The Elgar Guide to Tax Systems 0857933884, 9780857933881

Citation preview

THE ELGAR GUIDE TO TAX SYSTEMS

The Elgar Guide to Tax Systems

Edited by

Emilio Albi University of Madrid (Complutense), Spain

Jorge Martinez-Vazquez Georgia State University, USA

Edward Elgar Cheltenham, UK • Northampton, MA, USA

© Emilio Albi and Jorge Martinez-Vazquez 2011 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2011926844

ISBN 978 0 85793 388 1 Typeset by Servis Filmsetting Ltd, Stockport, Cheshire Printed and bound by MPG Books Group, UK

Contents

List of contributors Acknowledgments List of abbreviations

vii ix x

Introduction Emilio Albi and Jorge Martinez-Vazquez 1

2

1

Tax systems in the OECD: recent evolution, competition, and convergence Vito Tanzi

11

Direct versus indirect taxation: trends, theory, and economic significance Jorge Martinez-Vazquez, Violeta Vulovic, and Yongzheng Liu

37

3

Individual income taxation: income, consumption, or dual? Robin Boadway

4

The challenges of corporate income taxes in a globalized world Emilio Albi

93

130

5

Wealth and wealth transfer taxation: a survey Helmuth Cremer and Pierre Pestieau

183

6

Value-added tax: onward and upward? Jorge Martinez-Vazquez and Richard M. Bird

218

7

The economics of excise taxation Sijbren Cnossen

278

8

The scale and scope of environmental taxation Agnar Sandmo

300

9

Financing subnational governments with decentralized taxes Roy Bahl

328

The administration of tax systems John Hasseldine

360

10

v

vi

The Elgar guide to tax systems

11

Political regimes, institutions, and the nature of tax systems Stanley L. Winer, Lawrence W. Kenny, and Walter Hettich

376

12

Tax system change and the impact of tax research Richard M. Bird

413

Index

457

Contributors

Emilio Albi Complutense University of Madrid, Spain Roy Bahl Andrew Young School of Policy Studies, Georgia State University, USA Richard M. Bird Professor Emeritus, University of Toronto, Canada Robin Boadway Queen’s University, Canada and CESifo, Germany Sijbren Cnossen CPB Netherlands Bureau for Economic Policy  Analysis,  University of Pretoria, South Africa, University of Maastricht, Netherlands, and Erasmus University Rotterdam, Netherlands Helmuth Cremer Toulouse School of Economics (GREMAQ, IDEI and IuF), France John Hasseldine Associate Professor of Accounting at Whittemore School of Business and Economics, University of New Hampshire, USA (from August 2011) Walter Hettich California State University, Fullerton, USA Lawrence W. Kenny University of Florida, USA Yongzheng Liu International Studies Program and Department of Economics, Andrew Young School of Policy Studies, Georgia State University, USA Jorge Martinez-Vazquez International Studies Program and Department of Economics, Andrew Young School of Policy Studies, Georgia State University, USA Pierre Pestieau CREPP, University of Liège, Belgium, and CORE Agnar Sandmo Norwegian School of Economics, Norway Vito Tanzi Former Director of the Fiscal Affairs Department of the IMF and former Undersecretary for Economy and Finance in the Italian government

vii

viii

The Elgar guide to tax systems

Violeta Vulovic International Studies Program and Department of Economics, Andrew Young School of Policy Studies, Georgia State University, USA Stanley L. Winer Carleton University, Canada, CESifo, Germany, and ICER, Italy

Acknowledgments

The papers in this volume were presented and discussed at the conference ‘Tax Systems: Whence and Whither’ organized by FUNCAS (the Spanish Savings Banks Foundation) and UNICAJA (an important savings bank in the south of Spain) and held in the city of Malaga (Spain) over a period of three days in September 2009, and later revised with the comments received at the conference. We are most grateful to Victorio Valle, General Director of FUNCAS and Braulio Medel, President of UNICAJA for their guidance and for graciously sponsoring the conference. We also want to thank José M. Dominguez-Martinez and José A. Antón for all the logistical and support work involved in the successful celebration of the event. All the authors at the conference benefited significantly from the comments received from discussants, session chairs, and others attending the conference. We are also grateful to all of them and their contributions can be found in the double special issue of Papeles de Economía Española, No. 125/126, 2010.

ix

Abbreviations

ACC ACE AETR AF ASE BBLR BEA BEIT BLU BVT CBIT CCCTB CCE CEN CEPII CESifo C-FS CIT CORE CREPP DEA DIT ECJ FDI FRT FUNCAS GFS GMM GREMAQ GST HT-SMEs ICBIT

allowance for corporate capital allowance for corporate equity average effective tax rate apportionment formula allowance for shareholders equity broad-based low-rate US Bureau of Economic Analysis business enterprise income tax best, linear, unbiased business value tax comprehensive business income tax common consolidated corporate tax base credit for corporate equity capital export neutrality Centre d’Etudes Prospectives et d’Informations Internationales Center for Economic Studies and Ifo Institute for Economic Research cash-flow statements corporate income tax Center for Operations Research and Econometrics Center of Research in Public Economics and Population Economics data envelopment analysis dual income tax European Court of Justice foreign direct investment flat rate tax Savings Banks Foundation of Spain Government Finance Statistics of the IMF generalized method of moments Le Groupe de Recherche en Economie Mathématique et Quantitative goods and services tax home taxation for SMEs CBIT with immediate expensing of investment x

Abbreviations xi ICER ICRG IDEI IDT IFRS IFS IMF IRAP ISS IVFE LSE MCPF METR MRS MRT MST OECD OLG OLS OT OTPR PCSE PE PNS RST SADC SUR TMTR UNICAJA VIES VIF WDI

International Center for Economic Research International Country Risk Guide Industrial Economic Institute inheritance and donation tax International Financial Reporting Standards Institute for Fiscal Studies International Monetary Fund Italian regional tax on productive activities (Imposta Regionale sulle Attività Produttive (Italy) tax on services (Imposto Sobre Serviços) in Brazil instrumental variable fixed effects large-scale enterprise marginal cost of public funds marginal effective tax rate marginal rate of substitution marginal rate of transformation manufacturer’s sales tax Organisation for Economic Co-operation and Development overlapping generations ordinary least squares optimal tax theory Office of Tax Policy Research panel corrected standard error permanent establishment public sector, not-for-profit activities and charities retail sales tax Southern African Development Community seemingly unrelated regression top marginal tax rates financial institution in Malaga, Spain VAT Information Exchange System variance inflation factor World Development Indicators

Introduction Emilio Albi and Jorge Martinez-Vazquez

Tax systems, how governments raise money to finance public expenditures, are central to the economic and political institutions and the overall day-to-day and long-term performances of every country. How much is raised in taxes to be spent on public services and infrastructure and how much is left to the private market to provide is eminently a political decision about which economics has little to say. However, taxes can cause important distortions on the behavior of economic agents, and the same exact amount of tax revenue can be raised with very different additional excess burdens to society depending on how tax systems are structured. In addition, all taxes can have significant effects on the distribution of income and even poverty levels depending on their final economic incidence, or who actually bears the burden in terms of reduced disposable income. These two aspects, the economic efficiency and the distributional equity of taxes, have been at the core of the theory and practice of public finance over the past several centuries, from the classical writings of David Ricardo and John Stuart Mill, to Richard Musgrave’s Fiscal Systems in 1969, to the influential reports produced by the Meade Commission in the UK and the Bradford Commission in the US in the late 1970s, to the most recent 2010/2011 comprehensive report from the Institute of Fiscal Studies in the UK, Reforming the Tax System for the 21st Century (the Mirrlees Review). Tax systems around the world have changed considerably in the past three decades. Although direct taxation continues to be mainly levied on income, personal and corporate income taxes are today quite different from what they used to be just a few decades ago. Social Security contributions have increased rapidly in importance, especially in the developed world, while wealth taxes have all but disappeared. Change has taken place at an even faster pace in indirect taxation where in most countries, with just a few exceptions, notably the US, the value-added tax (VAT) has become by far the most important indirect tax, accompanied still by an incipient use of environmental taxes that supplement classical excises, and a rapid decline in the importance of customs tariff taxes. Personal income taxes show flatter marginal rates than 30 years ago, with moderately rising average tax rates, and taxing capital income with generally lower rates than wage earnings, as is the case of the dual income 1

2

The Elgar guide to tax systems

tax approach. Corporate income taxes apply relatively lower statutory rates on broader bases than in the past. Internationally, there has been an increase in country competition and corporate taxation increasingly follows the source-based criterion, with limited gains in the coordination of corporate taxation. In practice, the powerful revenue potential of the VAT, sometimes dubbed a ‘money machine’, has been limited by design issues – multiple rates, narrow base, and so on – in developed countries and enforcement issues in developing countries. All these fundamental changes found in tax systems have been the result of powerful economic developments such as increased economic globalization observed in the last 30 years, but also the consequence of new political stances toward the role of the public sector, and quite possibly also the result of developments in the theory and practice of public finance. With this background, it seems clear that the time is ripe to take stock of the changes, to fully understand where we stand now on the subject of tax systems, and to consider what types of reforms are feasible and worth undertaking in the next decade. These perspectives and prospective views need to cover not only tax policy issues but also focus on the important challenges arising in the field of tax administration and enforcement, in the design of sub-national taxes, and in the analysis of the political economy of taxation. The papers in this volume are intended to shed light on and advance our understanding of these questions. In the following paragraphs we briefly describe the contributions of the different authors. Over the years, many tax innovations have taken place in OECD countries, and trends in the evolution of tax systems in these countries to a large extent provide an advanced look into what we would expect to see happening in middle-income countries and even in some low-income countries in the coming decades, if not years. Chapter 1 by Vito Tanzi reviews the changes in OECD countries for the last four decades. He finds important increases in tax levels, significant and generally positive changes in the structure of VAT and personal and corporate income taxes and VAT, and also changes in tax composition. In this last respect, Social Security contributions and the VAT have continued to gain significance over other taxes. Tanzi also detects several trends that will make things harder for tax policy-makers, including increasing trends towards tax decentralization and heightened international tax competition, accompanied by tax evasion and avoidance schemes. However, for Tanzi, the most important challenge facing tax systems in the OECD is the strong lobbying forces working to increase tax complexity, which, in the end, work as a ‘hidden and regressive tax’, imposing high costs to society. One of the most fundamental policy decisions to be made in the design of tax systems is that of direct–indirect tax mix, or how much to rely on

Introduction

3

direct taxes, such as personal and corporate taxes, versus indirect taxes, such as VAT and excises. Although this choice is generally framed as a choice of the degree of progressivity or distributional impact of tax systems, the choice actually can have many other important implications, including economic growth, macro-stability, and the flow of foreign direct investment (FDI). In Chapter 2 Jorge Martinez-Vazquez, Violeta Vulovic, and Yongzheng Liu revisit this theme by first looking at some definitional issues, since what is a direct tax versus an indirect tax is not always clear, and concluding that, based on current theory, there is no superiority of one type of taxation over the other. Using data for 116 developed and developing countries for the years 1972–2005, these authors find that the direct to indirect ratio has clearly increased, mostly due to increases in Social Security contributions in developed countries and to decreases in customs tariff revenues in developing countries. They also document a policy trade-off. Higher reliance on indirect taxation supports faster economic growth and increases a country’s competitiveness for FDI, but this higher reliance on indirect taxation also reduces the effectiveness of automatic stabilizers for macro-policy and weakens the strength of income redistribution policies. The personal income tax has been for many decades a key component of tax systems in developing countries, raising significant revenues, on average 10 percent of GDP, but also very useful in playing significant roles in the pursuit of redistribution objectives (including the use of negative income tax credits), and macroeconomic stabilization objectives. This role has been much more subdued in developing countries, where this tax raises proportionally much less revenue (between 1 and 2 percent of GDP) mainly because of the lower availability of automatic withholding systems and more widespread informality and evasion. But even in developed countries the personal income tax has gone through a lot of questioning and actual transformation in recent decades. In Chapter 3, Robin Boadway reviews the arguments for and against the choice of income or consumption as the base for direct individual taxation and concludes that nearly all the arguments ‘based on administrative ease, efficiency, and equity conspire against comprehensive income as an ideal form of taxation’. A consumption base scores well in many respects, especially in administrative matters, but capital income taxation is supported by solid arguments based on efficiency, equity, or political economy. This explains why the alternative of the dual income tax (or, more generally, schedular taxation) with lower flat rates for capital income is becoming increasingly attractive as a useful compromise for this tax. Boadway also reviews the thorny issue of the degree of progressiveness of individual income taxes, a matter that depends not only on rates but also on base composition,

4

The Elgar guide to tax systems

and the reasons for limiting high income tax rates. A number of special issues, which clearly complicate the implementation of this tax, are also addressed in this chapter including the relationship between individual and corporate income tax, coordination with sub-national income taxes, the treatment of pensions, human capital investment, wealth transfers, and the tax treatment of family size. The corporate income tax has been the constant complement of the personal income tax, acting as a withholding mechanism for some forms of capital, but always with more volatile revenue performance, as its base of corporate profits is more sensitive to the business cycle. In recent times, this tax has been under attack from globalization trends, heightened international capital mobility, income shifting, and cross-country tax competition. In Chapter 4 Emilio Albi makes a good case that the corporate income tax is here to stay and even prosper more than in recent times but in adapted forms, especially focusing on the experience of 15 European Union (EU) countries. Complexity is the unavoidable rule of corporate income tax structures, and therefore reforms have steered toward less distortionary and costly solutions, within the established constraints, as opposed to more perfect, albeit impracticable, alternatives. The classical questions surrounding the corporate income tax such as the ultimate incidence of the tax, distortions of investment and financing decisions, discouragement of dividends distribution, business form, avoidance scheme, or compliance and enforcement costs still have as much relevance today as many decades ago. The future likely holds a continuing trend in base broadening and further tax rate competition with new targets for tax credits and taxation of shareholders; international coordination even at the EU or OECD levels is far more problematic, as are the prospects for radical reform. One form of direct taxation that has experienced a radical decline in recent times is that of wealth taxes, and more so in the case of annual taxes on net wealth than in the case of the transfer of assets (as in inheritances or donations). Besides raising little revenue, higher capital mobility and competition may have played a role, but unpopularity and ideology have probably been key. In Chapter 5, Cremer and Pestieau focus on the efficiency and equity properties of these taxes from the normative perspective of optimal taxation and in the context of the entire tax system, dominated by income and commodity taxation. This is important because wealth taxes have been traditionally thought of as complementary to income taxes. Cremer and Pestieau show that a crucial factor in designing the tax structure, according to the desirable efficiency and equity properties of the different tax instruments, is the motive underlying wealth accumulation and transfers. Motivations are important because they determine the

Introduction

5

way individuals react to taxation, and therefore what the excess burdens involved may be. Among the traditional motivations, they consider consumption smoothing (retirement, children’s education, precautionary motives), prestige or status, intergenerational transfers triggered either by pure altruism, imperfect altruism (joy of giving) or by exchange motives (e.g., the so-called strategic bequests). Uncovering the true motivations of taxpayers unfortunately provides a difficult basis for policy design in practice. This may help explain the contemporary lack of appeal of different forms of wealth taxation. The single most important transformation of tax systems around the world in the last several decades has been the massive adoption of an invoice-credit consumption-based VAT. In a relatively short period of time the VAT has become the workhorse of most tax systems around the world. And yet its rather stellar performance has not been free of problematic issues. In Chapter 6, Martinez-Vazquez and Bird provide an analysis of how well the VAT has actually performed, discuss several important policy issues related to the structure of the VAT, and take a closer look at some administrative challenges of the VAT, in particular the issue of fraud. To analyze VAT performance, Martinez-Vazquez and Bird examine the determinants of three different measures of VAT efficiency (VAT efficiency ratio, C-efficiency ratio, and VAT gross collection ratio) using a sample of 107 countries, between 1990 and 2008 and find the share of agriculture in GDP, urbanization, and tax morale to be significant determinants of VAT collection efficiency. They also find that on average the adoption of the VAT led to an increase in revenue collections by 12 percentage points of GDP but with the effect being most significant in developing countries, with positive effects in the collection of income taxes and very substantial substitution for collections from customs duties. In most countries VAT may still be the most economically desirable and administratively effective way to increase revenue collections, but the reality is that the VAT is not always well designed, not that well administered, and that the political economy issues continue to make VAT reform difficult. Excise taxes are among the oldest forms of taxation and continue to have a significant presence in most tax systems not only as reliable sources of revenues but also as instruments of social engineering for addressing different assortments of negative economic externalities. In his Chapter 7 Sijbren Cnossen examines the rationale of excise taxation by reference to its non-revenue objectives, compares the effectiveness of excise duties (with specific or ad valorem rates) vis-à-vis regulations and permits in the pursuit of those non-revenue objectives, and reviews the issues of discrimination, coordination, and earmarking that are often connected with excise taxation. Although the adoption of excise taxes is directly connected

6

The Elgar guide to tax systems

to their ease of administration, they also have the economic rationale of falling on products with inelastic demands and therefore represent low distortions or excess burdens. But this low elasticity, although a cause of revenue productivity, also means lower effectiveness in discouraging behaviors from negative externalities – using tobacco products, drinking alcoholic beverages or using motor fuels. Often, regulations (smoking bans, breath tests, etc.) can be a much more appropriate alternative. Cnossen also points out that using excise-type charges on luxury goods (with high income demand elasticities) for making the tax system more progressive is much less justified or effective. Raising in some ways similar issues to those associated with excise taxes, environmental or ‘green’ taxes remain the unrealized promise of most tax systems. In Chapter 8, Agnar Sandmo reminds us that, although with clear roots in Pigou’s contribution to the theory of externalities, environmental taxation is a relatively recent innovation in tax systems and largely the creature of academic researchers. So it may take time to overcome practical issues, such as the selection of the ‘right’ bases and measurement of polluting activities, and political economy issues, such as perceptions about the distributional impact of green taxes. Sandmo also reminds us of the usefulness of analyzing environmental taxes in the overall context of tax systems and public expenditures and that the ‘double dividend’ of environmental taxes (they improve the environment and also generate revenue that can be used to reduce other taxes with significant distortionary effects) is not always assured. The last part of this chapter examines the relationship between ‘green’ taxes and regulatory and administrative aspects of environmental protection, the interactions between ‘intrinsic’ (moral) and ‘extrinsic’ (tax or regulatory) incentives to act in a socially rational way in environmental issues, the global aspects of green taxes, and the political economy of environmental policy. As many more countries have embarked on fiscal decentralization reforms over the last three decades, the tax systems of these countries have increased in complexity because of the necessary assignment of taxing powers to sub-national government units, raising issues of tax competition, vertical externalities on tax bases, and so on. However, the scope and quality of those assignments vary considerably across countries. In Chapter 9, Roy Bahl examines the case for assigning taxing powers to sub-national governments, and asks whether the international trend in tax assignments is in step with what economists have prescribed. While sub-national taxes in OECD countries average about 8 percent of GDP, in developing countries it is less than one-half as much and there is much higher dependency on intergovernmental transfers. The core benefit from increased tax autonomy is the enhanced accountability of sub-national

Introduction

7

government officials toward the needs and preferences of their constituencies and increased fiscal responsibility. Although it is not entirely clear how much revenue autonomy is needed to realize those benefits, it would seem clear that many tax systems fall short. The choice of decentralized taxes is much more limited in developing countries than in industrialized countries; for example, personal income taxation of any form can rarely be found at the sub-national level in the developing world. Given that corporate income tax is a clearly bad choice for assignment to sub-national governments, and that the sub-national VAT can be too complex, Bahl explores other likely candidates for providing sub-national tax autonomy beyond the already well accepted property tax and vehicle taxes. The list of good possible candidates is short: destination-based excises and perhaps residence-based payroll taxes. Tax systems can be thought of as the interaction of two fundamental components: tax policy and tax administration. And it is typically the latter that is the most difficult to get right. There is wide consensus among taxation experts that, without an effective and efficient tax administration, no tax systems can function adequately regardless of the quality of the design of tax policy and that this design must always pay close attention to administrative feasibility. In Chapter 10, John Hasseldine reviews the recent evolution and present state of tax administration theory, discussing the separation of operational tasks (audit and risk evaluation, tax payer services, etc.) and the broader, also important, issues of internal management (organizational matters, personnel and information technology management, strategic policy formulation). In the review of the current context of tax administration, especially in Europe, Hasseldine finds increasing emphasis on compliance and risk management issues, greater reliance on the cooperation and service paradigm, and the important role played by the systematic measurement of performance and the sharing of best practices. Regarding this latter, the roles of the OECD and the IMF have been fundamental. Improvements in the future will come through the continued sharing of information and knowledge between tax agencies and from providing more attention to cost-effective management and the successful introduction of technology. Tax policy reform can easily be one of the most contentious political issues any country may face. Most of the tax systems we observe are the result of complex, highly bargained political equilibriums. And one way to get tax laws approved is to give something to many different interest groups; hence, the complexity we observe in most modern tax systems and the importance of political economy and institutional considerations in the formulation of tax policy. In Chapter 11, Winer, Kenny, and Hettich examine the relationship between the types of political regimes

8

The Elgar guide to tax systems

and institutions and the nature of tax systems. To that end they utilize a simple model for explaining tax system outcomes as a political equilibrium, where they show that all types of political regimes will have to face different tax rate–revenue relationships for particular tax bases, although these may assume different shapes for dictatorial and democratic societies. Dictatorships are more likely to emphasize enforced rather than voluntary compliance. Winer et al. do in fact find empirical support for the proposition that democratic regimes rely more heavily on taxes requiring voluntary compliance, such as income taxes, suggesting that the degree of consent may play a vital role in the nature of observed tax systems. The type of institutions within democratic systems also seems to matter. Countries using proportional electoral systems make heavier use of Social Security, payroll, and domestic trade taxes while countries using majoritarian electoral rule rely more on individual income, corporate, and international trade taxes. The reasons behind those results await further research in the future. The last question is whether economists in volumes like this one or in other forums can actually influence the shape and future of tax systems. The main issue discussed by Richard Bird in Chapter 12 is whether tax systems in the last few decades have been much influenced by tax research or, from an opposite mirror perspective, whether the economic literature actually has focused on the problems and issues that really matter to policy-makers and that finally shape tax systems. This is an issue that has been discussed recently in different forums, as if there were a sudden need to evaluate the effectiveness of academic economists’ efforts now that society may be turning away from the profession in the aftermath of the global financial crisis. Bird’s scope of tax systems is wide: OECD countries, where research and most advice comes from, and, perhaps for the most part, non-OECD tax structures. It is the experience in the non-OECD world that may provide ‘a particularly clear test’ of whether economists’ advice has been effective. From an optimistic viewpoint, it is true that the influence of tax theory may be found in taxation changes in recent decades. Examples could be the international expansion of VAT, the incipient interest in environmental taxation, the sizeable reduction of income tax rates and the broadening of tax bases, the spread of dual income taxes with lower flat rates on capital income, or even flat rate income taxes. But as Bird puts it, even in these cases ‘fiscal history suggests that much tax research takes place precisely because countries change tax policies for their own reasons. Often, tax researchers are not so much leading the reform elephant as mopping up behind it’. All this ultimately means a more humble but nevertheless useful role for economists helping to shape tax systems. Those interested in improving tax systems should

Introduction

9

focus not on the short-term political game within which policy decisions are inevitably made in all countries but rather on the long-term game of building up institutional capacity, both within and outside of governments, to articulate relevant ideas for change, to collect and analyze relevant data, and to assess and criticize the effects of such changes as they are made.

REFERENCES Mirrlees, J. et al. (eds) (2010/2011), Reforming the Tax System for the 21st Century (2 vols; Dimensions of Tax Design and Tax by Design), Oxford: OUP for IFS. Musgrave, R.A. (1969), Fiscal Systems, New Haven, CT: Yale University Press.

1

Tax systems in the OECD: recent evolution, competition, and convergence* Vito Tanzi

1

INTRODUCTION

The popular view is that taxes are imposed to finance the activities of governments and public sectors. They are the prices paid for the (generally free) services that citizens receive from their governments. This view remains popular and finds its way into the writings of economists and tax experts. However, with the passing of time it has become a little detached from reality for at least two reasons. First, governments finance at least part of their activities with non-tax revenue such as public debt, sale of assets, earnings from publiclyowned natural resources, incomes from publicly-owned monopolies, fees, fines, and so on. Second, tax policies have increasingly been used to promote non-revenue objectives. These may be to stimulate or, at times, slow down some activities, enterprises, regions, social behavior, and so on. Revenue needs, to finance public sectors activities or government objectives, change over time, for the same countries, and, at a given time, are likely to be different across countries. The broader the economic role of the state, the more revenue a country’s government will need at a given time. Much of this revenue will come from taxes. Thus, both the share of total tax revenue into gross domestic product – T/GDP – and the share of public spending – G/GDP – will have to change. At the same time, reflecting the countries’ preferences as well as their economic and other characteristics, the structure of their tax systems will also change. In this chapter the focus of attention will be on the 30 countries that belong to the grouping that goes under the name of the Organisation for European Co-operation and Development (OECD). This international organization started as a European entity but over the years it has expanded its membership to include several countries from other continents. Its current membership includes countries from all continents with the exception of Africa.

11

12

The Elgar guide to tax systems

2

TAX LEVELS IN OECD COUNTRIES, 1965–2007

In this section we focus on the evolution of tax levels in OECD countries between 1965 and the present. The latest year for which data are now available is 2007, thus, we are able to cover a 42-year period during which many significant changes took place. We shall focus our discussion first around Table 1.1, which provides data on tax levels for all 30 OECD countries over the period. For countries that joined the OECD in later years, such as Mexico, Korea, and the countries from East Europe (Czech Republic, Hungary, Poland, and the Slovak Republic), the data are available only for later years. In 1965, the highest tax levels were those of Sweden (35.0 per cent of GDP), France (34.1 per cent of GDP), and Austria (33.9 per cent of GDP). A few other countries had T/GDP ratios of 30 per cent or higher. These were the Netherlands, Germany, Belgium, UK, Finland, and Denmark. The lowest levels were those of Turkey (10.6 per cent of GDP), Spain (14.7 per cent of GDP), Portugal (15.9 per cent of GDP), Switzerland (17.5 per cent of GDP), Greece (17.8 per cent of GDP), and Japan (18.2 per cent of GDP). It is difficult to conceive today how the governments of these countries managed with such low tax ratios. For the whole group of OECD countries the unweighted average, for the share of total taxes to GDP, in 1965 was 24.2. By 2007 only Mexico and Turkey had tax ratios lower than the 1965 average. After 1965 the changes in tax ratios were mostly upwards, at least until recent years. In recent years there is some indication that tax levels started to go down; or, at least, they stopped going up. We shall return to this point below. By the year 2000 taxes had gone sharply up in the majority of the OECD countries and especially in Spain where the share of total taxes into GDP had increased from 14.7 per cent to 34.2 percent, more than in any other country in the table over that period. The unweighted average tax level for the OECD increased by 11.9 per cent of GDP, reaching 36.1 per cent in 2000, above any country’s level in 1965. In some countries, the T/GDP had reached extraordinarily high levels. For example, in Sweden it was almost 52 percent, probably a record in the history of the world. In four other countries (Denmark, Finland, Belgium, and France) the T/GDP had come close to, or had exceeded, 45 per cent of GDP. Excluding a few countries with different economies or traditions, such as Mexico, Korea, and Turkey, and, perhaps some of the Nordic countries that had to finance expensive welfare states, the differences in tax levels among the countries of Europe had been much reduced, showing a significant degree of convergence in tax levels over this period. Before leaving this discussion of tax levels in OECD countries, it may be

Tax systems in the OECD 13 Table 1.1

Total tax revenue as percentage of GDP, 1965–2007

Country

1965

1975

1985

1990

1995

2000

2005

2006

2007 Provisional

Canada Mexico United States Australia Japan Korea New Zealand Austria Belgium Czech Republic Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Luxembourg Netherlands Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom OECD Total

25.7 n.a. 24.7

32.0 n.a 25.6

32.5 17.0 25.6

35.9 17.3 27.3

35.6 16.7 27.9

35.6 18.5 29.9

33.4 19.9 27.3

33.3 20.6 28.0

33.3 20.5 28.3

21.0 18.2 n.a 24.0 33.9 33.1 n.a.

25.8 20.9 15.1 28.5 36.7 39.5 n.a.

28.3 27.4 16.4 31.1 40.9 44.4 n.a.

28.5 29.1 18.9 37.4 39.6 42.0 n.a.

28.8 26.8 19.4 36.6 41.2 43.6 37.5

31.1 27.0 23.6 33.6 42.6 44.9 35.3

30.8 27.4 25.5 37.5 42.1 44.8 37.5

30.6 27.9 26.8 36.7 41.7 44.5 36.9

n.a. n.a. 28.7 36.0 41.9 44.4 36.4

30.0 30.4 34.1 31.6 17.8 n.a. 26.2 24.9 25.5 27.7 32.8 29.6 n.a. 15.9 n.a.

38.4 36.5 35.4 34.3 19.4 n.a. 30.0 28.7 25.4 32.8 40.7 39.2 n.a. 19.7 n.a.

46.1 39.7 42.8 36.1 25.5 n.a. 28.2 34.6 33.6 39.5 42.4 42.6 n.a. 25.2 n.a.

46.5 43.5 42.0 34.8 26.2 n.a. 30.9 33.1 37.8 35.7 42.9 41.0 n.a. 27.7 n.a.

48.8 45.7 42.9 37.2 28.9 41.3 31.2 32.5 40.1 37.1 41.5 40.9 36.2 31.7 n.a.

49.4 47.2 44.4 37.2 34.1 38.0 37.2 31.7 42.3 39.1 39.7 42.6 31.6 34.1 33.8

50.7 43.9 43.9 34.8 31.3 37.2 40.7 30.6 40.9 37.8 38.8 43.5 32.9 34.7 31.8

49.1 43.5 44.2 35.6 31.3 37.1 41.5 31.9 42.1 35.9 39.3 43.9 33.5 35.7 29.8

48.9 43.0 43.6 36.2 n.a. 39.3 41.4 32.2 43.3 36.9 38.0 43.4 n.a. 36.6 29.8

14.7 35.0 17.5 10.6 30.4

18.4 41.2 23.9 11.9 35.2

27.6 47.3 25.5 11.5 37.6

32.5 52.2 25.8 14.9 36.1

32.1 47.5 27.7 16.8 34.5

34.2 51.8 30.0 24.2 37.1

35.8 49.5 29.2 24.3 36.3

36.6 49.1 29.6 24.5 37.1

37.2 48.2 29.7 23.7 36.6

24.2

29.4

32.7

33.8

34.8

36.1

35.8

35.9

n.a.

Note: n.a.: indicates not available. Source:

OECD: Revenue Statistics, 1965–2007: 2008 Edition (OECD, 2008).

worthwhile to call attention to a recent trend that has attracted less attention than it deserves. This is the inversion in the upward movement of tax levels that had taken place until the end of the millennium. Until around the year 2000 the trend in tax levels had been clearly and continually

14

The Elgar guide to tax systems

upward. However, in the more recent years, the earlier, uniform, upward movement came to an end. This change can be shown, as in OECD (2008),1 by taking several periods, 1965–75, 1975–85, 1985–95, 1995–2006, and illustrating graphically the increases and the decreases in T/GDP for each country and for each of these periods. The OECD charts show that, as one moves from the earliest to the latest period, the number of countries that show some reduction in T/GDP increases. There is, perhaps, a more powerful way to show the change in this trend. It is by taking, for each country, the year in which the T/GDP reached the highest level and comparing that level with the level reached in the latest available year. This is a better approach because it recognizes that the change in trend in different countries may occur at different times, though mostly around 1998. The results of this exercise are shown in Table 1.2. The second column shows the highest tax levels reached and the years when those levels were reached. The third column shows the most recent levels, generally for 2007 except for a few countries for which only data for 2006 are available. The last column (column 4) shows the differences between the highest levels reached by the T/GDPs and those reached in the most recent year. 1998 was the average year when the highest levels of T/GDP were reached. However, the highest levels were reached earlier by some countries (New Zealand, Ireland, the Netherlands, Norway, Sweden, and to some extent the UK), and later by other countries (Spain, Portugal, Mexico, Italy, and Iceland). The reduction in tax levels by the first group were generally connected with major economic reforms aimed at reducing the spending role of the state in these countries. The earlier increases in the other group were mainly connected with attempts at enlarging the spending role of the state. As shown by Table 1.2, 24 countries, out of 30, had in 2007 levels of taxation lower than the highest levels reached in earlier years. Six countries had either higher levels in 2007 than in the past, or the same levels. The largest reductions in T/GDPs were shown, in decreasing order by the countries in Table 1.3. The only country showing a significant increase in T/GDP in recent years was Korea, which, as shown by Table 1.1, had had an unusually small tax level. It had also gone through a major financial and economic crisis in the late 1990s. It came out of that crisis with a high fiscal deficit and significant expenditure needs. It should be noted that several of the previously high tax countries reduced the tax level in recent years. In 2007, the highest tax levels in OECD countries were in Denmark (48.9 percent) and Sweden (48.2 percent). The lowest levels were in Mexico (20.5 percent) and Turkey (23.7 percent). There is of course no theory that can tell us

Tax systems in the OECD 15 Table 1.2

Total tax revenue in selected years (percentages of GDP)

Country

1965

Highest Level and Year Reached

2007 Provisional

Level in 2007 Less Highest Level

Canada Mexico USA Australia Japan Korea New Zealand Austria Belgium Czech Republic Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Luxembourg Netherlands Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom Average

25.7 n.a 24.7 21.0 18.2 n.a 24.0 33.9 31.1 n.a 30.0 30.4 34.1 31.6 17.8 n.a 26.2 24.9 25.5 27.7 32.8 29.6 n.a. 15.9 n.a. 14.7 35.0 17.5 10.6 30.4 25.6

36.7 20.6 29.9 31.1 27.3 24.6 38.0 44.0 45.2 37.8 50.1 47.2 45.1 37.2 35.9 45.7 41.5 36.8 43.3 39.4 45.5 44.5 38.8 36.6 36.5 37.2 52.2 30.0 26.1 38.1 38.1

33.3 20.5 28.3 30.6a 27.6a 28.7 36.0 41.9 44.4 36.4 48.9 43.0 43.6 36.2 31.3a 39.3 41.4 32.2 43.3 36.9 38.0 43.4 33.5a 36.6 29.8 37.2 48.2 29.7 23.7 36.6 34.4

−3.4 −0.1 −1.6 −0.5 +0.3 +4.1 −2.0 −2.1 −0.8 −1.4 −1.2 −4.2 −1.5 −1.0 −4.6 −6.4 −0.1 −4.6 0.0 +0.5 −7.5 −1.1 −5.3 0.0 −6.7 0.0 −4.0 −0.3 −2.4 −1.5 −3.7

(1998) (2006) (2000) (2000) (2001) (2004) (1989) (1998) (1999) (2004) (1999) (2000) (1999) (2000) (1996) (1993) (2006) (1988) (2007) (1998) (1987) (1986) (1993) (2007) (1998) (2007) (1990) (2000) (2001) (1986) (1998)

Note: a. 2006. Source:

Adapted from Revenue Statistics, 1965–2007: 2008 Edition (OECD, 2008).

what should be the optimal level of taxation for a country. That level would depend on (1) how well governments use the tax revenue; (2) how good are the tax laws used to collect the revenue; (3) how good is the tax administration; and (4) how citizens react to the inevitable disincentive

16

The Elgar guide to tax systems

Table 1.3

Countries that experienced the largest reductions in taxes

Netherlands Slovak Republic Hungary Poland Ireland Greece Finland Sweden Canada Turkey Austria New Zealand

−7.5 −6.7 −6.4 −5.3 −4.6 −4.6 −4.2 −4.0 −3.4 −2.4 −2.1 −2.0

Source: See text and Table 1.2.

effects that high tax rates generate. Danish taxpayers are likely to react differently from Mexican or even US taxpayers. Today the name of Keynes is often called on to justify high public spending and high tax levels. However, it may be worthwhile to mention that in 1944, in a letter sent by Keynes to the then influential Australian economist Colin Clark, Keynes agreed with Clark, who had argued, in a publication, that 25 per cent of GDP was probably the maximum level of taxation that the citizens of countries would tolerate, before various psychological and economic reactions against high taxes would be set in motion.2 As Table 1.1 shows, with the exception of Mexico and Turkey, countries are now well above that maximum tolerable level. I have argued elsewhere that with a spending of 30 per cent of GDP, governments should be able to promote most of the objectives that can be legitimately and efficiently assigned to them (see Tanzi and Schuknecht, 2000).

