The Size of Government: Measurement, Methodology and Official Statistics

THE SIZE OF GOVERNMENT

Austrian Economics

Series Editor: Per Bylund

The Austrian school of economics offers a powerful framework for understanding major economic events such as the fall of socialist economies in the early 1990s and the financial crisis of 2007–08 and thanks to this promise, and to the limitations of mainstream economics, the Austrian tradition has attracted increasing interest from a new generation of economists and social scientists. This series seeks to capture this renewed interest by publishing original research within the modern Austrian tradition.

Published

Capitalism, Socialism and Property Rights

Mateusz Machaj

The Size of Government

Václav Rybáček

THE SIZE OF GOVERNMENT

Measurement, Methodology and Official Statistics

VÁCLAV RYBÁČEK

© Václav Rybáček 2020

This book is copyright under the Berne Convention.

No reproduction without permission.

All rights reserved.

First published in 2020 by Agenda Publishing

Agenda Publishing Limited

The Core

Bath Lane

Newcastle Helix

Newcastle upon Tyne

NE4 5TF

www.agendapub.com

ISBN 978-1-78821-010-2

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Typeset by Newgen Publishing UK

CONTENTS

Introduction

Chapter 1 The size of government in economics

1.1 Economic theory

1.2 Economic research

1.3 Economic freedom and macroaggregates

1.4 Holism and macroeconomics

1.5 Austrians and macroeconomics

Chapter 2 The role of measuring government in economic policy

2.1 Macroaggregates and the European Union

2.2 The banks

2.3 All for good figures

2.4 Lesson learned, or ignored?

Chapter 3 A short trip to the past

3.1 Growing public

3.2 Disappearing government

3.3 Rediscovered government

Chapter 4 The current approach

4.1 Taxes

4.2 Expenditures

4.3 Cost of government

4.4 Delimitation of the general government sector

Chapter 5 What’s wrong with the current approach?

5.1 Fifty is not enough

5.2 Market forces at work: reality or illusion?

5.3 Central banks

5.4 What market is a real market?

5.5 What is missing?

5.6 Where is the dividing line?

Chapter 6 The size of government and GDP

6.1 The informal and illegal economy

6.2 Imputed values

6.3 Government in GDP

6.4 GDP and the size of government

6.5 GDP recalculated

Chapter 7 Is the size of government underestimated?

7.1 Government expenditures extended

7.2 The government sector extended

7.3 Further methodological issues

Chapter 8 Recalculation of the size of government

8.1 Government expenditures recalculated

8.2 Recalculated indebtedness

8.3 Estimated tax basis and the size of government

8.4 Going conventional

8.5 Should the indicators be redefined?

Conclusion

Notes

Bibliography

Index

INTRODUCTION

Everyone knows that government has grown, but no one knows exactly how much.

Robert Higgs

In 2016, the debt of Germany stood at 68.3 per cent of GDP with a downward trend. Among highly indebted European countries, Germany was considered one of the finest. However, what if we said that the debt of Germany stood actually at 314.9 per cent, with an upward trend? Would that be insanity? Not necessarily. Statistical reality is only a matter of statistical methodology and statistical theory. We, as data-users with critical thinking, should always ask ourselves a simple question to begin with: how can statisticians know that? We shall keep asking throughout this book: what methodology stands behind those figures? What are its weaknesses? And ultimately, are we really getting the information that we expected?

Questions of this kind might be asked in any statistical area. As Morgenstern (1963) rightly pointed out, statistical observations are deliberately designed, guided by constructed statistical theories which make macroeconomic aggregates graspable, but with aggregation comes costs. Errors of observations, calculation errors, misreporting and many other factors can introduce high uncertainty into final figures and their explanatory power. However, this typically remains hidden from data-users who get little idea of what degree of uncertainty is inherent in the final aggregates.

Let’s suppose that the most recent rate of GDP growth released by statisticians reached 2 per cent annually. During its calculation and the balancing process, a certain discrepancy between production and expenditure inevitably occurred. Statisticians then pick a method of how best to reconcile both sides of the balance. The final growth rate depends, of course, on their decision. They can increase total expenditure or to lower production, or vice versa, in all cases influencing the annual growth rate. It is entirely conceivable that the range of potential results obtained can vary from zero to 4 per cent growth. Therefore, we could have concluded that economic growth had come to a standstill, but we could have equally well concluded that we are experiencing solid economic growth.

It would be more honest if statisticians shared confidence interval with users, but statisticians prefer to keep them in the drawer. The purpose of this book is to open up this drawer in the area which has recently received wide attention: government statistics. We shall demonstrate how currently available data can be rearranged and definitions extended, to arrive at more or less different statistical representations of government. We shall also assess the relevance of the statistical theory from the perspective of the Austrian school of economics, which will help us fully appreciate the limitations of statistics as well as efforts to figure out the optimal size of government.

