securities firms. Statistical
tests indicate no significantly adverse effects of the entry by foreign securities
firms into this field on the profits or stock prices of domestic securities
firms.
Samuel L. Hayes, in his comments on the article by Nachtmann and
Phillips-Patrick, focuses on the composition of equity underwriting that is
subject to foreign competition. Hayes assumes that, in the United States
as in other countries, equity underwriting is done primarily by domestic
underwriters. Thus, Hayes assumes that securities firms primarily underwrite
debt issues outside their home markets. Hayes expresses doubts about the
ability of empirical tests to indicate the competitive effects of entry by
foreign securities on domestic firms.
R. Alton Gilbert
Assistant Vice President
I
1 THE OPENING OF NEW
MARKETS FOR BANK ASSETS
Gary Gorton
George G. Pennacchi
Introduction
When a corporation seeks to raise resources, it can issue securities such

as bonds or equities, or alternatively, it can borrow from a bank. These
two methods of raising capital are not identical. Empirical evidence sug-
gests that a bank loan is significantly different from a marketable secur-
ity.l A key implication of theories explaining the uniqueness of bank
lending is that bank loans should not be resold once originated by banks.
Bank loans are assumed to be difficult to value by investors, and there-
fore, should be nonmarketable. The illiquidity of bank loans underlies the
rationale for bank regulation and deposit insurance because, historically,
the absence of information-revealing markets for these securities led
to banking panics. 2 In fact, one might say that if banks were able to sell
loans, it would be the end of banking. (Possibly bank regulation would
also end.)
If banking is not coming to an end, at least it is changing significantly.
In the last few years many banks have been able to sell substantial amounts
of loans. The growth of markets for bank assets has been remarkable.
3
4 THE CHANGING MARKET IN FINANCIAL SERVICES
In a few short years the outstanding principal of repackaged automobile

loans and credit card receivables sold has grown from nothing to about
$50 billion. At the same time, the volume of commercial and industrial
loan sales has grown from insignificant amounts to $240 million in 1989.
Growth in volume has been accompanied by a deepening of the new
markets as maturities of securitization contracts have lengthened, an
increase in the average risk of the underlying loans, and the beginnings
of secondary markets for the claims on loans.
However, the ratio of loans originated to loans sold varies significantly
by loan type, suggesting that certain kinds of loans are more attractive
candidates for resale. There also appears to be important differences in
the types of contracts used to sell loans, as well as differences in the types
of investors who buy them. Loans such as mortgages and consumer re-
ceivables are typically pooled and then sold as mortgage- or asset-backed
securities.3 Asset-backed securities typically represent senior claims on a
pool of homogeneous consumer bank loans such as automobile loans or
credit card receivables. These pools of loans are "credit-enhanced" by
being guaranteed or insured by a third party guarantor and almost always
receive triple A ratings. The primary buyers include professional money
managers, pension funds, bank trust departments, and wealthy individuals.
In contrast to mortgages and consumer loans, commercial and indus-
trial loans are almost never pooled when sold. These loan sales, or sec-
ondary loan participations, are proportional (equity) claims to the cash
flow from a single commercial or industrial loan. No recourse, insurance,
guarantee, or credit enhancement is included. There is usually no rating.
The single loan underlying a loan sale is often the unsecured obligation
of a non-investment grade firm. The typical buyers of commercial and
industrial loans are foreign banks and smaller domestic banks, although
recently, about one quarter of the buyers have been nonfinancial firms.
Banking regulations require that for loan selling banks to remove loans
from their balance sheet, so that the proceeds obtained from loan buyers
are not classified as deposits subject to required reserves and required
capital, the loans that are sold must be sold without recourse to the selling
bank.4 Hence, if a bank is motivated to sell loans as a way of avoiding the
financing costs associated with required reserves and capital, it must not
provide any guarantee of a loan's quality to either a third party buyer (in
the case of commercial and industrial loan sales) or a third party guaran-
tor (in the case of asset-backed securities). Since nearly all loans sales are,
in fact, sold without recourse to the originating bank, they seem to contradict
the nonmarketability of bank loans. 5
Recent information-based theories of financial intermediation argue
THE OPENING OF NEW MARKETS FOR BANK ASSETS 5
that banks provide special services that cannot be obtained when securities
are offered to the public. One service deals with the production of infor-
mation regarding the credit risk of a potential borrower. Another service
refers to the monitoring of borrowers so that loan covenants can be enforced
while loans are outstanding. Individual investors may have a very difficult
time undertaking these activities because of "free riding" by investors.
Free riding prevents individual investors from recovering the costs of
producing information or monitoring. Realizing that this problem exists,
no individual would be willing to pay the costs. Even if individuals were
willing to bear the costs, a superior arrangement would be to delegate the
tasks to another agent, namely, a bank, rather than duplicate each others'
costly efforts. 6
The difficulty with delegating these services to banks is that depositors
are unlikely to have the ability to directly observe banks' performance. In
other words, an information asymmetry exists between banks and depositors
(and between banks and other outsiders such as potential loan buyers).
In such an environment what guarantees that banks will actually choose
to produce information and monitor borrowers? Incentive compatibility
between the bank (equity holders) and the depositors requires that the
bank be a residual claimant on the loan's return, such as the case when
the loan is held on the bank's balance sheet or sold with recourse or a
guarantee. This insures that the bank equity holders will see to it that the
bank will, in fact, perform these services (since otherwise, they will be the
first to suffer losses).7 If loans were sold without recourse, the bank would
have no incentive to perform these services, that is, equity holders would
no longer be at risk for nonperformance. These loans would decline in
value because potential loan buyers would realize this incentive problem
exists and offer less to buy the loans. This would have the effect of inducing
banks to hold the loans rather than sell them. Thus, the existence of
banks is synonymous with the creation of assets which must be held until
maturity. This is the basis for the term intermediary.
The analysis equates the existence of banks with the nonmarketability
of their loans. But, this argument appears to be directly contradicted by
the recently observed experience of banks selling loans without recourse
or guarantee. The argument is also contradicted by the willingness of
third parties to credit enhance asset-backed securities. Apparently, banks
can be induced, at least partially, to perform the services associated with
loans without being required to hold a claim on the loans until maturity.
What is this new mechanism for insuring incentive compatibility? In other
words, why are third party buyers, or third party guarantors in the case
of asset-backed securities, willing to rely on the bank to continue to perform
on the original loan contract after the loan has been sold? This question
is the focus of this article.
A number of hypotheses may explain how markets for bank assets
could develop.8 This paper considers the development of the market in
which banks sell their loans. We explore three hypotheses about mechanisms
that could make loan sales incentive compatible. First, there is the possi-
bility that the bank offers an implicit (unobservable) guarantee or insur-
ance on the value of the loan. Perhaps the bank is willing to buy the loan
back, bearing the losses, if the credit risk of the underlying borrower
increases. Such a promise would have to be implicit since an explicit
promise would not allow the bank to remove the loan from its balance
sheet. Since this arrangement would be implicit, it would likely be hard
to detect though Gorton and Pennacchi (1989) claim to have found some
evidence for this contract feature.
A second possible mechanism for making loan sales incentive compat-
ible involves the selling bank holding a fraction of each loan sold. If the
selling bank sells only a part of the loan, retaining the rest, then the bank
will continue to have at least a partial incentive to perform on the loan.
Either of these two possibilities, or the two features combined, could
explain the apparent paradox of loan sales. However, it is not clear, as we
shall see, how these two contract features could be enforced. Nothing in
written loan sales contracts, at least for commercial and industrial loan
sales, requires selling banks to hold a fraction of the loan, and nothing
requires the bank to guarantee the value of loans sold. If these implicit
contracts exist, then market forces (as opposed to legal forces) must enforce
them.
There is a third possibility that could help explain the feasibility of
loan sales. It may be that there has been a significant reduction of the
agency problem between loan buyers or third party loan guarantors and
loan selling banks. Perhaps, in contradiction of the assumptions of aca-
demic researchers and the implicit assumptions of policymakers, loan buyers
or guarantors can observe whether or not banks produce information and
monitoring. Reductions in the cost of information production and trans-
mission would be at the root of such a change. If the banks' actions,
information production and monitoring, are observable (unlike in the
past), then it may be possible to enforce bank performance without
directly requiring them to risk their equity. Determination of the cause of
the opening of bank asset markets is a necessary part of addressing the
question of whether there are limits to the securitization trend.
A number of articles have addressed banks' motivation for loan sales
and have pointed out the incentive compatibility problems associated with
selling bank assets;9 however, there has been little empirical work inves-
tigating the possibly implicit contractual features that banks might use to
alleviate these incentive problems. One reason for this is a lack of data
to analyze. In this article we review and summarize our attempts to dis-
criminate between the above hypotheses regarding the mechanisms whereby
loan sales could be feasible. While the focus of the article is on loan sales,
the same issues arise for asset-backed securities. So, we begin in the second
section by briefly explaining asset-backed security contracts and by re-
viewing recent developments in asset-backed security markets. Section
three is devoted to briefly tracing and explaining the growth of the (in-
dividual) loan sales market. Section four begins the discussion of the
relative merits of the first two hypotheses explaining the existence of
these markets. The first tests aimed at testing these hypotheses are reviewed.
Section five introduces a simple model that provides the basis for more
sophisticated tests. We then review some earlier work that has tested the
various hypotheses. In Section six we analyze the hypothesis that tech-
nological change has resulted in a significant reduction of the information
asymmetry between banks and outsiders. The opening of new markets for
bank assets has implications for the regulation of banks. These are briefly
discussed in the concluding comments of Section seven.
Asset-backed Securities
An asset-backed security is a claim on a portfolio of underlying bank

loans that have some common features, for example, the underlying loans
may all be automobile loans or credit card accounts. In fact, car loans and
credit card receivables are the two leading examples of loan types used to
construct asset-backed securities.lO The underlying loans have the feature
that they are fairly standardized loan contracts and have well known risk-
return characteristics (including historic default experience and prepayment
propensities). Asset-backed securities, as with mortgage-backed securities,
are generally issued in one of two forms, either pass-through or pay-
through. A pass-through structure allows for a sale of the receivables or
loans to the investors, while a pay-through security is a debt of the issuer
backed by the receivables.
Excluding mortgage-backed securities, the total issuance of asset-backed
securities during the years 1985 to 1988 was $36.2 billion. Of this, $23.1
billion (64 percent) was automobile loan-backed securities, $9.8 billion
(27 percent) was credit card receivable-backed securities and the remainder
of $3.3 billion (9 percent) was backed by other types of loans. In 1989 the
Table 1-1. Asset-backed Securities Volume (Issued Annually).

Total Car Loan Credit Card Total Number
Year Amount* ($ bit.) Backed Securities Backed Securities of Issues
1985 1.237 0.9 0.0 7
1986 10.041 9.9 0.0 16
1987 9.134 6.7 2.4 31
1988 14.939 5.5 7.4 59
1989 9.54** NA NA 27
* Principal Amount
** As of August 30th.
NA: Not Available
Source: Goldman, Sachs & Co.
Table 1-2. Asset-backed Securities Issuance by Type of Issuer (1988) ($ billions).

Commercial Dealer Manufacturing
Backing Loans Banks Thrifts Affiliates Subsidiaries Retailers
Car Loans 3.0 1.3 0.1 1.2 0.0
Credit Card
Receivables 5.3 0.0 0.0 0.0 2.1
Other Loans 0.6 0.5 0.0 0.3 0.0
Source: Goldman, Sachs & Co.
total issuance was over $50 billion. Table 1-1 provides the data for the
recent history of the asset-backed market. Table 1-2 provides information
about the types of issuers of these securities.
One important feature of asset-backed securities, mentioned above,
is the presence of credit enhancement devices in the contract. Credit
enhancement is provided by a third party credit facility (such as a letter
of credit or surety policy) as well as over-collateralization. Banks must rely
on third party credit enhancement (rather than issuing their own guarantees
as is sometimes the case with non-bank issuers) because a guarantee by
the issuing bank would not allow the deal to qualify as a sale of assets
according to Regulatory Accounting Principles (RAP). Credit protection
provided by a third party usually covers all losses up to a fixed percentage
of the principal. This percentage is usually significantly in excess of the
loans' historic default rates. In some issues of asset-backed securities, a
straightforward recourse provision is included under which the third party
guarantor looks to the issuer to back up the guarantee. Banks, however,

cannot use such a mechanism since it is inconsistent with the regulatory
definition of an asset sale; hence, a potential incentive compatibility problem
exists between the issuing bank and the third party guarantor. However,
this problem is likely to be minimized due to the ability to objectively
classify the credit risk of these rather homogeneous loans. Initial agree-
ments can be made such that loans added to the pool are required to
surpass a certain quality based on analysis from credit scoring models.
If loans are later found to have not met these initial requirements, the
issuing bank can be forced to buy back these below standard loans.
Another important feature of asset-backed securities is the existence
of secondary markets. These markets appear to be increasingly liquid.
Automobile Receivable-backed Securities
The automobile receivable-backed securities market began in May 1985

with a $60 million offering by Marine Midland Bank. Since then, the
market has grown rapidly. In the period from 1985 to 1987 there were 27
offerings. Importantly, 19 of these offerings were by financial subsidiaries
of major automobile companies and not by commercial banks. In particu-
lar, General Motors Acceptance Corporation had 16 offerings during this
period. During 1988 there were 27 new offerings. In that year commercial
banks were the major issuers. The largest of these were by Citicorp which
(through its subsidiary, Citicorp Acceptance Company, Inc.) issued two
automobile pass-through securities totaling $1.078 billion. ll
For automobile-backed securities, the pass-through, utilizing a grantor
trust, is the most common format because it achieves sales treatment for
generally accepted accounting practice (GAAP), regulatory accounting
practice (RAP), tax, and legal purposes. In a typical pass-through, auto-
mobile receivables are sold to the grantor trust, which then issues certifi-
cates representing undivided interests in the trust assets. The simplest
pass-through structure has certificates paying a lower coupon than the
rate paid on the receivables. Principal and interest are passed through
monthly as received (with a lag of 24 days). The spread between the
average loan rate and the certificate coupon rate is used to pay servicing
fees, guarantee fees, and other expenses. Sometimes the spread is used to
support the credit enhancer. The trust is administered by an independent
trustee.
Recent innovations in automobile-backed securities include the use of
a senior/subordinated structure and reductions in the extent of the credit
enhancement. These developments seem consistent with a growing maturity

of this market since these changes parallel the previous transformation of
the mortgage-backed securities market.
Credit Card Receivables
Credit card receivable-backed securities are claims on a designated (and

selected) pool of outstanding credit card accounts. The cash flows from
these accounts consist of principal and interest. Each month's activity in
each account can vary, so that the principal and interest cash flow from
the total pool will also change. Investors purchase a fixed amount of
principal with a fixed coupon rate. Variations in the principal amounts are
handled via a two-class payment structure: an investor class and a seller
class. The investor class has a fixed principal amount, usually representing
60 percent to 80 percent of the original pool principal. Interest is paid
monthly at a fixed coupon rate. The seller class must absorb variations
in the size of pool balances. 12 The seller also receives any increases in
receivable balances as well. However, the seller class is not subordinate,
since all the risks of the pool are shared with the investor class pro rata.
There are generally two forms of distribution to investors: (1) interest
only (or nonamortization), and (2) principal pay-down (or amortization).
Credit card accounts, being revolving credits, have no natural maturity,
unlike most other loans. Therefore, credit card securities have the dis-
tinguishing feature of a period during which only interest is paid (usually
eighteen months to three years), followed by an amortization period. During
the interest only period, all principal payments are allocated to the seller
class and are used to purchase all the additional receivables which have
been added to the designated accounts.13
The Loan Sales Market
Sales of commercial and industrial loans, or secondary loan participations,

are quite distinct from asset-backed securities because they typically do
not involve the cash flows from a pool of loans, but from a single loan. It
appears that these loans are not pooled for at least two reasons. First,
commercial and industrial loans tend to be of much larger denomination
than most consumer loans and mortgages, so that there is less need for
pooling to obtain economies of scale in selling them. Second, their lack of
homogeneity makes them less easily classified by credit scoring methods
than consumer loans. They do not necessarily have standardized contracts

with readily known risk-return characteristics. Today, less than half of
these loans are the obligations of publicly rated investment-grade firms.
This would appear to make it more difficult and costly for a third party
guarantor to analyze the overall credit risk of a pool of these loans.
Background on the development of this market can be found in Gorton
and Haubrich (1988).
The important feature of loan sales, however, that distinguishes them
from asset-backed securities is the fact that loan sales offer no explicit
mechanism for reducing the risk to the loan buyer, that is, they come
without explicit issuer or third party guarantees. The key provision of the
secondary loan participation contract is the following, taken from a typical
contract:
Each Participation purchased by the Participant hereunder will be without
recourse to the Bank and for the Participant's own account and risk, and the
Bank makes no representation or warranty as to, and shall have no responsibility
for, the due authorization, execution or delivery by any Borrower of any
Participated Loan or Loan Documents, the legality, validity, sufficiency, en-
forceability or collectibility of any Participated Loan or Loan Documents, the
value, validity, perfection or priority of, or any other matter concerning, any
collateral or other support for any Participated Loan, any representation or
warranty or information made or furnished by any Borrower, the performance
or observance by any Borrower of any of the provisions of any Loan Docu-
ments, the financial condition of any Borrower or any other matter relating to
any Borrower, Loan or Loan Documents.
In other words, if the underlying buyer of the loan fails (as has occurred
in several cases) the loan buyer has no recourse to the selling bank. 14
Moreover, the loan buyer has no legal relationship with the underlying
borrower since the loan sale contract is strictly between the bank and the
loan buyerY
In the early 1980s the incentive problems associated with loan sales did
not seem as great as they do now. Initially, the loans that were sold were
very short-term obligations of banks' best corporate customers, that is,
very well-known firms with triple A credit ratings. In 1985, 80 percent of
the loan sales contracts had maturities of ninety days or less. By mid-1987
over half of the loan sales contracts had maturities exceeding one year,
indicating a lengthening of loan maturities. Also, by mid-1987 less than
half the loans sold were the obligations of investment-grade borrowers,
and by 1989 only 35 percent were of investment grade borrowers. 16 While
this fraction rose somewhat in 1990 to 44 percent, the overall trend has
been toward the sale of higher risk 10ansY
The ability of the loan sales market to incorporate increasingly risky

loans has been accompanied by an enormous growth in volume over a
very short period of time. This growth is shown in table 1-3. The outstanding
volume of commercial and industrial loan sales has grown from minuscule
amounts of loans, essentially sold through the traditional correspondent
network, to over $200 million. In 1990, nine money center banks accounted
for approximately 69 percent of loan sales. Table 1-4 provides informa-
tion about the loan sales volume of the top twenty-five loan selling banks
in the second quarter of 1990.
To date, most buyers of loans have been other banks. Over three-
fourths of loan sales are purchased by foreign banks (38.3 percent) or
domestic banks (37.6 percent), though the proportion purchased by
nonbanks is becoming significant. 18 Thus, loan sales do not appear to
represent a simple growth of the correspondent banking network. (Table
1-4 provides the amounts of loans purchased by the top twenty-five loan
selling banks in the second quarter of 1990.)
Because loan sales seem to be such a clear contradiction of the
nonmarketability of bank assets, they are of most interest to us. In addition,
it is the historical illiquidity of these bank assets that has been the basis
of deposit insurance. Paradoxically, however, loan sales are not obviously
an innovation since the loan sales contract itself has been in use for decades,
although the volume sold under the contract was not particularly import-
ant. Thus, the change allowing for the sudden spurt in sales of loans
cannot be accounted for by any particular change in the contract per se.
Something in the environment must have changed to allow this existing
contract to support the market.
Implicit Loan Sales Contracts: The First Tests
Loan buyers must be assured that the bank originating the loan will continue
to behave as if the loan had not been sold. If incentives for the selling
bank are not maintained, the value of the loan will decline. In the past it
appears that it was not technologically possible to provide such assurances
without the bank directly risking its equity. To analyze the existence of
markets for bank assets, we will first proceed by attempting to identify the
mechanism that makes loan sales incentive compatible. If we can identify
this mechanism, we will then ask how it is enforced. As mentioned in
the Introduction, three possible mechanisms are: (1) implicit insurance by
the selling bank against possible losses to the buyer due to failure of the
underlying borrower, (2) maintenance by the selling bank of a stake in
Table 1-3. Quarterly Outstanding Loan Sales of Commercial Banks* ($ billions).
Date Loan Sales Loan Purchases
198302 26.7
03 26.8
04 29.1
198401 32.8
02 33.3
03 35.5
04 50.2
198501 54.6
02 59.9
03 77.5
04 75.7
198601 65.4
02 81.2
03 91.3
Q4 111.8
198701 162.9
02 195.2
03 188.9
Q4 198.0
198801 236.3 16.64
02 248.4 16.22
03 263.0 17.65
04 286.8 19.29
198901 272.7 16.16
02 276.5 18.20
03 290.9 17.82
04 258.7 19.89
199001 228.3 16.07
02 190.2 15.94
* Sales reported are gross and exclude sales of consumer loans and mortgage loans. Also
excluded are loans subject to repurchase agreements or with recourse to the seller.
Source: FDIC Call Reports, Schedule L.
Table 1-4. Top 25 Loan Selling Banks in 199002.
DSB Number Name Loans Sold Loans Purchased
12060730 Security Pacific National Bank 53,602,449 o
2364840 Bankers Trust Company 34,602,000 419,000
2364900 Citibank, N.A. 14,526,000 768,000
2364985 Chemical Bank 10,461,000 62,000
2364965 Chase Manhattan Bank, N.A. 8,835,613 226,602
4426723 Mellon Bank, N.A. 8,038,004 80,343
2365328 Morgan Guaranty Trust Company of New York 5,591,960 o
2365290 Manufacturers Hanover Trust Company 5,507,000 4,000
2365441 Security Pacific National Trust Company 4,520,819 o
7171630 First National Bank of Chicago 4,369,881 72,632
7171560 Continental Bank, N.A. 3,915,535 96,244
12061400 Bank of America, NT & SA 3,458,000 150,000
3420204 Philadelphia National Bank 2,696,399 57,091
5512345 Signet Bank-Virginia 2,425,409 132,137
2364880 Bank of New York 2,018,508 62,131
2361310 Marine Midland Bank, N.A. 1,817,920 36,500
12060755 First Interstate Bank of California 1,217,951 41,442
1250487 Bank of New England, N.A. 1,116,253
5512430 Crestar Bank 1,104,857 24,197
°
2364893 Bank of Tokyo Trust Company 1,048,930 113,259
1250370 First National Bank of Boston 816,384 9,236
9274037 First Bank, N.A. 590,536 52,280
7171650 Harris Trust & Savings Bank 581,274 94,685
2342200 Midlantic National Bank 531,009 o
5370355 NCNB National Bank of North Carolina 530,072 267,167
Source: FDIC, Call Reports, Schedule L.
any loan sold, and (3) a lack of information asymmetry between loans
selling banks and loan buyers. We first consider implicit insurance or
guarantees.
Implicit Guarantees By the Selling Bank
The possibility of implicit insurance is clearly not motivated by any ob-

served feature of the contract; the contract explicitly denies the existence
of such recourse. It is, however, motivated by some empirical observa-
tions, though we cannot document the observations. Anecdotal evidence
suggests that banks sometimes buy back loans they have sold. 19 Two
factors combine to suggest the possibility of an implicit guarantee on
loans sold. First, banks, because of regulatory restrictions, cannot be
explicit about guarantees. Both selling banks and loan buyers might prefer
that guarantees be explicit, but this is not possible if the bank wants to
remove the loans from the balance sheet and thus avoid reserve and
capital requirements. Second, loan buyers must be concerned with the
possibility that they may wish to sell a loan they have purchased. The
question of the existence of secondary markets is usually important in any
security market. Again, because of securities-law restrictions, loan sales
contracts explicitly forbid resale of the participation. 20 Thus, if the buyer
needs cash, once having bought the loan, only the selling bank is likely to
buy the loan back. 21
The issue of implicit insurance or a guarantee, then, can appear in
the following way. Suppose a loan buyer discovers that the underlying
borrower's credit risk has deteriorated, causing the value of the buyer's
participation to unexpectedly decline. This buyer may approach the selling
bank and ask the bank to buy back the loan, claiming that there is a need
for cash. The bank, wanting to provide some liquidity for its participations
has already (informally) said that it will repurchase its participations (even
though the participation contract is explicit about the bank not being
required to repurchase loans).22 The bank also knows that the borrower
is more risky now than originally. At what price does the bank buy the
participation back, if it does buy it back?
There are no data to directly address this question. However, Gorton
and Pennacchi (1989) indirectly tested for the presence of such an implicit
guarantee. The intuition for their procedure was as follows: If loan buyers
believe the participation contract literally, so that there is no implicit
guarantee or insurance of the loan, then loan sales prices should contain
no component which values the default risk of the selling bank. If there
is an implicit guarantee, then the value of this guarantee would be related

to the default risk of the selling bank since the guarantee would be worth
less if the bank defaulted. In this latter case, loan sales prices should be
related to the selling bank's default risk.
This same logic clearly applies to commercial paper. The guarantee is
explicit in the case of commercial paper since commercial paper typically
requires some kind of credit enhancement from a bank. Almost all com-
mercial paper is "backed" by a line of credit, an irrevocable standby letter
of credit, a direct pay commitment, or an insurance company indemnity
bond. 23 Commercial paper yields should contain a component which
essentially prices the default likelihood of the bank that provides this
backup credit.
Gorton and Pennacchi (1989) posited that the yields on loans sold and
on commercial paper should be a function of: (1) the opportunity cost of
funds to loan buyers and commercial paper holders; (2) the risk of the
firm issuing the paper or receiving the loan; and (3) the quality of the
guarantee given by the bank to buy back the loan sold or provide a
backup credit line in the case of commercial paper. In particular, consider
the following model:
(1)
Yjtis the yield at time t on a debt security of type j, j = 1 in the case of
a loan sale and j = c in the case of commercial paper. f, is a measure of
the default risk of the borrowing firm; b t is a measure of the default risk
of the bank providing the guarantee on the security; r t is the opportunity
cost of an equivalent maturity riskless investment at time t. The aj/s are
assumed to be constant and E jt is an independently distributed error term.24
The risk premium attached to the borrowing firm is the coefficient on the
measure of the firm's likelihood of default, the variable f,. In equation (1)
the coefficient a jl would be less than 1 if the bank partially guarantees the
debt. If the bank's guarantee is not only credible, but insures that the
debt is riskless, then a jl would equal zero. The bank's guarantee, how-
ever, will be limited by the likelihood of the bank defaulting and, thus,
being unable to fully honor the guarantee. A risk premium corresponding
to this probability of failure is measured by the variable bt •
Unfortunately, equation (1) cannot be directly estimated for either
commercial paper or loan sales because there is no data available for
calculating the underlying firms' default likelihoods. Gorton and Pennacchi
(1989), however, were able to estimate the following equation which
eliminates the unknown variable f, by substitution:
Equation (2) is designed to test the relative strengths of guarantees given

by banks on selling loans versus guarantees in the form of backup credit
lines for commercial paper. 2S The coefficient all/ac! measures the relative
strength of the weight placed on the borrower's default risk. The more
complicated coefficient on b l measures the relative strength of the guar-
antees provided by banks on the two instruments.
Empirical tests of the model, a total of 32 observations, were carried
out using weekly data from July 1987 to March 1988. The model was
estimated for ninety day loan sales for the cases of Al \Pl rated borrowers
and separately for A2\P2 rated borrowers. The details are provided by
Gorton and Pennacchi (1989). Briefly, the results were as follows: The
estimated coefficient all/ac! is about 0.8 for each quality borrower (that
is, Al\Pl and A2\P2); both coefficients were significantly different from
zero at the five percent confidence level. This result suggests that inves-
tors place less weight on borrowers' default risk in the case of loans sold
relative to commercial paper.
The estimated coefficient on the bank default risk variable, b l , how-
ever, was significantly negative in both cases. Omitting the details, this
would, in contradiction to the first set of estimates, seem to imply that a
stronger guarantee is given on commercial paper that for a loan sold.
Gorton and Pennacchi (1989) offered several reasons why these appar-
ently contradictory results might be consistent with the hypothesis of the
existence of implicit guarantees. One point raised concerns the fact that
the guarantee on commercial paper is a European option, while the
guarantee on loan sales may be an American option.26 Basically, however,
the data set, averaged loan sales yields collected from a survey by Asset
Sales Report, was not fine enough to unambiguously answer the question
of whether implicit guarantees are at the root of the loan sales market.
Gorton and Pennacchi (1989) were limited by the available data. Not
only did an indirect approach have to be taken to address the question of
interest, but the data set used did not reveal the second possible mechanism
for insuring incentive compatibility. The possibility that the fraction of a
loan sold which was retained might playa significant role, even though
there is nothing in the contract about this, only became clear when more
data became available.
Maintenance of a Stake in the Loan Sold
The motivation for the possible existence of the second incentive com-
patibility mechanism is empirical. It turns out that selling banks often
retain a fraction of the loan sold. In a sample of 872 loans sold by a major
money center bank during the period from January 20,1987 to September
1,1988, the selling bank held a portion of 360 of these loans (41 percent).
Of those loans for which the bank held a positive share, its share aver-
aged approximately 59 percent.27 Table 1-4 provides more detailed infor-
mation on the fraction retained.
Causal observation of Table 1-4 suggests that the fraction retained
increases as maturity increases, holding the rating of the underlying
borrower constant. Perhaps less obvious is whether the fraction retained
increases as the underlying borrower's rating worsens, holding maturity
constant.28
The behavior of the fraction retained is difficult to understand without
consideration of the prices at which loans were sold, and possibly without
consideration of whether or not an implicit guarantee is offered. Certainly,
an implicit guarantee and maintenance of a fraction of the loan by the
selling bank are not mutually exclusive possibilities. Also, the extent of
the fraction sold and the extent of the guarantee are not independent of
the price of the loan sale.
The fraction of the loan sold by the selling bank was not a variable that
was considered by Gorton and Pennacchi (1989). The behavior of this
variable could explain the results. To see how the fraction of the loan sold
affects the results, Gorton and Pennacchi (1990A) rerun regression (1)
with the new data set, adding the fraction of the loan sold, s, as an ex-
planatory variable. Also, with the new data set the risk of the underlying
borrower defaulting can be directly measured as the spread of the loan
rate over the riskless rate (proxied by LIBOR) of the same maturity,
'b - 't· The results of that regression are as follows:
'1 - 't = .00098 + .2388 ('b - 't) + .0079 b -
.00051 s
(.00009) (.0118) (.0506) (.00010)
No. of obs. = 872. R2 = .329. (Standard errors in parentheses).
Note that the loan sale spread, '1 - 't, is significantly related to the spread
paid by the borrowing firm. It is also positively related to the risk of bank
failure, but this coefficient is not significantly different than zero.29 The
fraction of the loan sold is, paradoxically, negatively related to the spread
Table 1-5. Fraction Sold.

Maturity (days)
Rating 0-5 6-15 16-30 31-60 61-90 90+

A1+
Average Fraction Sold 1 1 1 .843 .556
Number of Observations 1 1 1 9 3 0
Al
Average Fraction Sold .917 .867 .746 .455 1
A2
Average Fraction Sold .733 .826 .810 .746 .600 .608
A3
Average Fraction Sold .778 .771 .625 1
NR
Average Fraction Sold .889 .784 .738 .703 .707 .750
Source: Money Center Bank.
on the loan sale. This is, however, not necessarily, a puzzle. It may be the
case that banks choose to sell larger fractions of less risky loans. Perhaps
some loans require less monitoring by the selling bank. Selling a larger
fraction of these loans does not increase the yield paid to loan buyers
as much as more risky loans.
Clearly, introduction of the new variable, the fraction sold, s, so suf-
ficiently complicates the analysis that a model is required to be clear
about the issues. Gorton and Pennacchi (1990A) introduce a model aimed
at more rigorously addressing the question of the existence and possible
interaction between the fraction sold, an unobservable guarantee, and the
price of a loan sold.
Loan Sale Incentive Compatibility: More Sophisticated

Tests
This section summarizes the model presented in Gorton and Pennacchi

(1990A).30 The model provides a basis for empirical work, the results of
which are summarized here.
A Model of Loan Sales
In this model, banks have an incentive to sell loans in order to avoid the
costs of required reserves and required capital. Banks can improve the
expected return on loans by monitoring borrowers, but bank monitoring
of borrowers is assumed to be unobservable so that banks and loan buy-
ers cannot write contracts contingent on the level of monitoring. Thus, as
discussed above, since banks provide a special service, namely, monitor-
ing, a moral hazard problem exists. The bank may not monitor at the
most efficient level after having sold its loans. Since loan buyers are
rational, and understand the moral hazard problem, they will only buy
loans if there is some way of forcing banks to monitor.
The bank's problem is to maximize expected profits from the sale of a
particular loan. 31 A bank loan is assumed to require "one dollar of initial
financing. It produces a stochastic return, x, at the end of 'l" periods, where
x e [0, L], and where L is the promised end-of-period repayment on the
loan. The return, x, has a cumulative distribution function of F(x, a), where
a is the level of monitoring by the bank. The bank has a constant returns
to scale technology for monitoring loans. The cost function is given by
c(a) = c*a.
The bank can sell a portion, s, of the loan where s e [0, 1]; the bank
retains the portion (1- S).32 Risk neutral loan buyers require an expected
return on loans purchased of rf' The bank finances its portion with a
weighted average cost of deposit and equity financing given by rio Also,
the bank can offer an implicit (partial) guarantee against default of a loan
that it sells. The proportion of a loan sale that the bank promises to
guarantee is r e [0, 1]. Such implicit guarantees are costly because regu-
lators do not approve of them,33 The cost to the bank is assumed to be
given by k(r), where k' > 0 and kIf ~ O. The bank can fulfill this guarantee
only if it is solvent at the time the loan matures. The probability that the
bank is solvent, p, is assumed to be uncorrelated with the return on the
loan.
The optimal loan sale contract requires the bank to choose a level of
monitoring, a, the fraction of the loans to be sold, s, and the fraction of

the loan to be implicitly guaranteed, r:
max fOL (1- s)xdF(x, a) - srp fOL (L - x)dF(x, a) - c(a) - k(r) - er/t ]
where:
] = 1- e- rfT [f: sxdF(x, a) + sr fOL (L - x)dF(x, a)]
subject to: (i) foL [(1-s)+srp]xdFa (x,a)=c'(a)

(ii) s :5; 1
(iii) y:5;1 (3)
In the bank's problem, (3), the first term of the objective function is the
expected return on the portion of the loan return held by the bank. The
second term is the expected value of the guarantee the bank gives to the
loan buyer. ] is the amount of internal (debt and equity) funding which
the bank provides when fraction s of the loan is sold. Constraint (i) is the
incentive compatibility constraint.
Omitting the explicit steps needed to solve the bank's problem, (3), at
the optimum the following relation holds between the choice of the frac-
tion to be sold, s, and the size of the implicit guarantee, r:
sp( (JL - J.L) = (k'( r) + e)(1 - rp) (4)
where (J == exp[(,/ - ,,)-r], J.L is the Lagrange multiplier associated with
constraint (ii) of the problem, and e is the Lagrange multiplier on con-
straint (iii) of the problem. The relation (4) relates the endogenous vari-
ables of the fraction sold and the size of the guarantee.
To obtain a testable implication of the model, it is necessary to adopt
an explicit functional form for the relation between the level of monitor-
f:
ing and the expected return on the loan. The form we assume is 34
xdF(x, a) = L(1- ae-pa ) (5)
Using this form, the first order conditions for problem (3) also yield
another expression:
(Je-(r,,-rf)T - J.LIL
s = -------:-'----:----
(1- yp)[1 + () - e-(r,,-rf)'r - J.LIL]
s= '[-'f-J.LI(-rL) (6)
(1- rp)[,[ -'f + 'Is -'f - J.LI( -rL)]
In expression (6) note that when s is less than one, so that f.l = 0, the
fraction of the loan sold is approximately proportional to the ratio of the
excess cost of internal financing (over the risk free rate) to the excess cost
of internal financing plus the excess cost of funds received from the loan
sale.
Equations (4) and (6) provide the basis for the first empirical tests in
Gorton and Pennacchi (1990A). The two relations define the links between
the bank's optimal choice of sand y. By choosing a parametric form for
the guarantee cost function k(r), the equilibrium values of sand rcan be
determined. However, note that if one assumes that the bank is constrained
to give the same partial guarantee, r, on all the loans that it sells, then
equation (6), alone, is sufficient to determine the bank's optimal share of
each loan sold given the guarantee.
We now turn to summarizing the results of estimating the model. Details
are provided in Gorton and Pennacchi (1990A).
Empirical Results 1: Equal Guarantee on All Loan Sales
The model is tested using the sample of 872 individual loan sales made by
a large money center bank during the period January 20,1987 to September
1, 1988. If the loan selling bank's partial guarantee, r, is the same for each
loan, then we can treat it as a parameter to be estimated using equation
(6).35
The empirical results in this case do not support this hypothesis. The
basic result obtained is that the term (1- rp) in equation (6) appears with
the wrong sign or that the point estimate of ris negative and significantly
different from zero. This result is nonsensical, and so, offers no support
for the hypothesis. However, we do find empirical support for the propo-
sition that the bank's choice of the fraction sold is a decreasing function
of the loan sale spread, indicating that the loan sales buyers are aware of
the implications of the fraction sold for the bank's incentives.
Empirical Results 2: Different Guarantees on Loan Sales
When the bank is assumed to be able to make different levels of partial

guarantees on different loans that it sells, then it will optimize using equa-
tion (4), as well as (6). In this case, the parameter r cannot be directly
estimated. Instead, we obtain estimates of the regulatory cost function
k(r). We assumed that this function had a simple quadratic form: k(r) =
ko + k 1r+ (l/2)k2r2. Based on the parameter estimates for k1 and k2 the

optimal level of the loan sale guarantee can be calculated for each obser-
vation. 36 The "test" of the hypothesis is essentially to examine whether
the estimates of the guarantees on the individual loans are sensible.
The results again turn out to be unsupportive of the existence of an
implicit guarantee hypothesis, as the implied values for r for the different
loans are predominantly negative.
Summary
There appears to be no evidence of an implicit guarantee or implicit

insurance in loan sales contracts. There is weak evidence that the fraction
sold is used by selling banks in a way which is consistent with incentive
compatibility. It is important to keep in mind, however, that the results
are not only tests of the hypothesis of immediate interest, but also of joint
hypotheses concerning specific functional forms for the monitoring tech-
nology and for the function determining the regulatory cost of issuing
implicit guarantees.
Technological Change Creates Symmetric Information
The results of estimating the model in search of evidence for an implicit

guarantee are distinctly negative, though there is some weak evidence for
the use of the fraction sold as a device for insuring incentive compatibility.
An important feature of the model, however, is the fact that it assumes
that loan buyers can force the bank to honor the guarantee or enforce the
restriction that not all of the loan be sold. As noted in the Introduction,
however, there is nothing in the participation contract that specifies that
the selling bank is, in fact, providing a guarantee, or any contract provision
which restricts the proportion of the loan which the bank may sell. On the
contrary, the contract suggests that the selling bank is not providing any
guarantee. Thus, it is not clear how either of these two mechanisms would
be enforced. This observation is true for any implicit contract which is
proposed as an explanation of the opening of the loan sales market.
Since any implicit contract feature would have to be enforceable by
market forces, the question is raised of whether there is some change in
the underlying contracting technology that, by itself, could explain the
incentive compatibility of loans sales.
Various sources of indirect evidence suggest that significant changes
affecting the underlying contracting technology have occurred. While it

is difficult to measure precisely, the tremendous advances in computer
technology and telecommunications during the past two decades have
undoubtedly led to a drastic reduction in the cost of gathering, analyzing,
and transmitting information. This is not to say that this lower cost of
collecting and disseminating information has reduced information
asymmetries between all borrowers and investors in the economy by equal
amounts. One would expect that institutional investors who acquired the
necessary computer and telecommunications technologies would have
benefitted more than most individual investors. Intermediaries, such as
commercial banks, thrifts, mutual funds, insurance companies, and pen-
sion funds were likely to be most affected by this technological progress,
in part because they were positioned to benefit by economies of scope
from investing in computers and telecommunications. Computer and
telecommunications systems are vital to lowering the cost of banks' and
thrifts' transactions services, the cost of mutual funds' servicing of
shareholder's accounts, the cost of insurance companies' accounting and
analysis of customers' risks, and the cost of pension funds' servicing their
participants' accounts.
Informal Evidence of Technological Change
If institutional investors have benefitted the most from technologies that

have reduced information costs, one would expect that they should hold
a large proportion of those assets where information acquisition would be
critical to valuation. High yield bonds or "junk" bonds are assets that
would appear to fit into this category. The junk bond market grew dra-
matically during the 1980s. Prior to 1981, annual new issues were less than
$1.5 billion, but peaked to over $30 billion new issues in 1986, and settling
in the range of $25 to $30 billion through the end of the decade?7 Im-
portantly, as of year end 1988, three-quarters of the stock of junk bonds
was held by insurance companies, money managers, mutual funds, or
pension funds. Individual investors owned only 5 percent of the stock of
junk bonds.38
Given that financial institutions have invested relatively greater amounts
in information technology and have therefore experienced relatively greater
reductions in information costs, what effect would this have on bank
lending? First, if other financial institutions can directly acquire infor-
mation about borrowers at low cost, there should be less need for banks
to provide information production and monitoring services for many bor-
rowers. Banks' comparative advantage in eliminating duplication of infor-

mation services or free-riding problems by multiple investors is likely to
be reduced when these multiple investors have low costs of information
acquisition. Therefore, if regulations such as reserve and capital re-
quirements increase the costs of funds of banks who already must pay
competitive rates for deposit and equity financing, these banks would
become uncompetitive as a source of financing for many borrowers. The
previously mentioned growth in the junk bond market might reflect this
phenomenon. There is also evidence that banks' comparative advantage
has been reduced even more at the other end of the debt risk spectrum.
The ratio of nonbank commercial paper to banks' commercial and industrial
loans rose from less than 10 percent in 1959 to over 75 percent in 1989,
indicating a migration of large and medium sized corporations from bank
financing to publicly issued securities. 39
Second, even if many financial institutions could not directly acquire
information about certain classes of borrowers at low cost, they may be
able to inexpensively verify the information collected by another bank
regarding these borrowers. This would be expected to produce two effects
on the market for loan sales. One effect would be that third party guar-
antees on asset-backed securities would become feasible if the third party
bank or insurance company could verify the accuracy of the originating
bank's credit analysis. This would explain the growth of the asset-backed
securities market. The other effect would be that individual financial in-
stitutions could verify the accurate production of credit information and
monitoring by a bank wishing to sell a single loan that it has originated.
This would explain the ability of banks to sell single commercial and
industrial loans to other institutions. To take this idea a step further, if we
make the logical assumption that banking institutions, being in the same
line of business as a loan selling bank, are able to verify the credit analysis
and monitoring of the loan selling bank at lower cost than nonbank financial
institutions, this would explain why over three-fourths of loan buyers are
other banks.
Empirical Results 3: Observability in the Loan Sales Market
A final test is performed in Gorton and Pennacchi (1990A) that provides

some evidence of the ability of loan buyers to verify the performance of
loan selling banks. Recalling the model described in the previous section,
note that the incentive compatibility constraint (3i) makes the assumption
that the bank's level of monitoring is not directly observable by the

loan buyer. If monitoring were observable, the bank and loan buyer could
contract to set the level of monitoring at its most efficient level, namely
the level which would satisfy (3i) where s = 0, that is, the level of moni-
toring the bank would choose if it had not sold the loan. This suggests
that a direct empirical test of the incentive compatibility constraint might
be able to shed light on the question of whether or not bank monitoring
of borrowers is observable by loan buyers. Using constraint (3i), as well
as the assumed functional form for the expected return on the loan, equa-
tion (5), the following relationship can be derived:
P= c 1- rp (7)
L(1_e-(rl,-rf)~) [1-s(1- yp)]
The left hand side of equation (7) equals the loan specific parameter
Pwhich is a measure of a given loan's benefit from greater monitoring.
On the right-hand side are two multiplicative ratios. The first, which is
independent of the fraction of the loan sold, s, and the fraction of the loan
guaranteed, r, is the relationship between the loan sale yield, 'Is> and P
that would hold if monitoring were observable. The second term on the
right hand side denotes the effect of unobservability. We can then test the
relationship given in equation (7) using the borrower's commercial paper
rating as a proxy for Pi' This was done using 360 of the 872 observations
on loan sales for which the borrower reported a commercial paper rating.
Equation (7) was estimated in log form as a probit model, where a
borrower's commercial paper rating was assumed to be a discrete measure
of its benefit from bank monitoring. It was assumed that the fraction
of the loan guaranteed (if any) was the same for all loan sales, so that r
was treated as a parameter.
If monitoring is observable, there should be a positive and significant
relationship between P and the first term on the right-hand side of (7),
while there should be an insignificant relationship between Pand the second
term. Both terms should show a significant relationship.to Pif monitoring
is unobservable.
The empirical results support observability to the extent that the first
term on the right hand side of (7) was significantly related to Pwhile the
second was not. However, the effect of the first term was not as strong as
the theoretical model would suggest. 40 We interpret these results as sug-
gesting that loan buyers have the ability to verify, at least partially, the
performance of loan selling banks.
Concluding Comments: Marketable Bank Assets and

Regulation
Technological change appears to have resulted in the opening of markets

for bank assets. Previously nonmarketable assets can now be sold. The
loan sales market has the important feature that bank assets are sold
without the creation of a contingent liability, that is, there is no explicit
contract feature which eliminates the risk that the selling bank will fail to
perform on the original contract with the underlying borrower. While this
development is not yet well understood, it would seem to be a rather
important shift in the way banking is conducted.
The existence of well functioning markets for bank assets, like those
that appear to be developing, does not mean that intermediation per se
is ending. All the explanations for loan sales considered above imply that
banks still offer services for certain classes of borrowers that cannot be
obtained in capital markets via issuance of open market securities.41 The
asset sales contracts mean, however, that it is no longer necessary for
banks to hold loans until maturity, risking their equity during the life of
the asset created.
If bank assets can be sold in fairly liquid markets, then the rationale
for bank regulation is called into question since it is fundamentally based
on the illiquidity of bank assets. Deposit insurance was, at least originally,
aimed at providing the public with a circulating medium which did not
expose people to losses either due to better informed traders or because
of banking panics. If markets for bank assets open, the market incom-
pleteness necessitating government intervention would seem to be gone.
While the loan sales market is sizeable, it is by no means clear that the
requisite volume of bank assets are marketable.
The technology underlying banking is changing more generally. The
same technological forces that have allowed the markets for asset sales to
open, have also allowed for a much larger set of nonbank possibilities for
the private provision of liquidity. (These are discussed in Gorton and
Pennacchi (1990C).) In fact, as we have argued elsewhere, the combina-
tion of illiquid assets with demandable debt liabilities appears to be an
arrangement which is no longer the most efficient way of providing either
a circulating medium or monitoring services. Technological change is forcing
a separation between these two sets of activities.
The regulatory issues which must be confronted in the banking indus-
try are larger than can be discussed here in detail. Suffice it to say that
technological forces cannot and should not be regulated away.
Notes
1. Evidence that bank loans are unique is provided by James (1987), Lummer and
McConnell (1989), and Fama (1985). Using event study methodology, James shows that
there is an (abnormal) positive return on stocks when nonfinancial firms announce that they
have obtained a bank loan. The stock market reaction (in the sense of abnormal return)
upon the announcement of offerings of common stock, preferred stock, convertible preferred
stock, convertible bonds, and straight bonds is negative, though not always statistically
significant. (See Smith (1986).) A similar study by Lummer and McConnell discriminates
between new and revised bank loan announcements. Their results show that only revised
loan announcements have a positive effect on stock returns, suggesting that banks learn
about borrowers after an initial loan is made. Also, both Fama and James point out that
bank borrowers, not bank depositors, bear the cost of bank reserve requirements. Yield
spreads between bank CDs, which are subject to reserve requirements and other money
market instruments, are not significantly different. This implies that firms are paying more
to borrow from a bank than they would if they borrowed the same dollar amount by issuing
securities on the open market. Presumably, firms would only do this if bank loans involved
the production of some services not obtainable by issuing securities in the capital market.
2. All theories of banking panics crucially depend on the illiquidity of bank assets either
by assuming that there is an information asymmetry between banks and depositors or by
assuming that there is a cost to liquidating bank assets. See Calomiris and Gorton (1990) for
a discussion of the causes of banking panics.
3. Throughout this essay we will ignore mortgage-backed securities, at least those secur-
ities that are backed by mortgages carrying government (e.g., VA and FHA) guarantees
against default. This government credit enhancement appears to be the primary determinant
of their marketability.
4. Apparently, the logic behind this regulation is that if a bank sells a loan with recourse
and thus retains the loan's risk, it should be required to hold the same amount of capital and
reserves as if it had retained the loan's risk by not selling it.
5. Since nearly all loans are sold without recourse or guarantee, this supports the hypo-
thesis that bank loan sales are motivated, at least in part, by a desire to avoid the regulatory
costs associated with required reserves and required capital. Note that this hypothesis is
consistent with the observation that most loan buyers are other banks. Pennacchi (1988)
shows that banks possessing considerable loan making opportunities but competitive deposit
markets (money center banks) will find it profitable to sell loans while banks possessing
limited loan making opportunities but having market power in their local deposit markets
(foreign banks and smaller domestic banks) will find it profitable to buy loans. It follows
that the recent increase in aggregate loan sales volume can be explained by the recent
increase in deposit market competition faced by many banks.
6. The formal arguments, and details are provided by Campbell and Kracaw (1980),
Boyd and Prescott (1986), Diamond (1984), and Gorton and Haubrich (1987).
7. Presumably this is part of the logic behind capital requirements for commercial banks.
8. Some hypotheses explaining the existence of markets for bank assets we do not ex-
plore. For example, one possibility is that bank loan sales and asset-backed securities simply
represent bank underwriting and are strictly limited to assets which do not involve any
information production or monitoring. This hypothesis argues that asset sales represent an
attempt by commercial banks to compete with investment banks. We believe this explanation
cannot be totally satisfactory. First, most loans that are sold are the obligations of borrowers
who do not have commercial paper ratings, suggesting that they do not have ready access
to public securities markets. (See Gorton and Haubrich (1989).) Secondly, it seems unlikely
that, lacking any special services by the bank, this would be a successful way to compete.
In a loan sale, if the underlying borrower fails, the holder of the secondary participation has
no legal connection with the failing firm and hence cannot be represented in bankruptcy
court. Holders of marketable securities, thus, have important rights which holders of sec-
ondary loan participations do not have.
9. Flannery (1989), Benveniste and Berger (1987), Cumming (1987), Greenbaum and
Thakor (1987), Kareken (1987), James (1988), Pennacchi (1988), and Boyd and Smith (1989)
suggest a variety of possible motivations for loan sales.
10. There are a large number of other types of loans that have formed the basis for
asset-backed securities. Examples include light truck loans, motorcycle loans, boat loans,
trade receivables, and equipment leases.
11. Other offerings of automobile securities by commercial banks in 1988 included: two
issues totaling $480 million by Marine Midland Bank, two issues totaling $470 million by
Chemical Bank, a $250 million issue by Huntington National Bank, and a $181 million issue
by Signet Bank (Virginia). Thrifts also issued automobile securities during 1988. Empire of
America Federal Savings Bank made two offerings of automobile securities totaling $631
million. Western Financial Savings Bank came to market four times in 1988 for a total
issuance of $460 million. Rochester Community Savings Bank offered $125 million of these
securities which featured a floating rate coupon.
12. In other words, suppose the initial pool was $100 million of principal with $75 million
sold to the investor class and $25 million sold to the seller class. Then if, at the end of a
month, the principal balance of the pool falls to $95 million, the investor class remains at
$75 million, while the seller class would drop to $20 million.
13. Details and terminology vary with the issuer. For further details see "Credit Card
Backed Securities: An Introduction," Goldman, Sachs & Co., January 1987; "Securitization
of Credit Card Receivables Using a Senior/Subordinated Structure," Goldman, Sachs &
Co., December 1988; "Rating of Credit Card Receivables" Duff & Phelps Inc., (no date);
"Credit Card Receivables: Moody's Examines the Risks," Moody's Investors Service, January
1987.
14. Of the $80.2 billion dollars of outstanding loans sold as of June 30, 1990 by banks
responding to the Federal Reserve System's Senior Loan Officer Opinion Survey of Lending
Practices, 3.5 percent were nonperforming loans. It is not clear that this entire amount
represents defaults since some loans may have been nonperforming when they were sold.
One year earlier 1 percent of loans were nonperforming out of a total outstanding amount
of $72.2 billion for survey respondents.
15. Some loan sales take the form of assignments, rather than participations. Assign-
ments are quite strong contracts because the assignees have all the rights and responsibilities
of the lead or originating bank. In practice these rights vary a great deal depending on the
complexity of the contract and the sophistication of the parties involved. Importantly, many
assignments specify a minimum amount which must be held by the originating bank. See
Gorton and Haubrich (1989) for further discussion.
16. This is based on the sample of banks surveyed in the Federal Reserve System's
Senior Loan Officer Opinion Survey on Bank Lending Practices for various years. See also
Gorton and Haubrich (1989).
17. About 37.5 percent of the outstanding loans sold as of June 30, 1990 were merger and
acquisition related. (See the Federal Reserve Senior Loan Officer Opinion Survey of Bank
Lending Practices, August 23, 1990.)
18. According to the Federal Reserve's August 1990 Senior Loan Officer Opinion Survey
on Bank Lending Practices, 11.8 percent of loans were purchased by domestic banks with
assets under $2 billion while 25.8 percent of loans were purchased by domestic banks with
assets over $2 billion. Nonbank buyers accounted for 24.1 percent of loan sales, with 7.9
percent of this being purchases by nonfinancial corporations and 16.2 percent being purchases
by other nonbank buyers.
19. No bank we contacted was willing to provide any hard evidence on the amounts of
loan repurchases or the prices of repurchases, though banks were often willing to admit that
repurchases of loans were not infrequent.
20. By the G1ass-Steagal Act banks are prohibited from issuing securities. In order to
avoid blatant violations of Glass-Steagal secondary loan participation contracts explicitly
forbid secondary market sales. The existence of such sales plays a role in the Icgal deter-
mination of what constitutes a security. See Gorton and Haubrich (1989).
21. Investment banks have recently started to make a secondary market in some
participations. See Wall Street Journal, January 4, 1991, p. 1.
22. Investment banks that underwrite commercial paper will usually agree to repurchase
this paper prior to maturity for the expressed purpose of providing liquidity.
23. See Gorton and Pennacchi (1989) for further discussion.
24. The linear specification is ad hoc, but dictated by data limitations as will become
apparent.
25. The data on loan sales yields and LIBOR was collected by a survey of a few money
center banks by an industry publication called Asset Sales Report. Loan sales yields were
classified by maturity and commercial paper rating of the underlying borrower. For details
see Gorton and Pennacchi (1989). The premium for bank risk, b" was calculated using bank
stock prices as described in Gorton and Pennacchi (1989).
26. The guarantees are options, but they are vulnerable options, that is, the bank writing
the option may default. Vulnerable options have different comparative statics than stand-
ard, exchange traded, options. The main point is that an American option can be valued
significantly higher than a European option because the right to exercise early increases in
value when the likelihood of default by the option writer increases. The holder of commercial
paper can only call on the guarantor at the maturity of the paper, whereas the loan sale
holder may be able to exercise at any time. This means that the effective premium that loan
sales buyers place on bank risk, the b, variable, will be quite small relative to commercial
paper buyers. This can explain the negative coefficient obtained on that variable. See Johnson
and Stulz (1987) on vulnerable options.
27. Hence, the overall average of the fraction of loans sold that were retained by the
bank was approximately 24 percent. This data set was obtained somewhat subsequent to the
study by Gorton and Pennacchi (1989). Interestingly, the bank involved was motivated to
provide the data mostly because it was convinced that there was no implicit guarantee. In
the data set, the average maturity of the loans sold was 28.04 days. The average maturity
of the loan sales contracts was 27.63 days. The average loan interest rate was 7.53, while the
average loan sale interest rate was 7.41. Further details on the sample can be found in
Gorton and Pennacchi (1990A).
28. The following OLS regression supports these observations. Letting fr = the fraction
of the loan sale retained by the bank, maturity = the maturity (in days) of the loan, and rate
= {4 if no rating, 3 if A3, 2 if A2, I if AI, and 0 if AI+J, we have:
fr = .17890 + .00151 maturity + .00742 rate