3

EVOLUTION OF TAX STRUCTURES

The previous section addressed the evolution of tax levels in OECD countries for the period from 1965 to the present and showed the remarkable increases in those levels that occurred over the period in the majority of OECD countries. It can be expected that the changes in tax levels would be accompanied by significant changes in tax structures. The evolution of tax structures is the focus of attention in this section. Table 1.4 shows the shares of the major tax categories in total tax revenue over the 1965–2006 period. Because the taxes are aggregated in

Tax systems in the OECD 17 Table 1.4

Shares of major tax categories in OECD, 1965–2006 (percentages of total taxes) 1965

1975

1985

1995

2006

Personal income tax Corporate income tax Social Security contributionsa (Employee) (Employer) Payroll taxes Property taxes General consumption taxes Specific consumption taxes Other taxes

26 9 18 (6) (10) 1 8 14 24 1

30 8 23 (7) (14) 1 6 15 18 0

30 8 22 (7) (13) 1 5 16 16 1

27 8 25 (8) (14) 1 5 18 13 3

25 11 25 (9) (15) 1 6 19 11 3

Total

100

100

100

100

100

Note: a. Includes contributions by independent workers. Source:

OECD, Revenue Statistics, 1965–2007: 2008 Edition (OECD, 2008).

major categories, the table gives an impression of more stability in the structures than in fact took place. However, as we shall show, this is largely a statistical illusion because, as with icebergs, much of the action may take place below the visible part of tax systems. One of the conclusions of this chapter will be that economists have concentrated too much on broad or macro-variables and have ignored the many micro- and often largely invisible changes that go on continually below the visible parts of tax systems and that, over many years, may significantly change the systems. In any case, the macro-changes are described first. The main changes shown by Table 1.4 are the following. First is the decline in importance of the personal income tax since the decades of the 1970s and 1980s. The other two categories that also declined in importance over the period are ‘specific consumption taxes’, which saw their share fall from 24 per cent of the total in 1965 to 11 per cent in 2006, and ‘property taxes’, which declined from 8 per cent in 1965 to 5–6 per cent in later years. Categories that increased significantly are Social Security contributions and general consumption taxes. These two categories, combined, saw their share of total taxes grow from 32 per cent in 1965 to 44 per cent in 2006. Because these were growing shares of growing total taxes in GDP, the importance of this change is particularly significant. A few general comments may be helpful in explaining some of the above changes. First, the increase in the share of ‘general consumption taxes’ was

18

The Elgar guide to tax systems

largely the outcome of the introduction of the value-added tax (VAT) in the 1960s and later decades. The VAT replaced, in many OECD countries, the turnover taxes and some specific excises. The introduction of the VAT was one of the truly major ‘technological innovations’ in the tax systems for both OECD and developing countries. It provided an alternative to the personal income taxes that had been popular and had been pushed by US tax advisors in their technical assistance to countries. The VAT was first introduced in France as early as 1948 and then, in the 1960s, in Denmark (1967), Germany (1968), the Netherlands and Sweden (1969), and, in the 1970s, in Luxembourg and Norway (1970), Belgium (1971), Ireland (1972), Austria, Italy, and the UK (1973). In Spain it was introduced later (in 1986).3 Except for the United States, all OECD countries now use VATs that generate revenues that often exceed 7 per cent of GDPs and in some cases (Denmark, Iceland, Norway) approach 10 per cent of GDPs. These latter countries use the highest nominal rates. In Europe the ‘standard’ VAT rate tends to be around 20 percent. A few countries (Belgium, Denmark, Finland, Iceland, Ireland, Norway, and Sweden) have higher rates. A few other countries (Greece, Spain, Switzerland, and the UK) have rates somewhat lower. The highest rate now in use in OECD countries is 25 per cent (Denmark, Norway, and Sweden). Many countries apply reduced rates on some basic products believing that this helps poorer families. Tax experts have been generally skeptical regarding these claims and have supported VATs with single rates. Administratively a VAT with a single rate on the broadest possible base is considered preferable. The VAT has provided countries with a powerful instrument to raise revenue but also with a potentially very good instrument to pursue stabilization policy, a fact that has not attracted attention. Unlike personal income taxes that generally include many features that would need to be assessed when making policy changes, VAT revenue can be changed by modifying a single feature (its rate). Furthermore, the change can have an immediate effect on revenue because the revenue from the VAT is not subjected to any significant collection lag as is the case for personal and corporate income taxes.4 The effect on revenue of a VAT rate change is thus almost immediate. Another desirable feature of the value-added tax is that it lends itself less easily to ‘social engineering’ on the part of policy-makers bent on influencing the behavior of citizens. ‘Tax expenditures’ and other tax preferences can be more easily promoted through income taxes. However, the campaign on the part of the French government to reduce the VAT rate applied to restaurants, on the argument that it would stimulate employment, may be interpreted as an example of a ‘tax expenditure promoted through the VAT’. The same is for the zero-rating of food on the part of the British.5

Tax systems in the OECD 19 In recent years, because of the elimination of customs controls on trade between countries belonging to the European Union, there has been increasing concern and evidence of tax evasion and, especially, of tax frauds linked to fake exports within EU countries. The increasing use of Internet commerce, combined with the increasing trade in virtual products that do not have a physical content has also facilitated tax evasion. These concerns have raised questions about one of the claimed virtues of the VAT, namely its self-enforcing quality. Several studies of various countries have attempted to estimate tax evasion for the VAT (see Ceriani, 2009). The growing globalization of the world economies, and the fall in the share of GDPs accounted for by the sale of large industrial enterprises, have contributed to the difficulties encountered in the use of value-added taxes. Still, the VAT remains one of the important workhorses of tax systems and is likely to remain so. Let us now turn our attention to the personal income tax. There was a time, especially in the 1950s and 1960s, when the personal income tax, especially in its ‘global, comprehensive and progressive tax’ version was seen as the fairest and the best of all taxes. At the time taxpayers’ surveys in some countries indicated the strong preference that citizens had for this tax while tax experts promoted its comprehensiveness by attempting to include in its base all incomes, including unrealized capital gains. The income tax could be made as equitable as desired through its progressive rates; it could allow some governments (and especially those of AngloSaxon countries) an alternative to high public spending, through the provision of ‘tax expenditures’. At that time there were almost no studies that had quantified negative incentive effects that might be attributed to high tax rates. Furthermore, most incomes received by citizens were from domestic sources and, had a physical content. Thus, they were, presumably, easier to ascertain and tax. Phenomena such as tax evasion (both domestic and foreign) and underground economic activities and tax havens had not yet attracted the attention of economists and policymakers.6 Finally, conservative economists and policy-makers had not acquired the influence that they would acquire in later years. With the passing of time, and starting in the 1970s, attitudes vis-à-vis highly progressive, personal income taxes started changing and became progressively more negative.7 Also, the ongoing process of globalization started to have an impact on tax systems. Economists started to find growing evidence of tax evasion and of underground economic activities stimulated by high tax rates (see Tanzi, 1980). Questions started to be asked as to whether the existing personal income taxes were truly ‘the fairest of all taxes’. Pressure groups (lobbies, etc.) and politicians had discovered that the income tax was an ideal tool for permanently promoting certain

20

The Elgar guide to tax systems

preferred activities, through the use of ‘tax expenditures’ rather than through public spending that needed legislative approval each year. ‘Tax expenditures’ were justified on grounds that they improved equity and promoted some worthwhile objectives. This made income taxes progressively more complex and, horizontally and perhaps vertically, less equitable. So-called ‘second generation’ econometric studies, which used more sophisticated econometric techniques, started to discover disincentive effects connected with these taxes, effects that had been missed by earlier studies. These disincentives had to do with work participation (especially for second workers within family units), with number of hours worked, with the choice of activities, with the propensity to save and to invest, with the choice of assets in which to invest, and so on. What came to be called the ‘supply-side revolution’, the powerful political and intellectual movement that became popular especially with the election of President Reagan and Prime Minister Thatcher, and the increasing influence of some prominent conservative economists, such as Milton Friedman, generated politically powerful economic concepts, such as the ‘Laffer curve’ and similar ones, that started to influence tax policy in several countries. President Reagan and Prime Minister Thatcher became strong opponents of large public spending and high taxes, and especially of progressive income taxes. At the same time, an increasing number of individuals started receiving their incomes, or part of their incomes, from foreign sources, as a consequence of the accelerating pace of globalization. This was also the period when corporations extended their reach globally. Multinational corporations with ‘global reach’ became normal features. These developments complicated the task of the tax administrators and opened the way to international tax competition and to international tax evasion. In some way, the tax base that countries, and especially small countries, could try to tax started to extend to the global economy. The US 1986 Tax Reform promoted by President Reagan became a watershed event that within a short time set in motion major changes to the tax systems of many countries. The 1986 Tax Reform dramatically reduced the marginal tax rate for the personal income tax in the United States from 50 per cent to 28 percent. It also aimed at widening the national income tax base by eliminating tax incentives and ‘tax expenditures’ and helped establish a strong preference for low tax rates. This reform had a powerful ‘demonstration effect’ on other countries (see Tanzi, 1987). It set in motion a process that within a few years would bring a sharp fall in income tax rates in many OECD and other countries. Table 1.5 provides information, for selected years, for the top marginal personal income tax rates for the 1975–2007 period. Comparing 1975 with 2007 it is easy to see the remarkable reductions in these rates. In 1975

Tax systems in the OECD 21 Table 1.5

Top marginal personal income tax rates, 1975–2007: selected years (percentages)

Country Australia Austria Belgium Canada Czech Republic Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Japan Korea, Republic of Luxembourg Mexico Netherlands New Zealand Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom United States

1975

1981

1986

1992

1996

2000

2007

65 62 60 47 –

60 62 76.32 62.78 –

60 62 71.13 34 –

47 50 55 29 55

47 50 55 29 40

48 45 64 46 32

46.5 50.0 53.5 46.4 32.0

40 51 60 56 63 – – 77 72 75 –

70 51 60 56 60 – – 60 72 93 79.05

39.6 51 61.35 56 63 – – 60 62 70 55

68 39 56.5 53 50 40 33 52 51 50 50

65 39 – 53 45 48 – 48 51 50 40

60 55 5. 54 45 40 45 44 46 50 44

59.7 50.5 47.8 47.5 40.0 36.0 35.7 41.0 44.9 50.0 38.5

57 – 71 60 73 – – –

58.43 55 72 60 65.4 – 77.5 –

57 55 72 66 40 – 60 –

50 35 60 33 13 40 40 –

50 35 60 33 13.7 45 40 42

47 40 60 39 48 40 35 35

38.9 28.0 52.0 39.0 40.0 40.0 42.0 19.0

62 87 44 68 83

65.6 85 42 – 60

66 80 11.5 50 60

53 20 11.5 50 40

56 30 11.5 55 40

48 55 43 36 40

43.0 56.5 42.1 35.6 40.0

70

69.13

50

31

39.6

47

41.4

Note: The rates include sub-national income taxes. For specific comments on the data, the original sources must be consulted. Sources: 1975–99: World Tax Database, Office of Tax Policy Research; 2000–07: OECD Tax Database, Table 1.4; data from Office of Tax Policy Research, University of Michigan Business School.

22

The Elgar guide to tax systems

there were still many countries with rates equal to or above, 70 percent. Two countries (Sweden and the United Kingdom) had rates still above 80 percent. In the 1960s the rates had been even higher. By 2007 there was no country with rates higher than 60 per cent and only seven countries had rates equal to, or higher than, 50 percent. Only Denmark, Belgium, the Netherlands, and Sweden had rates above 50 percent. There was also significant convergence in the top rates. Ignoring the countries form East Europe, most countries had marginal income tax rates in the 40–50 per cent range. At the same time the number of income brackets to which different rates were applied had been significantly reduced in most countries. Countries attempted to maintain needed tax revenue by widening the tax bases but they were not always successful. Statutory tax bases remained smaller than theoretical, fully-comprehensive bases. Furthermore, revenue requirements to finance higher different spending needs in some countries left many rates higher than policy-makers would have liked. When the personal income taxes are combined with Social Security taxes, the tax wedges on most dependent workers in many countries are very high.8 Table 1.6 shows the tax wedges in 2007, for a worker with an average income, in OECD countries. These wedges vary from 21 per cent in Mexico, and 30 per cent in Korea, to more than 60 per cent in Belgium, Hungary, Sweden, and Austria. Workers with higher than average incomes have higher tax wedges. These wedges indicate the extent to which the workers’ degrees of freedom, in how to spend their before-tax incomes, are reduced by these taxes. In 2006 tax revenue from personal income taxes, as shares of GDP, ranged from very low levels in East European countries to 24 per cent in Denmark, 15.7 per cent in Sweden, and 14.9 per cent in New Zealand. The shares were also more than 10 per cent of GDP in Canada, the USA, Australia, Belgium, Finland, Iceland, Italy, Switzerland, and the United Kingdom. The OECD average was 9.2 per cent of GDP. For Spain it was 6.9 percent. Some of the countries with the highest personal income tax shares in GDP, such as Australia, New Zealand, and Denmark, do not have Social Security taxes because they finance basic public pensions with revenue from personal income taxes or from general taxation rather than from Social Security taxes. For this reason, the tax wedges for these countries are not higher than in other countries. In fact for Australia and New Zealand, they are lower. As mentioned earlier, tables such as Table 1.5 provide useful information but they may still hide some significant differences or some major developments that may be going on below the visible surface. A first aspect to mention is that the same marginal tax rate may be more or less harmful to the incentives of individuals, and more or less productive of tax revenue,

Tax systems in the OECD 23 Table 1.6

Total tax wedges for an average worker in OECD countries in 2007 (percentages)

Country

Tax Wedge

Australia Austria Belgium Canada Czech Republic Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Japan Korea Luxembourg Mexico Netherlands New Zealand Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom United States Source:

35.4 60.1 66.4 40.8 51.4 49.2 55.1 51.2 55.7 53.4 64.8 37.6 33.2 53.6 34.1 30.0 54.3 20.9 50.2 33.0 51.1 45.2 47.1 44.4 48.2 63.4 36.0 44.5 40.6 43.3

OECD: Taxation of Wage Income (2007).

depending on the threshold of income at which it is applied. Table 1.7 provides some information related to this aspect. It shows that while the top marginal tax rates (TMTRs) that applied to employees with average wages varied significantly among countries (from 28 per cent in Mexico to 56.5 per cent in Sweden), the level of income at which the rate applied also varied significantly (from 0.3 times the average wage in Iceland to 8.7 times average wage in the USA). Naturally, given the TMTR, the higher

24

The Elgar guide to tax systems

Table 1.7

Top marginal tax rates (TMTRs) and threshold levels at which they apply: 2007 in OECD countries

Countries Australia Austria Belgium Canada Czech Republic Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Japan Korea Luxembourg Mexico Netherlands New Zealand Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom United States

TMTR (Percentage)

Threshold (Multiple of Average Wage)

46.5 50.0 53.5 46.4 32.0 59.7 50.5 47.8 47.5 40.0 36.0 35.7 41.0 44.9 50.0 38.5 38.9 28.0 51.0 39.0 40.0 40.0 42.0 19.0 43.0 56.5 42.1 35.6 40.0 41.4

2.6 1.9 1.0 2.9 1.5 1.0 1.9 2.8 5.9 3.7 0.8 0.3 1.1 3.5 4.5 3.2 0.9 1.4 1.3 1.3 1.5 3.3 4.4 0.5 2.6 1.4 3.5 3.0 1.2 8.7

Note: The TMTR is for employees. Source: OECD: Taxation of Wage Income (2007).

the threshold at which it is applied the less damaging it must be to incentives. This aspect, while perhaps obvious, does not seem to have attracted the attention that it deserves on the part of tax experts. There must be some income level at which incentives are especially sensitive to the tax rate.9 This

Tax systems in the OECD 25 level and not the marginal tax rate per se independently of the income level to which it is applied, should be the focus of attention of tax analysts.10 Table 1.7 indicates that the United States, Germany, Japan, Portugal, Greece, Italy and Switzerland have the highest thresholds, while Iceland, Slovak Republic, Hungary, Luxembourg, Belgium, and Denmark have the lowest. Increases in thresholds may have essentially the same impact on the incentives of many individuals as reductions in top marginal tax rates. Unfortunately, there is no easily accessible information on what has happened to the levels of thresholds, over the years. Three other developments related to personal income taxes merit some mention, although a full discussion of them is not possible in this context. These developments signal major departures from the thinking that prevailed in the past. These are: 1.

2. 3.

The progressive abandonment of the attempt to introduce comprehensive, progressive, income taxes. There has been over the years a gradual movement toward the schedular income taxes that had been common in continental Europe in the early part of the last century. This movement has been in part forced by globalization and by tax competition. The introduction of dual income taxes (DITs) by Nordic countries, first, and subsequently, by other countries, including Spain. The growing popularity of flat-rate taxes (FRTs) and their introduction in a number of countries.

The growing movement towards a schedular approach to income taxation is evident from several changes introduced in tax systems in recent years. For example, there has been little interest in taxing unrealized capital gains while this issue monopolized the attention of many tax experts in the 1960s. Additionally, several countries now tax realized capital gains at different, and often more favorable, rates than labor income; dividends are also often taxed differently. DITs are conceptually consistent with a schedular approach. The dual income taxes (DITs) impose lower flat tax rates on capital income while they maintain higher and progressive rates on labor income. The DITs have been introduced to simplify tax administration and to reduce the incentives for capital flight toward tax havens and low-tax countries. These incentives are particularly strong when capital incomes are taxed at the top marginal tax rate within a comprehensive income tax (see Sørensen, 1994 and Boadway, 2005). The assumption behind the DITs has been that labor is internationally less mobile so that high, progressive tax rates on labor income do not encourage the emigration

26

The Elgar guide to tax systems

of workers, especially when the high taxes are used to pay for good, free, social services that the workers receive.11 It should be added that especially the Nordic welfare states could more easily simplify their tax systems because they had less need to provide ‘tax expenditures’ for health, education, housing, and so on because these services were provided free or were highly subsidized by the state. Thus DITs, with withholding at the source for capital incomes, and with few ‘tax expenditures’ for labor incomes, could be made simple administratively. Of course, highly skilled individuals who can earn high incomes globally and who are less likely to benefit from public services may still have strong incentives to emigrate. As the world becomes more globalized, this becomes more of an issue (see Tanzi, 1995, Chapter 4). Flat-rate taxes can be considered children of the supply-side revolution (see Hall and Rabushka, 1985 for the original proposal). They are often linear taxes that maintain a nominally fixed personal exemption (i.e., a zero rate bracket) and tax the income above the exemption at one, flat rate. This provides them with some progressivity. However, given the need for tax revenue, the higher the size of the exemption, the higher must be the tax rate.12 These taxes are based on a conservative philosophy that assumes that a flat, and presumably low tax rate, reduces the pressures by lobbies and special interest groups to ask for preferential tax treatments. It is claimed that this makes it possible to simplify the tax system while making it more efficient and more pro-growth (see, for example, Slay, 2009). However, the claim of simplicity has little to do with whether an income tax system has one or more tax rates but with whether the tax base can be made comprehensive. Most complications come from the identification of the tax base and not from the number of rates. Even with a single rate, the higher the personal exemption and the greater the country’s revenue needs, the higher must be the tax rate and the stronger the pressures for special tax treatments. Also a single tax rate that is relatively high, and that applies at a relatively low level of income, may create strong disincentive effects. It might apply at the level of income where the disincentive effects are most sensitive to the tax rate. There are now 24 countries that have introduced flat-rate taxes. They include Iceland, Turkey, and Slovakia among the OECD countries. Given future revenue needs of most OECD countries, and the growing concerns about poverty and income distribution in many of them, it is not likely that flat-rate taxes will become a common feature among OECD countries. If they did become common features in OECD countries, the flat rate would have to be very high. Russia, not a member of OECD, and Turkey are the only major countries that have introduced flat rates. The corporate income tax (CIT) is the next tax that will be discussed

Tax systems in the OECD 27 briefly. In spite of its relatively modest contribution to revenue, this tax now receives more attention than any other tax. It has been claimed, in several theoretical papers, that it is destined to ‘disappear’ because of global tax competition. Some theoretical papers have also been written about its presumably strong, negative impact on economic growth. Some economists and business pressure groups consider it the most damaging tax. Over the years, some economists have questioned the rationale for its existence. If the owners of corporations could be fully taxed for the corporate earnings, the rationale for the CIT would largely disappear. However, this imputation of earnings to individual taxpayers is very difficult especially when many of the owners are foreigners. We shall not discuss these theoretical questions but only report on a few points connected with the use of corporate income taxes in OECD countries. Let us start with the tax rates. In 2008, the CIT rates for OECD countries ranged from a low rate of 12.5 per cent in Ireland, to maximum rates of around 39 per cent in Japan and the United States. These rates include taxes levied by sub-national governments. Larger countries have higher rates and smaller countries have lower rates. This is consistent with the conclusions of economic theory.13 Ten countries had rates of 30 per cent or above. Six countries had rates of 20 per cent or below. The latter are all small countries in terms of income if not in terms of population, because they include Turkey and Poland. The tax rates provide evidence of tax competition. The very low rates in Ireland (12.5 percent) and, to a lesser extent, in Iceland (15.0) have attracted a lot of attention. Some studies have shown that some profits earned in other countries are often allocated to these countries, where they are taxed at lower rates. Since the 1980s, the tax rates on corporate income have fallen considerably. The trend continued in 2008 when several countries cut the CIT rate, some significantly (Germany by 8.7; Italy by 5.5; and Canada, Spain, UK, and the Czech Republic by 2–3 points). However, the significant cuts in tax rates over the years have not resulted in reductions in tax revenue from these taxes. For the OECD as a group, the tax revenue collected from CITs has grown continually since the mid-1960s. In 2006, the last year for which data are available, CITs generated almost 4 per cent of GDP for the OECD countries. The range was from 1.5 per cent of GDP for Turkey to 12.4 per cent for Norway, the latter high revenue due to taxes on profits from high oil prices. Excluding Norway, the largest revenues were received by Australia (6.6 percent), New Zealand (5.8 percent), Luxembourg (5.0 percent), and the Czech Republic, Denmark, Spain, and the UK (4 per cent or more). There is relatively little correlation for the OECD countries between tax revenue, as percentage of GDP, and tax rates (see Figure 1.1 and Table

28

The Elgar guide to tax systems 14.0 Norway

Tax revenue in 2006 (percentage of GDP)

12.0

10.0 y = 2.3448 + 0.0559x R² = 0.0356

8.0 Australia

6.0

New Zealand Czech Rep. Denmark

4.0

Ireland

2.0

Luxembourg Japan

Korea U.K. Netherlands Sweden Finland Switzerland Portugal Poland Slovak Rep. Greece Iceland Hungary Austria

Spain

Belgium

Canada Italy

U.S.A. France Germany

Turkey

0.0 0.0

5.0

10.0

15.0

20.0 25.0 Tax rate 2007

30.0

35.0

40.0

45.0

Source: Table 1.8.

Figure 1.1

CIT: tax rate and tax revenue in OECD countries (percentages)

1.8). It has been argued that in OECD countries the revenue-maximizing corporate income tax rate is 33 per cent (see Clausing, 2007). If this estimate is correct, it implies that most OECD countries could increase their revenue by increasing the rate to around 33 percent. However, another study has found that while in the 1980s the revenue-maximizing corporate tax rate was 34 percent, it has now declined to 26 per cent (see Brill and Hasset, 2007). Thus, especially the United States and Japan should reduce the rate to get more revenue, according to this study.

4

GROWING SAND IN THE MACHINERY OF TAX SYSTEMS

Several of the changes introduced over the years in the tax systems of the OECD countries have been in what many tax experts would consider the right direction. For example, the replacement of many excise and turnover (cascade) taxes with non-cascading value-added taxes was probably a major step in the right direction, because VATs distort less the prices at

Tax systems in the OECD 29 Table 1.8

CIT: tax rate and tax revenue in OECD countries (percentages)

Country Canada Mexico USA Australia Japan Korea New Zealand Austria Belgium Czech Republic Denmark Finland France Germany Greece Hungary Iceland Ireland Italy Luxembourg Netherlands Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom

Tax Rate in 2007

Tax Revenue in 2006 (percentage of GDP)

36.1

3.7

39.3 30.0 39.5 27.5 33.0 25.0 24.0 21.0 25.0 26.0 34.4 38.9 25.0 20.0 15.0 12.5 33.0 30.4 25.5 28.0 19.0 26.5 21.0 32.5 28.0 21.2 20.0 30.0

3.3 6.6 4.7 3.8 5.8 2.2 3.7 4.8 4.3 3.4 3.0 2.1 2.7 2.3 2.4 3.8 3.4 5.0 3.4 12.9 2.4 3.0 2.9 4.2 3.7 3.0 1.5 4.0

which goods are sold. The greater dependency on the VAT, rather than on income taxes, was also probably a step in the right direction at least in terms of allocative efficiency because the VAT exempts savings. The same can be said for the lowering of tax rates, for both individuals and corporate taxpayers because this lowers the welfare costs of income taxes. The reduction in the number of tax rates, in the tax systems of the countries’ personal income taxes, was also probably a change in the right direction

30

The Elgar guide to tax systems

even though one can be skeptical that fewer rates automatically mean simpler tax systems. The introduction in some countries of dual income taxes and the recognition that in today’s world it is no longer possible (or desirable) to aim for a ‘global, comprehensive, and progressive personal income tax’ have also been considered positive developments. Finally, the fact that inflation came down significantly in most OECD countries over the years, eliminated a problem that was considered serious in the past. This was the way in which inflation distorts tax systems (see Tanzi, 1980). Finally, by and large, at least in a superficial sense, there has been some convergence among countries in both tax levels and tax structures. While many of the above changes in tax systems, introduced in the last two decades, have been in what many would consider a clearly desirable direction, other problems have appeared that have complicated matters. I shall mention some of these problems and then focus on a couple of them, namely tax competition and especially, growing tax complexity. A first problem to mention is that the trend toward growing fiscal decentralization and toward fiscal federalism that has characterized many OECD countries in the past couple of decades (Belgium, Italy, Spain, Mexico, others) is having and will continue to have consequences for the tax systems. When sub-national governments acquire political control over some tax decisions, or over shares of some taxes, the national governments inevitably lose some of the degree of freedom in tax matters they had before, as happened in India, Argentina, Brazil, and some other countries. This may have negative consequences for tax systems. Fiscal federalism can also create tax competition among sub-national governments, leading to potential problems. This has happened in several countries, both within and without the OECD.14 A second problem is that of the increasing overlap of ‘national’ tax bases with the ‘global’ tax base. The global tax base has acquired some characteristics of ‘commons’, that is, of common grounds that can be exploited by, especially, small countries. This leads to attempts by some countries to export taxes by attracting foreign capital, foreign consumers, and foreign high-income individuals including pensioners. This ‘tax competition’ is aimed especially at mobile tax bases. It can be pursued by lowering tax rates on capital (but also on labor) income and on consumption for highly priced products. It can also be pursued through the erosion of the taxable base, especially for corporate income. This erosion explains in part the large differences that exist between corporate income tax rates and shares of corporate income taxes into GDP for some countries. For the corporate income tax, tax competition today is taking place through base erosion as well as through rate reduction. Base erosion is less visible and because of that it attracts less attention. However, it increases

Tax systems in the OECD 31 significantly, especially for small enterprises, the compliance cost of operating in different countries. Just think of a small enterprise operating in many countries and expected to deal with the complexity of each country’s tax system. Large corporations can afford to have accounting departments prepared to deal with different tax systems. Small enterprises do not have that luxury. It should be possible to deal with this kind of tax competition through international agreements. The definition of the base for taxing enterprise profits should be a purely technical matter. Therefore, it would be advantageous and, perhaps, possible, especially for countries belonging to the European Union or the OECD, to agree on the definition of the corporate income tax base. Then, if countries wished to compete, they could do it more transparently through the tax rate.15 There are different shades of tax competition, from the transparent one carried out through the level of tax rates; to the less transparent one, carried out through the manipulation of tax bases; and through the ‘unfair competition’ carried out through bank secrecy laws and other tools used by tax havens and by off-shore centers. This latter kind of tax competition has opened the way to major cross-country tax avoidance and global tax evasion, now estimated in the hundreds of billions of euros or dollars. Tax evasion and tax avoidance are increasingly becoming global phenomena rather, than largely domestic ones as in the past. They are often connected with bank secrecy and with special regulatory treatments of foreign companies. In a recent statement, President Obama noted that 18 857 [sic] US businesses had reported being housed in one single building in the Cayman Islands! They had all reported the same address. It must have been a very large building! Bloomberg News (4 May, 2009) and The Financial Times (5 May, 2009) reported that in 2003 a third of the foreign profits of US corporations came from just three small countries: Bermuda, Netherlands, and Ireland. It was also reported that US companies paid an effective tax rate of just 2.3 per cent on the $700 billion earned in foreign profits in 2004. This 2.3 per cent must be compared with the almost 40 per cent statutory US corporate income tax rate. A huge problem for tax systems in a globalized world is that it is now possible for multinational corporations to earn their profits in one country but, through various accounting strategies or, in some cases, by simply ‘checking a box’, to allocate the profits to another country where the tax rates are very low or zero or where it is possible to hide the identity of the real beneficiaries of the profits. As I argued almost two decades ago, it is time to reconsider some of the rules that have guided tax arrangements in recent years (see Tanzi, 1995). The reliance on the residence principle, without effective exchange of information, can lead to large tax avoidance. The future is likely to see

32

The Elgar guide to tax systems

a move towards arrangements that depend on source-based taxation and/ or on a fuller exchange of information. This, however, will require that institutions such as the European Union, the OECD, the United Nations or others can force reluctant countries to behave in a responsible way in a more globalized world. The existence of a World Tax Organization capable of monitoring tax developments and facilitating international agreements would help (see Tanzi, 1999 and 2008). A final problem to be mentioned and one that, in my judgment, is a major and growing one is tax complexity. As was mentioned earlier, economists tend to focus on the general structures of tax systems while many micro-changes in those systems are promoted by interest groups and are implemented by tax lawyers. Many of these changes are individually so small and seemingly insignificant that they often do not attract the attention of citizens or even of economists. Over the years these changes can have the cumulative effect of termites. They weaken the structure, increase the complexity of the systems and the cost of compliance, and reduce the efficiency and the equity of the tax systems. This issue is discussed briefly in the next section.

5

GROWING COMPLEXITY IN TAX SYSTEMS

The complexity of tax systems has different origins and different consequences. As Gary Becker, the Nobel Prize winner in economics in 1992, put it: Complications in the tax code are an excellent example of the conflict that sometimes arises between what is rational at the individual level, and what is rational to society as a whole. Each interest group lobbies to promote the interests of its members, although their interests advance usually at the expense of the interests of others. When many groups succeed in promoting their interests, losers vastly outweigh winners since each group gains from what they do, but loses from what is done to them by hundreds of other powerful interest groups. (Becker, 2006.)

Putting it differently, simplicity in taxation is a pure public good and, like most public goods, does not have a specific constituency to promote it. At the same time, there are many strong constituencies that push for provisions (tax incentives, tax expenditures, etc.) that benefit them but create complexity. These provisions are always justified on grounds of promoting equity, efficiency, and other worthwhile objectives. The work of the constituencies is helped by the existence of asymmetric information vis-à-vis the specific tax changes that are advocated. Those who push for

Tax systems in the OECD 33 them know their content much better than the general legislatures that must approve them.16 The changes are often pushed by specific members of parliament, or, at times, by specific ministers, who are responding to the pressures of the vested interest behind the changes.17 The above argument implies that there is a cumulative process at work that, with the passing of time, brings increasing complexity to the tax system, unless specific governmental action puts a stop to it. Attempts to stop it, however, are not likely to succeed because of the difficulties involved, the time and effort required, and the absence of political pay-off for government to try. Reagan tried in 1986 and failed. Within a few years the simplification of the 1986 Reform was erased. The Bush administration gave up before even trying in 2005 (see Advisory Panel on Federal Tax Reform, 2005). The governments of some countries, including Australia and France, expressed the intention to simplify their tax systems. They did not get far (for Australia, see Krever, 2003 and McKerchar, 2007). For the United States there are some data provided by the Tax Foundation, by tax analysts, by the Urban Institute, the Cato Institute, and other institutions that provide some information on tax complexity. A report of the Tax Foundation, for example, states that: ‘In 2005 individuals, businesses and non profit [organizations] will spend an estimated 6 billion hours complying with the federal income tax code, with an estimated compliance cost of over US$265.1 billion’ [or about 22 per cent of the income tax collected].’ The report estimates that 56 per cent of the total compliance costs are borne by businesses and the rest by individuals. It also estimates that the compliance costs are regressive. They amount to 5.9 per cent of income, for taxpayers with adjusted gross income less than US $20 000, and 0.5 per cent of income for those with income above US $200 000. Thus, complexity becomes a regressive tax. The complexity of the tax system affects the growth of the economy because of overhead costs and opportunity costs. The overhead costs consist of (1) tax planning; (2) tax audits and litigation; and (3) tax compliance. Opportunity costs consist mainly of time and effort allocated to tax matters. The report states that complexity is caused mostly by difficulty in defining income and of determining to whom to assign income and expenses. Chris Edwards (2006) of the Cato Institute has estimated the number of pages of US Federal income tax rules between 1913, when the income tax was introduced, and 2006. In 1913 there were 400 pages. These rose to 504 by 1939. By 1945 there were 8 200. By 1984 there were 26 300. These had increased to 66 498 by 2006. The Special Report of the Tax Foundation has calculated that the number of words in the US Tax Code and IRS

34

The Elgar guide to tax systems

Regulations between 1955 and 2005 increased from 718 000 to 7.1 million. The increase was once again due to difficulties in defining the tax base. The problem of growing complexity is, of course, not just a US problem. It affects most countries. We report below a few comments on some other countries to give a sense of how the problem is seen elsewhere. In Britain, instability, ‘inefficiency, and absence of fiscal coherence characterizes the tax system’ (Steinmo, 1993, p. 48). ‘No one would design such a system on purpose and nobody did. Only a historical explanation of how it came about can be offered as justification. That is not a justification, but a demonstration of how seemingly individually rational decisions can have absurd effects in aggregate’ (Kay and King, 1978, p. 1). This is essentially the point made by Gary Becker. According to a 2001 report to ministers, the New Zealand Tax Code instills ‘anger, frustration, confusion, alienation’ (‘The Case for Flat Taxes’, The Economist, page 59, 16 April 2005, p. 59). Some years ago, Eugenio Scalfaro, when he was president of Italy, stated publicly that the income tax declarations that Italian taxpayers were required to complete could only have been designed by ‘lunatics’. It was also reported that some Italian taxpayers have taken advantage of tax amnesties, not necessarily because they had evaded taxes, but because they were not sure whether they had succeeded in complying with all the complex rules and requirements of the Italian tax laws. By paying a penalty for a violation that they may not have committed, they eliminate the risk and the worries of being audited and penalized at some later date.

6

CONCLUDING REMARKS

This chapter has dealt with various developments in the tax systems of the OECD countries in recent decades. It has shown the remarkable growth of the tax levels over five decades, the main changes in the tax structures, the process of gradual amalgamation among the countries’ tax structures, the growing tax competition, associated with globalization, and some other significant developments. The chapter has stressed that the superficial similarity in developments may hide micro-changes that were not the same for all the OECD countries. The chapter concluded by calling attention to a disturbing development that could have damaging consequences in the future: the continuously growing complexity of the tax systems. This complexity has the effect of a hidden and regressive tax. Over time it can make the systems less efficient and less fair. In my view, this is an aspect that deserves urgent attention on the part of tax experts and economists.

Tax systems in the OECD 35

NOTES *

1. 2. 3. 4. 5. 6. 7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17.

Paper presented at the conference on ‘Tax Systems: Whence and Whither’ (Recent evolution, current problems and future challenges’), sponsored by FUNCAS and UNICAJA, Malaga (Spain) 9–11 September, 2009. I wish to thank the tax program of the OECD for the provision of data. See Revenue Statistics, 1965–2007: 2008 Edition (OECD, 2008, pp. 49–52). Keynes had agreed that ‘25 per cent [of GDP] as the maximum tolerable proportion of taxation may be exceedingly near the truth’. See Clark (1964, p. 21). The concept of the VAT was also exported to Africa and Latin America in the 1960s. In most countries, those who withhold the VAT taxes are required to submit the taxes withheld to the tax authorities within a short time, normally 30 to 60 days. There are now several good treatments of the VAT. See Ebrill et al. (2001); Schenk and Oldman (2007); and Bird and Gendron (2007). I was always puzzled by the fact that the main public finance or taxation books of the time did not have tax evasion in their indexes. For a very early attempt at determining the (negative) impact of the income tax on economic growth in industrial countries, see Tanzi (1969). Tax wedges are the differences between the cost of labor to the employers and the takehome labor incomes that workers receive in cash. It can be argued that at very high levels, incomes become almost rents because the individuals who receive those incomes would probably continue performing their activities with the same intensity even if they were subjected to higher tax rates. This may not be true for lower income levels. It is doubtful that, say, top tennis players or opera singers, or heads of corporations, would reduce their effort if the marginal tax rates were higher. Marginal tax rates may be more important for individuals on their way to positions that pay high incomes rather than for individuals who have already reached those positions. Thus, the level at which the marginal tax rate becomes effective is very important. It can be argued, also, that when incomes become very high they contain an element of rent because those who receive them would not change their profession or effort if part of the income were taxed away. However, DITs do not prevent the migration of labor into underground economic activities. This can make the step between untaxed income and taxed income quite steep, thus potentially encouraging tax evasion. Small countries may gain more by attracting foreign capital to them with low rates than they lose in revenue because of the low rates. For a brief survey of the US experience, see Chapter 3 in Tanzi (1995). It could be argued that the tax base may not be the same in different countries. For example, depreciation allowances may have to reflect weather conditions. But agreements could take account of these differences. There is a lot of evidence that many of those who vote do not even read the bills on which they are voting unless they have a specific interest in them. At times the changes are hidden in bills that may appear to have little bearing on taxation. By the way this is a far greater problem for income taxes than for other taxes.

REFERENCES Advisory Panel on Federal Tax Reform (USA) (2005), ‘Simple, Fair and Pro-Growth: Proposals to Fix America’s Tax System’, Washington, DC: GPO. Becker, Gary (2006), ‘Tax Complexity and the Cost of Compliance’, The Becker-Posner Blog, 16 April.