Under the term the size of government, we can understand a variety of measures. Most frequently, it is regarded as the amount of financial means redistributed through the government accounts. However, we can conceive the size equally well as stock of scarce resources like labour force, non-financial and financial assets, or even liabilities which are under the control of government. The idea of how the size of government should be approximated is not deep-seated. As we shall show, the choice of a relevant indicator differs across the economic literature.

At the very start of our discussion, it is worth mentioning that using aggregates for an assessment of government power and influence on the economy¹ is misleading irrespective of the specific methodological rules or choice of indicator. Aggregates keep many relevant aspects from observation. For instance, even a relatively small government, in accounting terms, might exert a broad influence over the allocation of scarce resources. The central banks can serve as a perfect example of nominally or statistically less important public institutions legally separated from the state exerting a far-reaching influence over any economy through the centralized monetary system.

So, there is inevitably a certain kind of discrepancy between the nominal size of government, in accounting terms, and its real size or powerfulness. Due to that, Higgs (1991) rightly points out that the dynamic of the government power cannot be inferred from government aggregates themselves. For example, government can substantially extend its power by adoption of a new law enforced by current government employees, thus potentially without any budgetary implications. Conversely, a drop in the expenditures does not necessarily go hand-in-hand with a shrink in coercive power. The interpretation of changes in monetary aggregates is thus ambiguous. Looking at the data only, this cannot itself reveal whether the extent of coercive power has been extended or not, this simply is kept from observation.

Macroeconomic figures should therefore be read with extreme caution. Even in an economy run by a statistically smaller government, economic freedom might be strongly restricted so that the citizens are not free to run their own small-size business or it is extremely costly to do so. Privately-owned factories can be forced to adjust their production in line with government desires and needs, as in war time, or to follow a politically correct pattern in recruitment. On the other side of the market, people can be restricted in their positions as consumers, having limited choices in what goods and services they can buy. In other words, it is the government authority over decision-making of its citizens rather than its command over financial resources which makes government big or small.

Let’s illustrate this point by one of the most provocative examples, the economy of Nazi Germany. Without question, the Nazi regime exercised control over virtually all economic resources, aligning the structure of production to the general government policy. Importantly, most industries in Nazi Germany were left in private hands.² Hitler didn’t see an urgent need for expropriation: Why nationalize industry when we can nationalize the people? Politicians, of course, do not need to own enterprises in order to exert pressure on them.

The Nazi regime illustrates that the public control over economic resources – a key feature of socialism (Mises 1922) – cannot be understood only in terms of direct ownership. Structure of production, level of prices or structure of investments is affected by regulatory acts of government and other regulators, such as central banks, legally entitled to exercise public control. This makes the private and public interest overlapping and blurs the borderline.³ We shall consider existing institutional environments when discussing the range of government activities and, more importantly, the size of government and the public sector.

The existence of legal restrictions on the use of private property had existed, of course, long before the Second World War and they have not ceased to exist since the war came to an end. Rather the reverse is true; the scope of legislation has been steadily growing since then. The Czech Republic, a member state of the European Union, is an example of how the scope of legislation can reach an ungraspable magnitude. In 2015, there were more than two million legal norms – laws, decrees, etc. – to which citizens are expected to obey. Figure 0.1 shows this development over the long term.

Figure 0.1 Regulatory acts in the Czech Republic, 1945–2010

Source: Ministry of Internal Affairs, Czech Republic.

It might at first sight appear surprising that the number of legislative acts has strikingly started to grow after the fall of the communist regime in 1989. It suggests that direct government ownership was presumably replaced, to some extent, by indirect public control through legislation. Expanding regulation has been further enhanced by supra-national bodies like the European Commission issuing many other pieces of legislation, after the Czech Republic joined the EU in 2004. As of 2015, the EU had contributed to this trend by adopting 40,000 legal acts or 62,000 international standards, which require compliance by companies operating within the EU and, of course, by its citizens.

Influence over economic decisions through regulation opens the discussion of where private control ends and public control begins. We might ask for example, provocatively, if the market for banking services is a genuine market or a pretend market setting tightly hedged by excessive regulation. Do the existing entry barriers into the banking industry, which protect current competitors from newcomers, square with our understanding of a true market? Are not the commercial banks only implementers of the monetary policy of central banks? These questions are not only intellectual exercises, but the way of answering them can have a considerable impact on consequent actions of authorities and, of course, statistics. Put another way, legislation and existing institutions necessarily affect statistics, its methodological concepts and ultimately the aggregates themselves.