(.02965) (.00046) (.00762)
No. of Obs. = 872. R' = .012. (Standard errors in parentheses.)
Thus, point estimates suggest the fraction retained increases with maturity and the risk of
the rating, but only the maturity variable is significant.
29. When the fraction sold is omitted and equation (1) is estimated with this data set the
results are:
r" - r, = .00061 + .2355 (rb - r,) + .0068 b
(.00004) (.0120) (.0512)
No. of Obs. = 872. R' = .311. (Standard errors in parentheses.)
Thus, the loan sale spread is, as expected, significantly positively related to the spread paid
by the borrowing firm. But, the coefficient on the probability that the selling bank will fail
is insignificantly different from zero, though it is of the expected positive sign.
30. The reader is referred to Gorton and Pennacchi (1990A) for the details of the model.
Our purpose here is only to lay bare the issues by summarizing the model.
31. This problem is independent of the bank's choice of the number of loans to orig-
inate. See Pennacchi (1988) for an analysis of the initial portfolio choice.
32. This means that the loan sale contract is restricted to a proportional equity split
between the bank and the loan buyer. For example, senior/subordinated positions are not
allowed. The assumption is consistent with the observed contracts in this market. Presumably,
regulatory constraints prevent other contract forms. These constraints include the requirements
for taking the loan off the balance sheet, prohibitions on explicit guarantees, and securities
laws.
33. The cost of implicit guarantees is assumed to take the form of pressure by regulators.
34. The parameters a and /3 are assumed to be positive and loan specific. The parameter
a is also assumed to be less than unity. Note that if no monitoring is done, the expected
return is L(1 - a). The parameter /3 is a measure of the marginal increase in expected return
on the loan from additional monitoring. Note that the assumed function implies that as the
level of monitoring rises towards infinity, the expected return on the loan asymptotically
approaches the promised payment, L, with the speed determined by the parameter /3.
35. The model is estimated in log form by assuming that the log of the fraction of a loan
sold is equal to the log of the right-hand side of equation (6) plus a normally distributed
error term. The model does not explicitly account for the noise term. The idea is that this
noise term can capture the influence of missing factors, assumed to be uncorrelated with the
right-hand side of (6). Note that since the fraction of the loan sold, s, takes on a value that
is at most equal to 1 (and hence the In(s) is at most zero), we estimate the model as a non-
linear Tobit model since the model is censored (the right-hand side of (6) is observed while the
left-hand side is not) when s = 1. The likelihood function is given in Gorton and Pennacchi
(1990A).
36. This is done by solving for y using equations (4) and (6). The reader is referred to
Gorton and Pennacchi (1990A) for details.
37. See Becketti (1990) for more description of recent developments in the junk bond
market.
38. These figures are from United States Securities and Exchange Commission (1990).
For purposes of comparison, statistics from the Federal Reserve Bulletin show that individual
investors held more that 10 percent of all outstanding Treasury securities as of year-end 1988.
39. See Gorton and Pennacchi (1990B,C) for the theoretical rationale and empirical
evidence of a shift from bank financing to direct financing.
40. Since estimation was in log form, the right-hand side of (7) was transformed into two
additive terms. The second term's coefficient estimate was of the wrong sign (negative) but
insignificantly different from zero (its theoretical value under the hypothesis of full ob-
servability). The first term had a coefficient estimate that was of the correct sign (positive)
and significantly different from zero (.37 with a standard error of approximately .06) but
also significantly different from its theoretical value of unity.
41. There is, however, abundant evidence that the demand for bank provision of these
services has fallen. The rise of the commercial paper market and the medium term note
market suggest that the same technological forces which make loan sales feasible have
allowed directly marketable instruments to compete more effectively with bank loans.
References
Becketti, Sean. 1990. "The Truth about Junk Bonds," Federal Reserve Bank of
Kansas City Economic Review 75(4), (July/August): 45-54.
Benveniste, Lawrence and Allen Berger. 1987. "Securitization With Recourse: An
Investment that Offers Uninsured Bank Depositors Sequential Claims," Journal
of Banking and Finance 11, (September): 403-24.
Boyd, John and Edward Prescott. 1986. "Financial Intermediary-Coalitions," Journal
of Economic Theory 38, 211-32.
Calomiris, Charles and Gary Gorton. 1990. "The Origins of Banking Panics: Models,
Facts, and Bank Regulation," in Financial Markets and Financial Crises, ed. by
Glenn Hubbard (University of Chicago Press).
Campbell, Tim and William Kracaw. 1980. "Information Production, Market
Signalling, and the Theory of Financial Intermediation," Journal of Finance 35(4)
(September), 863-82.
Cumming, Christine. 1987. "The Economics of Securitization," Quarterly Review
(Autumn), Federal Reserve Bank of New York, 11-23.
Diamond, Douglas. 1984. "Financial Intermediation and Delegated Monitoring,"
Review of Economic Studies LI (July): 393-414.
Fama, Eugene. 1985. "What's Different About Banks?," Journal of Monetary
Economics 15 (January): 29-39.
Flannery, Mark. 1989. "Capital Regulation and Insured Banks' Choice Individual
Loan Default Risks," Journal of Monetary Economics 24, 235-58.
Gorton, Gary and Joseph Haubrich. 1990. "The Loan Sales Market," Research in
Financial Services, ed. by George Kaufman, Vol. 2, JAI Press Inc: Greenwich
CT.
- - . 1987. "Bank Deregulation, Credit Markets, and the Control of Capital,"
Carnegie-Rochester Conference Series on Public Policy 26 (Spring): 289-333.
Gorton, Gary and George Pennacchio 1989. "Are Loan Sales Really Off-Balance
Sheet?," Journal of Accounting, Auditing, and Finance 4(2), 125-45.
- - . 1990A. "Banks and Loan Sales: Marketing Nonmarketable Assets,"

unpublished working paper.
- - . 1990B. "Financial Intermediation and Liquidity Creation," Journal of Finance
45(1),49-72.
--.1990C. "Financial Innovation and the Provision of Liquidity Services," Reform
of Deposit Insurance and the Regulation of Depository Institutions in the 1990s,
ed. by James Barth and Dan Brumbaugh. Forthcoming.
Greenbaum, Stuart and Anjan Thakor. 1987. "Bank Funding Modes: Securitization
versus Deposits," Journal of Banking and Finance 11, (September): 379-401.
James, Christopher. 1988. "The Use of Loan Sales and Standby Letters of Credit
by Commercial Banks," Journal of Monetary Economics 22, (November): 395-
422.
- - . 1987. "Some Evidence on the Uniqueness of Bank Loans," Journal of Financial
Economics 19, (December): 217-35.
Johnson, Herb and Rene Stulz. 1987. "The Pricing of Options With Default Risk,"
Journal of Finance 42(2), (June): 267-80.
Kareken, John. 1987. "The Emergence and Regulation of Contingent Commitment
Banking," Journal of Banking and Finance 11,359-77.
Lummer, Scott and John McConnell. 1989. "Further Evidence on the Bank Lending
Process and the Capital Market Response to Bank Loan Agreements," Journal
of Financial Economics 25, 99-122.
Pennacchi, George. 1988. "Loan Sales and the Cost of Bank Capital," Journal of
Finance 43, 375-95.
Smith, Clifford. 1986. "Investment Banking and the Capital Acquisition Process,"
Journal of Financial Economics 15,3-29.
U.S. Securities and Exchange Commission. 1990. "Recent Developments in the
High Yield Market," Staff Report, March 15, 1990.
COMMENTARY
Stuart I. Greenbaum
Gorton and Pennacchi provide a service by directing the attention of

academe to loan sales and securitization, arguably one of the more signifi-
cant recent developments in the practice of financial intermediation,
particularly in the United States where these innovations were pioneered.
Of course, this same recent period was characterized by explosive growth
in Euromarkets, commercial paper, junk bonds, financial options, and
futures as well. Moreover, there is nothing terribly new about banks selling
loans.
Nevertheless, the documented growth of loan sales, and securitization
to a lesser degree, is stunning; it has fostered the decomposition of the
credit transaction into its more elemental components. This has facilitated
specialization among providers of financial services with unmistakably
positive implications for the efficiency of the intermediation process.
A compelling way to begin a research agenda is to ask the question
why. What explains the striking growth of this phenomenon at this par-
ticular time and place? In addressing this question, it seems to me that the
authors fall into a paradigmatic snare. As generally known, if there is
anything to be explained about banking, capital markets, or financial inter-
mediation more generally, the answer will be found in an information
35
asymmetry. If it is not a moral hazard, then surely it is a self-selection

problem.
From the authors' perspective the earlier absence of loan sales was
attributable to an insurmountable moral hazard-the loan originator's
failure to exercise the same due diligence and monitoring that it would
provide as funder of the assets. The more recent growth of loan sales
must then be attributable to some mitigation of the moral hazard. Three
possibilities are considered. First, by retaining a fraction of the loan, the
originator may align its own incentives with those of the loan purchaser.
Second, the originator may provide implicit guarantees against losses on
loans sold. Finally, the authors consider the possibility that technological
advances may have reduced informational asymmetries, thereby weaken-
ing the earlier basis for principal/agent type moral hazard.
The three hypotheses are lucidly articulated and then confronted with
sketchy data relating to loan sales and securitization. While the latter
exercise can only be described as disappointing, this conclusion should
not be considered a criticism of the authors. But nothing seems to matter
very much, and the authors tentatively conclude that the technological
advances hypothesis seems worthy of respect because the data fail to
support the alternative two hypotheses. I doubt that this is a very robust
conclusion and the authors, to their credit, seem to share that misgiving.
However, if one is precommitted to an information argument, the Gorton!
Pennacchi conclusion could have been inferred without the benefit of the
data manipulation. After all, there was nothing new in the idea of re-
taining a fraction of the loan or of granting an implicit guarantee. Hence,
these provide no basis for temporally anchoring the observed loan sales
growth. The only possibility is that advances in technology overwhelmed
the inhibiting influences of informational asymmetries.
The more important point, however, is that the growth of loan sales
may have had nothing to do with informational asymmetries, and this is
a possibility the authors strain to ignore, (note 5 notwithstanding). For
example, suppose we have two more or less autarkic countries that could,
but for whatever reason choose not to, trade very much. And further
suppose that these two countries have approximately the same savings
rates. Now for demographic reasons, the savings rate in one country triples
and in the other it falls in half. What sort of responses would we expect?
Presumably, asset prices in the high savings country would rise and real
interest rates would fall. The obverse would be expected in the low sav-
ings country. One might, under the circumstances, expect investors from
the high savings country to seek out assets in the low savings country
because of their relatively attractive prices. However, one impediment to
the purchase of debt by investors from the high savings country is their
lack of asset-originating capability in the low savings country. It would be
plausible for the investors from the high savings country to engage agents
(brokers) whose role it would be to find attractive assets for the high
savers to purchase. To be sure, the high savings country investors will not
be oblivious to the possibility of being exploited by myopic or cynical
brokers. The investors will, therefore, be attentive to the brokers' repu-
tations, and they may seek assurances, both written and implied. However,
the larger the spread between the home- and foreign-country interest
rates, the more compelling the foreign investment, despite the inhibiting
effects of private information. It is then possible that nothing changed in
the informational sphere and loan sales might have exploded (owing to
nothing more profound than changing population age profiles and savings
proclivities) !
Likewise, consider two countries that have chosen to subsidize their
banks by restricting entry, suppressing deposit interest rates, and under-
pricing governmental deposit insurance. It does not really matter why the
two countries might have chosen to adopt such a policy; only the fact that
deposits are subsidized matters. The deposit subsidy will prompt banks to
accept all proffered deposits, and we may even observe banks competing
in oblique ways to attract deposits, quite irrespective of investment op-
portunities. With sufficiently generous subsidies, the banks could profit-
ably hold risk-free assets, exclusively. In such an environment we would
expect the banks to develop asset origination skills in order to deploy the
proceeds of deposit liabilities.
Now, suppose that, for whatever reason, the deposit subsidies in one of
the countries disappears. Instead of being able to borrow at some sub-
risk-free interest rate, the impacted banks would now face market interest
rates. To make matters really dreadful, suppose that these suffering banks
are downgraded by rating agencies in light of a perceived excess supply
of banks and bankers. After all, the size of the industry was previously
bloated by the subsidies directed to banks.
Would it then seem reasonable for the impacted banks to sense an
excess capacity to originate assets relative to their funding capabilities?
Would it not seem reasonable then to originate assets for sale to others,
perhaps to banks in the country still enjoying deposit subsidies? What, if
anything, would such sales have to do with informational asymmetries?
To be sure, informational asymmetries will impede the asset sales, but in
what sense do the asymmetries explain the growth of loan sales?
I have one more question regarding the Gorton/Pennacchi paper. It is
correctly noted that public regulation of banks is related to the traditional
illiquidity of bank assets. Indeed, this is a defining characteristic of tra-

ditional banks: they are repositories of illiquid financial claims. Their raison
d'etre was to swap liquid for illiquid claims with a clientele displaying a
preference for the former.
In explaining the nexus between illiquidity and public regulation, how-
ever, the authors stress the role of deposit insurance. The link is more
basic than that. Illiquidity of bank assets gives rise to the need for a
lender of last resort. But, the introduction of the lender of last resort
weakens the bank's incentive to hold cash assets, and this shifts deposit
seigniorage from the public sector to the privately-owned banks. This
moral hazard is solved with the introduction of cash-asset reserve re-
quirements, and we consequently have endogenous regulation, the inevit-
able accompaniment of the lender of last resort.
In closing, I applaud the Gorton/Pennacchi effort. It directs attention
to an important phenomenon and it should excite scholarly interest. This
is presumably how we learn! My major reservation, as already indicated,
relates to over-reliance on the information paradigm. To be sure, infor-
mation arguments have been illuminating, and therefore seductive, espe-
cially in banking. But informational asymmetries should not be expected
to explain everything, for then the paradigm becomes a toy instead of a
tool.
II
2 INTERSTATE BANKING, BANK
EXPANSION AND VALUATION
Gerald A. Hanweck
The field cannot well be seen from within the field.

-Ralph Waldo Emerson
Introduction
After a decade of deregulation in banking, the field remains unseen. There

is in place a haphazardly developed set of state laws regulating interstate
banking, Federal law that is piecemeal and that is being eroded by the
liberalization of banking powers by the courts and Federal Reserve, and
an expansion by banking companies and others via the acquisitions of
failed thrifts and banks, that further undermil'es law and regulation.
Additionally, there is no public policy addressing Loncentration of bank-
ing resources through acquisitions, which singly pose no anticompetitive
threat, but collectively may lead to significant entry barriers and threats
to competition in the future. To date, there appears to be a rapid, but
strained, pace toward geographic expansion by larger banking companies;
however, as this article points out, had these companies and the banking
industry been more profitable over the past decade to fuel expansion, the
number of banks may have declined substantially to the point of concern
for regional and local market competition for banking services.
With the expansion of interstate banking opportunities in recent years,
41
more than forty-seven states and the District of Columbia have permitted
bank holding companies headquartered in other states to own banks in
those states. The remaining states with no interstate banking statutes are
expected to enact such legislation soon. Since geographic banking oppor-
tunities have been altered, the recent changes in state laws regarding
interstate banking should be expected to affect the valuation of commercial
banks. To the extent that such dramatic changes in permissible bank
expansion affect valuation, it is likely that such effects would vary with
the location, nature of banks, and the type of state law change. As this
article states, these changes in state law are not uniformly reflected in
bank valuation changes.
Even though the expansion by the major banking companies has been
constrained, much interstate expansion of banking services has taken place
since the 1970 amendments to the Bank Holding Company (BHC) Act
via nonbank subsidiaries of bank holding companies (section 4(c)(8) of
the Bank Holding Company Act as amended) and loan production offices
of banks. By the end of 1988 there were 6,778 interstate nonbank offices
of banking companies and 7,492 domestic and foreign interstate banking
offices.' In many respects, the presence of a nonbank office, such as a
mortgage banking company, can be just as effective as the presence of a
full-service bank, since many wholesale and retail customers are shopping
for a particular type of credit arrangement or other financial service. The
major handicap to the lack of a full-service banking office is the solicitation
and provision of many deposit services. The proximity of a bank for cus-
tomers to make deposits and receive cash is important for retailers and
households, and special service offices and automated teller machines
(ATMs) are not convenient substitutes. Thus, even though there has been
considerable geographic nonbank expansion by BHCs, this is not an
effective substitute for interstate banking. It is likely, then, that the trend
will be Lor banking companies to expand geographically through full-
service banking company acquisitions rather than nonbank subsidiaries
of the holding company or bank. Also likely, is that the trend will accel-
erate by early 1991 when approximately thirty-two states adopt some
form of national interstate banking, with twelve of these having no recip-
rocal restrictions. 2
As will be developed in a discussion of the history and state of inter-
state banking, the movement toward full-scale interstate banking has been
hampered by state legislation that has attempted to protect the market
value of existing banking companies in the states. By early 1991, thirty-
two states will have some form of national interstate banking with twelve
of these having no reciprocal restrictions. As a result, almost any large
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 43
banking company with sufficient resources can acquire a bank in these

states. Though the development of interstate banking is poised to accelerate,
the profitability of many of the "superregional" banks, having recently
grown rapidly through interstate expansion, is deteriorating. The result is
that these banks, like their larger money-center counterparts, will not
have the profitability, resources, and capitalization to expand in the near
future? Consequently, only a handful of banking companies, including
some foreign banks, will be able to expand geographically to any great
extent and much of this expansion is likely to be by way of acquisitions
of failing banks and thrifts. With this as a prospect, it is imperative that
federal legislation be enacted to eliminate McFadden Act provisions to
interstate branching and the Douglas Amendment to the BHC Act. This
Act also should be thoroughly amended to encourage interstate banking
through branching and acquisition and to promote a long-run competitive
environment through transition rules limiting local market and state shares
and, through existing regulation, excessive expansion.
The second section of this article summarizes and reviews the legal
background and the state of interstate banking. The third section considers
the public policy and competitive issues surrounding interstate banking
and reviews some of the relevant literature and studies that have addressed
these areas. In the fourth section is developed hypothesized reactions
of banks, in terms of profitability and valuation, to changes in interstate
banking legislation when they are within the affected state or region and
outside a state or region enacting interstate banking legislation. This section
reports on several recent studies that have attempted to evaluate interstate
banking law effects on the profitability and valuation of banking companies.
The fifth section will consider the view that banking profitability will be
the constraining factor in the future development of interstate banking.
This section will assess the proposition that banking resources will be so
limited and the overhang of the real estate market recession will be so
substantial that bank expansion geographically or into alternate product
lines will be significantly hampered for the next several years. The final
section expresses some conclusions and suggestions for further research
and public policy toward interstate banking. As suggested, the implications
of balkanized legislation, and a Congress preoccupied with the growing
costs of the savings and loan (S&L) bailout crisis, continue to hamper
the development of interstate banking and expanded product lines and
will contribute to the deteriorating profitability of banking companies,
which will only serve to encourage bankers to seek profitable investment
opportunities elsewhere such as in the new markets emerging in Europe
after 1992.
The History and State of Interstate Banking
Prior to the mid-1960s, banks had little incentive to expand beyond state
boundaries. However, changing economic and technological conditions
have stimulated commercial banks to reconsider their perceived optimal
structural organizations, and many of the larger banks have attempted to
circumvent the legal restrictions on interstate banking by way of nonbanking
affiliates of bank holding companies. 4 State banks are chartered to oper-
ate within a single state. The introduction of national banks by the Na-
tional Banking Act of 1863, and subsequent legislation and interpretations,
prohibited branching by national banks. The McFadden Act of 1927 and
the Glass-Steagall Act of 1933 allowed national banks branching oppor-
tunities equal to that permitted state-chartered banks. This legislation
eliminated, in effect, branching across state lines.
Through the bank holding company organizational form, commercial
banks have been able to circumvent the legal restrictions on interstate
branching. Prior to 1956, bank holding companies were unregulated and
some banks used this vehicle to establish a multistate presence. The Glass-
Steagall Act of 1933 provided for limited regulation of bank holding
companies by the Federal Reserve System, but did not limit interstate
expansion.
The Bank Holding Act of 1956, however, prohibited further establish-
ment of interstate banking offices but grandfathered the seven existing
domestic interstate banking organizations at that time. Foreign banking
organizations were permitted to establish full banking operations across
state lines until the International Banking Act of 1978 attempted to equalize
opportunities available for foreign and domestic banks. Though prohib-
ited from expanding full-service offices across state lines, this legislation
permitted interstate foreign banks to retain existing networks. In addition
to banking companies, there are several types of limited service facilities,
such as Edge Act Corporations, loan production offices, foreign bank
branches and agencies, non bank banks, and savings and loans,S which have
been permitted on an interstate basis.
Section 3( d), the Douglas Amendment, of the Bank Holding Company
Act of 1956 permitted states the power to legislate out-of-state bank holding
companies to acquire, but not charter, and operate banks within each
state. Seven domestic bank holding companies and five foreign banking
companies were permitted to continue their interstate banking activities
through a "grandfather" provision of this Act. Maine was the first state
to pass a law permitting interstate acquisitions in 1975, later liberalized
in 1978 to national reciprocal, and in 1984 to open entry by banks
headquartered in any state. No other states followed Maine's lead until

1982 when New York and Alaska enacted liberalized interstate banking
legislation. In rapid succession many states followed so that currently
forty-seven states and the District of Columbia have some type of law
permitting interstate banking through bank holding companies (see table
2-1). As of August 1990, assets controlled by banks from other states are
15.2 percent of total banking assets nationwide with control as high as
over 84 percent in Maine and 79 percent in Washington. By early 1991,
thirty-two states will have effective national interstate banking laws, of
which twelve will have no reciprocal restrictions (see tables 2-1 and 2-2).
The 1970 Amendments to the Bank Holding Company Act redefined
a BHC to include one-bank holding companies and also established
procedures to determine permissible nonbanking activities for BHCs. Prior
to passage of this law, many of the nation's largest banking companies
had established one-bank holding companies that were exempt from cov-
erage of the 1956 BHC Act and had entered into nonbanking activities
which were of questionable legality if done through the bank and, more
importantly, often crossed state lines.
As technology, communication, and transportation have improved within
the past thirty years, it has become more advantageous for banks to conduct
their business across state lines. Other companies that appear to compete
directly with commercial banks in some services, such as securities firms
(for example, Merrill Lynch's cash management account), finance companies
(GMAC), and retail stores (Sears), are not limited by state lines and can
follow their markets to any geographic location. Commercial banks have
not, as discussed above, been able to establish offices that accept deposits
and make commercial loans across state boundaries. In response to the
changing financial marketplace environment and in order to compete more
effectively with their nonbank financial rivals, commercial banks have
found a number of ways to maintain operations in more than one state.
Two of these ways have been the grandfathered BHCs and the nonbanking
BHC subsidiaries. The current status of all the major exceptions to the
interstate banking restriction can be classified into three categories: (1)
full-service bank-charter offices; (2) limited service (nonbank bank) banking
offices; and (3) nonbanking offices.
Since the focus of this paper is on interstate banking in a full-service
office context, we consider more prominently the first group of exceptions.
The seven interstate domestic holding companies grandfathered by the
Douglas Amendment are First Interstate Bancorp (formerly Western
Bancorporation and once part of Bank of America), First Bank System,
Northwest Bancorporation (now Norwest), Otto Bremer Foundation, Credit
Table 2-1. Interstate Banking Legislation by State (As of September 5, 1990).
State Legislation in Effect Area
Alabama Currently Reciprocal. 13 States and DC
(AR, FL, GA KY, LA, MD,
MS, NC, SC, TN, TX, VA, WV,
DC).
Alaska Currently National, no reciprocity.
Arizona Currently National, no reciprocity.
Arkansas Currently Reciprocal. 16 States and DC
(AL, FL, GA, KS, LA, MD,
MO, MS, NC, NE, OK, SC, TN,
TX, VA, WV, DC).
California Currently Reciprocal. 11 States (AK, AZ,
CO, HI, ID, NV, NM, OR, TX,
UT, WA).
January 1, 1991 National, reciprocal.
Colorado Currently Reciprocal. 7 States (AZ, KS,
NE, NM, OK, UT, WY).
January 1, 1991 National, no reciprocity.
Connecticut Currently National, reciprocal.
Delaware Currently National, reciprocal.
District of Columbia Currently Reciprocal. 11 States (AL, FL,
GA, LA, MD, MS, NC, SC,
TN, VA, WV).
Florida Currently Reciprocal. 11 States and DC
(AL, AR, GA, LA, MD, MS,
NC, SC, TN, VA, WV, DC).
Georgia Currently Reciprocal. 10 States and DC
(AL, FL, KY, LA, MD,
MS, NC, SC, TN, VA, DC).
Idaho Currently National, no reciprocity.
Illinois Currently Reciprocal. 6 States (lA, IN,
KY, MI, MO, WI).
December 1, 1990 National, reciprocal.
Indiana Currently Reciprocal. 12 States (lA, IL,
KY, MI, MN, MO, OH, PA,
TN, VA, WI, WV).
July 1, 1992 National, reciprocal.
Iowa January 1, 1991 Reciprocal. 6 States (IL, MN,
MO, NE, SD, WI).
Kentucky Currently National, reciprocal.
Louisiana Currently National, reciprocal.
Maine Currently National, no reciprocity.
Table 2-1. (Cont.)

Maryland Currently Reciprocal. 14 States and DC
(AL, AR, DE, FL, GA, KY,
LA, MS, NC, PA, SC, TN, VA,
WV,DC).
Massachusetts Currently National, reciprocal.
Michigan Currently National, reciprocal.
Minnesota Currently Reciprocal. 14 States (CO, lA,
ID, IL, IN, KS, MO, MT, ND,
NE, SD, WA, WI, WY).
Mississippi Currently Reciprocal. 13 States (AL, AR,
FL, GA, KY, LA, MO, NC, SC,
TN, TX, VA, WV).
Missouri Currently Reciprocal. 8 States (AR, lA,
IL, KS, KY, NE, OK, TN).
Nebraska Currently Reciprocal. 10 States (CO, lA,
KS, MN, MO, MT, ND, SD,
WI, WY).
Nevada Currently National, no reciprocity.
New Hampshire Currently National, no reciprocity.
New Jersey Currently National, reciprocal.
New Mexico Currently National, no reciprocity.
New York Currently National, reciprocal.
North Carolina Currently Reciprocal. 13 States and DC
(AL, AR, FL, GA, KY, LA,
MD, MS, SC, TN, TX, VA,
WV, DC).
Ohio Currently National, reciprocal.
Oklahoma Currently National. After initial entry,
BHC must be from state
offering reciprocity or wait 4
years to expand.
Oregon Currently National, no reciprocity.
Pennsylvania Currently National, reciprocal.
Rhode Island Currently National, reciprocal.
South Carolina Currently Reciprocal. 12 States and DC
MD, MS, NC, TN, VA, WV,
DC).
South Dakota Currently National, reciprocal.
Tennessee Currently Reciprocal. 14 States and DC
Table 2-1. (Cont.)

(AL, AR, FL, GA, IN, KY, LA,
MD, MO, MS, NC, SC, VA,
WV, DC).
Texas Currently National, no reciprocity.
Utah Currently National, no reciprocity.
Vermont Currently National, reciprocal.
Virginia Currently Reciprocal. 12 States and DC
MD, MS, NC, SC, TN, WV,
DC).
Washington Currently National, reciprocal.
West Virginia Currently National, reciprocal.
Wisconsin Currently Reciprocal. 8 States (lA, IL, IN,
KY, MI, MN, MO, OR).
Wyoming Currently National, no reciprocity.
Note: Several states prohibit acquisition of banks in operation for less than a specified
number of years.
Source: Financial Structure Section, Board of Governors of the Federal Reserve System.
and Commerce American Holdings (formerly Financial General

Bancshares), General Bancshares Corporation, and First Security Corpo-
ration (see table 2-3). These organizations have full-service banking
offices in twenty-four states with a total of over two hundred banks as of
1983 (since then some have consolidated within a state). Three of these
BHCs are headquartered in Minnesota and one each in California, Mis-
souri, Utah, and the Netherland Antilles (formerly in the District of
Columbia). As of 1982, when most states began to change their interstate
banking legislation, the extent of full-service banking geographic opera-
tion is shown in table 2-3. This table also indicates all other banking
organizations with subsidiary banks in more than one state and those that
had been approved.
The next group of BHCs in table 2-3 are the foreign banking organ-
izations with interstate operations. Prior to 1978, foreign banks were not
subject to the interstate restrictions affecting domestic organizations. In
order to equalize competitive advantages, the International Banking Act
of 1978 prohibited foreign banks from expanding interstate in ways not
permitted for domestic banks, but allowed them to keep existing interstate
networks. The foreign BHCs listed in table 2-3 are those retaining full-
service banks in multiple states, but note that there are numerous other
banks which control various types of entities in multiple states.
The final group of banks in table 2-3 are those that had expanded
interstate as state laws had begun to change. Since 1982, many states for
example, South Dakota and Maryland, have permitted out-of-state BHCs
to establish banks with limited powers. These banks with limited powers
serve customers of their parent organizations in many states, but the
limitations on their powers keep them from competing with local banks
for a full range of banking services. The primary objective of the states in
permitting entry by the out-of-state BHCs was to attract capital and jobs
to the states. Some states have actively competed to attract large banks.
Money-center banks have pushed for further relaxation of these restric-
tions, and have succeeded in getting reciprocity legislation passed in
several states.
The expansion of banks interstate via chartered banks has been rapid
since 1982. In 1983 there were twenty-six domestic banking companies
operating i full-service banks interstate. Four years later, the Federal
Reserve reported that fifty-one banking organizations had subsidiary
banks in one or more states beside their home state (Savage 1987). Today
this number is approximately 180 banking companies after accounting for
interstate acquisitions since 1987. This number includes expansion by
acquisition of failing banks. This rapid expansion is also reflected in the
number of banking offices controlled by interstate banking companies.
Domestic banking companies with interstate banking operations have
increased the number of banking offices outside their home states under
their control from 1,258 in 1983 to 7,364 in 1988 (see table 2-4). In terms
of banking assets, these banking companies had $442.4 billion in interstate
bank assets as of March 31, 1990, which represents an increase from $148.4
billion as of June 30,1986. These data indicate nearly a tripling of banking
assets associated with interstate banking within the past four years and
nearly a six-fold increase in banking offices since 1983. Both these figures
suggest a very rapid expansion of interstate banking via full-service banks.
Some Explanations For the Rapid Spread of Interstate

Banking
The rapid expansion in interstate activity has explanations in fundamental

economic reasons for banking managements' desire to expand their mar-
kets. A review of the states with the greatest interstate banking activity
in terms of asset or office growth reveals that Florida, a state with a rapid
growth in banking assets, has shown a remarkable increase in interstate
VI
0 Table 2-2. Interstate Banking Legislation by State (As of February 1, 1989).
Number of
State Effective Date Area Partner States
Alabama Currently Reciprocal, 12 states and DC (AR, FL, GA, KY, LA, MD, 13
MS, NC, SC, TN, VA, WV).
Alaska Currently National, no reciprocity. SO
Arizona Currently National, no reciprocity. SO
Arkansas Currently Reciprocal, 16 states and DC (AL, FL, GA, KS, LA, MD, MS, 17
MO, NE, NC, OK, SC, TN, TX, VA, WV). Reciprocity hinges
on commitments to community reinvestment.
California Currently Reciprocal, 11 states (AK, AZ, CO, HI, ID, NV, NM, OR, 11
TX, UT, WA).
Jan. 1, 1991 National, reciprocal.
Colorado Currently Reciprocal, 7 states (AZ, KS, NE, NM, OK, UT, WY). 7
Connecticut Currently Reciprocal, 5 states (MA, ME, NH, RI, VT). 5
Delaware Currently Reciprocal,S states and DC (MD, NJ, OH, PA, VA). 6
Special-purpose banks permitted.
June 30, 1990 National, reciprocal.
District of Columbia Currently Nationwide, no reciprocity if community development SO
commitments are made.
Florida Currently Reciprocal, 11 states and DC (AL, AR, GA, LA, MD, MS, 12
NC, SC, TN, VA, WV).
Under a 1972 law, NCNB and Northern Trust Corporation are
grandfathered and can make further acquisitions.
Georgia Currently Reciprocal, 10 states and DC (AL, FL, KY, LA, MD, MS, NC, 11
SC, TN, VA).
Hawaii None 0
Idaho Currently National, no reciprocity. 50
Illinois Currently Reciprocal, 6 states (lA, IN, KY, MI, MO, WI). 6
Nationwide, organizations may acquire failed institutions if the 6
failed institution is larger than $1 billion in assets.
Under a 1981 law, General Bancshares Corporation is
grandfathered and can make further acquisitions in the state.
Dec. 1, 1990 National, reciprocal.
Indiana Currently Reciprocal, 11 states (lA, IL, KY, MI, MO, OH, PA, TN, VA, 11
WI, WV).
July 1, 1992 National, reciprocal.
Iowa 1972 Under a 1972 law, Norwest Corporation is grandfathered and 0
is permitted to acquire banks in Iowa.
Kansas None 0
Kentucky Currently National, reciprocal. 31*
Louisiana Currently National, reciprocal. 29*
Maine Currently National, no reciprocity. 50
Maryland Currently Reciprocal, 14 states and DC (AL, AR, DE, FL, GA, KY, LA, 15
MS, NC, PA, SC, TN, VA, WV) and special-purpose banks.
Massachusetts Currently Reciprocal, 5 states (CT, ME, NH, RI, VT). 5
Michigan Currently National, reciprocal. 20*
Minnesota Currently Reciprocal, 11 states (CO, IA, ID, IL, KS, MO, MT, ND, SD, 11
WA, WY).
Mississippi Currently Reciprocal, 4 states (AL, AR, LA, TN). 4
July 1, 1990 Reciprocal, 13 states (AL, AR, FL, GA, KY, LA, MO, NC,
SC, TN, TX, V A, WV).
VI
......
VI
IV
Table 2-2. (Cont.)

Number 0/
State Effective Date Area Partner States
Missouri Currently Reciprocal, 8 states (AR, lA, IL, KS, KY, NE, OK, TN). 8
Montana None o
Nebraska Currently Special-purpose banks. 0
Jan. 1, 1990 Reciprocal, 10 states (CO, lA, KS, MN, MO, MT, NO, SO,
WI, WY).
Nevada Currently National, no reciprocity. 50
New Hampshire Currently Reciprocal,S states (CT, MA, ME, RI, VT). 5
New Jersey Currently National, reciprocal. 21 *
New Mexico Currently Nationwide acquisition of failing banks. 50
Jan. 1, 1990 National, no reciprocity.
New York Currently National, reciprocal. 19*
North Carolina Currently Reciprocal, 12 states and DC (AL, AR, FL, GA, KY, LA, 13
MD, MS, SC, TN, VA, WV).
North Dakota Currently A grandfathered interstate banking organization is permitted to 0
sell its North Dakota banks to out-of-state bank holding
companies.
Ohio Currently National, reciprocal. 23*
Oklahoma Currently National, no reciprocity. 50
Oregon Currently 8 states, no reciprocity (AK, AZ, CA, HI, 10, NV, UT, W A). 8
July 1, 1989 National, no reciprocity.
Pennsylvania Currently Reciprocal,7 states and DC (DE, KY, MD, NJ, OH, VA, 8
WV).
March 4,1990 National, reciprocal.
Rhode Island Currently National, reciprocal. 23*
South Carolina Currently Reciprocal, 12 states and DC (AL, AR, FL, GA, KY, LA, 13
MD, MS, NC, TN, VA, WV).
South Dakota Currently National, reciprocal and special-purpose banks. 21*
Tennessee Currently Reciprocal, 13 states (AL, AR, FL, GA, IN, KY, LA, MO, 13
MS, NC, SC, VA, WV).
Texas Currently National, no reciprocity. 50
Utah Currently National, no reciprocity. 50
Vermont Currently Reciprocal,S states (CT, MA, ME, NH, RI). 5
Feb. 1, 1990 National, reciprocal.
Virginia Currently Reciprocal, 12 states and DC (AL, AR, FL, GA, KY, LA, 13
MD, MS, NC, SC, TN, WV).
Washington Currently National, reciprocal. Failing institutions may be acquired by 21*
organizations from any state.
West Virginia Currently National, reciprocal. 29*
Wisconsin Currently Reciprocal, 8 states (lA, IL, IN, KY, MI, MN, MO, OH). 8
Wyoming Currently National, no reciprocity. 50
* Does not count the two states where nationwide entry by acquisition of failing banks is possible.
Source: Federal Reserve Bank of Atlanta, Economic Review (May/June): 35-36, 1989.
VI
W
Ul
.j::o.
Table 2-3. Bank Holding Companies with Subsidiary Banks in More Than One State (District of Columbia and Puerto Rico
included as states for this table).
Name BHC Location States in Which Banks are Located

First Interstate Bancorp. Los Angeles, CA AZ, CA, CO, ID, MT, NM, NV, OR, UT, WA, WY
First Bank System, Inc. Minneapolis, MN MN, MT, ND, SD, WI
Northwest Bancorporation Minneapolis, MN lA, MN, MT, NE, ND, SD, WI
Otto Bremer Foundation St. Paul, MN MN,ND, WI
Otto Bremer Company St. Paul, MN
Credit and Commerce American Holdings Curacao, NA DC, MD, NY, TN, VA
Credit & Commerce American Invest. Amsterdam
Financial General Bankshares, Inc. Washington, DC
General Bancshares Corporation St. Louis, MO IL, MO, TN
First Security Corporation Salt Lake City, UT ID, UT
Bank of Montreal Montreal, Canada CA,NY
Canadian Imperial Bank of Commerce Toronto, Canada CA,NY
The Bank of Tokyo, Ltd. Tokyo, Japan CA,NY
Barclays Bank Limited London, England CA,NY
Carclsys Bank International, Ltd. London, England
The Sumitomo Bank, Lts. Osaka, Japan CA,HI
The Royal Bank of Canada Montreal, Canada NY, PUERTO RICO
Banco Central, S.A. Madrid, Spain NY, PUERTO RICO
Citicorp New York, NY DE, NY, SD, (acqd. DE 9/14/82)
J.P. Morgan & Co. New York, NY DE, NY, (acqd. DE 12/21/81)
The Girard Company Bala Cynwyd, PA DE, PA, (acqd. DE 12/30/81)
NCNB Corporation Charlotte, NC FL, NC, (acqd. FL 01108/82)
Chase Manhattan Corporation New York, NY DE, NY, (acqd. DE 02/11/82)
Provident National Corporation Philadelphia, PA DE, PA, (acqd. DE 03110/82)
Northern Trust Corporation Chicago,IL FL, IL, (acqd. FL 04/05/82)
Maryland National Corporation Baltimore, MD DE, MD, (acqd. DE 03115/82)
NEWCO Philadelphia, PA DE, PA, (approved 03131183)
Philadelphia National Corporation Philadelphia, PA DE, PA, (acqd. DE 06/01182)
First Maryland Bancorp. Baltimore, MD DE, MD, (acqd. DE 05/10/82)
Equitable Bancorporation Baltimore, MD DE, MD, (acqd. DE 06111182)
Chemical New York Corporation New York, NY DE, NY, (acqd. DE 10/01182)
Manufacturers Hanover Corporation New York, NY DE, NY, (acqd. DE 01103/83)
Pittsburgh National Corporation Pittsburgh, PA DE, PA, (acqd. DE 01120/83) PNC Financial Corp to
acquire HC 0465
Mellon National Corporation Pittsburgh, PA DE, PA, (acqd. DE 04/06/83) Mellon to acquire HC 0463
Ranier Bancorporation Seattle, WA AK, WA, (approved 03/29/83)
Norstar Bancorp, Inc. Albany, NY ME, NY, (acqd. ME 06/01183)
First National of Nebraska, Inc. Omaha,NE NE, SD, (approved 04/14/83)
Bank America Corporation San Francisco, CA CA, WA, (approved 06/02/83)
Source: Board of Governors of the Federal Reserve System, Bank Holding Companies and Subsidiary Banks as of December 31, 1982.
Vl
Vl
Table 2-4. Changes in Interstate Banking Presence (1983-88).

Number Number Change in
Reported Reported Number Percent
Type of Office in 1988 in 1983 Reported Change
Bank Offices Controlled by
Domestic Bank: Holding Company 7,364 1,258 6,106 485
Bank Offices Controlled by
Foreign Bank Holding Company 128 148 -20 -14
Total Bank Offices 7,492 1,406 6,086 433
Offices of Foreign Banks 302 241 61 25
Domestic Edge Act Corporations 79 143 -64 -45
Total Offices for Foreign Transactions 381 384 -3 -1
Section 4(c)(8) Offices 6,446 5,500 946 17
Loan Production Offices 332 202 130 64
Total Nonbank Offices 6,778 5,702 1,076 19
Total Offices of Banks 14,651 7,492 7,159 95
Thrift Institutions 1,616 N.A.
Total Interstate Offices 16,267 N.A.
Source: "Interstate Banking Developments in the 1980s," Economic Review, Federal
Reserve Bank of Atlanta (May/June) 1989.
activity. With the exception of Texas, Florida's interstate bank assets at

$55.9 billion as of August 1990 are by far the largest of any state (see
table 2-5). Therefore, states with potential markets and potential merger
partners are the more attractive markets and this should be reflected in
entry by out-of-state banks once formidable legal barriers to entry are
reduced.
Most interstate banking legislation has limited entry by acquisition
of an existing bank rather than by de novo bank formation. This has
stimulated potential acquired banks to choose between being an acquisition
target or an acquirer. The effect has been to encourage out-of-state banks
to bid quickly for those banks they wish to acquire in any given state for
fear that the price might be bid up by in-state banks desiring to position
themselves for an interstate acquisition.
More liberal interstate banking laws for partner states encourage out-
of-state bank entry. States such as Maine and Arizona, which allow nation-
wide, nonreciprocal entry, or states in the Mid-Atlantic or southeastern
compacts (Georgia, Florida, and Maryland) have experienced considerable
out-of-state entry (see tables 2-1, 2-2 and 2-5). Thus, the lower the barriers
to entry, the more entry activity should be expected. Accordingly, as more
states move to nationwide, nonreciprocal interstate banking, the more

vigorous should interstate activity become.
The weakness of banks and thrifts in certain regional economies in the
United States has also encouraged interstate acquisitions. Since Garn-St.
Germain was passed in 1982, federal law has become progressively liberal
in allowing interstate acquisitions of failing banks and thrifts as a means
of reducing the cost of these failures to the Federal Deposit Insurance
Corporation (FDIC). The Southwest, and especially Texas, has seen a
large number of interstate banking acquisitions of failing or failed banks
and thrifts. Between mid-1986 and mid-1990, assets held by out-of-state
banking companies in Texas increased from zero to $68.9 billion (see
table 2-5). This increase, the largest of any state, is substantially due to
acquisitions such as those by BancOne (Ohio) and NCNB (North Carolina)
of large, failing Texas banking companies. As other regional economies
begin to experience recessions and as banks in these areas suffer losses
from the increasingly severe real estate downturn, they will become targets
for acquisition because their market values have fallen (making them
more attractive takeover candidates) and/or the FDIC will encourage
regulatory mergers. Given the size of many of these banking companies
(those in the Northeast), only acquisitions by an out-of-state superregional
bank would be viable. Consequently, interstate bank expansion is likely
to continue at a rapid pace, but with an altered regional focus.
Limited-Service Banks, Nonbank Companies, and Foreign

Bank Interstate Expansion
Banks have been able to offer limited interstate banking services through
various means since the passage of the so-called Edge Act in 1919. An
Edge Act corporation can be established outside a bank's home state to
deal primarily with loans and deposits related to international trade. The
International Banking Act in 1978 liberalized the extent to which Edge
Act corporations of banks could offer loans and deposits to customers,
but the international trade motivation of the relationship was supposed
to be maintained. These institutions grew rapidly in the 1970s and early
1980s, but have recently shown a considerable slow down. Since 1983, the
number of domestic Edge Act corporations has declined from 143 in
eighteen states to 79 in sixteen states, as of 1988 (see table 2-4). This
decline is attributed to both an expansion in full-service banking offices in
many areas via interstate banking and a reduction in the attractiveness of
international trade (see King, et ai., 1989).
Many banks maintain loan production offices outside of their home

state in order to have a lending sales representation in a local area with-
out the need for a large physical presence. A Federal Reserve Bank of
Atlanta survey of large banks reported that these offices have increased
from 202 in 1983 to 332 in 1988 (see table 2-4). One interpretation of this
64 percent increase is that it reflects the continuing need for even the
largest banking companies to have a local presence in order to effectively
service local loan customers. Since deposit solicitation cannot be accom-
plished directly from these offices, they do not serve as an effective means
to raise deposit funds and are not superseded as easily as Edge Act
corporations when full-service banks are acquired in a local market.
Finally, one of the dominant means by which banking companies have
expanded interstate is through nonbank subsidiaries of bank holding
companies. Since the 1970 Amendments to the Bank Holding Company
Act of 1956 were adopted, the Federal Reserve Board has permitted
BHCs to engage in certain nonbank activities pursuant to section 4(c)(8)
of that Act.6 The most prominent of these nonbank activities by assets
have been commercial finance, mortgage banking, consumer finance, and
securities brokerage, accounting for $146 billion in nonbanking assets in
BHC affiliate nonbank activities in 1987 (see Liang and Savage 1990, 5).
To a great extent, the activities of bank holding company nonbank affili-
ates are in areas traditionally conducted by banks: commercial finance,
mortgage banking, consumer finance, and leasing. However, nonbank
affiliates have the advantage of unrestricted interstate operations. Many
BHCs have taken advantage of this fact and since 1970 have expanded
their operations interstate so that by 1983 there were 5,500 offices of
BHC nonbank affiliates associated with 4(c)(8) activities and 6,446 by 1988
(see table 2-4). Much of this growth was concentrated in the activities
previously mentioned.
Even though the recorded expansion of interstate nonbank affiliates
has not been as dramatic since 1983 as interstate bank expansion, the data
may not reflect the full character of interstate nonbank holdings. As sug-
gested by King, et al. (1989), interstate acquisitions of bank holding
companies transfer nonbank ownership to the acquiring company; therefore,
nonbank subsidiaries of the acquired company in the acquired company's
state will not be counted as interstate nonbank companies. Offsetting this
to some extent are those nonbank affiliates in the acquired BHC's state
that will be counted as interstate companies after an interstate BHC ac-
quisition. The net effect has not been determined, but it is likely that
there may be a net underestimation of interstate 4(c)(8) expansion. With
these qualifications, the data in table 2-4 indicate that BHC's use of nonbank
Table 2-5. Assets by State and of Interstate Banks ($ Billions).
1987 1990
Total Total
Interstate Interstate Total %
State Assets Assets Assets Interstate
Alabama 0.0 0.2 36.1 0.4
Alaska 0.7 0.9 4.2 21.3
Arizona 11.7 15.6 27.1 57.5
Arkansas 0.0 0.1 19.6 0.6
California 0.0 0.4 291.3 0.1
Colorado 2.6 4.2 25.9 16.4
Connecticut 11.0 24.5 36.8 66.5
Delaware 0.0 3.7 64.9 5.8
District of Columbia 1.9 8.2 17.4 46.9
Florida 30.3 55.9 133.8 41.8
Georgia 10.0 15.9 65.4 24.3
Hawaii 0.0 0.0 15.6 0.0
Idaho 2.5 3.9 8.1 48.3
Illinois 0.9 15.0 173.4 8.6
Indiana 3.8 14.0 55.1 25.3
Iowa 2.2 3.0 32.6 9.1
Kansas 0.0 0.0 27.1 0.0
Kentucky 1.0 15.1 39.4 38.3
Louisiana 0.0 1.3 36.3 3.6
Maine 4.7 6.7 8.0 84.3
Maryland 7.9 14.9 53.6 27.8
Massachusetts 2.6 0.0 94.6 0.0
Michigan 0.4 2.4 87.9 2.8
Minnesota 0.0 1.0 50.0 2.4
Mississippi 0.0 0.5 20.3 2.6
Missouri 0.0 0.1 58.0 0.2
Montana 3.2 2.4 6.9 36.2
Nebraska 1.4 1.5 18.5 8.3
Nevada 3.1 5.3 16.0 33.0
New Hampshire 0.0 2.1 9.2 22.9
New Jersey 2.1 14.4 92.3 15.7
New Mexico 0.9 0.9 10.7 8.6
New York 1.0 15.2 420.9 3.6
North Carolina 0.0 0.1 75.5 0.1
North Dakota 2.5 2.3 7.0 32.8
Ohio 0.4 7.6 106.5 7.1
Oklahoma 1.8 1.5 26.0 5.6
Table 2-5. (Cont.)
1987 1990
Total Total
Interstate Interstate Total %
State Assets Assets Assets Interstate
Oregon 5.8 9.5 22.3 42.5
Pennsylvania 0.0 21.0 160.6 13.1
Rhode Island 3.6 4.7 17.5 27.0
South Carolina 6.9 11.4 24.0 47.7
South Dakota 2.9 2.9 18.0 16.1
Tennessee 0.8 14.2 45.9 31.0
Texas 0.0 68.9 167.2 41.2
Utah 1.3 3.6 12.2 29.8
Vennont 0.0 1.0 5.8 18.1
Virginia 2.6 4.6 67.4 6.9
Washington 12.2 29.8 37.8 78.7
West Virginia 0.0 0.2 17.0 1.2
Wisconsin 1.3 8.4 44.2 19.1
Wyoming 0.5 1.4 4.3 33.0
Total U.S. 148.4 442.4 2,915.9 15.2
Source: Federal Reserve Bulletin, February 1987, Table 2A, and Financial Studies Section,
Board of Governors of the Federal Reserve System, September 1990.
affiliates continues to be strong even though banking alternatives have

been the dominant means for interstate expansion since 1983. The prospects
for bank expansion via nonbank BHC subsidiaries will likely hinge on the
willingness of Congress and/or the Federal Reserve Board to permit
expanded nonbank powers. If full investment banking powers were per-
mitted, there might be a flurry of interstate acquisitions of regional invest-
ment banking companies.
Competition and Other Public Policy Issues Arising

From Interstate Banking
Dating back to the previous century, interstate banking has been con-
sidered by many banking company managements as a preferred means
of conducting banking business. This is evidenced not only by the current
rush by banks to form interstate relationships, but by the number of
interstate banking companies established before the 1956 Bank Holding