36

The Elgar guide to tax systems

Bird, Richard M. and Pierre-Pascal Gendron (2007), The VAT in Developing and Transitional Countries, New York: Cambridge University Press. Brill, Alex and Kevin A. Hassett (2007), ‘Revenue-Maximizing Corporate Income taxes: The Laffer Curve in OECD Countries’, AEI Working Paper No.137, July. Boadway, Robin (2005), ‘Income Tax Reform in a Globalized World: The Case for a Dual Income Tax’, Journal of Asian Economics, 16(6), 910–27. Ceriani, Vieri (2009), ‘Comments on Methodologies for Measuring the VAT Gap’, mimeo, Rome, 9 February. Clark, Colin (1964), Taxmanship: Principles and Proposals for the Reform of Taxation, Hobart Papers No.26, London: Institute of Economic Affairs. Clausing, Kimberly A. (2007), ‘Corporate Tax Revenue in OECD Countries’, Int. Tax Public Finance, 14(2), 115–33. Ebrill, Liam, Michael Keen, Jean Paul Bodin and Victoria Summers (2001), The Modern VAT, Washington, DC: IMF. Edwards, Chris (2006), ‘Income Tax Rife with Complexity and Inefficiency’, Tax and Budget, Bulletin of the Cato Institute, No. 3, April, Washington, DC. Hall, Robert and Alvin Rabushka (1985), The Flat Tax, Stanford: Hoover Institution Press. Kay, J.A. and M.A. King (1978), The British Tax System, Oxford: Oxford University Press. Krever, R. (2003), ‘Taming Complexity in Australian Income Tax’, The Sydney Law Review, 25(4), 67–505. McKerchar, Margaret (2007), ‘Tax Complexity and its Impact on Tax Compliance and Tax Administration in Australia’, The IRS Research Bulletin, Publication No.1500, Proceedings of the 2007 IRS Research Conference. OECD (2008), Revenue Statistics 1965–2007: 2008 Edition, Paris: OECD. Schenk, Alan and Oliver Oldman (2007), Value-Added Tax: A Comparative Approach, New York: Cambridge University Press. Slay, Ben (2009), ‘Some Early Lessons on Tax and Social Reform from New EU Member States’, paper presented at a conference on ‘Tax Policy Options in the Wake of EU Accession for Turkey’, Ankara, 11–12 January. Sørensen, Peter B. (1994), ‘From the Global Income Tax to the Dual Income Tax: Recent Reforms in the Nordic Countries’, International Tax and Public Finance, 1(1), 57–79. Steinmo, Sven (1993), Taxation and Democracy: Swedish, British and American Approaches to Financing the Modern State, New Haven, CT: Yale University Press. Tanzi, Vito (1969), The Individual Income Tax and Economic Growth: An International Comparison, Baltimore: The John’s Hopkins Press. Tanzi, Vito (1980), Inflation and the Personal Income Tax, Cambridge, UK: Cambridge University Press. Tanzi, Vito (1987), ‘The Response of Other Industrial Countries to the U.S. Tax Reform Act’, National Tax Journal, XL(3), 339–55. Tanzi, Vito (1995), Taxation in an Integrating World, Washington, DC: The Brookings Institutions. Tanzi, Vito (1999), ‘Does the World Need a World Tax Organization?’, in Assaf Razin and Efraim Sadka (eds) The Economics of Globalization, Cambridge, UK: Cambridge University Press. Tanzi, Vito (2008), ‘Globalization, Tax Systems, and the Architecture of the Global Economic System’, in Vito Tanzi, Alberto Barreix and Luiz Villela (eds) Taxation and Latin American Integration, New York: IDB and Harvard University. Tanzi, Vito and Ludger Schuknecht (2000), Public Spending in the 20th Century, Cambridge, UK: Cambridge University Press.

2

Direct versus indirect taxation: trends, theory, and economic significance* Jorge Martinez-Vazquez, Violeta Vulovic, and Yongzheng Liu

1

INTRODUCTION AND SOME DEFINITIONS

One of the oldest questions in the theory and practice of taxation is that of the appropriate mix of direct and indirect taxes. The choice between direct and indirect taxes has contributed to a long animated debate, in political and academic circles, regarding the virtues and defects of those two forms of taxation. In this chapter we provide an overview of the evolution of the ratio of direct taxes to indirect taxes across countries over the past three decades, the theorizing that has gone behind the alleged superiority of one form of taxation or the other, the determinants that appear to be behind the intensity with which both forms of taxation are used, and the economic relevance of the choice of tax structure in terms of economic growth, macroeconomic stability, the distribution of income, and the flow of foreign direct investment (FDI). To get started it is helpful to have a working definition of direct and indirect taxes. Following Atkinson (1977) we will define as direct taxes those that may be adjusted to the individual characteristics of the taxpayer and as indirect taxes those that are levied on transactions irrespective of the circumstances of buyer or seller. Thus, conventional income taxes can be classified as direct taxes and the same can said for most taxes on assets and wealth as long as there are potential adjustments for the characteristics of owners. For example, property taxes on owner-occupied housing may be adjusted for the personal characteristics of owners but that is not always the case. Property taxes on commercial buildings, motor vehicles, and the like are hardly ever adjusted for personal or household characteristics and therefore those can be considered indirect taxes. In this category of indirect taxes are most taxes on transactions with differentiated rates (sales, value-added tax [VAT], excises, customs tariff, etc.). But, as indicated by Atkinson, there are what may called ‘transitional’ taxes between the two categories; in particular, a uniform general sales tax can be easily transformed into a general consumption or expenditure tax, which can be adapted to personal or household characteristics.1 37

38

The Elgar guide to tax systems

Over the last three decades the average ratio of direct to indirect taxes for a sample of 116 countries has been on the increase and these changes have been more pronounced for developed countries than for developing countries. As we will see in detail in the next section, in the case of developed countries it has been the numerator of the ratio that has moved the most, with the main driver being increases in the relative importance of Social Security contributions, while smaller relative increases in corporate income taxes have been offset by also smaller relative decreases in personal income taxes; this has been accompanied by a relatively flat performance of domestic consumption taxes. In the case of developing countries, it has been changes in the denominator of the ratio that has had the largest impact. Fairly large decreases in the relative importance of customs taxes have been only partially offset by increases in the relative importance of domestic consumption taxes, while at the same time a small decrease in income taxes has been more than offset by an increase in the relative importance of Social Security contributions. In the economics literature a theoretical debate has accompanied over the years the choice between direct and indirect forms of taxation. The choice of direct versus indirect taxes is fundamental to the optimal design of tax structures since those forms of taxation may affect differently the goals of efficiency and equity. While some early contributions strove to demonstrate the superiority of direct over indirect taxes under specific conditions (Hicks, 1939),2 most of the focus early on in the optimal tax literature was on separate forms of taxation (e.g., Ramsey, 1927; Diamond and Mirrlees, 1971). A key development in the optimal tax literature from the perspective of the optimal tax mix was Atkinson and Stiglitz’s (1976) seminal paper. These authors, who for the first time considered the interaction of direct and indirect taxes in the attainment of efficiency and equity goals, reached a powerful result. The Atkinson and Stiglitz theorem states that, in an economy where individuals differ only in their earning abilities, government can impose a general income tax, and where the utility function is separable between labor and all commodities, then in the optimum tax design there is no need to employ indirect taxation. This important result was followed, as we will see in the overview of the theoretical literature below, by a significant number of other theoretical contributions showing how important aspects of the economy (e.g., the scope of tax evasion) and heterogeneity among taxpayers would justify the existence side by side of direct and indirect forms of taxation. This is comforting since basically all economies employ together broad forms of direct and indirect taxation even though we are far from fully understanding what the main determinants of the direct to indirect tax mix are (Kenny and Winer, 2006).

Direct versus indirect taxation 39 With the coexistence of direct and indirect forms of taxation explained in the theoretical optimal tax literature, the big question that has remained largely unanswered is that of the economic consequences of different mixes of direct and indirect taxes. For example, from the perspective of economic growth, in a neoclassical framework, the tax structure, and in particular the tax mix, has no permanent effects on the growth rate, although changes in tax policy can have transitory effects.3 But in the context of endogenous growth models even stable tax structures can impact the growth rate due to the externality effects on the accumulation of human and physical capital. As we review below, an increasing number of studies find important effects of the tax mix on the rate of economic growth. The choice of the direct–indirect tax mix is also likely to have, as we review below, important consequences in other dimensions of the economy including macroeconomic stability, disparities in income distribution, and foreign direct investment flows. All those, including economic growth, will be revisited in this chapter. There are several other potential effects of the choice of tax mix, including the impact on risk taking and entrepreneurship or taxpayers’ moral and voluntary tax compliance. As Atkinson (1977) points out, supposedly taxpayers may show preference for indirect taxation on the grounds that it offers them choice and some politicians may have similar preferences because indirect taxes may be perceived by the public as being less visible.4 None of these other possible effects will be explored further in this chapter. The rest of the chapter is organized as follows. In Section 2 we provide an overview of the international trends in the use of direct versus indirect forms of taxation over the last three decades. In Section 3 we review the theoretical literature on optimal tax design and the more recent empirical literature on the economic consequences of the choice of tax structure. In Section 4 we revisit the issue of the determinants of tax structure with international panel data from the perspective of the direct to indirect tax ratio. In Section 5, using the same international panel data set, we explore the effects of the direct to indirect tax mix on economic growth, macroeconomic stability, income distribution, and foreign direct investment flows. In Section 6 we conclude.

2

TRENDS IN DIRECT VERSUS INDIRECT FORMS OF TAXATION

In this section we provide as background information an overview of the evolution of the average direct to indirect tax ratio over the period 1972–2005 for a sample of 116 developed and developing countries. Figure

40

The Elgar guide to tax systems

2.5

2

1.5

1

0.5

Full Sample

Developed Countries

Developing Countries

2005

2004

2003

2002

2001

2000

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

0

Countries in Transition

Note: a. Based on a sample of 116 countries (number of countries in the sample varies across years) (note change in GFS methodology after 1990); property taxes included in direct taxes; for 1990–94 data not available. Source: IMF GFS Database.

Figure 2.1

Average annual direct to indirect tax ratio,a 1972–2005

2.1 shows the trend when property taxes are classified as direct taxes, but the trends are maintained when property taxes are classified as indirect taxes, with tax ratio having a lower value.5 This figure omits observations for the period 1990–99 because a change of classification in the GFS (Government Finance Statistics) from the IMF led to irregular country reporting over the period, which distorts the average figures.6 Several significant trends are observed. For developed countries the ratio has steadily increased over the period by over 50 per cent while for developing countries has roughly stayed the same, with an average ratio that is about one-third of its value for the average of developed countries. For both developed and developing countries there tends to be somewhat of a jump in the direct to indirect ratio in the 2000s but without a clear trend. Some of this increase in the tax ratio is no doubt due to the changes in the definition of GFS, which most substantially represented a more explicit separate accounting for Social Security taxes, which before 1990 had been classified as non-tax revenues and also partially as income taxes. For the full sample, there is correspondingly also an increase of the tax

Direct versus indirect taxation 41 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05

Income Taxes Property Taxes

Payroll Tax Customs

Social Security Contributions

2005

2004

2003

2002

2001

2000

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

0

Consumption Taxes

Other Taxes

Note: a. Based on a sample of 116 countries (number of countries in the sample varies across years) (note change in GFS methodology after 1990); property taxes included in direct taxes; for 1990–99 data not available. Source:

IMF GFS Database.

Figure 2.2

Average annual tax structure as a share of total taxes,a 1972–2005

ratio from roughly a value of 0.75 during the 1970s and 1980s to almost 1.0 in the most recent years. To understand better what is driving the behavior of the direct to indirect tax ratio, we show the historical evolutions of the share of each of the main taxes as a ratio of total taxes over the 1972–2005 period for the full sample, and for developed, developing, and transition countries in Figures 2.2, 2.3, 2.4, and 2.5, respectively. We should note that GFS reporting is fairly aggregate in some cases and so, for example, we are not able to distinguish, for a number of countries, between personal and corporate income taxes or in the case of domestic consumption taxes, between VAT and excises. Although one should not pay much attention to fluctuations over short periods of time, which can be due to, among other things, sample composition, these figures are useful to identify some trends. In terms of indirect taxes, for the full sample we observe an increase in consumption taxes, supposedly driven by increases in VAT collections. This increase in the relative importance of consumption taxes is apparent in the

42

The Elgar guide to tax systems

0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05

Income Taxes Property Taxes

Payroll Tax Customs

Social Security Contributions

2005

2004

2003

2002

2001

2000

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

0

Consumption Taxes

Other Taxes

Note: a. Based on a sample of 32 developed countries (number of countries in the sample varies across years) (note change in GFS methodology after 1990); for 1990–99 data not available. Sources: IMF GFS Database and World Bank World Development Indicators.

Figure 2.3

Average annual tax structure as a share of total taxes in developed countries,a 1972–2005

groups of developing and transition countries; for the case of developed countries their importance remains fairly flat. Another noticeable trend is the drop in the relative importance of customs taxes, especially in developing countries. For many decades now a standard policy recommendation for developing countries, from the IMF, the World Bank, and many other sources, has been to promote trade liberalization by implementing a revenue-neutral reform reducing the customs tariff and increasing domestic consumption taxes, mostly the VAT. However, this policy thrust has been shown only partially successful in actual implementation in a number of recent empirical studies.7 Keen (2008) provides reasons why it is difficult for developing countries to replace the loss of trade taxes with increased VAT revenues, and Baunsgaard and Keen (2005) find that the degree of revenue recovery through domestic taxes is significantly less in lower-income versus middleand high-income countries. While for high- and middle-income countries, this revenue recovery effect is generally effective, for low-income

Direct versus indirect taxation 43 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05

Income Taxes Property Taxes

Payroll Tax Customs

Social Security Contributions

2005

2004

2003

2002

2001

2000

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

0

Consumption Taxes

Other Taxes

Note: a. Based on a sample of 75 developed countries (number of countries in the sample varies across years) (note change in GFS methodology after 1990); for 1990–99 data not available. Sources: IMF GFS Database and World Bank World Development Indicators.

Figure 2.4

Average annual tax structure as a share of total taxes in developed countries,a 1972–2005

countries, however, this effect is weak, that is, less than 30 per cent of the trade tax loss could be offset by the increase of the domestic consumption tax. And there is no evidence supporting that the presence of a VAT will bring a significant difference to the degree of recovery in low-income countries. In terms of direct taxes, the big mover and shaker is Social Security contributions, which experienced a significant increase over the last two decades, especially in developed countries and less of an increase in developing and transitional countries. Income taxes have decreased in relative importance in developing and transition countries, but remained rather flat in the case of developed countries. Using OECD data for developed countries shows that for this group, while personal income taxes have decreased, corporate income taxes have increased.8 The increases in corporate income taxes have taken place despite the fact that statutory corporate tax rates have declined internationally as a response to the increasing

44

The Elgar guide to tax systems

0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 2000

2001

Income Taxes Property Taxes

2002 Payroll Tax Customs

2003

2004

Social Security Contributions

2005 Consumption Taxes

Other Taxes

Note: a. Based on a sample of nine countries in transition (number of countries in the sample varies across years) (note change in GFS methodology after 1990). Source: IMF GFS Database and World Bank World Development Indicators.

Figure 2.5

Average annual tax structure as a share of total taxes in transition countries,a 2000–05

mobility of capital and firms in the last two decades in an attempt of many governments to remain attractive to international capital. A substantial body of research has put forward explanations for this apparent paradox. First, the broadening of the corporate tax base by changes in the laws has played an important role in offsetting the reduction of statutory tax rates (Devereux et al., 2002; Simmons, 2006; Sørensen, 2006; Piotrowska and Vanborren, 2008). Second, income shifting from personal to corporate tax bases, or from non-corporate to the corporate sector due to the incentive effect of the low tax rate in the corporate sector has been suggested as another explanation for the paradox (Devereux and Sørensen, 2005; De Mooij and Nicodème, 2008). Third, an increase of corporate profitability and the size of the corporate sector may have increased the effective tax rate and, therefore, tax revenues (Devereux et al., 2002; Devereux and Sørensen, 2005; Auerbach, 2006; Simmons, 2006; Clausing, 2007). The relative shares of property taxes and other taxes (environmental levies, etc.) have been fairly constant over time.

Direct versus indirect taxation 45

3

THE CHOICE OF DIRECT VERSUS INDIRECT TAXATION: AN OVERVIEW OF THE THEORY AND EMPIRICAL FINDINGS

A voluminous literature has developed over the last decades on the optimal design of tax systems and more in particular on the choice of direct versus indirect forms of taxation. In this section we give an overview of the main developments in these literatures and where the debate stands today. Optimal Tax Theory: What Role for Indirect Taxes? The Atkinson-Stiglitz theorem The starting point in the optimal tax literature is the well-known Atkinson-Stiglitz (1976) theorem, which states that when the government may choose a general income tax function, individuals differ only on wage earning ability, and the utility functions are separable between labor and all commodities, then no indirect taxes need be employed. This theorem implied, as Atkinson-Stiglitz (1976) noted, that the extent to which indirect taxes are employed may depend on the (more complex) form of consumer preferences and possibly on restrictions on the type of income taxation that can be employed; for example, horizontal equity considerations can introduce constraints on the structure of income taxes. The costs of tax administration are not recognized either; allowing for cost differences for separate taxes could also affect the optimal tax structure. The Atkinson-Stiglitz theorem shaped the research agenda on optimal tax structures for many years to come. But, it is important to note that Atkinson-Stiglitz (1976) saw their analysis as being more useful in shaping the structure of the argument regarding the choice of optimal tax structure than in providing policy advice. What followed Atkinson-Stiglitz’s work has been a series of important papers showing how indirect taxes may be justified in an optimal tax structure if some of the explicit and also implicit assumptions in their work are relaxed. Role of tax evasion and avoidance It turns out that considering the administration of taxes, in particular enforcement and evasion issues have important consequences for the optimal tax mix of direct and indirect taxes. Boadway et al. (1994) show that if different taxes have different evasion characteristics, some optimal tax structure with a meaningful role for indirect taxes emerges naturally. Assuming that only income tax can be evaded (or can be evaded more easily)9 the authors analyze the case for supplementing optimal (non-linear) income taxation with commodity

46

The Elgar guide to tax systems

taxation and develop conditions under which commodity taxation should not be at the same rate.10 Role of uncertainty Cremer and Gahvari (1995) show that in the presence of uncertainty, where otherwise identical individuals are uncertain about the wage they would earn, differential commodity taxation is a necessary component of an optimal tax structure. Role of the production side Naito (1999) shows that, even when the government is using a Pareto-efficient non-linear income tax system under weak separability of workers’ utility functions, imposing a non-uniform commodity tax can improve welfare, once the assumption of constant marginal cost of production is abandoned and the production side of the economy is explicitly introduced in the analysis. Role of heterogeneity Cremer et al. (2001) show that when individuals differ in several unobservable characteristics (productivity and endowments), differential commodity taxes do have a role to play as instruments of optimal tax policy – an optimal (general) income tax will not suffice, while the optimal commodity tax rates follow traditional Ramsey rules. Papers by Saez (2002) and Balestrino et al. (2003) make contributions along similar lines. Role of endogenous human capital accumulation Naito (2004) finds that using a commodity tax can increase social welfare in the presence of a non-linear income tax system when human capital accumulation is endogenous. In particular, assuming that individuals with greater ability have comparative advantage in accumulating skilled human capital, Naito shows that indirect redistribution such as imposing a tariff on unskilled human capital-intensive goods can increase the efficiency of, and complement, an income tax system. Transparency Dahlby (2003) argues that levying both direct and indirect consumption taxes could improve the transparency of the tax system, especially when there are several tiers of government with autonomous taxing powers. Impact on Economic Activity: Does the Selection of Direct to Indirect Tax Ratio Matter? Alongside the theoretical modeling on optimal tax structure an empirical literature has developed over the past several decades examining the

Direct versus indirect taxation 47 impact of the direct to indirect tax ratio on economic activity. The empirical findings are varied and not always consistent. While older studies tend to find less significant economic effects of taxes, more recent studies tend to find significant effects of the direct versus indirect tax mix on various outcomes. These differences in results have to do with the sample period of the studies but also with the methodology employed. Impact on labor supply, prices, and output An earlier paper in this literature is by Atkinson and Stern (1980), who use an extended linear expenditure system with United Kingdom Family Expenditure Survey data to examine the impact of a reduction in income taxes and an increase in the VAT on labor supply and welfare. For labor supply they find a net increase in labor supply by those with the highest wage rates, with the income tax cut increasing hours and the VAT change reducing them. The analysis of welfare changes shows that the benefits of a switch from income tax to VAT would flow to those with higher wages. A second paper by Poterba et al. (1986) uses quarterly data from the United Kingdom and the United States to investigate how shifts in the direct versus indirect mix affect wages, prices, and output. The period studied for the United Kingdom was 1963:3 to 1983:4, while for the United States it was 1948:1 to 1984:3. For the United Kingdom the results suggest that shifts from direct to indirect taxation in the short run leads to an increase in prices and after-tax wage and reduces real output, but that in the long run the shift from direct to indirect taxes seems to have no significant effects. The results obtained for the United States are very similar to those for the UK. Madsen and Damania (1996) augment Poterba et al.’s (1986) work to explore the impact of switches from direct to indirect taxes on both wages and output levels for 22 OECD countries over the period 1960 to 1990. They conclude that for the majority of countries in the sample a revenue-neutral switch from direct to indirect taxes has no impact on the level of long-run economic activity. However, they also find that in some economies those tax changes have resulted in increases in output levels and lower nominal wages in the long run. More recent studies have found quite different results. A study by the European Commission (2006) simulates the macroeconomic effects of a revenue-neutral shift in taxation from direct to indirect taxes, using the QUEST model and shows that the shift in taxes might indeed strengthen economic growth and increase employment. In a more recent paper, Johansson et al. (2008) analyze the effects of changes in tax structure on GDP per capita for 21 OECD countries over the period 1970 to 2005. These authors find that consumption and property taxes have a significantly less adverse effect on GDP per capita than taxing income and that

48

The Elgar guide to tax systems

corporate income taxes appear to have a particularly negative impact on GDP per capita. Impact on economic growth The strongest evidence yet that direct versus indirect tax choices matter is in the context of dynamic endogenous growth settings; this evidence points to the fact that switching the tax mix toward consumption taxation and away from income taxation has very significant growth effects or dynamic efficiency gains (Kim, 1998; Dahlby, 2003; Li and Sarte, 2004).11 In the paragraphs below we survey some of this empirical work, which has been mainly carried out with data from OECD countries. Kneller et al. (1999), using five-year average data for 22 OECD countries for the period 1970–95, find that while income taxes reduce growth, consumption taxes do not. Widmalm (2001), using panel data for 23 OECD countries between 1965 and 1990, finds that that the proportion of tax revenue raised by taxing personal income is robustly, negatively correlated with economic growth. Widmalm also finds evidence that tax progressivity, measured in terms of the long-run income elasticity of tax revenue, tends to reduce economic growth and that progressivity affects growth, not so much via physical capital accumulation, as through the accumulation of human capital. Padovano and Galli (2001), also using panel data for 23 OECD countries covering the 1950s to the 1980s, find robust results that high marginal income rates and progressivity are negatively correlated with economic growth. The same conclusions are reached in Padovano and Galli (2002) with an updated panel of 25 industrialized countries covering 1970 to 1998. Li and Sarte (2004) find evidence that the decreases in progressivity associated with the Tax Reform Act of 1986 (TRA-86) in the US lead to small but non-negligible increases in US long-run growth (from 0.12 to 0.34 percentage points). Finally, Lee and Gordon (2005), using panel data for 70 countries covering the period 1970–97, find in cross-sectional regressions and fixedeffects regressions that higher corporate tax rates are associated with lower growth rates. Impact on income distribution The interest in the impact of tax structure on income distribution dates back to Meltzer and Richard’s (1981) work on the majority rule and the median voter model, predicting that when the mean income rises relative to the median income (that of the decisive voter), taxes rise, and vice versa. However, their model does not unbundle the different taxes, although the presumption would likely be that the rise in taxes should take more the form of direct taxes (mostly paid by higher income groups) as opposed to indirect taxes (more evenly

Direct versus indirect taxation 49 distributed across all taxpayers).12 Although there is a fairly large applied literature on tax incidence, allocating tax burdens among different income groups according to a conventional set of assumptions about tax shifting,13 there has been less empirical work on the impact of the tax structure, in particular the direct to indirect tax mix on the distribution of income. Li and Sarte (2004) find that the progressivity change associated with the TRA-86 in the United States had a significant effect on income inequality, resulting in a 20 to 24 per cent increase in the Gini coefficient of income. More recently, Weller (2007) uses cross-country data from 1981 to 2002 and finds positive effects of progressive taxation on income distribution. An important handicap, explaining the few studies available, is the difficulty of putting together compatible panel data on income distribution. Duncan and Sabirianova Peter (2008) examine whether income inequality is affected by the structural progressivity of national income tax systems and find that while progressivity reduces observed inequality in reported gross and net income, it has a significantly smaller impact on true inequality, approximated by consumption-based measures of Gini. Impact on macroeconomic stability Even though the built-in stabilizing properties of tax structures have been a noted issue since Musgrave (1959), little empirical work has been conducted to estimate the impact of different tax structures and in particular the role of the direct to indirect tax mix in increasing macroeconomic stability, the presumption being that tax systems that rely more heavily on direct taxation will contribute more effectively to macroeconomic stability.14 Auerbach and Feenberg (2000) examined the tax system’s potential to stabilize income fluctuations in the US economy since the early 1960s and find that that automatic stabilization of aggregate demand probably offsets as much as 8 per cent of the initial shocks to GDP. In addition, they find that there has been relatively little net change in the role of the tax system as an automatic stabilizer; the US tax system effectiveness in stabilizing aggregate demand in 1995 was roughly the same as in the early 1960s, but lower than at its estimated peak in 1981. In a more recent study, Weller (2007), using cross-country data for 1981 to 2002, finds the relationship between progressive taxes and growth volatility to be ambiguous. In Section 5 of this chapter we revisit the questions of the potential economic impact of the direct to indirect tax mix on economic growth, macrostability, and income distribution using a unified international panel data set. But before doing that we examine in the next section the determinants of the tax mix ratio.

50

The Elgar guide to tax systems

4

THE DETERMINANTS OF THE DIRECT TO INDIRECT TAX RATIO

In this section we examine the determinants of the direct to indirect tax ratio, building on recent work by Kenny and Winer (2006) and Hines and Summers (2009) on the determinants of the different components of tax structures. Our central question is: in practice, what are the main determinants of the proportion in which direct and indirect taxes are used? This is a broad question and we are interested in the different aspects of the economy and societal institutions that may bear on this issue. In the first part of the section we discuss a number of methodological issues and in the second part we present the panel data set and the estimation results. Empirical Approach We estimate the following model using a two-stage least squares (2SLS) methodology with panel corrected standard errors,15 including country dummies to account for any potential individual fixed effects: TaxRatioit 5 Xitb 1 ui 1 eit; i 5 1, ., n, t 5 1, . . ., T

(2.1)

where i indexes country and t indexes year, and ui represents the countryspecific fixed effects. The TaxRatio is measured as the ratio of direct taxes (personal and corporate income tax, payroll tax, Social Security contributions, and property tax) and indirect taxes (taxes on goods and services, taxes on international trade, and other taxes). The tax data represent consolidated general government data and are drawn from the IMF GFS Database. Given that certain types of property taxes can be treated as direct and some as indirect taxes, and because we are not able to distinguish among different types within the data we have, we will alternatively estimate the model using a dependent variable where the property tax is included as an indirect tax in the denominator. Alternative definitions of the dependent variable, the direct to indirect tax ratio are possible. For example, Poterba et al. (1986) use in their analysis of how tax systems may affect wages, prices, and output a direct to indirect tax mix variable defined as the difference between the direct and the indirect tax rates computed as (t 2 q)/(1 + q) where t is the direct tax rate and q the indirect tax rate, and these tax rates computed, respectively, as total direct and indirect taxes divided by nominal GDP. This alternative definition is highly correlated with our measure of TaxRatio, the simple correlation coefficient for the two measures in our panel data set being 0.841.16

Direct versus indirect taxation 51 The set of observable characteristics Xit that we hypothesize to affect the tax ratio is selected following the work in Kenny and Winer (2006) and Hines and Summers (2009).17 The first paper by Kenny and Winer (2006) examines the determinants of the structure of tax systems using a sample of 100 democratic and non-democratic countries over the period 1975–92. For estimation purposes, Kenny and Winer (2006) use an SUR (seemingly unrelated regression) approach to test for whether and how the set of explanatory variables matters for each of the tax instruments in a country’s tax system. Since our variable of interest is the ratio of direct to indirect taxes rather than individual taxes per se, we should not expect to find the same relationships (signs and significance) between the respective explanatory variables and our dependent variable based on Kenny and Winer’s (2006) results. Nevertheless, their study provides a very useful guide on the channels through which particular determinants may be expected to influence the direct to indirect tax ratio. The second study by Hines and Summers (2009) examines the effect of globalization on tax design using cross-country data over the period 1972 to 2006. In cross-sectional regressions for 1973, 1985, and 1999 they find that the reliance on income taxes (personal income taxes and corporate income taxes) on total taxes is higher the larger the country (log population) and the wealthier the country (log per capita income) with this reliance increasing over time.18 For expenditure taxes (taxes on goods and services and international trade taxes), the cross-sectional regressions for 1973, 1985, and 1999 suggest that country size and per capita income are consistently associated with smaller ratios of expenditure taxes to total tax revenues. The panel evidence is quite consistent with the cross-sectional evidence. Growing income levels are associated with reduced reliance on expenditure taxes (44.2 percent), and population growth is likewise associated with less use of expenditure taxes. The determinants of the tax mix ratio may be categorized into ‘demand’ factors and ‘supply’ factors. By demand factors we mean those that pull the level of certain taxes or the overall level of taxation up because of preferences or the overall budget constraint of the public sector; if more public goods and services are desired, more taxes on private income will need to be raised. Supply factors represent those that facilitate the collection of certain taxes or all taxes in general, such as the availability of tax bases or ‘tax handles’, and institutional and structural features that facilitate tax administration and enforcement. Among the demand factors, we identify first several forms of ‘scale effects’. The size of total revenue to GDP measures how much overall government a particular society wishes to have. As the size of government gets larger, it is likely that most or all revenue categories (measured as a

52

The Elgar guide to tax systems

share of GDP) will need to rise, but there is no clear reason why direct or indirect taxes would have to rise faster. There are also additional scale effects arising from the size of the country measured by population and from the degree of decentralization in a country. A larger population and thus more congestion may lead to higher tax levels, and with decentralization the consolidated government sector is also likely to be larger; both of these factors are likely to lead to a more intense use of different tax sources but without a clear decantation a priori for higher use of either direct or indirect taxes. Another demand factor is that of ‘political preferences’. For example, repressive regimes may turn away from sources requiring higher degrees of citizen cooperation or voluntary compliance, such as income taxes; for the opposite reasons, more democratic regimes may turn toward those types of taxes. Thus, we may expect that higher degrees of democratic liberties may lead to higher direct to indirect tax ratios. But the political color of democratic regimes may also have an impact on the direct to indirect ratio. Kenny and Winer (2006) find that socialist governments substituted toward corporate taxes on goods and services, which does not lead to a clear prediction in terms of direct versus indirect taxes. Another important political factor may be collective preferences for redistribution and overall more equitable societies. We may assume that ‘redistribution’ is a normal or even superior good with income elasticity positive or greater than one; if so, the variable per capita income may capture this effect. Moving on to the ‘supply factors’, we need to identify features that make it easier (more difficult) to raise tax revenues from different sources. In the list are those that Kenny and Winer call ‘tax base effects’, meaning that countries will be attracted to use taxes for which there are relatively larger tax bases available. For example, major oil-producing countries may have larger non-tax revenues shares and also easy access to additional revenues via the corporate income tax due to the profits from the exploitation of oil reserves. In this case we would expect the significance of oil production in a country to be associated with higher direct to indirect tax ratios. Similarly, a higher direct to indirect ratio may come from relatively larger tax bases for personal income tax (measured by real GDP per worker), and payroll taxes (proxied by the labor force participation). On the other hand, taxes on domestic goods and services have larger bases in the formal sector in countries in which more people live in urban areas. This may lead to a lower direct to indirect tax ratio. Similarly, countries with more open economies would tend to rely more on trade and other indirect taxes given the easier collection of VAT and excises at the ports of entry. An additional set of supply factors, not entirely distinguishable from the previous one, is that of ‘administration costs’, including among other things the ability to provide taxpayer services and conduct tax

Direct versus indirect taxation 53 enforcement activities. Urbanization may capture the effect of administration costs on tax structure. Because of the higher population density in urban areas, monitoring of tax compliance may become less expensive, implying overall higher tax compliance. However, the impact of urbanization on tax compliance may be more complicated than that (Kau and Rubin, 1981). Because people live close to their neighbors in urban settings, informal transactions become more feasible, which in turn will tend to reduce tax collections of both indirect and direct taxes. Summarizing, the set of observable characteristics Xit we include as explanatory variables in our analysis of the tax mix is as follows: 1.

Demand factors: – total revenue (including tax and non-tax revenue) to GDP ratio; – log population, normalized by dividing it by the annual mean of this variable; – dummy for country’s formal federal structure; – expenditure decentralization, calculated as the ratio of state and local expenditures to total expenditures; – democracy index; – dummy for socialist government; – log GDP per capita, normalized by dividing it by the annual mean of this variable;

2.

Supply factors: – domestic crude petrol per capita production; – labor force participation; – trade openness, measured as the ratio of imports plus exports to GDP; – share of agriculture in GDP19 – globalization index; – percentage of urban population.20

To be sure, there are several departures in our approach from that used by Kenny and Winer (2006). Besides the different dependent variables, we employ a slightly larger sample of 116 developed, developing, and transitional countries and observe a longer time period, between 1972 and 2005. Furthermore, we utilize a different regression specification based on some theoretical assumptions on the determinants of the tax mix and use annual data rather than creating subsample averages.21 Our analysis covers the full sample of countries but we also run separate regressions for developed

54

The Elgar guide to tax systems

and developing countries to check for potentially separate effects due to differences in economic structure. Like Kenny and Winer, we allow for the endogeneity of certain righthand variables. But, in addition, we correct for autocorrelation. Let’s first address the possible presence of endogeneity among some of the explanatory variables. Kenny and Winer (2006) account for the possible endogeneity of government size (proxied by the ratio of total revenue to GDP) although with or without correction for endogeneity, the inclusion of this variable in the regression has very little impact on their results.22 Given that our dependent variable is the ratio between direct and indirect taxes rather than individual tax instruments, the reverse causality between the direct to indirect tax ratio and total revenue to GDP variable is less likely to be present. We test for endogeneity in total revenue to GDP using the same instruments as Kenny and Winer (2006), absolute latitude of the country’s largest city, scaled to take values between 0 and 1, and voter turnout rate, but fail to detect it.23 A second issue is the need to correct for autocorrelation.24 Since we detect the existence of the first-order panel-specific autocorrelation in our model, we estimate the model with panel corrected standard errors (PCSEs), as suggested by Beck and Katz (1995).25,26 Estimation Results Table 2.1 presents the estimated effects obtained by using the annual data and applying panel corrected standard errors to the full sample to correct for panel-specific autocorrelation.27 The highly significant and positive estimated effect of total revenue to GDP ratio suggests that countries with larger government size tend to rely more on direct taxes (10 percentage points increase in total revenue to GDP leads to an increase in the direct to indirect tax ratio by between 2.1 and 3.7 percentage points). For population size, recent evidence suggests that countries with smaller populations have relatively mobile tax bases and as a result they rely relatively less on corporate and personal income taxes than other countries (Hines and Summers, 2009). These countries instead rely more on expenditure-type taxes, tax on goods and services, and import tariffs. Our results strongly support those previous findings. The significant results for the federal structure dummy variable suggest that federal countries tend to rely relatively more on direct taxation. Furthermore, the degree of expenditure decentralization seems to be on average not significant in deciding the tax mix, but when we observe developed and developing countries separately, we find the expenditure decentralization to be significant in both subsamples, although the effect

55

Labor force participation (LFP)

Supply factors Tax base effect Crude petrol

Log(GDP per capita)

Socialist

Political preferences Democracy

Decentralization

Federal

Log(population)

Demand factors Scale effect Revenue to GDP

0.01 (0.01) 0.01 (0.01)

0.38* (0.22) −0.36 (0.34) −2.26 (2.20)

0.00 (0.01) 0.03** (0.01)

−0.42 (0.55) 0.00 (0.00) −2.92 (3.88)

1.45*** (0.42) 6.68 (14.80) 1.85 (3.83) −0.03*** (0.01)

Developed

Full

2.20*** (0.50) 29.39*** (9.86) −3.59*** (1.24) −0.01* (0.00)

(2)

(1)

0.06 (0.10) −0.01 (0.02)

0.72** (0.33) −0.25 (0.41) −1.45 (3.05)

2.15** (1.03) 67.08*** (24.23) −18.63*** (6.46) 0.01* (0.01)

Developing

(3)

0.01 (0.01)

0.02 (0.14) −0.16 (0.31) −2.41* (1.44)

1.59*** (0.42) 19.67*** (7.36) −13.36* (7.96) −0.01** (0.00)

Full

(4)

0.02** (0.01)

−0.18 (0.48) 0.00 (0.00) −6.56*** (2.41)

1.70*** (0.48) −4.37 (9.53) 22.76* (12.08) −0.02*** (0.00)

Developed

(5)

0.02 (0.01)

0.04 (0.14) −0.24 (0.37) −0.71 (1.74)

0.76 (0.70) 39.39*** (13.09) −36.76*** (13.85) 0.00 (0.00)

Developing

(6)

Table 2.1 Determinants of tax mix: 1972–2005, fixed effects, annual data (dependent variable: direct to indirect tax ratio)

56

Source:

Note:

−7.14*** (1.76) −19.56** (9.21) 437 41 0.91

−11.41*** (2.73) 6.71 (17.30) 227 17 0.97

0.30 (0.22) −12.50*** (3.54) −1.52 (1.12)

Developed

Full

0.33** (0.14) 0.62 (1.48) −2.77*** (0.70)

(2)

(1)

−9.47*** (2.83) −45.44** (17.77) 210 24 0.81

0.35** (0.16) 0.27 (1.69) −2.09** (0.82)

Developing

(3)

−6.14*** (1.61) 0.00 (0.00) 635 63 0.93

0.13 (0.12) −0.29 (1.02) −1.37*** (0.53)

Full

(4)

−10.61*** (1.70) 0.00 (0.00) 328 24 0.96

−0.39** (0.20) −7.81*** (2.08) 0.01 (0.71)

Developed

(5)

Authors.

Panel corrected standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%.