Here we are approaching the key issue, which we shall discuss in depth: the statistical treatment of market behaviour. Conjectures and disagreements are guaranteed and inevitable. While left-leaning economists mostly follow a misleading view that identifies the market with private ownership only, right-leaning economists rather doubt that or even reject this view. Ambiguity and heated discussions between statisticians in this matter only mirror the wide difference of views across economic schools on the features of a truly market-based system.

In the first two chapters, we shall highlight the far-reaching impact of statistical data on economic research, theory and economic policy, in effect to answer the question why it is important to deal with this issue at all. In the third chapter, we shall undertake a short trip into the recent past to consider how methodology has evolved alongside the development in the institutional setting of modern societies. Arriving at the present, we shall discuss in Chapter 4 the current way of measuring government. In the next chapter, we shall confront the current approach with market-based economic theories enabling us to reveal weaknesses and limitations of the current statistical rules.

After this theoretical-methodological part, we shall examine the (re)calculations of key headline figures. First, we focus on the untouchable or even sacred indicator: gross domestic product (GDP) and attempt to reformulate and recalculate one of the great inventions of the twentieth century (Samuelson & Nordhaus 1999) to better fit with the government’s size measurement and an evaluation of the government’s fiscal condition. After that, we shall turn our attention to the definitions of government aggregates such as total expenditures, final consumption and debt. In the course of doing so, we do not leave the standard way of their computation intact. Modifying these aggregates, we recalculate numerators and denominators together to provide a different perspective on the size of European governments.

Before presenting the recalculated values, it is worth emphasizing the point that we can hardly ever arrive at precise figures; but, neither do the official macroeconomic statistics. Actually, the official figures released by government bureaus have always been and will continue to remain, only rough quantitative estimates of real economic phenomenon. This book purports to show that this official data is ill-defined in certain aspects, and therefore increases the risk of wrong conclusions by researchers, or inappropriate policy responses by economic authorities. We are thus not about to criticize mainstream economics and statistics, let alone statisticians, but to examine the uncertainty inherent in the quantifications, the soundness of theoretical grounds on which the methodology stands, and the appropriateness of rearranging data to arrive at a more realistic picture.

By recalculating the aggregates, we will not, of course, change their holistic nature.⁴ We will only show how different treatment of the market at the microeconomic level might bring macroeconomic figures into a sharply different perspective, and to outline possible future directions for the methodology to evolve to generate as accurate data as possible. Interestingly, even macrostatistics methodology itself has taken a number of steps toward a microeconomic approach. This general trend has been ongoing since the 1970s and is undoubtedly on the rise in the measurement of government size, notably by focusing on the institutional environment in which a particular producer is operating. This has opened up further discussions and investigations that we are going to consider.

In order not to get lost in terminology, a short note is worth making here. Normally, data-users don’t spend much time thinking of differences between the following terms: the state, government, the government sector, or the public sector. Indeed, they are often considered synonymous. The term state, the narrowest entity, historically covers a legislature, an executive (government) and a judiciary, operating under a state budget. Over the years, the state has established thousands of legally separated institutions with their own budgets and mandates to pursue the social and economic policy of government. The state, the government and the public sector are thus distinguished in the macroeconomic statistics covering a different range of units involved in the figures.

It is reasonable to go beyond the conception of the state in our analysis since it is impossible in our everyday lives to avoid using at least some public-sector facilities – roads, public transport, education, healthcare – which are invariably run by organizations not under the state budget themselves, or reflected in the accounting of the state in their entirety. In addition, the state extends its hold on the economy by way of grants, licences, contracts, tenders or even a delegation of legal authority to regulate certain market-segments.⁵ Consequently, even private companies are engaged in the provision of public goods and services where the main responsibility currently lies with public institutions.

All this considered, our analysis of the size of government cannot think in terms of the state or the state budget only. The border between all concepts has become considerably blurred and the size of the public and the government sector in its entirety virtually untraceable. Many economists and statisticians have never got past this situation and have already offered different views. Just for starters, Light (2003) suggests that by extending US figures on government employees to cover private contracting partners of government, the true size of government would jump from 12 million employees to nearly 17 million. Clearly, an underestimation of the government’s size can be substantial. We are at the beginning of an interesting journey.

CHAPTER 1

THE SIZE OF GOVERNMENT IN ECONOMICS

Let’s start with the question, why the discussion on the measurement of government should be important at all. Following fundamental changes in the field of mainstream economic theory in the first half of the twentieth century, the role of government in economic affairs has become one of the predominant concerns of economists. So, economic researchers have put a great deal of effort into finding out an optimal size or scope of government, which draws on mutual causal links between relevant macroeconomic aggregates like government expenditures and economic growth, or government indebtedness and inflation rates. Undoubtedly, statistics, that is the quantitative measurement of abstract economic concepts, is the sine qua non of empirical research.