Company Act and the one-bank-holding-company attempts to expand
interstate and by product line in the 1960s via nonbank holding company
subsidiaries. Further evidence of the desire for interstate banking rela-
tionships is the extent of the complex of correspondent banking relationships
that continue even in the era of extensive electronic funds transfer and
information media.
The motivations for banking management to expand interstate are
numerous-not the least of which is the desire to secure a monopolistic
position in a banking market or many banking markets. Complementary
to this motivation is that of establishing a dominant market position through
absolute firm size and/or a large market share. In this section the analysis
turns to considering the competitive and other public policy issues arising
from interstate banking and to public policy recommendations for
improving the transition to a desirable competitive environment once
interstate expansion has reached a steady state. The issues considered
are those raised in the Treasury Report on Geographic Restrictions on
Commercial Banking (1981):
1. Competition and concentration of financial resources

2. Effective provision of banking services in local communities
3. Viability of small banks
4. Safety and soundness of the banking system
5. The dual banking system.
The Competitive Issues and the Evidence
The interstate banking movement is a transition from a condition of re-

stricted (practically blockaded) entry of banking companies headquartered
in one state into another. As suggested above, this condition represents
a disequilibrium caused by governmental prohibitions to bank expansion.
Once these restrictions were lifted, many banking company owners and
managers found it in their best interest to expand interstate.
Under the present structure of state laws governing interstate banking,
de novo entry and interstate branching are not possible. The effect of this
legal structure implies that banking companies must acquire existing banks
or bank holding companies in order to expand interstate. Because of the
resources required for such expansion and the managerial experience to
coordinate the expanded organization after the merger, only the largest
and better capitalized bank holding companies have been the major
participants in interstate expansion. Many of these companies such a

NCNB (North Carolina), First Interstate (California), Citicorp (New York),
BancOne (Ohio), and Sovran (Virginia) have expanded into multiple states
forming a network of interstate banking offices.
The possible competitive effects, if this form of expansion continues,
are mixed. Interstate banking acquisitions or mergers are primarily of the
market-extension type, even though some are among banks competing in
the same local markets (for example, mergers between several Virginia
bank holding companies and Maryland and District of Columbia banking
companies were among banks competing in the Washington, D.C. metro-
politan area). This type of expansion and the lowering of interstate entry
barriers may be thoroughly consistent with fostering better banking services,
reducing monopolistic profits, and lessening the concentration of banking
resources. The affirmative of this position was taken by the Treasury
Department in a Report of the President, (1981, 12) after making a study
of geographic restrictions on commercial banking.
There has not been sufficient time for interstate banking to produce a
record of effects on local market and statewide concentration. However,
several authors (Rhoades 1985, Savage 1987, and Amel and lacowski
1989) have attempted to address this issue. Rhoades compares local mar-
ket concentration as measured by the three-firm concentration ratio and
the Herfindahl index from 1966 to 1981 and concludes that, for both
SMSA and non-SMSA markets, local market concentration has declined.
He also finds that there is no difference in the average local market con-
centration among states with different interstate banking laws, with the
exception of states with strict unit banking laws and prohibitions on
multibank holding companies. Rhoades concludes from these data that
interstate banking will likely have little effect on local and statewide banking
market concentration if the same banking company merger standards are
enforced. However, Rhoades does not consider the possibility of banks
within states or regions consolidating in order to increase their size and
value to ward off potential unwanted acquirers. These acquisitions may
be considered more favorably by banking authorities if interstate banking
were in place and result in an increase in local market and statewide
concentrations.
Some evidence of a reversal of the declining trend of local banking
market concentration has been reported by Amel and lacowski (1989,
132), particularly in urban markets. As measured by the three-firm con-
centration ratio, the average value increased from a low of 65.8 percent
in 1982 to a value of 67.5 percent by 1986 (the latest date for which data
are available). Additional evidence by state reveals that the rise in local
banking market concentration has been widespread. These data indicate

that thirty-three states experienced an increase in the average three-firm
concentration ratio in urban markets of those states from 1981 to 1986
while only seventeen states recorded a decrease. As the authors remark,
these data suggest that the trend reversal in local concentration occurs at
the time when many states began to adopt liberalized interstate banking
laws. By themselves, interstate banking acquisitions and mergers should
not have a direct effect on local banking market concentration. As was
suggested above, there had been many intrastate bank and banking holding
company mergers and acquisitions in preparation for interstate banking.
These consolidations may have been the cause of the rise in local market
concentration beginning in 1982.7
The undue concentration of aggregate, national financial and, particu-
larly, banking resources has been of serious political and economic concern
in the United States since the beginnings of the nation. The balance between
a concentrated federal system and a more dispersed state-controlled sys-
tem is seen nowhere more sharply than in the U.S. dual banking system
with its federal and state chartering and regulatory control of banks and
thrifts. Additionally, the structure of the Federal Reserve is a compromise
between those wanting a strong central bank and those wanting regional
control of this power. Interstate banking does not often change state or
federal control with one exception: It gives the Federal Reserve author-
ity, via the BHC Act as amended, over larger banking companies.
These concerns over an undue concentration of resources in banking
may presently be misplaced. However, there has occurred over the past
five years a marked increase in the share of banking assets controlled by
the largest firms. The most recent data show that the share of banking
assets of the twenty-five largest domestic banking companies has increased
from 33.1 percent in 1985 to 36.5 percent in 1990, and for the largest fifty
banking companies from 45.7 percent in 1985 to 51.7 percent in 1990 (see
table 2-6). The greatest increase, however, has been in the share of the
one hundred largest banking companies which rose from 57.7 percent in
1985 to 64.2 percent in 1990, despite remaining practically constant over
the previous fifteen years. From these data, the largest increases have
taken place among those banking companies in the superregional group
and outside the top ten inasmuch as there has been little change in the
share of the top ten, (table 2-6). As was discussed in the previous section,
the non-money-center banking companies have been the most aggressive,
and generally better capitalized, in interstate banking expansion. Coupled
with interstate banking laws of many states placing limitations on money-
center bank entry into these states and large banking company failures in
Table 2-6. Shares of Domestic Commercial Banking Assets Held by Largest

Banking Organizations 1 (Percent).
Year Top 5 Top 10 Top 25 Top 50 Top 100

1970 14.0 21.4 32.8 41.1 50.4
1975 13.7 21.3 32.6 41.1 50.8
1980 13.5 21.6 33.1 41.6 51.4
1985 12.8 20.3 33.1 45.7 57.7
1990 2 13.1 21.7 36.5 51.7 64.2
1 Banks are ranked by banking assets. Only insured commercial banks are included; non-
deposit trust companies are excluded.
2 June 1990 Report of Condition
Source: Federal Reserve Bulletin, February 1987, Table 3, and Board of Governors of the
Federal Reserve System, Financial Studies Section, September 1990.
the Southwest, there is little wonder that nationwide concentration has

increased in the particular way it has.
These results need not be interpreted as alarming or as a substantial
degradation of national competition. Indeed, these developments in
nationwide concentration might be considered healthy. These results
demonstrate that a group of superregional banks are developing, willing
to take risks to expand nationwide and ready to challenge the traditional
money-center banks for banking customers throughout the nation. Such
developments may increase competition for medium- to larger-size busi-
ness and household customers. These larger superregional banks with
banking offices in a number of states may be more able to compete for
customers over a wider region or nationally than they had been able to
do with a narrower, regional setting.
In contrast to this sanguine outlook for competition in the face of
increasing national and statewide concentration, is the view that anti-
competitive effects may arise from "linked oligopoly" or multimarket
interdependence among competitors. This hypothesis, attributed to Corwin
Edwards (1955) and applied to banking, is that as multimarket firms meet
in numerous markets, they will develop a mutual awareness and inter-
dependence and avoid aggressive behavior in all mutual markets and
retaliation in any market. Within the context of interstate banking, this
hypothesis would suggest that as banks expand interstate they will begin
to be competitors or rivals in numerous markets throughout the United
States. As the progress of interstate banking proceeds, more banking
companies that have not previously been direct competitors will become
so in many local markets. The development of mutual forbearance among

these various competitors is then likely to become established. The com-
petitive effect is to blunt the pro-competitive aspects of interstate banking
such as new entry by independent banking companies into previously
blockaded markets and the potential competition of new firms not presently
in these markets.
No research has directly addressed this competitive issue within an
interstate banking framework. However, there has been research in this
area using data from intrastate bank expansion into local markets (Rhoades
1983, and Mester 1987). To a great extent this research has not provided
much support for this hypothesis. As Rhoades (1983, 9) finds, there is no
systematic evidence of multimarket linkage among competitors and various
measures of local market rivalry such as mobility and rates of turnover.
This conclusion is supported by Mester (1987).
A related hypothesis maintains that the absolute size and financial
strength, so called "deep pockets," of multimarket banking companies is
a source of barriers to entry into local markets and may reduce interfirm
rivalry (see Rhoades 1983, Curry and Rose 1984, and Hanweck and Rhoades
1984). Research in this area has been mixed, but at least the study by
Hanweck and Rhoades found a significant relationship between firm
dominance and rivalry within local banking markets. This suggests that as
interstate expansion proceeds, there may result large, nationwide banking
firms which, by acquisition and from a pre-existing market structure, gain
dominance within local markets. If the dominant firm-deep pockets-
hypothesis is correct, substantial barriers to entry may be erected leading
to significantly higher than competitive prices for banking services in these
markets. Not only does further research need to be done on this and the
multimarket interdependence hypothesis, but, without clear evidence of
anticompetitive effects, antitrust policy and enforcement are impotent in
redressing competitive harm from interstate banking.8
The Survival of Small Banks
Interstate banking is, as the evidence presented in this study shows, giving
rise to greater consolidation among banking companies and to more large-
size banking firms nationwide. This result has not been unanticipated.
Many observers have predicted that small local banks will not be able to
successfully operate in direct competition with large superregional and
money-center banks. Their ability to survive such competition under
interstate banking will depend on their cost efficiency compared to much
larger banks, their ability to generate profits, and their efficiency in adopting
new technology. Employing a simple rule of inference, small banks are
likely to survive interstate banking if they have been able to survive the
intrastate expansion that has taken place in a number of states over the
past thirty years. In this regard, the evidence since 1985 (Amel and Jacowski,
1985, 125) clearly shows that the small bank is far from becoming extinct
(see Rhoades, 1985, 1130, for a similar pre-1985 assessment).
Of the studies that have recently addressed interstate banking com-
petitive effects directly, all have found no meaningful anticompetitive or
procompetitive effects. The study by Phillis and Pavel (1986), mentioned
earlier, found that banks expanding interstate preferred to acquire banks
with larger branching networks, strong local deposits bases, and a greater
concentration in consumer lending. A study by Goldberg and Hanweck
(1988) compared the performance of those domestic BHCs grandfathered
by the 1956 Bank Holding Company Act with those banks and BHCs
operating in the same states but without interstate banking privileges.
The study confirms that the grandfathered BHCs experienced a statisti-
cally significant decline in the share of state deposits and a homogeniza-
tion in their profitability performance and portfolio composition compared
with their peers in the same states over the 1960 to 1983 period. By 1970
the extent of the homogenization is nearly complete, and it is difficult to
find much difference between grandfathered BHC affiliates and banks of
the same size in the same state; by 1983 the homogenization process
appears to be complete. These results are similar to those of other related
studies (Rhoades 1984, and Golembe 1979) that support the view that it
is unlikely that interstate banking will result in the dominance of local or
regional banking markets by large interstate banking organizations
(Goldberg and Hanweck, 1988, 67).
Productive Efficiency and Safety and Soundness Issues
Bankers have contended that interstate expansion is necessary to achieve

economies of scale and scope in the production and distribution of bank-
ing services and achieve greater geographic and product diversification. 9
The achievement of scale and scope economies is of dubious worth inas-
much as the overwhelming number of studies have found economies of
scale in banking to exist only up to at most $500 million in assets (see
Benston, Hanweck and Humphrey 1982, Berger, Hanweck and Humphrey
1987, and Clark 1988 for a review of this literature). Economies of scope
fare no better. Studies by Berger, Hanweck and Humphrey (1987) and
Gilligan and Smirlock (1984) found little to suggest that scope economies
are important in banking. Most of these studies do suggest that unit banking
is inefficient, resulting in significant diseconomies of scale at larger sizes
and which were not exhibited by branching banks (Berger, Hanweck and
Humphrey 1987). Several recent studies have focused on scale economies
using only the largest banks in their estimates. Their results suggest
economies of scale among the largest banking companies at about $20
billion in real assets. The problem with these studies is that there is little
factual rationale for such economies particularly in light of the fact that
automation costs have plummeted relative to computing power or service
over the past fifteen years.
The safety and soundness of the banking system may be enhanced
through interstate banking expansion. Banks can now become more
geographically diversified in both their deposit liabilities and lending in
commercial and retail markets. Because of the larger available markets,
interstate banking companies may be able to rely more upon less volatile
core deposits related directly to their banking business in local markets.
Interstate banking provides a stop-gap avenue for acquisition of failing
banks that, when unavailable, raise the cost to taxpayers and the FDIC of
failed bank resolutions. Finally, on a less positive note, interstate banking
is creating many more superregional and large size banking companies
with wide-ranging networks of bank and non-bank affiliation throughout
the nation. The failure of anyone of these will tax the FDIC bank reserve
fund (BIF) and require the Federal Reserve to provide the support to
avoid liquidation until a merger partner can be secured. With the continued
increase of larger banking companies, the FDIC will become increasingly
vulnerable to such failures resulting in greater uncertainty and instability
within the banking system. lO
Interstate Banking and Banking Company Valuation
Nationwide interstate banking may increase the number of bidders so

that the prices of banks that might be targeted for acquisition, and the
market value of those banking companies being potential acquirers, will
be bid up. Analysis of the effects of state laws upon bank valuation must
distinguish among the types of state laws and among the types and loca-
tions of banks. As discussed above, some states have tried to attract entry
by banks throughout the country by not placing restrictions on interstate
banking. Many states have reciprocity provision that prohibit entry by
banks from states that prohibit interstate banking. Finally, a number of
states have combined into regional compacts, permitting entry only by

bank holding companies within the region. This type of rule is intended
to deter money-center banks from entry into these regions. This method
has not been entirely successful as the money-center banks have obtained
entry into some of the restrictive states through other means, such as
acquisition of failing institutions. 11 The effects on bank valuation is likely
then to vary by type of state law.
The effects of state law changes on bank valuation should vary also by
type of bank. Studies attempting to assess the effect of interstate banking
law changes must be able to examine the effects of announcements of
these changes not only on banks eventually acquired, but on all banks in
all states that might be affected. A large volume of literature on mergers
generally finds that acquired firms experience increases in stock value, but
the results are mixed and usually insignificant for acquiring firms.12 In
contrast, a number of studies on bank mergers have found positive
announcement effects on returns for acquiring firms as well as acquired
firms (Desai and Stover 1985, James and Weir 1987, and De and Millon
1988). In one of the studies addressing interstate banking acquisitions, De
and Millon (1988) find positive announcement effects on returns for both
acquired and acquiring firms in interstate banking acquisitions. Based on
these findings, they recommend relaxation of interstate restrictions, and
particularly advocate entry on a national rather than a regional basis.
In a recent study, Adkisson and Fraser (1990) consider the effects of
the removal of barriers to entry by liberalized interstate banking laws on
premiums (price paid less book value) paid by acquiring banks in banking
mergers and acquisitions. They test two separate, but related, hypotheses.
The first is that the removal of interstate banking restrictions increases
the number of potential bidders for target acquired banks, resulting in
higher merger premiums. The second hypothesis is that the reduction in
entry barriers, as a result of reducing the local market power and potential
excess profits of target acquired banks, lowers their merger premiums.
Additionally, according to this hypothesis, target banks in states with less
restrictive intrastate banking laws and, thus more market opportunities
for expansion by potential acquirers, will tend to have lower merger
premium offers.
Adkisson and Fraser use an analysis of covariance and a sample of 174
bank holding company acquisitions in 1985 and 1986. Their choice of time
period was made to mainly include acquisitions that occurred after the
Supreme Court in June 1985 upheld the constitutionality of regional banking
pacts in the Northeast Bancorp decision. The authors' results support the
first hypothesis since they find statistically significant higher acquisition
and merger premiums associated with acquisitions in states allowing

interstate banking and for those states with unlimited intrastate holding
company expansion. Since the authors provided no evidence of effects on
premiums of acquisitions in restrictive intrastate banking and liberalized
interstate banking state or vice versa, these interactions cannot be as-
sessed even though conceptually meaningful in determining bid pricing.
As was mentioned, studies have shown that interstate acquirers prefer
banking companies with more extensive branching systems in place (see
Phillis and Pavel 1986, Savage 1987, and Rose 1990). If, as the authors
conclude, their results confirm the common finding that excess returns are
negative for acquiring companies in bank acquisitions, the advantage of
interstate banking to acquiring banking companies is doubtful or extremely
long-term and beyond the horizons of most investors.
Evidence supporting the view that even longer term advantages of
interstate banking are not necessarily forthcoming to those companies
expanding interstate is found in Goldberg and Hanweck (1988). This study
examined the performance of the seven grandfathered interstate banking
organizations and found them to differ little from their peers. In fact,
from 1960 to 1983 these organizations experienced a statistically signifi-
cant decrease in their share of state deposits. This suggests that acquiring
organizations would not be expected to gain significant advantages. Un-
less surviving banking companies can operate more efficiently than the
acquired banks, gain scale or scope economies, or reduce risk, there is
unlikely to be an increase in value for surviving banks in the long-run.
Potential increases in value may be competed away in the bidding for
acquisitions and later by the greater competition for banking services
arising from reduced entry barriers, made possible by liberalized interstate
banking laws.
The issue of motivations for interstate expansion, in face of substantial
evidence of unprofitable and perhaps even negative net present value
acquisitions and mergers, is considered in a recent study by Rose (1990).
In this study the author compares factors possibly influencing the risk and
expected value of the firm to shareholders in the case of seventy-seven
acquired banking companies and ninety-two acquiring companies involved
in interstate acquisitions. All acquisitions involving a failed or failing
banking company were eliminated from the respective samples.
Rose (1990, 39) summarizes the results from the study as follows:
The Study finds that, on average, across the entire sample of banking firms
reviewed, interstate acquirers generally performed less well than comparably
sized and located noninterstate banking companies displaying greater average
risk to capital and lower average asset returns and operating efficiency. The
same deficiencies appeared to characterize the firms targeted for acquisition

across state lines relative to peer institutions ....
Rose (1990, 39-40) concludes that:
While these 'mergers of equals' may thus be expected to have relatively few
benefits from a diversification standpoint, recent interstate acquisitions may
nevertheless represent a managerial strategy to change the structure of bank-
ing markets served in order to create an environment more conducive to higher
returns on equity capital, or reduced risk exposure, or greater price discrimi-
nation, etc., for acquisition-bound banking companies.
It may well be that full-service interstate bank expansion is less a vehicle for
the aggressive pursuit of profit and value maximization through expanded
revenues and greater cost economies and more a defensive reaction to in-
creased risk occasioned by deregulation and the emergence of new domestic
and foreign competitors ....
These results suggest that both the motivation and effects on banking
company valuation from interstate expansion will result in a deterioration
in banking company valuation. Consequently, the markets at the time of
the announcements of liberalized interstate banking law changes should
have taken these factors into account. Rose's results suggest that there
should have been either no change in bank valuation or a decline in value
when these laws were first passed and when they became effective.
The effect on banking company valuation from changes in interstate
banking laws is evaluated in a recent study by Goldberg, Hanweck and
Sugrue (1990). They hypothesize that announcement of a liberalization in
interstate banking laws would positively affect the stock prices of banks
in that state, especially for those banks most likely to be acquired. An
increase in the number of potential bidders for a bank would likely increase
the value of a bank in a liberalizing state. In addition, a new law might
trigger reciprocity provisions in other states and thus increase expansion
opportunities to lucrative out-of-state markets. From another perspective,
the liberalization of interstate banking might be regarded as a stimulant
for economic development in a state. Maine, for example, the first state
to adopt interstate banking legislation, was anxious to encourage the New
England banks to bring resources to the state. Alternatively, the value of
established firms in a state could be reduced by increased competition
from out-of-state entrants.
The study by De and Millon (1988) on interstate banking acquisitions
indicates that acquiring firms have increased in value. Therefore, banks
outside of states with changing laws would increase in value if interstate
expansion were deemed to be a valuable option and would not change if
interstate expansion were considered to have no special advantages. Banks

in nearby states should be affected to a greater extent, especially when a
regional interstate banking law is the applicable law. In fact, if the regional
interstate law is expected to allow smaller regional banks to grow interstate
and become stronger competitors with money-center banks, the value of
money-center banks might very well decrease. In addition, money-center
banks might be expected to be less affected by the opportunity for interstate
banking expansion for several reasons. First, they have already been able
to engage in various interstate operations so the new powers would not
provide relatively as much in increased opportunities as for smaller banks.
Second, a physical presence could be expected to be relatively more
important for smaller banks. Third, new interstate operations would be a
smaller share of a money-center bank's operations and thus would have
less of an impact upon firm value. In the empirical analysis, the authors
separate out-of-state and out-of-region money-center banks from other
commercial banks in order to uncover the different effects of the laws.
Goldberg, Hanweck and Sugrue (1990) measure the reactions to the
new state laws in two ways. Passage of the law reflects anticipation of the
law's effect. Enactment of the law represents the realized effect on bank
value. The results reported in this study are similar for both sets of dates.
The authors estimate the effect of interstate banking law changes for
thirty-five passage events and thirty-five effective events using a sample of
131 banking companies with continuous trading over the period of February
1, 1982, to September 1, 1987. Estimates were made of the fifty-one
nonoverlapping passage and effective dates of interstate banking law
changes using daily stock return data and the Zellner (1962) seemingly
unrelated regression technique. The matrix by state and event date of
interstate banking law changes is presented in table 2-7 and demonstrates
the complex of events that took place for which effects on bank valuation
must be estimated.
This study differs from most other event studies in that it attempts to
simultaneously estimate the effects of numerous events affecting many
banking companies over an extended period of time. Not only are the
events clustered, but they are expected to affect some banks differently
than others. Banks are identified not only if they are in a state that had
an interstate banking law change, but if they are in a state with a regional
interstate banking compact. In addition, the study is able to determine the
effects on those banks outside of states and regions for each separate
event period and, therefore, can measure separate effects for all banks at
different times.
The results of the study by Goldberg, Hanweck and Sugrue (1990)
Table 2-7. Banking Law Dates and Reciprocal Provisions.
A A A A C CCDDFH G I I I I K K L MM
State Passage Effective L K Z R A 0 T E C L I A D L N A S Y A E D
NY 820628 820628
AK 820701 820701
MA 821230 830701 X X
RI 830517 840701 X X
870701
cr 830608 830608 X
ME 840207 840207
GA 840405 850101 X X X X
UT 840406 840415
KY 840407 840715 X X
860715
SC 840521 860101 X X X X X X X X
FL 840522 850701 X X X X X X
NC 840707 850101 X X X X X X X X
OR 850312 860701 X X X X X
ID 850312 850701
VA 850324 850701 X X X X X X X X
AZ 850418 861001
IN 850418 860101 X X
TN 850501 850701 X X X X X X X
WA 850516 870701
NV 850613 850701 X X X X X
MD 850621 850701 X X
870701
OH 850718 851017 X X X X X X
DC 851008 851122 X X X X X
IL 851125 860701
MI 851205 860101 X X
MO 860218 860813 X X X X
AL 860221 870701 X X X X X X X
MN 860318 860701
W1 860331 870101 X X X X
PA 860626 860825 X X X
LA 860709 870701 X X X X X X X
TX 860924 870101
CA 860929 870701 X X X X X
WY 870316 870523
NH 870514 870901 X X
MMMMMMNNNNNNNNOOOPRSSTTUVVWWWW
A INS 0 T E V H J M Y C D H K R A leD N X T T A A V I Y
x x x
x x x
x x x x
x x x x
x x x x x
x x x x x
x x x x x x
x x x x
x x x
x x x x x
x x
x x x x x x
x x x x x
x x
x x x x x x x x
x x x x x x
X X X
X X
X X X X X X
X X X X
X X X X
X X X X X X X X
X X X X X
X X X
support those of Adkisson and Fraser (1990) and contradict, to some

extent, those of Rose (1990). The authors find that in-state and in-region
banks benefitted most from the liberalization in interstate banking that
has taken place since 1982. Both the increase in the number of potential
acquiring firms and the added expansion opportunities for these banks
would be expected to increase their value. This has apparently outweighed
the negative impact on value of increased competition. Except for money-
center banks, out-of-state and out-of-region banks appear also to have
benefitted from passage and enactment of the liberalized interstate banking
laws. This can be due to increased expansion opportunities and to triggering
of reciprocity provisions which may increase the availability of potential
acquirers of the out-of-state banks. Because money-center banks already
have some interstate operations and the new opportunities, as a result of
the legal changes, do not represent a relatively large part of their busi-
ness, they do not appear to be positively affected by the liberalization of
interstate banking. Finally, the generaf findings of this study suggest that
investors in banking companies believe that the liberalization of interstate
banking will have a positive impact on the future ability of banking
companies to increase their profitability and value. It is in this regard that
these results are in direct contrast with those of Rose for acquired and
acquiring banks over the same time period.
Future Interstate Bank Expansion and Bank Profitability

and Capitalization
This study documented in sections two and three the rapid expansion of
interstate banking and the substantial consolidation of banking resources
that has been taking place since about 1982 in the largest fifty and one
hundred banking organizations. From 1980 to 1989 there have been over
ninety-five Federal Reserve approved bank mergers and holding com-
pany acquisitions among banking companies, with each having in excess
of $1 billion in domestic deposits. This total excludes acquisitions by foreign
banks. Most of these domestic banking acquisitions have taken place since
1982 and about thirty-two are acquisitions of banks interstate. The char-
acter of this expansion has been such that regional banks have been at its
forefront of this expansion. In view of adverse developments in profitabil-
ity, market valuation, and capitalization over the past year affecting many
of the "superregional" banking companies, the sustainability of this rapid
expansion is in question and with it the future growth in interstate bank-
ing. Coupling this with the forecast recession in the general economy, and
Percent
ROA
1.20
ROA-Small ROABanks
1.00 Banks $300M to $5B in
-l Assets
0.80 --- t
--------~~o;--------
.
ROA - All Banks -~_ -----....
- - - l . _ _-7'-~",,--\ ................
0.60 /
/ ......-.:-...... "" ,""
,,,.,,.~
-------------~-- ------------\
0.40
----_J /
/
,\
0.20 ROA Other Large \
Banks ,,
\
,,
0.00
,,
, ,,
-0.20 ,,
,,
,,
-0.40
,,
,
-0.60 \ I
ROA 9 Money \ :
Center Banks - \ :
-0.80
'J
-1.00
1983 1984 1985 1986 1987 1988 1989

Source: Board of Governors of the Federal Reserve System
Figure 2-1. Rate of Return on Assets (ROA) by Bank Size
real estate in particular, and stiffer regulatory capital requirements and

oversight of commercial real estate lending, many banks will find it almost
impossible to expand without greater equity capital and/or regulatory
assistance in FDIC-assisted acquisitions.
Bank Profitability and Cash Flow
During the 1980s, overall bank profitability has both declined, in terms of
either rates of return on assets (ROA) or on equity (ROE), and demon-
strated a high degree of variability (see figures 2-1 and 2-2). Over the
period from 1983 to 1989, for which consistent data are available for
banks by size classes, the ROA for all banks varied from a high of 0.84
Percent
ROE
25.00
20.00
ROE Banks
15.00 ROE - All Banks
..
$3OOM to $5B in
.. :....
." .-----------,1'-
/ ----
................... ::\... Assets
10.00
-........
/
/ .~
... ---------,;...
....
, ....... ~
----- ..
-:,......... t
----
.... ~~~~
J
---- ./ T ....-,--- - •
5.00 • - ROE-Small \ •
',~ 'I :
ROE Other Large Banks
'/:
.
\
Banks \ \/'
0.00 ,, ".,:
\ \ /
,, ,
,, ,,
,
,, ,
,,,
-5.00
,,
,, ,,
-10.00 ,, ,,
, ,
-15.00
ROE 9 Money \
Center Banks \ :
'
., '
\ :
"'I
-20.00 +,---+-----+---4-----+----+----f...-----i
1983 1984 1985 1986 1987 1988 1989
Figure 2-2. Rate of Return on Equity (ROE) by Bank Size
percent to a low of 0.11 percent Likewise, ROE for all banks over the
same period ranged from 13.7 percent to 1.94 percent These values are
in sharp contrast with bank earnings performance even during some of
the turbulent periods of the 1970s.
Much of this variation and general decline is due to the performance
of large banks. As figures 2-1 and 2-2 demonstrate, ROA and ROE for
smaller banks, those with assets less than $5 billion, may have declined
over the period, but even facing difficult earnings pressures from agriculture
and energy price declines, their variability was not excessive. Although
it is not the purpose of this study to dissect the banking industry, and
particularly larger banks, to determine the causes of the present profitability
problems, it is sufficient to point out that these problems have persisted
for much of the 1980s. For larger banks, their sources are well known and
Percent
NIM to Assets
4.50
NIM-Small
......__ ...,................................ Banks ....... _
4.00 ... -............ __ i ____--------......
Banks $3OOM to
$5B in Assets
t ~~####~---------------------- ... ---- --_ .. -----------
3.50
----------............. NIM . All Banks
--
t
/'
~------...--...-
3.00 ,-}
/' T ~
...-..-"- Other large " ..................
...--- Banks /' --........._
~~
2.00
---------------------
.
----------... ............
9 Money Centet----_ .."
Banks
"'
I" ""
1.50
1.00
1983 1984 1985 1986 1987 1988 1989

Figure 2-3. Net Interest Margin (NIM) to Assets by Bank Size
stem largely from the overhang of less developed country (LDC) loans
originated in the late 1970s and regional real estate recessions.
In order to emphasize this point, the net interest margin to total assets
(NIM) is presented in figure 2-3 for all banks and those of the four size
classes. Without exception, there has been a slight upward trend in NIM
over the 1983 to 1989 period. Only for the nine money-center banks has
there been any significant variability. These data suggest that banks of all
sizes have been able to maintain a consistent spread between interest
income and interest expense despite deregulation and greater competi-
tion from nonbank and foreign banking sources.
The principal cause of the variation and downward trend in bank
Percent
LLP to Assets
2.50
•
2.00
" I,,
I
I '- 9 Money Center
I
I \, Banks
I ,
I ,
I ,
I ,
I ,
I ,
1.50 I 1'\ \
I 1\\ I
I
I I \' I
I
I I \ \ I
I
I I \ \ I
I
I I \' I
I I ' I
llt \
I
1.00 LLP-SmaU
II
I
II \'\
\ '
I
I
/
:~~Large '~
I /
l/
........:: .. '/
t ,..,.. . . . . . . . . . . . . . . . . . . . . . ,,-';:.----- ~................... \\ l/
. . . . . .". ,,*' ,'. . . . . /~ . . . . . . . . . - ". . . . . . . . ~ /~"",-
LLP to Assets - .............
0.50
"
....~",,','
..
... '
"",
All Banks ' ,
.....t1---__
____~-- • ,I
....... Banks $3OOM to 'I
$5B in Assets
0.00
1983 1984 1985 1986 1987 1988 1989