Observations Number of id R-squared

Constant

Administration costs Urbanization

Globalization

Agriculture

Supply factors Tax base effect Openness

Table 2.1 (continued)

−6.60*** (1.90) 0.00 (0.00) 307 39 0.83

0.13 (0.14) −0.60 (0.95) −2.03*** (0.64)

Developing

(6)

Direct versus indirect taxation 57 has the opposite signs in the subsamples, negative for developed countries and positive for developed countries; but note that the economic effect is quite small in both cases.28 For factors representing political preferences, we find that on average more democracy implies higher direct to indirect tax ratios; however, for the subsamples, the coefficient for developed countries is also significant but takes an unexpected negative sign.29 We find no evidence that countries in transition from socialism tend to show a marked reliance on either direct or indirect taxes.30 The estimated coefficients for GDP per capita are not statistically significant, except for developing countries, which takes a negative sign. On the supply side, the effect of globalization on the tax ratio appears to be statistically significant and negative, which is consistent with the widely accepted conjecture that with increasing globalization all countries are becoming small open economies being forced to lower their reliance on direct taxes vis-à-vis indirect taxes.31 Furthermore, in line with the expectations, taxes on domestic goods and services are more important in countries in which more people live in urban areas. Our results suggest a very significant negative and robust effect of urbanization, our proxy for domestic indirect tax base, on the direct to indirect tax mix.32 Finally, a more educated population can facilitate the implementation of taxes, such as the personal income tax, that require more ability to fill out sophisticated tax forms. Our results indicate that increased education leads to greater reliance on direct taxes. This result is quite robust to alternative specifications.

5

RELEVANCE OF THE DIRECT TO INDIRECT TAX RATIO IN THE REAL ECONOMY

In this section of the chapter we use a fairly large panel data set of developing and developed countries to explore the empirical significance of the direct to indirect tax choices countries make for their tax systems on four important dimensions of macroeconomic performance: economic growth, macro-stability, income inequality, and foreign direct investment. On Economic Growth With little doubt the most commonly thought but nevertheless controversial effect of high reliance of tax systems on direct taxes versus indirect taxes is its negative impact on economic growth. In the review of the literature above we have seen that a number of recent studies provide empirical

58

The Elgar guide to tax systems

evidence, albeit not always consistent, of the negative long-term growth effects of direct taxes, particularly corporate income taxes and progressive personal income taxes. Our goal here is to explore more specifically the potential role of the direct to indirect tax ratio on economic growth. To extrapolate from the most recent empirical literature we should anticipate that higher direct tax to indirect tax ratios should lead to lower rates of economic growth all other things being equal. The empirical literature on economic growth is vast and ever growing.33 Our analysis in this section builds on a fairly recent study by Lee and Gordon (2005), analyzing the potential role of corporate taxes on economic growth and based on a panel data set for 70 countries over the period 1980–97. In this section, besides adding the direct to indirect tax ratio variable, we introduce several other modifications to Lee and Gordon’s (2005) approach.34 First, we extend the sample period by eight years to 1972–2005, while we proceed to divide it into seven subsample periods: one three-year period (1972–74), five five-year periods (1975–79, 1980–84, 1985–89, 1990–94, 1995–99), and one six-year period (2000–05); following Lee and Gordon we regress the average subsample GDP (real) per capita growth rate on the tax variable and the other control variables. Second, we expand the sample size from 70 to 116 countries. We proceed to estimate the following equation: GDPgit 5 aTaxRatioit 1 Xitb 1 ui 1 eit, i 5 1, . . ., n, t 5 1, . . ., T (2.2) where i indicates country and t denotes subsample period, GDPg represents average subsample GDP (real) per capita growth rate, TaxRatio is the average subsample direct to indirect tax ratio, Xit represents a set of control variables affecting GDP growth, including: GDP per capita in the initial subsample year in US$ 10 000, the initial subsample year top marginal corporate tax rate, the initial subsample year of the primary school enrollment, average subsample openness (measured as sum of import and export to GDP), the average subsample International Country Risk Guide (ICRG) index, the average subsample population growth rate, and average subsample inflation rate. Before we proceed, we need to address several issues concerning the estimation strategy. First, there is the possibility that the direct to indirect tax ratio variable is endogenous; for example, countries with faster growth may increasingly rely on direct taxes for equity or economic stability reasons. In order to address this issue, we use an instrumental variable for the tax ratio variable that is calculated in a similar way to the instrumental variable for the corporate tax rate used by Lee and Gordon (2005). In particular, we first instrument each direct to indirect tax ratio observation

Direct versus indirect taxation 59 with the weighted average of the tax ratios for all other countries in the corresponding year, where the weights are the inverse of the distance (as described below) between the two countries. The value of the tax ratio instrumental variable for country i in year t, TaxRatioIVit is, therefore, calculated as: TaxRatioIVit 5

1

n 1 TaxRatiojt; i 2 j 1 a j51 dj a j51 d j

(2.3)

n

where dj is the distance between the largest cities in country i and country j, and TaxRatioijt is the tax ratio in country j in year t. The underlying intuition for using this particular instrument is that economic growth in a country relative to others generally should not have an effect on the design of the tax mix of those other countries, so the dependent variable should not be correlated with the instrument. On the other hand, the design of the tax mix in a country should be affected by the design of the tax mix in the neighboring countries, this effect being especially strong in the case of small countries.35 Because we use the corporate tax rate in our regressions, which is the tax variable of interest for Lee and Gordon (2005), we also reproduce their steps regarding the instrumentation of the corporate tax rate variable. Second, before applying the instrumental variable methodology, we perform a Hausman test for endogeneity concerning the direct to indirect tax ratio variable and the corporate tax rate. The Hausman tests reject the null hypothesis that OLS is a consistent estimator, providing support for using instrumental variables methodology. The overidentification test has a P-value of 0.9, suggesting that we fail to reject the hypothesis that all excluded instruments are exogenous. Third, following Lee and Gordon (2005) we use a battery of estimation approaches: first, we employ ordinary least squares regression, robust regression, and median regression to check for the robustness to the outliers; second, we use panel estimation including fixed effects36 regression and the instrumental variable regression with country dummies. The estimation results are shown in Table 2.2 for the case when our main independent variable of interest, the direct to indirect tax ratio, includes property taxes as direct taxes.37 The most relevant finding from our perspective is that higher direct to indirect tax ratios appear to have a significant and negative impact on economic growth. From the robust regression and median regressions in Table 2.2 we can see that the estimated coefficient for the tax ratio is quite robust to outliers. After controlling for individual country effects, the impact of the tax ratio variable on

60

Primary enrollb Av. openness

GDP per capb

Corp tax rateb

Tax ratioa

Table 2.2

−0.323** (0.147) −0.034** (0.014) −0.775*** (0.246) 0.016 (0.016) 0.641** (0.285)

Robust

OLS

−0.248 (0.179) −0.028* (0.015) −0.891*** (0.243) 0.026 (0.017) 0.672** (0.332)

(2)

(1)

−0.338* (0.178) −0.031* (0.017) −0.929*** (0.319) 0.041** (0.020) 0.569 (0.375)

Median

(3)

−0.872*** (0.284) −0.052*** (0.019) −1.924*** (0.549) −0.035 (0.030) 3.825*** (1.156)

Fixed Effect

(4)

−3.910** (1.575) −0.092*** (0.033) −1.654*** (0.559) −0.089** (0.045) 4.475*** (1.327)

Full

(5)

(8)

−5.575** −2.429 (2.774) (2.791) −0.055* 0.057 (0.032) (0.099) −2.401*** −11.247* (0.705) (6.304) −0.141** −0.076 (0.070) (0.052) 2.282 3.880 (1.527) (4.101)

−3.805 (2.382) 0.070 (0.125) −2.373 (5.287)

20.107 (20.683)

Full

Country Dummies + IV

(7)

Developed Developing

(6)

(10)

0.272 (4.058)

−1.733*** −12.399*** (0.557) (3.565) −0.145* −0.076** (0.081) (0.038) 3.279*** 2.179 (0.981) (1.743)

−4.293 (3.321)

Developed Developing

(9)

Direct indirect tax ratio and economic growth regressions for subsample periods, 1972–2005 (dependent variable: GDP per capita growth rate for subsample periods)

61

0.316 0.319* 0.499** 0.417 (0.195) (0.170) (0.221) (0.393) −1.211*** −1.107*** −1.057*** −1.084** (0.227) (0.177) (0.231) (0.425) –0.007*** –0.006*** −0.006*** −0.003** (0.001) (0.002) (0.002) (0.002) 2.337 3.325* 0.302 8.288** (1.924) (1.722) (2.230) (3.471) 197 197 197 197 0.37 0.34 0.28

Source:

Authors.

Notes: Robust standard errors in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. a. Property taxes treated as direct taxes. b. These variables take values at the initial subsample year.

Observations R-squared

Constant

Av. inflation

Av. ICRG Index Pop. gr. rate

0.826* −0.791 (0.449) (0.713) −1.461*** −0.838 (0.518) (0.759) −0.002** −0.280*** (0.001) (0.091) 14.446*** 27.024** (5.395) (11.832) 197 120 0.73 0.57

1.018 −3.081 (0.887) (2.942) −4.337*** 0.620 (1.585) (1.527) −0.002 −0.008** (0.002) (0.003) 17.142* −9.339 (9.300) (18.071) 77 275 0.87 0.60

−0.489 (0.629) −0.544 (0.956) −0.137** (0.055) 29.040** (13.760) 135 0.54

−0.153 (1.220) −2.077*** (0.770) −0.006** (0.002) 19.791*** (5.755) 140 0.66

62

The Elgar guide to tax systems

economic growth remains negative and significant and this overall result is also maintained after we control for the potential endogeneity of the tax ratio and corporate tax rate variables. However, when we divide the full sample into subsamples for developed and developing countries some of the results change. In the case of developed countries, the direct to indirect tax ratio continues to have a negative and highly significant effect on economic growth. In the case of developing countries, even though the coefficient is negative, it is not statistically significant. For the rest of the control variables, we obtain comparable results to those in the previous literature including Lee and Gordon (2005). The coefficient on initial subsample GDP per capita is negative and significant, which is consistent with the assumption of the conditional convergence of growth rates reported in previous studies (Barro, 1991; Mankiw et al., 1992; Kneller et al., 1999). Inflation affects economic growth rate negatively, supporting the hypothesis that, among other things, inflation increases investment uncertainty and, therefore, reduces economic agents’ incentives to invest (Padovano and Galli, 2001 and 2002; Romero-Avila and Strauch, 2008). Trade openness has a positive and significant effect on the growth rate, which is consistent with previous findings (Dollar, 1992; Edwards, 1998; Frankel and Romer, 1999; and Dollar and Kraay, 2003). The results for institutional factors (measured by the ICRG index) are not robust to changes in estimation methodology; there is also less consensus in the empirical literature concerning the role of these factors.38 Lastly, note that the results for the rest of the control variables are overall of lower statistical significance for the subsample of developing countries. Because of the very high standard errors and high R-squared of the regression we may suspect the presence of multicollinearity.39 On Macroeconomic Stability One of the well-known benefits of direct taxes is that they can act as automatic stabilizers.40 Progressive personal income taxes tend to withdraw proportionally more private income during economic expansions and less so during contractions of the economy. Similarly, corporate income taxes yield higher revenues when profits are high in the expansion phase of the business cycle but they drop considerably in the contraction phase. On the other hand, indirect taxes, such as the VAT or excises, lack these stabilizing features. To explore the role of tax structure in terms of the direct to indirect tax composition on macroeconomic stability, we employ a simple regression model in which we regress the volatility of economic

Direct versus indirect taxation 63 growth, measured by the standard deviation of GDP growth rate within each subsample period, on the direct to indirect tax ratio and a set of other explanatory variables. For the basic specification of the regression equation we follow the work of Easterly et al. (2000) and Beck et al. (2001). These other control variables include the ‘volatility of inflation’ (measured by the standard deviation of the inflation rate within the subsample period), ‘average openness’, and ‘average GDP per capita’. The direct to indirect tax ratio captures the effect of automatic stabilizers on economic stability while average openness and the volatility of inflation are included as proxies for the degrees to which the economy is exposed to real and monetary shocks; average GDP per capita is intended to capture any possible relationship between wealth and economic volatility. For estimation purposes we use the sample of 116 developed, developing and transitional countries, over the period 1972–2005 and as in other sections we divide the sample in seven subsample periods (one three-year period [1972–74], five five-year periods [1975–79, 1980–84, 1985–89, 1990–94, 1995–99], and one six-year period [2000–05]). For the independent variable of interest, the direct to indirect tax ratio, we use two alternative measures, one treating property tax as a direct tax (‘tax ratio 1’) and the other treating it as an indirect tax (‘tax ratio 2’); below we only report the results for ‘tax ratio 1’ since the results obtained using ‘tax ratio 2’ are fundamentally the same. We proceed to estimate the following equation: SD_GDPgit 5 a1TaxRatioit 1 a2TaxRatio2it 1 a3TotalTaxit 1 a4TaxRatioit * TotalTaxit 1 Xitb 1 eit; i 5 1, . . ., n, t 5 1, . . . T (2.4) where i indicates country and t denotes subsample period. The dependent variable, SD_GDPg, is measured as the subsample standard deviation of annual GDP (real) per capita growth rate. TaxRatio is the average subsample direct to indirect tax ratio, and TotalTax is the average subsample total tax to GDP. Finally, Xit represents a set of control variables affecting GDP growth volatility, including: the subsample standard deviation of M1 annual growth rate,41 the average subsample openness (measured as sum of import and export to GDP), and the average subsample GDP per capita. For the choice of the correct panel data estimation procedure, we perform a Hausman test for selecting between the fixed and random effects approaches and fail to reject the null hypothesis that the coefficients estimated by the efficient random effects estimator are the same as the ones

64

The Elgar guide to tax systems

Table 2.3a

Tax ratio and economic stability, subsample periods, 1972– 2005, random effects estimation (dependent variable: standard deviation of annual GDP growth rate) (1)

(2)

(3)

(4)

Random Effects Full Tax ratio

−0.934 (0.663) Tax ratio 0.246** squared (0.110) Total tax 1.445 (3.693) Tax ratio −0.028 * Total tax (0.021) SD M1 1.909 (11.644) Average 1.061** openness (0.422) Average GDP 0.042*** per capita (0.003) Constant 60.346 (71.999) Observations 256 Number of id 89 R-squared 0.69

Developed Developing −1.043 −1.186** (2.092) (0.543) 0.631* 0.004 (0.351) (0.095) 18.693 −1.076 (13.715) (2.977) −0.104* 0.033 (0.057) (0.020) −36.364 −3.915 (710.348) (9.405) −0.970 0.091 (1.539) (0.369) 0.025*** 0.080*** (0.008) (0.005) 237.137 36.391 (349.230) (53.486) 59 197 17 72 0.76 0.82

(5)

(6)

Random Effects IV Full

Developed Developing

−1.556* (0.841) 0.240* (0.130) −4.449 (5.245) 0.006 (0.028) 5.428 (12.329) 3.902*** (1.331) 0.039*** (0.004) −50.070 (106.041) 256 89 0.57

−0.896 −3.383** (2.118) (1.651) 0.640* 0.201 (0.353) (0.210) 21.537 −19.281 (14.712) (12.496) −0.111* 0.091 (0.059) (0.055) 137.684 16.403 (780.403) (19.007) −2.009 8.664* (2.440) (4.887) 0.023*** 0.025 (0.008) (0.031) 237.816 −131.435 (350.788) (151.119) 59 197 17 72 0.75 0.31

Note: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%. Source: Authors.

estimated by the consistent fixed effects estimator, allowing us to use the random effects procedure.42 We may suspect the potential endogeneity of the trade variable. For example, countries with more stable output growth may be more inclined to liberalize trade barriers. Even though the Hausman test fails to reject the hypothesis of no endogeneity for trade openness, we perform both random effects estimations, with and without an instrumental variable for trade openness. The results obtained using random effects are reported in Table 2.3a, while Table 2.3b presents the results obtained by applying the fixed effects estimation methodology. We instrument trade openness with the weighted average of the trade openness for all other countries in the

Direct versus indirect taxation 65 Table 2.3b

Tax ratio and economic stability, subsample periods, 1972–2005, fixed effects estimation (dependent variable: standard deviation of annual GDP growth rate) (1)

(2)

(3)

(4)

Fixed Effects Full Tax ratio

−1.037 (1.517) Tax ratio 0.248 squared (0.204) Total tax 12.643 (11.136) Tax ratio −0.017 * Total tax (0.047) SD M1 7.572 (14.603) Average 4.922 openness (4.642) Average GDP 0.002 per capita (0.018) Constant −243.824 (308.514) Observations −1.037 Number of id (1.517)

(5)

(6)

Fixed Effects IV

Developed

Developing

Full

Developed Developing

1.485 (8.445) 1.088 (0.723) 25.674 (58.855) −0.125 (0.223) −544.237 (855.686) 0.390 (8.625) 0.010 (0.032) −423.950 (1,913.955) 1.485 (8.445)

−4.546 −1.037 1.485 (5.028) (1.517) (8.445) 0.719 0.248 1.088 (0.564) (0.204) (0.723) 6.059 12.643 25.674 (26.234) (11.136) (58.855) −0.043 −0.017 −0.125 (0.120) (0.047) (0.223) 15.437 7.572 −544.237 (23.429) (14.603) (855.686) 24.257 4.922 0.390 (23.924) (4.642) (8.625) −0.295 0.002 0.010 (0.273) (0.018) (0.032) −671.252 −243.824 −423.950 (694.173) (308.514) (1,913.955) −4.546 −1.037 1.485 (5.028) (1.517) (8.445)

−4.546 (5.028) 0.719 (0.564) 6.059 (26.234) −0.043 (0.120) 15.437 (23.429) 24.257 (23.924) −0.295 (0.273) −671.252 (694.173) −4.546 (5.028)

Note: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%. Source:

Authors.

corresponding subsample period, where the weights are the inverse of the distance (as described below) between the two countries. The value of the trade instrumental variable for country i in year t is, therefore, calculated as: TradeIVit 5

1

n 1 Tradejt; i 2 j 1 a j51 dj a j51 d j

(2.5)

n

where di is the distance between the largest cities in country i and country j, and Tradejt is the trade openness of country j in year t. In the regressions we allow for a non-linear relationship between the tax

66

The Elgar guide to tax systems

ratio variable and economic stability by including a squared term. In addition, the different specifications are estimated using separate subsamples for developed and developing countries to allow for potentially different responses due to different fundamental economic structures. Overall, the results provide strong evidence that the direct to indirect tax ratio has a significant negative effect on economic volatility. There is only weak evidence that this relationship between the direct to indirect tax ratio and economic volatility may be quadratic.43 When the quadratic term is significant, the implied threshold for the tax ratio starting to have a positive effect on economic instability has an improbable value as high as 6.5 times greater reliance on direct taxes relative to indirect taxes.44 Furthermore, within the subsample of developed countries, the direct to indirect tax ratio has more automatic stabilizing power in countries with higher total tax to GDP ratios. For the other control variables, it appears that the volatility of M1 has no significant effect on economic stability. On the other hand, trade openness is shown to be positively correlated with economic volatility in the cases of the full sample and the subsample of developing countries; this suggests that the more exposed the economy is to outside real shocks, the more prone the economy is to volatility, as previously found in Easterly et al. (2000) and Beck et al. (2001). Last, average GDP per capita has a positive effect on economic volatility, and this effect is more pronounced among the subsample of developed countries. On Income Inequality Our interest here is to investigate the importance of the direct to indirect tax ratio as a determinant of income inequality in a country. The general presumption is that greater vertical equity and more equal income distributions require a more progressive tax system, which means that direct taxes (generally expected to be progressive) would need to be relatively more important than indirect taxes (typically expected to be regressive or less progressive) in tax systems. The evidence in the empirical literature on this issue is mixed,45 and our own empirical findings in this section do not offer any strong support to the conjecture that the direct versus indirect composition of taxes plays an important role on observed inequality in distribution of income. However, this conclusion is subject to the important caveat of the difficulties involved in measuring inequality in income distribution across countries and over time. In investigating the importance of the direct–indirect tax balance on income inequality and redistribution we focus on the evolution of the Gini coefficient for income distribution. The Gini coefficient is computed on the

Direct versus indirect taxation 67 basis of income distributions using different concepts of income, including gross income, net income, and consumption. This presents some measurement and comparability issues that we can only partially address below. We are interested in finding out how the direct–indirect tax mix and a set of other explanatory variables has affected the Gini coefficient over time in our sample of countries. The empirical model we estimate is: Giniit 5 a1TaxRatioit 1 a2TotalTaxit 1 a3TaxRatioit*TotalTaxit 1 Xitb 1 GiniConceptit 1 eit; i 5 1, . . ., n, t 5 1, . . ., T

(2.6)

where i indicates country and t denotes years. Gini is the Gini coefficient as a measure of income inequality46 over time and across countries; Xit is the set of observable characteristics that affect income inequality, including: the initial Gini coefficient, the direct–indirect tax ratio, total tax collection to GDP, GDP per capita growth rate, private credit as a percentage of GDP, labor force participation, openness (measured by the ratio of import plus export to GDP) dependency ratio, and dummy for EU15. The set of explanatory variables, except for the direct–indirect tax ratio, represents a consensus specification in the empirical literature on aggregate income distribution. In the estimation we employ annual data for a sample of 116 developed, developing, and transitional countries, over the period 1972 to 2005. For the estimation we apply a 2SLS procedure to control for the potential reverse causality between income inequality and the financial system (measured by the share of private credit in GDP) and between income inequality and the direct to indirect tax mix. As suggested by Beck et al. (2004), the reverse causality between income inequality and private credit might take the form that reductions in inequality may lead to higher demand for more efficient financial systems. Following La Porta et al. (1999) and Beck et al. (2004), we use as instrumental variables for the financial system, latitude (the scaled absolute value of) as well as legal origin (English, French, and German). We have already discussed in the previous section the rationale for using these variables as instruments for GDP per capita growth; a similar intuition applies to the case of financial development and so it will not be repeated here. Furthermore, the potential endogeneity between income inequality and the tax mix may arise from the fact that countries with higher income inequality may tend to rely more on direct taxes in order to reduce it. In order to test and correct for endogeneity in the tax mix ratio, we instrument the direct–indirect tax ratio with the weighted average of the tax ratio for all other countries in

68

The Elgar guide to tax systems

the corresponding year, where the weights are the inverse of the distance (as described below) between the two countries. The value of the tax mix instrumental variable for country i in year t is, therefore, calculated as:

TaxRatioIVit 5

1

n 1 TaxRatiojt; i 2 j a j51 n 1 dj a j51 d j

(2.7)

where di is the distance between the largest cities in country i and country j, and TaxRatiojt is the direct–indirect tax mix of country j in year t. The Hausman test for endogeneity rejects the null hypothesis that OLS is a consistent estimator for both private credit and the tax ratio, providing support for using the 2SLS procedure. For the instrumental variables 2SLS procedure we estimate first stage equations as below with latitude, and legal origin as instruments in the private credit equation and the weighted tax ratio for all other countries as the instrument in the tax mix equation:47 Creditit 5 Zitd 1 ai 1 uit; i 5 1, . . ., n, t 5 1, . . ., T

(2.8)

TaxRatioit 5 Zitg 1 ai 1 nit; i 5 1, . . ., n, t 5 1, . . ., T

(2.9)

where Zit includes all exogenous variables from Equation (2.6) plus the instruments, and E (Zrituit) 5 E (Zritnit) 5 0. To carry out the panel estimation we perform a Hausman test for selecting between fixed and random effects estimation on the basis of the second-stage equation, which fails to reject the null hypothesis that the coefficients estimated by the efficient random effects estimator are the same as the ones estimated by the consistent fixed effects estimator, allowing us to use the random effects procedure. The estimation results are presented in Table 2.4. Our main interest is in the relationship between income inequality and the direct to indirect tax structure. The expectation, based on the past literature, is that of a negative relationship, albeit possibly weak between the direct to indirect tax mix and income distribution. The results in Table 2.4, overall, provide at best weak support for the conjecture. Our results suggest that the effect of the tax ratio on income inequality depends on the size of the taxation system: in countries with relatively smaller tax systems, the tax ratio tends to have a positive effect on income inequality, whereas its negative (equalizing) effect increases with enlarged tax systems. For the full sample, the tax ratio mix has negative effect on the Gini coefficient (reducing income inequality) in countries with shares of total taxes in GDP larger than 0.29.

Direct versus indirect taxation 69 Table 2.4

Direct to indirect tax ratio and income inequality, 1972–2005, 2SLS estimations (dependent variable: Gini coefficient (%))

Initial Gini Tax ratio Total taxes Tax ratio * Total taxes Private credit GDP per capita growth LFP Dependency ratio EU15 Openness Net income Gini concept Consumption Gini concept Constant Observations Number of id R-squared

(1)

(2)

(3)

Full Sample

Developed Countries

Developing Countries

0.74*** (0.09) 10.04* (5.95) 60.28* (33.84) −35.21* (19.58) −4.73* (2.53) −0.02 (0.11) 0.06 (0.08) 3.55 (6.29) −3.48** (1.48) 2.04** (0.80) −2.11*** (0.77) −3.69*** (0.88) −7.18 (17.42) 447 62 0.64

0.30** (0.13) 15.05** (6.68) 95.02*** (36.77) −44.31** (18.86) 3.51 (3.20) 0.10 (0.16) −0.07 (0.13) 22.29* (11.79) −0.19 (1.89) 0.35 (1.43) −1.55 (1.12) −2.53** (1.20) −19.66 (24.30) 274 25 0.31

0.65*** (0.13) 1.17 (6.67) 8.74 (25.84) −5.95 (24.48) 0.60 (4.69) −0.04 (0.12) −0.11 (0.09) −9.66 (8.20) 0.00 (0.00) 0.57 (1.74) 0.10 (0.77) −2.82*** (0.71) 29.49** (12.58) 173 37 0.65

Note: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%. Source:

Authors.

This threshold is larger in developed countries (0.34); for the subsample of developing countries there is no statistically significant effect. This latter pattern seems to fit the conventional wisdom on the low impact of tax systems on the distribution of income for developing countries (Bird and

70

The Elgar guide to tax systems

Zolt, 2005; Harberger, 2006). But we must note that once we control for unobserved individual country effects by adding the individual country dummies, the importance of direct to indirect tax ratio in terms of income inequality seems to practically vanish in all equations. For the other control variables, the results largely follow those in the previous empirical literature on the determinants of aggregate income distribution.48 Following Beck et al. (2004), we include the initial level of the Gini coefficient at the beginning of the observation period as one of the explanatory variables to capture the country’s initial conditions. The initial level of inequality has a strong positive effect, especially for developing countries, on observed inequality. The coefficient for the level of financial development takes a negative and significant sign, as expected, but only for the full sample. There is a broad literature emphasizing the role of education as one of the major factors affecting the degree of income inequality. Even though policy-makers usually justify higher educational spending as an effective tool for reducing income inequality, the theoretical predictions about this relationship are ambiguous and the empirical findings are not consistent.49 Years of schooling in the total population as a measure of education has a positive and significant effect on income inequality but only for developing countries. We also include a dummy for the 15 old European Union member states to control for the generally higher social welfare expenditures in those countries; however, this variable is not statistically significant. The dependency ratio variable appears to have a highly positive and significant effect on inequality in developed countries. The evidence in the literature on the effect of trade openness on income inequality is inconclusive. Barro (2000) finds a positive relationship between trade openness and income inequality, while Calderon and Chong (2001) and Dollar and Kray (2002) do not find any significant relationship. Our results provide some evidence of a positive effect of trade openness on income inequality, but this effect vanishes in the subsamples. Finally, we control for variations in the conceptual measurement of the Gini coefficient, and as expected, we find that measured income inequality is significantly smaller when net income or consumption are used vis-à-vis gross income. On Foreign Direct Investment Because of globalization and the increasing international mobility of the factors of production, especially of capital, there has been a lot of interest in the literature studying FDI (foreign direct investment) flows and how

Direct versus indirect taxation 71 corporate income taxes and other direct taxes may affect these flows.50 Thus, the choice of tax structure and in particular the direct to indirect tax ratio can be anticipated to have potentially significant effects on FDI flows. That is the question researched in this section. In this section, again, our strategy in analyzing the potential role of the direct to indirect tax ratio in FDI flows consists of including the tax ratio with a set of other control variables in a general specification that has been commonly used in the empirical literature on the determinants of FDI. Because data availability is more of a problem with respect to FDI, we are limited to using an annual panel data set for 53 developed and developing countries over the period 1984–2005. We use two alternative measures of FDI: global net FDI inflows to GDP ratio from UNCTAD,51 and the ratio of net FDI inflow from the United States to GDP from the Bureau of Economic Analysis (BEA). As we have done in all the previous sections, we employ two alternative definitions of the direct to indirect tax ratio, depending on whether property taxes are categorized as direct or indirect taxes. In addition, the analysis will be performed for the full sample of countries and for the two subsamples of developed and developing countries. The empirical model we estimate is FDIit 5 Xitb 1 ui 1 eit; i 5 1, . . ., n, t 5 1, . . ., T

(2.10)

where i is an index for country and t an index for year, and FDI is the net foreign direct investments inflow (total or from the US) to GDP over time and across countries; Xit is the set of observable characteristics that affect FDI inflow, including: the direct to indirect tax ratio, GDP per capita, average effective tax rate, infrastructure (proxied by the number of telephone lines), education, and political and institutional variables (democracy, corruption, and bureaucracy). The set of explanatory variables, except for the direct to indirect tax ratio, represents a consensus specification in the empirical literature on foreign direct investments. Finally, ui represents time-invariant individual country effect. In terms of estimation issues, the Hausman test for fixed versus random effects rejects the null hypothesis that the coefficients estimated by the efficient random effects estimator are the same as the ones estimated by the consistent fixed effects estimator, indicating the need to apply the fixed effects procedure. In order to account for individual country effects, we include a set of country dummies in our estimation model. Next, we test for the presence of a non-linear relationship between the tax mix and FDI but fail to detect it. Finally, we detect the existence of panel-specific

72

The Elgar guide to tax systems

Table 2.5

Tax ratio and foreign direct investments: 1984–2005 (dependent variables: total FDI net flow and net flow from the US) (1)

(2)

Full Sample

Tax ratio GDP per cap Labor cost Average effected tax rate (AETR) AETR sq Telephone landlines Secondary school enrollment Democracy Corruption Bureaucracy Constant Observations Number of id R-squared

Total FDI

FDI from US

−0.057* (0.034) −0.010 (0.016) 0.002 (0.010) −0.845** (0.406) 0.780** (0.374) 0.165* (0.095) −0.001 (0.002)

(3)

(4)

Developed Countries

(5)

(6)

Developing Countries

Total FDI

FDI from US

−0.033** (0.014) 0.013* (0.008) −0.007 (0.004) −0.294** (0.125)

−0.157** (0.073) −0.015 (0.018) 0.004 (0.012) −1.148* (0.600)

−0.063** (0.027) 0.015* (0.008) −0.009* (0.005) −0.441** (0.181)

0.002 (0.002) 0.003 (0.003) 0.001 (0.000) 0.016 (0.040)

−0.014 (0.012) −0.020** (0.009) −0.001 (0.001) −0.275 (0.193)

0.302** (0.122) 0.068** (0.027) −0.001 (0.000)

0.957* (0.539) 0.197** (0.090) −0.001 (0.001)

0.406** (0.172) 0.072** (0.030) −0.001 (0.001)

−0.024 (0.038) 0.004 (0.007) 0.000* (0.000)

0.269 (0.186) 0.079*** (0.021) −0.001 (0.001)

−0.040 −0.113* (0.073) (0.062) 0.007 0.008 (0.021) (0.007) 0.124** −0.032 (0.062) (0.029) −0.200 0.295*** (0.231) (0.107) 379 374 42 42 0.30 0.48

−0.292 −0.173 (0.353) (0.145) 0.032 0.005 (0.031) (0.009) 0.295** −0.075 (0.121) (0.047) 0.000 0.000 (0.000) (0.000) 257 253 25 25 0.33 0.57

Total FDI FDI from US

0.012 −0.100 (0.021) (0.067) −0.000 0.014 (0.003) (0.009) −0.001 0.019*** (0.003) (0.006) −0.048** 0.117 (0.020) (0.073) 122 121 17 17 0.58 0.07

Note: Panel corrected standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%. Source: Authors.

autocorrelation so we use throughout the PCSEs to correct for autocorrelation (Beck and Katz, 1995). Table 2.5 presents the results obtained when property taxes are classified as direct taxes in the numerator of the tax ratio variable.52 Our results show that that the direct to indirect tax ratio, our variable of interest, as

Direct versus indirect taxation 73 expected, affects both total and FDI from the United States inflows negatively with the coefficients being statistically significant for the full sample and the developed country subsample. However, the coefficients are statistically insignificant for developing countries. For the other control variables, the results reported in Table 2.5 are fairly standard in the FDI empirical literature. The coefficient for GDP per capita takes a positive sign whenever significant in the case of full sample and developed countries, suggesting that high-income countries tend to attract more investments from the United States, whereas the results for developing countries suggest the opposite, lower-income countries tend to attract more foreign direct investments from the United States. Positive and statistically significant results hold for telephone lines, suggesting that foreign investors are more attracted to a better infrastructure. However, the coefficients for the labor cost variable are mostly insignificant, except for FDI from the United States for developed countries where cheaper labor attracts more FDI. For the average effective tax rate (computed from the Bureau of Economic Analysis data for US firms), we find a statistically significant and robust non-linear relationship. Foreign investors are discouraged by higher average effective rates but this is so at a decreasing rate. In the regressions we also control for the effect of political and institutional variables (democracy, corruption, and bureaucracy). Corruption, measured by an index from 0 to 6, with 6 denoting least corruption, takes the expected positive sign, but it is mostly statistically insignificant. Bureaucracy, also measured by an index from 0 to 6, with 6 denoting the highest quality, has the expected positive sign whenever significant. Finally, we estimate very consistent negative and in some cases significant effects for democracy, suggesting that less democratic countries may be able to attract more foreign direct investment.53 This is in accord with some results in the previous literature.54

6

CONCLUSION

In this chapter we have examined the evolution and economic consequences over the last three decades of the direct to indirect tax ratio in the tax systems of a large number of developed and developing countries. Over this time period the average ratio of direct to indirect taxes for a sample of 116 countries has been on the increase and this movement has been more pronounced for developed countries than for developing countries. The underlying reasons for these trends have differed between the two groups of countries, with increases in Social Security contributions

74

The Elgar guide to tax systems

being the main driver in the case of developed countries and a fairly large decrease in the relative importance of customs taxes that has been only partially offset by an increase in domestic consumption taxes in the case of developing countries. While the optimal tax literature never provided quick or exact recipes to be followed in the design of tax structures, it has been understood that optimal tax design requires the use of both direct and indirect taxes leaving open what the optimal tax mix should be. The more recent empirical evidence based on endogenous growth models tells a robust story on the negative effects on the rate of economic growth of heavy reliance on different forms of direct taxation. But as we saw in several sections of this chapter the choice between direct and indirect forms of taxation may not be so clear. While lowering the direct to indirect tax ratio, it seems, would bring advantages in terms of economic growth and an enhanced competitive stand regarding FDI, it would also dampen the ability to rely on automatic stabilizers for the macroeconomy and possibly reduce the scope or ability for income redistribution policies. In terms of those potential trade-offs, it is interesting to note that developing countries, by choosing on average a much lower direct to indirect tax ratio than developed countries, seem to be giving a much heavier weight to economic growth and FDI flows than to the potential distributional and macroeconomic control issues. But, of course, the choice of tax mix by developing countries is also significantly based on administration and capacity issues. Our empirical findings provide a first-order approximation for quantifying the types of trade-offs policymakers would face in making choices on the overall tax mix. From our estimates and provided that the tax mix ratio is within some expected bounds, a 10 percentage point increase in the direct to indirect tax ratio on average would reduce economic growth and FDI inflows by 0.39 per cent and 0.57 per cent respectively, but at the same time it would also reduce economic volatility by 0.15 per cent and income inequality by about 1 percent. However, we need to recall that the equalizing effect of higher direct to indirect tax ratios on the income distribution is partially dependent on the size of tax system; the larger the ratio of taxes to GDP, the larger the equalizing effect of the tax mix ratio. We find a tax to GDP ratio threshold of 0.29 for the tax mix ratio to have equalizing effects on income distribution; this pattern seems to fit the conventional wisdom on the low impact of tax systems on income inequality, especially in the case of developing countries because of their generally smaller tax systems.

Direct versus indirect taxation 75

NOTES *

1.

2. 3. 4. 5. 6. 7.

8. 9. 10.

11.

12. 13. 14. 15. 16.

Presented at a Conference sponsored by the Savings Banks Foundation of Spain FUNCAS, and UNICAJA, Malaga, Spain, ‘Tax Systems: Whence and Whither (Recent Evolution, Current Problems and Future Challenges)’, 9–11 September, 2009. We are thankful to Jesus Ruiz-Huerta for helpful comments. Other definitions of direct and indirect taxes could be used that would likely produce similar classification results. For example, Poterba et al. (1986) define direct taxes as taxes on individuals, including income taxes and employee contributions for social insurance, and indirect taxes are defined as those collected from firms, including sales and value-added taxes, employer contributions for social insurance, and various excise taxes. For empirical estimation purposes in this chapter, given the data available, we will allow for several groupings of direct and indirect taxes. Essentially Hicks (1939) assumed identical individuals with perfectly inelastic labor supply (Atkinson, 1977). See, for example, the discussion in Lee and Gordon (2005). However, Watrin and Ullman (2008) using an experimental approach find that participants are less compliant with consumption taxes than with income taxes. Trends in Figures 2.1–2.5 are based on five-year moving averages. See Table 2A.3 in the Appendix at the end of the chapter for an accounting of the number of observations for each period and a discussion of the changes in definitions and reporting in the GFS data set (Box 2A.1). The policy thrust has also been the subject of theoretical criticism. For example, Emran and Stiglitz (2005) have argued that a revenue-neutral shifting from tariff to VAT is welfare worsening because of the existence of a large informal sector in developing countries; Munk (2008) has argued along similar lines because the allocation benefits from domestic taxes may be outweighed by increasing administrative costs. See also Kreickemeier and Raimondos-Møller (2008) on the negative effect on market access and questionable welfare effects. See also Johansson et al. (2008). Part of the reduction in the share of personal income taxes can be explained by a change towards flatter personal income tax schedules and a reduction in the top statutory income tax rates. Alternatively, their model can be interpreted as one of tax avoidance with different compliance costs. Some other authors have questioned the premise that direct taxes may be more difficult to evade than indirect taxes. See, for example, Kesselman (1993). Without evasion, there is equivalence between a uniform commodity and an income tax. However, with tax evasion, that equivalence is gone. Dahlby (2003) argues that both forms of consumption taxation, direct consumption tax in the form of expenditure tax and indirect consumption tax in the form of a sales tax, are needed because both types of taxes are subject to somewhat different forms of tax avoidance and tax evasion behavior. For a dissenting view see Mendoza et al. (1997) who provide evidence in support of Harberger’s (1964) claim that, although theory may predict that the mix of direct and indirect taxes is an important determinant of long-run growth and investment rates, in practice plausible changes in tax rates are unlikely to affect growth, even if they can alter moderately the investment rate. Borge and Rattso’s (2004) work for Norwegian local governments in 1996 supports the Meltzer–Richard hypothesis. See, for example, Martinez-Vazquez (2008). Nevertheless, Mino and Nakamoto (2008) warn that in the presence of heterogeneous agents with different preferences, the stabilizing power of progressive income taxation demonstrated in representative-agent models may not always be effective. Clustered by country. See Table 2A.4 in the Appendix for the results obtained using this definition of the dependent variable.