To bring economic concepts into existence by way of appropriate quantifications is left almost exclusively to statisticians. Admittedly, after many decades of rapid development in economics and statistics, we have arrived at the point where economists rely on statistical quantifications, but are at the same time little aware of how these figures have been compiled.¹ On the other hand, most statisticians have little idea of all the purposes for which the result of their work is used. This situation is apparently further enhanced by a precipitous development on both sides, i.e. a growing complexity of the methodology of macroeconomic data and a broadening of its use in economic policy.

In the case of macroeconomic aggregates, we are not talking about monetary categories easily observable. As Morgenstern (1963) reiterates, statistical observations are based on the constructed, deliberately designed and invented theory. The compilation carried out by professional statisticians is a lengthy and highly complicated process. Lack of information and methodological complexity creates an environment in which an occurrence of measurement errors is inevitable. In his famous book, Morgenstern (1963) further enumerates several sources of error negatively affecting the accuracy of economic calculations. For the sake of our argument, chiefly the lack of clear definitions of the government sector, one of the sources of errors stated by Morgenstern, is of vital importance. Morgenstern rightly notes that weak and unsteady definitions cannot provide a stable ground for reliable quantifications.²

1.1 Economic theory

Macroeconomic textbooks abound with terms like government, fiscal expansion, fiscal restriction, government debt, but it is only exceptionally accompanied by any thorough reference to the substance of these terms. So, what is meant under the term government? Here we are referring to an entity whose actions generate effects for the economy through, in the words of modern macroeconomics, expansionary or restrictive fiscal policy.

The role of government is illustrated by the total of government consumption expenditures, the famous G. Through management of its consumption expenditures, government is supposed, especially according to Keynesian theory, to stabilize business cycles. As Keynesian theory holds, G is supposed to be a stabilizing part of GDP whereas this deliberate government policy is necessitated by presumed instability of investment expenditures (I). As the theory goes, if aggregate demand falls short of that level required to maintain full employment, government shall step in and counterbalance falling demand even at the cost of deficit. On the contrary, if aggregate demand threatens to cause price inflation by way of excessive employment, i.e. above the full level target, government shall drain income to cool down an overheating economy, i.e. to create a budget surplus.

The aggregate G thus stands in the centre of those considerations. We can illustrate this by using the famous neo-Keynesian Mundell–Fleming model which deals with the effects and efficiency of fiscal (and monetary) expansion under different circumstances like capital (im)mobility or fixed/flexible currency regimes. Because the model is based on the basic Keynesian model of output and income, it counts with government expenditures, which are treated as a part of total output (or income). As this theory holds, the fiscal expansion, i.e. growth in G, is effective in the short term if some specific circumstances are met, for example, if exchange rate is flexible and capital is perfectly mobile. A growth in planned expenditures (G) gives rise to a shift in output; because of a government action and multiplication of newly injected expenditures, the economy will find its new equilibrium at a higher level of economic activity and income. This is illustrated by Figure 1.1.

Figure 1.1 Impact of increase in government expenditure on interest rate and income

To make these concepts applicable, workable and testable, even this model needs to be loaded with data. There are, here, at least two methodological issues that go mostly unnoticed by economists or, to rephrase that, most economists are unconcerned with them. First, as Auerbach, Leibfritz and Kotlikoff (1999) point out, Keynesian economists take a cash flow measure of policy. That is to say, the fiscal expansion or restriction is achieved through cash expenditures or accumulation of cash in the hands of government. This cash-approach is deeply rooted in the key Keynesian assumption on the short-sightedness of people in economic affairs. Auerbach, Leibfritz and Kotlikoff further draw attention to the fact that the financial consequences of government behaviour are spread over time affecting also welfare of future generations. Even if Kotlikoff’s central point is different from ours,³ this brings us to the important issue of an accounting principle on which fiscal figures should be based.

For the sake of the quantifications of macroaggregates, relying on the cash approach would tell use little about actual economic development. Economic activity might be carried out no matter how resulting transactions are settled. Payments of invoices might be postponed, the existence of long-term contracts or the settlement might be executed through other means like barter exchanges. The national accounts methodology thus applies the accrual principle recognizing creation of values irrespective of the way of settlement. This supposedly makes the results more realistic. Nevertheless, this might pose difficulties for judging whether an economic boom is sustainable, or predominantly debt-driven, or accompanied by a