Figure 2-4. Loan Loss Provisions (LLP) to Assets by Bank Size
profitability is due to loan losses. Even though banks have been able to
maintain consistently strong net interest margins, as demonstrated by loan
loss provisions as a proportion to assets, loan losses have cut deeply into
this basic source of bank earnings. With the exception of the smallest
banks, those with assets less than $300 million, which have shown a re-
versal since 1986, provisioning for loan losses remains on an upward trend
(see figure 2-4). Taking into consideration the possibility of a general
economic recession in 1990 and the spreading regional real estate recessions,
there is little likelihood that this trend will be reversed in the next several
years.
The variability and downward trend in bank earnings has limited bank-
Percent
Equity to Assets
9.00
Equity to Assets
Small Banks _------
8.00
---
~~~-----
j
------- -_...............
7.00 Banks $300M to
'_"_"_'_"-'~":::"'::--::-~-----------------------------~-!~~7----------~
-.
.
6.00
Equity to Assets
All Banks ~-----~~- ~~~~
.
5.00
" ~/ --------~
"" Other Large
--_...............-
,"'" _--------------............. Banks ..... _
--.".-~--- ..................
............ 9 Money Center .. ~------
Banks
4.00
3.00
1983 1984 1985 1986 1987 1988 1989

Figure 2-5. Equity to Assets by Bank Size
ers' abilities to add to capital from internal funds. Despite this, banks
have managed to add to equity in sufficient amount to generally show an
upward trend in equity to assets on a book value basis. As figure 2-5
portrays, the smaller banks have the highest ratios of equity to assets,
nearly double that of the nine-money-center banks. With the high degree
of earnings variability demonstrated by the largest banks in recent years,
investors in these banking firms and their large uninsured depositors will
demand more protection in the form of greater equity. In addition, bank
regulators have adopted more stringent capital requirements and en-
forcement procedures. All of these factors contribute to pressure on banks,
especially the largest banks, to add to equity capital.
Percent
Cash Dividends to Assets
0.55
0.50
0.45
0.40
/
r ,I'.............. ........
0.35 ,,/ /
0.30 Dividends to
• / / Other Large
/ . . . f"" """l/'
Assets - All // Banks
.... Banks / "
0.25 ~,;:;::::::.. ,,','
.... ----....
....t. . . . .,,'' "
0.20 9 Money Center
Banks
0.15
0.10
1983 1984 1985 1986 1987 1988 1989

Figure 2-6. Cash Dividends to Assets by Bank Size
Despite the obvious need for equity by banks of all sizes, there has
been an alarming upward trend in cash dividends paid by banks (see
figure 2-6). For example, in 1989 small banks paid out 58 percent of after-
tax earnings as cash dividends and the nine money-center banks, exper-
iencing an after-tax loss, paid 0.36 percent of their assets as cash dividends
or about $2.3 billion. For many banks, cash dividends are paid to the
BHC parent holding company. For many of these holding companies,
their banking subsidiaries are their largest holding and cash dividends
their predominant source of revenue. Although these dividends may be
earnings retained by BHCs, it is likely that they are used to meet holding
company expenses associated with nonbank enterprises and payouts to
shareholders of the BHC. Regardless of their use, they are a source of
cash flow to banks that is not being returned to them in the face of
growing demands for bank equity capital.
Banking Company Capitalization and Valuation
For the banking industry and particularly the largest banking companies,
market valuation has not kept pace with other industries in the United
States. At about the time of the start of the bull stock market in 1983,
bank stock values began to lag behind those in the general market.
In figure 2-7, the S&P 500 stock price index is compared with S&P
stock price indices for thirteen major regional banking companies (regional
banks) and seven money-center banking companies (money-center banks).
These data clearly show the divergence in valuation that began to take
place in the early 1980s. To emphasize the divergence and clarify the
timing, figure 2-8 shows the ratio of the S&P 500 index to the two
banking company indices. These data show that, for the regional
banks, the divergence from the S&P 500 began about 1980 or 1981.
For the money-center banks, this divergence began much later in 1985 or
1986. Based on these data, the market valuation and capitalization of major
U.S. banking companies have not kept pace with a broad range of firms in
other industries.
There are certainly individual exceptions to this generalization, but
data for a broader spectrum of banking companies support this same
conclusion. As reported in Business Week (April 13, 1990, 219-222),
ninety-two major banking companies lost 9 percent in market value in
1989, while all industries showed a gain of 15 percent and general nonbank
financial services (insurance, brokerage, etc.) had a 14 percent increase.
Of the ninety-two banking companies reported, thirty-three showed market
value declines. Some of the major gainers were BankAmerica (California),
Crestar Financial (Virginia), Bank of New York, NCNB (North Carolina),
and Banc One (Ohio), while some of the larger losses were experienced
by Bank of Boston, Shawmut National (Massachusetts), Citicorp, Southeast
Banking (Florida), and First Interstate (California). By way of contrast,
1988 was an improvement for most banking companies, while in 1987
banking company market value over a broad range of companies showed
a decline of about 11 percent (Business Week, April 15, 1988,234).
Smaller regional banking companies have fared no better than major
regional and money-center banks since July 1989. Figure 2-9 compares
the monthly S&P's stock indices for the money-center and major regional
banks with the National Association of Securities Dealers Automated
Stock Price Index

350
300
250
S&P500-
200
150
Regional
1\
100
.j
"
V, ~ Banks A
1V
\;,J,' '\ l
,~\
',_\
I " , \
I \
I '
,'\ ,,--...;
50 ,. ____"J
,I' '.,.... __......
........ _ _ _ _ , '
,....,,1
,_""_-""--............ '.,iJ' Money Center
,---' Banks
o
1958 1962 1966 1970 1974 1978 1982 1986 1990
Source: Standard and Poor's Analyst's Handbook: OffIcial Series, 1989.
Latest data plotted: August 29, 1990
Figure 2-7. Bank Stock Indices Value vs S&P 500
Quotations (NASDAQ) Bank stock index composed of 209 regional

banking companies (NASDAQ Regional Banks). From July 1989 to
November 1990, the stock value of major regional and money-center
banks declined by 45 percent, and the value for the smaller regional banks
declined by 47 percent. These are substantially greater declines than the
7 percent decline in the S&P 500 stock index over the same time period.
The marked decline in larger banking company market valuation relative
to other U.S. industries, and the erosion of the profitability of these compa-
Percent
S & P to Bank Indices
4.5
3.5
S&P to Regional Banks

2.5 t
1.5
0.5
o
1~ 1~ 1~ 1~ 1m 1m 1m 1m 1~ 1~ 1~ 1~
Source: Standard and Poor's Analyst's Handbook: Official Series, 1989.
Figure 2-8. S&P 500 Index to Bank Stock Indices
nies, will impede the expansion of banking companies interstate. In view

of these data and the serious problems arising in commercial real estate
in a number of formerly real-estate-boom regions of the United States,
banks will need to devote their resources to supporting loan losses and
employ managerial expertise to survival and consolidation of operations.
Unfortunately, the valuation of smaller banking companies has not fared
much better than that of the largest companies. However, these smaller
companies have not suffered as severe erosion in profitability as the larger
companies. For those banking companies that are presently well-capitalized
and have a strong cash flow, the next few years may provide an exceptional
Bank Stock Indices
530
500 1\
,I I,A.
, \ II ~,
470 I \-J I II
440 IA'I
/Vil
II \I\ '('Jt\l\
I
, 1
\
410 J\(J
'I
II
I I
\I
I I: II
380 I I!
I " \
350
)
I
\,
320 \1\
NASDAQ Regional Banks,J
290
t,N
, I
/ I
260
r/ \V
I
230
200 /"" {'

j"' y
('
,
170
,J
,"'
140
;J(~j
110
r~J \J
80
50
204--+--+--r--~~-+--+--+--~~-+--+--+--~4--+--+--r~~4--+--
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
Source: National Association of Securities Dealers and 51
Figure 2-9. Bank Stock Indices January 198Q--November 1990

opportunity for expansion interstate because of more banks being offered

for take over and fewer competitors for them. However, the data presented
in this section suggest that there are not many larger banking companies
in such an enviable position.
Conclusion and Public Policy Recommendations
Liberalized interstate banking legislation since 1982 has presented banking

in the United States with an opportunity for the expansion of geographic
markets, a consolidation of banking resources, and increased competition
via entry by acquisition. All of these may have some benefit to the public
in terms of less expensive and more competitive financial services. The
most notable observation that can be made of the expansion of interstate
banking is that there has been a substantial consolidation of banking
resources within the largest fifty and one hundred banking companies.
There has been little evidence that the provision of financial services has
become any more competitive or that banks have improved their operat-
ing efficiency, profitability, or safety and soundness. Of the numerous
studies reviewed, several indicated that banks have incrementally improved
their market values as a result of the passage and effective implementation
of liberalized interstate banking laws. However, there are those studies
that suggest that acquiring banking companies will not fare as well as they
expect once expansion is completed.
Under present Federal law, banking companies cannot open offices
outside their home states unless explicitly permitted by state law. States
presently do not allow banks in other states to enter de novo and do not
allow branches to be operated in a state unless explicitly provided by
state law. Consequently, there may be little reason to expect that entry
restricted only to acquisition will result in improved competition in local
banking markets.
Additionally, there may also be little reason to expect noticeable im-
provements in the efficiency of bank operations. Few studies have been
able to convincingly demonstrate that there are significant scale and/or
scope economies in banking to cause an improvement in operating effi-
ciency as a result of simply increasing size. In fact, many studies suggest
that without proper diligence by management, rapid expansion into markets
that are poorly understood is a prescription for failure.
The principal policy prescription from this study is that there needs to
be a thorough review of federal law dealing with interstate banking. Current
law in most states has made the prohibitions of the Douglas Amendment
to the Bank Holding Company Act of 1956 virtually obsolete. In order to
get the greatest competitive benefit from interstate banking, banks must
be free to open offices and banks de novo within any state, much like most
other businesses (including bank loan production offices and finance
company subsidiaries of bank holding companies). Such legislation has
been introduced by Congress in the past and debated to some degree.
Some of this legislation prescribes that there be safeguards, such as the
limitation of a banking company gaining excessive share of any single
local market or state. These issues need to be considered and dealt with
within the broader spectrum of antitrust legislation and enforcement. The
financial condition of the banking system may not hinge on whether there
is more liberal interstate banking legislation, but as the profitability and
valuation picture painted in this study suggests, banking companies will
not have the resources to maintain the interstate expansion pace of the
past eight years. Without an alternative means of expansion that requires
less resources than under the present system and allows smaller, more
profitable banking companies to become active in interstate expansion,
further benefits from interstate banking will be long in coming.
Notes
* The author wishes to give special thanks to Don Savage of the Financial Structures
Section of the Board of Governors of the Federal Reserve System for thoughtful discussion
and valuable data.
1. King, Tschinkel, and Whitehead (1989), Table 2, 49.
2. Clair and Tucker (1989), 14.
3. On September 21,1990, Chase Manhattan, the second-largest U.S. banking company,
announced a severe retrenchment by a cut of 5000 staff or 12 percent of its work force by
December 1990. This announcement was part of a major restructuring that includes a larger-
than-expected $600 million addition to loan loss reserves and a cut in dividend. This is
expected to reduce the losses of $665 million taken in 1989. (See the Washington Post,
September 22, 1990, C1).
4. See Goldberg and Hanweck (1988) for a discussion of the legal framework of inter-
state banking and a detailed description of various exceptions to, and means of, circumventing
interstate banking laws.
5. However, the expansion of powers granted S&Ls in the 1980s allowed them powers
similar to banks until they were restricted in 1989 under FIRREA. Nonbank office expan-
sion by bank holding companies will be discussed below.
6. See King, Tschinkel, and Whitehead (1989, 48-50) for a listing of the approved
activities and those prohibited under section 4(c)(8).
7. It is important to note that statewide concentration, as measured by the average five-
firm ratio, increased from 54.7 percent in 1976 to 60.9 percent in 1987. These results support
the evidence for local market concentration increases.
8. Rose (1985) addresses the problems of antitrust enforcement in market extension

mergers and acquisitions. The use of the potential competition doctrine in its two forms has
not been successful in the courts, and the Federal Reserve Board virtually abandoned the
doctrine after the Mercantile Texas case decision in 1981. (Mercantile Texas Corp. v. Board
of Governors, 638 F.2d 1255, 1268 (5th Cir. 1981». Also see Rhoades (1985).
9. For example, one of the reasons given for the Continentallllinois failure in 1984 was
that it was a large unit bank located in a state that did not permit branching or, at the time,
multibank holding companies, so that it was unable to acquire stable core deposits and
needed to rely heavily on volatile purchased funds.
10. At the present time there is considerable concern in the ability of the FDIC to meet
a single failure of any of the large New England banks. The Bank Insurance Fund, BIF, has
an estimated value of $11 billion, after accounting for estimated FDIC losses of about $3
billion in 1990. A failure of the size of the Bank Of New England, with deposits of $9 billion
as of year-end 1989, may result in a cost to the FDIC of $2-$5 billion.
11. See Amel (1988) for a discussion of the motivations for some of the interstate bank-
ing laws.
12. See Jensen and Ruback (1983) and Jarrell, Brickley and Netter (1988) for summaries
of this literature.
References
Adkisson, A. and D. Fraser. 1990. "Competitive Effects of Interstate Banking:
The Impact on Bank Acquisition Markets." Presented at the Bank Structure
Conference, Federal Reserve Bank of Chicago. (May): Mimeograph.
Amel, D. F. 1988. "State Laws Affecting Commercial Bank Branching, Multibank
Holding Company Expansion, and Interstate Banking." Board of Governors of
the Federal Reserve System. (May): Mimeograph.
Arne!, D.F., and Michael Jacowski. 1989. "Trends in Banking Structure since the
Mid-1970s." Federal Reserve Bulletin. Board of Governors of the Federal Reserve
System (March): 120-133.
Ahrony, Joseph, Anthony Saunders, and Ithzak Swary. 1988. "The Effects of
DIDMCA on Bank Stockholders' Returns and Risk." Journal of Banking and
Finance 12: 317-331.
- - . 1986. "The Effects of a Shift In Monetary Policy Regime of the Profitability
and Risk of Commercial Banks." Journal of Monetary Economics 17 (May)
363-377.
Beatty, Randolph P., John F. Reim, and Robert F. Schapperle. 1985. "The Effects
of Barriers to Entry on Shareholder Interstate Wealth: Implications for Banking."
Journal of Bank Research (Spring): 8-15.
Benston, G.J., Gerald A. Hanweck, and David B. Humphrey. 1982. "Scale
Economies in Banking: A Restructuring and Reassessment." Journal of Money,
Credit, and Banking XIII (November Part I): 435-56.
Berger, A., D. Humphrey, and J. Frodin. September 1986. "Interstate Banking:
Impacts on the Payments System." Research Papers in Banking and Financial
Economics. Board of Governors of the Federal Reserve System, Washington,
D.C. #90.
Berger, A, G. Hanweck, and D. Humphrey. 1987. "Competitive Viability in

Banking: Scale, Scope, and Product Mix Economies." 20 December Journal of
Monetary Economics 20: 501-520.
Berger, A and T.H. Hannan. 1988. "The Price-Concentration Relationship in
Banking." Finance and Economics Discussion Series, No. 23, Federal Reserve
Board (April).
Binder, J. 1985a. "Measuring the Effects of Regulation with Stock Price Data."
Rand Journal of Economics 16 (Summer): 167-183.
- - . 1985b. "On the Use of the Multivariate Regression Model in Event Studies."
Journal of Accounting Research 23 (Spring): 370-383.
Cheng, David C., Benton E. Gup, and Larry D. Wall. 1989. "Financial Determinants
of Bank Takeovers." Journal of Money, Credit, and Banking XXI 4 (November):
524-536.
Clair, Robert T., and Paula K. Tucker. 1989. "Interstate Banking and the Federal
Reserve: A Historical Perspective." Economic Review. Federal Reserve Bank
of Dallas (November): 1-20.
Clark, Jeffrey A 1988. "Economics of Scale and Scope at Depository Financial
Institutions: A Review of the Literature." Economic Review. Federal Reserve
Bank of Kansas City (September/October): 16-33.
Curry, Timothy J. and John T. Rose. 1984. "Diversification and Barriers to Entry:
Some Evidence from Banking." Antitrust Bulletin XXIX (Winter). pp. 759-74.
De, S. and M.H. Millon. 1988. "An Examination of Stock Market Reactions to
Interstate Bank Mergers." (July): Mimeograph.
Demsetz, Harold. 1973. "Industry Structure, Market Rivalry, and Public Policy."
Journal of Law and Economics 16 (April): 1-9.
Desai, Anand S. and Roger D. Stover. 1985. "Bank Holding Company Acquisitions,
Stockholder Returns and Regulatory Uncertainty." Journal of Financial Research
8 (Summer): 145-156.
Dunham, Constance R. 1986. "Interstate Banking and the Outflow of Local Funds."
New England Economic Review. Federal Reserve Bank of Boston (March/April):
7-19.
--and Richard F. Syron. 1984. "Interstate Banking: The Drive to Consolidate."
New England Economic Review. Federal Reserve Bank of Boston (May/June):
11-28.
Evanoff, Douglas D. and Diane Fortier. 1986. "The Impact of Geographic Expansion
in Banking: Some Axioms to Grind." Economic Perspectives. Federal Reserve
Bank of Chicago (May/June): 24-38.
Flannery, Mark J. 1982. "Retail Bank Deposits as Quasi-Fixed Factors of
Production." American Economic Review (June): 527-536.
Frieder, L.AI987. "Interstate Banking: Landscape, Policy, and Misconceptions."
Review of Research in Banking and Finance (Fall): 1-17.
Department of the Treasury. Jan 1981. Geographic Restrictions on Commercial
Banking in the United States, The Report of the President. Including the attached
compendium of research prepared by the federal banking agencies. Washington,
D.C.
Gilbert, R. Alton. 1984. "Bank Market Structure and Competition: A Survey."

Journal of Money, Credit, and Banking 16 (November Part 2): 617-644.
Golembe Associates. 1979. A Study of Interstate Banking by Bank Holding
Companies. For the Association of Bank Holding Companies. Washington,
D.C. (May).
Goldberg, Lawrence G. and Gerald A. Hanweck. 1988. "What We Can Expect
From Interstate Banking." Journal of Banking and Finance 12 (March): 51-
67.
- - , G.A. Hanweck, and T. Sugrue. 1990. "Differential Impact on Bank Valuation
of Interstate Banking Law Changes." Department of Finance, George Mason
University. (September): Mimeograph.
Hanweck, G.A. and S.A. Rhoades. 1984. "Dominant Firms, 'Deep Pockets', and
Local Market Competition in Banking." Journal of Economics and Business
XXXVII (December): 391-402.
Hunter, William C. and Stephen G. Timme. 1986. "Technical Change, Organizational
Form, and the Structure of Bank Production." Journal of Money, Credit, and
Banking XLIII (May): 152-66.
James, Christopher M. and Peggy Weir. 1987. "Returns to Acquirers and
Competition in the Acquisition Market: The Case of Banking." Journal of
Political Economy 95 (April): 355-370.
Jarrell, G., 1. Brickley, and J. Netter. 1988. "The Market for Corporate Control:
The Empirical Evidence Since 1980." Journal of Economic Perspectives 2:
49-68.
Jensen, M. 1986. "The Takeover Controversy: Analysis and Evidence." Midland
Corporate Finance Journal (Summer): 6-32.
- - , and Richard S. Ruback. 1983. "The Market For Corporate Control: The
Scientific Evidence." Journal of Financial Economics 11 (April): 5-50.
King, B. Frank, Sheila L. Tschinkel, and David D. Whitehead. 1989. "Interstate
Banking Developments in the 1980s." Economic Review. Federal Reserve Bank
of Atlanta. (May/June): 32-5l.
Liang, N. and D. Savage. 1990. "New Data on the Performance of Nonbank
Subsidiaries of Bank Holding Companies." Staff Study 159, (February) and
Summary, Federal Reserve Bulletin (March) Board of Governors of the Federal
Reserve System.
- - , and S.A. Rhoades. 1988. "Geographic Diversification and Risk in Banking."
Journal of Economics and Business 40: 271-84.
Maddala, G.S. 1986. Qualitative Variables in Econometrics. Cambridge, MA.,
Cambridge University Press.
Mester, Loretta J. 1987. "Multiple Market Contact Between Savings and Loans."
Journal of Money, Credit, and Banking XIX, 4 (November): 538-49.
Palepu, Krishna G. 1986. "Predicting Takeover Targets." Journal of Accounting
and Economics (March): 3-35.
Phillis, D. and C. Pavel. 1986. "Interstate Banking Game Plans: Implications for
the Midwest." Economic Perspectives. Federal Reserve Bank of Chicago.
Rhoades, Stephen A. 1982. "Structure-Performance Studies in Banking: An
Updated Summary and Evaluation." Staff Study. 119, Board of Governors of

the Federal Reserve System, August.
- - . 1983. "A Note on the Resource Allocation Efficiency of MBHCs Versus
Independent Banks." Quarterly Review of Economics and Business XXIII
(Summer): 112-117.
- - , and Arnold A. Heggestad. 1985. "Multimarket Interdependence and Per-
formance In Banking: Two Tests," Antitrust Bulletin (Winter): 975-995.
--.1979. "Geographic Expansion of Banks and Changes in Banking Structure."
Staff Study 102. March Board of Governors of the Federal Reserve System.
- - . 1985. "National and Local Market Banking Concentration in an Era of
Interstate Banking. Issues in Bank Regulation (Spring): 29-36.
- - . 1987. "Determinants of Premiums Paid in Bank Acquisitions." Atlantic
Economic Journal, XV (March): 20-30.
- - . 1985. "Interstate Banking and Product Line Expansion: Implications from
Available Evidence." Loyola of Los Angeles Law Review 18(4): 1115-1164.
- - . 1981. "Can Small Banks Compete?" Bankers Magazine (Jan.lFeb.): 59-65.
Vol 164
Rogowski, RJ. and D.G. Simonson. 1987. "Bank Merger Pricing Premiums and
Interstate Bidding." Paper Presented at the Financial Management Association
Meetings, Las Vegas, Nevada. October 14-17, 1987.
Rose, John T. 1985. "Interstate Banking, Potential Competition, and the
Attractiveness of Banking Markets for New Entry." Antitrust Bulletin (Fall):
729-743.
- - . 1986. "Interstate Banking and Small Business Finance: Implications from
Available Evidence." American Journal of Small Business (Fall): 23-39.
Rose, P.S. 1987. "The Impact of Mergers in Banking: Evidence from a Nationwide
Sample of Federally Chartered Banks." Journal of Economics and Business
XXXIX, (4 November): 289-312.
--.1989. The Interstate Banking Revolution: Benefits and Costs for Business and
Consumers. Greenwood Press, Quorum Books. Westport, Conn.
- - . 1990. "The Consequences of Interstate Banking Deregulation for Competition,
the Structure of Service Markets, and the Performance of Interstate Banking
Financial-Service Firms." Paper presented at the Bank Structure Conference.
Federal Reserve Bank of Chicago. (May).
Savage, D.T. 1987. "Interstate Banking Developments." Federal Reserve Bulletin
(February): 79-92.
Shaffer, S. and D. Edmond. 1986. "Economies of Superscale and Interstate
Expansion." Research Paper #8612. Federal Reserve Bank of New York.
Stigler, G.J. 1971. "The Theory of Economic Regulation." Bell Journal of Economics
and Management Science 2 (Spring): 3-21.
Syron, Richard F. 1984. "The 'New England Experiment' in Interstate Banking."
New England Economic Review. Federal Reserve Bank of Boston. (MarchI
April): 5-17.
Theil, H. 1971. Principles of Econometrics. John Wiley & Sons, Inc., New York.
Trifts, lW., and K.P. Scanlon. 1987. "Interstate Bank Mergers: The Early Evidence."
Journal of Financial Research X (Winter): 305-31l.
Thompson, R. 1985. "Conditioning the Return-Generating Process on Firm-Specific
Events: A Discussion of Event Study Methods." Journal of Financial and
Quantitative Analysis 20 (June): 151-168.
Whitehead, D. 1983. A Guide to Interstate Banking. Federal Reserve Bank of
Atlanta.
Zellner, A. 1962. "An Efficient Method of Estimating Seemingly Unrelated
Regression and Tests for Aggregation Bias." Journal of the American Statistical
Association 57 (June): 348-368.
COMMENTARY
Peter S. Rose
Gerald A. Hanweck presents an interesting and thoughtful article on the

recent developments in interstate banking and their effects on bank stock
values and bank market structure. The article is a combination of thoughtful
normative economics and positive economic analysis, and addresses a
number of emerging trends in local, state, and national market structure
that are at least partly the result of recent interstate banking legislation.
One of these potentially disturbing trends is an apparent increase in
both local and national banking concentration. Hanweck cites a study
by Amel and Jacowski (1989) that points to an apparent reversal of the
historical trend toward lower local banking market concentration, espe-
cially in urban areas. Amel and Jacowski calculate that thirty-three states
experienced a rise in their mean three-firm urban banking concentration
ratios from 1981 to 1986, with only seventeen states experiencing a decline
in concentration. Similarly, the shares of nationwide banking assets held
by the top five, ten, twenty-five, fifty and one hundred domestic commer-
cial banking organizations have all risen since 1985. Moreover, the aggregate
shares of all but one of these groups (the top five banks) have been
increasing nationwide since 1970.
Most of the concentration increases appear to have occurred in the
93
decade of the 1980s-the period over which most of the interstate bank-
ing legislation has appeared. Perhaps, as Hanweck suggests, we need to
do more research focusing on the possible application of the dominant-
firm and linked-oligopoly hypotheses specifically to interstate banking to
determine if the apparent trend toward concentration is damaging to the
public welfare.
I share Hanweck's concern for this apparent rise in banking market
concentration, but I'm not yet convinced that this trend is happening to
the degree that his analysis implies or primarily because of interstate
banking. For example, if you look at the article's numbers and citations
(particularly pages 62-65 and in table 2-6), you discover quickly that
the concentration figures are for domestic commercial banking firms only.
There is no adjustment for foreign-bank penetration of United States
markets, which in selected states has become a potent competitive force,
not only in corporate loans and credit guarantees, but also in consumer
and small business credit. Nor is there any adjustment for nonbank financial-
service competitors, ranging from thrifts, finance companies, credit -card
firms (including most recently AT&T), and insurance companies to security
dealers. We must also hasten to point out that the total assets of banks
that Hanweck focuses upon do not describe the structure of any particular
financial-services market.
There may be another explanation besides interstate banking for the
apparent trend toward greater banking concentration. The 1970s and 1980s
represent a period of sweeping changes in state branching laws. At the
beginning of this period the fifty states were roughly equally divided between
states sanctioning statewide branching activity, those allowing limited
branching within city, county, or district boundaries, and unit banking
states where full-service branch offices are prohibited. By the end of the
period, however, the majority of states had adopted statewide branching
regimes, either de novo or by acquisition, and the number of strict unit
banking states had dwindled to less than five. Historically, bank concen-
tration ratios have always averaged highest in statewide branching states,
followed by states with limited branching rules, and lowest in unit banking
states. Therefore, the trend toward national banking concentration that
Hanweck cites may reflect, in large measure, the shift in state banking
rules applying to intrastate expansion, and not primarily to interstate bank
acquisitions.
I'm also not as eager to discard the research evidence on economies of
superscale as Hanweck may be, particularly for those financial services
that appear to be uniquely important among the very largest banks,
including standby credit guarantees, corporate security underwriting,

correspondent banking, swap contracts, and other risk-hedging products-
to name just a few. Scale economies research to date has several glaring
holes and one of these centers upon the often enormous differences in
large bank and small bank product lines. If there is a significant con-
centration trend overall in financial services, and not just in traditional
banking, that trend may have an operating efficiency basis that we, with
our imperfect models and statistical tools, simply have not been able to
competently measure.
I do share Hanweck's concern with the path the federal government
and the states have recently chosen to deregulate geographic expansion in
banking. As he correctly notes, geographic financial-services deregulation
in the United States has, for all intents and purposes, devolved upon
the states because Congress has not yet moved to amend or abrogate the
McFadden and Glass-Steagall Acts or the Douglas Amendment to the
Bank Holding Company Act which prohibit full-service interstate expansion.
The states, for their part, have captured the interstate initiative with at
least forty-seven states now permitting some form of interstate expansion.
However, as Hanweck notes, interstate full-service expansion is limited
essentially to acquisitions. Neither de novo charters nor branching across
states lines are permitted in most instances and the latter is probably a far
less costly vehicle for expansion than acquisitions.
Such a restriction tends to limit interstate expansion to the largest and
most capital-endowed banking firms, due to the greater costs involved,
and may not generate the optimal market structure needed to spur
competition. There is, for example, the risk of dominant banking firms
stretching across the nation's landscape to confront each other in market
after market, practicing mutual forbearance and limiting competitive
rivalry among themselves. If, in fact, Hanweck is right in his assessment
of economies-of-scale studies, we would perhaps benefit more in a com-
petitive sense if we permitted de novo entry and full-service branching
across state lines, giving smaller banks a much better chance at full
participation in the interstate movement.
His article contains a very competent and well-written literature re-
view of the interstate field and it blends in rather neatly the data on
recent changes in bank returns and stock values with research evidence
from the period of most intense interstate banking activity.
What is somewhat disappointing about the article lies in what it does
not do with its central hypothesis. The author begins with the observation
that:
Recent changes in state laws with respect to interstate banking should be expected
to affect the valuation of commercial banks ... (and) it is likely that such
effects should vary with the location, nature of banks, and the type of state law
change ... As this paper reports, these changes in state law are not uniformly
reflected in bank valuation changes.
Certainly the article reports such a result but it does not rigorously test
that result. There is no model per se and there are no statistical tests (though
the time-series graphs on bank financial performance are both interesting
and well-presented). The closest that we get to a rigorous statement of
hypotheses about how interstate banking laws might impact the valuation
of banks and rigorous testing of those hypotheses is a summary of an
unpublished article prepared earlier by Goldberg, Hanweck, and Sugrue
(1990). This recent unpublished work is deemed to support the view that
interstate banking legislation has mostly benefitted those banks situated
in states that have recently changed their interstate banking laws and
banks located in the same region as the state that made the change. The
rationale given for this finding is that new interstate banking laws in-
crease the number of acquiring firms and broaden geographic expansion
opportunities sufficiently to outweigh the value-reducing impact of greater
competition once interstate entry is allowed. From this we are led to the
general conclusion that "the liberalization of interstate banking will have
a positive impact on the future ability of banking companies to increase
their profitability and value."
These hypotheses may be true (and there is a substantial body of
financial theory with which they appear to be consistent), but there is no
way of evaluating the quality and power of the author's test of them
without seeing the seemingly unrelated regressions model and the sampling
scheme to which that methodology is applied. Moreover, as plausible as
Hanweck's hypotheses are, not all the evidence that we currently have is
totally consistent with the conclusions reached.
There is, for example, the argument that interstate bank expansion is
a selective process; only certain kinds of banking firms are likely to par-
ticipate in it. Indeed, this paper provides us with one of the reasons for
this phenomenon: the peculiar nature of interstate banking laws passed
during the 1980s. The banks involved thus far in interstate acquisitions
have apparently not been "average" in a number of key structural
characteristics or performance dimensions (such as asset returns, loan-
loss risk exposure, and operating efficiency). A recent study by this au-
thor (May 1990) finds that, on average, both interstate acquirers and their
targets have tended to perform less well than comparably sized firms. It
is not intuitively obvious that the announcement of an interstate acquisition
of a below average performing banking firm by another below average

performer will generate positive returns. Perhaps capital market investors
have seen too many acquisitions by major banking firms go awry to believe
that positive returns will automatically follow the passage of a new interstate
banking bill. Or, perhaps these investors have read an excellent article by
Goldberg and Hanweck (1988) that does not emerge with optimistic
findings regarding the performance of grandfathered interstate banking
firms.
About a year ago this author conducted a study of the premiums paid
by about five hundred acquiring banks involved in acquisitions and merg-
ers over the 1978 to 1987 period (February 1990). The study drew upon
the developing theory of auctions and bidding to estimate the impact on
bank merger pricing from both structural factors (such as changes in state
banking laws) and individual bank risk-return measures. As expected, this
study suggested that individual firm profitability and risk are important
in explaining a substantial proportion of the variability in observed bank
merger premiums. But there was something else-evidence of a statisti-
cally significant negative relationship between interstate banking legisla-
tion and merger premiums posted. There is a branch of auction and bidding
theory that argues that the winning valuation of an asset subject to sale
at auction depends not only on how many bidders (acquirers) there are,
but also upon the supply of alternative assets available for auction. It may
be that a change in state banking law that significantly alters both the
number of potential acquisition targets, as well as the number of potential
acquirers, will yield value effects that could be either positive or negative-
again, depending upon the relative change in potential acquirers versus
acquisition targets. I certainly agree with Hanweck's overall observation
that the character of a state's interstate banking law can have a significant
impact on bank asset valuation, but, the direction of that impact may be
ambivalent. More research is clearly needed here.
The other point about that earlier auction and bidding study that seems
relevant here is that the merger sample used in that study contained a
substantial number of banks from states suffering severe economic dislo-
cations in the wake of the energy, farm, and real estate crises of the mid-
1980s. Many of these states passed reciprocal interstate banking laws at
the time when the market values of their banks were falling sharply. The
condition of each state's economy may well be an important valuation
factor that could also account for the estimated negative link between merger
premiums and the presence of new interstate banking legislation. This,
too, is an area where more research is clearly warranted.
Overall, the article on interstate banking by Gerald Hanweck makes a
worthwhile contribution to the literature and especially to discussions of

appropriate public policy in the banking structure field. It forcefully re-
minds us that the structure of banking in the United States is changing
rapidly without benefit of a cohesive national policy that places the welfare
of the consuming public first and without full awareness of the consequences
for competition and efficiency that only careful research can provide.
References
Goldberg, L.G. and G.A. Hanweck. 1988. "What Can We Expect from Interstate
Banking?" Journal of Banking and Finance 12: 51-67.
Hanweck, Gerald A. 1990. "Interstate Banking, Bank Expansion and Valuation."
Proceedings of the Fifteenth Annual Economic Policy Conference, Federal
Reserve Bank of St. Louis. (October 18)
Rose, Peter S. 1989. The Interstate Banking Revolution: Benefits, Risks, and Tradeoffs
for Bankers and Consumers. Westport, Conn.: Quorum Books.
--.1990. "Bidding Theory and Bank Merger Premiums: The Impact of Structural
and Regulatory Factors." Unpublished Paper, Department of Finance, Texas
A&M University (February).
- - . 1990. "The Consequences of Interstate Banking Deregulation for Competition,
the Structure of Service Markets, and the Performance of Interstate Financial-
Service Firms." Proceedings of the Conference on Bank Structure and
Competition. Federal Reserve Bank of Chicago (May).
3 THE MARKET FOR HOME
MORTGAGE CREDIT: RECENT
CHANGES AND FUTURE
PROSPECTS
Patrie H. Hendershott
Three major changes occurred during the 1980s in the market for home
mortgage credit: the securitization of fixed-rate mortgages, the develop-
ment of a national primary market for adjustable-rate mortgages, and the
decimation of the savings and loan industry. These changes and their im-
pacts on various financial industries and homebuyers are the subjects of this
article. The three major changes are summarized before a detailed analy-
sis is presented.
The Federal Housing AdministrationlVeterans Administration (FHA/
VA) fixed-rate mortgage (FRM) market was integrated with the capital
market gradually throughout the 1970s via increased usage of the Ginnie
Mae pass-through program, and the conventional FRM market was in-
tegrated in the first half of the 1980s with the development of active
markets for the mortgage pass-through securities of Freddie Mac and
Fannie Mae. Integration was stimulated in the 1980s by the deregulation
of deposit rate ceilings and the erosion of thrift tax subsidies, two de-
velopments that switched thrifts from low cost funders of mortgages to
high cost funders. As a result, coupons on new issue FRMs now swiftly
adjust to changes in other capital market rates, and mortgage funds are
readily available at all times and places at going market rates.
99
Nationwide authority to originate adjustable-rate mortgages (ARMs)

was not given to federally chartered thrifts until 1979, and ARMs with
loose per adjustment period and life-of-Ioan rate caps were not permitted
until 1981. In response to this authorization, ARMs accounted for two-
fifths of the total conventional loan volume originated by all lenders over
the 1984 to 1989 period, and by 1989, half of the thrift home loan portfolio
was in ARMs. The widespread existence of ARMs reduced the previous
cross subsidy from more mobile to less mobile households and increased
the ability of many households to qualify for larger loans.
The large losses of the savings and loan (S&L) industry in the 1980s
stemmed first from badly mismatched asset and liability portfolios in the
late 1970s and early 1980s, when interest rates skyrocketed, and then
from rapid S&L growth and accelerated investment in nontraditional, and
expost nonprofitable, real estate loans in 1983 and 1984. In the 1985 to
1988 period, the industry grew at a relatively slow rate, and since passage
of The Financial Institutions Reform, Recovery and Enforcement Act of
1989 (FIRREA), the industry has been in freefall. Between the end of
1984 and the end of 1990, the S&L share of home mortgages outstanding
has declined from 43 percent (also the average share in the second half
of the 1970s) to 25 percent. As I will discuss, the S&L problems in the
early 1980s substantially raised conventional FRM rates, but the recent
decline in the S&L home mortgage share has not raised FRM rates, at
least for the 90 percent of FRMs that can be sold to Fannie Mae and
Freddie Mac.
The remainder of this article is divided into six sections. The first two
describe the development of mortgage securitization and its impact on
mortgage rates. The next two document the growth in the ARM market
and the decline in the S&L industry. Section five speculates on future
prospects for the market for home mortgage credit, and Section six sum-
marizes the paper.
The Development of Mortgage Pass-Through Securities
In 1968, the Government National Mortgage Association (Ginnie Mae)

was formed within the United States Department of Housing and Urban
Development to administer government mortgage support programs. Two
years later Ginnie Mae began guaranteeing mortgage-backed pass-through
securities, Government National Mortgage Association Securities (GNMAs)
representing shares in pools of FHA/V A loans. Investors in pass-throughs
receive a pro rata share of the payments, both scheduled and early (in the
THE MARKET FOR HOME MORTGAGE CREDIT 101
event of prepayment or default), on the underlying mortgages. While

investors in whole FRANA loans are insured by FRA or VA against loss
of principal and interest, investors in GNMAs are guaranteed the full
timely payment of principal and interest.
In 1970, the Federal Rome Loan Mortgage Corporation (Freddie Mac)
was chartered to spur the development of a secondary market for con-
ventional mortgages. Freddie Mac introduced the first conventional mort-
gage pass-through in 1971, the Mortgage Participation Certificate (PC).
Because the underlying conventional mortgage is not itself fully insured,
in contrast to FRANA loans, the Freddie Mac guarantee adds more
value to the underlying mortgage than does the Ginnie Mae guarantee,
even though Freddie Mac doesn't have a full faith and credit Federal
guarantee. In 1981, the Federal National Mortgage Association (Fannie
Mae) initiated a conventional mortgage-backed security (MBS) program
similar to Freddie Mac's PC program. Fannie has intermediated in the
more traditional sense, buying mortgages and issuing its own debt since
1938, and it has an implied guarantee comparable to Freddie Mac's.
The size of the FRANA and conventional loan volume that can be
securitized by Ginnie Mae and the sponsored agencies (Fannie and Freddie)
is restricted by limits on the dollar value of loans that can be pooled into
the various pass-through securities. The dollar limit on GNMAs follows
from the limit on the underlying FRA and VA loans. The 1990 limit
varies regionally from $67,500 to $124,750. The lower limit was not changed
in the 1980s, but the upper limit, which was $90,000 between 1980 and
1987, was increased in both 1988 and 1989. Moreover, limits were raised
in numerous regions many times during the 1980s. The dollar limit on
conventional loans that Fannie Mae and Freddie Mac can purchase, the
"conforming" limit, changes annually with a house price index but does
not vary regionally. The 1989 limit was $187,600, up 63 percent since 1985
(the limit was virtually unchanged in 1990). In 1987, over 90 percent of
home mortgage loans (80 percent of dollar volume) were eligible for
pooling by the agencies, and this percentage has been fairly constant.
The markets for fixed rate FRANA and conforming conventional loan
pass-throughs developed at different rates. Table 3-1 presents data on the
growth in the securitization of fixed-rate FRANA loans. The importance
of securitization to the new origination market is measured as the ratio of
GNMA issues backed by single family loans to total originations of these
loans (Ginnie Mae is prohibited from securitizing FRANAs over eighteen
months old). By the second half of the 1970s, two-thirds of FRANA
originations went into GNMA pools; by the early 1980s four-fifths did;
and since 1982 all FRANA have gone into GNMAs.!
Table 3-1. The Growth in the Securitization of FHAIVA 1-4 Family Loans.
1 2 3 = 2/1 4
FHA/VA GNMA Share of Mortgage
Originations Issues Originations Banker Share of
($ bi/.) ($ bi/.) Securitized Originations
1971-73 15.6 2.7 .17 .70
1974-75 13.5 5.8 .43 .75
1976-79 28.3 17.6 .62 .78
1980-82 21.5 16.6 .77 .81
1983-86 55.0 55.6 1.01 .78
1987-89 54.9" 70.3 1.28" .71"
• Mortgage banker issues are likely understated. Thus originations and the mortgage
banker share are too low, and the share of originations securitized is too high.
Source: Hendershott and Van Order (1989), updated from DataBase, Secondary Mort-
gage Markets, FHLMC.
As the data in table 3-2 indicate, the pass-through market for conforming
conventional loans developed less rapidly. The best measure of the agencies'
presence in this market is the share of new fixed-rate conventional FRMs
(generally defined as less than one year since origination) eligible for
agency securitization (under the conforming limit) that is, in fact, securitized
by Freddie Mac and Fannie Mae. The upper part of table 3-2 lists total
agency pass-through issues, those backed by seasoned FRMs, new ARMs
and multifamily loans, and, by subtraction, new FRMs. The lower part
puts the agency issues backed by new FRMs into a market perspective.
Total single family conventional originations are multiplied by an esti-
mate of the fraction that had a fixed rate, and the result is divided by
agency issues backed by new FRMs to obtain the agency share of the
total FRM market. Of course, the agencies can only participate in the
conforming end of the market. Assuming this end to be 75 percent of
total FRM volume, the last column is an estimate of the percentage of
new FRM conforming loan volume that is sold directly to Fannie and
Freddie. As can be seen, this estimate rose from 4 percent in the 1977 to
1981 period, to almost 25 percent in the 1982 to 1985 period, and to over
50 percent since 1986, including 69 percent in 1989.
The difference in the development of FHAIVA and conventional pass-
throughs in the 1970s and early 1980s stems largely from the historical
differences in the origination of FHAIV A and conventional mortgages.
Mortgage bankers have tended to dominate the FHA/V A market,
Table 3-2a. The Agencies Increased Role in the Conforming FRM Market.
Pass-Throughs Issued by FHLMC and FNMA ($ bU.)
Pass- Throughs Backed-By
1 2 3 4=1-2-3
Total Seasoned New ARMs, Multis, New
Issues FRMs andFHAIVAs FRMs
1977-1981 4.6 0.6 4.0

1982 38.2 28.8 9.4
1983 33.0 17.1 1.8 14.1
1984 32.2 17.7 3.7 10.8
1985 62.3 25.5 5.1 31.7
1986 160.1 29.7 10.7 119.7
1987 138.2 21.2 16.3 100.7
1988 94.7 4.5 29.8 60.4
1989 145.1 11.7 36.2 97.2
Table 3-2b. Percentage of New 1-4 Family Conventional Originations Securitized.

5 6 7=5x6 8=417 9 = 8/0.75
% New
Total Fraction FRM % New Conforming
Originations Fixed Originations FRMs FRMS
($ bil.) Rate ($ bil.) Securitized Securitized
1977-1981 125.0 1.00 125.0 3 4
1982 77.8 0.65 50.6 19 25
1983 154.2 0.69 106.4 13 17
1984 176.0 0.46 81.0 13 17
1985 204.6 0.55 112.5 28 37
1986 362.1 0.77 278.1 43 57
1987 376.2 0.66 248.3 41 55
1988 338.5 0.47 159.1 38 51
1989 296.0 0.63 186.5 52 69
Source: Hendershott (1990a, Table 4).
accounting for 70 to 80 percent of originations (see table 3-1) versus only

7 percent of conventional originations prior to 1982 and 15 percent since, and
they sell virtually all their originations to other investors. Thus, when an
improved method for selling mortgages became available, mortgage bankers
quickly took advantage of the opportunity. By the early 1980s, virtually
all of mortgage banker originations were sold into Ginnie Mae securities
or to Fannie Mae and Freddie Mac (some conventionals were sold to
Fannie Mae for its portfolio). In contrast, depository institutions have
dominated conventional originations (80 percent to 90 percent), and they
have tended to keep their originations as portfolio investments. Thus, an
improved selling method alone was not sufficient to stimulate the conven-
tional pass-through market.
Portfolio restrictions on savings and loans (no corporate loans, bonds,
or equity issues) encouraged investment in residential mortgages prior to
the 1980s. Moreover, these investments were especially profitable to thrifts,
owing to special tax advantages. The tax preference was the ability of
thrifts to compute loan-loss reserves that far exceeded a reasonable
provision for normal losses, as long as thrifts invested a large fraction of
their assets in housing related loans or liquid assets (Hendershott and
Villani, 1980, Appendix). In effect, thrifts were allowed to transfer large
portions of their before-tax income to reserves, thereby avoiding taxes.
Between 1962 and 1969, the transfer was limited to 60 percent of taxable
income; between 1969 and 1979, the fraction was gradually reduced to 40
percent; the Tax Reform Act of 1986 lowered the fraction to 8 percent.
The incentive provided by the extraordinary loan-loss provisions for
investment in residential mortgages depends on the expected level of
thrift taxable profits over the expected life of the investment (with no
profits now or in the future, the incentive is zero), the income tax rate,
and the statutory fraction of income that can be transferred to reserves.
Assuming a 1 percent net pretax return on assets, the incentive was
substantial in the 1960s and 1970s (Hendershott and Villani, 1980). In the
1960s when the transfer fraction was 60 percent, savings and loans would
have accepted a three-quarters percentage point lower pretax return on
tax preferred housing-related assets than on comparable non-preferred
assets. By 1979, when the transfer fraction was down to 40 percent, they
would have accepted a half percentage point less.
Two major factors have driven the increase in conventional loan
securitization in the 1980s. First, thrifts maintained their share of mort-
gage originations but reduced their relative investment in home mort-
gages (have sold some of the originated mortgages). Most strikingly, the
share of savings and loan total assets in home mortgages and agency se-
curities (largely Fannie and Freddie pass-throughs) fell from 72 percent

to 59 percent during the 1982 to 1984 period. This portfolio shift reflected
the reduced profitability of savings and loans, first due to high interest
rates and a maturity mismatch and then due to disinflation, credit losses,
and the expansion of savings and loan asset powers. The reduced
profitability eroded the tax incentives for residential mortgage investment,
while the expansion of powers encouraged thrifts to invest more widely.
Second, a Freddie Mac PC or Fannie Mae MBS is excellent collateral for
borrowing through FHLB advances and security repurchase agreements,
and in the 1980s these became cheaper marginal sources of funds than
deposits for many savings and loans. During the 1984 to 1988 period,
savings and loans increased such debt by over $150 billion; that is, some
loans were simply swapped for pass-throughs, and the pass-throughs were
retained in portfolio and "repoed" or used to increase advances.
Recent years have also marked a surge in the resecuritization of fixed-
rate mortgage-backed securities, the slicing up of these securities into
different maturity tranches (CMOs or REMICs) or into interest or principal
strips (lOs and POs) . The first resecuritization (issuance of a multiclass
mortgage security) was a Freddie Mac CMO issue in 1983. Issues have
risen from $16 billion in 1986 to $100 billion in 1989. Initially, issues were
dominated by private sector participants; Fannie Mae and Freddie Mac
accounted for only 2 percent of issues in 1986 to 1987. By 1989, though,
these agencies accounted for 83 percent of issues. Throughout the 1986 to
1989 period, pass-through securities, rather than whole mortgage loans,
have been the collateral for 96 percent of multiclass issues.
The securitization of conventional conforming ARMs by Fannie Mae
and Freddie Mac is less prevalent. It appears that only 2 percent to 3
percent were securitized in 1984 to 1985 and that the percentage is still
only 10 percent to 12 percent. The greater securitization of fixed-rate than
adjustable-rate mortgages likely reflects both the greater standardization
of FRMs than ARMs and the greater desire of originators to hold ARMs
than FRMs in their portfolio.
Finally, some investment banks and large thrifts also securitize home
mortgages, but these institutions largely (possibly exclusively) limit
themselves to nonconforming or jumbo loans and they likely securitize
only 10 percent to 20 percent of the market (Woodward, 1988). Fully
private mortgage securitizers cannot compete with Fannie Mae and Freddie
Mac for two reasons: (1) The agencies do not have to maintain as much
capital as fully private institutions; and (2) The agencies' costs of securitizing
are lower. The latter stems from lower explicit costs (exemption from
SEC requirements, state and local taxation, etc.) and from the economies-
of-scale achieved as a result of being the low-cost securitizer in a large

part of the market.
The Impact of Securitization on Mortgage Coupons
A clear implication of mortgage securitization is that mortgage yields

should be more closely connected to capital market rates with securitization
than without. The expected impact of securitization on the general level
of mortgage rates is more complicated. Empirical evidence relating to
each of these impacts is discussed below.
Timing of Conventional Mortgage Rate Adjustment to

Capital Market Rates
The perfect mortgage market model says that the coupon rate for a near-
par mortgage depends on a small number of general capital market
variables, such as the six-month and seven-year Treasury rates and the
volatility of these rates (Hendershott and Van Order, 1989). Moreover,
the response to changes in those variables is predictable and fast. In con-
trast, twenty years ago mortgage lending was tied to the thrifts, which
were heavily regulated and also tax-advantaged to invest in mortgages,
and connections of mortgage and capital markets were tenuous and gradual.
Hence, most researchers focused on things peculiar to the thrift industry,
such as deposit rates and deposit flows, rather than general capital market
conditions as determinants of mortgage rates. If one regressed actual
mortgage rates during such a: period on fictional mortgage rates predicted
by the perfect market model, one would expect to obtain a poor fit.
Moreover, to the extent that the predicted rate had any effect, it would
be a lagged one.
Roth (1988) analyzed the integration of mortgage and capital markets
by looking at trends in the month-to-month correlation of changes in
coupon rates on conventional mortgage commitments and ten-year Treas-
uries annually from 1972 to 1987. His results are reproduced and extended
to include 1988 and 1989 in table 3-3. Prior to 1982, the correlation of the
changes ranged from -0.5 to +0.5 and was never statistically different from
zero. After 1981, though, the correlation was never less than 0.58 and was
always statistically positive. Moreover, in the last three years the correla-
tion has averaged 0.90.
A potential problem with Roth's analysis is that the mortgage rate
Table 3-3. Correlation Between Mortgage Rates and Capital Market Rates.
Year Correlation*
1972 -0.22
1973 0.19
1974 0.46
1975 -0.18
1976 0.16
1977 -0.49
1978 0.42
1979 0.34
1980 0.33
1981 0.42
1982 0.80**
1983 0.81**
1984 0.65**
1985 0.76**
1986 0.58**
1987 0.91**
1988 0.87**
1989 0.92**
* Correlations are between month-to-month changes in the FHLMC commitment rate
and the 10-year Treasury rate.
** Significantly different from zero at a 5 percent confidence level.
Source: Roth (1988, Table 1) for 1972-87; 1988 and 1989 computed by the author.
incorporates a call-premium while the Treasury rate does not, and in

some periods the value of the call-premium may have changed markedly,
possibly disguising a close relationship between the non-call component
of the mortgage coupon and the Treasury coupon. Hendershott and
Van Order (1989) attempted to correct for this potential problem by
constructing a perfect mortgage-like capital market rate and estimating
the adjustment of conventional commitment mortgage rates to this per-
fect rate (rather than to a Treasury rate). The analysis consisted of two
parts. First, they estimated a price equation for GNMAs, set the price
equal to the new-issue price, and solved the equation for the perfect-
market retail coupon rate. Second, they regressed conventional commit-
ment mortgage coupon rates on current and past values of the estimated
perfect-market coupon rate taken from the GNMA equation.
The price equation was estimated on weekly GNMA price and coupon
data from the January 1981 to July 1988 period. In this equation, the
Table 3-4. The Time Response of Conventional Commitment Rates to Fictional

Perfect Market Rates.
Adjustment to One Point Rise in Perfect Rate
Time Period Current 3 weeks 5 weeks 7 weeks 9 weeks
1986-88 .59 .95 .96 .87 .84
1983-85 .16 .55 .68 .83 .88
1980-82 .08 .45 .75 .93 1.05
1976-79 .01 .36 .62 .66 .86
1971-75 .06 .17 .37 .56 .74
Source: Hendershott and Van Order (1989, Table 5).
GNMA price was regressed on the coupon (adjusted to a bond-equivalent

basis), the seven-year Treasury rate, and two determinants of the value of
the borrower's call option-the term structure slope (seven-year rate less
six-month rate) and an estimate of the volatility of the seven-year rate.
Various interactions of these variables were included to allow for nonlinear
price responses.
To obtain the perfect-market rate, the estimated price equation was
solved for the coupon rate after the mortgage price was set equal to one
hundred, less the actual points charged in the conventional market (less
one point presumed to equal origination costs). This coupon was then
converted to a mortgage (rather than bond-equivalent) basis, and fifty
basis points were added for servicing and other costs. As the degree of
integration increased, changes in the perfect-market coupon rate should
have been reflected more quickly in the conventional commitment rate. 2
Retail conventional commitment rates were regressed on the current
and lagged one-to eight-week values of the perfect-market rates for various
parts of the 1971 to 1988 period. Table 3-4 reports the cummulative
adjustment of the commitment rate concurrently and over two, four, six,
and eight week lags. The shift toward integrated markets is striking. The
percentage of the change in the perfect-market rate that is reflected in-
stantaneously in the retail conventional rate rose monotonically from
effectively zero in the 1970s to 8 percent in the 1980 to 1982 period, 16
percent in the 1983 to 1985 period, and 59 percent in the 1986 to 1988
period. The fraction of the change in the perfect-market rate reflected in
the conventional rate within two weeks rose monotonically from a sixth
in the first half of the-1970s, to almost half in the early 1980s, to over half
in the 1983 to 1985 period, and to nearly one in recent years.
Table 3-5. Actual and Perfect Market Effective Conventional Commitment Rates
(%).
Actual Perfect Market Difference

1971 7.54 8.33 -.79
1972 7.38 7.92 -.53
1973 8.04 8.97 -.93
1974 9.19 9.78 -.60
1975 9.05 9.92 -.87
1976 8.86 9.22 -.35
1977 8.84 9.09 -.24
1978 9.64 10.08 -.44
1979 11.20 11.34 -.14
1980 13.76 14.24 -.48
1981 16.69 16.55 .13
1982 15.97 15.24 .73
1983 13.23 12.86 .37
1984 13.89 13.52 .37
1985 12.43 11.95 .48
1986 10.19 9.69 .49
1987 10.21 10.01 .20
1988 10.23 10.21 .02
Source: Hendershott and Van Order (1989, Table 6).
Securitization and the Level of Mortgage Rates
Probably the best starting point is a comparison of the actual conven-

tional commitment rate and the Hendershott-Van Order fictional perfect-
market rate. Table 3-5 lists annual values of these rates and the difference
between them for the 1971 to 1988 period. 3 The precise differences are,
of course, subject to some error: the actual rate is a survey rate, and the
perfect rate is computed from an empirical equation estimated with error.
Nonetheless, the overall pattern of the differences seems both systematic
and plausible. The actual rate was three-quarters of a percentage point
below the perfect-market rate in the 1971 to 1975 period; a third of a
point below in the 1976 to 1980 period; and roughly half a point above the
perfect rate in the 1982 to 1986 period.
As explained above, the low mortgage rates in the 1970s can be attributed
to tax advantages for thrift mortgage investments and portfolio restrictions
against nonmortgage investments, and the switch in the 1980s reflects a
Table 3-6. Effective Loan Rates for California FRMs with Loan-to-Value Ratios
of 75 and 80 Percent by Loan Size, Selected Years.
Percent of Conforming Loan Limit 1978 1985 1986 1987

0.0-50.0 10.12 11.75 10.65 9.83
50.1-67.0 10.04 11.87 10.53 9.82
67.1-80.0 9.97 11.98 10.51 9.77
80.1-90.0 9.97 11.66 10.40 9.63
90.1-100.0 9.95 12.22 10.36 9.62
100.1-115.0 9.94 11.13 10.62 9.63
115.1-130.0 9.97 11.46 10.65 9.91
130.1-145.0 9.95 11.39 10.68 9.80
Over 145.0 9.94 10.97 10.70 9.83
Number of Loans 3,590 710 1,157 1,706
Percent of Dollar Volume Securitized 4 37 57 55
Source: Hendershott (1990b, p. 156).
sharp relative shift of thrifts out of home mortgage investments owing to

the reduced (non) profitability of savings and loans and the expansion of
savings and loan asset powers. The half percentage point premium in the
early 1980s provided the incentive for the securitization of conforming
conventional FRMs. The premium covered the start-up costs of the
securitizers and the liquidity premium demanded by investors.
Beginning in the middle of 1987, the actual rate is very close to the
perfect-market rate, the conventional conforming mortgage market
seemingly being fully integrated into capital markets. Thus, as the volume
of mortgage pools grew, bid/ask spreads were bid down (and thus the
liquidity premium fell), and the per dollar costs of the securitizers declined.
This suggests that the rates on conforming loans, which are eligible for
purchase by the agencies, should have declined relative to rates on non-
conforming or jumbo loans.
The raw data displayed in table 3-6 suggest that beginning in 1986
yields on conforming loans were lowered relative to those on nonconforming
loans. Average effective rates are listed for loans of increasing size (percent
of the conforming loan limit) with similar loan-to-value ratios (75 percent
to 80 percent) originated in California in the May to June periods of 1978,
1985, 1986, and 1987. Holding the loan-to-value ratio and state of origi-
nation constant controls for default risk. California was chosen because it
accounts for roughly a quarter of the dollar value of all conventional
FRMs closed in the United States and more than half of all jumbo loans
(those over the conforming limit). The number of loans in the samples
and the percent of the eligible dollar volume securitized by FNMA and
the FHLMC are reported at the bottom of the table (see Hendershott and
Shilling, 1989, for more detail on the loan samples).
In general, one would expect the loan rate to decline with loan size
because the costs of originating and servicing loans decrease per dollar of
loan as the loan size increases. This is clearly the case for loans below the
conforming limit in all years except 1985, where the limited number of
observations results in much noise in the averages. Of most interest, though,
is what happens to the loan rate when the loan size increases above the
conforming limit. Prior to 1986, the loan rate is either flat (1978) or still
decreasing (1985). But in 1986, the loan rate jumps at, and in 1987 just
above, the loan limit; that is, rates on loans below the loan limit are
noticeably lower than those above the limit. The most likely cause is, of
course, the expanded activities of Fannie Mae and Freddie Mac. These
expanded activities probably also influence rates on loans just above the
limit because such loans are likely to be conforming within a year
(Woodward, 1988). Thus, the low rate for loans that are 100 percent to
115 percent of the limit in 1987 may not be as anomalous as it first appears.
Hendershott and Shilling (1989) estimated, separately on data for 1978