76

The Elgar guide to tax systems

17.

A few other papers have examined the composition of tax structures. Aizenman and Jinjarak (2006) evaluate the impact of globalization throughout the 1980s–90s on the vector of taxes collected by countries at varying stages of development. At the subnational level, Geys and Revelli (2009) investigate the economic and political determinants of the local tax mix in the Flemish region of Belgium. Hines and Summers (2009) also use the interaction of the natural logarithm (ln) of income and ln population as an additional explanatory variable; all variables are normalized by dividing them by their annual means. We re-estimate the model using percentage of shadow economy as an alternative variable for agriculture. However, due to the quite small number of observations on this variable, the sample size reduces significantly, causing some of the coefficients to lose the statistical significance. The simple correlation between agriculture and shadow economy is 0.5. The list of variables and data sources and descriptive statistics are presented in the Appendix, Tables 2A.1 and 2A.2, respectively. The exact replication of methodology and specification of Kenny and Winer (2006) but using the direct to indirect tax ratio as the dependent variable and five-year subsample periods rather that annual data produced similar but generally weaker results. We would expect that the effect of some of the explanatory variables on individual tax categories may be neutralized when the dependent variable is the tax mix ratio. The intuition behind potential endogeneity in size of government is that more efficient tax structures will lead to the growing size of the government sector. To correct for potential endogeneity in government size, Kenny and Winer use two instrumental variables, absolute latitude of the country’s largest city and voter turnout. Latitude is used because temperate zones have climate that is more agriculturally productive and less severe, enabling countries located in these zones to develop their economies faster (Landes, 1998; La Porta et al., 1999). Beck et al. (2004) also apply this same intuition to the case of financial development. The North Temperate Zone extends from the Tropic of Cancer (at about 23.5 degrees north latitude) to the Arctic Circle (at approximately 66.6 degrees north latitude). The South Temperate Zone extends from the Tropic of Capricorn (at approximately 23.5 degrees south latitude,) to the Antarctic (at approximately 66.6 degrees south latitude). The intuition behind using voter turnout is from Mueller and Stratmann (2003) who find that countries with higher voter turnout rates have more equal income distribution and larger government size (measured by expenditures and transfers to GDP). Note that for the voter turnout instrument we had to adapt the variable to our annual observations by interpolating the annual data between the election dates. We also test for but find no heteroskedasticity in our model. See Beck and Katz (1995) who show that the ordinary least squares (OLS) with the PCSE’ is the most proper approach for data sets with relatively many cross-sectional units (N) and relatively short time period (T). This approach is compatible with unbalanced panel estimation. The panel corrected standard errors are robust in the style of Huber-White standard errors. However, using the Huber-White rather than PCSE formula to calculate standard errors would be wrong because it ignores the fact that we assume there is a common variance structure within a cross-section unit and that the correlation across units follows a very specific pattern – equal covariance between any two units for any particular time. The Hausman (1978) test for fixed/random effects fails to reject the null hypothesis that the coefficients estimated by the efficient random effects estimator are the same as the ones estimated by the consistent fixed effects estimator, allowing us to apply the fixed effects procedure. However, since the Hausman test may be misleading due to the presence of autocorrelation, we include a set of individual country dummies in our regression model to control for individual unobservable fixed effects. Property taxes are included in the direct taxes. Results obtained using the alternative definition of tax ratio that includes property taxes in indirect taxes are available from the authors upon request.

18. 19.

20. 21.

22.

23. 24. 25.

26.

27.

Direct versus indirect taxation 77 28. 29. 30. 31. 32. 33.

34. 35.

36. 37. 38.

39.

40. 41.

42.

We re-estimate the model by using the alternative definition of tax ratio where property taxes are included in indirect taxes but obtain very similar results. We re-estimated the model by using political rights and civil liberties separately, rather than combined in the democracy index and find no significant change in the results. This is in line with some of the previous literature claiming that political factors are not important in determining the actual shape of tax mix (Volkerink and de Haan, 1999; Geys and Revelli, 2009). However, Messere (1993) finds no evidence that increasing economic integration would greatly affect the tax mix in OECD countries. See Kau and Rubin (1981) for an elaboration on the argument that urbanization positively affects taxes on goods and services because of the potential less monitoring costs on tax compliance in urban areas. Since the pioneering contributions by Solow (1956), Swan (1956), Barro (1990), Barro and Sala-i-Martin (1992, 1995), King and Rebelo (1990), and Lucas (1990), empirical research on economic growth has been extended in various fiscal dimensions, including public expenditure and taxation (Jones et al., 1993; Mendoza et al., 1997; Kim, 1998; Dahlby, 2003; and Lee and Gordon, 2005). In turn, Lee and Gordon’s (2005) estimating equation, except for the tax variables, is based on the specification used in Mankiw et al. (1992) and Barro (1992). The smaller the size of country i, the relatively shorter the distance between its largest city and largest cities in neighboring countries, implying relatively stronger effect of their tax ratios on the tax ratio in country i. The source for the distance measure between two countries is CEPII (Centre d’Etudes Prospectives et d’Informations Internationales, http://www.cepii.fr/). Geodesic distances are calculated following the great circle formula, which uses latitudes and longitudes of the most important cities/ agglomerations in terms of population. We used a Hausman test to check for the appropriateness of fixed effects estimation approach. Practically identical results are obtained when property taxes are included as indirect taxes; these results are available from the authors. For example, Acemoglu and Verdier (1998) argue that corruption facilitates economic growth because it helps government officials become more efficient in approving the project process. On the other hand, Mauro (1995) and Knack and Keefer (1995) claim that corruption increases uncertainty in decision-making and in the costs of conducting business, and, therefore, that it reduces economic growth. To investigate the presence of multicollinearity, we calculate the tolerance and variance inflation factor (VIF) for each explanatory variable and find that almost all variables have tolerance higher than 0.4 and a low VIF value, suggesting a low degree of multicollinearity, if any. We also perform a sample estimation of the correlations between the independent variables. Only three correlation coefficients satisfy the ‘conservative’ requirement of 0.5 or larger, involving the corporate tax rate. When the corporate tax rate is excluded from the regression, the R-squared for the subsample of developing countries drops substantially (from 0.87 to 0.66), but the coefficients for school enrollment, openness, and inflation now become statistically significant. In the case of developed countries, the exclusion of the corporate rate does not cause any significant changes. The literature on this issue is large, going back to Musgrave and Miller (1948), Brown (1955), Musgrave (1959), and Pearse (1962). Money is the sum of currency outside banks and demand deposits other than those of central government. This series, frequently referred to as M1 is a narrower definition of money than M2. Data are in current local currency. For more information, see Table s: WDI 4.15 (http://data.worldbank.org/indicator/FM.LBL.BMNY.CN). Since we are allowed to use random effects rather than fixed effects, we further consider whether there are any unobserved effects at all. If this were the case, we could use pooled OLS, which would offer two important advantages: it would provide a gain in

78

43. 44.

45. 46.

47. 48.

49.

50. 51.

52. 53. 54.

The Elgar guide to tax systems efficiency because we would not have to allow for within-group correlations, and we could use its finite sample properties rather than relying on asymptotic properties of random effects. However, the Breusch-Pagan Lagrange Multiplier test rejects the null hypothesis that OLS is consistent, so we stay with the random effects procedure. Even though the squared terms are frequently not individually statistically significant, the level and the squared term are often jointly significant at the 10 per cent level. However, not including the squared term may bias the estimates of the level term upwards. The correlation within each pair (dependent variable versus included variable, dependent variable versus omitted variable, and omitted variable versus included variable) is positive, implying that the estimate is upwardly biased. The evidence on redistributive effects of taxes is especially weak for developing countries (Bird and Zolt, 2005; Martinez-Vazquez, 2007; and Harberger, 2008). To control for the fact that income distributions across countries are based on different measurements of income, including gross income, net income, and consumption, we include in our empirical model a set of dummies for net income and consumption definitions, and use gross income as the base category. The over-identification test has a P-value of 0.9, suggesting that we fail to reject the null hypothesis that the instrumental variable for the tax ratio is exogenous. Sala-i-Martin (1997) finds that larger government size, measured by social transfers, reduces income inequality, while Landau (1985), Peden and Bradley (1989), Fölster and Herekson (2001) find that resources are allocated less efficiently within larger governments, with government size having no or a negative effect on income inequality. As education expands, income distribution may become more unequal, which is particularly important in countries with very low levels of education. However, as more people receive education, the return to education will generally decline, reducing income inequality (Schultz, 1960; Becker, 1964; Mincer, 1974; Knight and Sabot, 1983; and Gregorio and Lee, 2002). See, for example, Devereux and Griffith (1998, 2002), Buttner (2002), De Mooij and Ederveen (2003, 2005), Bénassy-Quéré et al. (2005, 2007), Razin and Sadka (2006). According to the UNCTAD definition, FDI flows consist of the net sales of shares and loans (including non-cash acquisitions made against equipment, manufacturing rights, etc.) to the parent company plus the parent firm´s share of the affiliate´s reinvested earnings plus total net intra-company loans (short and long term) provided by the parent company. For branches, FDI flows consist of the increase in reinvested earnings plus the net increase in funds received from the foreign direct investor. FDI flows with a negative sign (reverse flows) indicate that at least one of the components in the above definition is negative and not offset by positive amounts of the remaining components. The corresponding results with the alternative definition of the tax ratio are fairly similar and therefore not reported but available from the authors upon request. The democracy variable measures the existence of civil rights and liberties and is calculated as (14 – civil liberties – political rights)/12, where both ‘civil liberties’ and ‘political rights’ are scaled from 1 (most free) to 7 (least free). See Adam and Filippaios (2007) for a review of the literature on this issue.

REFERENCES Acemoglu, D. and T. Verdier (1998), ‘Property Rights, Corruption and the Allocation of Talent: A General Equilibrium Approach’, The Economic Journal, 108(450), 1381–403. Adam, A. and F. Filippaios (2007), ‘Foreign Direct Investment and Civil Liberties: A New Perspective, European Journal of Political Economy, 23(4), 1038–52. Aizenman, J. and Y. Jinjarak (2006), ‘Globalization and Developing Countries – A Shrinking Tax Base?’, NBER Working Paper No.11933.

Direct versus indirect taxation 79 Atkinson, A.B. (1977), ‘Optimal Taxation and the Direct versus Indirect Tax Controversy’, The Canadian Journal of Economics/Revue canadienne d’Economique, 10(4), 590–606. Atkinson, A.B. and N.H. Stern (1980), ‘On the Switch from Direct to Indirect Taxation’, Journal of Public Economics, 14(2), 195–224. Atkinson, A.B. and J.E. Stiglitz (1976), ‘The Design of Tax Structure: Direct versus Indirect Taxation’, Journal of Public Economics, 6(1–2), 55–75. Auerbach, A.J. (2006), ‘The Future of Capital Income Taxation’, Fiscal Studies, 27(4), 399–420. Auerbach, A. and Feenberg, D. (2000), ‘The Significance of Federal Taxes as Automatic Stabilizers’, Journal of Economic Perspectives, 14(3), 37–56. Balestrino, A., A. Cigno and A. Pettini (2003), ‘Doing Wonders with an Egg: Direct and Indirect Taxation when Households Differ in Market and Non-market Abilities’, Journal of Public Economic Theory, 5(3), 479–98. Barro, R.J. (1990), ‘Government Spending in a Simple Model of Endogenous Growth’, Journal of Political Economy, 98(5), 103–26. Barro, R.J. (1991), ‘Economic Growth in a Cross Section of Countries’, The Quarterly Journal of Economics, 106(2), 407–43. Barro, R.J. (1992), Human Capital and Economic Growth, Proceedings, Federal Reserve Bank of Kansas City, pp. 199–230. Barro, R.J. (2000), ‘Inequality and Growth in a Panel of Countries’, Journal of Economic Growth, 5(1), 5–32. Barro, R.J. and J.-W. Lee (2000), ‘International Data on Educational Attainment: Updates and Implications’, manuscript, Harvard University. Barro, R.J. and X. Sala-i-Martin (1992), ‘Convergence’, Journal of Political Economy, 100(2), 223–51. Barro, R.J. and X. Sala-i-Martin (1995), Economic Growth, New York: McGraw-Hill. Baunsgaard, T. and M. Keen (2005), ‘Tax Revenue and (or?) Trade Liberalization’, IMF Working Paper No. 05/112. Beck, N. and J.N. Katz (1995), ‘What To Do (and Not To Do) with Time-Series CrossSection Data’, The American Political Science Review, 89(3), 634–47. Beck, T., A. Demirgüç-Kunt and R. Levine (2000, revised: November 2008), ‘A New Database on Financial Development and Structure’, World Bank Economic Review, 14, 597–605. Beck, T., A. Demirgüç-Kunt and R. Levine (2004), ‘Finance, Inequality and Poverty: CrossCountry Evidence’, World Bank Policy Research Working Paper No. 3338. Beck, T., M.K. Lundberg and G. Majnoni (2001), ‘Financial Intermediary Development and Growth Volatility: Do Intermediaries Dampen or Magnify Shocks?’, World Bank Policy Research Working Paper No. 2707. Becker, G.S. (1964), Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education’, New York: NBER and Columbia University Press. Bénassy-Quéré, A., L. Fontagné and A. Lahrèche-Révil (2005), ‘How Does FDI React to Corporate Taxation?’, International Tax and Public Finance, 12(5), 583–603. Bénassy-Quéré, A., N. Gobalraja and A. Trannoy (2007), ‘Tax and Public Input Competition’, Economic Policy, 22(50), 385–430. Bird, R.M. and E.M. Zolt (2005), ‘Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries’, UCLA Law Review, 52(6), 1627–95. Boadway, R., M. Marchand and P. Pestieau (1994), ‘Towards a Theory of the DirectIndirect Tax Mix’, Journal of Public Economics, 55(1), 71–88. Borge, L.-E. and J. Rattso (2004), ‘Income Distribution and Tax Structure: Empirical Test of the Meltzer–Richard Hypothesis’, European Economic Review, 48(4), 805–26. Brown, C.E. (1955), ‘The Static Theory of Automatic Fiscal Stabilization’, Journal of Political Economy, 63(5), 427–40. Buttner, T. (2002), ‘The Impact of Taxes and Public Spending on the Location of FDI: Evidence from FDI-flows within Europe’, ZEW Discussion Paper No. 02-17. Calderon, C. and A. Chong (2001), ‘External Sector and Income Inequality in Interdependent Economies using a Dynamic Panel Data Approach’, Economic Letters, 71(2), 225–31.

80

The Elgar guide to tax systems

Clausing, K.A. (2007), ‘Corporate Tax Revenues in OECD Countries’, International Tax and Public Finance, 14(April), 115–33. Cremer, H. and F. Gahvari (1995), ‘Uncertainty, Optimal Taxation and the Direct Versus Indirect Tax Controversy’, The Economic Journal, 105(432), 1165–79. Cremer, H., P. Pestieau and J.C. Rochet (2001), ‘Direct versus Indirect Taxation: The Design of the Tax Structure Revisited’, International Economic Review, 42(3), 781–99. Dahlby, B. (2003), ‘Restructuring the Canadian Tax System by Changing the Mix of Direct and Indirect Taxes’, in H.G. Grubel, Tax Reform in Canada: Our Path to Greater Prosperity, Vancouver: The Fraser Institute, pp. 77–108. De Mooij, R. and S. Ederveen (2003), ‘Taxation and Foreign Direct Investment: A Synthesis of Empirical Research’, International Tax and Public Finance, 10(6), 673–93. De Mooij, R. and S. Ederveen (2005), ‘Explaining the Variation in Empirical Estimates of Tax Elasticities of Foreign Direct Investment’, Tinbergen Institute Discussion Paper No. 108/3. De Mooij, R. and G. Nicodème (2008), ‘Corporate Tax Policy and Incorporation in the EU’, CPB Discussion Paper No. 97. Devereux, M. and R. Griffith (1998), ‘Taxes and Location of Production: Evidence from a Panel of US Multinationals’, Journal of Public Economics, 68(3), 335–67. Devereux, M. and R. Griffith (2002), ‘The Impact of Corporate Taxation on the Location of Capital: A Review’, Swedish Economic Policy Review, 9(1), 11–33. Devereux, M. and P.B. Sørensen (2005), ‘The Corporate Income Tax: International Trends and Options for Fundamental Reform’, EPRU-Analysis No. 24. Devereux, M., R. Griffith and A. Klemm (2002), ‘Corporate Income Tax Reforms and Tax Competition’, Economic Policy, 17(35), 451–95. Diamond, P.A. and J.A. Mirrlees (1971), ‘Optimal Taxation and Public Production I: Production Efficiency and II: Tax Rules’, American Economic Review, 61(1), 8–27. Dollar, D. (1992), ‘Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976–1985’, Economic Development and Cultural Change, 40(3), 523–44. Dollar, D. and A. Kray (2002), ‘Growth is Good for the Poor’, Journal of Economic Growth, 7(3), 195–225. Dollar, D. and A. Kraay (2003), ‘Institutions, Trade and Growth’, Journal of Monetary Economics, 50(1), 133–62. Duncan, D. and K. Sabirianova Peter (2008), ‘Tax Progressivity and Income Inequality’, Andrew Young School of Policy Studies Research Paper Series No. 08-26. Easterly, W., R. Islam and J.E. Stiglitz (2000), Shaken and Stirred: Explaining Growth Volatility, Washington, DC: The World Bank. Edwards, S. (1998), ‘Openness, Productivity and Growth: What Do We Really Know?’, Economic Journal, 108(447), 383–98. Emran, M.S. and J.E. Stiglitz (2005), ‘On Selective Indirect Tax Reform in Developing Countries’, Journal of Public Economics, 89(4), 599–623. European Commission (2006), ‘Macroeconomic Effects of a Shift from Direct to Indirect Taxation: A Simulation for 15 EU Member States. Note Presented by the European Commission Services (DG TAXUD) at the 72nd meeting of the OECD Working Party No. 2 on Tax Policy Analysis and Tax Statistics, Paris, 14–16 November 2006’, available from: http://www.oecd.org/dataoecd/43/56/39494151.pdf; accessed 17 April 2011. Fölster, S. and M. Herekson (2001), ‘Growth Effects of Government Expenditure and Taxation in Rich Countries’, European Economic Review, 45(8), 1501–20. Frankel, J.A. and Romer, D. (1999), ‘Does Trade Cause Growth?’, American Economic Review, 89(3), 379–99. Gastil, R.D. (various years), Freedom in the World: Political Rights and Civil Liberties, New York: Freedom House. Geys, B. and F. Revelli (2009), ‘Decentralization, Competition and the Local Tax Mix: Evidence from Flanders’, Department of Economics, University of Torino, Working Paper No. 2/2009.

Direct versus indirect taxation 81 Gregorio, J.D. and J.W. Lee (2002), ‘Education and Income Inequality: New Evidence from Cross-Country Data’, Review of Income and Wealth, 48(3), 395–416. Griffiths, A.L. and K. Nerenberg (2005), Handbook of Federal Countries, 2005, Montreal & Kingston: McGill-Queens University Press. Harberger, A. (1964), ‘Taxation, Resource Allocation and Welfare’, in J. Due (ed.), The Role of Direct and Indirect Taxes in the Federal Revenue System, Princeton, New Jersey: Princeton University Press. Harberger, A.C. (2006), ‘Taxation and Income Distribution: Myths and Realities’, in J.  Alm, J. Martinez-Vazquez and M. Rider (eds), The Challenges of Tax Reform in a Global Economy, New York: Springer. Harberger, A.C. (2008), ‘Corporation Tax Incidence: Reflections on what is Known, Unknown and Unknowable’, in J.W. Diamond and G.R. Zodrow, Fundamental Tax Reform: Issues, Choices and Implications, Cambridge, MA: MIT Press. Hausman, J.A. (1978), ‘Specification Tests in Econometrics’, Econometrica, 46(6), 1251–71. Hicks, J.R. (1939), Value and Capital, Oxford: Oxford University Press. Hines, J.R. and L.H. Summers (2009), ‘How Globalization Affects Tax Design’, NBER Working Paper No. 14664. Johansson, A., C. Heady, J. Arnold, B. Brys and L. Vartia (2008), ‘Taxation and Economic Growth’, OECD Economics Department Working Papers No. 620. Jones, L., R. Manuelli and P. Rossi (1993), ‘Optimal Taxation in Models of Endogenous Growth’, Journal of Political Economy, 101(3), 485–519. Kau, J.B. and P.H. Rubin (1981), ‘The Size of Government’, Public Choice, 37(2), 261–74. Keen, M. (2008), ‘VAT, Tariffs, and Withholding: Border Taxes and Informality in Developing Countries’, 92(10–11), 1892–1906. Kenny, L.W. and S.L. Winer (2006), ‘Tax Systems in the World: An Empirical Investigation into the Importance of Tax Bases, Administration Costs, Scale and Political Regime’, International Tax and Public Finance, 13(2), 181–215. Kesselman, J.R. (1993), ‘Evasion Effects of Changing the Tax Mix’, The Economic Record, 69(205), 131–48. Kim, S.J. (1998), ‘Growth Effect of Taxes in an Endogenous Growth Model: To What Extent do Taxes Affect Economic Growth?’, Journal of Economic Dynamics and Control, 23(1), 125–58. King, R. and S. Rebelo (1990), ‘Public Policy and Economic Growth: Developing Neoclassical Implications’, Journal of Political Economy, 98(1), 126–51. Knack, S. and P. Keefer (1995), ‘Institution and Economic Performance: Cross Country Tests Using Alternative Institutional Measures’, Economics and Politics, 7(3), 207–27. Kneller, R.M., F. Bleaney and N. Gemmell (1999), ‘Fiscal Policy and Growth: Evidence from OECD Countries’, Journal of Public Economics, 74(2), 171–90. Knight, J. and R.H. Sabot (1983), ‘Educational Expansion and the Kuznets Effect’, American Economic Review, 73(5), 1132–6. Kornai, J. (1992), The Socialist System: The Political Economy of Socialism, Princeton, NJ: Princeton University Press. Kreickemeier, U. and Raimondos-Møller, P. (2008), ‘Tariff-Tax Reforms and Market Access’, Journal of Development Economics, 87(1), 85–91. Landau, D. (1985), ‘Government Expenditure and Economic Growth in the Developed Countries: 1952–76’, Public Choice, 47(3), 459–77. Landes, D. (1998), The Wealth and Poverty of Nations, New York: W.W. Norton. La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R.W. Vishny (1999), ‘The Quality of Government’, Journal of Law Economics and Organization, 15(1), 222. Lee, Y. and R.H. Gordon (2005), ‘Tax Structure and Economic Growth’, Journal of Public Economics, 89(5–6), 1027–43. Li, W. and P.-D. Sarte (2004), ‘Progressive Taxation and Long-run Growth’, American Economic Review, 94(5), 1705–16. Lucas, R. (1990), ‘Supply-side Economics: An Analytical Review’, Oxford Economic Papers, 42(2), 293–316.

82

The Elgar guide to tax systems

Madsen, J. and D. Damania (1996), ‘The Macroeconomic Effects of a Switch From Direct to Indirect Taxes: An Empirical Assessment’, Scottish Journal of Political Economy, 43(5), 566–78. Mankiew, N.G., D. Romer and D.N. Weil (1992), ‘A Contribution to the Empirics of Economic Growth’, The Quarterly Journal of Economics, 107(2), 407–37. Martinez-Vazquez, J. (2007), ‘Budget Policy and Income Distribution’, International Studies Program Working Paper Series No. 0707, International Studies Program Andrew Young School of Policy Studies, Georgia State University. Martinez-Vazquez, J. (2008), ‘The Impact of Budgets on the Poor: Tax and Expenditure Benefit Incidence Analysis’, in Blanca Moreno-Dodson and Quentin Wodon (eds) Public Finance for Poverty Reduction, World Bank. Mauro, P. (1995), ‘Corruption and Growth’, The Quarterly Journal of Economics, 110(3), 681–712. Meltzer, A.H. and S.F. Richard (1981), ‘A Rational Theory of the Size of Government’, The Journal of Political Economy, 89(5), 914–27. Mendoza, E.G., G.M. Milesi-Ferretti and P. Asea (1997), ‘On the Ineffectiveness of Tax Policy in Altering Long-Run Growth: Harberger’s Superneutrality Conjecture’, Journal of Public Economics, 66(1), 99–126. Messere, K.C. (1993), Tax Policy in OECD Countries: Choices and Conflicts, Amsterdam: IBFD Publications B.V. Mincer, J. (1974), Schooling, Experience and Earnings, New York: Columbia University Press. Mino, K. and Nakamoto, Y. (2008), ‘Progressive Taxation, Wealth Distribution and Macroeconomic Stability’, Discussion Papers in Economics and Business No. 08-22, Osaka University, Graduate School of Economics and Osaka School of International Public Policy. Mueller, D.C. and T. Stratmann (2003), ‘The Economic Effects of Democratic Participation’, Journal of Public Economics, 87(9–10), 2129–55. Munk, K.J. (2008), ‘Tax-Tariff Reform with Costs of Tax Administration’, International Tax and Public Finance, 15(6), 647–67. Musgrave, R. (1959), The Theory of Public Finance, New York: McGraw-Hill. Musgrave, R.A. and M.H. Miller (1948), ‘Built-in Flexibility’, American Economic Review, 38(1), 122–8. Naito, H. (1999), ‘Re-examination of Uniform Commodity Taxes under a Non-linear Income Tax System and its Implication for Production Efficiency’, Journal of Public Economics, 71(2), 165–88. Naito, H. (2004), Atkinson and Stiglitz Theorem with Endogenous Human Capital Accumulation, Institute of Social and Economic Research, Osaka University and Department of Economics University of California Irvine. Padovano, F. and E. Galli (2001), ‘Tax Rates and Economic Growth in the OECD Countries (1950–1990)’, Economic Inquiry, 39(1), 44–57. Padovano, F. and E. Galli (2002), ‘Comparing the Growth Effects of Marginal vs. Average Tax Rates and Progressivity’, European Journal of Political Economy, 18, 529–54. Pearse, P.H. (1962), ‘Automatic Stabilization and the British Taxes on Income’, Review of Economic Studies, 29(2), 124–39. Peden, E.A. and M.D. Bradley (1989), ‘Government Size, Productivity and Economic Growth: The Post-war Experience’, Public Choice, 61(3), 229–45. Piotrowska, J. and W. Vanborren (2008), ‘The Corporate Income Tax Rate–Revenue Paradox: Evidence in the EU’, European Commission Taxation Papers No. 12. Poterba, J.M., J.J. Rotemberg and L.H. Summers (1986), ‘A Tax-Based Test for Nominal Rigidities’, The American Economic Review, 76(4), 659–75. Ramsey, F. (1927), ‘A Contribution to the Theory of Taxation’, Economic Journal, 37(145), 47–61. Razin, A. and E. Sadka (2006), ‘Vying for Foreign Direct Investment: An EU-type Model of Tax Competition’, NBER Working Paper No. 11991.

Direct versus indirect taxation 83 Romero-Avila, D. and R. Strauch (2008), ‘Public Finance and Long-Term Growth in Europe: Evidence from a Panel Data Analysis’, European Journal of Political Economy, 24(1), 172–91. Saez, E. (2002), ‘The Desirability of Commodity Taxation under Non-linear Income Taxation and Heterogeneous Tastes’, Journal of Public Economics, 83(2), 217–30. Sala-I-Martin, X. (1997), ‘Transfers, Social Safety Nets and Economic Growth’, IMF Staff Paper No. 44(1),81–102. Schultz, T.W. (1960), ‘Capital Formation by Education’, The Journal of Political Economy, 68(6), 571–83. Simmons, R.S. (2006), ‘Does Recent Empirical Evidence Support the Existence of International Corporate Tax Competition?’, Journal of International Accounting, Auditing and Taxation, 15(1), 16–31. Solow, R.M. (1956), ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal of Economics, 74(1), 65–94. Sørensen, P.B. (2006), ‘Can Capital Income Taxes Survive? And Should They?’, CESifo Economic Studies, 53(2), 172–228. Swan, T.W. (1956), ‘Economics Growth and Capital Accumulation’, Economic Record, 32(3), 334–61. Volkerink, B and J. de Haan (1999), Political and Institutional Determinants of the Tax Mix: An Empirical Investigation for OECD Countries, Research Report No. 99E05, University of Groningen, Research Institute SOM (Systems, Organisations and Management). Watrin, C. and R. Ullmann (2008), ‘Comparing Direct and Indirect Taxation: The Influence of Framing on Tax Compliance’, The European Journal of Comparative Economics, 5(1), 33–56. Weller, C.E. (2007), ‘The Benefits of Progressive Taxation in Economic Development’, Review of Radical Political Economics, 39(3), 368–76. Widmalm, F. (2001), ‘Tax Structure and Growth: Are Some Taxes Better than Others?’, Public Choice, 107(3–4), 199–219.

84

The Elgar guide to tax systems

APPENDIX Table 2A.1

Variables description and sources

Variable

Description

Source

Agriculture

Share of agriculture in GDP

Average effective tax rate

Average effective tax rate = foreign income taxes/ (foreign income tax + net incomes) of all affiliates for US firms operating abroad in each country Bureaucratic quality index, ranging from 0–6, with 6 denoting the highest quality (14−political rights−civil liberties)/12 Scale from 1 (most free) to 7 (least free) Scale from 1 (most free) to 7 (least free) Top marginal statutory Corporate income tax rate in the initial year of the corresponding period Corruption index, ranging from 0–6, with 6 denoting least corruptive Per capital crude petroleum production (in thousands of metric tons) Age dependency ratio (dependents to working-age population) Average years of schooling in the adult population 25+ years old State and local expenditure to total (central, state, local) expenditure Foreign direct investment flows from US firms divided by GDP

World Development Indicators (WDI) Bureau of Economic Analysis (BEA)

Bureaucracy index Democracy index Political rights Civil liberties Corporate tax rate

Corruption index

Crude petrol

Dependency ratio Education

Expenditure decentralization FDI from US to GDP (net)

International Country Risk Guide (ICRG) 2009 Freedom House: Authors’ calculations Freedom House Freedom House Office of Tax Policy Research (OTPR)

International Country Risk Guide (ICRG) 2008 UN Energy Statistics Database World Development Indicators (WDI) Barro and Lee (2000)

IMF GFS: Authors’ calculations Bureau of Economic Analysis (BEA)

Direct versus indirect taxation 85 Table 2A.1

(continued)

Variable

Description

Source

Total FDI to GDP (net) Federal

Total foreign direct investment flows divided by GDP = 1 if country has formal federal structure GDP per capita in current local prices GDP per capita in 2000 US$ Real per capita GDP growth rate Gini coefficient

UNCTAD

GDP per capita (current prices) GDP per capita (real) GDP per capita growth rate Gini

Globalization Inflation, consumer prices Labor cost

Labor force participation Labor force participation female Latitude

Legal origin

M1 growth rate

Population

KOF Index of Globalization Inflation, consumer prices (annual %) Wages of employees working in US companies’ foreign affiliates (000 $US/year) Labor force participation rate, total (share of total population ages 15–64) Labor force participation rate, female (share of female population ages 15–64) The absolute value of the latitude of the country, scaled to take values between 0 and 1 The legal origin of the Company Law or Commercial Code of each country: English, French, or German Commercial Code The annual growth of the sum of currency outside banks and demand deposits other than those of central government. Population size

Handbook of Federal Countries, 2005 World Development Indicators (WDI) World Development Indicators (WDI) World Development Indicators (WDI) UNU-WIDER World Income Inequality Database, May 2008 ETH Zürich KOF Konjunkturforschungsstelle World Development Indicators (WDI) Bureau of Economic Analysis (BEA) World Development Indicators (WDI) World Development Indicators (WDI) La Porta et al. (1999)

La Porta et al. (1999)

World Development Indicators (WDI)

World Development Indicators (WDI)

86

The Elgar guide to tax systems

Table 2A.1

(continued)

Variable

Description

Source

Population growth rate Private credit

Population growth rate

World Development Indicators (WDI) Beck et al. (2000/2008)

School enrollment primary School enrollment secondary Socialist

Tax ratio 1

Tax ratio 2

Telephone landlines Total tax to GDP Trade openness Urbanization

Private credit by deposit money banks and other financial institutions to GDP Primary enrollment rate (%) (gross) Secondary enrollment rate (%) (gross) Countries having either a socialist economic system or a mixed socialist economic system and a socialist or communist political Direct (income tax, payroll tax, Social Security contributions, property tax) to Indirect (taxes on goods and services, taxes on int’l trade, other taxes) tax ratio Direct (income tax, payroll tax, Social Security contributions) to indirect (taxes on goods and services, taxes on int’l trade, other taxes, property tax) tax ratio Telephone landlines (per 1000 people) Share of total (tax and non-tax) revenue in GDP in current prices (Imports + exports)/GDP Urban population (share of total)

UNESCo Institute of Statistics UNESCo Institute of Statistics Gastil (various years); Kornai (1992)

IMF GFS: Authors’ calculations

IMF GFS: Authors’ calculations

World Development Indicators (WDI) IMF GFS, WDI: Authors’ calculations World Development Indicators (WDI) World Development Indicators (WDI)

Direct versus indirect taxation 87 Table 2A.2

Variables descriptive statistics

Variable

Obs

Mean

St. Dev.

Min

Agriculture to GDP Average effective tax rate Bureaucracy index Corporate tax rate (%) (subsample initial year) Corporate tax rate IV (subsample initial year) Corruption index Crude petrol per capita (000 of metric tons) Democracy index Political rights Civil liberties Dependency ratio EU15 Expenditure decentralization Federal GDP (real) per capita ($) GDP (real) per capita growth rate (%) Gini coefficient (%) Net income Gini concept Gross income Gini concept Consumption Gini concept Globalization Inflation, consumer prices (%) Labor cost (000 $US) Labor force participation Latitude Legal origin English Legal origin French Legal origin German M1 growth rate Net FDI from the US to GDP Net FDI to GDP Openness Population Population Growth Rate Private credit to GDP School enrollment primary School enrollment secondary

3205 1152 1114 453

0.17 0.33 2.84 35.14

0.15 0.20 1.08 11.79

0.00 −0.28 0.00 0.00

0.94 0.98 4.00 60.00

454

35.36

3.87

25.62

43.54

1912 1798

3.37 2.61

1.45 10.38

0.00 0.00

6.00 196.24

3394 3396 3394 3831 3944 1487 3944 3501 3295

0.60 3.35 3.14 0.68 0.13 25.22 0.16 6995.34 1.86

0.33 2.16 1.86 0.19 0.34 17.31 0.36 8770.67 3.70

0.00 1.00 1.00 0.31 0.00 0.56 0.00 56.45 −9.54

1.00 7.00 7.00 1.17 1.00 87.00 1.00 51 673.98 9.26

1302 1302 1302 1302 3429 3192 1154 2938 3944 3944 3944 3944 2964 1114

35.85 0.52 0.31 0.17 0.52 10.65 25.80 0.69 0.32 0.29 0.46 0.04 0.29 0.11

10.22 0.50 0.46 0.38 0.18 13.45 17.80 0.09 0.20 0.46 0.50 0.20 1.75 0.29

16.60 0.00 0.00 0.00 0.09 −21.68 2.59 0.46 0.01 0.00 0.00 0.00 −0.42 −0.01

73.90 1.00 1.00 1.00 0.93 99.88 89.96 0.93 0.72 1.00 1.00 1.00 67.25 2.92

1166 2974 3933 3830 3040 1107 639

Max

0.04 0.22 −0.16 4.97 0.77 0.54 0.07 4.32 26 900 000 84 500 000 40 130 1 100 000 000 0.01 0.01 −0.04 0.04 0.45 0.38 0.01 3.45 85.55 18.12 9.48 104.57 84.92 24.86 19.00 161.66

88

The Elgar guide to tax systems

Table 2A.2

(continued)

Variable

Obs

Mean

St. Dev.