and 1986, the relationship between rates on loans closed and the follow-
ing factors: loan-to-value ratio, loan size, precise month the loan was
closed, dummy variables for geographic regions in the state, and whether
the loan was on a new property, was under the conforming limit, or was
just above the limit. The loan-to-value ratio had the expected positive
impact; the loan size and the new property dummies had the expected
negative impacts; and the responses in the two years were remarkably
similar. For those two years, however, the effects of the conforming limit
differed markedly. In 1986, conforming loans had a thirty basis point
lower rate than well-above-the-limit loans, and soon-to-be-conforming loans
had a fifteen basis point lower rate (standard errors were only five basis
points). In 1978, however, the point estimate for the conforming loan
coefficient was only three basis points.
It should be emphasized that the perfect-market rate listed in table
3-5 is computed from a GNMA price equation, not an equation explain-
ing prices on seven- or ten-year Treasury bonds. The working assumption
of Hendershott and Van Order (1989) was that the GNMA market has
been integrated with capital markets since 1981. This assumption seems
plausible because GNMAs have full faith and credit guarantees and have
traded like Treasuries, with comparably low transactions costs and high
volume, at least since 1981. However, some would argue that the
resecuritization phenomenon, particularly the surge in CMOs and REMICs,

has lowered GNMA yields, and thus conventional mortgage rates, rela-
tive to Treasury yields. This might suggest an additional quarter point
relative decline in the cost of FRMs since the first half of the 1980s (0 'Keefe
and Van Order, 1990).
The agency securitization explosion and the resultant reduction in FRM
yields has affected a number of financial firms importantly. Most obvi-
ously, the shareholders of Fannie Mae and Freddie Mac have benefitted.
One guide is the movement in Fannie Mae's stock price, which quadrupled
from 8 percent in early 1983 to 32 percent in the summer of 1990.4 Also
benefitting were those investors who were holding FRMs when the agen-
cies were lowering FRM rates and thus mortgage prices were rising. On
the other hand, the thirty basis point reduction in FRM yields hurts ongoing
portfolio lenders, both those investing in FRMs and those wishing to
originate more ARMs for their portfolios. The latter group includes both
S&Ls and commercial banks. Finally, securitization has led to greater
specialization in separate phases of the mortgage process; different firms
are concentrating on origination, servicing, investing and insuring (Fannie
and Freddie).
The Growth in Adjustable-rate Mortgages
Prior to the 1980s, there existed one mortgage type-the standard fixed-
payment mortgage (FRM). Periodically in the 1960s and 1970s, increases
in interest rates reminded thrifts of the problems of borrowing short and
lending long. At these points (1971 and 1974 specifically), federally char-
tered thrifts lobbied for permission to offer borrowers an alternative choice,
adjustable rate mortgages (ARMs), that would reprice more in line with
thrift deposits. Congress made clear to the regulatory body (then the
Federal Home Loan Bank Board) that it did not want borrowers to have
that choice (Cassidy, 1984). In December 1978, an exception was made
for federally chartered S&Ls in California, allowing them to compete
with state chartered S&Ls, and in July 1979, nationwide authority to
invest in ARMs was first granted. However, annual and life-of-Ioan rate
increases were restricted to one-half and two-and-a-half percent, respec-
tively, and the choice of rate index was greatly restricted.
Ironically, about the same time that interest rates were peaking in the
early 1980s, congress gave thrifts the opportunity to invest in ARMs that
might be attractive to both lenders and borrowers. In April 1981, fairly
liberal regulations were implemented for federally chartered S&Ls and
savings banks, and in August 1982, these were loosened further and ex-
tended via the Deposit Institutions Act to all state chartered institutions
(although individual states could override the regulations during a three
year period). Thrifts, indeed, took advantage of this opportunity. In the
middle of 1982, ARMs were only 10 percent of the single family mortgage
portfolio of FSLlC-insured S&Ls. By March 1989,48 percent of the thrift
single family loan portfolio (including mortgage-backed securities) was in
ARMs (Hendershott and Shilling, 1990). Moreover, over the 1984 to 1989
period, ARMs accounted for 43 percent of the conventional single family
loan volume originated by all lenders (computable from table 3-2).
The coupon on a FRM reflects the pure cost of "intermediate" term
(say seven-year) money plus the cost of a call or prepayment option.5 Altern-
atively, the FRM coupon can be thought of as the pure cost of short-term
money (say one-year) plus the cost of an intermediate term (say seven-year)
rate cap that will pay the borrower the difference between the actual short-
term rate and the initial one should interest rates rise. 6 The sole existence
of the FRM caused cross-subsidization among borrowers because all
borrowers pay virtually iq,entical costs for the call option or rate cap even
though the option/cap is certainly more valuable to less mobile households
than to more mobile households because longer lived options/caps have
greater value than shorter lived options/caps. That is, with the FRM, more
mobile households, who pay more for the call option/rate cap than it is
worth to them, subsidize less mobile households who pay less than it is
worth to them. The existence of adjustable-rate mortgages (ARMs) would
give mobile households an alternative to overpaying for the FRM and
would reduce the subsidy to less mobile households who choose the FRM.
A second potential advantage of ARMs for at least some borrowers
stems from the facts that initial coupon rates on ARMs are less than those
on FRMs and lenders qualify borrowers based on the lower rates. Thus,
borrowers financing with ARMs can obtain larger loans and purchase
larger houses. The latter suggests that the ARM share of conventional
single-family mortgage originations will vary cyclically as the FRM-ARM
rate spread varies, and figure 3-1 shows that this variation has occurred.
Large rate spreads (two-and-a-half percentage points) in 1984 to 1985 and
mid-1987 to the end of 1988 were associated with 40 percent to 60 percent
ARM shares, while small spreads (one-and-a-half points) in 1986 to mid-
1987 and late 1989 were associated with 20 percent to 25 percent shares.7
Another important determinant of the ARM/FRM choice is the level
of the FRM rate (Brueckner and Follain, 1989); the higher the FRM rate,
the more desperately do households need to obtain a lower initial rate on
which they can qualify for a reasonably sized loan. The sharply reduced
ARM Share Rate Spread

0.7 r-----....----,-----r-----.---~---__. 3.0
0.6
,,
Rate Lagged I
, I
,
I
One Quarter I
I
I 2.5
I
I
I
0.5 -- I
I
I
I
I
I
I
I
I
I
I
I
0.4 I
I
2.0
I
I
,.', I
I
I
.
l " I
,
I
I
I
0.3 ,
', I
\I
"•
1.5
0.2
0.1 L-l---1.....J.......J........L....L-..L..-.L....JL....l---1.....J.......L....L....L-..L..-J......JL....l---L.-L....J......L-...1-..L..J 1 .0
841 843 851 853 861 863 871 873 881 883 891 893 901
842 844 852 854 862 864 872 874 882 884 892 894
Source: FHLMC
Figure 3-1. ARM share and FRM minus ARM rate spread quarterly, 1984-1989.
ARM share in 1986 as related earlier years reflected, in part, the marked
decline in FRM yields in 1985 and 1986.
The successful securitization of the FRM market, but not the ARM
market, has likely reduced the FRM rate as related to the ARM rate. This
can significantly alter the ARM/FRM mix. For example, using an FRM
rate of 10.50, an ARM rate of 8.50 and the mean values of the other
demographic variables relevant to the ARM/FRM choice, Brueckner and
Follain's (1989) estimation equation implies a 23 percent probability of
this household choosing an ARM. However, a thirty basis point higher
FRM rate raises the ARM probability to 32 percent.
0.8r---------.--------r-------,--------~------__,
S&L Portfolio Share
0.7
0.6
0.5 .....................,
,---,' ' ..
" ' ...
"" .........,
.......... ----........... "S&L Share of Mortgage Market"------"
M ,,
......
--...,,,,
......
,,
0.3 ,
\,
0.2~ __ __
~ ~~~~ __ __ ~ __ __
L_~~_L __ ~ ~~ _L~
1965 67 69 71 73 75 77 79 81 83 85 87 1989
Source: Flow of Funds Accounts. First quarter of 1990 annualized.
Figure 3-2. Share of S&L TFA in home mortgages, S&L share of total home
mortgage market, and ratio of S&L TFA to total home mortgages outstanding.
The Savings and Loan Industry
The institutions that have been most affected by, or causally linked to,
changes in the home mortgage market in the 1980s have been savings and
loans (and Fannie Mae and Freddie Mac). A shrinkage in the savings
and loan industry in the very early 1980s and a relative shift of S&Ls out
of home mortgages in the 1982 to 1984 period raised conventional FRM
rates from one-half percentage point below perfect-market levels in the
1970s to one-half point above perfect-market levels in the early 1980s. This
induced securitization, which successfully lowered conforming FRM rates
back to perfect market levels.
Figure 3-2 provides data on both savings and loan behavior and the
relative role of S&Ls as home mortgage investors over the past quarter
century. The behavior of S&Ls is reflected in the proportion of S&L

assets invested in home mortgages either directly or indirectly through
holdings of agency securities. As can be seen, this proportion varied within
a fairly narrow 72 percent to 74 percent range until 1981, before plummeting
to 59 percent at the end of 1984. The ratio slipped further to 57 percent
at the end of 1987, but has since risen to 61 percent. The sharp decline in
1983 and 1984 reflected accelerated growth in the S&L industry (18 per-
cent annual growth rate), not an actual shift out of mortgages.
The S&Ls presence in the home mortgage market is measured as the
ratio of S&L total (direct and indirect) home mortgage holdings to total
home mortgage debt outstanding. This presence is the product of the
other two series in the figure, the fraction of S&L assets invested in home
mortgages and the ratio of S&L total assets to the book value of all
outstanding home mortgages. Beginning at 0.43 in 1965, the S&L presence
rose gradually throughout most of the 1970s, reaching a peak of 0.51 in
1977. Since then, the S&L presence has been halved. The increase between
1969 and 1977 and subsequent decline through 1981 reflected swings in
the size of the S&L industry relative to the size of the home mortgage
market (recall that the S&L mortgage portfolio share was constant over
this period). The ratio of S&L total financial assets to total home mortgage
debt outstanding rose from 0.56 in 1969 to 0.70 in 1977, before declining
to 0.63 in 1981.
While S&L total financial assets grew between 1984 and 1988, they
grew at a slower rate than the home mortgage market; with a constant
S&L mortgage portfolio share, the S&L share of the home mortgage
market fell from 73 percent to 63 percent. Since early 1989, S&L assets
have been shrinking rapidly; in just a year-and-a-half the ratio of S&L
total assets to home mortgage debt has fallen from 0.36 to 0.25. In this
year-and-a-half, S&Ls have liquidated nearly $90 billion in agency securities
and over $50 billion in direct home mortgage holdings. The recent decline
is, of course, related to the Financial Institutions Reform, Recovery and
Enforcement Act of 1989 (FIRREA), although that act was itself largely
a belated recognition of economic realities.
FIRREA has offsetting effects on home mortgage investment by S&Ls.
The act strongly encourages such investment in two ways. First, a
strengthened qualified thrift lender (QTL) test directly mandates more
investment. S&Ls must now keep 70 percent of assets in qualified loans
rather than 60 percent, and fewer non-housing-related loans are now clas-
sified as qualified than was previously the case. Second, restrictions on
nonhousing related loans are substantially increased. For example, limi-
tations are tightened on commercial real estate loans, tangible personal
property holdings, commercial and corporate debt, and wholly-owned

service corporations, and existing junk bond holdings must be liquidated
entirely. While some of these provisions are not yet fully in force Gunk bonds
do not need to be fully liquidated until July 1, 1994, although regulators
are forcing undercapitalized thrifts to liquidate sooner), the decline in the
home mortgage portfolio share of S&Ls has already been arrested.
The negative effect of FIRREA on S&L home mortgage investment
comes from the increased capital requirements that are already leading to
a downsizing of the S&L industry. S&Ls are shrinking and being sold off
wholesale to nonS&Ls. To date, this second effect, the downsizing, has
dominated and the S&L role in home mortgage financing has plummeted.
It is noteworthy, though, that interest rates on conforming home mortgages
do not appear to have risen relative to Treasury rates, which is what one
would expect given the full integration of the market. Figure 3-3 contains
monthly average values for the Freddie Mac commitment rate and the
seven-year Treasury rate. The commitment rate has continuously tracked
the seven-year rate, and the difference in 1990 is about the same as in 1988
and is less than in 1989. While simply comparing these rates is generally
inappropriate because the call premium built into mortgage coupon rates
can change, sharp changes in the slope of the term structure and the vola-
tility of interest rates do not appear to have occurred in the 1988 to 1990
period. Whether ARM rates or jumbo FRM rates have been affected is
unknown.
A quick look at the policy mistakes that caused the crippling of the
S&L industry is probably worthwhile. The Federal Savings and Loan
Insurance Corporation (FSLIC) debacle is generally viewed as occurring
in two stages (Kane, 1989). First, sharply rising interest rates easily
eliminated the net worth of most S&Ls owing to their asset liability
mismatch (borrowing short and lending long). This mismatch was, of course,
aided and abetted by congress, which prevented the widespread introduction
of ARMs by a decade. Second, S&Ls then took substantial risks (doubled
their bets) in the 1980s, as one might expect. The latter was made easier
by the increase in deposit insurance coverage from $40,000 to $100,000
per account, the use of brokered deposits, and the enactment of new asset
powers (including additional flexibility in writing mortgage contracts).
Regulatory forbearance and loose oversight, this time aided and abetted
by both Congress and the Administration, encouraged this risk-taking and
led, in conjunction with the generous tax provisions of the 1981 Tax Law,
to substantial overbuilding throughout the United States. As a result,
"good" real estate investments have turned bad, and commercial banks
and other real estate investors are starting to take large real estate losses. 8
11.5~1--------------------------------------------------------------------------------~
10.5 FRM Commitment Rate
9.5
7 Year Treasury Rate //"'\
,..._./ \
/
A\ /
/ \
/ \. / \
/
r' . . . / ,v/ \\ /
//...... ---."\
/ \ / - -
8.5 / \ /
\ / \ /
\ / \ /', /
\ // \y/.... " /
~ ~~
7.5 I I I I
0188 0388 0588 0788 0988 1188 0189 0389 0589 0789 0989 1189 0190 0390 0590 0790
Figure 3-3. FRM and treasury rates.

The cost to taxpayers and investors of forebearance is truly going to be

enormous.
Complicating matters is the erosion in the basic profitability of S&L
mortgage portfolio lending. Owing to increases in the costs of deposits
and advances relative to Treasury rates, S&Ls haven't been the low cost
supplier of home mortgage credit for some time (Hendershott, 1990b).
With a low profit stream, untoward events (credit problems, rising inter-
est rates) quickly reduce capital, rather than just lowering dividends, and,
with little capital, shortly increase taxpayer liabilities.
We are now supposedly solving the S&L problem by recapitalizing or
shrinking (forbearance is out), reregulating (new assets powers are out)
and reintroducing strict oversight. It is noteworthy, however, that the
original source of the problem, the vulnerability of S&Ls to periods of
sustained increases in interest rates, has not been removed. In early 1989,
S&Ls were still using roughly 40 percent of their short-term deposits to
fund long-term fixed-rate mortgage investments; the $400 billion so funded
is slightly more than was so funded in 1978. Moreover, many of their
adjustable-rate mortgages have rate caps that will bind in a period of
sustained interest rate increases. If interest rates should repeat their 1977
to 1986 pattern, taxpayers would certainly lose another $50 billion or
more in present value dollars (Hendershott and Shilling, 1990).
Future Prospects
Three questions follow from our earlier analysis:
1. Will the advantages of Fannie and Freddie be reduced, possibly

leading to an increase in FRM rates and to greater fully-private
securitizing?
2. Will there be an increase in ARM securitization, leading to lower
ARM yields and expanded use of ARMs?
3. What will happen to the S&L industry and what impact will this
have on mortgage rates?
These questions are addressed in turn.

As noted, the agencies have not had to hold as much capital as fully
private institutions and the agencies have lower explicit costs. User fees
have been proposed to compensate the government/taxpayer (the in-
surers of the agencies) for the greater risk associated with lower capital
requirements and to recapture for the taxpayers some of the advantages

extended. Alternatively, increases in required capital have been advocated
to reduce the insurer's risk. Each of these proposals would tend to raise
the coupon rate on new-issue conforming FRMs and thus reduce the tilt
of new originations from ARMs to FRMs.
To the extent that the agencies can fully pass through their costs to
borrowers, the fifteen basis point annual user fee on agency mortgage-
backed securities (MBSs) proposed by the Bush Administration would
raise FRM rates by fifteen basis points. Because not all of the cost is
necessarily shiftable, maybe rates would only rise by ten basis points. A
doubling of capital from roughly 0.75 percent of assets to 1.5 percent,
which would seem to be the minimum required if the Treasury's proposal
were adopted, would have about the same effect. Maybe half of the thirty
basis point fee required by the agencies (the differences between the
mortgage rate, net of servicing, the borrowers pay and the yield investors
in MBSs receive) is return on equity, with the rest covering explicit costs
and a fair charge for the default guarantee the agencies provide MBS
investors. The doubling in capital would then necessitate a fifteen basis
point increase in the agency fee, and thus in FRM rates, if the required
return on capital investment in the agencies did not change. However, the
required return will be less because the agencies would be only half as
levered. Thus, again, about a ten basis point rise in FRM rates would
occur.
Securitization of ARMs would be stimulated by a more attractive FHA-
ARM. Currently, FHA-ARMs constitute only about 1 percent of the total
ARM market because the FHA's one-fifth program (maximum one per-
centage point rate adjustment per period and five percentage points over
the life of the loan) is unattractive to both borrowers and lenders. Bor-
rowers dislike the one-fifth loan because lenders charge a higher initial
coupon to compensate for the tight rate caps, and the primary advantage
of an ARM to a borrower is a low coupon that increases affordability. If
FHA were given a 2/6 ARM program, the FHA share of the ARM
market would rise significantly.
Successful securitization of an FHA-ARM would likely lead to
securitization of a similar conventional ARM. However, never should we
expect as great ARM as FRM securitization. Possibly the most popular
one or two ARMs will become widely (above 50 percent) securitized, but
other less popular products will always exist and these will not have suf-
ficient volume and thus liquidity to be widely securitized. Nonetheless the
secondary market would provide a ceiling for yields on the popular ARMs
and this, in turn, would provide some discipline for yields on other ARMs.
The decline in the S&L industry owing to FIRREA has just begun. A
further 30 percent to 50 percent decline over the next five years should
be expected. While some of the S&Ls will be sold to commercial banks,
many banks also face increased capital requirements, and all face higher
deposit insurance premiums. The depository institutions sector generally
is likely to shrink over the next five years. On the other hand, there is no
evidence that the relative decline in the S&L industry since 1984 or the
absolute decline since early 1989 has raised home mortgage rates relative
to capital market rates.
Conclusions
The market for fixed-rate mortgages is now fully integrated with capital
markets. Since 1986 the share of newly originated conventional conform-
ing fixed-rate mortgages securitized by the Fannie Mae and Freddie Mac
has ranged from 50 percent to 70 percent, and virtually all FHAIVA
mortgages go directly into GNMA mortgage pools. Moreover, empirical
estimation implies that conventional yields adjust fully to changes in capital
market rates within two weeks.
Mortgage rates are currently about what one would expect given capital
market rates. In contrast, conventional rates were a half percentage point
"too low" in the 1970s, owing to thrift tax advantages and portfolio re-
strictions, and a half point "too high" in the 1982 to 1986 period because
thrift profits and portfolio restrictions had effectively disappeared. This
half point "excess" return on mortgages stimulated development and use
of the Freddie Mac and Fannie Mae pass-through programs. Since the
middle of 1987, rates on the conforming FRMs that these agencies can
buy have been "about right", while rates on nonconforming or jumbo
FRMs are about thirty basis points higher. Enactment of user fees on agency
mortgage-backed securities or higher capital requirements for the agen-
cies would likely raise coupons on conforming FRMs.
ARMs have become a major factor in the conventional mortgage market,
with the ARM share of total originations periodically swinging between
a quarter-and-a-half, depending largely on the level and term structure of
interest rates. Borrowers are more likely to choose ARMs the less affordable
is housing (the higher are interest rates) and the more ARMs allow
borrowers to solve the affordability problem (the lower is the ARM rate
relative to the FRM rate). ARM securitization, however, has lagged behind
FRM securitization owing to the lack of both an attractive FHA-ARM
and standardization of conventional ARMs. Introduction of a more
attractive FHA-ARM would likely lead to greater securitization of the

comparable conventional ARM product, as well as the FHA-ARM.
The S&L industry has been shrinking relatively since 1984 and abso-
lutely since passage of FIRREA. Another 25 percent to 50 percent decline
is likely. The past shrinkage does not appear to have raised home mortgage
rates, and there is no reason to believe that the future shrinkage will.
Notes
1. The agencies also securitize multifamily mortgages. Between 1975 and 1982, 8 percent
to 16 percent of FHA multifamily mortgages were securitized (Seiders, 1983,278).
2. The retail commitment rate and points are those obtained by the Federal Home Loan
Mortgage Corporation in a weekly survey of 125 major lenders conducted since the spring
of 1971.
3. The rates in this table are not adjusted for points, that is, they are the coupon rates
consistent with whatever points were charged. The adjustment would not affect the differ-
ences between actual and perfect rates because the adjustment to both rates would be
identical.
4. The one-year Treasury rate was down to 9 percent in early 1983, versus 8 percent in
the summer of 1990, and the Fannie Mae share price was up from a low of two-and-one-half
in 1981-82. Freddie Mac stock was not freely traded until around year end 1988. Since
then, the share price has moved much like Fannie Mae's.
5. The FRM also reflects the cost of the borrower's default option to the lender.
6. This overstates the cost to the borrower because the actual FRM coupon rate does
not costlessly decline when interest rates fall and this gain to the lender should be priced
in a lower coupon. The same gain means that the cost of the borrower's call option is less
than if the borrower could costlessly refinance.
7. Changes in FRM-ARM rate spreads reflect changes in the Treasury term structure
(seven-year rate less one-year), in the values of the FRM call option and the ARM rate
caps, and in the initial ARM discount. Thrifts increased the attractiveness of ARMs throughout
1987 and 1988 by raising the average initial discount from one-half to three-and-one-half
percentage points (Peek, 1990). With thrifts no longer able to qualify borrowers for larger
loans based upon deep discounts, the initial discount plummeted in 1989 to less than one
percentage point.
8. The 1986 Tax Law has also contributed to these losses, but overbuilding is the pri-
mary culprit.
References
Brueckner, Jan K. and James R. Follain. 1989. "ARMs and the Demand for
Housing." Regional Science and Urban Economics 19(2) (May): 163-187.
Cassidy, Henry J. 1984. "A Review of the Federal Home Loan Bank Board's
Adjustable-Rate Mortgage Regulations and the Current ARM Proposal."
Research Working Paper No. 113. Federal Home Loan Bank Board. (August).
Hendershott, P.H. 1990a. "The Composition of Mortgage Originations in the Year

2000." Forthcoming in FNMA publication.
- - . 1990b. "The Future of Thrifts as Home Mortgage Portfolio Lenders." The
Future of the Thrift Industry. Proceedings of the 14th Annual Conference. Federal
Home Loan Bank of San Francisco. (Dec. 8-9): 152-163.
Hendershott, P.H. and J.D. Shilling. 1990. "The Continued Interest Rate
Vulnerability of Thrifts." Presented at NBER Conference on Financial Crisis.
(March). Also forthcoming in NBER volume.
- - . 1989. "The Impact of the Agencies on Conventional Fixed-Rate Mortgage
Yields." The Journal of Real Estate Finance and Economics 2: 101-115.
Hendershott, P.H. and R. Van Order. 1989. "Integration of Mortgage and Capital
Markets and the Accummulation of Residential Capital." Regional Science and
Urban Economics 19, 2 (May): 188-210.
Hendershott, P.H. and KE. Villani. 1980. "Secondary Residential Mortgage Markets
and the Cost of Mortgage Funds." AREUEA Journal 8 (Spring): 50-76.
Kane, Edward J. 1989. The S&L Insurance Mess: How Did It Happen? Washington,
D.C.: The Urban Institute Press.
O'Keefe, Michael and Robert Van Order. 1990. "Mortgage Pricing: Some
Provisional Empirical Results." AREUEA Journal 18 (Fall): 313-322.
Peek, Joe. 1990. "A Call to ARMs: Adjustable Rate Mortgages in the 1980s."
New England Economic Review (March/April): 47-61.
Roth, H.L. 1988. "Volatile Mortgage Rates-A New Fact of Life?" Economic
Review. Federal Reserve Bank of Kansas City. Roth: Volume 73, No.3 (March):
16-28.
Seiders, D.S. 1983. "Mortgage Pass-through Securities: Progress and Prospects."
AREUEA Journal 11 (Summer): 264-87.
Woodward, Susan E. 1989. "Policy Issues in the Privatization of FNMA and
FHLMC." Expanded Competitive Markets and the Thrift Industry. Proceedings
of the 13th Annual Conference. Federal Home Loan Bank of San Francisco.
(Dec. 10-11): 169-174.
COMMENTARY
Herbert M. Kaufman
It is always a pleasure to read one of Patrie Hendershott's articles. I

always learn something from them, they are always well prepared, and
conscientiously argued. Having said that of course, I wouldn't be a true
discussant if I now sat down and left it at that. This particular article
embraces many issues in mortgage credit and is really a summary of sev-
eral developments that have taken place in the mortgage market. I think
if Hendershott can be faulted for anything, it could be for being overly
brief in this article and attempting to deal with too many issues at the cost
of reduced attention to each. In particular, I want to focus on a couple of
issues that he raised, and see if I can carry them somewhat further with
respect to the arguments made, and with which, by the way, I am in
agreement.
Let me start with securitization. Hendershott does a very thorough job
reviewing the history of securitization and analyzing the impact of
securitization. The role that Ginnie Mae, Fannie Mae, and Freddie Mac
played in the development of securitization to its current prominent position
is well presented and argued. What I would like to have seen developed
further is, not the issue of the integration of the mortgage and capital
markets as a result of securitization, (which is handled quite well-but,
125
which because of his previous work and that of others has become well
known) but rather where we go from here with regard to the particu-
lar status that the government-sponsored agencies now have in this
environment.
It is unarguable that the success of mortgage market securitization,
especially in the decade of the 1980s, was due to the special status of
sponsored agencies. The view of their obligations as implicitly government
guaranteed, their specialist status in the market, and their expertise in
handling securitization all were critical to the rapid development and
expansion of securitization. These advantages are no longer necessary, or
certainly no longer necessary in the same form for continued development.
Hendershott alludes to the proposals that have been made for capital
requirements and user fees. He quite correctly points out the benefits that
have accrued to stock holders from the existence of the subsidies that are
implicit in agency activity from the government. Additional arguments
for benefits accruing or stemming from the existence of these implicit
subsidies can be made for other players, including bond holders, but also
including managers of these companies. This should not be a goal of
public policy. It strikes me that the time has come to suggest a winnowing,
and significant winnowing away, of the government-sponsored agencies
(Fannie, Freddie) from the kinds of explicit subsidies and implicit guar-
antees that they have available to them. There is no pressing reason to
believe that these institutions could not operate effectively on their own,
after some transitional period, nor is there any reason to believe that the
gains that have been made, which are significant and positive from their
role in integrating the capital and mortgage markets, will be reversed.
Therefore, meeting private capital requirements at a minimum is neces-
sary as a first step. The current proposal of the General Accounting Office
moves in the right direction and logically occurs prior to their being
subjected to bond rating requirements. Hendershott's work here and
elsewhere with various colleagues suggests that integration has reached,
at least with regard to the fixed rate mortgage market, a substantial level,
and will not likely be reversed. I will have more to say about adjustable
rate mortgages and securitization further into my commentary.
Furthermore, I would argue that continuing sponsorship of agencies
has had a negative impact in artificially segmenting the mortgage market
somewhat because of the pressure for conforming limits on conventional
loans. As Hendershott's work at least suggests, although the evidence is
mixed, nonconforming conventional loans seem to bear an interest rate
penalty for exceeding conforming limits. This is not sensible or defensi-
ble, in my judgment, and is primarily because of the political pressures
that are on these agencies to keep conforming loans at some reasonable

level with which middle America is comfortable. It is unlikely that until
these agencies are fully privatized that constraints such as these, which
are artificial and suggest non-market determined differentials, will be
removed.
I now want to tum to a couple of other issues that Hendershott raises.
One is the relative lack of securitization in the ARMs market and the
differences that dictate the choice of ARMs vs. fixed-rate mortgages. Again,
I agree with the thrust of his analysis that rate differentials are driving the
choice of ARMs vs. fixed-rate mortgages. However, certainly the exist-
ence of the more complete integration with general capital markets shown
in his work has led to increased production in securitized fixed-rate mort-
gages and ceteris paribus lower-rate differentials relative to ARMs because
of the reward for standardization. It strikes me that he is probably overly
pessimistic, however, about the development of securitization for ARMs.
If the market can successfully securitize credit cards and automobile loans, a
way to securitize ARMs effectively can be developed. I think over time,
as the interest in doing so becomes more apparent and profitable, the
development of large-scale securitized ARMs will be forthcoming.
Finally, I want to make one point about the demise of the thrift industry.
There is no question that Hendershott is right, that this erosion will con-
tinue. He expresses apparent surprise in the body of the paper, but not
in the conclusion, that the demise of savings and loan participation in the
mortgage market has resulted in no basic rise in mortgage rates. However,
his work itself is an argument for why that has happened, namely we in
fact have mortgage credit availability from the capital markets in general.
If one player's participation in the mortgage market declines, everything
else the same, the flow of funds into the market should not be materially
affected. It is, therefore, not at all surprising that in an integrated en-
vironment the impact on mortgage rates will be minimal from the decline
of thrifts.
4 EQUITY UNDERWRITING RISK
J. Nellie Liang
James M. O'Brien
Introduction
Risk in securities activities such as underwriting and dealing has become

of increasing interest as the separation between commercial and invest-
ment banking erodes. In January 1989, the Federal Reserve permitted
commercial bank holding companies to engage in limited amounts of
corporate securities underwriting and dealing and has recommended to
Congress the removal of the Glass-Steagall separation of commercial and
investment banking. As banks expand their securities powers, the associ-
ated risks have a policy significance because of deposit insurance and the
safety net provided to large commercial banks by the federal government.
Furthermore, understanding the risks associated with underwriting and
dealing can be important in understanding what determines the demand
for these services. In this study we examine equity underwriting risk within
the context of earlier work on price risk associated with equity offerings
and underwriting risks over time.
In an important study of equity underwriting risk, Giddy (1985) estimated
returns to firm-commitment equity underwritings, that is, underwritings
for which the underwriter purchases the issue from a corporation and
129
offers it at a pre-set price to the public. A principal result was that, de-
spite evidence of substantial price uncertainty when the offer price is set,
underwriters are able to keep the relative frequency and magnitude of
losses to very low levels. He provided further evidence to suggest that
equity underwriting may not be any riskier than bond underwriting. Armed
with these findings and an analogy between underwriting and bank lending,
Giddy concluded that underwriting stock may not be any riskier than
lending. His conclusions also raise a question about the traditional view
that a principal service provided by the underwriter in a firm-commitment
underwriting is to bear price risk, as developed, for example, in Mandelker
and Raviv (1977). If this were the case, one would expect a large part of
the price risk associated with issuing stock to fall on the underwriter.
However, if other services are principal, such as marketing and distribution
services (Benveniste and Spindt, 1990) or "certification" services (Booth
and Smith, 1986), underwriters may obtain substantial protection against
price risk, thus minimizing the risk-bearing function.
While Giddy is one of the few studies to have estimated equity under-
writing risk per se, his results are not without some conflicting evidence.
Boyd and Graham (1988) have argued from a time series perspective that
investment banking is one of the more risky enterprises in the financial
services industry.! The variance of the stock returns and market f3s of
investment banks are near the high end of the range for financial firms
and much higher than for commercial banks. Because investment banks
are also less highly leveraged than are commercial banks, the stock market
evidence suggests that the variability of the return to investment bank
assets also may be greater. 2
Whether Boyd and Graham's results on the risks of investment banking
actually conflict with Giddy's is unclear. Equity underwriting is a relatively
small part of major investment bank activities and most investment banking
risk may be coming from other activities. 3 However, Giddy's results apply
only to the price risk on individual underwritings and this risk may under-
state the risk associated with equity underwriting over time. We first
reexamine and extend Giddy's results on equity underwriting price risk.
We then consider equity underwriting risk from a temporal perspective
and examine the return variability to equity underwriting over time.
In our reexamination and extension of Giddy, the sample period is
increased and the effects of price risk on underwriting returns are tested
over several subperiods. Tests of the sensitivity of underpricing (setting
the offer price at less than the expected market price) and the underwriter's
spread (the difference between the offer price and the price paid to the
issuer) to the price risk of the issue are also expanded. As a further test
EQUITY UNDERWRITING RISK 131
Table 4-1. Characteristics of Samples of Seasoned and IPO Common Stock

Issues for 1977 to 1987.
Characteristics Seasoned IPO

Number of issues 2,580 1,641
Total value of issues ($ mil.) 89,643.9 21,743.3
Mean offer size ($ mil.) 34.7 13.3
Mean public offering price ($) 20.78 9.78
Mean firm size ($ mil.) 1,409.0 54.1
Mean underwriting spread (%) 5.48 8.13
Std. dev. of underwriting spread 2.13 1.58
Source: Securities and Exchange Commission's Registration Offerings Statistics.
of the importance of price risk, we examine the significance of low returns

to equity underwritings on the stock market returns for a small number
of investment banks.
Following the analysis of price risk, equity underwriting risk is refor-
mulated as return uncertainty over time and measured as a time series
variance of dollar returns from equity underwriting. This time series vari-
ance is decomposed and the effects and relative importance of price risk
and other factors on the variance are analyzed. We then test for the
significance of the time series of equity underwriting returns on the stock
returns for our sample of investment banks. If variability in equity under-
writing returns over time contributes significantly to the (high) variance
of investment bank stock returns, the time series of equity underwriting
returns should bear a significant relation to the investment banks' stock
returns. Overall conclusions and some implications for commercial bank
expansion into equity underwriting appear in the final section.
Characteristics of Underwritten Equity Issues
To examine equity underwriting risk, we constructed a data set of com-

mon stock issues underwritten over the period 1977 to 1987. Included
in our sample are all domestic issues registered with the Securities and
Exchange Commission (SEC) with gross proceeds of at least $1.5 million,
negotiated, and made under a firm-commitment offering during 1977 to
1987, for which data were available. 4 Some characteristics of this sample
of common stock issues, with issues divided between seasoned and initial
public offerings (IPOs), are shown in table 4-1. The sample contains 4,221
Table 4-2. Characteristics of Samples of Seasoned and IPO Issues for 1977 to
1979 and 1984 to 1986.
1977-1979 1984-1986
Characteristics Seasoned 1PO Seasoned 1PO
Number of Issues 321 46 694 683
Total Value of Iss. ($ mil.) 9,889.1 263.6 25,887.7 11,694.9
Mean offer size ($ mil.) 30.8 5.7 37.3 17.1
Mean public offer price ($) 20.2 12.6 20.1 9.3
Mean firm size ($ mil.) 1,346.0 17.3 1,957.7 103.6
Mean underwriting spread (% ) 4.8 7.8 5.6 8.0
Std. dev. of underwriting spread 1.9 0.7 2.1 1.4
Source: Securities and Exchange Commission's Registration Offerings Statistics.
offerings, of which 2,580 were seasoned, and 1,641 were IPOs.5 On aver-
age, IPOs are smaller and have lower public offering prices than seasoned
issues. IPOs also tend to be made by smaller firms.6 In addition, under-
writing spreads as a percentage of the offer price are significantly larger
for IPOs than seasoned issues and have less variability.
To investigate the sensitivity of underwriting returns to different mar-
ket conditions, subsamples of common stock issues are also constructed
for the periods 1977 to 1979 and 1984 to 1986. During 1977 to 1979, the
stock market was relatively fiat, with the S&P price index of common
stocks growing by less than 4 percent over the three year period. The
period was also one of low stock price variability. The 1984 to 1986 period
was one of substantial growth, with the S&P 500 stock market index in-
creasing by about 50 percent, and with much greater price volatility than
in the earlier period.
Characteristics of the underwriting data for the two subperiods are
shown in table 4-2. The table shows that during 1977 to 1979 relatively
few issues were made. Our sample consists of only forty-six IPOs and 321
seasoned issues over the three year period. In contrast, 683 IPOs and 694
seasoned issues were made over the three-year period from 1984 to 1986.
The difference in the number of issues and amounts offered between the
two periods may be explained by the incentives of firms to issue equity in
a rising market. Additionally, the average offer size and average firm size
are greater in 1984 to 1986 than in 1977 to 1979, particularly for IPO
issues. Average underwriting spreads and their standard deviations are
also somewhat larger in the latter period.
Price Uncertainty and Risk on Individual Equity

Underwritings
To analyze the price risk associated with a firm-commitment underwrit-

ing, the gross underwriting return is defined as the offer price times the
number of shares sold at the offer price plus the market price times
the number of shares sold at the market price, minus the issue price times
the number of shares issued. This is a gross return because there is no
deduction of expenses incurred by the underwriter in bringing the issue
to market. Formally, the dollar underwriting return, R, is
R = sQ if the market price is greater than or equal to
the offer price (p 2: 0)
=pQp + sQ if the market clearing price is less than the
offer price (p < 0)
where Q == dollar amount of the offering (offer price times number of
shares issued)
Qp == dollar amount of the offering times fraction sold at the
market price if p < 0
p == the difference between the market and offer price per
dollar of the offering price
s == the underwriting spread between the offer and issue price
per dollar of offering price. (1)
The uncertainty of the return at the time the offer price is set will depend
on the uncertainty of the market price, and hence p, given the offer price,
the underwriting spread, and the amount of the offering.
In this section we examine two factors that can reduce the under-
writer's price risk: (1) underpricing the offering (defined as the expected
value E(p) > 0) and (2) making the underwriting spread(s) sensitive to
price risk. Underpricing reduces the underwriter's price risk by increasing
the probability that the subsequent market price of an issue is greater
than the offer price. 7 ,8 The underwriter then earns the entire spread. Also,
if the underwriting spread includes a cushion for price risk, the market
price can fall below the offer price by the amount of the cushion before
cutting into direct costs.
Underpricing and Market Price Uncertainty
Giddy's results indicated that price uncertainty for underwritten issues is

substantial, with that for IPOs being much greater than for seasoned
Table 4-3a. Percentage Differences between Day-After Market and Offer Prices
for Seasoned and IPO Issues. 1
Seasoned Issues IPO Issues
1977-87 1977-79 1984-86 1977-87 1977-79 1984-86

Mean 1.0* 0.2 2.0* 8.4* 9.3* 5.9*
Std. dev. 9.7 5.2 11.7 22.3 14.9 20.0
Range -58,163 -24,45 -36,163 -50,212 -15,50 -38,212
Distribution of Differences Between Market and Offer Prices (percent)
positive 39.7 27.4 44.4 53.4 63.0 51.0
negative 40.0 46.1 36.9 30.1 30.4 32.9
same 20.3 26.5 18.7 16.5 6.5 16.1
Table 4-3b. Conditional and Unconditional Negative Percentage Differences

between Day-After Market and Offer Prices.

1977-87 1977-79 1984-86 1977-87 1977-79 1984-86
Conditional: Market Prices Below Offer Prices!
Mean -3.69* -2.24* -3.14* -7.60* -5.03* -6.79*
Std. dev. 5.40 2.62 3.73 7.98 3.86 7.23
Unconditional 2
Mean -1.47* -1.03* -1.16* -2.29* -1.53* -2.24*
Std. dev. 3.83 2.10 2.72 5.59 3.13 5.23
* Significantly different from zero at 1 percent.
1 Market prices: closing prices taken from the CRSP daily stock files.
2 For the unconditional, the percentage difference equals the actual percentage difference
if the market price is below the offer price; otherwise, the percentage difference equals zero.
issues. In this section, we expand the analysis of the importance of

underpricing in reducing the underwriter's exposure to price risk. Per-
centage differences between the day after market and offer price, specified
as p in equation (1), are presented in tables 4-3a and 4-3b for the entire
sample and for the sub-samples. The presentation of table 4-3a follows
that of Giddy (tables 6.3 and 6.6, 156, 162) except for different sample
periods. Table 4-3b provides additional information on the distribution of
differences between market and offer prices when the market price is
below the offer price.
The mean percentage difference between the day-after market and
offer prices, presented in row one of table 4-3a, is our measure of

underpricing. Underpricing is significant but modest for seasoned issues
over the eleven year period. 9 For IPOs, underpricing is substantial, 8.4
percent of the offer price over the entire sample period. The standard
deviation of the differences between market and offer prices, presented
in row two, is our measure of the underwriter's price uncertainty. As
Giddy found with his sample, price uncertainty appears to be substantial
for both seasoned issues and IPOs, although it is much greater for IPOs.
However, the greater underpricing for IPOs as compared to that for sea-
soned issues, reduces the magnitude of negative differences between the
market and offer prices. These reductions can at least partly explain the
comparability between the lower ends of the ranges for seasoned and IPO
price differences, in the third row of table 4-3a, despite the higher stand-
ard deviations for IPO price differences. lO
Underpricing can also explain differences in the frequency distributions
for IPOs and seasoned issues presented in the lower half of table 4-3a. As-
suming that positive and negative expectations errors on the market prices
are equally likely, actual market prices would be above offer prices with
the same relative frequency as they would be below offer prices in the
absence of underpricing. Thus, the difference between the relative fre-
quencies of market prices being above and below the offer prices is an
indication of the effects of underpricing in reducing the probability of the
market price falling below the offer price. Under this reasoning, the small
amount of underpricing on seasoned issues does not appear to materially
affect the likelihood of the market price falling below the offer price for
seasoned issues. However, underpricing substantially reduces the likeli-
hood of the market price falling below the offer price for IPOs. Indeed,
as shown in the table, the likelihood of a negative difference is smaller for
IPOs than for seasoned issues.
The top half of table 4-3b presents means and standard deviations for
percentage differences between the market and offer prices when the market
price is below the offer price. For all three periods, the conditional mean
differences for IPOs are considerably lower (more negative) than for
seasoned issues, with the differences in means significant at the one percent
level. Thus, when market prices fall below offer prices, the expected
percentage difference is substantially greater for IPOs than for seasoned
issues. However, while the conditional price risk is greater for IPOs than
for seasoned issues, the probability that market prices will be less than
offer prices is smaller for IPOs, as shown in the lower half of table 4-3a.
These counterbalancing effects are reflected in the unconditional mean
negative price differences which are equal to the conditional mean dif-
ferences multiplied by the probability of getting a negative difference.

Based on these unconditional mean values, shown in the bottom half of
table 4-3b, the differences between seasoned issues and IPOs are more
moderate, though differences in means are still mostly significant at the
one percent level (the exception being the 1977 to 1979 period).
With respect to the sub-periods, the standard deviations and ranges for
both IPOs and seasoned issues and the conditional and unconditional
means for the negative price differences indicate greater price risk in the
latter period. This suggests that underpricing also should be greater in the
latter period, ceteris paribus. The average underpricing of seasoned issues
in the latter period is significantly greater than in the former period at the
one percent level. However, the average underpricing of IPOs is smaller
in the latter period than the former period, although the difference is
significant only at the 10 percent level.
In sum, our evidence of substantial market price uncertainty on stock
offerings and underpricing is consistent with Giddy's results. Our analysis
of the effects of underpricing indicates that underpricing for IPOs is
important in modifying these issues' downside price risk. While this makes
the price risk for underwriting IPOs more comparable to that for under-
writing seasoned issues, IPOs still have greater downside price risk.
Underwriting Spreads and Market Price Uncertainty

In addition to underpricing, risk-related underwriting spreads will also
reduce the sensitivity of underwriting returns to price risk. If underwrit-
ing spreads are higher for issues with greater price risk, price variability
produces smaller underwriting return variability since issues for which
realized market prices are low will tend to have higher underwriting
spreads. We are aware of only a few studies that have attempted to find
evidence of risk premiums in equity underwriting spreads, and the evi-
dence from these studies is inconclusive. ll Giddy, however, reported that
underwriting spreads for IPOs and seasoned issues combined were posi-
tively associated with the absolute differences between offer prices and
subsequent market prices. He concluded from this that risk-sensitive
underwriting spreads help to stabilize the underwriter's returns (p. 156
and figure 6.8).
Following Giddy, we use the absolute percentage difference between
the offer and (day after) market price as an ex post measure of uncer-
tainty about the market price when the offer price is set. That is, the
greater is ex ante uncertainty about the market price, the larger is the
absolute difference between the realized market price and the offer price
likely to be. This price difference will contain, in addition to the unex-
pected deviation between the market price and its expected value, an
(anticipated) underpricing component. As described above, this compo-
nent is also likely to vary with the issue's risk, ceteris paribus. We do not
model the joint determination of the spread and underpricing. For this
analysis, differences between the realized and the expected market price
are assumed to be the dominant component explaining variation between
the market price and the offer price.
In expanding on Giddy's test, we first regress the underwriting spread(s)
on the absolute value of the percentage difference between the offer price
and the day after market price (AP+ 1), controlling for issue size (IS), firm
size (FS), and (through separate regressions) whether the offering is IPQ
or seasoned. Issue size and firm size are control variables because they
may be inversely correlated with price uncertainty and positively related
to underwriting costs (Pugel and White 1985, 119). Thus, any relation
between spreads and our measure of price uncertainty cannot be interpreted
unambiguously if firm or issue size are omitted from the regression.
The regression results presented in table 4-4a provide strong evidence
that underwriting spreads contain a risk premium that increases with the
price uncertainty of the issue. For both the seasoned and IPQ equations,
the coefficients on the absolute value of the percentage change in price
are positive and highly significant. The negative coefficients on issue size
suggest that there are fixed costs, and hence scale economies, associated
with equity underwriting.
More directly related to the underwriter's price risk is whether the
underwriting spread tends to be larger when the market price is below the
offer price. Table 4-4b presents regression results based on the subsample
of issues for which the day after market price fell below the offer price.
Again, for both the seasoned issues and IPQs, the coefficients on price
uncertainty, measured by the absolute percentage price difference, are
highly significant. These results indicate that underwriting spreads reduce
the downside price risk of underwriting and help to stabilize the under-
writer's return.
Risk of Loss on Equity Underwriting
The tendency of underpricing and underwriting spreads to increase with

the price uncertainty of the issue being underwritten can be expected to
reduce the sensitivity of underwriting returns to price risk. In this section,
we examine estimated returns on underwritings and their sensitivity to
price risk using the procedures developed by Giddy.
Following Giddy, we assume that the entire issue is sold at the market
price if the price is less than the offer price (that is, in equation (1), Qp
Table 4-4a, b. Effects of Price Uncertainty on Underwriting Spreads. 1
All Issues
IPO: s = 8.10 + .016*AP+1 - .024*IS + 9.84E-5*FS
(160.3) (7.46) (-13.07) (1.66)
Seasoned: s = 5.58 + .069*AP+1 - .011 *IS + 9.30E-6*FS R2 = .20
(121.3) (14.78) (-15.59) (1.64)
Issues with Negative Day-After Market and Offer Price Differences
IPO: s = 8.06 + .085* AP+1 - .040*IS - .002*FS R2 = .24
(57.7) (7.36) (-5.80) (-1.84)
Seasoned: s = 5.33 + .134*AP+1 - .010*IS - 1.45E-6*FS R2 = .22
(65.8) (10.73) (-9.56) (-0.21)
1 t-statistics in parentheses. Spread (s) and absolute price differences (AP+l) are per-
centages. Issue size (IS) and firm size (FS) are in millions of dollars.
= Q if p < 0). Since part of an issue is usually sold at the offer price even
when the market price is below the offer price, this assumption may
substantially understate the underwriter's returns on poorly received
issues. On the other side, the use of the gross underwriting spread, which
does not net out the underwriter's expenses, overstates the net return on
an issue.
Giddy made three hypothetical calculations of the underwriter's gross
dollar return to each underwritten issue, depending on market conditions
one, five, and ten days after the offer date. Our calculations are returns
per dollar of the amount offered. Using the return notation in equation
(1) and the assumption that Qp = Q if p < 0, our three calculations are as
follows:
Calculation Market Condition Underwriter Return
1 P+1 ~ 0 s
otherwise P+1 + s
2 P+l or P+s ~ 0 s
otherwise P+s + s
3 P+l' P+s, or P+lO ~ 0 s
otherwise P+lO + s
Table 4-5. Hypothetical Underwriter Returns for Seasoned and IPO Common
Stock Issues for 1977 to 1987 (percent).

1 day 5 days 10 days 1 day 5 days 10 days
Mean 4.0* 3.6* 3.4* 5.8* 5.3* 4.8*
Std. dev. 4.0 4.7 5.3 5.4 6.5 7.6
Range -48,21 -49,21 -49,21 -42,20 -39,20 -46,20
Percent of
Gains 93.6 87.7 86.0 91.2 87.1 84.5
Losses 6.2 12.1 13.8 8.2 12.4 15.1
Breakevens 0.2 0.2 0.2 0.6 0.5 0.5
where P+i is the value for P using the closing market price i days after the
offering date.
These three calculations are substitute estimates of an underwriter's
return on an issue. The first calculation assumes that, if the market price
one day after the offer date is less than the offer price, the underwriter
sells the entire issue at the market price one day after the offering. The
next two calculations assume that, if the market price initially falls below
the offer price, the underwriter will hold onto the issue for a limited
number of days before selling the issue at the lower of the offer or market
price. The assumption of a five or ten day holding period is arbitrary and
simply a rough approximation to the gradual unloading of an issue that
is getting a poor reception.
Underwriter returns calculated for seasoned issues and IPOs separately
for our full sample are presented in table 4-5. The mean returns, reported
in the first row, equal the mean underwriting spreads (table 4-1) plus the
mean unconditional expected losses from selling at a market price below
the offer price (table 4-3b). The calculations show that the mean under-
writing returns are positive and significant for both seasoned issues and
IPOs, with somewhat higher average returns on IPOs. The standard de-
viation of returns for IPOs is only moderately higher than that for
seasoned issues, even though the standard deviation of the differences
between market and offer prices, shown in table 4-3a, is much higher for
IPOs. When estimated returns are negative using day-after market prices,
the average return is moderately lower for IPOs, -9.07 percent, as com-
pared to -6.26 percent for seasoned issues (not shown in the table).
In the lower half of table 4-5, the relative frequency of estimated losses
using next day market prices is seen to be about 6 percent for seasoned
issues and 8 percent for IPOs. These estimated loss probabilities are about
double those reported by Giddy (tables 6.4 and 6.6). Nonetheless, the
probabilities are still low and contrast sharply with the 40 percent and 30
percent probabilities for market prices falling below offering prices for
seasoned issues and IPOs, respectively, shown in table 4-3a. These loss
probabilities based on next day prices indicate that, for most issues, market
price declines are within the range of the underwriting spread, despite the
presence of very substantial price uncertainty. Further, the probability of
loss is only moderately greater for IPOs than that for seasoned issues,
despite the much greater price uncertainty of IPOs.
As the length of the holding period increases, average underwriting
returns for both types of issues decline, the standard deviations increase,
and the loss probabilities increase. These results may simply reflect the
random nature of stock price changes and the upper bound on the under-
writer's return. However, the results may also be influenced by under-
writer attempts to stabilize prices early in the offering period by holding
on to part of an issue getting a poor reception. To the extent that this
occurs, the estimated returns using the day-after prices are upward biased
and the risk of loss is understated. On the other side, the estimated returns
tend to be overstated by not recognizing the partial selling of an issue at
the offer price when the market price is less than the offer price.
Estimated returns are presented separately for the sub-periods 1977 to
1979 and 1984 to 1986 in table 4-6. There are no substantial differences
in returns across the periods, although the number of offerings differ
substantially. The results show, however, slightly greater return variability
and loss frequencies in the latter period as compared to the former. Also,
the mean returns, conditioned on a negative return, were lower (more
negative) in the latter period for both IPOs and seasoned issues (not
shown). This is consistent with the greater market price uncertainty in the
latter period shown in table 4-3a. On the whole, however, these results do
not indicate a lot of sensitivity of underwriting returns to general market
conditions. For both periods, the estimated loss frequencies are moderate
and are in the same general range as the frequencies previously reported
for the longer sample period.
Investment Bank Stock Returns and Low Returns on Equity

Underwritings
The previous tests provide evidence that price risk is modest for firm-
commitment equity underwriting. These tests required assumptions about
Table 4-6a, b. Hypothetical Underwriter Returns for Seasoned and IPO Issues
1977 to 1979 and 1984 to 1986 (percent).
1977-79
Seasoned Issues (N = 321) IPO Issues (N = 46)
Mean 3.8* 3.2* 2.9* 6.3* 5.6* 5.5*
Std. dey. 2.3 3.2 3.9 3.3 5.3 6.1
Range -15,10 -15,10 -16,10 -8,10 -14,10 -19,10
Percent of
Gains 97.2 88.5 84.7 93.5 87.0 91.3
Losses 2.8 11.2 14.3 4.3 13.0 8.7
Breakevens 0.0 0.3 0.9 2.2 0.0 0.0
1984-86
Seasoned Issues (N = 694) IPO Issues (N = 683)
Mean 4.5* 4.2* 3.9* 5.8* 5.4* 5.0*
Std. dey. 3.1 3.8 4.8 4.9 5.7 6.5
Range -26,18 -26,18 -31,18 -28,16 -28,16 -32,16
Percent of
Gains 96.0 90.8 88.3 91.5 87.6 84.9
Losses 3.9 8.9 11.5 7.8 11.9 14.2
Breakevens 0.1 0.3 0.1 0.7 0.6 0.9
the amount sold at the market price and how quickly the issue was dis-
posed of when the market price was less than the offer price. An alter-
native test of the importance of price risk is to test whether the occurrence
of such events has a significantly unfavorable effect on the investment
bank's stock return. Such an effect may occur if losses on average tend to
be much larger than the calculations in the last section suggest.
For the purpose of this test, daily stock returns were taken for four
investment banks, E.F. Hutton, Merrill Lynch, and Paine Webber over
the eleven-year period and First Boston since mid-1983P For each of the
individual banks, the daily returns were regressed on a market return
index and "low return" variables for seasoned and IPO underwritings in
which the bank was the lead underwriter (information where the bank
was only a participating underwriter is not available)Y The "low return"

variable equals the number of shares offered multiplied by the difference
between the market price and the offer price, when this difference is
negative, and zero otherwise. This is a low return definition and not a
"loss" definition because it does not take into account the underwriter's
spread. Its advantage is that it gives us a meaningful number of observations
of low returns since estimated losses are so infrequent for individual lead
underwriting firms. Nonetheless, results from using a loss variable, in-
stead of a low return variable, were similar.
The regression results are reported in table 4-7 using alternatively the
investment bank and market return on the day of the offering and on the
next day.14 The coefficient on the low return variable is expected to be
positive if low or negative returns to equity underwritings have a signifi-
cant negative impact on the bank's profits and hence stock returns. While
the majority of the low return coefficients have the expected positive sign,
the t-statistics are low and only one coefficient is significant at usual test
levels. IS These results suggest that the market typically does not regard
low returns on equity underwritings as having much significance for the
investment bank's profits and reinforces our conclusions from our earlier
tests that price risk may not be significant. 16
Conclusions and Limitations
The results presented here support the findings and conclusions of Giddy
concerning the significance of price risk on stock offerings. We find strong
evidence that underpricing and underwriting spreads are sensitive to price
risk which enable underwriters to translate substantial price uncertainty
into small loss probabilities. Also, estimated loss probabilities do not appear
to be very sensitive to general market conditions, as indicated by the
results from our subsamples. Furthermore, in the infrequent instances
when they occur, low underwriting returns or losses do not appear, on
average, to have very significant effects on the profits of large investment
banks.
These results do not imply that there is no potential for large losses.
Underwriters cannot entirely eliminate their exposure to sharp declines
in market prices and there is the element of choice in determining risk
exposure. Vulnerability to market declines is illustrated by a number of
estimated underwriting losses in the week just prior to the October 19,
1987 market crashP These estimated losses produced an estimated nega-
tive dollar return on seasoned equity underwriting for the fourth quarter
Table 4-7. Effects of Low Returns to Equity Underwritings on Investment Bank
Stock Returns. 1
Coefficients
Number of
Seas IPO Low Returns
Market Low Low
Firm Constant Return 2 Return Return R2 Seas IPO
a. Day-of-Offering Stock Returns 3
E.F. .0005 1.009 -.88E-5 .12E-5 .11 28 6
Hutton (1.03) (18.83) (-.85) (.09)
First -.0004 Ll42 .02E-5 1.70E-5 .34 13 3
Boston (-.81) (24.52) (.09) (2.25)
Merrill .0001 Ll29 .02E-5 -.06E-5 .17 110 15
Lynch (.26) (23.92) (1.60) (-1.30)
Paine- .0003 1.098 .02E-5 .04E-5 .12 13 4
Webber (.67) (19.82) (.59) (.74)
b. Day-After Stock Returns 4
E. F. .0005 1.010 -.05E-5 .06E-5 .11
Hutton (1.06) (18.85) (-.55) (.43)
First -.0004 Ll42 .07E-5 .08E-5 .34
Boston (-.89) (24.45) (.28) (.11)
Merrill .0001 Ll31 .01£-5 .05E-5 .17
Lynch (.26) (23.92) (.97) (LlO)
Paine .0003 1.099 -.38E-5 .57E-5 .12
Webber (.64) (19.86) (-.12) (1.20)
1 t-statistics in parentheses.