Min

Max

Socialist Tax ratio 1 Tax ratio 1 (Poterba et al., 1986) Tax ratio 1 IV Tax ratio 2 Tax ratio 2 (Poterba et al., 1986) Tax Ratio 2 IV Telephone landlines (per 000 people) Total revenue to GDP Urbanization

3944 3944 1773

0.12 3.32 0.02

0.32 1.97 0.10

0.00 0.02 −0.25

1.00 4.87 0.44

3944 3944 1773

1.52 3.21 −0.03

0.45 2.03 0.06

0.22 0.00 −0.26

5.17 4.50 0.39

3944 1166

1.28 2.81

0.37 2.15

0.20 0.02

3.10 7.97

1715 3944

0.28 0.54

0.13 0.24

0.03 0.03

0.64 1.00

Direct versus indirect taxation 89

BOX 2A.1

NUMBER OF COUNTRIES PER YEAR FOR WHICH THE GOVERNMENT FINANCE DATA IS AVAILABLE IN THE SAMPLE: ISSUES WITH INTERNATIONAL MONETARY FUNDS’ GOVERNMENT FINANCE STATISTICS DATA

Data on taxes are downloaded from the International Monetary Fund’s (IMF) Government Finance Statistics (GFS) Database, which provides data with consistent definitions across countries and years. However, a change in methodology from GFS1986 to GFS2001 in 2001, with the data from 1990 onward being reclassified from the old to the new methodology, has made historical data (1972–89) not comparable with new data (1990 onwards). Another issue is that coverage for particular regions and individual years may be limited. Given that data collection is through a questionnaire filled out each year by local ministries in member countries, data availability in the GFS Database primarily depends on filer responsiveness. Table 2A.3 documents the data availability for each year in the sample. The classifications of revenue are substantially different in the two manuals. Revenue in the 1986 GFS Manual is classified as tax, non-tax, or capital revenue. Grants form a separate, non-revenue category of receipts. In the revised GFS Manual, revenue is subdivided into taxes, social insurance contributions, grants, and other revenue. In more detail: taxes exclude Social Security contributions in the revised GFS Manual, but include them in the 1986 GFS Manual; social insurance contributions in the revised GFS Manual include Social Security contributions, which are classified as taxes in the 1986 GFS Manual, and contributions to social insurance schemes operated for the benefit of government employees, which are classified as non-tax revenue in the 1986 GFS Manual. Source:

2001 GFS Manual: p. 158.

90

21 25 26 26 26 26 26 25 27 27 27 27 27 27 27 27 27

Developed Countries 31 42 43 48 52 50 51 52 54 51 49 51 53 56 56 50 53

Developing Countries 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Countries in Transition 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Year 77 4 4 4 4 5 16 17 18 26 32 41 47 59 65 62 50

Full Sample 26 0 0 0 0 0 9 9 10 17 20 26 28 29 28 27 21

Developed Countries

Government Finance Statistics Manual 2001 (GFSM 2001), IMF Statistics Department, 2001.

52 67 69 74 78 76 77 77 81 78 76 78 80 83 83 77 80

1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988

Source:

Full Sample

Number of countries in the sample

Year

Table 2A.3

51 4 4 4 4 4 6 7 7 8 11 12 15 24 28 26 21

Developing Countries

0 0 0 0 0 0 0 0 0 0 0 3 4 6 9 9 8

Countries in Transition

91

Supply factors Tax base effect Crude petrol

Log(GDP per capita)

Socialist

Political preferences Democracy

Decentralization

Federal

Log(population)

−0.0014 (0.0009)

0.0200 (0.0162) −0.0057 (0.0214) 0.0338 (0.1281)

−0.0014** (0.0006)

−0.0302 (0.0410) 0.0000 (0.0000) −0.3018 (0.3292)

0.3802*** (0.0365) −0.9733 (1.3251) 0.1462 (0.1016) −0.0009* (0.0005)

Developed

Full

0.3410*** (0.0427) −0.0379 (0.7455) 0.3055 (0.8275) 0.0002 (0.0003)

(2)

(1)

0.0022 (0.0050)

0.0558*** (0.0171) 0.0100 (0.0280) 0.0758 (0.1431)

0.1240** (0.0570) 3.5780*** (1.3630) −0.9824*** (0.3706) 0.0010** (0.0005)

Developing

(3)

−0.0048 (0.0113) −0.0034 (0.0203) −0.0666 (0.1098)

0.2803*** (0.0368) −0.1186 (0.6610) −0.0423 (0.1772) −0.0000 (0.0002)

Full

(4)

−0.0239 (0.0293) 0.0000 (0.0000) −0.0824 (0.1681)

0.4120*** (0.0422) −0.1269 (0.8749) 0.2423*** (0.0889) −0.0007** (0.0003)

Developed

(5)

(6)

0.0082 (0.0085) 0.0148 (0.0253) 0.0706 (0.1057)

0.0162 (0.0411) 1.6323* (0.8352) 0.3030*** (0.0778) 0.0003 (0.0003)

Developing

Determinants of tax mix: 1972–2005, fixed effects, annual data (dependent variable: tax ratioa)

Demand factors Scale effect Revenue to GDP

Table 2A.4

92

−0.515*** (0.1262) 0.0000 (0.0000) 437 41 0.91

−0.3224 (0.2161) 1.4358 (1.5138) 227 17 0.94

−0.4416*** (0.1174) −2.7267** (1.1227) 210 24 0.87

−0.0002 (0.0008) 0.0397*** (0.0101) −0.0319 (0.1102) −0.1850*** (0.0575)

Developing

(3)

−0.434*** (0.1057) 0.4681 (0.8042) 635 63 0.91

0.0014** (0.0006) 0.0187** (0.0081) 0.0272 (0.0670) −0.0730** (0.0349)

Full

(4)

−0.5364*** (0.1284) 0.3053 (0.9675) 328 24 0.93

0.0011 (0.0009) −0.0261*** (0.0094) −0.1493 (0.1382) 0.0941 (0.0574)

Developed

(5)

−0.3398*** (0.0950) −1.9395* (0.9989) 307 39 0.90

0.0012* (0.0006) 0.0275*** (0.0086) −0.0499 (0.0561) −0.1726*** (0.0393)

Developing

(6)

Source:

Authors.

Notes: Panel corrected standard errors in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. a. Dependent variable calculated as (t – q)/(1 + q) where t is the direct tax rate and q the indirect tax rate, and these tax rates computed, respectively, as total direct and direct taxes divided by nominal GDP, as in Poterba et al. (1986).

Observations Number of id R-squared

Constant

Developed

Full

0.0001 (0.0011) −0.0407** (0.0207) −0.7812*** (0.2980) 0.0384 (0.1019)

(2)

(1)

0.0010 (0.0007) 0.0380*** (0.0099) 0.1667 (0.1066) −0.145*** (0.0487)

(continued)

Administration costs Urbanization

Globalization

Agriculture

Openness

Supply factors Tax base effect LFP

Table 2A.4

3

Individual income taxation: income, consumption, or dual?* Robin Boadway

1

INTRODUCTION

The individual income tax constitutes, along with the value-added tax (VAT), the bulwark of the revenue-raising system for industrialized countries. While the VAT serves as an efficient revenue-raiser, the income tax is the main tax instrument used to achieve broader objectives of the tax-transfer system, such as distributive equity, equality of opportunity, social insurance, and social policy. Not surprisingly, given such diverse and value-driven objectives, the system can be both complex and controversial. This can be seen in the many aspects of income tax design that policymakers and tax specialists must assess and address. Some key ones include the following: 1.

2.

3.

Choice of the base. At the most basic level, the individual tax base could include income broadly defined, it could be restricted to consumption, or it could include elements of both. The choice may be driven partly by matters of economic principle, such as efficiency, equity, and tax base mobility, or it may be driven by administrative necessity, such as measurability, collection and compliance costs, and possibilities for evasion. Social and other objectives. The individual tax may be used to contribute to objectives such as equal opportunity for children and for those capable of pursuing higher education, job market participation, and economic security. Behavioral incentives. Elements of the individual tax may be intended to influence the behavior of taxpayers in ways thought to be desirable, perhaps because of perceived market failures. For example, there may be tax incentives for retirement saving, home ownership, and/or post-secondary education. Or, there may be incentives for entrepreneurship and innovative activity. Other examples include incentives for contributions to charity or to political parties, or incentives for environmental conservation. 93

94

The Elgar guide to tax systems

4.

Relation with the corporation income tax. One role of the corporation tax is to act as a backstop or withholding device for the individual tax. Individuals owning shares of corporations can shelter their income by retaining and reinvesting it rather than receiving dividends. Since the former give rise to capital gains that are taxed on a realization basis, taxation of income earned through corporations can effectively be postponed with discretion while it accumulates. The corporation income tax undoes this postponement by taxing the income at source. To that extent, the corporation tax can be viewed as a withholding tax, and such withholding should be recognized by providing a credit against individual taxes when these taxes are eventually paid on corporate source income. Tax systems typically do so by mechanisms for integrating individual and corporation taxes, such as dividend tax credits. The choice of integration devices is a difficult one, especially in a world of international capital mobility. Rate structure. Regardless of the individual tax base, the rate structure can in principle be chosen independently. A key issue is the extent of progressivity. At one extreme, the tax rate might be constant, with progressivity induced by the accompanying system of credits or exemption (the so-called flat tax or linear progressive tax). Alternatively, the marginal tax rate may vary with the tax base, typically, though not necessarily, in an increasing fashion. Moreover, the marginal tax rate may be negative at the bottom end, as in some OECD countries. Next, the tax structure might be schedular in the sense that a different rate structure applies to different components of the base. A relatively recent schedular tax innovation is the dual income tax system introduced initially in some Nordic countries. This applies a different rate structure to capital income than to labor income and transfers. Finally, the tax schedule may differ according to the family status of individuals. Thus, there may be separate tax schedules for families than for individuals, or individuals’ tax liabilities may be simply conditioned on their family circumstances. Relation with wealth and wealth transfer taxes. Individual income tax policy might be influenced by whether or not wealth or wealth transfers are taxed. For example, the manner in which accrued capital gains are treated on death might depend on whether inheritances or bequests are taxed. Also, the case for preferential treatment of capital income depends on the taxation of wealth or wealth transfers. Thus, dual income tax systems that tax capital income at low proportional rates as in the Nordic countries might be more attractive if wealth transfers are taxed. Relationship between national and sub-national tax systems. In federal

5.

6.

7.

Individual income taxation 95

8.

nations, as well as in some unitary ones, individual taxes might be used at more than one level of government. In such cases, the harmonization of tax bases that are co-occupied serves as a means of reducing collection and compliance costs as well as coordinating the share of taxes going to each level of government and addressing the potential equity and efficiency effects associated with decentralizing revenue-raising. Transfers delivered through the tax system. An important innovation in individual income taxation has been the use of refundable tax credits, that is, tax credits that are refundable to taxpayers with little or no tax liabilities against which to set them. This has added an element of progressivity to the tax system by bringing those with limited or no income into the personal tax-transfer system net, thereby complementing standalone transfer systems such as welfare, unemployment insurance, and state pensions. Refundable tax credits are also a vehicle for addressing some of the objectives mentioned earlier, such as support for children and for low-income workers for whom participation in the labor market is marginal. Finally, refundable tax credits have been used as devices to undo adverse equity effects on low-income persons from broad-based VATs that apply to their consumption purchases and would otherwise be regressive.

In the course of our study of individual tax systems, we necessarily touch on all of the above issues. Before doing so, it is worth emphasizing that the personal tax is one of three main broad-based taxes whose bases overlap to a considerable extent, the others being the VAT and payroll taxes. The bases of the latter two taxes are roughly similar in present value terms. They differ only to the extent that an individual’s net inheritances (the present value of inheritances less bequests) and other net transfers are positive.1 Both are essentially taxes that distort the labor–leisure choice (including the participation decision), at least to the extent that payroll taxes are not used to finance actuarially fair transfer programs. Thus, if payroll taxes finance the equivalent of fully contributory pension funds, they are unlikely to impose a distortion on the labor–leisure choice. In practice, payroll taxes are usually not earmarked to individual accounts so this is not an issue. Since for most taxpayers the bulk of taxable income consists of labor income, there is considerable overlap among the three bases. The main difference is that individual taxes might include elements of capital income in the base (for which the overlap might be with wealth or property taxes). That being the case, the overall tax rate faced by individuals includes all three tax rates.

96

The Elgar guide to tax systems

Nonetheless, there are some issues at stake in choosing the tax mix among these three taxes. For one thing, redistributive equity is influenced by the share of the individual tax in the tax mix since this tax offers the most scope for progressivity. While the VAT may achieve some progressivity through exemptions of some goods, payroll taxes are often somewhat regressive since they have an upper limit on tax payments. Second, given the tax mix, the tax system is effectively a schedular one to the extent that capital income is part of the individual income tax base. Capital income is thus implicitly taxed at a lower rate than labor income. Next, although payroll taxes and VATs distort the labor–leisure margin, they can also have different effects on saving behavior. Life-cycle saving principles would suggest that saving is lower under payroll taxes than under the VAT since tax liabilities tend to come earlier in the life cycle for the former. Moreover, to the extent that disposable income affects saving (for example, because of liquidity constraints), payroll taxes will involve less saving. Finally, a mix of taxes can mitigate tax evasion by reducing the tax rate on any one base and spreading the tax around. Thus, those who might find it easy to evade the VAT might not find it easy to evade the payroll tax, and vice versa. By lowering the tax rate on all bases, the incentive to evade from any one base might be lower. In what follows, we begin with an outline of the issues involved in the choice of a tax base for direct individual taxation. Then, in Section 3, we consider the rate structure and extent of progressivity. Section 4 discusses a number of important special issues in the design of the individual income tax, and Section 5 considers what all this implies for the choice of an individual tax system. Our focus is almost entirely on ideas rather than on facts.

2

CHOICE OF THE INDIVIDUAL TAX BASE

The choice of a base for direct taxes on individuals is the most fundamental one for tax design and one that has received considerable attention over the past several decades. Traditionally, the choice has been between two ideal contenders: comprehensive income and consumption. Comprehensive income, associated with Schanz (1896), Haig (1921), and Simons (1938), is the sum of consumption plus net accruals of wealth in a given tax period. Its appeal is as a measure of the ability to pay taxes by individuals. Consumption, originally advocated by Kaldor (1955) and subsequently by the US Treasury Blueprints (1977) and the Meade Report (1978),2 is meant to be a better index of individual welfare than comprehensive income. In fact, neither base is perfect from the latter perspective, nor can either be easily implemented, especially comprehensive income.

Individual income taxation 97 An ideal consumption base would include aggregate consumption services over the taxpaying period. While it is not feasible to measure consumption services directly, reasonably good indirect methods are available that can capture the equivalent of a substantial element of consumption services. Thus, consumption expenditures can in principle be obtained by deducting savings from income in any tax period. If a taxpayer saves an amount St out of earnings in a given year, that amount is deductible. If the asset purchased is held and the capital income reinvested in any given year, the capital income constitutes both a source of income and of savings, so can be ignored for tax purposes. When the asset is sold and its principle and accumulated interest used for consumption, these are added to the tax base. This method of deducting savings from the tax base and adding both the principle and interest when asset wealth is drawn down is referred to as taxing assets on a registered basis. Not all forms of consumption can be captured by treating savings on a registered basis. For one thing, some assets yield returns in imputed forms that are difficult to measure. Housing and other consumer durables are cases in point. For these assets, so-called tax-prepaid treatment accomplishes an equivalent effect in present value terms. Tax prepayment would give no deduction for saving or acquisition of the asset, and would not include any consumption from the imputed return or sale of the asset in the future. In other words, the acquisition of tax-prepaid assets would have no tax consequences. In the case of consumer durables, this accounting for consumption yields a measure of consumption expenditures that in present value terms is equivalent to consumption services. The combination of registered and tax-prepaid treatment allows most forms of consumption to be included in the tax base. Moreover, the taxpayer could even be allowed some discretion over the division of savings between the two categories, subject to the fact that some assets naturally fit into one category rather than the other (e.g., consumer durables as tax-prepaid assets, private businesses and human capital accumulation as registered assets). In effect, the taxpayer can choose the time profile of tax liabilities over the life cycle. At the same time, some forms of consumption are difficult to bring into the tax base, especially those that do not go through the market. Leisure and household production are examples. Another might be consumption financed out of inheritances, unless the latter are included as elements of income. These same problems apply to comprehensive income since it also includes consumption as an element of the tax base. However, comprehensive income encounters the additional substantial difficulty of including capital income in the tax base. To see this, it is useful to redefine

98

The Elgar guide to tax systems

comprehensive income as follows, using the individual’s per period budget constraint: Ct 1 DAt 5 Et 1 rAt where Ct is consumption purchased on the market, At is asset wealth, Et is earnings and r is the interest rate. (Any net transfers received should also be included in Et for completeness.) The left-hand side is comprehensive income, which would be difficult to measure, so the income tax base is defined by the right-hand side, which involves measuring capital income. There are many difficulties with including all sources of capital income in the tax base, and that alone makes comprehensive income an impossible ideal, even if it might be desired on economic grounds. Some forms of capital income are very difficult to measure accurately. In addition to the imputed returns on housing and durables already mentioned, examples include the imputed return on human capital investment and on individual businesses, both of which can be readily treated on a cash flow basis under a consumption base. There is also a problem with including asset returns that take an accrued form, such as capital gains. Inflation also compromises capital income taxation, some forms of which should ideally be indexed to avoid introducing elements of the principle into the tax base. Finally, relative to earnings, capital income is much more amenable to evasion, especially if it can be held abroad. The implication of this discussion is that comprehensive income is an unattainable base for taxation, much more so than consumption or labor income. The tax design problem then becomes one of comparing a compromised form of income taxation with reasonably complete consumption taxation. With this in mind, what does normative tax analysis suggest about the case for the choice between consumption and income taxation, or equivalently, the case for taxing capital income? From a theoretical perspective, taxing capital income is equivalent to taxing future consumption at a higher rate than current consumption. Whether this is justified depends on the relationship between present consumption, future consumption, and leisure, which is untaxed. The optimal tax literature is agnostic on this question, though it has identified some arguments that would influence the case for capital income taxation, albeit at different rates than labor income taxation. The simplest arguments go back to the optimal commodity tax literature, particularly the famous Corlett and Hague Theorem. Corlett and Hague (1953), and later Harberger (1964) in an explicitly optimizing framework, showed that in a world with a representative taxpayer consuming two taxable goods and leisure, a higher tax rate should be imposed on the good that is most

Individual income taxation 99 complementary with leisure. If the two goods are interpreted as present and future consumption (as in King, 1980), the case for capital income taxation would require that future consumption be more complementary with variable leisure than present consumption, which is not particularly plausible given that variable leisure in this case is present period leisure (future leisure being fixed). Matters become more complicated when one extends the model to several periods, each of which has consumption and variable leisure (Erosa and Gervais, 2001). In this setting, zero capital income taxation is optimal if the (1) utility discount factor equals the interest rate, (2) the wage rate is the same in both periods, and (3) preferences are additively separable and identical over time. Moreover, in this case, labor income taxes are constant over time. Alvarez et al. (1992) show in a related context with the wage rate constant over the two-period life cycle, optimal labor income taxes should decline with age if the interest rate exceeds the utility discount factor, and vice versa. On the other hand, if labor income taxes cannot be made age-specific, there should be a positive interest income tax if the interest rate exceeds the utility discount factor, and vice versa. A more plausible setting is when the intertemporal model involves not a representative household, but heterogeneous households so taxation serves redistributive objectives. Here the benchmark case is the AtkinsonStiglitz, Theorem (Atkinson and Stiglitz 1976), which shows that in a static optimal nonlinear tax setting à la Mirrlees (1971) with multiple commodities, no differential commodity taxation is called for if preferences are weakly separable with leisure. Adapting this to an intertemporal setting is not straightforward, but some insights have been obtained by Golosov et al. (2007), Diamond (2007), and Banks and Diamond (2008). If one takes a discrete-type version of the Mirrlees model and extends it to a multiperiod life-cycle setting with heterogeneous households, the AtkinsonStiglitz Theorem remains intact if (1) household per period utilities are additive in consumption and labor, (2) a household’s ability or wage rate is the same in each period, (3) there is no uncertainty, (4) the government can impose nonlinear taxation within each period, and (5) the government can fully commit to a tax policy announced at the beginning of life cycles. In this case, no capital income taxation is called for. However, relaxing those assumptions in particular ways generates a case for some capital income taxation. Some cases are as follows. If high-wage households have lower utility discount rates (which Saez, 2002a argues is plausible), capital income should be taxed in an optimal tax system (Diamond, 2007). (This argument might, however, be viewed as penalizing persons according to their preferences, since persons with a lower utility discount rate are treated as

100

The Elgar guide to tax systems

having higher utility from a social welfare perspective.) If future abilities are uncertain and households cannot insure against such uncertainty, some capital income taxation is useful. In this case, capital income taxation is a form of social insurance. If abilities change over the life cycle, and rise more steeply for high-ability persons, however, capital income taxation is not required if the earnings tax can be made age-dependent. Despite arguments for age-dependent taxation (e.g., Banks and Diamond, 2008), personal taxation is in practice blind to age. In these circumstances, taxation of capital income may be a fallback policy. While models such as these can be constructed that make the case for capital income taxation as an element of an optimal nonlinear tax system, it would be a stretch to use such results as a basis for actual policy recommendations. They are after all based on models in which governments are choosing nonlinear income taxes optimally and with full commitment, and in which particular specifications of household preferences apply.3 At the same time, there might be other arguments based on more practical considerations that might support some capital taxation. As mentioned above, some capital income might be from inheritances and inter vivos transfers that have not been included in the tax base. Imposing a tax on capital income will be an indirect way of taxing consumption financed by inheritances even if there is no a priori reason for taxing capital income. Of course, taxing inheritances directly circumvents this problem, but most inheritance tax systems apply only to large inheritances. In addition, some persons, especially the self-employed, have some discretion to convert labor income into capital income. Taxing the latter offsets this advantage. There may be constraints on the progressivity of the individual tax that precludes the government’s redistributive objectives from being achieved. These may be a consequence of taxpayer mobility or incentives to evade. To the extent that capital income accrues disproportionately in the hands of high-income persons, this might be an indirect way of achieving redistributive objectives that might otherwise be ideally addressed by progressive consumption taxation. Finally, there may be more positive-based arguments for taxing capital income. To the extent that governments cannot commit to future taxes, even benevolent governments cannot avoid taxing capital income. That being the case, the manner in which capital income enters the individual tax base may be important. Along the same lines, political economy considerations may make it difficult for governments to avoid taxing capital income. The upshot of this discussion is that a consumption tax has a number of attractive features relative to an income tax. It is much easier to implement from a tax administration point of view, and it has normative properties that are attractive. Nonetheless, there are some arguments for

Individual income taxation 101 taxing capital income at least to some extent (and recognizing that not all forms of capital income can easily be taxed). If future wages are uncertain, taxing capital income can be optimal, at least in the absence of full insurance (Golosov et al., 2007). And, of course, if the government is unable to commit to future tax rates, it will have an incentive to tax capital income at any point in time since there will always be some capital assets that have been accumulated and represent an inelastic source of tax revenues. Many of these arguments arise not because of the desirability of taxing future consumption differentially per se, but because of shortcomings in taxing consumption fully. We return to these issues below.

3

THE RATE STRUCTURE AND THE EXTENT OF PROGRESSIVITY

The structure of tax rates applying to individuals in the direct tax system comprises several elements. First, there is the rate structure per se, which is typically a piecewise linear system consisting of a number of brackets with marginal tax rates escalating from bracket to bracket. The rate structure may apply to individuals, or it may apply to families. Next, there are various tax credits, deductions, and exemptions, whose purpose is to provide tax relief on the basis of personal circumstances. Some of these are refundable, and others not, with refundable credits sometimes being income-tested so as to increase the degree of progressivity. The most prominent sorts of tax credits are those based on family circumstances, especially the number of children, those targeted to the low-income employed, and those for the elderly. Third, the extent of progressivity of the individual tax is also affected by the sheltering of various forms of income, especially capital income. Sheltering is typically afforded to limited amounts of saving for retirement, to home ownership, and to investment in human capital. Income received as inheritances, gifts, and inter vivos transfers may also be treated favorably in some tax systems. The same is the case for some forms of capital income, such as capital gains or dividends. Finally, different rate structures may apply to different sources of income. In particular, capital income may be subject to a separate tax structure than labor income and transfers, as is the case in dual income tax systems. This complicated mix of measures that underlie individual tax rate structures reflects the fact that arguments for progressivity vary considerably according to individual circumstances and sources of income. The standard theory of optimal redistribution through the income tax-transfer system, based on the original work of Mirrlees (1971), Atkinson (1973),

102

The Elgar guide to tax systems

and others, suggests a number of qualitative features of the optimal income tax, at least as the tax applies to earnings. First, it will typically be efficient to have some persons at the lower end of the skill distribution not working but receiving transfers. For those who are working, marginal tax rates would rise at a fairly steep rate at the lower end of the skill distribution and then approach a plateau near the top, not unlike tax structures actually observed. Even with very egalitarian social welfare functions, such as maximin, the degree of progressivity is apparently reasonably limited. This reinforces Mirrlees’ original finding that the rate structure was not appreciably different from linear above the lower part of the income distribution. However, these results are very model-specific, relying on settings in which the only source of inequality is endowed ability; persons are able to vary their labor supplies at will (along the so-called intensive margin); there is no saving or uncertainty; households have identical preferences; and governments are benevolent and can observe incomes perfectly. In the real world, there are a number of complicating factors that should influence the desired degree of progressivity. Some of them are as follows. The standard model emphasizes the responsiveness of labor supply to after-tax wages as workers adjust their hours of work at the margin. This is a misleading caricature of the endogeneity of labor supply for a number of important reasons. First, for most workers, varying hours of work is not the relevant option since hours of work are largely fixed. The main options they have are discrete choices of participation in the labor force (the extensive margin) or occupational choice. Recent literature on the participation decision (Diamond, 1980; Keen, 1997; Boadway et al., 1999; Saez, 2002b) finds that when the participation decision is the relevant margin, negative marginal tax rates are desired. Since this is applicable for lower-income workers, it provides some justification for the use of income-conditioned refundable tax credits for low-income employed taxpayers. Occupational choice, on the other hand, leads to similar prescriptions as the case with intensive labor supply margins (Saez, 2002b). At the upper end of the income distribution, variability of labor supply and occupational choice are more important options, especially for persons that are self-employed or entrepreneurs. One might expect that this would suggest lower tax rates at the top end if all sources of income are included in the tax base. However, it is not apparent how responsive labor supply at the top end is to after-tax earnings. Income effects may be important for these persons. Indeed, their income relative to others may be more important for them than their absolute income, as findings in happiness and behavioral economics suggests (Layard, 2005; Besley and Layard, 2008), in which case variability of earnings would call for more

Individual income taxation 103 progressivity rather than less. Also, for many high-income persons, the distinction between labor and capital income may be opaque. More important for practical purposes, an income tax that taxes both capital income and earnings at the same rate constrains the extent of progressivity that might be attained. Capital income might be much more elastic with respect to tax rates than labor income, especially as it accrues to higher-income taxpayers. To accommodate this, the income tax rate structure chosen would be less progressive than if it applied to labor income alone. Schedular approaches that tax capital income at a lower rate than labor income might serve to lift what would otherwise be an impediment to high marginal tax rates on earnings at the upper end. In any case, on the basis of these considerations, it is not clear that measures as extreme as linear progressive taxation are called for, especially if capital income is separated from labor income for tax purposes. On the other hand, other sources of elasticity of the supply of income at the upper end can temper progressivity. For one, high-income earners may be more internationally mobile, in which case standard fiscal competition models lead to lower average tax rates at the top. (This will also be relevant for progressive taxes used by sub-national jurisdictions in federations.) For another, it will be easier for higher-income persons to understate their incomes for tax purposes (i.e., evade taxes or masquerade earnings as capital income) than low-income earners, whose tax is typically withheld by employers. More generally, income tax evasion reduces the extent of optimal progressivity, possibly dramatically. For example, Chander and Wilde (1998) argue that even if incomes are fixed, the possibility of evasion can lead to a regressive tax structure if penalties for evasion are reasonably restricted. A further complicating factor is the fact that income inequality may be influenced by things other than exogenous abilities or productivities combined with endogenous work effort. Individual outcomes may be to some extent the result of pure luck, resulting from terms-of-trade effects, productivity shocks, and industrial innovation. Given that these shocks are beyond the control of individual taxpayers, progressive taxation will have limited efficiency costs, and will serve as an insurance mechanism at the same time.4 Desired progressivity can also be influenced by the fact that skills are to some extent endogenous. Individuals can augment their human capital by education and training. The outcomes from investment in human capital are to some extent risky, in which case taxing their returns, which accrue disproportionately to high-income persons, is a form of social insurance. Perhaps more important, the cost of education, including post-secondary education, are borne partly by the public sector (apart from forgone earnings, which are a substantial part of the cost). Recouping these costs by

104

The Elgar guide to tax systems

a progressive tax on earnings seems reasonable, and is an imperfect substitute for policy instruments like income-contingent loans that are often advocated for financing post-secondary education. There are also some arguments for restricting the progressivity of the individual tax. High-skilled and low-skilled labor might be complementary in the production process. If so, progressive taxes that reduce the supply of high-skilled labor have as a general equilibrium effect the reduction of low-skilled wages (Stiglitz, 1982). Progressivity might be limited on these grounds. Also, earnings might vary considerably over the life cycle, both because of a relation between age and earnings and because of temporary fluctuations. Given that, annual income gives a misleading impression of permanent income or consumption. To the extent that averaging of the tax base is feasible, either through the tax system itself or via self-averaging as can take place through a consumption tax system of the sort outlined above, this will not be a problem. However, few tax systems allow averaging on a lifetime basis. In addition, at least some income inequality can reflect different preferences for leisure. If otherwise equally productive persons make different choices about how to use that productivity, it could be argued that those that choose to exert more effort to earn more income should not be penalized relative to those who prefer to consume leisure. More generally, it is argued that persons should be compensated for characteristics over which they have no control (e.g., their endowed ability) but not characteristics over which they do have control (such as their choice of leisure). The former is the principle of compensation and the latter the principle of responsibility (Roemer, 1998; Fleurbaey and Maniquet, 2007). Though the implications of this point of view for taxation have not been fully exploited, they do suggest some constraint on progressivity to avoid penalizing those with relative preferences for work (Boadway et al., 2000; Choné and Laroque, 2009). The principle of responsibility has other potentially far-reaching consequences when individuals make different decisions when confronted with similar opportunities. These include the tax implications of one’s family status, over which one presumably has considerable control; the tax treatment of saving and risk-taking, which reflect one’s discount rate and risk aversion; and the tax treatment of bequests and inheritances, which are largely voluntary to the donor but beyond the control of the recipient. If one takes the principle of responsibility seriously, the tax system might (1) be immune to family circumstances, so be based solely on individual incomes or consumption, (2) neither penalize nor reward saving or risktaking, in which case consumption might be favored, and (3) ignore the giving of bequests or gifts, but include inheritances fully as income.5

Individual income taxation 105 One final consideration in the choice of progressivity of the income tax is the recognition that at least as much redistribution takes place on the expenditure side of the budget as on the tax side. Expenditure programs such as public education, health care, and other social insurance programs like pensions and unemployment insurance are significantly redistributive. Indeed, that is one of the main reasons why they are delivered through the public sector. Once these are factored into overall redistribution, the need for progression of the income tax might be reduced. A special problem arises for persons at the bottom end of the income distribution, both because they might be more deserving and because they earn little or no income so are not liable to pay income taxes. These persons will likely be transfer recipients, and in principle the transfers may either be delivered through the income tax system as refundable tax credits or separately through standalone transfer programs. The increasing use of refundable tax credits has been an important innovation in individual income tax systems. They have been used to target transfers to low-income families with children as a way of improving opportunities for such children. As mentioned, they have also been used to support lowincome workers partly to encourage participation in the workforce. They have the advantage of being easily administered, taking advantage of the income tax administrative machinery and the ability to target according to individual or family income. They provide an enhanced degree of progressivity to the tax system by being geared to income, unlike other tax credits that are equal per capita and not refundable. To the extent that all tax credits and exemptions were made refundable, this would turn the income tax system into a full-fledged negative income tax system. At the same time, refundable tax credits co-exist with other separately administered transfer programs, such as welfare transfers, disability transfers, unemployment insurance, and transfers to the elderly. In some cases, the distinction between transfers delivered through the income tax system and those delivered separately simply reflects institutional arrangements that existed before refundable tax credits became widely used. However, there are some differences between the two methods of delivery that are important. Transfers delivered through the tax system are based on the self-reporting mechanism used in tax systems whereby taxpayers report their incomes and personal circumstances, and any monitoring is done ex post by selective audits and possible penalties. The use of the income tax collection system to deliver transfers can be quite efficient, and may reduce the stigmatization that often accompanies transfers. At the same time, such a system may have some disadvantages as a way of delivering transfers. There may be both type I errors (inadequate coverage due to deserving persons not getting the proper transfer) and type II errors

106

The Elgar guide to tax systems

(leakages due to undeserving persons receiving transfers). The former arise partly because of failure to apply and to file income tax system forms, forms that may not have been required for low-income persons who are in a taxpaying position for whom taxes are deducted at source, and partly because of errors of administration. Type II errors occur because of errors in application, changes in circumstance, or outright dishonesty. The system of ex post monitoring and sanction is of limited use here, unlike for the taxpaying population. It is difficult to recoup transfers that have been erroneously made to low-income persons simply because they are likely to have spent them in the meantime. Related to this is that the responsiveness of refundable tax credits to changing circumstances, both positive and negative, is limited. Tax credits are typically based on the previous year’s taxable income, so cannot easily cope with the levels of income volatility that many low-income persons face and that are otherwise very costly to them from a self-insurance point of view (Chetty and Looney, 2006). Standalone transfer systems that monitor recipients ex ante for eligibility and assess their needs are more responsive, albeit not perfectly so given the possibility of errors, agency problems with program administrators, and the likelihood of stigmatization. Standalone transfer systems also have the advantage that they are able to take account of eligibility criteria other than income-based ones, and also to deliver benefits that include more than money transfers. Thus, transfer eligibility may differ for those who are able to work and those who are unemployable (e.g., the disabled). There may be conditions attached to the receipt of transfers, such as the requirement to participate in job search, training, or workfare programs. Program benefits may include in-kind transfers in addition to income, including health care and pharmaceuticals, counseling and rehabilitation, housing, food, transportation, daycare and so on. Assessing the needs for these ancillary items presumably requires discretionary ex ante monitoring rather than self-reporting. In practice, there will be a balance between transfers delivered through the individual tax system and those delivered separately. What seems clear is that, at least in some countries, the tax-transfer system has not been particularly successful at alleviating poverty at the lower end of the income distribution, as recently documented by OECD (2008). Enhancing refundable tax credits delivered through the tax system to complement standalone transfer programs could go some way to addressing that issue. In summary, the issue of progressivity of the individual tax system is multi-dimensional and value-laden. It involves not only the choice of the rate structure that has been the focus of much of the public economics

Individual income taxation 107 literature, but also the composition of the tax base, and the various issues associated with targeting benefits to those most in need.

4

SPECIAL ISSUES IN INDIVIDUAL TAX DESIGN

Individual taxation fulfills important functions both as a revenue-raising device and as an instrument for achieving social and economic goals. As such, it is one component of a larger arsenal of fiscal policies. There are many special design issues that arise in addressing multiple objectives and in coordinating the individual tax with other fiscal instruments that are complementary in objective. In this section, a number of these special issues are considered. Relation with the Corporation Income Tax As mentioned, one role that the corporation tax fulfills is as a withholding device against capital income generated at source in domestic corporations. To the extent that individuals own corporate shares, income accruing on them can be sheltered from income tax by retaining profits within the corporation and allowing capital gains to accrue from retained earnings investment without realization. As well as facilitating tax sheltering, this would provide an advantage to mature firms with accumulated cash flows for whom internal financing would be a less expensive form of finance than for firms who are forced to raise external funds. To avoid this problem, the corporation tax can be designed as a withholding device to ensure that corporate-source income pays some tax as the income accrues rather than when it is repatriated to the shareholders through dividends. To the extent that the corporation tax is taken to serve a withholding function, credit should be given to domestic shareholders when funds are taken out of the corporation, and this is usually done by some tax crediting device, such as a dividend tax credit against individual income taxes. To the extent that corporations operating domestically are foreignowned, the corporation tax also withholds income from foreign shareholders. This can be done without discouraging foreign investment if such firms are able to credit taxes paid abroad against tax liabilities in their home country. In such cases, the corporation tax essentially shifts tax revenues from foreign treasuries to the domestic one. The ability to do this varies from country to country, so the corporation tax may end up repelling foreign investment from countries that do not offer full foreign tax credits. In that case, the argument for using the corporation tax as a withholding device against domestic shareholders is compromised.