2 The market return index is the CRSP value weighted index of returns from the NYSE
and AMEX exchanges.
3 Sample size is 2,780 for all investment banks except First Boston which has a sample
size of 1,174.
4 Sample size is 2,779 for all investment banks except First Boston which has a sample
size of 1,173.
of 1987. Nonetheless, this adverse experience is relatively modest when

compared to the size of the October stock market decline.
There is, however, a limitation in just using price risk to measure the
risk associated with underwriting. While our results on price risk were not
very sensitive to the sample period examined, the number of underwritings
varied greatly between the two subsamples, and underwritings fell sharply
in the fourth quarter of 1987 following the stock market crash. This partly
reflects a reduced supply of offerings, but it may also be influenced by a
reluctance to underwrite in order to reduce price risk. This reduced volume
of underwriting also reduces underwriting returns and should be taken
into account in measuring underwriting risk. This and other aspects of
uncertainty over time in underwriting returns are taken up in the next
section.
Underwriting Returns Over Time
Return uncertainty in an underwriting business comes from more than

price risk on individual issues. Other sources of uncertainty include unex-
pected variability in the number of stock offerings, the size of offerings,
and underwriting spreads. There is also the potential for incurring multi-
ple losses over a short period. The combination of these factors will
determine the variance of the firm's equity underwriting returns over
time. This variance is a more inclusive measure of underwriting return
uncertainty. In this section, we first specify the relation between the vari-
ance of dollar returns to equity underwriting over time and the various
sources of return uncertainty. The relative importance of different sources
are then estimated. Following this, we examine the relation between the
time series measures of equity underwriting returns and the stock returns
of the investment banks studied in section three. This will permit some
evaluation of the contribution of the variance in equity underwriting re-
turns over time to the variance of investment bank stock returns.
Temporal Uncertainty of Underwriting Returns
Time Series Variance of Underwriting Returns. To define underwriting

uncertainty on a temporal basis, let Rn be the gross dollar return on the
nth underwriting during time period t (t is one quarter). Underwriting
returns during time period t, R, are the sum of the returns on individual
underwritings
R= "",N R (2)
~n=l n
where N is the stochastic number of underwritings during the period. Let

V(R) be the variance of R which is our measure of underwriting uncer-
tainty over time. 1s The stochastic nature of the number of underwritings
per period complicates the determination V(R). To reduce this complex-
ity, assume that Nand Rn are independent. 19 As shown in Appendix A,
the variance of R under this assumption is
V(R) = E(N)(J2 + r2V(N)+2E[ L7L~ Cov(Rj, Rj )] (3)
where (J2 = V(Rn), r = E(Rn), and the last expectation on the rhs of (3)
is taken over the distribution of N.
Three sources of variability in the returns to underwriting over time
appear in equation (3). One is the variance of the return to individual
underwritings, (J2. This variance will depend on the occurrence of losses or
low returns on individual underwritings due to market price uncertainty,
given the dollar size of the underwriting and the underwriting spread. It
will also depend on (unexpected) variation in the dollar size of underwritings
and underwriting spreads. Implicit in the risk analysis of the last section
was the assumption that the size of the underwriting and underwriting
spread were nonstochastic so that their variances did not contribute to
underwriting return uncertainty.
A second source of variance in underwriting returns over time is the
variance in the number of underwritings, V(N). This source also is not
captured when studying returns to individual underwritings. The third
source, which is also not recognized, is the (expected) sum of the covariances
between returns on individual underwritings within the time unit. These
covariances depend on covariances across offerings between the market
prices of individual issues, the size of offerings, and spreads. The under-
writing return covariances may be positive or negative. 20
Relative Importance of Dift'erent Sources of Uncertainty. To estimate the

relative importance of the three sources of uncertainty identified in equation
(3), we first had to obtain a time series of underwriting returns. This was
done by time-ordering the previously estimated returns on the individual
underwritings for seasoned issues and IPOs separately.21 One quarter was
selected as the time unit and the time-ordered returns were aggregated
into quarterly returns for each of the two types of underwritings. 22 Each
time series can be viewed as representing the quarterly returns to a firm
that gets a pro rata share of the quarterly underwriting income in IPOs
Table 4-8a. Variance Decomposition of Quarterly Underwriting Returns.
Percentage Contribution of Variance from:!

Number of Offerings Individual Returns
Days After Offering [r2V(N)] [E(N)a 2] Residual
Seasoned Underwritings
1 72 13 14
5 63 16 21
Initial Public Offerings
1 70 12 18
5 66 12 22
Table 4-8b. Percentage Contribution of Variance of Number of Offerings to

Variances of Alternative Quarterly Underwriting Returns. 2
Firm-commitment
Best Best Efforts plus
Common Preferred Total Efforts Firm-commitment
75 58 74 67 72
r2 == square of the sample average dollar return per underwriting

V(N) == sample variance of the quarterly number of underwritings
E(N) == sample average of quarterly number of underwritings
a 2 == sample variance of the dollar return per underwriting
Residual == V(R) - r2V(N) - E(N)a 2.
2 The percentage contribution from rV(N), as in part a of the table. The number of issues
(average size per issue, $ mil.) are: Common firm-commitment 4221 (26.39), preferred firm-
commitment 844 (62.36), best efforts 2132 (3.03). For preferred firm-commitment, 812 were
seasoned issues; for best efforts, 2020 were IPO issues.
and seasoned issues. The time series variances are decomposed into the
three components identified in equation (3). Results are presented in
table 4-8a.
The column in table 4-8a labeled Number of Offerings is the percent-
age of the quarterly dollar underwriting return variance attributable to
the variance in the number of offerings as indicated in equation (3). For
the one-day-after market prices, 70 percent or more of the variance comes
from the variance in the number of offerings per quarter, and slightly less
for the five-day-after results. As shown in the next column, the variance
in returns to individual underwritings contributes less than 15 percent to
the quarterly underwriting return variance. Roughly 15 percent to 20 per-

cent of the quarterly return variance comes from the Residual. Some
additional calculations (reported in Appendix B) examine the residual
component further. These calculations suggest that dollar returns on in-
dividual underwritings have on average a small positive covariance. The
aggregation of these (small) covariances can account for a significant portion
of the residual variance component. Further calculations suggest a posi-
tive covariance between the number of IPO underwritings in the quarter
and the return per IPO underwriting. Although not explicitly allowed for
in our specification of the quarterly return variance, this positive covariance
may also be accounting for some part of the residual variance for IPOs.
The results in table 4-8a do not report the variance decomposition for
returns from underwriting both IPOs and seasoned common stock. Nor is
consideration given to firm-commitment preferred stock underwritings or
best efforts offerings. The underwriting firms in our sample that did firm-
commitment underwritings generally did both seasoned offerings and IPOs
for both common and preferred stock. However, the major underwriters
were associated with only a very small amount of best efforts offerings.
To determine the importance of the number of offerings in the variance
decomposition using these alternative and broader categories of under-
writing, returns were calculated for individual underwritings and time
aggregated for these alternative groups.23 The percentages of the under-
writing return variance accounted for by the number of offerings (r2V(N))
are presented in table 4-8b. The most important implication of the results
of table 4-8b is that broadening the types of stock underwritten does not
reduce the relative importance of the number of underwritings in accounting
for the variance of underwriting returns. As can be seen, for broader
groupings, over 70 percent of the return variance comes from the variance
in the number of underwritings.
The variance decomposition of returns reported in table 4-8a were
calculated using the entire eleven-year sample period. However, the relative
importance of the variability in the number of offerings over such a long
period may not be of much significance if it reflects long-term trends in the
volume of underwriting to which firms can easily adjust their scale of
operations. We thus calculated for the various underwriting categories
the relative importance of the number of underwritings for the return
variances over subperiods of three to four years. The results (not re-
ported) were the same as for the longer sample period in that over 70
percent of the return variance was coming from variability in the number
of underwritings.
In sum, the variance of the return on individual stock offerings, which
includes losses due to unfavorable market prices, appears to be a small

part of the variance of equity underwriting income over time. A small
positive covariance between returns on individual underwritings may also
contribute to the time series return variance. The major factor accounting
for the estimated quarterly underwriting return variability is the number
of underwritings. 24 This result holds for broader as well as narrower
measures of equity underwriting returns and for shorter and longer sam-
ple periods.
Quarterly Underwriting Returns and Investment Bank Stock

Returns
The results just presented suggest that factors other than price risk,
particularly the number of offerings, are the major determinants of return
variability to stock underwriting. Previously, we considered the sensitivity
of investment bank stock returns to only price risk. Here we consider the
sensitivity of investment bank stock returns to variability in underwriting
returns over time.
If unanticipated movements in the bank's underwriting returns over
time have an important effect on the bank's profits, they should be signifi-
cantly related to the bank's stock returns. However, variation in equity
underwriting returns may not be important because the variability is small
in relation to other activities, or because of diversification effects. These
effects would be enhanced if investment banks were able to shift resources
between equity underwriting and other activities in response to exogenous
influences on the returns to anyone activity.
Table 4-9a presents two sets of correlations between quarterly equity
underwriting returns and quarterly returns to the stocks of our sample of
four investment banks. In one set of correlations, the equity underwriting
returns (underwriting-specific) are estimated from issues (seasoned, IPO,
and preferred) for which the respective banks were lead underwriters. 25
We have no data on the returns to underwriting in which the investment
banks participated but were not lead underwriters. For the other set,
equity underwriting returns (aggregate) are estimated from underwritten
issues (seasoned, IPO, and preferred) for all investment banks. These
aggregate returns would proxy for the bank's equity underwriting returns
if, as lead and participating underwriter, the bank earned a pro rata share
of the aggregate returns.
For the four investment banks, the correlation coefficients in table 4-
9a are mostly small and negative. Except for Merrill Lynch, the correla-
Table 4-9a. Correlations Between Quarterly Equity Underwriting Returns and

Investment Bank Stock Returns. 1
Type of E. F. First Merrill Paine
Offering Hutton Boston Lynch Webber Average
Underwriter-specific Underwriting Returns 2
Seasoned .076 -.150 .391* .090 .102
IPO -.222 -.045 .008 -.133 -.098
Total Common -.002 -.141 .330* -.124 .016
Common + Preferred -.007 -.059 .320* -.006 .062
Aggregate Underwriting Returns
Seasoned .199 .119 .305* .204 .207
IPO -.081 -.212 -.020 -.051 -.091
Total Common .097 -.017 .194 .112 .097
Common + Preferred .097 -.032 .209 .107 .095
Table 4-9b. Correlations Between Innovations in Quarterly Equity Underwriting

Returns and Investment Bank Stock Returns. 1
Underwriter-specific Underwriting Returns 2
Common -.082 -.138 .269 -.128 .020
Common + Preferred -.008 .105 .257 .056 .103
Aggregate Underwriting Returns
Common .165 .376 .375* .316* .308
Common + Preferred .224 .485* .434* .369* .378
* Significantly different from zero at 5 percent. Significance not computed for averages.
1 Sample size is 40 for all investment banks except First Boston which has a sample size
of 18.
2 Underwriting returns are estimated quarterly returns for which investment bank was
lead underwriter.
tion coefficients are all insignificant. We have no ready explanation for

the significance of the results for Merrill Lynch. 26 With the possible ex-
ception of Merrill Lynch, the results presented in table 4-9a do not indicate
that variability in quarterly returns to equity underwriting contributes
importantly to variability in investment bank stock returns.
In table 4-9b, measures of unanticipated bank-specific and aggregate
equity underwriting returns are used, since it is the unanticipated component
that the market would be expected to respond to. The unanticipated
component is measured as the innovation in the equity underwriting re-
turn where the innovation is the residual from an autoregression of equity
Table 4-10. Effects of Innovations in Quarterly Underwriting Returns on Merrill

Lynch Stock Returns.'
Coefficients
Constant Market Return Underwriting Return
Underwriter-specific Common Stock Underwriting Returns
-.021 2.019 .205E-8 .65
~~~ ~n) (~~
Underwriter-specific Common + Preferred Stock Underwriting Returns
-.023 2.058 .307E-9 .64
(-.93) (7.66) (.14)
Aggregate Common Stock Underwriting Returns
-.023 2.060 .365E-1O .64
(-.92) (7.15) (.14)
Aggregate Common + Preferred Stock Underwriting Returns
-.021 2.009 .195E-9 .64
~~~ ~~ (~)
1 t -statistics in parentheses. Sample size is 40.
underwriting returns on its past values. 27 The results reported in table 4-

9b for the underwriter-specific returns show no significant correlations
between innovations in bank-specific underwriting returns and the invest-
ment banks' stock returns. However, the correlation coefficients using the
innovations in the aggregate underwriting returns, shown in the lower
half of table 4-9b, are higher and frequently significant.28
Because the innovations in quarterly equity underwriting returns are
significantly correlated with the market return index, it is possible that the
aggregate underwriting return measure is simply proxying for the market
return. Thus, we performed a further test to see if innovations not associated
with the market index were related to the investment bank's stock returns.
For this test we used only the stock returns for Merrill Lynch, since it was
Merrill Lynch whose stock returns were most highly correlated with the
underwriting returns. In table 4-10, regressions of the stock returns of
Merrill Lynch on a market index and innovations in the various aggregate
underwriting return measures are presented. As can be seen, the various
equity underwriting return measures have no significant effect on Merrill
Lynch's stock returns when market return effects are controlled for. While
it is still possible that variation in equity underwriting returns does help
to explain variation in the investment banks' stock returns, our tests can-
not separate any such effects from general market effects.
Conclusions
The results of our study reinforce the evidence and conclusions of Giddy
(1985) concerning price risk in equity underwriting. We find strong evi-
dence that underpricing and underwriting spreads are sensitive to price
risk. This evidence is consistent with further results that the risk of loss
on firm-commitment underwritings is small despite the presence of sub-
stantial market price uncertainty when the offer price is set. These results
are not sensitive to the sample period. In reformulating the underwriter's
risk in terms of a quarterly time series variance of equity underwriting
returns, we find that the major determinant of this time series variance is
the variance in the number of offerings. The return variance on individual
issues (which depends on uncertainty about the market price, the amount
to be underwritten, and the underwriting spread) and return covariances
between issues account for only a minor share of the time series return
variance.
The time series results could help to reconcile Giddy's conclusions that
equity underwriting has relatively low risk with the time series evidence
on the relatively high risk of investment banking. That is, "business risk,"
such as the frequency with which underwritings are done, may have a
much greater effect on the variance of underwriting returns than price or
"market risk", although the two types of risk need not be independent. In
this case, the variance of returns to equity underwriting over time, which
depends on both types of risk, might still be important in accounting for
variation in investment bank profits and hence stock returns. However,
we did not find consistent evidence that variability in equity underwriting
returns over time had a significant effect on the investment bank's stock
returns.
Giddy's and our results, which suggest that underwriting equity may
involve relatively small price risk, raises doubts about the importance
of risk-bearing by the underwriter as a motive for firm-commitment
underwritings. Recent literature has emphasized other motives for stock
underwriting. Baron and Holmstrom (1980), Baron (1982), and Benveniste
and Spindt (1990) explain underwriting and firm-commitment contracts in
terms of marketing and distribution services where investors have infor-
mation advantages over issuers on the demand for new issues. Booth and
Smith (1986) also suggest that underwriters provide a certification service
on security offerings when the issuer has an information advantage over

investors on the value of the firm. These studies emphasize that the
effective provision of these services requires an ongoing business so that
customer relations and reputation are established.
These arguments bear on the risks and opportunities for commercial
banks as they expand into the business of underwriting corporate equity.
Although our results suggest that underwriters are able to obtain strong
protection against price risk, the substantial uncertainty about market
clearing prices on stock offerings requires prudence in setting offering
prices (underpricing) and underwriting spreads, as well as the other terms
of underwriting contracts. Going beyond our results and utilizing suggestions
from the recent literature on underwriting, the potential for profit may lie
in having an established network for the marketing and distribution of
issues and a reputation for being able to evaluate the issuing firm and the
worth of the issue. Both of these "assets" of the underwriting firm can
reduce the underpricing of an issue and thus make the firm more attractive
to issuers.
While this suggests that underwriting is not a "quick entry" business,
large commercial banks would seem to be relatively well positioned to
gain entry in corporate securities underwriting, given their dealings with
institutional investors and their traditional role and customer relationships
in the financing of firms. Once established, variability in the frequency of
underwriting may be the greatest source of variability in equity under-
writing income. Although we did not find strong evidence that this in-
come variability has been important for investment banks, our results are
limited to equity underwriting. Income variance from activities such as
brokerage and dealing could be of greater importance for commercial
banks expanding their investment banking services.
Appendix A
Variance of a Stochastic Sum of Random Variables
The following definitions and assumptions are used:

Rn == a sequence of random variables with a common marginal
distribution
N == the number of variables in the sequence and is independent of
Rn
R ==
E(Rn) == r
r:=l Rn
V(Rn) == (J2
The variance of R appearing in equation (3) of the text is determined
as follows:
V(R) = EN[V(R)IN] + VN[E(R)IN]
= EN[ r: V(Rn)+2r~ r:jCOV(RbRj)IN J+ VN [r: E(Rn)INJ

= (J2EN[r: IN] + 2EN[ r~ r:jCOV(Rb Rj)INJ + r 2VN [r: IN]
= (J2E(N) + r2V(N) + 2E[ r~ r:jCOV(Ri, Rj )]'
Appendix B
Residual Variance in Quarterly Underwriting Returns
Under the assumption that the number of underwritings and the return
per underwriting are independent, the residual component in table 4-8, as
shown in equation (3), is
"Residual" = 2E [r~ r:Fov(R i , Rj )]
where E[ ] is taken over N. A rough attempt was made to determine the

importance of underwriting return covariances on the quarterly return
variance by examining covariances on returns to "nearby" underwritings.
For this purpose, seasoned issues and IPOs over our eleven-year sam-
ple were each chronologically ordered. For multiple underwritings on the
same day, the ordering was random. If i is the ith-ordered underwriting
and j is the jth-ordered underwriting, for j > i, the "distance" between i
and j is defined as j - i. For IPOs, we calculated correlations and covariances
between dollar returns to underwritings i and j for "distances" between i
and j of 1, 2, 3, ... , 37 (the average number of IPO underwritings per
quarter was 37).29 The average correlation was .018 and the average
covariance was $5.7(10)10. There was no readily apparent pattern in the
correlations and covariances as the "distance" increased up to 37. 30
Evaluating "Residual" at the mean covariance (with N = 37) yields
$756(10)11. This compares with the actual residual in table 4-8a of

$1841(10)11. As a rough indication, "Residual" evaluated at the mean
covariance suggests that a substantial fraction of the actual IPO residual
reported in table 4-8a can be attributed to small positive covariances
between returns on nearby underwritings.
The same exercise was repeated for seasoned offerings, except that the
correlations between returns on i and j were calculated for "distances" up
to 59 (the average number of seasoned offerings per quarter). The average
correlation and covariance were .008 and $4.8(10)10, respectively. Again,
there was no readily apparent pattern in the correlations and covariances
as the "distance" increased up to 59. The estimated "Residual" with the
covariances evaluated at the mean was $1628(10)11, which compares with
the actual residual in the table of $3409(10)11.
Finally, we calculated correlation coefficients between the average dollar
return per underwriting each quarter and the number of issues underwritten
in the quarter for IPOs and seasoned issues. If the dollar return on individual
underwritings within the quarter is independent of the number of
underwritings (as assumed in equation (3», we would expect these cor-
relations to be zero. The actual correlations were .39 for IPOs, which is
significant at the one percent level, and .12 for seasoned offerings which
is not significant even at the five percent level. These correlations suggest
that there may be some positive dependency between dollar returns on
individual IPO underwritings and the number of issues being underwritten
in the quarter. If so, this positive correlation w<;>uld also contribute to the
quarterly return variance through the residual component.
Notes
1. Kwast (1989) also reported that the standard deviation of asset returns across trading
departments of commercial banks, which would include all bank securities activities, was
much higher than that for the banks' return on assets excluding the trading department.
Benston (1990) provides a different view of the relative risks of commercial and investment
banking.
2. Allowance has to be made for leverage because of its positive effect on the firm's
equity return variance and market /3. The asset return standard deviation is important to
creditors (or a deposit insurer) concerned with the risk of default, as well as to stockholders.
In option based models of corporate debt yields, such as Merton (1984), the default premium is
an increasing function of the firm's asset return standard deviation.
3. Giddy (1985, pp. 166- 6) reports that underwriting revenue for NYSE firms doing a public
business averaged about 10 percent of total revenue between 1974 and 1982. Equity under-
writing revenue would be a smaller percentage.
4. The issuer, date of issue, public offering price, gross proceeds, underwriter spread,
managing underwriter, and total assets of the issuing firm are from the SEC Registration
Offerings Statistics (ROS) data. Market prices for the issues subsequent to the issue date are
from the CRSP data.
5. These issues account for 85 percent of the total amount of seasoned issues and 68
percent of IPOs registered with the SEC. All the issues could not be included because of
either the unavailability of CUSIP numbers or stock market price data for one, five, and ten
days after the issue date.
6. The mean firm size reported is based on the 86 percent of firms that reported total
assets.
7. Benveniste and Spindt (1990) also show that underpricing will increase the amount of
the issue that underwriters are able to informally pre-sell to "regular" customers.
8. Tinic (1988) has emphasized that underpricing provides protection against litigation
risk from dissatisfied investors. Litigation risks are not considered here.
9. While there is no allowance for a normal return between the time of the offer price
and when the subsequent market price is observed, using returns in excess of a normal
return would have no appreciable effect on the results. The one-day return for the CRSP
value-weighted market return index over our sample period is .05 percent.
10. The effects of underpricing will not be fully captured by the average underpricing
if underpricing on individual issues is positively related to the uncertainty of the market
price of the issue. Possibly greater price stabilization for IPOs poorly received by the market
also could be a factor reducing the lower end of the range for IPOs as compared to that for
seasoned issues (see Giddy, p. 155).
11. Logue and Lindvall (1974) found the underwriting spread was negatively related to
the issuing firm's sales which they interpreted as proxying for price uncertainty. Pugel and
White (1985) did not find a significant relation between the underwriting spread and the
mean square change in an over-the-counter stock index for the days preceding the offering,
which was their measure of price uncertainty. Booth and Smith (1986) did not find a relation
between underwriting spreads and the variance of the systematic component of the stock's
monthly returns prior to the offering. However, they did get some indication of a positive
relation between the spread and the ratio of the idiosyncratic and systematic variances.
12. Our sample was limited by the small number of investment banks who are frequent
underwriters of equity and met our criterion of having publicly traded stock over a substan-
tial part of our eleven-year sample period.
13. This is an alternative form of the "event study" model. It is recommended by Acharya
(1989) when the firm experiences repetitive events. For another application, see Malatesta
and Thompson (1985).
14. Even though the price risk variable uses the day-after market price, the market may
have information on the market's reception of the offering on the day of the offering.
15. The estimated market f3s using daily returns are considerably lower than those re-
ported for investment banks by Boyd and Graham (1988) (an average of 1.69 for securities
firms using annual return data). Estimated market f3s using daily returns tend to be biased
because of non-synchronous trading that arise when stocks are not traded every day (see
Brown and Warner (1985». When 5-day returns are used, the estimated market f3s are: First
Boston 1.67, Hutton 1.57, Merrill Lynch 1.46, and Paine Webber 1.72.
16. Correlations between "low return" variables and the market return are low, suggest-
ing that low returns are not explained primarily by the overall market return. Thus,
multicollinearity is not an issue in interpreting the insignificant coefficients on the low return
variables.
17. Since our underwriting data includes only issues of domestic corporations, our esti-
mated losses do not include those experienced by U.S. underwriters of British Petroleum
stock in late October 1987 which were reported to be in the millions of dollars. While these
losses are directly attributable to the stock market crash, they reflect extenuating cir-
cumstances concerning differences in underwriting practices between the United King-
dom and the United States and the willingness of the U.S. underwriters to take a large risk
exposure. Of particular importance was the setting of a firm-commitment offer price more
than two weeks before the offering date (and before the market crash).
18. The moments of R, N, and Rn should be conditioned on information available at the
time the expectations are formed. In general, the conditional expectations will be time
variant. In this presentation, we ignore explicit conditioning and time variation in the mo-
ments of the distributions.
19. This assumption is considered below.
20. For example, returns to underwritings offered on the same day may have a positive
covariance if returns to the underlying stocks have positive market f3s or if the amounts being
underwritten on individual issues have positive serial correlation. Alternatively, if the size
distribution of underwritings in a given period was fixed (e.g., there were a fixed number of
small underwritings and a fixed number of large underwritings) but the sequence of
underwritings of different amounts was stochastic, the sum of the covariances would be
negative, ceteris paribus.
21. The estimated returns are for firm-commitment underwritings of common stock.
Results are reported below for underwritings that include preferred stock and best efforts
offerings.
22. Variances were also calculated using annual returns. As would be expected, the
annual return variances were much higher than the quarterly return variances. However, the
relative importance of the different sources of the return variance was similar to that pre-
sented below for the quarterly returns.
23. Returns to firm-commitment preferred stock offerings and to best efforts offerings
contain only the dollar underwriting spreads earned on the offerings. Thus, for firm-
commitment preferred underwriting, there is no allowance for losses due to unfavorable
market prices. The evidence for common stock underwritings that losses due to unfavorable
market prices are infrequent and relatively small can be expected to carryover to preferred
stock offerings as well.
24. For an individual underwriter, uncertainty about the volume of underwriting may be
even greater than indicated by the variance in the quarterly aggregate underwriting volume
because of unanticipated changes in its market share.
25. Returns to best efforts offerings were not included because the four underwriting
firms were agents for only a few of these offerings.
26. Over the 11 year sample period, Merrill Lynch was the largest underwriter of sea-
soned issues of stock. While its presence in this market was greater than that of other
underwriters, the fraction of its annual revenues from all types of underwriting appeared to
be commensurate with those for other large investment banks. Annual reports show that
investment banking income for Merrill Lynch, which included merger activity as well as
underwriting, accounted for 10.4 percent of its gross revenues over the period 1983-87.
27. Underwriting returns were regressed on the returns over the previous four quarters.
Significant coefficients were almost always confined to the first and second lags.
28. There was substantial serial correlation in the aggregate underwriting returns but
very little in the bank -specific returns.
29. Because of multiple underwritings on some days and variation in the number in days
between other underwritings, j - i is not a constant time interval. It serves only as a rough
measure of the temporal "closeness" of individual underwritings.
30. As the "distance" was expanded further, the average correlations and covariances
declined, indicating that the correlations and covariances tended to zero for large "dis-
tances." This result held for returns to both IPO and seasoned offerings.
References
Acharya, Sankarshan. 1989. "Insiders Decision of Corporate Events: A Simple

Measure of Stock Price Effect, Conditional on Outsiders Information."
Unpublished paper (May).
Baron, David P. and Bengt Holmstrom. 1980. "The Investment Banking Contract
for New Issues Under Asymmetric Information: Delegation and the Incentive
Problem." Journal of Finance 35(5) (December): 1115-1138.
Baron, David P. 1982. "A Model of the Demand for Investment Banking Advising
and Distribution Services for New Issues." Journal of Finance 37(4) (September):
955-976.
Benston, George J. 1990. The Separation of Commercial and Investment Banking.
New York: Oxford University Press.
Benveniste, Lawrence M. and Paul A. Spindt. 1990. "How Investment Bankers
Determine the Offer Price and Allocation of New Issues," Journal of Financial
Economics 24 (October): 343-36l.
Booth, James R. and Richard L. Smith. 1986. "Capital Raising, Underwriting and
the Certification Hypothesis," Journal of Financial Economics 15 (January/
February): 261-28l.
Boyd, John H. and Stanley L. Graham. 1988. "The Profitability and Risk Effects
of Allowing Bank Holding Companies to Merge with Other Financial Firms:
A Simulation Study," Federal Reserve Bank of Minneapolis Quarterly Review
12(2) (Spring): 3-20.
Brown, Stephen J. and Jerold B. Warner. 1985. "Using Daily Stock Returns: The
Case of Event Studies." Journal of Financial Economics 14 (March): 3-31.
Giddy, Ian H. 1985. "Is Equity Underwriting Risky for Commercial Bank
Affiliates?" In Deregulating Wall Street, Ingo Walter, Editor. New York: John
Wiley & Sons: 145-69.
Kwast, Myron L. 1989. "The Impact of Underwriting and Dealing on Bank Returns
and Risks," Journal of Banking and Finance 13 (March): 101-125.
Logue, Dennis and John R. Lindvall. 1974. "The Behavior of Investment Bankers:
An Econometric Investigation," Journal of Finance 29 (March): 203-215.
Malatesta, Paul H. and Rex Thompson. 1985. "Partially Anticipated Events: A
Model of Stock Price Reactions with an Application to Corporate Acquisitions,"
Journal of Financial Economics 14 (June): 237-250.
Mandelker, Gershon and Arthur Raviv. 1977. "Investment Banking: An Economic
Analysis of Optimal Underwriting Contracts." Journal of Finance 32 (June):
683-694.
Merton, Robert C. 1974. "On the Pricing of Corporate Debt: The Risk Structure
of Interest Rates," Journal of Finance 29 (May): 449-70.
Pugel, Thomas A. and Lawrence J. White. 1985. "An Analysis of the Competitive
Effects of Bank Affiliates to Underwrite Corporate Securities." In Deregulating
Wall Street, Ingo Walter, Editor. New York: John Wiley & Sons: 93-139.
Smith, Clifford W. 1986. "Investment Banking and the Capital Acquisition Process."
Journal of Financial Economics 15: 3-29.
Tinie, Seha M. 1988. "Anatomy of Initial Public Offerings of Common Stock."
Journal of Finance 43 (September): 789-822.
III
5 THE COMPETITIVE IMPACT OF
FOREIGN COMMERCIAL BANKS IN
THE UNITED STATES
Lawrence G. Goldberg
Foreign commercial banks have grown rapidly in the United States in the
1980s and have had a significant competitive impact. The growing role of
foreign banks in domestic markets has stimulated claims of unfair com-
petition. Banking is experiencing the same apprehension that has arisen
from foreign entry into other industries. This article reviews the history
and status of foreign banks in the United States. Reasons for the growth
of foreign banks are closely related to their competitive impact; therefore,
the paper summarizes studies that have analyzed the motivation for growth
of foreign banks in the United States. Because foreign banks can employ
various organizational forms, have entered different geographic areas
unevenly, have originated from different countries at different rates, and
have faced a changing legal and economic environment, an extensive dis-
cussion of the institutional framework within which the foreign banks
operate is provided. Through the use of descriptive material, the review
of past empirical studies and the development of an original empirical
analysis, this article will attempt to assess the competitive impact to date
and the likely future role of foreign banks in the American markets.
Foreign banks have grown much more rapidly in the United States
than American banks have grown abroad. In earlier years, U.S. banks had
far greater assets abroad than foreign banks had in the United States.
In 1955 the gross assets of U.S. bank branches abroad were only $2 billion,
161
but by 1972 had grown to $77.4 billion (representing about 80 percent of

the total American presence abroad), and by 1988 to $318 billion. The
number of foreign branches of U.S. banks grew from seven in 1955 to 627
in 1972 and to 849 in 1988. Foreign subsidiaries of U.S. banks had total
assets of $146.6 billion in 1988. 1 The assets of foreign banks in the United
States grew from only $28.3 billion in 1972 to $751.0 billion in 1989.2 The
total number of foreign bank offices in the United States at year-end 1989
was 704. 3
This pattern of greater foreign direct investment in banking in the
United States, compared to American investment abroad in the 1980s, is
similar to that found in other industries. The increased internationaliza-
tion of business has resulted in increased competition in many markets
that were previously protected from entry by vigorous competitors from
other countries. Whereas in the early part of the 1970s complaints of
unfair competition were leveled against American banks abroad, the latter
part of the 70s and the 80s found more complaints from domestic American
institutions. In fact, the International Banking Act of 1978 was intended
to equalize competitive opportunities between domestic and foreign banks
in the United State and will be discussed later. In 1992 Europe will allow
banks to operate across country borders. Because of reciprocity questions,
this could affect competition within the United States between foreign
and domestic banks and is discussed in the last part of this article.
Foreign banks can operate within the United States by employing various
organizational forms, each with its own purpose and focus. The competitive
analysis must distinguish among the major organizational forms. Foreign
banks are attracted to different states for different reasons and conse-
quently may compete for different types of business by state. The major
states with foreign banks are analyzed separately here. Finally, banks
from different countries might behave differently. Japan, which has by far
the largest number of offices and the greatest value of assets of foreign
banks in the United States, has attracted much attention and is accorded
special consideration here.
This article seeks to answer several questions about foreign banks in
the United States including:
1. What has affected their growth?

2. In what areas do they compete?
3. Do they have unfair competitive advantages?
4. What has been their competitive effect?
5. What will be the role of foreign banks in the United States in the
future?
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 163
Even though this article is unable to answer all of these questions com-
pletely, it does provide a useful start for the competitive analysis of foreign
banks.
The first section discusses the institutional setting of foreign banks in
the United States, including the types of foreign banking organizations,
the legal framework, and the extent of the foreign bank presence in the
United States. The second section reviews empirical studies of the factors
that have affected the overall growth of foreign banks, the growth by
country, and growth into various states. The growth of Japanese banks is
also reviewed. In the third section the descriptive evidence of competitive
effects is presented and the allegations of unfair competitive advantages
evaluated. Balance sheet ratios of foreign banks are compared to domestic
banks in the fourth section in order to determine in which activities the
foreign banks are concentrating their efforts. The final section summarizes
the evidence with respect to the competitive impact of foreign banks and
assesses the prospects for the future for foreign banks, including an
evaluation of the changes proposed for Europe in 1992.
The Institutional Setting
The first foreign bank to establish an office in the United States was the
Bank of Montreal, which set up an agency in New York in 1818. Other
well known banks that established offices early on were: Hong Kong and
Shanghai Banking Corporation (1875), Lloyds Bank International (1856),
Barclay's Bank International (1890), and Mitsubishi Bank (1920). These
banks engaged primarily in trade finance and funds transfers and par-
ticipated in the New York stock and bond markets. One study has esti-
mated that through 1985, "of foreign commercial banks currently operating
in the United States 74 per cent (224 banks from sixty countries) first
established their U.S. presence after 1970."4 From table 5-1 it can be
seen that only twenty-seven banks from twelve countries were operat-
ing in the United States prior to 1946; of these, seven were British,
five Canadian, three Italian, and three Swiss. 5 Furthermore, the data in
table 5-2 indicate that most of the growth of foreign banks in the United
States has occurred since 1972. The ratio of foreign controlled to do-
mestically controlled assets has risen dramatically from 3.6 percent in
1972 to 21.4 percent in 1989. Every year has witnessed an increase in the
share of foreign banks. These data by themselves would suggest that for-
eign banks are attracting relatively more business and placing competitive
Table 5-1. Year of United States Entry by Foreign Banks.
Number of Number of
Period Foreign Banks Home Countries
Before 1946 27 12
1946-1970 48 18
1971-1985 229 60
Unclassified entry year 2 2
Source: Cho, Krishnan, and Nigh (1987) 61.
pressure upon domestic banks. The areas where the competitive pressure
has been most intense are examined later in the article.
Foreign banks can utilize several different organizational forms to operate
in the United States. There have been changes in the relative advantages
and disadvantages in each form over time, and consequently, the rate of
growth of the usage of each form has varied. Since different forms permit
different types of activities to be performed and require different levels
of commitment by the parent organization, the choices enable foreign
banks to tailor their American operations to their business desires.
The most limited, but the easiest to establish of the organizational
forms, is the representative office. These offices neither take deposits nor
make loans, but can act as agents for the foreign bank and forward payments
or loan documents to the home office. Representative offices are often
established as a precursor for further involvement, and one study has
calculated that 60 percent of foreign banks through 1985 used the rep-
resentative office as their initial entree into the American market. 6
Agencies are an integral part of foreign banks and may make commercial
and industrial loans; however, they cannot make consumer loans nor accept
deposits. They do maintain credit balances that are very similar to deposits,
but most payments cannot be made from these accounts. Funding is from
the parent bank or by borrowing in the Federal Funds or interbank markets.
The branch is the most important organizational form and is an integral
part of the parent bank. It can offer a full range of services and, like the
agency, it has traditionally been engaged mostly in wholesale operations.
The final major form is the subsidiary or commercial bank. Subsidiar-
ies have identical powers as domestic banks and are regulated in the same
manner. Many are oriented toward retail business. Foreign banks can
establish subsidiaries either through acquisition or de novo entry. The
acquisition by foreign banks of several large domestic banks, such as
Marine Midland and Crocker in 1979 and 1980, alarmed many people and
Table 5-2. Total U.S. Assets of Foreign-Controlled U.S. Banking Offices, 1972 to 1989 (Billions of Dollars).
u.s. Banking Offices
Owned by Foreign Banks Memo: Ratio of
u.s. Banks Foreign-Controlled
U.S. Branches Commercial Foreign to Total Domestic
Year and Agencies Banks Other * Individuals Total Banking Assests**
1972 22.2 4.4 1.1 0.6 28.3 3.6
1973 25.2 5.4 1.5 1.0 33.1 3.7
1974 34.0 10.8 1.9 0.6 47.3 4.8
1975 38.2 11.8 2.0 2.1 54.1 5.3
1976 45.7 13.8 1.5 2.7 63.7 5.9
1977 59.1 16.2 1.6 4.9 81.8 6.7
1978 86.8 20.7 2.0 6.4 115.9 8.4
1979 113.5 34.6 2.4 7.7 158.2 10.3
1980 148.3 68.1 2.8 10.2 229.4 13.5
1981 172.6 78.5 3.2 11.5 265.8 14.2
1982 208.2 90.9 3.9 17.5 320.5 15.3
1983 229.0 78.5 4.2 19.7 352.2 15.4
1984 273.2 103.1 4.5 19.8 400.6 16.1
1985 312.4 111.3 5.4 23.0 452.1 16.5
1986 398.1 109.5 5.3 27.0 539.9 17.9
1987 462.7 112.8 6.1 28.7 610.3 19.8
1988 515.3 123.7 6.6 28.3 673.9 20.6
1989 581.3 134.1 6.5 29.1 751.0 21.4
* Includes N.Y. Investment corporations and directly owned Edge corporations.
** Total domestic banking assets include the assets of domestic offices of insured commercial banks plus those of U.S. branches and
agencies of foreign banks. Consequently, in this table they also include balances booked in IBFs.
Source: Houpt (1988), p. 25, and additional information from J. Houpt. Original data were obtained from call reports.
stimulated several proposals to limit foreign bank growth in the United

States.
Other less important forms should be mentioned for completeness.
Investment companies are similar to agencies but are the only form
permitted to deal in securities. They are limited to New York state. The
International Banking Act permitted foreign banks to establish Edge Act
Corporations to maintain competitive equality with domestic banks. They
are chartered by the Federal Reserve Board and must engage in only
international banking activities outside the home state of the parent bank.
Table 5-3, taken from a recent survey by the American Banker, indi-
cates the size of foreign banks by type and in relation to domestic banks
for June 1989 and June 1988 as measured in four different ways: number
of offices, C&I loans, deposits, and assets. Clearly, foreign banks now
play an important role in the American banking scene. Significantly, it
should be noted that since 1982, by all four measures, foreign banks have
expanded steadily compared to domestic banks. In terms of assets in
1989, branches were by far the most important form with 63.8 percent of
total foreign assets, followed by commercial banks with 23.2 percent, and
agencies with 12.1 percent. The higher ratios of C&I loans, than other
measures, indicates that foreign banks may be more important in the
market for loans to large companies. This will be further explored when
the financial ratios of foreign banks in relation to domestic banks are
analyzed.
Foreign banks from more than sixty countries have some form of office
in the United States. Table 5-4 lists the assets by country of foreign banks
for 1988. The ten most important countries, however, account for almost
90 percent of the total assets of foreign banks in the United States, and
table 5-5 presents the total assets for banks from each of these countries
from 1980 to 1988. In 1988, the Japanese share was greater than 55 percent
of all foreign banks and was more than five times larger than the next two
most important countries, Canada and Italy. Note that there is quite a bit
of variability among the countries over time. Some, such as Italy, Japan,
and Canada, have grown rapidly, while others, such as West Germany,
France, and the United Kingdom have experienced a more moderate rate
of growth.
The dominance of the Japanese is further illustrated by a list in table
5-6 of the twenty-five top foreign banks in the United States in 1989. Of
the ten largest banks, eight are Japanese, and of the twenty largest, fifteen
are Japanese. Consequently, it is appropriate to examine closely the
Japanese banks in the United States.
Foreign bank activity in the United States varies widely by state. Part
of this variance is due to the types of laws erected by states to either

attract or deter foreign banks, but it is also due to the economic attrac-
tiveness of the states to foreign banks. Table 5-7 presents the total assets
of United States offices of foreign banks by state from 1980 to 1988. In
1988, only sixteen states plus the District of Columbia had foreign bank
offices, with New York, by far the most important, with greater than two-
thirds of total assets. California is the second most important state and
has been particularly attractive for Japanese banks. It is not surprising
that the top three states, New York, California, and Illinois, accounted for
94.6 percent of total foreign bank assets in 1988, since these three states
contain the major international financial centers in the United States. On
the other hand, there has been substantial growth of foreign banking
activity in other states. Florida, for example, has attracted almost as many
foreign banks as Illinois. Since Florida serves primarily as a point of entry
for foreign wealth from Latin America and the Caribbean, it has a major
foreign bank presence in deposit-taking, but relatively little in lending. In
the later analysis of foreign bank financial ratios, each of the four largest
states for foreign bank activity is examined separately.
In a discussion of the competitive impact of foreign banks in the United
States, it is important to recognize the legal environment in which the
foreign banks operate. The most important law of recent vintage is the
International Banking Act of 1978. Prior to 1978, many aspects of foreign
bank activity in the United States were unregulated and domestic banks
claimed that foreign banks had unfair competitive advantages. In particular,
foreign banks had significant authority to establish offices across state
lines, while domestic banks were severely limited. The main purpose of
the International Banking Act was to equalize the treatment of foreign
and domestic banks. The act limits the multi-state operations of foreign
banks by prohibiting them from establishing offices outside their own
home state. New branches and agencies established by foreign banks outside
their home states could only accept the same types of deposits accepted
by the Edge Act Corporation, a form available to domestic banks. However,
foreign banks were allowed to retain their existing interstate operations.
This led to a flurry of foreign bank expansion across state lines prior to
the effective date of the law. Federal chartering of branches and agencies
was established. The Federal Reserve was given authority to impose reserve
requirements on both federal and state chartered branches and agencies
and FDIC insurance was required for all branches taking retail deposits.
Foreign banks were made subject to the non-banking provisions of the
Bank Holding Company Act and thus could no longer engage in new
activities not permitted by domestic bank holding companies. Finally, the
.......
0\
00 Table 5-3. Foreign Banks in the United States, summary of Survey Findings. C&I Loans, Deposits, and Assets at U.S.
Offices of Foreign Banks (Dollar Amounts in Billions).
Number of Offices C&I Loans Deposits Assets
1. Agencies and Branches 6/30/89 565 $129.9 $247.1 $527.5
6/30/88 543 118.5 216.0 479.6
+22 +9.6% +14.4% +10.0%
1A. Agencies 6/30/89 201 $27.8 $22.4 $83.9
6/30/88 194 25.2 17.1 86.1
+7 +10.3% +31.0% -2.6%
lB. Branches 6/30/89 364 $102.1 $224.7 $443.5
6/30/88 349 93.3 198.9 393.5
+15 +9.4% +13.0% +12.7%
2. Commercial Banks (a) 6/30/89 100 $46.6 $121.3 $161.3
6/30/88 85 43.9 112.9 152.0
+15 +6.2% +7.4% +6.1%
3. Edge Act Banks (b) 6130/89 26 $0.26 $1.6 $2.0
6/30/88 28 0.29 1.9 2.4
-2 -10.3% -15.8% -16.7%
4. Investment Companies 6/30/89 11 $2.2 $0.4 $4.9
6/30/88 10 1.7 0.3 4.2
+1 +29.4% +33.3% +16.7%
Totals for U.S. Offices of Foreign Banks (c) 6/30/89 697 $178.9 $370.4 $695.6
6/30/88 666 164.3 331.2 638.3
+31 +8.9% +11.8% +9.0%
Total for All U.S. Banks (d) 6/30/89 13,639 $628.3 $2,138.2 $3,073.6
6/30/88 14,090 598.2 2,042.5 2,927.5
-451 +5.0% +4.7% +5.0%
Ratio of Foreign to All U.S. Banks (e) 6/30/89 5.1% 28.5% 17.3% 22.6%
6/30/88 4.7% 27.5% 16.2% 21.8%
6/30/87 4.4% 24.8% 16.3% 20.9%
6/30/86 4.3% 22.1% 14.2% 18.4%
6/30/85 4.1% 21.8% 15.6% 18.6%
6/30/84 3.6% 20.5% 14.9% 17.7%
6/30/83 3.6% 21.4% 12.4% 16.0%
6/30/82 3.2% 21.7% 11.2% 15.5%
All data totals and the number of offices shown here-under each office type-are for the banking offices providing data for the survey.
C&I loans, deposits and assets for June 30, 1989 and June 30, 1988 were used for all banking offices.
(a) Commercial bank number count includes bank head offices only. Branches of these banks are not included.
(b) Edge bank number count includes Edge branches. To avoid double counting, data for Edge banks, which are subsidiaries of foreign
bank-owned U.S. commercial banks, are excluded from the totals for C&I loans, deposits, and assets.
(c) The total number of U.S. banking offices of foreign banks-including those offices for which data were not available-were 704 on
6/30/89 and 675 on 6/30/88.
(d) The asset, deposit, and C&I data for all U.S. banks are from the Federal Reserve Bulletin. The number of U.S. banks includes all
U.S.-chartered, FDIC-insured commercial banks; the U.S. branches and agencies of foreign banks; all U.S. Edge Act bank offices; and New
York State investment companies. These data were provided by the Federal Deposit Insurance Corporation, and Federal Reserve Board.
(e) The ratio is all U.S. offices of foreign banks-including those for which data were not available-divided by all U.S. banking offices.
Source: American Banker, February 27, 1990, 18A.
.....
$
......
-..l
0
Table 5-4. Total Assets by Country of the U.S. Offices of Foreign Banks Listed by Country All Countries (1988). (Assets
in Thousands of Dollars).
Country Assets Country Assets Country Assets
Austria 1,197,628 Canada 39,252,400 Iran 17,266
Belgium 3,131,391 Argentina 864,630 Israel 9,249,139
Denmark 2,047,792 Brazil 4,282,035 Japan 360,853,869
Finland 1,419,688 Chile 85,969 Jordan 316,177
France 29,255,588 Colombia 825,477 Korea, South 4,616,861
Germany (FDR) 13,024,980 Dominican Rep. 29,154 Kuwait 993,404
Greece 1,224,959 Ecuador 187,652 Malaysia 525,101
Ireland 6,425,407 EI Salvador 874 Pakistan 669,442
Italy 29,859,425 Mexico 4,431,142 Phillipines 321,381
Luxembourg 756,076 Panama 136,424 Qatar 116,600
Netherlands 10,126,495 Peru 30,756 Saudi Arabia 743,798
Norway 1,153,410 Uruguay 214,022 Singapore 380,246
Portugal 658,407 Venezuela 2,602,703 China, Rep. of 2,744,694
Spain 6,085,742 Bermuda 137,868 Thailand 824,274
Sweden 1,210,392 Cayman Islands 130,387 United Arab Emir. 257,735
Switzerland 23,879,249 Bahrain 1,027,948 Egypt 174,952
Turkey 310,421 China, Peoples Rep. 964,112 Australia 3,930,744
U.K. 32,708,605 Hong Kong 25,732,010 New Zealand 892,436
Yugoslavia 655,590 India 750,346 Multiple Countr. 936,426
Other W. Europe 5,044,752 Indonesia 2,665,731
Grand TotaL for aLL Countries: $653,092,182
Source: Federal Reserve System.

Table 5-5. Total Assets of the U.S. Offices of Foreign Banks, 1980 to 1988. Ten Largest Countries (as of 1988). (Thousands
of Dollars)
Ten Largest Countries 1980 1981 1982 1983 1984 1985 1986 1987 1988
Canada 15,718,428 21,503,472 22,148,288 25,838,300 38,101,415 39,553,299 42,431,042 44,244,905 39,252,400
France 12,925,772 16,914,185 15,736,972 16,101,115 18,384,141 20,653,757 20,653,757 24,568,815 29,255,588
Gennany, West
(Federal Republic) 7,253,180 7,379,767 8,880,866 7,382,739 7,565,605 8,801,212 11,049,175 13,511,074 13,024,980
Hong Kong 11,920,682 12,983,516 16,707,658 19,705,400 17,288,573 23,376,817 24,920,412 25,561,199 25,732,010
Israel 4,097,400 4,239,426 6,084,051 7,096,780 7,863,308 7,812,780 8,074,245 8,349,632 9,249,139
Italy 9,216,825 10,891,512 14,718,455 17,523,890 23,931,971 29,090,071 36,445,780 41,014,763 39,859,425
Japan 72,484,137 88,646,854 113,005,182 125,982,961 151,259,221 178,761,678 245,571,205 294,499,442 360,853,869
Netherlands 36,681,150 4,813,409 5,284,834 4,894,644 5,335,755 7,134,185 8,539,678 8,684,287 10,126,495
Switzerland 11,312,568 11,225,989 12,929,098 13,215,445 15,280,357 18,338,243 24,518,535 27,957,911 23,879,249
United Kingdom 25,136,319 46,445,237 52,171,252 53,058,135 51,443,593 57,170,899 40,631,292 43,685,868 32,708,605
Totals:
Ten Countries 173,733,461 225,043,367 267,666,656 290,799,409 336,453,939 390,692,941 422,835,121 532,077,896 583,941,760
All Countries 198,115,287 251,217,538 301,021,581 333,336,188 378,313,617 441,525,633 526,589,545 594,077,896 653,092,182
Source: Federal Reserve System.
......
-..J
......
......
-.I
N
Table 5-6. Top twenty-five Foreign Banks in the United States.

Based on Business Loans booked through U.S. banking offices
June 30, 1989, compared with June 30, 1988 (dollar amounts in millions).
Totals on June 30, 1989 Totals on June 30, 1988
Rank Rank No. of % Chg. No. of % Chg.

6/89 6/88 C&I Loans Assets Deposits U.S. Off. C&I C&I Loans Assets Deposits U.S. Off. C&I
Bank of Tokyo Ltd. (a) $12,821 $47,210 $25,544 11 +41.67 $9.050 $32,215 $18,070 11 +37.2
2 4 National Westminster Bank Pic,
London 9,595 22,189 17,281 6 +16.66 7,523 17,742 14,799 7 +17.7
3 3 Mitsubishi Bank Ltd., Tokyo 8,357 31,742 15,681 5 +5.9 8,071 32,881 15,689 9 +52.2
4 2 Sumitomo Bank Ltd., Osaka 7,914 24,472 16,551 7 -1.45 8,251 22,722 14,040 8 +64.8
5 7 Industrial Bank of Japan Ltd., Tokyo 7,886 28,710 17,700 7 +31.69 5,988 25,625 12,024 7 +34.8
6 6 Sanwa Bank Ltd., Osaka 7,859 28,763 15,190 8 +13.98 6,895 25,228 13,628 8 +63.8
7 10 Fuji Bank Ltd., Tokyo 7,838 30,372 13,009 8 +40.07 5,595 29,124 12,514 8 +14.2
8 9 Dai-ichi Kangyo Bank Ltd., Tokyo 6,960 33,913 16,070 5 +20.83 5,761 29,361 13,588 5 +23.2
9 5 Hongkong & Shanghai Banking Corp. 6,457 25,380 18,490 18 +9.03 7,104 26,762 19,266 21 -4.9
10 11 Bank of Montreal 5,409 18,936 9,667 16 +0.64 5,374 18,725 10,881 14 +25.5
11 12 Long-Term Credit Bank of Japan,
Ltd., Tokyo 5,223 10,543 3,664 4 +29.62 4,030 8,514 2,616 3 +60.5
12 15 Tokai Bank, Ltd., Nagoya 4,066 17,663 7,729 4 +27.64 3,185 16,684 6,390 4 +35.0
13 13 Mitsui Bank Ltd., Tokyo 3,971 11,209 6,254 4 +6.68 3,722 9,680 2,648 4 +14.0
14 8 Swiss Bank Corp., Basle 3,661 11.398 6,444 7 -38.81 5,979 14,414 8,011 7 +36.3
15 22 Mitsui Trust & Banking Co. Ltd.,
Tokyo 3,647 12,081 7,177 4 +68.35 2,167 10,893 4,889 4 +19.3
16 19 Nippon Credit Bank Ltd., Tokyo 3,572 6,499 2,046 2 +47.41 2,424 5,155 1,575 2 +14.3
17 17 Taiyo Kobe Bank Ltd., Kobe, Japan 2,939 8,453 3,929 5 +8.03 2,720 7,747 3,235 4 +74.3
18 21 Mitsubishi Trust & Banking Corp.,
Tokyo 2,705 10,180 4,768 3 +16.44 2,323 11,797 5,716 3 +12.1
19 20 Banco di Nipoli, Naples 2,635 7,653 4,076 +13.06 2,330 5,136 2,925 -12.4
20 14 Bank of Nova Scotia, Toronto,
Canada 2,621 8,123 3,218 9 -28.03 3,642 10,370 2,853 9 +2.9
21 18 Daiwa Bank, Ltd., Osaka 2,378 11,671 6,296 4 -12.13 2,707 9,650 4,918 4 +23.6
22 25 Banco di Roma. Rome 2,102 8,786 3,468 5 +13.65 1,849 7,677 3,355 4 +65.7
23 26 Banque Nationale de Paris 2,012 9,261 5,230 8 +17.72 1,709 8,482 4,584 9 -23.9
24 33 Sumitomo Trust & Banking Co. Ltd.,
Osaka 1,926 9,830 3,218 3 +35.57 1,420 8,024 4,676 3 +24.9
25 23 Algemene Bank Nederland, Amsterdam 1,830 7,124 4,779 19 -12.71 2,096 5,133 2,648 14 +30.0
Totals for the top 25 foreign banks in the U.S. 126,394 442,661 237,479 173 +12.94% 111,915 399,741 205,538 173 +22.0%
Totals for all foreign banks in the U.S. 178,900 695,600 370,400 697 +8.90% 164,300 638,300 331,200 666 +18.6%
Top 25 banks' share of total for all foreign banks 70.6% 63.6% 64.1% 24.8% 68.1% 62.6% 62.1% 25.9%
Sources: (1) American Banker survey questionnaires and Federal Reserve Board data. The totals for each bank are based on the combined business loans (C&! loans), deposits, and assets
for all of the bank's U.S. banking offices, for which data were available. U.S. banking offices include: agency offices, branches, commercial bank subsidiaries, New York State investment companies,
and Edge banks. % Chg. C&I is the annual growth in commercial and industrial loans over the year ended 6130/89 and 6130/88, respectively. Note: percent changes based on unrounded figures
and may differ slightly from dollar figures shown (a)-The data for 6130/88 does not include the acquisition of Union Bank, Los Angeles, which was completed on 10/31188.
(2) American Banker, February 27, 1990, p. 17A.
>-'
-l
tN
.....
-.l
+:>.
Table 5-7. Total Assets of the U.S. Offices of Foreign Banks, 1980 to 1988. (Thousands of Dollars)
State 1980 1981 1982 1983 1984 1985 1986 1987 1988
New York 132,185,600 149,759,632 190,238,583 214,902,610 243,689,943 293,966,846 367,366,289 412,144,700 453,101,639
California 50,051,420 79,694,302 86,269,854 88,555,387 93,341,093 100,852,400 99,066,950 106,682,400 112,895,868
Illinois 8,323,716 9,514,448 11,100,842 11,905,420 22,814,261 26,126,252 32,816,602 40,387,480 48,124,821
Florida 671,301 1,358,633 3,064,671 3,995,317 4,668,136 6,040,459 7,662,917 9,387,411 9,306,516
Georgia 1,809,330 2,015,763 2,401,384 2,579,083 3,531,717 3,904,019 4,593,574 5,617,684 6,043,273
Maryland 0 0 0 2,980,514 3,393,462 3,819,452 4,539,586 4,709,410 5,485,122
Oregon 3,731,203 5,996,888 4,819,417 4,745,088 3,359,525 2,001,934 2,615,469 2,970,110 3,849,959
Texas 48,216 59,936 95,776 103,085 67,432 196,431 2,079,911 2,427,095 3,055,482
Delaware 0 0 0 124,166 161,327 776,348 2,007,401 2,648,545 3,250,668
Washington 1,630,005 1,571,729 1,662,378 1,883,312 1,806,100 1,650,241 2,005,073 2,355,573 2,093,898
Pennsylvania 360,299 698,339 709,686 883,355 797,578 974,223 822,471 1,052,836 384,189
Massachusetts 278,170 280,435 340,099 335,573 343,980 405,656 511,127 821,641 1,005;386
District of Columbia 5,355 173,175 178,062 263,632 299,137 418,055 435,804 0 368,434
New Mexico 0 0 0 0 0 360,603 362,273 339,012 284,376
Hawaii 22,699 24,707 28,099 36,750 37,926 30,714 48,174 48,782 117,283
Louisiana 0 69,551 113,530 142,896 2,000 2,000 2,000 2,000 0
Arizona 0 0 0 0 0 0 0 0 25,740
North Carolina 0 0 0 0 0 0 0 0 249,128
Totals 199,115,314 251,217,538 301,022,381 333,436,188 378,313,617 441,525,633 526,935,621 591,204,885 649,641,782
Source: Federal Reserve System

power of Edge Corporations was expanded and foreign banks were

permitted to own Edge Corporations. This act went a long way toward
equalizing the competitive opportunities available to foreign and domestic
banks, but once again, arguments are being made that foreign banks have
competitive advantages.
The Motivation for Foreign Bank Growth in the United

States
In order to evaluate the competitive situation, it is important to first

understand the factors that have affected the growth of foreign banks in
the United States. Since 1981, a number of studies have been done to
empirically isolate the factors most responsible for foreign bank growth.
Several of the major studies are summarized here, including studies of
overall foreign bank growth, by country of origin, and by state. In addi-
tion, we examine Japanese bank growth in the United States because of
the import presence of Japanese banks.
The first study estimating the factors that determine the level of for-
eign bank activity in the United States was Goldberg and Saunders (1981a).
The study employed quarterly-time-series data from 1972 through 1979 to
estimate the following model of foreign bank growth:
y\ = f(RDIFF1, PIE, G, MN, FDI, CA, $IDM, DUM77),
where
Y\ = foreign banks' share of total U.S. commercial bank assets.
RDIFF1 = difference between the quarterly average Federal funds rate
and the 3-month Eurodollar rate.
PIE = the price-earnings ratio for bank stocks.
G = the growth of U.S. GNP relative to the growth of OECD
countries' GNP.
MN = U.S. imports divided by U.S. personal income.
FDI = net foreign domestic investment in the U.S.
CA = U.S. balance of payments (balance on current amount).
$IDM = the exchange rate between dollars and German marks
(dollars per mark).
DUM77 = dummy variable for anticipation of International Banking
Act. (1)
The regression estimates are presented in equation (2).

Y1 = -0.0030 - 0.00268 RDIFF1 - 0.0023 8 PIE + 0.0004 G + 0.1654 MN
(0.093) (-5.112) (-2.521) (0.509) (0.774)
+0.00006 8 FDI - 0.00001 b CA + 0.00198 $IDM - 0.0103" DUM77
(2.690) (-1.939) (6.303) (-2.540)
D - W = 1.8130
R2 = 0.9332
" Significant at 5 percent level.
b Significant at 10 percent level. (2)
Results using other dependent variables were very similar. The study
concluded "that the most important factors determining foreign bank growth
over the period of the study were (i) the size of interest differentials
between United States and foreign deposits and loans, (ii) the falling PI
E ratios for United States bank stocks, (iii) the increased size of (net)
foreign domestic investment in the United States, (iv) the persistent de-
preciation in the dollar, and (v) expectations that the IBA would have a
restrictive effect on foreign bank activity in the United States.,,7
A second study by Goldberg and Saunders (1981b) examined the growth
of foreign banks by agency, branch, and subsidiary with a similar model
of four explanatory factors. "The empirical results suggested that these
factors tended to impact on agencies in a different manner from branches
and subsidiaries, with agencies more affected by international business
considerations and more concerned with current, or short-run, profitabil-
ity. Also, the IBA appeared to have a negative affect on the underlying
growth of agencies and branches leaving subsidiaries relatively unaffected.,,8
Recently, Hultman and McGee (1989) updated these studies employ-
ing a model with three independent variables. They examined foreign
bank activity in the United States between 1973 and 1986 and found the
growth of foreign branches and agencies positively related to changes in
foreign direct investment in the United States, the value of the dollar, and
the International Banking Act of 1978. The growth of subsidiaries was
positively related to the first two of these variables and negatively related
to the bank price earnings ratio.
Useful insights can be obtained by examining the factors that have
resulted in differential growth across countries of origin of the foreign
banks. Grosse and Goldberg (1991) use pooled time-series cross-section
data from 1980 to 1988 to determine the factors affecting growth of foreign
banks from source countries. "The results indicate that foreign banks are
THE COMPETITIVE IMPACf OF FOREIGN COMMERCIAL BANKS 177
drawn to the United States to service customers from the home country,
since bank presence is positively related to foreign direct and portfolio
investment and foreign trade with the United States. Larger banking sectors
in the foreign countries encourage greater investment in banking operations
in the United States. Less stable countries are more likely to have greater
presences in the U.S. banking sector, holding other factors constant.
Finally, the greater the geographic or cultural distance of the source
country from the United States, the more likely are banks to build their
asset bases in the U.S. market.,,9
Until recently, foreign bank data by country were only available for
Japan, Canada, and all of Europe. Hultman and McGee (1989) have used
the data for Japan to identify factors affecting the growth of Japanese
banks in the United States. They find that the share of total banking
assets of Japanese banks in the United States is positively related to
Japanese foreign direct investment in the United States, and the Japanese
Banking Act of 1982 that opened Japan somewhat to foreign banks, and
negatively related to the value of the yen in relation to the dollar. Terrell,
Dohner, and Lowrey (1990) find that "U.S. activities of Japanese banks
during this period appeared strongly related to Japanese domestic financial
variables as well as to conditions in the U.S. market. Commercial and
industrial loans at these offices responded both to expansions in Japanese
trade and to restraints on domestic Japanese interest rates, while interbank
trading at U.S. offices of Japanese banks responded to both price and
quantity restraints on domestic Japanese banking activity.,,10
It is also useful to examine the factors causing differential foreign bank
growth across states. Goldberg and Grosse (1990) find that the foreign
bank presence, whether measured by assets or number of offices, is drawn
to states with large bank market sizes and low levels of regulation of
foreign banks. The level of international trade of the state was also im-
portant in explaining foreign bank assets. New York was entered as a
dummy variable because many foreign banks are primarily interested in
wholesale business and much of the foreign exchange and money market
business is concentrated in New York. The coefficient was the expected
positive sign.
These studies suggest that a number of economic factors and an im-
portant regulatory change affect the growth of foreign banks. The im-
portance of foreign trade and foreign direct investment suggest that foreign
banks frequently are serving customers from their home countries. Rela-
tive profitability, the cost of acquiring banks relative to earnings, and the
value of the dollar appear to motivate banks in the expected directions.
Since past regulatory changes have affected foreign bank growth, it might
be expected that future changes in the regulatory environment could also

affect their growth.
The Competitive Impact of Foreign Banks
Many observers have commented on the competitive impact of foreign

banks. In addition, many have claimed that foreign banks have a number
of competitive advantages over domestic banks because of regulatory factors
and differences in economic factors in the home countries of the parent
organizations. This section first assesses the arguments about the com-
petitive advantages of foreign banks and then indicates the specific areas
in which foreign banks have presumably provided competitive alterna-
tives to domestic banks. The major area of competitive impact is commercial
and industrial loans, and this activity is highlighted. The section concludes
with an analysis of legal differences among states and the relationship to
the doctrine of national treatment espoused by the International Banking
Act of 1978. The following section provides empirical evidence from the
balance sheets of both foreign and domestic banks indicating the areas
where foreign banks have the greatest competitive presence.
It has been argued that foreigners have been able to penetrate American
banking because of greater funds availability. Most foreign countries have
higher savings rates than the United States, and certain countries have
earned large trade surpluses and are trying to find places to invest. ll Note
that these arguments, as well as other arguments, are most applicable to
the case of Japan. For example, Zimmerman (1987) asserts that Japanese
banks in California may be able to borrow at lower cost from their parent
organizations because of greater funds availability. In their lending activities
foreigners are said to be more accommodating on paperwork and not as
rigid on documentation as U.S. banks.12 Finally, foreign banks are alleg-
edly more flexible in how they are compensated for loans. They allow the
borrower to determine how the price of the loan will be paid by a com-
bination of fees and balances.13
It is widely claimed that foreign banks as new entrants have priced
their products (particularly commercial and industrial loans) below domestic
competitors in order to obtain business. They have been willing to accept
smaller profit margins than their domestic competitors. It is claimed that
they can do this because of lower capital requirements and a greater
ability to utilize leverage. This also relates to claims that lax regulation
of banks by foreign governments provides a competitive advantage over
U.S. banks. Excessive regulation in home markets has clearly stimulated
u.s. bank activity in the Eurodollar markets l4 and Japanese bank activity
in Eurodollar and European markets,IS but the evidence is less clear re-
garding competitive relationships in the United States. Baer (1990) claims
that "fears regarding the competitive advantages conveyed by lax regu-
lation at home may be justified, particularly with respect to banks owned
by foreign governments. And although no objective rankings exist, this
concern would also appear to be valid where privately-owned foreign
banks enjoy stronger guarantees from their governments than U.S. banks
enjoy from the U.S. government. Whatever the particulars of the com-
plaint, it ultimately boils down to the assertion that foreign banks are able
to hold less capital per dollar of risk or pay less for the capital that they
raise.,,16
The evidence on capital levels is mixed. Book values of Japanese banks,
for example, are very low compared to American banks. The Japanese
banks, however, have large gains on equity holdings that are not recog-
nized on their balance sheets. When market capitalization is considered,
the Japanese banks have more capital than do banks from any of the
major countries. Recent international efforts have been made to equalize
capital requirements of banks across countries, but implementation of
equalized risk-based capital requirements will be difficult because of dif-
fering accounting practices across countries.
The major competitive impact of foreign banks has been in commercial
and industrial loans. From table 5-3 it can be seen that foreign banks have
a disproportionate amount of their assets concentrated in commercial and
industrial loans. Moreover, the foreign bank share of these loans has risen
steadily since 1982. In order to attract new business in these loans, foreign
banks, and especially the Japanese, are offering lower interest rates. "Where
spreads of seventy-five basis points over prime were once considered the
absolute minimum, some banks, notably the Japanese again, are earning
a reputation for ruthlessness. Comments John Miller, vice president in
the New York office of San Francisco-based Wells Fargo Bank: 'The
foreign banks are cutting their spreads down as low as ten basis points
over prime just to get the business.' ,,17 This has presented domestic banks
with a major competitive challenge.
Much of the increase in C&I loans held by foreign banks has resulted
from the sale of domestic C&I loans by U.S. commercial banks. These
purchases of existing loans have apparently become important since the
mid-1980s and in 1988 accounted for 2.5 percent of total C&I loans. "Thus,
U.S. banks have been directly responsible for over two-fifths of the 5.8
percentage point increase in the market share of U.S. branches of foreign
banks that occurred between 1980 and 1988."18 Generally, smaller foreign
banks have been more interested in loan participations because they have
less access to major corporate borrowers.
Increasingly, the larger foreign banks are moving toward competing
with domestic banks in a larger range of financial services. They have
moved away from serving only trade finance and large corporations
toward serving small and mid-sized companies. Cho, Krishnan, and Nigh
(1987) have surveyed 271 foreign banks in the United States: " ... 152
banks (56 percent) pointed out trade financing as one of the main areas
of specialization in their U.S. offices. One hundred and nineteen banks
(44 percent) listed corporate banking and 81 banks (31 percent) listed
foreign exchange trading as one of their main areas of specialization. ,,19
Other areas mentioned were money market activities, retail banking, and
business development.
There is much anecdotal evidence of foreign bank participation in a
number of banking activities. Foreign banks participate heavily in finan-
cial loans such as loans to depository institutions and real estate loans.
Increasing participation during the 1980s in real estate lending has not
gone unnoticed. Foreign banks have been able to obtain some corre-
spondent banking business because small banks consider them lesser threats
for stealing customers than large domestic banks. The foreign banks have
engaged in off-balance sheet activities such as support arrangements for
commercial paper. They have had some success in the mergers and ac-
quisitions advisory business, mostly with companies from their home
country. On the deposit side, they have appeared to concentrate on
nontransactions deposits. The only way to assess the competitive impact
of foreign banks in these areas is to examine empirical data. The next
section includes some necessary analysis and suggests other types of
empirical examinations to ascertain competitive impact.
The International Banking Act of 1978 was intended to equalize the
treatment of foreign banks and domestic banks in the United States;
however, because of varying state laws, foreign banks have differing
opportunities in different states. "Many states employ reciprocity provi-
sions, asset maintenance or deposit requirements, and geographic re-
strictions to influence the activities of foreign banks within their borders.
Tax laws may also have some influence on foreign banks' decisions regarding
where to locate and what type of office to establish. Few states pursue an
open policy that could be considered equivalent to a national treatment
approach.,,20 As can be seen in table 5-7, only a limited number of states
have foreign banking offices, and the activity is heavily concentrated in
the top three states. This uneven distribution is due largely to the moti-
vation of foreign banks to concentrate in the types of business done in the
largest cities, but it is probably also somewhat affected by state laws

prohibiting some or all types of foreign banking organizational forms.
These restrictions may affect the ability of U.S. banks to expand in Europe
in the future as will be discussed in the last section of the paper.
Balance Sheet Differences Between Foreign and

Domestic Banks
Several studies in the past have compared financial ratios of foreign banks
to domestic banks. These studies provide some information on the activities
engaged in by foreign banks. In the earliest study, Terrell and Key (1977)
compare a number of balance sheet items for three dates (November
1972, November 1974, and May 1977) between U.S. offices of foreign
banks and the weekly reporting banks. They also divide foreign banks by
organizational form, parent country (Japan, Canada, Europe, and rest of
the world) and state (New York, California, and Illinois). They note a
growing importance of C&I loans for foreign banks and find that some
offices offer a wide range of both wholesale and retail services.
A report by the General Accounting Office (1979) examines foreign
bank growth up to 1979. The study finds that foreign banks have become
more competitive with domestic banks in C&I loans from 1972 to 1979.
Trends in the growth of several balance sheet items are presented.
A third study by Houpt (1980) examines the characteristics and per-
formance of U.S. banks acquired by foreign banks and U.s. banks estab-
lished de novo by foreign banks. Note that agencies and branches are not
treated in the study. Using a paired-bank approach he finds that foreign-
owned banks held considerably less state and municipal debt, depended
more on purchased funds, and were less profitable. Houpt qualifies his
results because the number of banks was small and most of the acquisi-
tions occurred since 1975.
In order to assess the competitive impact of foreign banks in the United
States, it is first necessary to identify the activities in which foreign banks
have participated. This article does this directly by comparing financial
ratios from the most recent available balance sheets of both foreign- and
domestic-owned banks participating in the American market. From tables
5-2 and 5-3 it has already been indicated that foreign banks do playa
significant role in the American banking market. Comparisons of finan-
cial ratios by type of financial institution indicate in which activities the
foreign banks could be expected to have the greatest relative competitive
import.
Sixteen financial ratios are chosen for analysis and are presented in
tables 5-8 through 5-15. The first thirteen measures represent particular
assets divided by total assets, while the last three represent individual
liabilities divided by total assets. Data in tables 5-8 through 5-14 are from
the December 1989 Call Reports, while table 5-15 contains data from
December 1980 for foreign banks in order to ascertain changes in foreign
bank activity over the last nine years. Changes in balance sheet defini-
tions and data presentation make it difficult to obtain comparable data
for early periods. Note that since the agency and branch data were obtained
from a different data file than were the full commercial banks, some of
the items on the balance sheet differed and judgment had to be used to
make ratio comparisons meaningful.
The three main types of foreign banking organizations in the United
States follow different regulations and differ in purpose; therefore, the
foreign banks have been separated into three categories in each table.
Since the larger American banks differ from the smaller American banks
in their business activities, the banks with over $1 billion in assets have
been separated from other domestic banks in the analysis. Consequently,
each table (except for table 5-15) contains five columns: agencies; branches;
subsidiaries (commercial banks with greater than 50 percent foreign
ownership); all domestic banks; and large domestic banks (only domestic
banks with greater than $1 billion in assets). All banking organizations
that have reported the relevant data to the bank regulatory agencies are
included in the calculation of the ratios for each type of organization.
The financial ratios for all banks in each category of banking organ-
ization are calculated in two different ways. The first method adds the
particular asset or liability value for all banking organizations in a category
and then divides by the sum of the assets for all organizations. The results
are then multiplied by 100 so as to represent percent of assets. This method
is used in table 5-8 and tables 5-10 through 5-15. The second method
calculates the financial ratio for each institution and then finds the aver-
age of these ratios. Once again, the ratios are multiplied by one hundred
so that the results appear as percentages. Table 5-9 presents these results
for all geographic regions for December 1989 so that the two methods of
calculating ratios can be compared. Ratios using this method have been
calculated for all divisions of banks but because of space constraints these
results are not presented here.
The first method gives greater weight to larger organizations, while the
second method gives equal weight to every organization. This, in effect,
could mean that Citibank is given the same weight as a single office $10
million bank. Examination of the individual ratios reveals many extreme
Table 5-8. All States-Ratio of Sums-December 1989 (Values are percent of total assets).
Agencies Branches Subsidiaries All Domestic Large Domestic
Cash 9.48 28.69 18.54 9.94 11.17
Total Loans 62.32 41.95 57.73 62.51 64.69
US Treas & Govt 0.31 1.63 7.25 11.51 8.56
Other Securities 3.61 4.81 3.38 3.91 3.89
Federal Funds Sold 3.45 3.46 2.65 3.90 2.81
Other Loans 0.27 0.50 2.56 2.66 3.65
Comm & Indus Loans 33.20 22.17 23.99 16.87 20.80
Bankers Acceptances 0.34 0.18 0.12 0.05 0.05
Commercial Paper 0.17 0.16 0.06 0.08 0.07
Real Estate Loans 8.64 4.73 15.91 23.55 21.41
Financial Institutions Loans 14.82 11.36 3.04 1.61 2.22
Foreign Govt Loans 4.21 2.69 1.62 0.82 1.20
Loans for Purchase of
Securities 0.88 0.36 0.73 0.55 0.74
Deposits 21.82 51.90 69.63 77.30 72.76
Federal Funds Purchased 10.83 8.20 5.51 5.69 7.73
Other Liabilities 39.83 19.18 6.98 3.53 4.92
......
00
w
Table 5-9. All States-Average of Ratios-December 1989.
Cash 9.08 19.62 16.56 8.11 9.24
Total Loans 60.97 50.26 52.32 53.07 65.06
US Treas & Govt 1.55 2.01 8.78 21.29 11.51
Other Loans 1.32 1.15 1.28 0.92 2.11
Comm & Indus Loans 34.84 28.07 21.53 13.88 17.55
Securities 0.16 0.17 0.31 0.18 0.43
Deposits 27.62 42.15 71.41 87.53 76.25
Table 5-10. New York-Ratio of Sums-December 1989.
Cash 7.65 29.59 21.99 13.29 13.49
Total Loans 55.56 39.16 51.80 61.24 61.31
US Treas & Govt 0.65 1.86 6.78 4.09 3.73
Other Loans 1.17 0.52 2.15 5.88 6.02
Comm & Indus Loans 20.05 19.59 24.14 22.42 22.76
Securities 0.00 0.42 0.92 1.01 1.00
Deposits 22.86 56.80 65.44 69.95 69.37
.....
00
Vl
......
00
0\
Table 5-11. California-Ratio of Sums-December 1989.

Cash 10.25 11.69 12.52 11.45 11.91
Total Loans 63.10 59.55 72.89 72.70 74.52
US Treas & Govt 0.26 0.24 4.53 5.05 4.30
Other Loans 0.05 0.14 3.35 2.58 3.07
Comm & Indus Loans 34.68 34.56 25.35 20.20 22.15
Securities 1.29 0.00 0.06 0.50 0.61
Deposits 14.45 24.08 77.22 81.90 80.46
Table 5-12. Illinois-Ratio of Sums-December 1989.
Cash 29.09 13.40 10.37 15.20
Total Loans 41.82 60.14 54.68 56.62
US Treas & Govt 1.65 12.61 16.71 7.57
Other Securities 4.79 5.51 4.85 2.67
Federal Funds Sold 3.51 2.76 4.19 2.82
Other Loans 0.50 7.25 2.63 6.89
Comm & Indus Loans 22.03 25.06 18.25 33.51
Bankers Acceptances 0.18 0.04 0.05 0.13
Commercial Paper 0.16 0.04 0.15 0.02
Real Estate Loans 4.64 13.09 19.77 9.77
Financial Institutions Loans 11.49 3.13 0.93 2.23
Foreign Govt Loans 2.67 0.61 0.36 1.05
Securities 0.36 1.95 1.06 3.00
Deposits 51.84 75.85 79.95 64.70
Federal Funds Purchased 8.28 7.07 3.66 9.63
Other Liabilities 19.30 2.43 1.94 5.48
.......
00
-..l
>-'
00
00
Table 5-13. Florida-Ratio of Sums-December 1989.

Cash 16.02 31.26 8.22 9.11
Total Loans 27.26 37.79 64.87 65.35
US Treas & Govt 0.70 6.84 11.76 11.18
Other Securities 1.64 4.00 4.19 4.16
Federal Funds Sold 4.01 3.35 4.88 4.51
Other Loans 0.99 0.58 0.90 1.18
Comm & Indus Loans 10.99 8.51 9.81 11.41
Bankers Acceptances 0.45 0.32 0.001 0.002
Commercial Paper 0.02 0.03 0.06 0.001
Real Estate Loans 2.57 13.35 36.61 34.57
Financial Institutions Loans 8.19 1.84 0.50 0.56
Foreign Govt Loans 4.10 1.80 0.08 0.13
Securities 0.00 0.00 0.14 0.19
Deposits 88.61 88.98 82.95 80.42
Federal Funds Purchased 1.23 0.06 3.26 4.31
Other Liabilities 5.26 0.14 1.12 1.63
Table 5-14. All States Except New York, California, Illinois and Florida Ratio of Sums-December 1989.
Cash 0.28 3.04 7.37 8.82 9.95
Total Loans 96.13 50.57 70.70 61.82 64.64
US Treas & Govt 0.01 0.75 12.62 14.24 11.47
Other Loans 0.19 0.56 0.76 1.85 2.71
Comm & Indus Loans 55.90 31.13 23.95 15.14 19.84
Securities 0.00 0.00 0.11 0.41 0.59
Deposits 0.76 56.03 72.01 78.17 72.55
.....
00
\0
,....
~
Table 5-15. All States-Ratio of Sums-December 1980.

Cash 7.57 16.78 22.86
Total Loans 54.00 66.83 54.46
US Treas & Govt 2.06 1.58
Other Securities 0.35 0.54
Federal Funds Sold 4.32 4.63 2.42
Other Loans 0.35 1.77
Comm & Indus Loans 31.84 30.74 19.71
Bankers Acceptances 4.02 2.66
Commercial Paper 0.19 1.14 0.09
Real Estate Loans 3.37 0.29 12.87
Financial Institution Loans 13.20 26.90 6.92
Foreign Govt Loans 4.62 6.64 2.73
Securities 0.63 0.51 0.76
Deposits 7.44 43.32
Federal Funds Purchased 8.20 5.99 6.81
Other Liabilities 47.57 17.57 1.85
values, particularly for smaller banks. Consequently, the use of the sec-
ond method allows these extreme values to playa larger role. If we view
these ratios as being a sample from a time continuum of different ratios,
then statistical tests of differences between categories can be conducted
using the second method; however, the standard deviations for most of
the ratios are high as compared to the means. Here it was decided to
compare population means rather than to do statistical testing.
Tables 5-8 and 5-9 present data from all states for December 1989.
Tables 5-10 through 5-13 present data from the four most important states
for foreign banks in the United States. Comparisons across states indicate
that foreign banks do not perform identically in different states. Table
5-14 indicates results for all states, excluding the four largest. The only
data presented for 1980 are for all states, but individual state data have
been obtained with similar differences among states for 1980 as in 1989.
The data in table 5-8 of the ratio of sums for all states indicate substantial
portfolio differences for agencies and branches as compared to the three
categories of full commercial banks. In addition, the foreign subsidiaries
also differ from the two categories of domestic banks in a number of
financial ratios.
Commercial and industrial loans is the activity where it has generally
been acknowledged that foreign banks have had the greatest competitive
impact. The results in table 5-8 confirm that all types of foreign banks
have emphasized C&I loans to a greater extent than have their domestic
counterparts. This appears to be especially true for agencies that have
33.20 percent of assets in C&I loans. Branches with 22.17 percent and
subsidiaries with 23.99 percent of assets in C&I loans also exceed the
proportions in C&I loans of all domestic banks of 16.87 percent. As ex-
pected, the larger domestic banks with 20.80 percent in C&I loans con-
centrate on these loans more than the average bank. Agencies and branches,
and to a lesser extent subsidiaries, have placed greater emphasis in banker's
acceptances, commercial paper, loans to financial institutions, and loans
to foreign governments and institutions. There is no clear pattern for
differences between domestic and foreign institutions for loans for the
purchase of securities. Though some observers have indicated that foreign
banks are playing a greater role in real estate loans, the data indicate that
at this time domestic banks are much more involved with these types of
loans. Branches hold substantially more cash and less loans than the other
categories. Subsidiaries, though not as cash-laden as branches, also hold
more cash and less loans than domestic banks, whereas agencies are similar
to domestic banks. Foreign banks, and particularly agencies and branches,
hold substantially lower portfolios of U.S. government securities than
domestic banks. Results for holding other securities and lending in the
federal funds market vary by type of foreign institutions. Agencies and
branches make fewer other loans, many of which are retail oriented, than
domestic banks. Subsidiaries are close to domestic bank levels in this
activity.
There are also some interesting comparative results on the liability side
of the balance sheet. As expected, because of legal restrictions, agencies
obtain far fewer of their funds from deposits and credit balances than
other categories of institutions. Branches are also considerably less de-
pendent on deposits, and subsidiaries slightly less dependent. Agencies
and branches are more active, however, in borrowing in the federal funds
market and in obtaining funds from other liabilities (borrowing in money
markets). Subsidiaries are more active (but to a lesser extent) in money
market borrowing. Note, that agencies and branches obtain much funding
from their parent organizations and that this is not reported in the tables.
Most of the above-mentioned relationships hold when the second
measure of financial ratios is used in table 5-9. However, there are some
important differences due to the heavier weighting of smaller institutions
in the second method. The major differences are highlighted here and the
analysis by state is only reported using the first method. Results by state
using the second method have been obtained but are not presented here.
These results are, nevertheless, very similar to those presented in the
article. The major result with respect to C&I loans is even more pronounced
using this method because the smaller domestic banks with fewer C&I
loans are given more weight. The foreign banks still have less in real
estate loans and more in loans to financial institutions and foreign govern-
ments. The results for cash are not as distinctive for branches using this
methodology. Because smaller domestic banks have fewer loans and more
government securities, in table 5-9 the domestic banks have a lower ratio
of total loans and higher U.S. government securities than in table 5-8. The
differences in ratios of total loans are reduced among categories of banks
using this methodology, but results for U.S. government securities are
similar to before. Foreign banks hold fewer other securities under the
second methodology. Results for other loans lose their distinctive pattern.
On the liability side there are fewer differences in results by using alter-
native methodologies. Note, however, that the ratio for domestic bank
borrowing in the federal funds market is dramatically reduced because
smaller banks do not borrow heavily in this market.
Comparing ratios across the four most important foreign banking states
and the residual states provides some interesting information about the
nature of foreign bank competition. There are distinct differences among
states in the activity emphasis of both domestic and foreign institutions.

The most noteworthy differences are discussed here. It should be noted
that there are no agencies in Illinois nor branches in Florida.
Regarding commercial and industrial loans, there are noticeable differ-
ences among states for domestic banks. All banks and large domestic
banks in New York, California, and Illinois engage in more C&I loans
than the average for all states. The ratio for Florida banks is substantially
below the average for all states and the residual states are slightly below
the average. In New York the foreign agencies and branches hold less in
C&I loans than do their domestic counterparts, but subsidiaries hold slightly
more. This may indicate that the foreign banks in New York have diver-
sified into other activities and may be playing a stronger competitive role
in these other areas than they were previously. In California the foreign
agencies, branches, and subsidiaries are heavily concentrated in C&I loans,
but in Illinois the branches and subsidiaries do more C&I loans than all
domestic banks, but not more than the large banks. In Florida, however,
the agencies and subsidiaries have emulated the domestic banks with low
concentrations in C&I loans. Apparently, there is less effort by foreign
banks to lend to the large corporations in Florida than in the other three
states. Finally, in the residual states the agencies and branches are more
heavily invested in C&I loans than the average.
A second activity with interesting differences among states is real
estate loans. The two rapidly growing states, California and Florida, have
large real estate sectors and their domestic banks participate heavily in
real estate loans. In New York and Illinois the domestic banks are less
involved in real estate. The California foreign banks are also heavily in-
volved in real estate, but the Florida foreign banks are not. The foreign
banks in the residual states also appear to be making extensive efforts in
real estate lending.
Foreign subsidiaries in Florida have substantially higher cash ratios
than any other banks and this may indicate services to customers with
relatively high demand for cash. Correspondingly, the Florida banks also
have fewer loans. Agencies in New York have lent proportionately higher
sums in the federal funds market. Agencies in the residual states have
high loan ratios and extremely low cash ratios.
On the liability side, Florida agencies and subsidiaries appear to greatly
differ from banks in other states. All domestic Florida banks have higher
deposit ratios than the national average, but the Florida foreign banks
obtain even more funds from deposits than the domestic banks; they also
obtain less from other sources. California agencies and branches and the
residual states agencies obtain substantially less funds from deposits. The
California agencies and branches are more heavily involved in money

market borrowing.
We have discussed the major differences among states, but there are a
number of other differences that perhaps deserve consideration, though
they are not discussed here. It is clear, however, that foreign banks be-
have differently in different states and that their competitive impact will
differ by state.
It is also illuminating to observe how the balance sheets of foreign
banks have changed from 1980 to 1989. Comparisons of tables 5-8 and
5-15 reveal some significant changes in the balance sheets of the foreign
bank over the nine-year period. Note that because of data comparability
problems, we have no information on certain financial ratios for subsidi-
aries at this time.
The proportion of C&I loans increased slightly for agencies and sub-
sidiaries but decreased by over 8 percent for branches. Although foreign
banks have grown so rapidly during this period compared to domestic
banks, they have, nevertheless, been able to provide extra competitive
pressure upon domestic banks in the important activity of commercial
and industrial loans. The activity that appears to exhibit the greatest change
is real estate loans. All categories of foreign banks have substantially
increased their proportion of assets in loans backed by property. This
change plus the overall growth of foreign banks explains the observations
mentioned earlier about the increased competitive impact of foreign banks
in real estate loans. The agencies and branches have raised their holdings
of other securities. On the other hand, there has been a large decrease in
holdings of banker's acceptance for agencies and branches and commercial
paper for branches. Loans to financial institutions have decreased for
branches and subsidiaries and loans to foreign governments have decreased
substantially for all categories. On the liabilities side, agencies and branches
have increased their reliance on deposits. The main conclusion that can
be reached from these two tables is that foreign banks have changed their
activity mixes over the nine-year period and this, along with their overall
growth, has increased their competitive impact in certain areas.
The ratio analysis above identifies activities that foreign banks have
pursued and gives some indications about where the competitive impact
of foreign banks can be found. It is beyond the capability of this type of
analysis, and beyond the scope of the article, to be able to measure the
degree of the competitive impact of foreign banks. Later, several ways to
measure this impact are suggested. Note, that the data requirements are
very severe for most of these methods and some may be impossible to do
at this point in time.
Market shares of foreign banks in different markets (local markets or

states) can be calculated over time using either assets, deposits, or C&I
loans. The change in shares can then be related to performance variables.
Domestic banks may become less profitable after foreign bank growth
due to added competitive pressure. As an alternative to examining the
effect on various measures of profitability, one could look directly at the
impact on loan and deposit pricing. There are several serious difficulties
with this type of analysis. It may be difficult to isolate the effects of
foreign bank competition because of the presence of other factors changing
in the economy at the same time. Profitability measures are subject to
distortion because of accounting procedures. Pricing data are difficult to
obtain. In addition, the pricing of loans and deposits involves other factors
besides interest rates, which are difficult to quantify, and may be changing
as the interest rates change.
One could alternatively examine changes in the balance sheet of do-
mestic banks after foreign bank entry and growth. If foreign banks do
have a competitive impact, one might expect domestic banks to emulate
what foreign banks have done. However, in this type of analysis one has
to be careful about other factors which are affecting the behavior of banks.
Finally, the differential effect on domestic banks, due to competitive
pressure by different organizational forms of foreign banks, must be taken
into account. For example, we would expect subsidiaries to have a larger
impact on retail banking and agencies a relatively larger impact on wholesale
banking.
The measurement of the competitive impact of foreign banks is important
to the government in formulating policies to regulate foreign banks. There
is still much work to be done to measure the competitive impact of foreign
banks.
Conclusions and Prospects for the Future
Foreign banks are playing an increasingly important role in the American

financial marketplace. They have apparently affected the behavior of
domestic banks, especially in the area of commercial and industrial loans.
They are providing services in many other areas and have expanded their
efforts in several areas, such as real estate loans. It is difficult to measure
their competitive efforts, and this is not done in the article, but the anec-
dotal and empirical evidence available indicates a definite competitive
impact. The analysis has indicated the balance sheet areas that the different
categories of foreign banks have emphasized and shown differences among
geographic areas. Foreign banks have grown rapidly both in absolute size
and in relation to domestic banks; therefore, an activity showing a de-
crease in its relative importance to foreign banks might still be an activity
in which foreign banks are having an increased competitive impact.
The studies analyzing the factors affecting the growth of foreign banks
in the United States find a number of economic and regulatory factors
that have been important in motivating foreign bank growth. Several of
these variables are likely to change in the future in a direction that will
further encourage the growth of foreign banks. Though foreign banks in
the United States should be playing a greater role in the future, there are
certain factors that could retard this development. Accusations that for-
eign banks have competitive advantages over domestic banks have been
made but, as discussed earlier, there does not appear to be solid evidence
to support these suggestions. Nevertheless, legislation may be proposed
to restrict foreign banks. In addition, proposed legal changes in Europe
may affect American banks in Europe, and this in turn could lead to
changes in the treatment of foreign banks in the United States.
The European Community (EC) is scheduled in 1992 to remove many
commercial barriers between countries, including restrictions on commercial
banks operating in other EC countries. Banks with charters in one EC
country will be allowed to operate in other EC countries and acquisitions
of existing banks by banks in other countries are likely to occur. The
structure of banking markets in Europe should, therefore, change dra-
matically. The resulting structural changes could impact significantly upon
the availability and nature of banking services in these countries. There
could also be important effects upon the ability of banks from non-EC
countries, particularly American and Japanese banks, to operate in these
markets.
Since the Treaty of Rome in 1957, the EC has been moving to eliminate
commercial restrictions among countries. In 1960 the EC adopted the
First Banking Directive that argued for financial integration but retained
national control. In 1988 the EC adopted the Second Banking Directive
that permits any banking organization licensed in any EC country to open
offices in other EC countries and to engage in certain acceptable lines of
business. The directive is intended to be implemented before January 1,
1993.
The directive deals with the treatment of non-EC banking organiza-
tions. Subsidiaries of non-EC banking organizations chartered in an EC
country would be considered to be EC companies regardless of ownership
and, thus, would be allowed to branch throughout the Community without
obtaining additional licenses. Future entry, though, would be subject to
reciprocity conditions. Approval would depend upon whether the other

country granted equal opportunities to Ee banks in that country as granted
to domestic banks, that is, whether genuine national treatment was granted
to Ee banks. This might provide a barrier to foreign entry or might serve
as a bargaining tool to open other markets to Ee banks.
The major banks in Europe already have nationwide banking organi-
zations within their home countries. However, there are still restrictions
on interstate banking in the United States, though the situation has been
greatly liberalized in the last decade. The Second Banking Directive
explicitly provides for reciprocal treatment of banks from non-Ee coun-
tries. The United States faced this problem in 1978 and opted for a policy
of treating foreign banks on a national treatment basis; that is, all foreign
banks were accorded equal treatment with domestic banks. However, this
still produced some inequities from the European point of view. Ameri-
can banks were not restricted within individual European countries geo-
graphically in most cases, but European banks were restricted to full
banking operations in only one state, except for those already present by
1978. If Europe is treated as one entity-then the treatment appears to be
reciprocal, but if each country is treated separately-it is not. The Ee
must assess whether the American treatment is reciprocal.
The interaction of legal changes could affect the ability of foreign banks
to compete in the United States. Retaliatory measures against European
restrictions on American banks could reduce the competitive importance
of foreign banks. On the other hand, though, greater liberalization of the
treatment of foreign banks could enhance their ability to provide mean-
ingful competition to American banks.
Notes
1. Houpt (1988, 7, 11) and additional updated tables provided to the author by James
Houpt.
2. Ibid., (25) and updated material supplied by James Houpt.
3. American Banker (February 27,1990, 18A). See Table 3, footnote (c), in this article.
4. Cho, Krishnan, and Nigh (1987, 60).
5. Ibid., (60).
6. Ibid., (64).
7. Goldberg and Saunders (1981a, 32).
8. Goldberg and Saunders (1981b, 372).
9. Grosse and Goldberg (1991,1109-110).
10. Terrell, Dohner, and Lowrey (1990, 48-49).
11. This is argued by Nadler (1989, 5).
12. Ibid.
13. Nadler (1989) provides an extensive discussion of how this practice works and the
implications for domestic banks.
14. See Baer and Pavel (1987).
15. See Terrell, Dohner, and Lowrey (1989).
16. Baer (1990, 25).
17. Jedlicka and Tobin (1988, 89).
18. Baer (1990, 25).
19. Cho, Krishnan, and Nigh (1987, 69).
20. Hultman (1987, 348).
References
Baer, Herbert L. 1990. "Foreign Competition in U.S. Banking Markets," Federal
Reserve Bank of Chicago. Economics Perspectives 14 (3) (May/June): 22-
29.
Baer, Herbert L., and Christine A. Pavel. 1988. "Does Regulation Drive
Innovation?" Federal Reserve Bank of Chicago. Economics Perspectives 12, 2
(March/April): 3-15.
Cho, Kang Rae, Suresh Krishnan, and Douglas Nigh. 1987. "The State of Foreign
Banking Presence in the United States." International Journal of Bank Marketing
5,2: 59-75.
General Accounting Office. 1979. "Considerable Increase in Foreign Banking in
the United States Since 1972." Report by the Comptroller General of the
United States. (August).
Goldberg, Lawrence G., and Anthony Saunders. 1981a. "The Determinants of
Foreign Banking Activity in the United States." Journal of Banking and Finance
5, 1 (March): 17-32.
- - . 1981b. "The Growth of Organizational Forms of Foreign Banks in the United
States." Journal of Money, Credit and Banking 13, 3 (August): 365-374.
Goldberg, Lawrence G. and Robert Grosse. 1990. "Distribution by State of Foreign
Bank Activity in the United States." (March) Mimeographed.
Grosse, Robert, and Lawrence G. Goldberg. 1991. "Foreign Bank Activity in the
United States: An Analysis by Country of Origin." Journal of Banking and
Finance 15, 6 (December): 1093-112.
Houpt, James V. 1980. "Foreign Ownership and the Performance of U.S.
Banks." Staff Study 109. Board of Governors of the Federal Reserve System.
Washington, D.C. (July).
--.1988. "International Trends for United States Banks and Banking Markets."
Staff Study 156. Board of Governors of the Federal Reserve System. Washington,
D.C. (May).
Hultman, Charles W. 1987. "The Foreign Banking Presence: Some Cost-Benefit
Factors." Banking Law Journal 104, 4 (July/August): 339-49.
Hultman, Charles W., and Randolph McGee. 1989. "Factors Affecting the Foreign
Banking Presence in the United States," Journal of Banking and Finance, 13,
3, (July): 383-96.
Jedlicka, John, and Mary Tobin. 1988. "Foreigners Ferocious Financial Threat."
Euromoney (November): 89-93.
Nadler, Paul. 1989. "Balances and Buggy Whips in Loan Pricing." Journal of
Commercial Bank Lending 72, 6 (February): 4-9.
Terrell, Henry S., Robert S. Dohner, and Barbara R. Lowrey. 1989. "The United
States and U.K. Activities of Japanese Banks: 1980-1988." Board of Governors
of the Federal Reserve System. International Finance Discussion Papers 361
(September).
- - . 1990. "The Activities of Japanese Banks in the United Kingdom and in the
United States, 1980-1988." Federal Reserve Bulletin 76(2) (February): 39-50.
Terrell, Henry S., and Sydney J. Key. 1977. "The United States Activities of
Foreign Banks: An Analytical Survey." Board of Governors of the Federal
Reserve System. International Finance Discussion Papers, No. 113. (November).
Zimmerman, Gary C. 1989. "The Growing Presence of Japanese Banks in
California." Federal Reserve Bank of San Francisco. Economic Review
(Summer): 3-17.
COMMENTARY
Gary C. Zimmerman
Goldberg's article is a timely one. It deals with important issues facing the
U.S. economy: foreign investment, foreign bank control, foreign bank
advantages, foreign bank funding-and examines them in the context
of the rapid growth of foreign banks in the u.s. market. He uses these
issues to analyze the competitiveness of domestic banks and foreign banks,
especially the major Japanese-owned banks operating in the United States.
Clearly, this is a subject that is of interest to many of us. Foreign bank
presence is frequently discussed in the press. And fortunately, we have
some measurable ways of examining their presence using asset size,
commercial lending or activity in the "guarantee" markets. These measures
highlight foreign bank penetration of the U.S. banking markets. For
example, in California $1 out of every $3 in banks' outstanding business
loans are held by Japanese-owned banking institutions. In the guarantee
market, foreign bank presence is even greater. Again, in California, foreign-
owned banks issue about 60 percent of all standby letters of credit issued
by banking institutions in the state.
201
Review
Goldberg's article provides us with a good background on the growth of

foreign banks in the United States, on their status and history, and on
their varied organizational forms. This latter information is helpful to our
understanding of the scope of their operations, their competitive impacts,
and their possible advantages. In addition, the article provides informa-
tion on the foreign bank presence by states, which is useful, because at
this time foreign-owned banks tend to be most active in several major
banking markets and their organizational form and business emphasis can
vary dramatically from state to state.
The article then moves into a review of the literature and empirical
studies that attempt to explain the growth of foreign banks in the United
States. The literature, to which Goldberg has been a significant contributor,
elaborates on a number of factors influencing foreign bank activities.
Featured prominently on that list are foreign trade activity, direct foreign
investment, domestic economic conditions, regulatory differences, inter-
est rate differentials across countries, exchange rates, and price/earnings
ratios across countries.
There is strong evidence that foreign banks follow trade and invest-
ment activities, and it is this economic process that I believe should be
emphasized. Japanese banks in particular have followed Japanese multi-
national firms to the U.S. market. This is a process that is familiar. A
generation or more ago many major U.S. banks expanded their overseas
activities and portfolios as their major corporate customers invested abroad.
Clearly, foreign-owned banks are likely to have a competitive advantage
in dealing with these "home country" customers; they have established
banking relationships, they understand their customs, and they speak their
language. It is not surprising that overseas, U.S. corporations often bank
with U.S.-owned banks, or that Japanese multinational firms' overseas
operations bank with Japanese-owned banks.
Rapid Growth
The article cites some dramatic figures on foreign bank growth in the
United States. However, the figures are somewhat misleading because
virtually all of the foreign bank expansion in the United States has been
by Japanese-owned banks. Thus, the article would be more useful if it
reported bank growth for three categories of banks, based on country of
ownership: United States-owned, Japanese-owned, and other foreign-owned
(excluding the Japanese).
Growth rates for the various groups illustrate this point. Over the
period from year-end 1985 to mid-year 1990, non-Japanese foreign-owned
banks, branches and agencies as a group have grown at about the same
average rate as domestic banks, about 5 percent per year. In contrast,
Japanese-owned banks, branches and agencies have expanded at close
to a 25 percent annual rate. All three types of Japanese-owned banking
institutions grew rapidly during the period. Clearly, the growth pattern
indicates that the focus should be on the exceptional growth of J apanese-
owned banking institutions in the United States, rather than the broad
grouping of all foreign banks.
After describing the rapid foreign bank growth and the reasons behind
it, Goldberg presents a descriptive section on the evidence of competitive
effects and on the allegations of "unfair" competition from foreign banks.
In this section he makes his point using developments in the commercial
lending market, which is a good way to handle the analysis. Commercial
lending is an important area for foreign-owned banks, and it is a market
in which they playa major role.
One criticism of this section is that it again needs to focus on the
Japanese-owned banking institutions. They account for the exceptional
growth of foreign banks' commercial lending in the United States.
Commercial loans have grown at about a 1 percent annual rate since 1985
for both domestic banks and non-Japanese foreign-owned banks. Yet,
business loans at Japanese-owned banks have grown at over a 30 percent
annual rate over this period. Moreover, they captured nearly a 9 percent
increase in market share over the same period (from 6.4 percent to 15.1
percent).
Competitiveness
The descriptive section of the article provides a series of brief explana-

tions and stylized facts concerning the competitiveness issue. While it hits
on important issues (such as differential capital requirements across
countries, less restrictive regulation at home, potential overseas funding
sources, lower cost of funds, less paperwork and documentation, and
willingness to offer lower loan rates and accept smaller margins), it does
not provide the reader with a clear sense of what the competitive advan-
tages might be or how significant they are. Again, by focusing on the
rapid growth of Japanese-owned banks we may gain more insight into the
competitiveness issue.
Moreover, there may be some stronger evidence that there are no

significant advantages, especially with respect to the funds availability
question ap.d the allegation that foreign-owned banks have a lower cost
of funds.
Henry Terrell (1990) in testimony before Congress provided evidence
that the growth of Japanese banks in the United States was NOT funded
by raising low cost funds in Japan and lending them through their affiliates
in the United States. In fact, Terrell suggested that net sources of funds data
indicate that just the opposite pattern took place. Terrell stated:
Until recently, restrictions on the ability of Japanese banks to offer their domestic
Japanese customers market-determined interest rates on deposits appear to
have had the effect of inducing Japanese banks in the aggregate to shift some
domestically oriented business to their U.S. offices because the low regulated
interest rates in Japan caused funding difficulties and Japanese banks actually
relied on their overseas branches for funds for use at their domestic offices.
This fact suggests that Japanese banks were not on balance borrowing low-cost
funds in Japan and relending them in the United States at low rates.
Thus, the situation facing Japanese banks in their home market ap-
pears to be much like the situation U.S. banks faced when they were
constrained by Regulation Q interest rate ceilings on deposits. For
example, before deregulation in the United States, banks substituted
wholesale borrowings (increasing their reliance on large certificates of
deposit, other money market borrowings, and borrowing from their overseas
affiliates) to make up for retail deposit shortfalls caused by below-market
rates on retail accounts. Faced with a similar constraint on retail deposits,
Terrell indicates that Japanese banks have responded by booking and
funding some loans in the United States and by increasing their borrowing
in offshore markets without interest rate controls.
Terrell also shows that agencies and branches of Japanese banks oper-
ating in the United States are primarily funded by interbank borrowing at
open-market rates from U.S. banks and from affiliates in London and
other offshore markets. Thus, he presents evidence that in the aggregate,
over the last five years Japanese banks overseas have been net lenders to
Japan.
Terrell also indicates that the amount of borrowing to fund their do-
mestic activities has fallen as the share of Japanese deposits subject to
liberalized interest rates has risen. Of course, this trend could have im-
portant implications for the continued growth of Japanese-owned banks
in the United States.
California
In my own work of studying the role of Japanese-owned banks in Cali-

fornia, I also looked at the level of gross borrowing by foreign-owned
banks from their overseas parents. I found that Japanese-owned subsidi-
ary banks generally had very small gross borrowings from their parents,
on average less than 5 percent of their assets were funded by borrowings
from parents. Thus, even if they were being funded by the parent organi-
zations, and the parents were willing to provide funds at below-market
rates, it would still represent only a small share of their total funds.
The next question is whether Japanese-owned bank subsidiaries actually
reported a lower cost of funds than domestic banks. In this case it was
possible to examine the cost of funds for interest-bearing deposits and for
other borrowings (which would include borrowings from parents). The
results showed no statistical evidence from the cross-sectional regressions
that Japanese-owned banks had a lower cost of funds for either total
interest-bearing deposits or for other borrowings. Thus, the California
evidence does not support either the assertion that Japanese-owned banks
are funded by their overseas parents and/or that they have a significant
cost of funds advantage over domestic banks.
Incorporating this work with the research by Terrell would provide
empirical evidence that Japanese-owned banks likely do not have a signifi-
cant advantage over domestic banks in terms of access to below market-
rate funds.
Financial Ratios
Goldberg's section on balance sheet differences moves on to a compari-

son of financial ratios for foreign banks and domestic banks. I found this
section especially useful from a regional perspective. There are some major
differences in the activity and organization of foreign-owned banking
institutions, and this section highlighted those differences, although the
sheer volume of information and tables may be a bit overwhelming.
The aggregate financial ratios presented in the tables do provide some
evidence of differences between foreign-owned banks and domestic banks;
however, it is hard to judge their importance. For example, there are
some sizeable differences between the ratios of various assets and liabilities
to total assets, but we have no idea whether differences between foreign
and domestic banks are statistically significant or not.
A stronger case could be made by examining empirical evidence from

a cross-section of banks, while controlling for ownership by groups (do-
mestic, Japanese, or other foreign). In my own research on the foreign
bank presence in California, I ran regressions on cross-sections of Cali-
fornia banks, controlling for differences in bank size, branching structure
and other bank specific factors, including ownership status. The results
suggested that as a group there were some statistically significant differences
among Japanese-owned banks, other foreign-owned banks, and domestic
institutions.
Japanese-owned banks had higher ratios of commercial and industrial
loans to assets, commercial real estate lending to assets, and total loans
to assets (reflecting their strong commercial presence), and those differ-
ences were statistically significant. In other areas (total real estate lending
and consumer lending) there were differences of several percentage points
in the aggregate loan-to-asset ratios; however, at the individual bank level,
those differences were not statistically significant.
Japanese-owned banks also reported more reliance on wholesale funding,
and these differences also were significant. On average, Japanese-owned
banks had lower ratios of both total deposits and retail deposits to assets.
Conversely, they had much higher ratios for both large time deposits and
money market borrowings to assets.
Thus, the California data provide stronger evidence that there are some
important portfolio differences between foreign-owned banks (especially
Japanese) and domestic banks, and suggests another empirical avenue for
examining the financial ratio data.
In spite of the statistically significant differences, it is important to note
that Japanese-owned banks and other foreign banks still are primarily
funded by deposits raised in domestic markets. Moreover, in California,
most foreign-owned banks have a large share of retail deposits, and in
fact, they tend to behave and hold portfolios that resemble other retail
banks in the state. As well as providing retail banking services at the
branch level, Japanese-owned banks in California also have an important
presence in the middle market lending area, like other banks in the state.
Still, there are some statistically significant differences that allow us to
make a more forceful statement about the portfolio differences of foreign-
owned banks and their effect on competition. I'd suggest that as an area
for additional research.
Further research is necessary if we are to get beyond the two schools
of thought expounded by many California bankers. One school suggests
that foreign-owned banks, especially the Japanese-owned institutions, are
important competitors for domestic banks. The other holds the view that
foreign-owned banks are just another group of competitors among many,

including other banks, thrifts, nonbank financial intermediaries and out-
of-state banks. One way to measure whether foreign banks are important
competitors, and if their presence is becoming more important, is to examine
industry market shares.
Market Share
Herb Baer of the Federal Reserve Bank of Chicago recently made an

interesting observation on the foreign bank presence in the United States.
Writing in the May/June 1990 edition of Economic Perspectives, he noted
that after leather goods, wholesale banking has the next highest foreign-
owned market share of major U.S. industries. Goldberg suggests that we
might look at market share data to help evaluate the competitive impact
of foreign-owned banks. I followed up on this suggestion, and include
here some of my observations.
Nationally, foreign banks' market share has risen sharply. Between
1985 and 1990, domestic banks lost 3.7 percent and 7.5 percent of their
market share as measured by assets and commercial loans, respectively.
Other foreign-owned banks lost market share in both categories as well.
In contrast, Japanese-owned banks added 4.4 percent to their market
share for assets and 9.1 percent of commercial loans. Clearly, in the area
of market share it is the Japanese-owned institutions, rather than foreign
banks in general, that have been building market share.
California appears to be the exception to the national trend. Over the
same period, there was a rapid growth of Japanese-owned banks' market
share. However, it was offset by reductions in other foreign-owned share,
so that total foreign ownership generally held in the 31 percent range for
assets.
Goldberg also suggested that we may wish to look at profitability to
see if profit levels varied with foreign ownership. In this case I did a quick
examination of industry profits by the three ownership groups (domestic,
Japanese-owned, and other foreign-owned) for several geographic regions
(the United States, New York, California, and Illinois). Nationally, over
the last nine years average return on assets (ROA) for foreign banks was
only about half that for domestic institutions. Neither Japanese-owned
banks nor other foreign-owned banks ever report a higher ROA than
domestic banks.
Of course, lower profits also would be consistent with the suggestion
that foreign banks are using "low margins" to build market share at the
expense of current profits, especially if there was evidence that foreign

banks offered lower rates on "similar" loans.
It may be possible to examine this competitive pricing issue, although
it clearly is beyond the scope of this commentary. Using the Federal
Reserve's individual bank data on commercial and agricultural loan rates,
one could test to see if significant differences in interest rates exist across
individual banks, depending upon ownership. While this research may
have a few unavoidable problems, like a relatively limited sample of foreign-
owned banks in the reporting panel, it may still be possible to test whether
foreign banks price loans differently than domestic banks. There may be
an interesting story here.
Summary
Where does all this take us? Goldberg has answered the questions he
posed earlier in his article: What factors affect foreign bank growth? In what
areas are foreign banks competitive? Do foreign banks have unfair com-
petitive advantages? Are there competitive effects? Thus, he takes us to
what I consider to be the "big" question: What will the future bring for
foreign-owned banks, that is, will their expansion continue or not?
His article provides a good perspective on the inroads of foreign banks,
especially the Japanese-owned banks, and on the importance they now
playas competitors in the U.S. banking system. I would suggest extending
the research and the outlook to include a number of changes currently
going on in the world, in Europe and in Japan, that may impact foreign
bank activity in the United States.
Again, Goldberg's article points us in the right direction, even though
it would benefit from focusing more on the Japanese-banks and whether
they can sustain their rapid growth trend of the past decade. That is
important, since Japanese banks currently hold over 10 percent of the
U.S. bank market share (over 25 percent in California), and yet their
market share has not increased significantly since 1988. Many of the fac-
tors noted in his article and my comments, as well as trends in the Japanese
banking system: from liberalization of interest rates in Japan reducing
pressure to shift funding to the United States and to borrow from the
United States; to uniform international capital standards for banks; to the
significant decline in the Japanese stock market and its reduction of
Japanese banks' "hidden reserves" and capital positions; to banks in Japan
losing their traditional customers to the capital markets (much as it has
happened in the United States); to a new interest in consumer banking in

Japan-all suggest that the growth of Japanese-owned banks in the United
States will slow down. A slowdown, or even a pause in the dramatic
growth of Japanese banks in the United States, will clearly slow the pace
of foreign bank growth in the U.S. market.
6 THE COMPETITIVE IMPACT OF
FOREIGN UNDERWRITERS IN
THE UNITED STATES
Robert Nachtmann
Frederick J. Phillips-Patrick
Introduction
Global integration of financial markets and the internationalization of the

modern corporation are indicators of the openness of the world economy.
Economists argue that an open world economy increases economic effi-
ciency by enabling resources to flow to their most productive uses. In-
vestment banking firms contribute to the enhanced efficiency of a global
economy by intermediating the process of capital formation, servicing the
rebalancing of investor portfolios, facilitating the transfer of corporate
control, and enabling risk freturn decomposition through financial product
innovation. The investment bank not only contributes to the process of
international integration. It is also affected by it.
This article investigates one business segment of the investment bank-
ing firm-the underwriting of corporate securities. It identifies those firms
that service global underwriting activity and attempts to evaluate the impact
of foreign underwriters on the domestic underwriting industry in the United
States. The article is organized as follows. The first section discusses the
major developments affecting the domestic underwriting business in
the 1980s. In the second section, we define and document the extent of
211
the market and identify underwriters for worldwide corporate offerings.

This section also provides evidence concerning performance and pricing
for the domestic and foreign segments of the underwriting industry.1 In
the third section, we examine underwriting performance as a function of
whether the underwriting firm is domestic or foreign. The fourth section
measures the impact of foreign entry and acquisitions of domestic rivals
on the equity values of publicly traded domestic underwriters. Finally, in
section five, we summarize our results and discuss directions for future
research.
Significant Developments in the 1980s
Domestic broker/dealers and investment bankers witnessed dramatic

changes in both their business and their business environment during the
turbulent 1980s. New products were introduced that now represent a sig-
nificant portion of the market. Shelf registration and Rule 144a changed
the underwriting environment. Eight years of economic expansion, accom-
panied with the dramatic growth in corporate consolidations and the
substitution of debt for equity in corporate balance sheets, produced an
unprecedented growth in new issues being brought to the market. This
domestic growth has been matched and exceeded in some foreign capital
markets. But as domestic firms have expanded overseas, so have foreign
firms entered the domestic market. Thus, while opportunities for new
business have expanded, so has the competition for such business. This
section briefly reviews some of the significant developments in the under-
writing business in the 1980s.
New Product Development