108

The Elgar guide to tax systems

There are a number of problems that arise with the use of the corporation tax as a withholding device. If the capital income of domestic shareholders is taxed at progressive rates, the corporation tax cannot withhold the appropriate amount from each shareholder. A case might be made for setting the corporation tax rate at the highest individual tax rate to ensure that enough tax is withheld: those with lower tax rates will eventually obtain a credit. However, such a tax rate might be too high given the effect the corporation tax has on firms whose shareholders cannot claim a credit. A particular problem might arise with shares held in tax-sheltered form, such as pension funds. In principle, dividend tax credits should also be available to these shareholders so that sheltering is in fact effective. A similar problem arises from the fact that different corporations might face different tax rates. For example, some might be in a non-taxpaying position because of cumulated losses. It is practically impossible to base dividend tax credits in the individual tax system on tax rates actually paid by corporations. An alternative approach might be to apply the integration at the corporation level by offering corporate tax credits on dividend payments. However, it would be difficult to do this in a way that differentiates between dividends paid to domestic and foreign shareholders. The case for integration is also undermined by open economy considerations. If the country were a small open one in international capital markets, domestic savings would be segmented from domestic investment in capital markets. In these circumstances, rates of return facing both savers and investors would be fixed by world capital markets (apart from a country risk premium), and any integration measure implemented at the individual tax level, such as by a dividend tax credit, would effectively subsidize saving without offsetting taxes paid by corporations (Boadway and Bruce, 1992). The corporation tax would distort investment decisions, and the dividend tax credit would distort savings decisions, compromising the usefulness of integration in the first place. This argument is contingent on the fact that international capital markets do segment domestic savings and investment, which is something that has been called into question since the original work of Feldstein and Horioka (1980). All of these problems would be avoided under a consumption tax system. Since capital income is not taxed at the personal level, there would be no need for integration on that account. Moreover, the corporation tax could be devoted solely to capturing a share of rents generated by the corporate sector. It is well-known how a tax might be designed with this purpose in mind. See, for example, Auerbach et al. (2008). A dual tax system that applies a constant tax rate to all capital income at the personal level can also avoid some of these problems. If the corporation tax rate is set equal to the individual tax rate on capital income, the

Individual income taxation 109 correct amount of tax would be withheld from all individual taxpayers, at least for taxpaying corporations. Of course, problems with integrating sheltered savings would still apply, as would the more general issue of whether integration is effective in an open economy. Sub-national Individual Taxation In federal nations and in some unitary nations as well, sub-national governments share individual tax bases with the national government. This can be done with varying degrees of independence ranging from completely separate sub-national tax systems to highly harmonized arrangements whereby sub-national governments impose surtaxes on national taxes. An intermediate arrangement might allow sub-national governments to apply their own rate structures to a common tax base used by the national government. There are some advantages to allowing sub-national governments to have access to individual taxation. It is a broad-based tax that can be used to generate tax revenue at relatively low administrative cost. The ability of sub-national governments to choose their own tax rates is said to enhance the accountability of sub-national governments by giving them responsibility for raising a substantial part of their own revenues. These accountability advantages can be achieved in ways that do not compromise national economic objectives. Collection and compliance costs can be minimized by having a common tax base, which allows a single tax collecting authority for both national and sub-national taxes. Sub-national governments could choose a single surtax rate or even a rate structure and that would be compatible with a common base and collection machinery. There are two main problems with decentralizing revenue-raising authority to sub-national governments. One is that different regions would likely have different revenue-raising ability and so could provide different levels of public goods and services at similar tax rates, or the same levels at different tax rates. This would have the effect of either inducing taxpayers to migrate to regions with higher fiscal capacities, which would be inefficient, or leaving otherwise identical taxpayers in different regions with different fiscal treatments. This might be viewed as a violation of horizontal equity or national solidarity. This problem is typically remedied in federations and unitary nations alike by a system of equalization transfers from the national to the sub-national governments based on differences in their fiscal capacities (Boadway, 2004a). A second potential problem is that the sub-national governments use their fiscal discretion to behave strategically to enhance their own fiscal

110

The Elgar guide to tax systems

capacity at the expense of other jurisdictions. This is more likely the more discretion they have. If they can choose their own rate structures, they might be tempted to make them less progressive in order to attract high-income persons and those with personal businesses.6 This effect can be mitigated to the extent that the national government retains enough individual tax room to be able to achieve whatever amount of progressivity it wants. Also, individual tax systems that leave less discretion for sub-national governments to influence capital income tax rates will reduce the incentive to compete for mobile capital. This would be the case if the individual tax base were consumption. It would also be the case under a dual income tax, especially if only the labor income portion were used by sub-national governments and not the capital income component. One final problem that arises with allowing sub-national governments to have access to individual taxation is the allocation of the tax base among the sub-national jurisdictions when taxpayers can change their jurisdiction of residence between tax years. In the case of annual income, this is not a major problem. Some arbitrary date in the tax year can be chosen to establish residency for tax purposes. More complicated is the treatment of sheltered savings that have been accumulated in one jurisdiction when a taxpayer moves to another. In the case of international migration, taxpayers who emigrate are often forced to withdraw their assets from sheltered forms and make them taxable. This seems less feasible for migration across regions within a country since the taxpayer would presumably want to keep sheltered savings intact. Any attempt to allocate tax liabilities among sub-national jurisdictions accurately will be more complicated than it is worth. It is better to rely on the equalization system to compensate for any fiscal capacity differences that arise from interregional migration. Tax Treatment of Pensions The pension system in most countries consists of various elements, most of which interact with the individual tax system. Following the taxonomy of the World Bank (1994), pension systems consist of three pillars: basic transfers for the needy elderly, contributory public pensions, and private pensions. Basic transfers fulfill redistributive objectives, and are often targeted on the basis of individual or family incomes for tax purposes. Contributory public pensions and private pensions are to some extent sheltered from taxation. Moreover, contributory public pensions are typically mandatory, while private pensions are most often not. Several issues arise in how the tax system treats pensions, or more generally saving for retirement, which can include other assets such as housing and annuities.

Individual income taxation 111 Mandatory participation in contributory public pensions for those employed is a common feature in many countries. The need to mandate saving for retirement is typically based on behavioral arguments, particularly the observation that many persons systematically set aside insufficient funds for their own retirement. This may be due to excessive discounting of the future, an inability to refrain from current consumption, or an anticipation of obtaining public support in the event of inadequate retirement income. At the same time, mandatory public pensions are typically inadequate by themselves. The levels of retirement income they offer are limited, and the coverage is restricted to those who have been employees. Basic public pensions provide one form of backstop, at least for low-income persons. Tax incentives for private, voluntary saving for retirement provides a backstop for middle- and upper-income taxpayers. Tax incentives for retirement savings mimic the treatment of savings under a consumption tax. Savings in private pension plans are typically treated as registered assets for tax purposes, although in some countries tax-prepaid retirement savings schemes are permitted. Savings in housing, which constitute a significant proportion of assets held by retirees, are treated on a tax-prepaid basis. Indeed, tax incentives for housing may be even more generous than that since some countries offer mortgage interest deductibility while at the same time not taxing the imputed return on owner-occupied housing. (On the other hand, housing may be subject to property taxation, which undoes some of the preferential tax treatment.) These tax incentives for retirement saving may vary according to the nature of the individual tax base. If a consumption tax were implemented, the issue of preferential tax treatment would not arise since all assets would be tax sheltered either as registered or tax-prepaid assets. Countries that deploy income taxes typically allow tax sheltering of retirement savings up to some specified limit, whether in registered or tax-prepaid forms. Indeed, the combined treatment of pension savings and housing makes the individual income tax much closer to a consumption tax than a comprehensive income tax in most countries. Similar issues arise in dual tax systems. Given that capital income is taxed at a lower rate to begin with, there may be less reason for offering tax incentives for retirement saving especially since doing so imposes inter-asset distortions in capital markets. An emerging policy issue is the adequacy of private pensions. In some countries, a substantial proportion of these are provided by employers on behalf of their employees. In recent years, the share of the workforce that has private pensions has declined significantly, partly because of changes in the nature of the workforce (e.g., part-time workers, self-employed and contract workers, high rates of turnover) and partly because of the cost and risk to firms of financing pension funds. Combined with the fact that

112

The Elgar guide to tax systems

workers have limited propensity to save for their own retirement even in tax-sheltered forms, there is a danger that large numbers of retired workers will have inadequate financial support. Given that, there has been some interest in providing further inducement to save for retirement short of forced saving. One proposal that has begun to be implemented in some countries is the introduction of occupational pensions, perhaps through the public sector, which workers automatically enroll in unless they choose to opt out. For a fully rational and far-sighted worker, the choice between opting-in and opting-out would make no difference. However, evidence suggests that a much smaller proportion of workers choose to opt out when they are otherwise involuntarily enrolled than choose voluntarily to opt in, reinforcing the behavioral view of saving for retirement. The issue of providing inducement to save for retirement over and above tax incentives is relevant whatever the choice of individual tax system. That is, it would apply equally well in a consumption tax system where all savings for retirement are tax-sheltered as under an income tax system, dual or otherwise, where retirement savings may be preferentially sheltered. Human Capital Investment The tax treatment of human capital investment raises a number of special issues that, like retirement savings, leads to various forms of preferential treatment. Broadly speaking, there are two types of costs of human capital investment: forgone earnings and resource costs, such as tuition, books, and supplies. These costs might be partly financed by student loans, and the outcomes of human capital investment are typically risky. In the case of forgone earnings, the tax treatment mirrors registered asset treatment in that it is based on cash flows. Forgone earnings are effectively like deductions from current income, while future earnings are taxed when they accrue. Thus, investment in the form of forgone earnings is fully taxsheltered. Of course, the tax rate applicable to forgone earnings may well be much less than that applicable on future earnings, so an ideal system would allow some averaging. To the extent that tuition and the cost of books and supplies are tax-deductible, that too would accord with cashflow tax treatment, analogous to registered assets. It is common to give some relief to such costs, although it may not be full, and it may be based on tax credits rather than deductions. In addition, income tax systems may give tax incentives for saving for post-secondary education, especially by parents. In any case, it is clear that a significant proportion of investment in human capital is tax-sheltered, further emphasizing the point that even

Individual income taxation 113 so-called income tax systems are in fact closer to consumption-based individual taxes. The case for tax incentives for human capital investment is not at all clear. No doubt there is an externality associated with human capital investment, but that is at least partly captured in the generous subsidies that governments offer to educational institutions. There may be behavioral reasons for encouraging investment in education on the grounds that students may discount the future benefits excessively. Three additional factors have a bearing on how the individual tax system treats human capital investment. One is that for many students human capital investment requires financing, and capital market constraints might make credit hard to come by or costly since future earnings cannot be fully used as collateral. A second is that, as mentioned, human capital investment is inherently risky, both in terms of the chances of success (e.g., grades) and the earnings if successful. Third, human capital investment addresses the objective of equality of opportunity, which is an objective that is sometimes found in national constitutions. These three factors can be addressed by a combination of policy instruments, some of which rely on the income tax system. Governments often enact student loan programs to deal with credit constraints, and may relate eligibility to family income. These programs can simultaneously address credit constraints and uncertainty by making their repayment incomecontingent. Income-contingent student loan programs can suitably be delivered through, or at least coordinated with, the individual tax system. Note that these programs ideally fulfill efficiency functions – undoing credit constraints and insuring risk – rather than redistributive ones, so in principle could be operated on a self-financing basis using actuarial principles. Equality of opportunity, however, is a redistributive objective. It can mean very different things to different persons. At one extreme, it could mean using the education system as a device for equalizing the ability of all persons to earn income (regardless of their endowed skill levels). This is probably much too ambitious a notion of equality of opportunity and would be very expensive. A more common, and limited, notion of equality of opportunity suggests that persons of equivalent ability ought to have equal opportunity to fulfill their potential. According to this notion, the role of government would be to ensure that persons should not be deterred from acquiring human capital because of inadequate resources, especially those provided by their families. Policies to address equality of opportunity in this sense involve targeted assistance (grants) on the basis of need (and qualification). Such student grant schemes can be operated on a standalone basis, but presumably with the size of grants conditioned on family income as reported to the tax system.

114

The Elgar guide to tax systems

Wealth Transfers Taxpayers may make or receive transfers of wealth in a variety of ways. These include voluntary transfers to charities or non-profit organizations, bequests to family members and others, and non-cash transfers, including transfers in kind and voluntary work. In the case of bequests, there may be a separate tax applied, but otherwise transfers have some implications for individual taxation. There are both efficiency and equity implications of wealth transfers. Consider these in turn. The efficiency consequences of wealth transfers depend partly on the motive for the transfer and partly on the form. In the case of the former, an important distinction is between whether the transfer is made voluntarily or not.7 For voluntary transfers, one might argue, following Kaplow (2008), that there is an externality involved. Both the donor and the recipient obtain welfare from the transfer (the donor by revealed preference), yet the donor takes into account only his or her own benefit. To the extent that the welfare of both the donor and the recipient should count from a social welfare perspective, a corrective subsidy on transfers could be justified. If this were done through the tax system, donors would receive a subsidy for voluntary transfers whether to charity or to their heirs. The income tax system typically does provide some such subsidy on charitable donations, though typically not on donations to heirs. On the other hand, to the extent that recipients of voluntary transfers do not have them included in their tax base (or otherwise fully taxed under a separate inheritance tax), some preferential treatment is indirectly given. Transfers may be involuntary for a couple of reasons. They may be unintended bequests of wealth that have been held until death for precautionary or other reasons. In this case, no externality is generated since there is no benefit to the donor so there is no case for subsidization. Another form of transfer is one that is made in return for some service rendered. So-called strategic bequests are those made in return for care provided to the donor by their heir. This is like any other transaction, albeit not at arm’s length on a market, so does not deserve special treatment. On the contrary, recipients of the transfer who provide the personal service to the donor ought in principle to treat it as taxable income. Like any non-market transaction, that is difficult to enforce. In practice, it will be difficult to distinguish voluntary from involuntary transfers, and that will compromise the desire to subsidize transfers, except perhaps those to charitable organizations. The equitable treatment of wealth transfers can be contentious. If transfers are voluntary, they give rise to benefits for both the donor and the recipient. If both of these benefits count from a social welfare perspective, that should be reflected in the tax liabilities they both face. The

Individual income taxation 115 transfer would be like any other act of consumption by the donor out of their income, so no credit would be given except to correct for the externality from the donation. (One could further argue that indirect taxes on consumption, such as the VAT, should also apply to voluntary transfers, although the non-market nature of these transfers makes that unenforceable.) For the recipient, the transfer is like any other receipt of income and should be included in the tax base (unless part of it is saved in sheltered form). This would also be the case for requited transfers, such as strategic bequests. The inclusion of transfer receipts in the individual tax base can also be justified on grounds of equality of opportunity. On the other hand, some might argue that the same transfer should not count twice from a social welfare perspective, once in the hands of donors and once for recipients (Cremer and Pestieau, 2006). If one took that view, no additional tax should be levied on wealth transfers. These issues of the tax treatment of wealth transfers are distinct from the treatment of capital gains that have accrued on transferred wealth. In some countries, transfers of wealth trigger the deemed realization of capital gains for tax purposes. This is most often the case for countries that do not otherwise tax wealth transfers. In fact, some countries that do tax wealth transfers at the same time rebase the value of transferred wealth each time it is transferred without realization. In principle, the issue of the tax treatment of accrued capital gains on transferred wealth and the taxation of wealth transfers themselves are quite distinct issues. It would be perfectly rational to both tax accrued capital gains on wealth transfers and tax the transfers themselves. Tax Treatment of Children The tax treatment of children gives rise to some comparable issues since children are heavily dependent on transfers of all kinds from their parents. If one follows a social welfare approach to individual tax design, a number of issues arise. Despite the fact that children are dependent on parents, one would want them to count from a social welfare perspective. That being so, they would typically be owed a transfer given their absence of earnings, though that transfer might well accrue to their parents. How big would the optimal redistributive transfer be? Following Kaplow (2008), children might need less consumption to generate a given level of utility than an adult, so the transfer could be scaled down accordingly (although to account for differences in the ability to generate utility by children would be inconsistent if one did not also do so for, say, the disabled or the elderly). This kind of argument would justify refundable child tax credits payable to parents, as is the case in some tax systems.

116

The Elgar guide to tax systems

At the same time, since children depend on voluntary transfers from their parents, similar issues about double counting arise as with other wealth transfers. On the one hand, intra-family transfers provide benefits both for the donor parent and the recipient child (and other family member, for that matter), so there is an a priori case for subsidy of such transfers to capture that externality (although parent–child transfers might be fairly inelastic in any case, so there is little gain from a corrective subsidy). On the other hand, if the benefits of the donor parents count for social welfare purposes, this would represent a gain in utility that would ordinarily attract taxation. Presuming that the efficiency case for encouraging transfers counters the equity case for additional tax on the donors on social welfare grounds, one is still left with a case for refundable tax credits for children. The case for making those refundable tax credits progressive (i.e., declining with family income) might be based on equality of opportunity grounds. Children obtain apparent advantages from being raised by more affluent and highly educated parents. Such parents can make greater transfers to children, and can also make highly valuable transfers of human capital (e.g., knowledge and skills as well as educational financing). While it would be difficult to compensate for these differences in childhood advantage completely, refundable tax credits that are geared to family income would be one component of an equal opportunity policy, along with educational policies from an early age and financial assistance for higher education.

5

INCOME, CONSUMPTION, OR DUAL INDIVIDUAL TAX?

What does this all imply for the choice of an individual tax system? We consider the three main options sequentially in what follows. Income Taxation Income taxation has been the dominant form of direct taxation in OECD countries. Its intellectual foundation is the notion of comprehensive income being the best measure of ability to pay, following the influential arguments of Schanz (1896), Haig (1921), and Simons (1938).8 More recent literature on tax theory and policy has called into question many of the premises that underlie the case for comprehensive income taxation as the ideal, and therefore the case for income even less comprehensively defined as the best actual choice of tax base. The arguments are based on

Individual income taxation 117 the three main criteria adduced to evaluate a tax system: administrative ease, efficiency, and equity. The administrative problems of implementing a comprehensive income tax are nigh insuperable. There are two main sources of problem: nonmarket goods and asset income. Comprehensive income, following Hicks (1946), consists of consumption and increments of wealth (saving) on a per period basis. Consumption should in principle include all sources of consumption including not only those purchased on the market but also those obtained off-market, including household production, shadow economy production, and leisure activities. The inability to include these off-market items in the tax base is a problem for virtually all direct tax bases, and this essentially turns the quest for an individual income tax base into a secondbest policy issue to which we return. The problem of taxing asset income is what distinguishes all main contenders for individual taxation as we have mentioned. There are some types of capital income that are difficult to measure and are often left out of actual tax bases. They include assets like housing and consumer durables whose return takes an imputed form, human capital that has been accumulated, and the return on personal businesses. Asset income that occurs in an accrued form, such as capital gains, is also difficult to include in the tax base. As mentioned, one of the main rationales for corporation taxation is to serve as a withholding device against capital gains that have accrued within the corporate sector. In addition, for policy purposes, many countries offer tax sheltering for some types of saving, especially saving for retirement. Apart from the problem of measuring asset income, there is the further problem of separating out real from inflationary income since the latter represents a repayment of principal rather than asset income. Finally, given the mobility of capital, some assets earn their income abroad where they may or may not also face taxation. These problems are insuperable enough that income tax systems treat different sources of asset income very differently. A substantial proportion of asset income – perhaps most – is fully sheltered from tax. Other parts are offered preferential treatment (e.g., capital gains). The result is a complex tax system, and one that invites tax planning opportunities and perhaps even tax evasion. These administrative problems in themselves tend to favor alternative tax systems that either avoid or simplify the taxation of capital income, such as those discussed below. Efficiency arguments also militate against income taxation, which requires that capital income be taxed at the same rate as labor income. The simplest way to think about this is by imagining a simple setting in which the taxpayer obtains welfare from present consumption, future consumption, and leisure. All three of our alternative choices of individual taxation

118

The Elgar guide to tax systems

– income, consumption, and schedular – apply to present and future consumption, but not leisure. Consumption taxation taxes present and future consumption at the same rate, while income and schedular (dual) taxation tax future consumption at a higher rate than present consumption. In the case of income taxation, the relative rate of taxation of present and future consumption is determined by the fact that capital and labor income are taxed at the same rate. The differential tax on future consumption tends to be less with dual taxation since capital income is generally taxed at a lower rate than labor income. While, in general, efficiency is not guaranteed by taxing present and future consumption at the same rate, it is also the case that the optimal differential can be positive or negative, and it is most unlikely to be as highly positive as comprehensive income taxation would require. One can construct reasonable technical arguments for avoiding taxation on capital income, though most involve empirical properties of individual utility functions that are typically not satisfied or verified. In theory, it depends on the substitute/complement relationships among present consumption, future consumption, and leisure. If future consumption is relatively more complementary with untaxed leisure than present consumption is, one would want to tax future consumption at a higher rate (i.e., tax capital income), and vice versa.9 The empirical evidence on this is not compelling. Even if future income were more complementary with leisure, it is probably more difficult to make a compelling argument on efficiency grounds that the relative tax rate on present and future consumption should be determined by taxing capital income at the same rate as labor income. One mildly compelling technical argument for not taxing capital income is as follows. If one does not know whether the optimal tax on capital income is positive or negative and either is roughly equally possible, it is most efficient to opt for zero taxation. The reason is that the deadweight loss of taxation is convex in the tax rate, so the expected deadweight loss will be minimized if the rate is set to zero. In addition, the fact that some forms of capital income are sheltered from tax makes the efficiency argument for income taxation even more suspect. On second-best grounds, one might expect that if some forms of capital income are not taxed, that would suggest taxing others at a lower rate to avoid exacerbating interasset distortions. Equity arguments parallel the efficiency ones. Ideally, the tax base should reflect an index of the welfare level of the individual, at least assuming one takes a welfarist point of view of the normative objective of government. Given the inability to tax non-market activities, including leisure, the issue is which tax base best reflects welfare in the second-best

Individual income taxation 119 sense. As in the efficiency analysis, the argument turns on the complementarity relationship between consumption and leisure. Applying the theory of optimal income taxation to an inter-temporal setting, the AtkinsonStiglitz Theorem suggests that if present and future consumption are weakly separable from leisure in the intertemporal utility function, capital income should not be taxed (Diamond, 2007). Taken together, these arguments based on administrative ease, efficiency, and equity conspire against comprehensive income as an ideal form of taxation. The one remaining argument that might be adduced in favor of income taxation is that it may be difficult in practice to distinguish between capital and labor income in some cases. This is particularly so for income earned in unincorporated personal businesses where sole owners act as managers as well as suppliers of the capital of the business. This has proven to be a challenge for dual tax systems that would tax capital and labor income at different rates (Christiansen, 2004), and would also be so for consumption-based tax systems. However, this is a relatively small income source and may not be of sufficient importance to serve as a determining factor for the choice of tax base. As a final point, it should be recalled that the full taxation of an individual includes not only the individual tax system but also indirect consumption taxes (VAT) and payroll taxes for which the individual is liable. Taken together with an income tax system, these result in an overall tax system that entails a lower rate of tax on capital income than on labor income. To that extent, efficiency and equity arguments against income taxation are blunted, though administrative costs remain an issue. Consumption Taxation The arguments for consumption taxation follow directly from the above discussion. A consumption tax base can be constructed by sheltering capital income using a combination of two equivalent methods. Starting from a tax base defined as nominal income, savings can be sheltered in registered form, in which case savings are deducted immediately and accumulate tax-free as long as they are held. When they are drawn down for consumption, the full amount taken (both principal and asset income) is added back into the tax base. Alternatively, savings can be tax-prepaid in which case no deduction is given for savings and no tax is paid on future asset income. Registered asset treatment is more suitable for some types of assets (personal business, human capital investment), while taxprepaid treatment is more suitable for others (housing and consumer durables). If the taxpayer is allowed to choose between registration and tax-prepayment for other assets, this can be done to facilitate averaging

120

The Elgar guide to tax systems

over the life cycle so as to avoid horizontal equity problems arising from progressive taxation. From an administrative point of view, consumption taxation avoids many of the problems of income taxation. There is no need to measure capital income, and no need to worry about inflation indexing of capital income (though indexing is still required for the rate structure). The tax base can be defined in cash flow rather than accrual terms, which greatly facilitates the taxation of personal businesses and human capital accumulation. Consumption tax treatment does not preclude mandatory saving schemes for retirement. Indeed, most such schemes use consumption tax principles already. The main administrative cost is the need to maintain a record of assets held in registered form so that tax will ultimately be paid when they are run down. Special problems arise in the case where taxpayers leave the country, which typically triggers de-registration. Death may do so as well. The assessment of the efficiency and equity advantages of consumption taxation is more ambiguous. As with income taxes, leisure and other nonmarket activities are not taxed, and both efficiency and equity considerations depend on the relationship between these untaxed goods on the one hand and present versus future consumption on the other. If non-market goods were weakly separable in the inter-temporal utility function from consumption purchases, a case can be made on theoretical grounds for consumption taxation, as we have seen. This is probably the strongest economic argument for consumption taxation. There are, however, other considerations. The self-averaging property of the consumption tax avoids the horizontal inequities that arise under an income tax system for persons whose income fluctuates over the life cycle. At the same time, there may be arguments for taxing capital income that go beyond the standard substitute–complement relationships. For one, to the extent that inheritances and other windfall transfer are not taxed, capital income taxation can be an indirect way of getting at consumption financed from these sources. For another, if household incomes are uncertain and full insurance is not available, taxing capital income can serve as a substitute form of insurance (Golosov et al., 2007). Finally, capital income taxation is useful in an optimal tax setting if more productive persons save more than less productive ones because they have lower discount rates (Diamond, 2007; Banks and Diamond, 2008). Two further points are worth mentioning before turning to the final tax alternative. The first is to recall that the rate structure can be chosen independently from the base. The income tax is, however, constrained to apply the same rate structure to capital and labor income. This might in turn constrain the rate structure to be less progressive than would otherwise

Individual income taxation 121 be desirable given the fact that capital income might be more elastic than labor income, especially given the greater mobility of the former and the greater possibilities of avoidance and evasion. Also, the case for taxing earnings more progressively than capital income is enhanced by the fact that earnings apparently account for much more inequality than does capital income, and earnings inequality to at least some extent reflects human capital investments that have been largely financed by the public sector. Second, the inability of governments to commit might make it difficult not to include capital income in the tax base. And, this applies no matter whether governments are benevolent or self-serving. That being so, consumption taxation might be viewed as infeasible in which case the choice of tax base then must take account of the least costly way of including capital income in the base. From that point of view, the dual income tax to which we now turn offers an attractive alternative. Dual Income Taxation Dual income taxation taxes capital and labor income according to different rate schedules, and as such it can take many different forms. We restrict attention mainly to the broad form that has been adopted in the Nordic countries, and which has come to be associated with the term dual income taxation. In this system, labor income and transfers are taxed according to a progressive rate schedule. Capital income is taxed at a flat rate, typically the rate associated at the lowest labor income tax bracket. Some forms of capital income may continue to be sheltered, such as housing and durables, human capital investment, and/or saving for retirement. Also, two backstop taxes may be in place. A corporate tax can be levied at the same rate as the tax on individual capital income, and it can be integrated with the latter. And, a tax on wealth transfers (or wealth itself) can add further progressivity to the tax system if applied on wealth transfers above some minimum size. A dual income tax is a useful compromise between an income tax that taxes capital income at the same rate as labor income, and a consumption tax that taxes all capital income at a zero rate. By applying separate rate structures, the relative rates of capital and labor income taxation can be chosen at will. The dual income tax offers a number of advantages (Boadway, 2004b). The schedular approach allows for very different rates of progressivity applied to labor and capital income. In particular, the rate structure applied to labor income can be more progressive than under an income tax, and that applied to capital income less progressive. As mentioned,

122

The Elgar guide to tax systems

this seems to be generally desirable given the inequality of the distribution of labor income relative to capital income, and the higher elasticity of capital income with respect to the tax rate. From an equity point of view, this might be the strongest feature of the dual income tax: the elimination of the constraint on progressivity imposed by including capital income in the base. The ability to set a separate tax rate on capital income addresses the typical optimal tax case for a relatively low rate on capital income. The constant tax rate applied to capital income also has some administrative advantages. It facilitates collection and compliance by allowing financial institutions to withhold the tax on behalf of individuals since all pay the same rate. It makes integration relatively easy since the corporate tax rate can be set equal to the personal capital income tax rate. Any perceived equity disadvantages of a low tax rate can be partly offset by a complementary tax on large inheritances. The administrative problems that remain are twofold. First, as with any capital income tax regime, there is always the possibility to avoid or evade taxation by holding wealth abroad. Even in this case, the dual income tax might be useful in that, if it is used by several countries, withholding can apply to international capital flows. The second problem is that with a relatively low tax rate on capital income, there is a strong incentive for those with personal businesses to take as much of their income as capital income rather than labor income. As mentioned, this has proven to be a problem in the Nordic countries, and it can really only be remedied by applying arbitrary rules to determine the division of personal business income into labor and capital components (Christiansen, 2004). How progressive the tax structure applied to earnings should be is a matter of judgment, and the optimal income tax literature is of limited help. The degree of progressivity depends on such things as the distribution of skills, uncertainty with respect to outcomes, the ability to vary income by effort and occupational choice, and the degree of aversion to inequality by policy-makers. The traditional optimal income tax literature found marginal tax rates to rise rapidly at low skill levels, moderately so in the middle of the skill distribution, and fall at the upper end unless the skill distribution was not truncated at the top (Mirrlees, 1971; Tuomala, 1990). In the latter case, optimal income tax rates might rise at the upper end and follow a U-shaped pattern (Diamond, 1998). In the extreme case where the government has an infinite aversion to inequality (i.e., a maximin social welfare function), the pattern of marginal tax rates can be declining throughout the skill distribution, leading to a strictly concave pattern of average tax rates (Boadway and Jacquet, 2008). The pattern at the bottom end changes if the margin of labor supply is the so-called extensive one

Individual income taxation 123 in which persons can choose not to participate in the labor market. In this case, the marginal tax rate can be negative at the bottom, gradually decreasing to become positive at moderate income levels (Saez, 2002b). The agnosticism of these results leaves the policy-maker ill-informed by the literature, and it is not surprising that some have advocated a linear progressive, or flat, tax with constant marginal tax rates (Hall and Rabushka, 1995). Such a tax is progressive owing to its exemption level (especially if there is refundability of negative tax liabilities), but it is rather less progressive than a schedule with rising marginal rates. The latter case with rising marginal rates accords better with the optimal income tax literature in terms of the progressivity achieved at both the top and bottom end. Moreover, with capital income segregated from labor income in the tax schedule, there is no apparent advantage to a constant marginal tax rate. Overall, the flexibility and administrative simplicity of dual income taxation makes it arguably the strongest candidate for personal taxation. It was one of the two options recommended by the President’s Advisory Panel on Federal Tax Reform (2005), the so-called Growth and Investment Tax Plan option. Other Issues Whichever of these individual tax bases might be preferred, a number of other issues must be addressed, including those outlined earlier. In the case of income tax bases that include capital income, some forms of capital income may still be sheltered either out of necessity or out of choice. In principle, the imputed rent on housing could be included in the base, although including it can be done only imperfectly. Moreover, one might want to shelter it as a way of encouraging persons to save in housing, given the behavioral tendency to under-save. Similarly, a case can be made for using tax sheltering as a means of encouraging saving for retirement. In this case, the sheltering can take the form of both registered and tax-prepaid treatment to ensure that the advantages of self-averaging are available. Arguments can also be made for sheltering human capital accumulation on the grounds of both feasibility and desirability. The cash-flow treatment of human capital investment is the most reasonable way to accomplish this. That part of human capital investment consisting of forgone earnings is already implicitly treated on a cash-flow basis, and the incremental financial costs (e.g., tuition, books) could be so treated as well by being fully expensed. Note that this requires that the expensing be refundable to the extent that incomes are insufficient to cover them. In addition, the

124

The Elgar guide to tax systems

individual tax system – whether based on income or consumption – could be used in a proactive way to address issues of risk, liquidity constraint, and equality of opportunity. As mentioned, an income-contingent loan scheme integrated with the tax system and run on an actuarial basis would deal with risk and liquidity constraint, while an income-targeted student grant scheme can be used for equality of opportunity objectives. The treatment of voluntary transfers must also be decided, whether for charitable purposes or as transfers to heirs and other family members. As we have seen, there is a prima facie efficiency case for subsidizing such transfers on the grounds that they create benefits both for the donor and for the recipient, while on equity grounds the recipients of transfers should be taxed on what is a windfall increase in income. In the case of transfers to charitable causes, the efficiency arguments might dominate the equity ones for a couple of reasons. For one, those who receive benefits from such transfers are likely to be low-income persons whose tax liabilities would be minimal in any case. For another, voluntary contributions to charity might be thought of as a substitute for government transfers. Given that the latter must be financed by distorting taxes and that the government might face difficult agency problems in allocating such transfers, social goals might be enhanced by subsidizing individual transfers to charity. Transfers to one’s heirs raise similar issues except that in this case, recipients are more likely to be taxpayers. To the extent that such transfers are involuntary or are payments for services, they should be treated as income in the hands of the recipients with no offsetting credit for donors. In the case of voluntary transfers, efficiency might call for some incentive to donors that might to some extent offset the benefit of the transfer to recipients. In practice, this consideration might rationalize the fact that inheritances are treated somewhat more leniently than other sources of income, especially those made inter vivos. Finally, there is the broader issue of the tax treatment of the family in the direct tax system. This involves settling a number of conceptually difficult issues. If one takes an individual welfare approach to the design of the income tax, tax liabilities of each family should be related to their individual utilities, which in turn are related to the income or consumption spending of the family. Translating this into practice involves the following considerations: 1.

How is the income shared among family members? One can assume that the family allocates its resources ‘rationally’ so as to maximize the aggregate welfare of family members. If so, adopting the family as a tax unit might be justified. Of course, this presumes that the definition of what constitutes a family is agreed.

Individual income taxation 125 2.

3.

What are the implications of household production and shared consumption? To the extent that one or more persons in a family engages in household chores that provide consumption services to all members, that should in principle be reflected in the tax liability. This consideration might justify differential taxation of families with a non-working adult relative to those in which all adults work and earn the same family income. Similarly, since consumer durables are shared among family members, there are economies of scale in consumption. This would suggest that two-adult families with a given family income pay more taxes than two single persons with the same aggregate income. (At the same time, one single person with a given income would pay more tax than two married persons with the same total income.) How should children affect tax liabilities? Various issues are involved here. First, children might require less consumption to achieve a given level of utility compared with adults. If so, policy prescriptions are ambiguous on social welfare grounds (Kaplow, 2008). If the society’s social welfare function were strictly utilitarian – so all that counts is the sum of utilities – children should be favored by the tax system since they generate more utility per euro than adults. On the other hand, if the social welfare function exhibits inequality aversion – so puts some weight on equalizing levels of utility – children should bear correspondingly heavier tax burdens. Second, children rely on transfers (cash, in-kind, and parental attention) from their parents, so general issues of how to treat voluntary transfers come into play. As we have discussed, there are two main considerations. On efficiency grounds, there is an externality associated with voluntary transfers that may call for a Pigouvian subsidy. Given that transfers to one’s child are likely to be inelastic, this is perhaps not a major consideration. On equity grounds, transfers give rise to multiple sources of utility: to the donor, to the child, and perhaps to third parties (other family members). These added sources of utility might be expected to attract higher tax liabilities. If so, families with children would be taxed more heavily than those with identical income but no children. Finally, equality of opportunity considerations might warrant giving special attention to children in less privileged families. An important way in which parents assist their children is by investing in their books, and so on. Children lucky enough to be born into higher income and higher educated families start life with an advantage. This is recognized by the provision of compulsory public education for all children. But in addition, many countries offer income-contingent transfers to families with children (e.g., refundable tax credits). Some

126

The Elgar guide to tax systems go even further. For example, in the UK, all children are given at birth a capital grant that accumulates with interest in their name and is available to them once they reach the age of majority. This is intended as a (modest) means of enhancing opportunities.

What all this implies for the design of actual tax systems is not clear. Many tax systems take account of at least some of these considerations. Transfers for children are common, and sometimes further rationalized on the basis of improving a nation’s demographic circumstances. Families with a non-working adult are often treated more harshly than those with the same family income and two working adults. And, economies of living together may be taken into account by two-adult families with a given income paying more than two singles with the same total income. Of course, all this assumes that one’s choice of family circumstances deserve compensation. As mentioned, it might be argued that, to the extent that choice of family size is a free choice, it should be neither rewarded nor punished.

NOTES * 1.

2. 3.

4.

5.

Prepared for ‘Tax Systems: Whence and Whither (Recent Evolution, Current Problems and Future Challenges)’, a conference sponsored by Savings Banks Foundation of Spain (FUNCAS) and UNICAJA, Malaga, Spain, 9–11 September, 2009. Thus, in the absence of inheritances or bequests and of transfers, the lifetime budget constraint of a household is simply that the present value of consumption expenditures equals the present value of earnings. With bequests and transfers, the present value of consumption equals the sum of the present value of earnings, transfers, and net inheritances. In fact, much earlier, consumption had been proposed as the ideal tax base by Mill (1848). Acemoglu et al. (2008) have recently constructed a political economy model of optimal taxation in which commitment is enforced by the voting mechanism. In particular, using a retrospective voting model based on Ferejohn (1986), voters vote on the basis of the past performance of politicians. Politicians are precluded from exploiting their inability to commit by being punished by being thrown out of office. However, such models require that both politicians and voters be infinitely lived (or behave as if they were). On the other hand, Low and Maldoom (2004) argue that progressive taxation, while reducing risk, also reduces the need for self-insurance. To the extent that this takes the form of precautionary labor supply, the case for progressivity is reduced. On the other hand, Sinn (1996) argues that the insurance affected by progressive taxation and the welfare state encourages risk-taking, which may be a good thing. On the other hand, there may be efficiency arguments for subsidizing bequests, as Kaplow (2008) argues. Voluntary bequests presumably give benefit both to the donor (by revealed preference) and to the donee. Assuming that both of these benefits count from a social welfare perspective, the giving of bequests is analogous to an externalitygenerating activity, so a corrective tax is warranted. This is separate from the case for taxing inheritances on redistributive grounds. We discuss this further below.

Individual income taxation 127 6. As the fiscal federalism literature has stressed, there may also be strategic interaction between the sub-national governments and the national government. Increases in subnational taxes that reduce the relevant tax base will also reduce national taxes, so-called vertical fiscal externalities (Boadway and Keen, 1996; Dahlby, 1996). Sub-national governments would have some incentive for excessive taxation. It is not clear that in practice this constitutes a serious case against revenue decentralization. 7. Transfers might even be forced as in the case of theft of property. Given that this is illegal, the tax consequences are perhaps less pressing than the role of detection and punishment. Nonetheless, to the extent that victims of crime suffer a property loss, one might argue that this should be deductible like any other loss in income. The perpetrators of the crime will presumably suffer the consequences if caught. 8. Interestingly, Wildasin (1990) has questioned whether Haig actually was a comprehensive income tax proponent. He argues convincingly that to Haig, income taxation was actually inferior to consumption taxation. Nonetheless, comprehensive income taxation is typically referred to as Schanz-Haig-Simons income taxation. 9. This argument is for the simple case where individual consumption can be aggregated into present consumption, future consumption, and leisure. In more complicated cases where consumption and variable leisure occur in each period, the case for taxing capital income becomes even more unclear (Erosa and Gervais, 2001).