Investment banks created or developed several new financial products

during the 19808 that have grown exponentially. Among them are mortgage-
backed and asset-collateralized securities, below investment-grade bonds
(junk bonds), and interest and currency swaps. For the new issue market,
junk bonds and mortgaged-backed securities have played an especially
large role. Both of these developments appear to be a way for assets
traditionally held by banks to become securitized.
In 1980, the new issue volume in mortgage-backed securities was $0.5
THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 213
billion; in 1989, it was $115.5 billion, representing more than a third of

the value of all new issues in 1989.2 Since 1985, growth in other types of
collateralized securities has been equally explosive, increasing from $1.23
billion in 1985 to $23.3 billion in 1989.
Below investment-grade bonds have also experienced dramatic growth
during the early and middle parts of the decade, only to fall off over the
last two years. In 1980, the annual new issue volume of junk bonds was
about $1 billion. It hit a high in 1986 at about $32 billion in new offerings.
Last year, about $25 billion, a substantial portion of which was just one
issue, were offered.
The impact of these two products can be seen in the new issue market
by their effect on the proportion of the equity offerings as a percent of
all offerings. From 1980 through 1982, equity offerings averaged about
$30 billion, well over 20 percent of the new issue market. In 1983, equity
offerings surged, with growth in IPOs far outdistancing secondary offer-
ings. The equity underwriting market collapsed in 1984, off by more than
two-thirds from the previous year. In 1986 and 1987, the new issue market
rebounded, about 18 percent of the total new issue market. After the
market crash of 1987, equity offerings dropped significantly; for 1989, equity
offerings represented only about 7.3 percent of the new issue market, far
below the average of the early years.
Changes in the Business Environment
The financial innovations and the rapid domestic growth of the 1980s
expanded domestic markets. Similarly, expansion of business activity abroad
brought additional demands for financial services, an incentive for invest-
ment banks to expand beyond their own domestic horizons. Financial
engineering and new product development required both specialization
and market size to provide an initial market and the liquidity necessary
to launch a new product. These competing demands appear to have led
to substantial growth among a few of the largest firms. For example,
according to Investment Dealer's Digest, in 1980,80 percent of new issue
debt was syndicated; in 1989, only 20 percent was syndicated.
Likewise, changes in the regulatory environment have altered the nature
of the underwriting business. One of the largest changes was the adoption
of Rule 415 in 1982. Rule 415 allows firms to register their offerings with
the Securities and Exchange Commission before they offer them to the
market. The effect of the rule is to permit firms more flexibility in timing
the market for their new issues. But once the decision has been made to
bring the issue to market, underwriters are under pressure to do so expedi-
tiously. One effect of Rule 415 has been to reduce the role of regional
underwriters in the new issue market, perhaps because they lack sufficient
size to bring a full issue to market promptly.
In April 1990, the SEC adopted rule 144a, which permits the trading
of unregistered securities among "sophisticated" investors, that is, those
with more than $100 million in securities under management. One an-
ticipated result of Rule 144a is the increased trading of foreign securities
within the United States. Likewise, the liquidity provided by the new
trading opportunity is likely to make private placements a more attractive
alternative to normal underwritten issues and again place emphasis on
market size as a driving concern.
Foreign Competition
In 1978 Credit Suisse first formed a domestic investment banking venture

with First Boston. This was not the beginning of foreign investment in
domestic underwriters, but it marked the beginning of rapid change and
consolidation in the domestic underwriting market. Throughout the decade
of the 1980s there were about 550 completed combinations involving
domestic dealer/brokers, according to IDD Information Services, repre-
senting over $36 billion in reported value. Of these 550 deals, some ninety-
four, or about 17 percent, involved foreign acquire. The highest percentage
concentration of foreign acquisitions took place during the period 1985 to
1987, although the highest number of acquisitions has occurred more
recently, with twenty-eight of the total ninety-four deals occurring in 1989
alone.
Mergers and acquisitions were not unidirectional. Recently, U.S. firms
have started to acquire foreign broker/dealers. For example, over the
period 1985 to 1989, domestic firms completed forty-seven deals involving
foreign target broker/dealers.
Whether foreign entry into the U.S. market signals increased com-
petition for other broker/dealers or increased expansion opportunities for
both domestic and foreign business is an empirical issue. The effect of
foreign acquisitions on the market values of other domestic broker/dealers
is considered in section four. In the next section, we describe the depth
and the extent of the market.
The Market
Market Identification
The target market for underwriting services is defined for the purposes of
this analysis to be corporate security distributions managed by broker/
dealers that are registered to do business in the United States, acquired
by them through negotiated deliberations, and publicly offered on a firm-
commitment basis.3 A brief explanation of these defining characteristics
follows.
First, we are only concerned with those financial contracts that identify
the transfer of capital to corporations. We do not therefore address
government or municipality financing for their own account. We do,
however, include the taxable issues of corporate agencies that are often
described as quasi-governmental organizations.
Corporations source external capital through the sale of financial
contracts. The sale may be conducted as either a public or private
transaction. Representative private transactions are the private placement
of securities and loan contracts. Private transactions are not considered.
Public offerings are brought to the public market either directly by the
issuer or indirectly through the services of broker/dealers. Examples of
direct offerings such as shareholder dividend reinvestment or stock purchase
plans and the direct sale of rights or warrants typify the direct issue from
the firm to the investor. Indirect public offerings engage the services of
broker/dealers who bring the issue to market at a publicly announced
offer price. Indirect public offerings form the base set of the target market
for underwriting services.
Indirect offerings employ the services of broker/dealers on either a
"best efforts" or a "firm-commitment" basis. Firm-commitment under-
writing involves the direct purchase of the offer from the issuer by one or
more broker/dealers who agree to resell the issue at a specified public
offering price. Firm-commitment underwriting, therefore, incorporates a
distinct element of insurability whose value would be incorporated in the
discounted price to the issuer from the public offer price. Best efforts
underwriting does not entail the offering's purchase by the broker/dealer.
Instead, the broker/dealer simple agrees to extend its "best effort" to the
task of selling the offered securities at the public offer price. Only firm-
commitment offerings are analyzed.
Issuers identify firm-commitment underwriters either through a nego-
tiated or a competitive bid process. 4 Competitive deals result from sealed
bid auctions in which the syndicate is completely specified and the firm
hires the syndicate with the highest bid. The typical firm commitment
offering is serviced by an underwriting syndicate under the control of a
lead manager.s Underwriting services that are provided through negoti-
ated dealings between the corporation and the broker/dealer syndicate
offer the managing underwriter degrees of freedom in distribution that
may not obtain through the competitive bid process. The negotiated deal
is the primary focus of this paper.6
Extent of the Market
The market for worldwide corporate offerings grew at an annual rate

of approximately 20 percent during the 1980s. This growth was fueled
primarily by the rate of growth in corporate debt issues. Overall, the
underwriting market accounted for the distribution of $2.9 trillion worth
of corporate securities over the period. The primary issuing group was
financial firms-$1.4 trillion or 48 percent of all corporate issues. Industrial
firms were the second largest issuing group--$0.9 trillion. Corporate agen-
cies and utilities accounted for approximately $0.4 and $0.3 trillion, re-
spectively.
Table 6-1 lays out the depth of the market over time and across secur-
ity types. All data were collected from IDD Information Services' Do-
mestic and International New Issue Database. The valuation of issues
entering into the volume totals is consistent with the IDD's yearly rankings
for underwriting business done by investment banking firms. Common
and preferred issues are valued at share price times the number of shares
distributed. Debt issues are valued at net proceeds to underwriter com-
puted as principal times price.
We classify separately the following types of offerings: common stock,
domestic debt, preferred stock, convertible debt, convertible preferred,
and international debt. Common, preferred, and convertible issues combine
both domestic and international offerings. Since information on interna-
tional debt offerings is not available until 1982, debt offerings are reported
separately. Totals (A) and (B) in table 6-1 are provided to account for the
staging of international debt.
Public offerings have increased not only in the aggregate but also by
security type. Only isolated incidents can be identified where the volume
of any issue in a given year was not higher than its previous year's level.
The most dramatic period of growth was 1983 through 1985. Issue volume
increased almost threefold (growth factor 2.81). If the cost to issuers in
Table 6-1. Trends in Public Offerings of Underwritten Corporate Securities 1 (1980-1989). U.S.$ Billions.
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
(1) Common Stock· 13.47 14.44 14.18 37.40 14.20 29.33 63.37 76.39 42.90 38.08
(2) Domestic Debt b 37.09 35.87 40.74 46.05 65.32 97.75 219.76 210.48 234.54 272.83
(3) Preferred Stocke 1.90 1.18 4.52 5.14 3.12 6.48 9.89 7.64 7.47 7.61
(4) Convertible Debt d 4.35 4.65 6.00 11.61 12.01 15.01 19.16 27.37 15.91 20.79
(5) Convertible Prefd 1.29 0.46 0.51 3.15 0.82 2.38 4.82 4.30 1.07 1.58
(A) Total (1-5) 58.10 56.60 65.95 103.35 95.47 150.95 317.00 326.18 301.89 340.89
(6) Int'l Debt" n.a. n.a. 62.31 62.47 90.41 151.13 205.30 155.43 207.93 231.14
(B) Total (1-6) 58.10 56.60 128.26 165.83 185.89 302.13 522.38 481.68 509.91 571.52
• Common stock offerings both domestic and foreign excluding debt for equity swaps and rights.
b Domestic, straight debt valued at offer price.
C Preferred stock both domestic and foreign valued at offer price.
d Convertible offerings include both domestic and foreign issues valued at offer price.
e International debt includes Eurobond and non-Yankee foreign bond issues, expressed in U.S. dollar equivalents at the time of
issuance.
I Public underwritten offerings includes firm commitments only and excludes private placements.
the aggregate is related to the volume of business being done in the

underwriting industry, the period beginning in 1985 and proceeding through
1989 should result in a reduction in issuers' cost.
Increasing volume for the industry could drive this result in two ways:
(1) fixed costs for the industry are now spread over a larger volume of
business; (2) perhaps more importantly, to the extent that the increase in
volume represents a more diversified set of issuers, the insurance premium
set by the underwriting should decrease given the larger pool of transactions.
The trend data in table 6-1 suggest that this cost effect is most compelling
in the debt markets.
The Industry
Underwriting Firm Identification. To evaluate the impact of foreign

underwriting firms on their U.S. counterparts, firms were classified into
domestic and foreign classes. The classification procedure was performed
in the following way. First, only registered broker/dealers with the U.S.
Security and Exchange Commission were considered. Second, classifica-
tion as a foreign firm required an equity ownership by foreign citizens or
firms of minimally 25 percent. Third, each firm, domestic or foreign, had
to have been represented in the top twenty-five underwriters of world-
wide corporate offerings for at least one year during the period 1980
to 1989. Top twenty-five underwriter status for worldwide corporate
offerings was identified from various issues of IDD. Identification as a
registered broker/dealer and the foreign/domestic status of the firm was
identified as follows.
The foreign/domestic status of registered broker/dealers is not cur-
rently required by the Securities and Exchange Commission. The New
York Stock Exchange does, however, identify a registered firm to be
foreign-controlled if 25 percent or more of the firm's equity is owned by
citizens or firms of foreign countries. Additionally, the Securities Industry
Association surveys members of the National Association of Securities
Dealers with regards to their ownership structure. These sources were
used for the identification of foreign-controlled registrants in the United
States and resulted in the identification of thirty-seven firms in 1980 and
121 firms in 1989.7 Appendix A lists the underwriting firms used in this
study.
Growth of Registered Broker/Dealers. Table 6-2 provides information

about the growth of registered broker/dealers as well as equity and asset
Table 6-2. Broker/Dealer Registration (1980-1989).
1980 1989 Growth Factor
Number
Total 5,714 9,021 1.58
Foreign 37 121 3.27
Assets ($ Bn)
Total 150.5 649.9 4.31
Foreign 12.4 143.5 11.56
% Foreign 8.2 22.1
Equity ($Bn)
Total 10.0 36.0
Foreign 0.5 4.4
% Foreign 5.0 11.4
Equity/Asset Ratio
Total .066 .055
Foreign .040 .031
information for firms registered in 1980 and 1989. Accounting informa-

tion was taken from annual reports filed by broker/dealers registered with
the Securities and Exchange Commission.8
In relation to the total number of firms in the industry, foreign firms
represent a small percentage (.7 percent in 1980; 1.3 percent in 1989). This
small number of firms, however, accounted for 8.2 percent of the industry's
assets in 1980 and 22.1 percent of the industry's assets in 1989. The assets
of the average firm in the industry in 1989 is $72 million while foreign firm
assets were on average $1.2 billion or seventeen times the industry average.
Equity to assets ratios indicate that the equity position of foreign firms
is approximately 60 percent that of the average firm. Since capital re-
quirements are a function of the riskiness of the business conducted, these
figures would suggest that, while the average foreign firm operates at a
scale commensurate with the larger broker/dealers in the industry, it holds
assets that are less risky than those of the average firm in the industry.
Table 6-2 does not necessarily provide information that is pertinent
solely to the business of underwriting corporate securities. The assets of
broker/dealers support other activities such as retail sales in secondary
markets. It does, however, support the premise that foreign entry into the
market for financial services in the United States is an important element
of the market. Furthermore, the position of the foreign registrant in terms
of firm number and firm size is shown to have increased substantially over
the past decade.
Performance of Industry Segments
This section provides evidence on performance of the underwriters listed

in Appendix A in accordance with their foreign/domestic classification.
The relative positions of the industry groups segmented by foreign/
domestic ownership is evaluated both by aggregate worldwide business
conducted as well as by security type.
Indicators of Operating Performance
Table 6-3 catalogues the trends in four measures of operating perform-

ance: volume of business, number of issues, average volume per issue, and
gross spreads. The volume of underwritten offerings for the top twenty-
five (all firms identified in Appendix A), and the domestic/foreign seg-
ments are presented in Panel 1. The number of issues placed and average
volume per issue are reported in Panels 2 and 3. Panel 4 reports the
average gross spread in percentage terms. Both industry groups increased
the level of business done on an annual basis in all but two years of the
sample (1987 and 1988). The domestic industry underwrites a greater
volume of new business than does the foreign sector throughout the sam-
ple. This comparison understates, however, the business of foreign under-
writers since it does not include strictly foreign domestic business. For
example, issues underwritten by Nomura Securities for a Japanese firm
distributed solely in Japan or an issue for a West German firm distributed
solely in Germany are not counted.
One indication of operational efficiency is the comparison of average
volume($) per issue. Panel 3 shows that foreign underwriters, on average,
tend to bring to market larger issue volume. The fixed costs incurred in
transacting would therefore be spread over a larger issue volume, increasing
the profitability of the transaction. If domestic and foreign underwriters
face the same fixed costs, then the foreign underwriters may enjoy a cost
side advantage.
A measure of revenue performance is given by the gross spread. Panel
4 provides the average gross spread for all deals conducted for each group.
The average gross spread of the domestic group is higher than the foreign
group in eight of the ten years. Domestic firms appear to have an advan-
tage with contracted premiums for the services that they provide. The
trend in gross spread figures is one indication of increasing competition
in the industry. The difference in spreads measured in absolute values
ranges from a high of 2.45 percent in 1982 to a low of 0.058 percent in
Table 6-3. Aggregate Offering Production and Associated Fees 1 for Domestic and Foreign Underwriters.
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
Volume ($Bn)
Top 25 57.54 56.04 95.35 138.28 152.87 254.43 446.32 417.49 442.96 505.93
Domestic Mgr 51.07 51.08 54.27 89.14 92.37 150.32 275.04 253.76 273.65 305.06
Foreign Mgr 6.47 4.96 41.08 49.14 60.50 104.11 171.28 163.73 169.31 200.87
Issues (#)
Top 25 986 1009 1478 2303 1700 2393 4052 3546 4657 5372
Domestic Mgr 908 940 958 1705 1130 1746 2645 2241 2281 3889
Foreign Mgr 78 69 525 598 670 1047 1407 1305 1376 1483
Average Volume ($ Mn)
Domestic Mgr 56.24 54.34 56.65 52.28 81.74 86.09 103.98 113.24 119.97 78.44
Foreign Mgr 82.95 71.88 78.25 82.17 90.30 99.44 121.73 125.46 123.05 135.45
Gross Spread ('Yo),
Domestic Mgr: Mean 3.646 4.138 3.552 4.483 3.319 3.004 2.938 2.661 1.921 1.599
Foreign Mgr: Mean 2.223 2.472 2.103 2.250 1.834 1.848 2.036 2.018 1.979 1.939
I Full credit for issue allocated to book manager. Top 25 Underwriter status satisfied at least once. Appendix A provides the list of
underwriters used for each year in the period studied .

• Gross spread is computed as the sum of managing, underwriting, and selling fees as a percent of offer price.
1988. The reversal in average profitability is evidence consistent with in-

creasing competition across groups.
Trends in Market Concentration
One indicator of relative competitiveness is market share. Table 6-4 and

6-5 provide market share information. Table 6-4 shows for each group
total production, while table 6-5 shows the same data conditional on
security type.
Panels 1 and 2 of Table 6-4 report underwriting totals and market
shares for all issues other than international debt. The top twenty-five
underwrote more than 90 percent of the market in each year, except for
1982. Surprisingly, the presence of foreign underwriters as an important
factor in the domestic market was evident for the entire decade. Foreign
firms' market share was approximately 10 percent in 1980; their share
increased over the period, achieving a high of 19 percent in 1987. It also
appears to be the case that the market share increase of foreign under-
writers represented a transfer of business from the domestic underwriting
group. In the early part of the decade, domestic underwriters achieved a
90 percent market share. This appears to have been reduced and ranged
from 73 percent to 80 percent for the second half of the decade.
Panels 3 and 4 provide share information for all offerings that are
inclusive of international debt. The period 1982 through 1989 shows a
somewhat startling result. Market shares for worldwide business appear
to be constant. Domestic underwriters achieve an average market share
of 52 percent, ranging from 50 percent to 54 percent. The average market
share of foreign underwriters is 33 percent ranging from 30 percent to 34
percent. This suggests that the exchange of market share observed in the
domestic market is counterbalanced by activity in the international market.
Although this use of market share figures presumes that the market has
been correctly specified, the stability of the market share figures at the
international level supports our specification.
Table 6-5 provides market share information that is conditional on
security type. The market shares of foreign underwriters increased sub-
stantially for the domestic market over the period. Common stock under-
writing increased from a 7 percent share to a 27 percent share in 1987 and
a 26 percent share in 1989. This represents an approximate increase in
business of 300 percent over the period. Domestic debt and preferred
stock underwriting have also been successful target markets for foreign
underwriters. Domestic debt shares ranged from 10 percent in 1982 to 20
Table 6-4. Volume Totals and Market Shares for ''Top 25" Underwriters. 1 Volume Totals-U.S. $ Billions; Market Share--%.
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
(A) Total (1-5)' 58.10 56.60 65.95 103.35 95.47 150.95 317.00 326.18 301.89 340.89
Underwriter Totals
Top 25 57.54 56.04 50.08 94.61 86.42 143.29 296.83 301.42 282.41 317.07
Domestic Mgr 51.07 51.08 43.62 82.75 72.38 115.99 241.75 238.38 242.62 265.52
Foreign Mgr 6.47 4.96 6.46 11.86 14.04 27.30 55.08 63.04 39.79 51.55
Market Share of (A)
Top 25 0.99 0.99 0.76 0.92 0.91 0.95 0.94 0.92 0.94 0.93
Domestic Mgr 0.88 0.90 0.66 0.80 0.76 0.77 0.76 0.73 0.80 0.78
Foreign Mgr 0.11 0.09 0.10 0.11 0.15 0.18 0.17 0.19 0.13 0.15
(B) Total (1-6)' 58.10 56.60 128.26 165.83 185.89 302.13 522.38 481.68 509.91 571.52
Underwriter Totals
Top 25 57.54 56.04 95.35 138.28 152.87 254.43 446.32 417.49 442.96 505.93
Domestic Mgr 51.07 51.08 54.27 89.14 92.37 150.32 275.04 253.76 273.65 305.06
Foreign Mgr 6.47 4.96 41.08 49.14 60.50 104.11 171.28 163.73 169.31 200.87
Global Market Share
Top 25 0.99 0.99 0.74 0.83 0.82 0.84 0.85 0.87 0.87 0.89
Domestic Mgr 0.88 0.90 0.42 0.54 0.50 0.50 0.53 0.53 0.54 0.53
Foreign Mgr 0.11 0.09 0.32 0.30 0.33 0.34 0.33 0.34 0.33 0.35
1 Full credit for issue allocated to book manager. Top 25 Underwriter status satisfied at least once .
• (A) and (B) Market Totals repeated here from table 6-1.
Table 6-5. Dollar Volume Concentration By Security Type and Manager Class. Volume Totals-U.S.$ Billions.
1980-81 1982 1983 1984 1985 1986 1987 1988 1989

Common
Domestic 24.51 13.01 34.76 8.68 22.10 40.65 41.27 31.68 23.77
% Mkt 0.88 0.92 0.93 0.61 0.75 0.64 0.54 0.74 0.62
Foreign 2.55 0.97 1.82 1.20 4.06 9.39 20.81 5.44 9.85
% Mkt 0.09 0.07 0.05 0.08 0.14 0.15 0.27 0.13 0.26
Market 27.91 14.18 37.40 14.20 29.33 63.37 76.39 42.90 38.08
Dom Debt"
Domestic 73.21 25.82 40.38 60.03 85.83 188.60 187.97 204.43 234.89
% Mkt 0.88 0.55 0.70 0.78 0.76 0.79 0.79 0.82 0.80
Foreign 8.48 4.86 9.21 12.49 22.52 44.57 39.61 32.68 39.56
% Mkt 0.11 0.10 0.16 0.16 0.20 0.19 0.17 0.13 0.13
Market 81.96 46.74 57.66 77.33 112.76 238.92 237.85 250.45 293.62
Preferred"
Domestic 4.38 4.50 7.50 3.61 8.06 12.49 9.16 6.51 6.85
% Mkt 0.91 0.89 0.90 0.92 0.91 0.85 0.77 0.76 0.75
Foreign 0.40 0.50 0.82 0.30 0.71 1.12 2.62 1.68 2.13
% Mkt 0.08 0.10 0.10 0.08 0.08 0.08 0.22 0.20 0.23
Market 4.83 5.03 8.29 3.94 8.86 14.71 11.94 8.54 9.19
In1'l Debt
Domestic 0.00 10.85 6.39 19.99 34.33 33.29 15.38 31.03 39.54
% Mkt 0.00 0.17 0.10 0.22 0.23 0.16 0.10 0.15 0.17
Foreign 0.00 34.62 37.28 46.46 76.81 116.20 100.69 129.52 149.32
% Mkt 0.00 0.56 0.60 0.51 0.51 0.57 0.65 0.62 0.65
Market 0.00 62.31 62.47 90.41 151.13 205.30 155.43 207.93 231.14
a Domestic Debt and Preferred combines straight and convertible issues.
percent in 1985, while preferred offerings ranged from 10 percent to 22

percent in 1988.
Foreign firms' rising share of common stock underwriting appears to
have been captured at the expense of domestic underwriters. This does
not appear to be the case for debt. The total share of the market in 1980
and 1981 was split 88 percent to 11 percent, domestic to foreign. In 1982,
domestic underwriters appear to have lost 33 percent of their debt market
but not to the foreign underwriter group. The domestic underwriters then
continued to increase market share throughout the decade as did the
foreign underwriters. Domestic underwriters simply failed to recover the
position they enjoyed at the beginning of the' decade.
The international debt figures appear to be relatively constant for for-
eign underwriters. No apparent trend appears to be in place for the domestic
group.
Pricing
The cost of underwriting has been documented in the finance literature

as systematically related to various attributes of the issue. Generally, for
equal amounts of capital raised, gross spreads tend to be smaller for debt
issues rather than for equity, and they tend to be smaller for seasoned
equity than for initial public offerings. Spreads have been found to be
inversely related to the volume of the issue and they tend to be smaller
for competitive versus negotiated offerings. 9 Here, we model the rela-
tionship between the gross spread accruing to the underwriting syndicate
and several factors that have been shown to be related to the spread.
Additionally, we attempt to identify systematic influences peculiar to the
market segments studied here.
We model the gross spread as a function of the volume of the issue, the
industry of the issuer, the status of the underwriter (domestic or foreign),
and the nationality of the issuer (foreign issuer or U.S. corporation).
Additionally we control for the initiation of shelf registration under SEC
Rule 415, effective March of 1982, and the initial public offerings of common
stock. Issue-specific information was collected from IDD Information
Services. Ordinary least-squares regressions were run separately for com-
mon stock, domestic debt, and international debt. The results are re-
ported in Table 6-6. Due to the use of qualitative, explanatory variables,
the interpretation of regression coefficients on the dummy variables must
be performed with care. Their influence on the gross spread is in relation
to that identified by the null (off) condition on qualitative determinants
Table 6-6. Results of OLS Regressions to Explain Gross Spreads.
Model Common Domestic Debt Int'l Debt
Gross Spread 5.76 1.02 1.93

Intercept 6.78 1.31 3.28
In (Volume) -0.80 -0.14 -0.27
Mgr (For = 1, Dom = 0) -0.91 -0.12 0.21
Issuer (For = 1, Dom = 0) -0.85 0.03 -0.20
SEC Rule 415 -0.39 0.16
IPO (Y = 1, N = 0) 1.99
Industry
Industrial (Y = 1, N = 0) 1.28 1.03 0.05 1
Financial (Y = 1, N = 0) 1.37 -0.09 -0.37
Agency (Y = 1, N = 0) -0.93 1 -0.14 0.04 1
Adj. Rsq 0.56 0.16 0.02
F-statistic 30.56 29.74 23.20
N 5938 8525 6900
1 Insignificant at p = .05 level.
of the spread. The most general case is that for common stock. We interpret
the intercept term as the average gross spread on a seasoned issue for a
domestic public utility managed by a domestic firm prior to the initiation
of shelf registration. The spread for all other combinations are condi-
tional on active qualitative variables and additive in them. A summary of
the results follows.
First, the fees represented by the spread are inversely related to issue
volume for each issue type. This general result is consistent with previous
research on underwriter fees. Industry type does not enter into the modeled
spread in a systematic way across security types. The incremental fee
charged for common stock issues is positive and statistically significant for
both industrial and financial firms. The level of the coefficient is higher
for financial sector firms, perhaps indicating the increase of risk in their
operations during the decade of the 1980s. Domestic debt demonstrates
a different pattern for industry type. All three groups have significant
coefficients, but financial firms and corporate agencies enjoy a reduction
in costs that may be attributable to governmental guarantees of some
form. The industry dummy on international debt is negative and only
significant for financial firms.
Shelf registration is not modeled for international debt since the data
series begins in 1982, the year that Rule 415 becomes effective. Shelf
registration should reduce the cost of issuance to managing underwriters.

This cost reduction is expected to be passed on to issuers and should be
independent of issue type. The findings presented here suggest that the
expected result is found for common stock issues but not for debt issues.
Initial public offerings, a relevant variable for common issues only, is
statistically significant and is the largest increment to average gross spreads.
This is as expected and consistent with extant research.
The coefficients on the status of the managing underwriters are consist-
ent with increased competition in the underwriting market. Both common
and domestic debt have significant and negative effects on fees. This suggests
that foreign underwriters are charging lower gross spreads on average for
business underwritten in the United States. Surprisingly, this is not the
case for international debt. Two interpretations are consistent with the
signs of these coefficients. First, foreign underwriters underprice domestic
issues as an entry cost to the domestic market but do not have to system-
atically underprice in their established markets. Alternatively, the results
could be consistent with the relative risk of the business conducted in
each market. Neither case can be refuted by these results.
Issuer status is also significant in each model but has varying signs.
Foreign issuers of common stock incur lower fees on average as indicated by
the issuer variable. Since most equity issues for foreign firms are in the
form of American Depository Receipts (ADR), this lower cost may be
capturing the sponsor's layer of intermediation. We do not control for
sponsored versus unsponsored ADRs.
In the next section we attempt to measure the impact of foreign com-
petition on the equity values of domestic rivals.
Foreign Competition and Stockholder Wealth
Stockholder wealth has been used in the industrial organization literature

to measure the impact of various corporate events, such as mergers and
acquisitions. Eckbo (1983), for example, uses stock price changes to test
whether horizontal mergers increase the probability of successful collu-
sion among the remaining rivals. Similarly, an overall industry pattern of
positive abnormal returns can indicate profitable investment opportuni-
ties, or a negative pattern, a reason for consolidation through mergers. In
this section, we examine patterns of abnormal rising and falling stock
prices for the underwriting industry over the 1980s as an indicator of
consolidation or entry. We look specifically at the effect of foreign entry.
We also examine some accounting measures of profitability to ascertain
whether the abnormal stock price behavior is confirmed by subsequent

accounting data. We then look at the effect of horizontal mergers be-
tween foreign and domestic firms on the remaining domestic rivals equity
value. Finally, we examine the stock price reaction among domestic rivals
to the entry of foreign competition into the domestic market through
their registration with SEC as a broker/dealer.
Data
According to the data in the previous sections, about fifty firms accounted
for more than 95 percent of the underwriting activity during the 1980s. Of
these fifty large underwriters, we found ten firms that had publicly traded
equity in the United States and whose prices were available on the Center
for Research in Security Prices daily data tapes during some part of the
period from 1976 to 1989. We assembled these firms' daily stock returns.
Table 6-7a lists the firms, their initial listing date, the market where they
traded, the date they stopped trading, and the reason they stopped trading.
We also gathered several measures of accounting profitability for these
firms over the ten-year span, 1980 to 1989, from Compustat. Table 6-7b
lists the averages (in percent) of the firms' yearly return on investment,
return on assets, return on equity, year-to-year percentage changes in
sales, cash flow, and assets. It also lists the ten year geometric averages.
Most of the averages suggest that the mid-years of the 1980s were the
most successful for the firms. However, the trend since then has generally
been down; most of the recent observations are below their averages. For
example, while sales continue to grow, internal cash flows have not kept
pace.
There were at least 550 completed deals during the 1980s involving
domestic broker/dealers and investment banks. Of these 550 deals, ninety-
four involved foreign acquire. Figure 6-1 shows that the number of for-
eign acquire has been growing over the decade, with 28 completed deals
in 1989 alone. We identified 47 announcement dates for deals of $25
million or more involving foreign acquire/investors in domestic broker/
dealer firms. These deals included partial investments, such as Nippon
Life Insurance Company's purchase of a 13 percent stake in Shearson-
Lehman for an estimated $535 million, and Yasuda Mutual Life Insur-
ance Company's purchase of an 18 percent voting stake in Paine-Webber
for an estimated $300 million.
Foreign firms may also enter the U.S. market directly by registering
with the SEC as a broker/dealer. We have identified several registration
dates for foreign firms from confidential internal sources.
Table 6-7a. Domestic Broker/Dealers with Publicly Traded Equity in the 1980s.
Name Date Listed Date Delist Reason/or

(Per CRSP) Delisting
Alex Brown, Inc. 02/28/86
Bear Stems Company 10/29/85
Citicorp 11101168
First Boston Inc. 12/14n2 12/22/88 Acquired by
Credit Suisse
Merrill Lynch & Co. 07127n1
Morgan Stanley Group Inc. 03/21186
Paine Webber Group 04/03n2
Piper Jaffray Inc. 02/03/86
Salomon Inc. before 7/62
Shearson Lehman 05/07/87
Table 6-7b. Various Accounting Rates of Return and Growth Rates. Average of
Ten Publicly Held Domestic Broker/Dealers.
Account ROA RaJ ROE Sales Cash Total

Year Flow Assets
1980 2.115 14.911 16.981 27.001 13.21 9.8
1981 1.426 12.862 17.746 14.905 -41.401 12.039
1982 1.224 16.768 23.936 10.24 26.226 70.811
1983 1.818 18.63 25.81 12.44 28.02 16.931
1984 0.398 5.35 8.537 4.071 -13.044 12.394
1985 0.829 8.808 15.399 8.731 50.108 24.685
1986 2.037 21.855 34.175 21.365 22.755 18.27
1987 1.471 13.437 20.63 10.992 25.621 -0.379
1988 1.101 9.391 20.43 6.993 2.382 18.922
1989 0.939 6.341 14.761 17.507 -0.807 12.723
Ten Year 1.407 13.736 21.903 14.005 8.416 20.524
Averages
Number of Completed Deals
160r-----------------------------------------------------~====~--~
140
1 :::r:::'::::::~ I 1
1201~------------------------------------------------~:
100r---------------------------------------------~~~.
M ~~
60 ~~~~~~r~~~t ::::::::
40 I f:-:':':':':':' x x
2: : 1'1 ~ II,,: =m tJ t'a<l ,
1980 81 82 83 84 85 86 87 88 1989
Figure 6-1. Completed mergers and acquisitions among domestic broker/dealers.

THE COMPETITIVE IMPACf OF FOREIGN UNDERWRITERS 231
Methodology
Figure 6-2 shows the cumulative average abnormal returns experienced

by the sample of underwriting firms during the 1980s. The cumulative
abnormal return (CAR) is computed in two stages. First, a market model
is estimated for each firm as,
'i,t = ai + ~l m,t + E.i,t (1)
where 'i,t and 'm,t are the daily, continuously compounded rates of firm i
and the equally weighted market portfolio. Each firm's {3, estimated in the
regression equation above, captures the firm's systematic movement with
changes in the market's return. The error term measures the firm's
nonsystematic (or idiosyncratic) risk component, and captures the effect
of industry- and firm-specific events on a firm's equity value.
Because each firm's systematic risk may be unstable, a rolling window
was used to estimate the market model coefficients. The a i and ~i are
estimated yeady for each firm using ordinary least squares and the stock
returns from the prior two years. A minimum of 250 returns had to be
available for the estimation procedure. Using these estimated coefficients,
abnormal returns (forecast errors) were computed for the current year,
as,
(2)
This same procedure was used repeatedly, advancing one year, using the
most recent two year span to estimate the coefficients, then calculating
one year's worth of abnormal returns.
In the second stage of the cumulative process the abnormal returns
were averaged across firms at each point in time, forming portfolio abnor-
mal returns, 'p,t. These averaged abnormal returns were then cumulated
across time, forming CARs. This procedure is similar to cumulating the
abnormal return of a portfolio with changing components that is con-
tinuously rebalanced to maintain equal weight in each component stock.
Results
Industry Patterns. A pattern of rising cumulative abnormal returns may

signal new profit opportunities and thus provide an incentive for entry.
A pattern of falling cumulative abnormal returns can indicate falling
demand or rising supply prices, and hence may be associated with
Cumulative Abnormal Return
0.7 I
0.6T'~~~~~-~'r'------~--------~
0.5 I I 1\ I
0.4 I II ~l l
0.3+ ,)' \ ill' II•. 1 ML,;ilF'\,/\
0.2 I I,' ' 11\;..111. fIJ"fr W. I

N'
tf k F\ I I
0.1--+1+1-+--
o I' '" I, Ill.. RWllfl 1\1 'l ,
-0.1
- 0.2 I , \ Af ,"", I
-0.3~--__~------~----~------~----~------~----~------~----~----~
81 82 83 84 85 86 87 88 1989
Figure 6-2. Cumulative abnormal returns. Underwriters during the 1980s.

consolidation within the industry. Figure 6-2 shows that the cumulative
abnormal return peaked late in 1982 at around 60 percent. The early part
of 1983 was disastrous for the underwriters as a group, which lost not only
the cumulated 60 percent but also an additional 10 percent by the end of
the year. The years 1984 and 1985 generally were good ones for investors,
only to be followed by a significant downturn early in 1987. Since the
October 1987 market break, the abnormal returns have been generally
positive, moving from a negative 20 percent cumulative abnormal return
to a positive position of about 15 percent at the end of the decade.
Accounting rates of return may provide additional insights into the
competitive position of domestic underwriters during the 1980s. Table
6-7b presents yearly average accounting rates of return (on investment
(ROI), on assets (ROA), and on equity (ROE» for the ten firms in our
sample. Rates of growth in assets and sales are also presented.
The stock market is forward looking, whereas the accounting data are
historically oriented. The CARs are likely to be leading indicators of
subsequent accounting results. Figure 6-3 shows the yearly change in the
cumulative returns (lead one year); the change in average rate of return
on equity; the average yearly growth rate in sales; and finally, the number
of completed acquisitions (both foreign and domestic) in the industry for
the year. In all years, increases (decreases) in CARs over the past year
were associated with increases (decreases) in sales growth in the current
year. In six of the eight years covered, increases (decreases) in last year's
CARs were associated with increases (decreases) in this year's return on
equity. Finally, in five of the seven years in which there were appreciable
mergers and acquisition activity, increases (decreases) in last year's CARs
were associated with decreases (increases) in this year's mergers and
acquisitions (M&A) activity.
The association of merger and acquisition activity with poor perform-
ance is consistent with the notion that mergers consolidate an industry's
assets, particularly an industry that faces significant fixed costs. Thus, this
evidence indicates that individual merger and acquisition in this industry
is likely to have a mixed impact on the rest of the firms in the industry.
In the next section we look at the impact of this activity on the equity
values of the industry's publicly traded firms.
The Impact of Foreign Entry. Unexpected changes in a firm's operating

environment can change the market value of its equity. An anti-competitive
merger might raise product prices, and thus increase the equity value of
the remaining firms in the industry. Here, however, the entry of a foreign
competitor is likely to signal increased competition in the output market
Changes in Rates
100
I] ROE §
I Sales
E
80
~CAR §
60
I3M&A ~
§
§
40
§
•
§ ~
Jl
20
j §
o m
r
~ AI
•
I
~ f;i;J
-20
~
-40
-60 I I I I I I I
1982 83 84 85 86 87 88 1989
Figure 6-3. CARs as leading indicators of changes in growth rates and M&A
activity.
(lower product prices) and/or in the input market (higher input prices).
These changes are likely to lower the market values of the entrant's
domestic rivals. We attempt to test for this effect by examining the abnormal
returns to a stock portfolio that includes the major publicly traded domestic
underwriters around the announcements of foreign entry into the domestic
market.
Of the 550 completed acquisitions in this industry, we found forty-
seven that involved foreign purchasers of domestic firms or significant
stakes in domestic firms, according to IDD Information Services. These
forty-seven deals were $25 million or larger. We identified the announce-
ment dates of these foreign acquisitions of domestic dealer/brokers and
isolated the abnormal returns (described above) to the portfolio of domestic
underwriters on those particular dates. A two-day window was used, the
day of the announcement and the next day, in case the information arrived
after the close of business on the announcement date. These forty-seven
two-day CARs were then averaged to measure the average impact of an
announcement of a foreign acquisition of a domestic dealerlbroker. Overall,
the standard deviation of two-day CARs is 1.65 percent. The two-day

cumulative average abnormal return on the announcement of a domestic
acquisition by a foreign firm is -0.23 percent, slightly negative, but not
significantly different from zero. [An examination of the two-day cumula-
tive returns for individual acquisitions shows significant variation, with
some as large as a positive 2 percent but, at the other extreme, negative
2 percent.] However, the net result is approximately zero. We find no
consistent reaction to foreign acquisitions of domestic broker/dealers.
A second avenue of foreign entry into domestic markets is to register
directly with the SEC as a broker/dealer. From confidential internal sources
we have identified seven registration dates involving major foreign
broker/dealers. Using the same methodology as outlined above, we found
negative average portfolio abnormal returns of -0.5 percent on the an-
nouncement date and a two-day average abnormal return of -0.7 percen~.
However, the standard deviation of the portfolio's daily abnormal return
was 1.17 percent; thus, neither negative abnormal return is statistically
significant.
In summary, we find that industry abnormal stock price performance
to be a leading indicator of merger and acquisition activity within the
group, consistent with consolidation motive for an industry that faces
fixed costs and rising competition. Although we find a negative impact on
the equity values of publicly traded domestic broker/dealers when foreign
firms enter the domestic market, either through acquisition or direct reg-
istration, these results are weak and not statistically significant.
Summary and Direction of Future Research
The overall picture one can take from the data presented here is that the
domestic underwriting business is very competitive, with average spreads
falling through most of the decade and with foreign entry through acqui-
sition or registration providing additional competitive pressure. However,
it also appears that the domestic underwriters have been successful in
expanding their business overseas. So, whereas the domestic market share
of the major domestic underwriters has fallen, their global market share
has expanded such that their total market share has remained relatively
constant.
Also salient is that changing nature of the underwriting business. Debt,
especially asset-backed debt, has become the dominant issue in the do-
mestic market. Financial innovations, such as mortgage-backed securities,
which were a negligible portion of the market in the beginning of the

decade, now represent a major portion of the underwriting business. It is
likely that new products and services, such as exchange offers on existing
debt, may well represent dynamic growth areas for the future.
Finally, we have attempted to provide an overall picture of the impact
of foreign competition on the domestic underwriting business. While we
have necessarily provided just summary statistics, they do indicate some
direction for future research. For example, while we have documented
the extent of foreign presence in the domestic market, we do not know
the source of the competitive advantage they might enjoy. Do underwriters
just service their own domestic clients and follow their customers overseas
when they expand into foreign market? Or do underwriters enjoy suffi-
cient economies of scale to overcome the natural cost disadvantage of
operating in a foreign market. We hope to pursue some of these issues in
our future research.
Appendix A
Name List of Underwriting Firms 1980-1989.
Domestic Foreign
Alex Brown & Sons Kidder, Peabody Algemene Bank Nederland Lloyds Merchant Bank
Allen Laidlaw Amsterdam-Rotterdam Bank Natwest Capital Markets
Bankers Trust Lehman Brothers Banca Commerciale Italiano Nikko Securities
Bateman Eichler L.F. Rothschild Banque Nationale de Paris Nomura Securities
Bear, Stearns Merrill Lynch Banque Paribas Capital Markets N.M. Rothschild & Sons
Citicorp Montgomery Barclays Merchant Bank RBC Dominion Securities
D.H. Blair Morgan Stanley Canadian Imperial Bank of Commerce Societe Generale
Dain Bosworth Moseley Securities Commerzbank Soditic
Dean Witter Paine Webber Credit Commercial de France Swiss Bank
Dillon, Read Piper Credit Lyonaise S.G. Warburg
Donaldson Prudential-Bache Credit Suisse Union Bank of Switzerland
Drexel Robertson Credit Suisse/First Boston Yamaichi International
E.F. Hutton Salomon Brothers Daiwa Securities
Faherty & Faherty Smith Barney Deutsche Bank
First Boston Stephens Dresdner Bank
First Jersey Thomson McKinnon Hambros
Goldman, Sachs Wertheim Schroder Indosuez
Hambrecht & Quist Wheat, First Industrial Bank of Japan
John Muir Wheat-Butcher Kleinwort, Benson
J.P. Morgan William Blair Lazard Freres
Notes
1. Worldwide offerings by underwriters registered to do business in the United States are all
publicly offered corporate securities that are distributed in the United States, as well as all
offerings issued internationally. We do not have access to strictly domestic, foreign issues, i.e.,
foreign securities only issued in their home country.
2. The source for most of the data in this section is various issues of Investment Dealer's
Digest and IDD Information Services.
3. See Hayes, Spence, and Marks (1983) for an in-depth analysis of the market for underwrit-
ing services during the decade of the seventies as well as for an excellent historical review of the
market and the investment banking sector. Market identifying characteristics applied here are
essentially the same as those described in Hayes, et.al. (1983).
4. Hansen and Khanna (1990) document that competitive bid deals registered with the U.S.
Securities and Exchange Commission for the period 1972-1986 accounted for no industrial firm
offerings (6807 deals registered) and 28.9 percent of public utility offerings (3589 deals regis-
tered). For the period studied here, offerings representing competitive bids account for 2.8
percent of the business done by the managing firms analyzed.
5. The influence and size of syndicates are decreasing over the period of this study. The
importance of the lead manager in negotiated transactions should not be underestimated. The
managing underwriter establishes the underwriting agreement with the issuer and identifies
the terms of the offering as well as each member's commitment. The lead manager controls the
"pot". Typically, each syndicate member controls only a portion of the securities which it has
agreed to underwrite. This portion is referred to as "retention". The balance is placed in a general
syndicate account-the pot-which can be reallocated by the lead manager to syndicate
members on the basis of their ability to sell. An aspect of the "pot" which has taken on increasing
importance in the eighties is that it provides institutional buyers with distributional conveniences
such as central billing and delivery. These institutional orders are typically placed with the
managing underwriter but may be earmarked by the purchaser for the account of designated
selling agents.
6. A form of negotiated, firm commitment underwriting which is not a part of the market
studied involves "swapping". Swap transactions utilize securities as the medium of exchange in
the purchase of underwritten securities. They provide the dealer with the ability to diversify
security holdings during the period of distribution and give institutions an alternative means of
payment if it faces cash flow constraints.
7. Confidentiality of foreign-controlled registrants provided by the NYSE was requested.
Only aggregated information is therefore presented for foreign-controlled firms. Foreign-
controlled firms which appear in Appendix A were identified through public sources independ-
ently of information provided by the NYSE.
8. The annual reports are referred to as FOCUS Reports, the Financial and Operational
Combined Uniform Single Report: SEC Form X-17A-5.
9. This has been documented primarily for the issues of public utilities.
References
Brown, Stephen J. and Jerold B. Warner. 1985. "Using Daily Stock Returns: The
Case of Event Studies." Journal of Financial Economics 14 (March): 3-31.
THE COMPETITIVE IMPACf OF FOREIGN UNDERWRITERS 239
Eckbo, B. Espen.1983. "Horizontal Mergers, Collusion, and Stockholder Wealth."

Journal of Financial Economics 11 (April): 241-273.
Ederington, Louis H.1974. "The Yield Spread on New Issues of Corporate Bonds."
Journal of Finance 29 (December): 1531-1543.
Hansen, R., and N. Khanna. 1990. "Why Negotiation with a Single Syndicate May
Be Preferred to Making Syndicates Compete." Working paper. (May).
Hayes, Samuel L., A. Michael Spence, and David Van Praag Marks. 1983.
Competition in the Investment Banking Industry. Cambridge, MA: Harvard
University Press.
Kidwell, David S., M. Wayne Marr, and G. Rodney Thompson. 1987. "Shelf
Registration: Competition and Market Flexibility." Journal of Law and
Economics 30 (April): 181-206.
Logue, Dennis E. 1973. "On the Pricing of Unseasoned Equity Issues: 1956-1969."
Journal of Financial and Quantitative Analysis 8: 91-103.
Marr, M. Wayne, R. Rogowski, and J. Trimble. 1989. "The Competitive Effects of
U.S. and Japanese Commercial Bank Participation in Eurobond Underwriting."
Financial Management 18 (Winter): 47-54.
Schwert, G. William. 1981. "Using Financial Data to Measure Effects of Regulation."
Journal of Law and Economics 24 (April): 121-158.
Stoll, H. 1976. "The Pricing of Underwritten Offerings of Listed Common Stocks
and the Compensation of Underwriters." Journal of Economics and Business
28 (Winter): 96-103.
Walter, I. (ed.) 1985. Deregulating Wall Street: Commercial Bank Penetration of
the Corporate Securities Market. New York: John Wiley & Sons.
COMMENTARY
Samuel L. Hayes
The NachtmannlPhillips-Patrick article makes a useful contribution to

understanding what is happening in the global underwriting area. I think
the article was mislabeled, however. Rather than being a study of domes-
tic U.S. competition, it's the study of global competition. It chronicles the
movement of issuer financing offshore. The fact that there is relatively
little in the article that deals with the U.S. domestic underwriting market
doesn't really surprise me. Buyers and sellers of capital are, in large part,
shopping globally.
Equity Market Shares
I want to give particular attention to the findings that the biggest change
in the market shares have accrued in the equity area. It looks more dra-
matic in the article than its actual market impact. The increase in percentage
of equity financings done by foreign vendors has got to be offshore
equity. The authors' statistics, as I understand it, lump international and
domestic equity financings together and they therefore were not in a position
to distinguish between those offerings that were done offshore and those
241
that were done in New York. I believe that disaggregated statistics would
show that foreign vendors have a negligible share of the U.S. equity market.
The weak position of foreign vendors in the New York market is mir-
rored in other national markets around the world. In those markets, home-
based vendors have, for the most part, a very firm grip on the equity
offering business undertaken there.
Among preferred offerings in the United States, I don't believe that
foreign vendors have a significant position, either. Thus, the explanation
for the authors' preferred stock market share figures must be due to the
inclusion of the Euromarket financings in their data.
Vendor Profiles
Characteristics of the vendors is another interesting aspect of this article.

The authors distinguish between firms that are owned and controlled within
the United States and firms where at least 25 percent of their equity is in
the hands of offshore owners. It is certainly noteworthy that the firms that
are foreign controlled are a lot larger than those that are U.S. controlled.
I wonder if that can't be explained in part by the fact that these foreign
controlled organizations are, in many cases, offshoots of much larger
financial service conglomerates? I don't know whether or not the data
that the authors were able to obtain isolated the underwriting subsidiaries
of these foreign parents' firms. But even if this was the case, my guess is
that their financial structures and operating policies have nonetheless been
influenced by the deep pockets of their parent organizations. It also seems
likely that foreign controlled vendors operating in the United States would,
on balance, be larger than the group of twenty-five leading domestic U.S.
securities firms. They are, after all, the ones who can afford to come and
establish a foothold position in a foreign market that's as competitive and
as expensive to operate in as the U.S. marketplace.
The authors note that the financial leverage of the foreign firms was
greater than that of the domestic firms; this is, I think, an interesting
observation. Assuming that we are talking about the balance sheets of the
securities underwriting subsidiaries of these foreign organizations, then
the authors are probably correct in attributing it to their mix of business.
If it is largely an agency business without the kinds of assets that are
vulnerable to substantial price fluctuation, that large an equity base is not
needed. The business mix of U.S.-controlled securities firms is probably
much more heavily weighed towards principal business, with assets sub-
ject to much greater fluctuations in value. Thus, a larger equity base would
be required by these vendors to ensure their viability. Although the New
York Stock Exchange (NYSE) has "haircut" requirements that require
the setting aside of certain amounts of equity for specific types of assets,
it is my own personal opinion that these minimum requirements have in
some cases proved to be inadequate.
Gross Spreads
The authors chronicle the squeeze in gross spreads over time. There is no
doubt that the marketplace is becoming increasingly competitive. In the
United States we used to have the standard seven-eights of 1 percent gross
spread for "plain vanilla" bond underwritings. That has been drastically
reduced over the past fifteen years, particularly with the advent of shelf
registrations. If we take the total mix of financings done both in the
United States and in the international markets (the authors' frame of ref-
erence), there has been an impressive squeezing down of the spreads.
The authors' statistical results are intuitively appealing. They are what
I would have expected to find with respect to differences between "plain
vanilla" bonds and such things as initial public offerings. I think that
Nachtmann appropriately cautioned that in looking at these gross spreads
and their changes over time, it is important to carefully note alterations
in the financing mix, particularly on the debt side. The heavy volume of
asset-backed securities and so-called "junk bonds," which carry very high
gross spreads, are bound to influence the overall aggregate results.
I have two further comments on gross spreads. With respect to the
shelf registrations, a recent study provides some evidence that gross spreads
are being influenced by the exposure of the vendor to legal risk. Vendors
don't often have the time for careful examination of the issuer's affairs
before bidding on the financing and, therefore, may be vulnerable to a
charge of lack of "due diligence" in the event that the issuer's affairs
subsequently take an unexpected tum for the worse. This phenomenon
might help to account for the authors' finding that debt gross spreads
didn't shrink quite as much as one might have expected.
The other comment relates to a statement in the article about equity
gross spreads that is probably a [misprint.] There is no evidence in the
authors' paper to suggest that the foreign vendors have cut gross spreads
on equity financings in the United States. The data is not there to reach
that conclusion.
Stockholder Wealth
Turning to the stockholder wealth question, I see a lot of these "event"

studies and I'm not at all surprised that the authors came up with an
empty bag. On the one hand, I think it was helpful to pinpoint the special
events that are particularly relevant to the underwriting business, as
opposed to using some broader index measure. It is also useful to note
the backward-looking nature of the accounting data issued each year, as
opposed to the premium returns earned by this group of public com-
panies during each subsequent year. The authors are therefore able to
confirm that the stock market is efficiently anticipating results not yet
formally released. But, it does seem to me that it was a long shot on the
authors' part to hypothesize that the influx of foreign vendors into the
United States would be likely to have much of an impact on the share
prices of the ten publicly held financial vendors used in their sample. The
composition of these firms' business is just too diverse. I would also take
exception to the statement in the article suggesting that one might find
that an influx of foreign competitors would cause the prices of the domestic
competitors' stocks to go up because the new entrants would invigorate
domestic vendors to greater innovative efforts. That may be true in an
idealized world, but I'm not too optimistic that it would work that way in
the real marketplace. I also had one other less-weighty question about
why Citicorp would be included in this group often U.S. securities vendors.
Citicorp's real investment banking activity, as a fraction of its total rev-
enues, is very small.
In closing, the article provides some useful perspectives on underwriting
competition. As a sort of satellite photograph, we can see some features
that are important to an understanding of the overall direction of the
public marketplace and the operations of vendors within that public
marketplace. The authors conclude that the competitive position of the
U.S. vendors has not been eroded by the increased competition during
the 19808. I believe this has been due, in part, to the U.S. investment banks'
aggressive efforts to position themselves in other offshore market centers
as they follow their customers. In a recent book (Hayes and Hubbard),
Philip Hubbard and I have concluded that the marketplace for money is
already global from the point of view of both buyers and sellers of money.
The "electrical grid" of the global capital market is pretty much in place
and, when you plug into one part of that grid, it courses throughout that
whole system. By contrast, the vendors serving this interlinked marketplace
are not yet global. While they are working toward that goal, there are a
number of barriers to entry into various marketplaces that are difficult to
overcome. And so I think it's a real tribute to the U.S. investment banks
that, despite these offshore barriers to entry, this paper suggests that they
have nonetheless been able to hold their own in the market environment.
Morgan Stanley, which had only several hundred employees in one New
York location in the early 1970s, now has more than 6,000 employees in
locations all over the world. Given these changed requirements, the
authors' comments about increased concentration in the U.S. securities
industry aren't surprising. There are only a half dozen or so firms that
really count in terms of competitive interactions with foreign owned
securities firms. And that handful of dominant U.S. firms continue to be
world-class competitors.
References
Hayes, Samuel L., and Philip Hubbard. Investment Banking: A Tale of Three Cities.
Harvard Business School Press, Cambridge: 1990.

Gorton, Pennacchi, Gilbert Eds - The Changing Market in Financial Services

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Gorton, Pennacchi, Gilbert Eds - The Changing Market in Financial Services

Caricato da

Copyright:

Formati disponibili

The Changing Market in Financial

Springer Science+Business Media, LLC

Economic Policy Conference of the Federal Reserve Bank of

Copyright © 1992 by SpringerScience+Business Media NewYark

AII rights reserved. No part of this publication may be reproduced,

Printed an acid-tree paper.

Contributing Authors vii

Commentary by Stuart I. Greenbaum 35

Commentary by Peter S. Rose 93

Commentary by Herbert M. Kaufman 125

Commentary by Gary C. Zimmerman 201

Commentary by Samuel L. Hayes 241

Lawrence G. Goldberg Patrie H. Hendershott

George G. Pennacchi Peter S. Rose

The articles and commentaries included in The Changing Market in

commercial banks. The authors conclude that technological progress has

When a corporation seeks to raise resources, it can issue securities such

In a few short years the outstanding principal of repackaged automobile

An asset-backed security is a claim on a portfolio of underlying bank

Table 1-1. Asset-backed Securities Volume (Issued Annually).

Table 1-2. Asset-backed Securities Issuance by Type of Issuer (1988) ($ billions).

guarantor looks to the issuer to back up the guarantee. Banks, however,

Automobile Receivable-backed Securities

The automobile receivable-backed securities market began in May 1985

enhancement. These developments seem consistent with a growing maturity

Credit Card Receivables

Credit card receivable-backed securities are claims on a designated (and

The Loan Sales Market

Sales of commercial and industrial loans, or secondary loan participations,

than consumer loans. They do not necessarily have standardized contracts

The ability of the loan sales market to incorporate increasingly risky

Implicit Loan Sales Contracts: The First Tests

Implicit Guarantees By the Selling Bank

The possibility of implicit insurance is clearly not motivated by any ob-

is an implicit guarantee, then the value of this guarantee would be related

Equation (2) is designed to test the relative strengths of guarantees given

Maintenance of a Stake in the Loan Sold

Table 1-5. Fraction Sold.

Rating 0-5 6-15 16-30 31-60 61-90 90+

Source: Money Center Bank.

Loan Sale Incentive Compatibility: More Sophisticated

This section summarizes the model presented in Gorton and Pennacchi

A Model of Loan Sales

monitoring, a, the fraction of the loans to be sold, s, and the fraction of

] = 1- e- rfT [f: sxdF(x, a) + sr fOL (L - x)dF(x, a)]

subject to: (i) foL [(1-s)+srp]xdFa (x,a)=c'(a)

xdF(x, a) = L(1- ae-pa ) (5)

Empirical Results 1: Equal Guarantee on All Loan Sales

Empirical Results 2: Different Guarantees on Loan Sales

When the bank is assumed to be able to make different levels of partial

ko + k 1r+ (l/2)k2r2. Based on the parameter estimates for k1 and k2 the

There appears to be no evidence of an implicit guarantee or implicit

Technological Change Creates Symmetric Information

The results of estimating the model in search of evidence for an implicit

affecting the underlying contracting technology have occurred. While it

Informal Evidence of Technological Change

If institutional investors have benefitted the most from technologies that

rowers. Banks' comparative advantage in eliminating duplication of infor-

Empirical Results 3: Observability in the Loan Sales Market

A final test is performed in Gorton and Pennacchi (1990A) that provides

that the bank's level of monitoring is not directly observable by the

Concluding Comments: Marketable Bank Assets and

Technological change appears to have resulted in the opening of markets

fr = .17890 + .00151 maturity + .00742 rate