REFERENCES Acemoglu, Daren, Michael Golosov and Aleh Tsyvinski (2008), ‘Dynamic Mirrlees Taxation under Political Economy Constraints’, mimeo. Alvarez, Yvette, John Burbidge, Ted Farrell and Leigh Palmer (1992), ‘Optimal Taxation in a Life-Cycle Model’, Canadian Journal of Economics 25(1), 111–22. Atkinson, Anthony B. (1973), ‘How Progressive Should Income Tax Be?’, in Michael Parkin and A. Robert Nobay (eds), Essays in Modern Economics, London: Longman, pp. 90–109. Atkinson, Anthony B. and Joseph E. Stiglitz (1976), ‘The Design of Tax Structure: Direct vs. Indirect Taxation’, Journal of Public Economics, 6(2), 55–75. Auerbach, Alan, Michael Devereux and Helen Simpson (2008), ‘Taxing Corporate Income’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, chaired by Sir James Mirrlees, London: Institute for Fiscal Studies, mimeo, available at: http://www.ifs.org.uk/mirrleesreview/press_docs/corporate.pdf; accessed 18 April 2011. Banks, James and Peter Diamond (2008), ‘The Base for Direct Taxation’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, chaired by Sir James Mirrlees, London: Institute for Fiscal Studies, mimeo, available at: http://www. ifs.org.uk/mirrleesreview/reports/base.pdf; accessed 18 April 2011. Besley, Timothy and Richard Layard (eds) (2008), Special Issue: Happiness and Public Economics, Journal of Public Economics, 92(8–9), 1773–862. Boadway, Robin (2004a), ‘The Theory and Practice of Equalization’, CESifo Economic Studies, 50(1), 211–54. Boadway, Robin (2004b), ‘The Dual Income Tax System – An Overview’, CESifo DICE REPORT, Journal for Institutional Comparisons, 2(4), 3–8. Boadway, Robin and Neil Bruce (1992), ‘Problems with Integrating Corporate and Personal Taxes in an Open Economy’, Journal of Public Economics 48(1), 39–66. Boadway, Robin and Laurence Jacquet (2008), ‘Optimal Marginal and Average Income Taxation under Maximin’, Journal of Economic Theory, 143(1), 425–41. Boadway, Robin and Michael Keen (1996), ‘Efficiency and the Optimal Direction of FederalState Transfers’, International Tax and Public Finance, 3(2), 137–55. Boadway, Robin, Nicolas Marceau and Motohiro Sato (1999), ‘Agency and the Design of Welfare Systems’, Journal of Public Economics, 73(1), 1–30.

128

The Elgar guide to tax systems

Boadway, Robin, Maurice Marchand, Pierre Pestieau and Maria del Mar Racionero (2000), ‘Optimal Redistribution with Heterogeneous Preferences for Leisure’, Journal of Public Economic Theory, 4(4), 475–98. Chander, Parkash and Louis L. Wilde (1998), ‘A General Characterization of Optimal Income Tax Enforcement’, Review of Economic Studies, 65(1), 165–83. Chetty, Raj and Adam Looney (2006), ‘Consumption Smoothing and the Welfare Consequences of Social Insurance in Developing Economies’, Journal of Public Economics, 90(12), 2351–6. Choné, Philippe and Guy Laroque (2009), ‘Negative Marginal Tax Rates and Heterogeneity’, Institute for Fiscal Studies Working Paper No. W09/12, London. Christiansen, Vidar (2004), ‘Norwegian Income Tax Reforms’, CESifo DICE REPORT, Journal for Institutional Comparisons, 2, 9–14. Corlett, W.J. and D.C. Hague (1953), ‘Complementarity and the Excess Burden of Taxation’, Review of Economic Studies, 21(1), 21–30. Cremer, Helmuth and Pierre Pestieau (2006), ‘Wealth Transfer Taxation: A Survey of the Theoretical Literature’, in L-A. Gérard-Varet, S-C. Kolm and J. Mercier Ythier (eds), Handbook of the Economics of Giving, Reciprocity and Altruism, Volume 2, Amsterdam: North-Holland, pp. 1107–34. Dahlby, Bev (1996), ‘Fiscal Externalities and the Design of Intergovernmental Grants’, International Tax and Public Finance, 3(3), 397–411. Diamond, Peter A. (1980), ‘Income Taxation with Fixed Hours of Work’, Journal of Public Economics, 13(1), 101–10. Diamond, Peter A. (1998), ‘Optimal Income Taxation: An Example with a U-Shaped Pattern of Optimal Marginal Tax Rates’, American Economic Review, 88(1), 83–95. Diamond, Peter A. (2007), ‘Comment on Golosov et al’, NBER Macroeconomics Annual 2006, 365–79. Erosa, Andrés and Martin Gervais (2001), ‘Optimal Taxation in Infinitely-Lived Agent and Overlapping Generations Models: A Review’, Federal Reserve Bank of Richmond Economic Quarterly, 87(2), 23–44. Feldstein, Martin and Charles Horioka (1980), ‘Domestic Saving and International Capital Flows’, Economic Journal, 90(358), 314–29. Ferejohn, John (1986), ‘Incumbent Performance and Electoral Control’, Public Choice 50(1–3), 5–25. Fleurbaey, Marc and François Maniquet (2007), ‘Compensation and Responsibility’, in K.J. Arrow, A.K. Sen and K. Suzumura (eds), Handbook of Social Choice and Welfare, Volume 2, Amsterdam: North-Holland. Golosov, Mikhael, Aleh Tsyvinski and Iván Werning (2007), ‘New Dynamic Public Finance: A User’s Guide’, NBER Macroeconomics Annual 2006, 317–63. Haig, Robert M. (1921), The Federal Income Tax, New York: Columbia University Press. Hall, Robert E. and Alvin Rabushka (1995), The Flat Tax, Stanford: Hoover Institution Press. Harberger, Arnold C. (1964), ‘Taxation, Resource Allocation and Welfare’, in J. Due (ed.), The Role of Direct and Indirect Taxes in the Federal Revenue System Princeton: Princeton University Press, pp. 25–70. Hicks, John R. (1946), Value and Capital, 2nd edition, Oxford: Clarendon Press. Kaldor, Nicolas (1955), An Expenditure Tax, London: Allen and Unwin. Kaplow, Louis (2008), The Theory of Taxation and Public Economics, Princeton: Princeton University Press. Keen, Michael (1997), ‘Peculiar Institutions: A British Perspective on Tax Policy in the United States‘, National Tax Journal, 50(4), 779–802. King, Mervyn (1980), ‘Savings and Taxation’, in C.A. Hughes and G.M. Heal (eds), Public Policy and the Tax System, London: Allen and Unwin. Layard, Richard (2005), Happiness: Lessons from a New Science, London: Penguin. Low, Hamish and Daniel Maldoom (2004), ‘Optimal Taxation, Prudence and Risk-sharing’, Journal of Public Economics, 88(3–4), 443–64.

Individual income taxation 129 Meade Report, Meade, J.E. (1978), The Structure and Reform of Direct Taxation, Report of a Committee chaired by Professor J.E. Meade, London: George Allen & Unwin. Mill, John Stuart (1848), Principles of Political Economy, with Some of their Applications to Social Philosophy, London: J.W. Parker. Mirrlees, James A. (1971), ‘An Exploration in the Theory of Optimal Income Taxation’, Review of Economic Studies, 38, 175–208. OECD (2008), Growing Unequal?, Paris: OECD. President’s Advisory Panel on Federal Tax Reform (2005), Simple, Fair, and Pro-Growth: Proposals to Fix America’s Tax System, Washington, DC: US Treasury. Roemer, John E. (1998), Equality of Opportunity, Cambridge, MA: Harvard University Press. Saez, Emmanuel (2002a), ‘The Desirability of Commodity Taxation under Non-linear Income Taxation and Heterogeneous Tastes’, Journal of Public Economics, 83(2), 217–30. Saez, Emmanuel (2002b), ‘Optimal Income Transfer Programs: Intensive vs Extensive Labor Supply Responses’, Quarterly Journal of Economics, 117(3), 1039–73. Schanz, George von (1896), ‘Der Einkommensbegriff und die Einkommensteuergesetze’, Finanzarchiv, 13, 1–87. Simons, Henry C. (1938), Personal Income Taxation, Chicago: University of Chicago Press. Sinn, Hans-Werner (1996), ‘Social Insurance, Incentives and Risk-taking’, International Tax and Public Finance, 3(3), 259–80. Stiglitz, Joseph E. (1982), ‘Self-Selection and Pareto Efficient Taxation’, Journal of Public Economics, 17(2), 213–40. Tuomala, Matti (1990), Optimal Income Tax and Redistribution, Oxford: Clarendon Press. US Treasury (1977), Blueprints for Basic Tax Reform, Washington, DC: Government Printing Office. Wildasin, David (1990), ‘R.M. Haig: Pioneer Advocate of Expenditure Taxation?’, Journal of Economic Literature, 28(1), 649–60. World Bank (1994), Averting the Old Age Crisis, Oxford: Oxford University Press.

4

The challenges of corporate income taxes in a globalized world Emilio Albi*

1

INTRODUCTION

This chapter reviews corporate income taxation in the context of the economic globalization experienced in the last 30 years. Given present flows of capital and income between nations and the importance of multinational firms, due consideration can no longer be given to corporate taxation without contemplating international issues. The main purpose of this chapter is to examine the current corporate tax trends derived from the changes that occurred in the last three decades with a view to defining potential policy prescriptions aimed at making corporate taxation less distortionary and costly. To make this issue more manageable, our analysis will focus on European Union (EU) countries, chiefly on EU(15) corporate taxation, paying special attention to Spain whenever appropriate. Other European or non-European OECD countries may be considered as terms of reference. The chapter will take an economist’s approach and I will attempt to provide a comprehensive summary of what we know about the economic issues of corporate income taxes (CITs) and their alternatives. The agents involved in tax systems, however, are far more diverse than economists, and have their own objectives, which will also be considered here. In particular, policy-makers and tax officials give priority to revenue, administration and enforcement issues, while the business community may pay more attention to the effects of statutory and average effective tax rates or compliance costs, and less to the distortions associated with marginal tax rates, a main interest of economists. Within the EU, common tax policy and the rulings of the European Court of Justice (ECJ) provide constraints applicable to national CITs. International tax competition and income shifting among nations, or the difficulties of shareholder taxation, are also aspects of considerable significance. Complexity, therefore, is the rule in tax systems, steering discussions toward less distortionary and costly solutions, within the established constraints, as opposed to more perfect, albeit impracticable, alternatives. 130

The challenges of corporate income taxes 131 Traditional CITs, levied on the return to equity, apparently create a differential burden on shareholders, in as much as CITs are not shifted away from corporations. The initial study by Harberger (1962) on the incidence of corporate income taxation, established that the tax burden affects capital throughout the economy, not just corporate equity. Corporate taxes cause capital to flow to unincorporated productive sectors, and capital and labour adjustments result in the tax burden affecting all capital, incorporated or not. Relaxing Harberger’s assumptions, and introducing more issues into his model, has resulted in a large number of subsequent studies that have changed our perspective of who bears the CIT burden. For our purpose in this chapter, we could well accept the balanced conclusion in the recent review of this topic by Auerbach (2005), who states that instead of all capital, the CIT burden may be borne by labour (if CIT hits new savings and investment, which reduces capital and leads to lower wages) and shareholders (especially in the short run, as shareholders may be unable to shift the tax on ‘old’ capital, or economic rents, as well as tax-induced managerial underperformance, which diminishes the advantages of corporate ownership). Classical corporate taxation problems still affect present day tax systems, although they are enhanced by international issues. CITs distort investment and financing decisions, and may discourage the distribution of dividends or, in some cases, encourage incorporation. In an international setting, the CIT induces tax competition among governments and income shifting between nations in the search for lower tax liabilities by multinational firms. Compliance and enforcement costs, or avoidance and anti-avoidance schemes, consume real resources and lead to reduced welfare. Nonetheless, CITs are well and alive in most tax systems. The traditional reasons for justifying corporate taxation have been enhanced by the internationalization of the economy. The withholding function of CITs, acting as a backstop to personal taxation, makes sense in the case of resident and non-resident shareholders, particularly if the latter are non-resident in name only. National and multinational corporations benefit from public expenditure where their operations are located, and a CIT is not a bad surrogate for user charges that may otherwise be impracticable to implement. Finally, CITs not only affect normal profits but also location-specific or firm-specific economic rents, so governments can easily tax all of them, thus avoiding differentiation problems and only taxing location-specific rents.1 The two types of rents are often difficult to distinguish unless the latter are derived from natural resources. These reasons are also valid for the alternatives to CIT considered in this chapter.

132

The Elgar guide to tax systems

In spite of the reasons for maintaining corporate taxation in a tax system (as a CIT or any of its alternatives), there is still pressure on such taxes to reduce distortions and other costs. A discussion of the challenges facing corporate income taxes in EU economies should start by an overview of the evolution of CITs in the last few years of economic globalization. This is best accomplished by analysing present CIT trends, focusing on the European tax systems, and this comprises Section 2 of this chapter. Section 2 reviews new trends in corporate income taxation, starting by considering traditional CITs and their alternatives, with some practical experiences discussed. Regarding existing CIT rate and revenue trends, we will focus on international tax competition (with the effects of different tax rates on corporate investment, location and income-shifting decisions) and the rate–revenue paradox. This is followed by a consideration of base broadening and the new targets for tax credits. Shareholder taxation is the fourth subject considered in Section 2. Changes in the taxation of shareholders may influence both corporation and shareholder behaviour. Last in this section, we also consider the concessions granted in many countries to small and medium-sized enterprises (SMEs). Section 3 summarizes the constraints imposed by the EU and ECJ on the tax systems of member countries and the impact of other OECD initiatives. Although these constraints are neither decisive nor numerous, they must be considered before we can continue. At the same time, this Section will describe cases where international coordination has been established in spite of the apprehensive reaction of governments to the EU or OECD when fiscal sovereignty is touched, however lightly. The proposed EU consolidated tax base is also considered. In Section 4, we shall be discussing current CIT challenges in a national and international setting, together with the different alternatives available in the academic literature or offered in practice in actual tax systems. Regarding these alternatives, we focus on the main question of how they perform in solving current corporate taxation problems or whether it is best to shy away from radical reform and continue with piecemeal changes and international coordination for traditional corporate income taxes. The chapter will end, in Section 5, with a summary of conclusions including policy prescriptions regarding possible future reforms of corporate taxes.

2

TRENDS IN CORPORATE INCOME TAXATION

Very few alternatives to CITs were available 30 years ago. CITs back then had fairly high tax rates, narrow bases and a rather limited scope.

The challenges of corporate income taxes 133 Revenues were lower than at present, especially in Europe. The incentives provided by CITs were in need of rationalization, and countries largely opted for imputation systems in the shareholder taxation area. All this has changed considerably. There are clear new trends in corporate income taxation. Traditional CIT as the Norm Versus its Alternatives Conventional CITs levied on the return to equity is the standard tax on corporations’ profits worldwide, and we do not often see alternatives. These alternatives can be divided into two groups according to the sort of corporate base subject to tax. Taking a conventional CIT as our reference, the principal characteristic of each group is either: (1) (2)

to maintain normal profits free of tax and therefore levy the tax only on economic rents over the normal return on investment; or to tax all returns on capital, irrespective of whether the source of financing is debt or equity.

Types of alternatives to CIT The main types of taxes to be included in group (1) are: a.

A CIT with an allowance for corporate equity (ACE). This tax follows a seminal idea by Boadway and Bruce (1984), which was detailed by the British Institute for Fiscal Studies (IFS) just a few years later (1991). ACE relieves the cost of equity, deducting from profits the result of applying a risk-free nominal rate of interest to the amount of equity, as explained by Bond and Devereux (1995). This approximates normal profits and therefore the CIT (ACE) is levied on rents (provided there is full loss compensation, so that the ACE is not reduced in present value terms if it goes unused in any year). b. Flow-of-funds corporation taxes (cash-flow taxes), as devised more than three decades ago by the Meade Committee (1978),2 where all expenses, including capital expenses, are tax deductible when incurred. These taxes can take different forms according to three wellknown bases (R, R+F or S – the last two being equivalent). The R base comprises real inflows (all sales, even of fixed assets) minus real outflows (all purchases, including fixed assets and wages). The R+F base also includes financial inflows net of financial outflows. The S base consists of share outflows (including dividends paid and reductions in own shares) less share inflows (dividends received and the amount of new share capital issued).

134

The Elgar guide to tax systems

Alternatively, flow-of-funds corporation taxes may be implemented using a VAT-type tax with a base formed by subtracting all input and fixed asset purchases and labour costs from sales. These two kinds of taxes (cash-flow or VAT-type) can be applied on a source base or a destination base and are levied on rents arising from real capital assets or real capital and financial assets, as investment costs are deductible when incurred. The tax is also levied on the total return on ‘old’ investments if, in a transitional period after its introduction, these investments do not obtain flow-of-funds treatment. The most important taxes included in group (2) are: a.

Comprehensive business income tax (CBIT), as proposed by the US Treasury (1992). This tax disallows interest cost deductibility from traditional CIT, putting debt and equity financing of investments on an equal footing. It has been described as similar to a ‘pure’ dual income tax – see Cnossen (1996) and (2000) or Auerbach et al. (2011), among others. b. Bird and Mintz (2001) went one step further with the business value tax (BVT), a tax on the value-added of a firm (net income sort) with a base comprising revenues less purchases of inputs and depreciation allowances (labour costs being non-deductible). One can envisage the base of a BVT as a conventional CIT base to which interest expenses and labour costs are added back. In fact, Bird and Mintz rationalize the Italian regional business tax (IRAP), which will be discussed below. Other different combinations of all these alternatives to CIT, or two-tier systems with a corporate and a personal level, have been proposed. These alternatives are discussed in Section 4.

Practical experience with alternatives to CIT What practical experience of the above taxes can be found in tax systems today? As of 2008–09, and chiefly considering EU(15) countries, Belgium, on a national level, and Italy, on a regional scale, have applied alternatives to a traditional CIT in their systems. The Belgian tax follows the ACE system and Italy uses a BVT.3 Furthermore, the UK operates a North Sea fiscal regime CIT with a 100 per cent capital allowance on most capital expenditure, and other components, so that this regime acts like a cash-flow tax. Without leaving the oil industry, Norway, a non-EU member, has a petroleum tax system with such generous depreciation allowances that, when added to interest payment deductibility, normal returns on investment in this sector are practically tax exempt.

The challenges of corporate income taxes 135 Besides these examples from the resource sector (which resemble other experiences in the Canadian provinces or Australia), within Europe only Estonia – EU(27) – has been applying a distribution tax in lieu of a CIT since the year 2000. This tax on distributed profits (including employees’ fringe benefits, transfer pricing adjustments, donations and other nonbusiness expenses) is somewhat similar to the S base of flow-of-funds corporation taxation. The EU authorities, however, have seen this tax more as a withholding tax on dividends in breach of the EC Parent–Subsidiary Directive regarding distributions of profits to parent companies from EU member states. As a result, the tax no longer has applied to such distributions from 1 January 2009 onwards. Another change recently is that the distribution tax will be levied on liquidations to shareholders exceeding initially paid in capital. In general, the Estonian experience appears to have produced good results for the country’s economy – see Funke (2005) and Angelov (2006). But at present it is quite an isolated case. Outside Europe, Mexico has been operating a corporation cash-flow tax – Impuesto Empresarial a Tasa Unica4 – with a base close to the R+F base proposed by the Meade Committee, since 2008. Since January, 2006, the Belgium CIT has been applying a ‘notional interest deduction’ similar to an ACE. An important requirement makes it only applicable to corporations in which the financial year is from 1 January to 31 December. The deduction is based on the corporation’s equity (share capital plus retained earnings) at the end of the previous year. The value of the allowance is calculated by multiplying the corporation’s equity by a rate equal to the average monthly reference indices, based on the interest rate of ten-year government bonds. The rate for assessment year 2009 is determined as the average monthly reference index for such bonds two years earlier, in 2007. The rate cannot exceed 6.5 per cent and the maximum deviation between the rates of two subsequent years must be 1 per cent or less. Unused parts of the allowance can be carried forward for up to seven years only, with no interest for deferrals. The revenue cost of the Belgian ACE has been estimated by an OECD (2007) policy study as 10.4 per cent of the pre-reform CIT yield (not considering dynamic effects) and the new deduction is shown to be more profitable for SMEs. There are ample possibilities for accelerated fixed asset depreciation, especially for SMEs, and loss carry-forward (carry-back is not allowed) compensates only the nominal value of losses, without interest being applied for any time delay. Rather than for efficiency-enhancing reasons, the Belgian ACE may have been a response to the European Commission (EC)’s decision of 12 February, 2003, establishing that the

136

The Elgar guide to tax systems

coordination centre regime is an incompatible state aid. At present, ACE is not granted (among other entities) to coordination centres, which have to relinquish their own special tax treatment in order to apply the deduction but are able to use it when the period to qualify for the favourable regime ends. Other prior experiences with ACE, within EU(15), have been the Italian ‘dual income tax’ (1997–2003), well reported in Bordignon et al. (2001), or the Austrian ‘notional interest’ (2000–04). Croatia, not an EU member, implemented an ACE tax (1994–2000), studied by Keen and King (2002) and by Klemm (2006). All these past experiences may have been repealed for the same reason: the use of relatively high nominal rates to prevent much revenue loss from application of the ACE. Outside Europe, Brazil has been applying ACE, under the name of ‘protective interest’, since 1996, but only when returns are distributed to shareholders – see Klemm (2006). ACE has a counterpart in the personal income tax. At an individual shareholder level, the allowance for shareholders equity (ASE), reported by Sørensen (2005a), introduces an allowance for normal returns on equity, in which case the income tax is equivalent to an expenditure tax. This system may favour investment neutrality (provided there is full loss compensation), as it exempts normal returns on investments, and it may also reduce income shifting by corporate owner-managers (active shareholders) given the double taxation of economic rents – Crawford and Freedman (2011) suggest a combined use of ASE and ACE. Since 1 January 2006, Norway has been applying the ASE system to the shares in national and foreign firms owned by resident taxpayers. The allowance is calculated by applying the interest rate for three-month government bonds, after taxes, to the capital invested in the shares. The system’s enforcement therefore requires an administrative registry of the acquisition cost of shares, which is not easy to achieve. In my opinion, this is the greatest difficulty involved with the ASE system. A variant of this system provides shareholders with an allowance per share equal to the value of ACE divided by the number of shares. Alternatively, shareholders can obtain a tax credit equal to the allowance per share multiplied by the corporate tax rate – credit for corporate equity (CCE). With regard to the Italian ‘regional tax on productive activities’ (IRAP), introduced in 1998, its tax base has been determined since 2008 by reference to the corporation’s accounting results (irrespective of the adjustments required for CIT). Banks and financial institutions are basically taxed, under specific rules, on the difference between interest receipts and payments. For other enterprises, the tax base is the difference between revenue (plus increase in inventory or work in progress) and production

The challenges of corporate income taxes 137 costs (intermediate goods and services, depreciation of tangible and intangible assets, provision for risk and other costs). Neither labour costs nor interest payments are deductible, unlike certain personnel costs: Social Security contributions, costs of R&D personnel, some of the costs of qualifying new employees and 5000 euros per employee with an indefinite labour contract. Taxpayers with activities in more than one region apportion their tax base among them on the basis of personnel remuneration in each region. The IRAP is very similar to the BVT proposed by Bird and Mintz (2001). Its base basically comprises wages, normal profits, rents and interest payments. It is an origin-based tax on the value-added by the corporation, of the net income type as depreciation allowances are deductible. Total profits and interest payments are equally taxed, with no discrimination between equity and debt. This tax resembles the single business tax applicable in Michigan since 1976 or the business enterprise tax of New Hampshire, enforceable since 1993, although it has a larger base. It is also similar to the Hungarian ‘local business tax’. Rates and Revenues: International Tax Competition and the Rate–Revenue Paradox Since just over 20 years ago, there have been two clear trends in corporate taxation in most OECD countries: a considerable cut in statutory tax rates coupled, oddly enough, with increased revenues. This has been repeatedly discussed in the literature –see OECD (2007) or Devereux (2006) for all the references – and we can skip a detailed review. Statutory tax rates started to fall towards the end of the 1980s and in the early 1990s, and this trend continued in 2000–08. Effective corporate tax rates (marginal or average) also fell in general in most countries, although less than nominal rates due to tax base increases and the reduction of tax credits to compensate for the revenue losses derived from lower rates – see, among others, Devereux and Sørensen (2006). The reduction in statutory CIT rates was significant and constant from the late 1980s up to 2008 both in Europe and the OECD. The US, on the other hand, maintained the nominal rate of its corporation tax system at 39.3 per cent from 2000 to 2008, although it had been cut sharply in both 1987 and 1988 (the 1986 rate was 11 percentage points higher than in 1990). The falling trend was maintained in 2008 and 2009 in the EU(15), with the average nominal rate falling from 28.44 per cent (2007) to 27.20 per cent (2008) and to 25.21 per cent (2009). This trend can also be found in average OECD European and total OECD rates. Meanwhile the revenue from corporate income taxes has tended to rise

138

The Elgar guide to tax systems

both as a percentage of the GDP and total taxation during much of the same period. The most important revenue growth is found, for Europe and the OECD, in the mid-1990s (with significant increases also registered in the second half of the 1980s). From 1990 to 2000, revenue increased by more than 1 per cent of GDP in the OECD and Europe (although only by two-tenths of a point in the US). No major changes are found in this century: from 2000 to 2005, revenue followed the same trend, with a slight increase in the OECD average and a clear increase in the US. The EU(15), however, registered a fall in average revenue. In any event, we will have to wait and see what happens with corporate tax revenues after 2008 in view of the current economic crisis. It does not seem possible to predict the relative movements and possible trend changes in CIT revenue in each country, national GDPs or other tax revenues in different parts of the world. We can be sure, however, that there will be substantial variations. International tax competition Is the persistent reduction of CIT rates in the recent globalization period basically due to tax competition between nations to attract capital and income flows? One traditional theoretical argument found in the literature in the last 20 years – Gordon (1986), among others – is that, when international free movement of capital and income is established, there is a ‘race to the bottom’ regarding tax rates in small, open countries, finally resulting in zero marginal effective rates. With capital mobility, if a small open economy applies a CIT, it will face capital outflows causing an increase in the before-tax rate of return of capital that compensates the tax, with effects on immobile production factors, labour and land, and production inefficiencies. What we expect, then, is that income from mobile factors such as capital will not be taxed, as it will be preferable to obtain revenue from the immobile factors that in any event bear the tax burden. Furthermore, attempts by countries to attract multinationals with good management methods, know-how and other intangible assets (which generate firm-specific rents) will also lead to a ‘race to the bottom’ process (Gordon and Hines, 2002). This kind of prediction has not been accurate in practice, and it is not known whether it would be strengthened by greater international capital mobility. Winner (2005), however, with panel data from 23 OECD countries and especially since the mid-1980s, finds that greater capital mobility has had a significant negative impact on the capital tax burden and a positive effect on labour taxation (including Social Security). In contrast, Garrett and Mitchell (2001) do not find a negative relationship between capital mobility and tax rates, so the empirical evidence is somewhat

The challenges of corporate income taxes 139 ambiguous. Indeed, a multinational may decide to establish itself in a country, in spite of taxes, to serve its consumers better or because of product transport costs. Other important factors in the location decision of foreign direct investment (FDI) are relatively low wages or a qualified workforce. Quality of infrastructures or the management skills of the country’s business community, with good goods and service providers, are also important for FDI location. These factors represent locationspecific rents, which governments will want to tax, and can be interpreted as ‘agglomeration effects’. Baldwin and Krugman (2004) explained that with a medium level of international economic integration, and consequently significant trade costs, rich countries with good infrastructures, supplier and customer networks, business expertise, educated labour and a good technological level will have taxes that will not particularly take tax competition arguments into account, as ‘agglomeration external effects’ protect their investments and support new capital inflows. This is because ‘peripheral’ countries, with less ‘agglomeration’, find it difficult to attract much foreign direct investment merely by lowering their CIT rates. Such lower rates, however, do create pressure on the taxes of rich countries as, with economic integration progress, lower trade costs limit the advantages of ‘agglomeration effects’. With this economic geography approach, the existing, but incomplete, integration of national economies does not lead to convergence of low tax rates, although it does limit these rates. Furthermore, the suggestion that capital levies vanish with increased mobility is counteracted by clear political arguments. The voters in a jurisdiction would not easily accept the disappearance of a CIT (unless, perhaps, it is replaced by one of the alternatives) when nationally- or foreign-owned firms are earning profits. From a policy-making perspective, it could also be best to keep a CIT in a small, open economy if it affects immobile production factors, as the revenue directly and indirectly obtained from such factors would be difficult to levy in a politically acceptable way by direct means. In sum, international capital mobility is unlikely to lead to the elimination of corporate taxation, although tax competition has limited and reduced both nominal and effective tax rates. Tax competition can be established through different CIT rates (statutory, marginal or average) according to the different factors involved in an FDI decision5 that governments could wish to affect. Having decided to enter a foreign market, one basic business decision refers to the domestic or foreign location of production (in the case of pure or financial services, a foreign location is often used to serve multiple markets). In relation to goods, they can be either exported or their production located in the actual

140

The Elgar guide to tax systems

marketplace. Transport or communication costs and all factors generating the aforementioned location-specific rents are of primary importance in such decisions. With regard to taxation (besides the effects of personal taxation on expatriate staff members), the primary aspect to be considered for discrete location decision is the profit to be obtained after tax, which depends on the average effective tax rate (AETR) of corporate taxation in the country where production is located. A parallel decision refers to the scale of the investment (if it does not depend on expected share of the target market, which in turn is affected by present or future competitors). This decision is related to, but separate from, the location question. If there are no restrictions, a firm should invest to the extent that the marginal product of the capital is equal to its cost. This decision considers a CIT’s effect on the cost of capital, its marginal effective tax rate (METR). A firm might also produce at other sites, in order to supply the market. If there is already spare production capacity, the new production could be spread among these locations as a marginal decision. A further aspect, which is given high priority by firms, has to do with the ease, or difficulty, involved in locating profits in countries other than the one in which they are obtained, in order to reduce the overall tax burden. Geographical transfer of income is not difficult for a multinational group, as profits are much more mobile than investment on an international scale. They can be transferred through financial transactions, locating debt and its respective financial expenditure in countries with high taxes, financed with loans from group companies operating in places with low taxes (or, with more sophisticated structures, obtaining a double dip interest deduction through countries with low taxes). Governments protect themselves from such practices with thin capitalization rules, which can be avoided without much difficulty, interest allocation rules or by limiting the interest expenses incurred to earn foreign-source incomes. The other major alternative is transfer pricing: the use of other than market prices in international transactions to shift profits to lower tax locations. Transfer pricing policies can be implemented by selling and buying goods and services among group firms, with cost distribution or charges for central services or the use of the multinational group’s intangible assets. The tax authorities attempt to control this process with legislation on not-at-arm’s length operations that establishes relatively strict limits to the prices applied in these operations, which can occasionally lead to double taxation in both jurisdictions. The basic factor in all these business practices is the statutory rate, together with the administration’s attempts and ability to combat such practices, as they are merely income transfers between countries and

The challenges of corporate income taxes 141 business decisions do not normally consider elements of the tax base, tax deductions or any other CIT peculiarity. On the other hand, decisions regarding the location of after-tax profits (whether to keep them as reserves, to be re-invested or lent to other locations, or to pay dividends) are affected by shareholder taxation, which we shall be considering later. In sum, the different aspects of an FDI decision are influenced by three different classes of CIT rates: the AETR for investment location, the METR for its scale or the statutory rate for income shifting between jurisdictions. What empirical evidence do we have on the impact of those rates? Devereux and Griffith (1998) showed that the decisions of US multinationals to establish their production in Europe or the US were not significantly affected by tax considerations, but that they do have an impact on where they are located inside Europe. However, the same authors (2002) sustain that, although tax policies are important for business location and investment decisions, the scope of their effect is unknown. Devereux et al. (2006) analyse the case of governments simultaneously competing with METRs and statutory rates, finding that influences between countries are greater for statutory rates, to attract profits, and much weaker with METRs, to attract capital. On the other hand, it seems clear that both METRs and AETRs have fallen less than statutory rates, because of base enlargements established to maintain revenue, as explained by Devereux et al. (2002) using forwardlooking measures. Grubert (2001), however, with a sample of 60 countries, determines that statutory rates have fallen less than AETRs, although the outcome is affected by the effects on AETRs (backward-looking measures) from different factors: profitability, loss compensation, cyclic factors or income shifting. The large amount of empirical work conducted in the last 15 years in this field has been studied in a meta-analysis by De Mooij and Ederveen (2003), further extended in (2008). In the 2008 results, business behaviour relative to income shifting (caused by statutory rate differences) and discrete location choice (where the relevant variables are AETRs) suggests a greater response to tax than other business decisions. For example, the semi-elasticity of the corporate tax base to tax rates in these two cases represents an effect of –1.2, whereas for investment in the margin (METR), the semi-elasticity values range from –0.3 to –0.6, with lower values for debt/equity discrimination. Distortions relative to the form of the business activity, corporate or individual, can be considerable according to some of the studies, although they have less impact on revenue as it is obtained both from the CIT and from personal income tax. All this suggests that tax competition is more important in relation to statutory tax rates or AETRs, and that multinationals often make use of

142

The Elgar guide to tax systems

income shifting – see Grubert (1998) or Weichenrieder (2009) for German data – as they are basically interested on after-tax profits. The same conclusion is reached when considering tax competition dynamics. Overesch and Rinke (2008), using panel data from 32 European countries from 1980 to 2007, and dynamic estimations, show that countries compete strongly in relation to statutory rates and AETRs although there is little interaction relative to METRs (however, they imply that the results depend on how inter-country interaction is modelled). Considering tax competition in general, it is clear that the idea that corporate taxation would vanish in small, open countries missed the mark. Low tax-rate jurisdictions, however, in a much more globalized world economy, have established limits and reduced CIT rates. There appears to have been stronger competition regarding statutory rates and AETRs than METRs, in line with what we know about the international behaviour of firms. Multinationals pay special interest to location decisions (influenced by AETRs) and income-shifting possibilities in order to locate profits in jurisdictions with low statutory tax rates. The rate–revenue paradox While tax rates have been falling, corporate tax revenue, in terms of the GDP and total taxation, has been increasing for the last 20 years in most OECD countries, although it has either remained the same or diminished on average in the EU(15) in the twenty-first century – see OECD (2007). This would appear contradictory, and therefore explanations have been sought to solve the puzzle. One first explanation of this inconsistency between the evolution of CIT rates and CIT revenues is related to the broadening of the base that accompanied the rate cuts, especially in the second half of the 1980s. In empirical studies, special attention was paid to the reduction, in most of the OECD, of the present value of the depreciation allowances produced by lower depreciation rates for tax purposes (reduction of the present value of these allowances has become smoother since the early 1990s, given the lower discount rates that went hand in hand with lower inflation). Changes in other aspects of different legal definitions of the tax base, which are much more difficult to compare between countries, such as the disappearance of exemptions or investment deductions (partially compensated by tax benefits for R&D or environmental issues) and how expenses or provisions, loss compensation or capital gains are treated, have received less attention. The breakdown by Clausing (2007) and Sørensen (2007) of the revenue/ GDP ratio in different factors has led empirical studies to consider other possible explanations for the rate–revenue paradox, such as the profitability of corporations or the size of the corporate sector in the economy.

The challenges of corporate income taxes 143 Non-significant results were obtained with regard to the first of these factors. In relation to the second, Sørensen mentions that agriculture, with a lower level of incorporation, is less important in present day economies, and Auerbach (2006) highlights the growth found in the financial or service sectors. De Mooij and Nicodème (2008) study the income shifting phenomenon from the personal sector via the progressive incorporation of business activity in the EU, which helps to understand revenue growth arguments – also see Gordon and Slemrod (2000). Another type of income (or investment) shifting aims at countries with low rates, such as Ireland inside Europe. Using sectional data from OECD countries, Bertelsman and Beetsma (2003) show that reported value-added is negatively related to statutory tax rates. And then we have income shifting towards tax havens. Becker and Fuest (2007a), however, show that internationalization had a positive impact on 16 German Länder (in a high tax country), which is not consistent with the idea that multinationals reduce their global tax burden by shifting profits to low tax countries. The efforts of the tax authorities to counteract income shifting, or tax avoidance in general, appear to have had good results over time, which could partially explain the conclusion reached by Becker and Fuest. The empirical studies conducted on this subject can be divided into two large groups: those that refer to specific countries and those that consider a broad sample of OECD members. Besides the paper about the German Länder, the most interesting country studies refer to the UK and the US. We have selected the most recent papers. Devereux, et al. (2004), for the UK, a country that registered important statutory rate cuts and high tax revenues from 1980 to 2004, show that the broadening of the tax base only partially explains the rate–revenue paradox. Other explanations can be found in the growth of the corporate sector and the large profits of the financial sector. Auerbach (2007) shows that most of the recent revenue increases in the US are due to constraints regarding loss compensation, which in turn increase the effective taxation of corporations. Clausing (2007) reviews a sample of 29 OECD countries (all of them except Mexico) for the 1979–2002 period. She finds a parabolic relationship between tax rates and corporate tax revenues as a share of GDP, with a 33 per cent revenue-maximizing corporate tax rate for the sample, although this rate depends in specific countries on their size and openness. In general, Clausing expects small, open economies to have a lower revenue-maximizing tax rate than larger, richer or more closed economies. This is compatible with the statistics – OECD (2007) – which show greater revenue increases in poorer countries with lower nominal rates than in rich countries (with higher nominal rates). It is also compatible with the aforementioned ‘agglomeration effects’ used by Baldwin and Krugman (2004),

144

The Elgar guide to tax systems

in as much as tax competition differs according to the country; peripheral countries gain revenue by applying lower rates but advanced core nations with higher tax rates are not much affected. Devereux (2006), using a slightly different approach, obtains similar results for a panel of 20 OECD countries in the 1965–2004 period, with a revenue-maximizing rate of around 30 per cent. However, none of the tax effects are significant in the presence of time dummies and other control variables. So Devereux (2006, p. 25) is not convinced that there is a ‘systematic relationship between tax rates and revenues across OECD countries’. The contributions of Clausing and Devereux, then, leave us without a clear explanation for the rate–revenue paradox. There may be different suggestions to explain the puzzle, such as the broadening of the CIT base to counteract rate cuts, the effects of income shifting or simply the better performance of tax agencies, but we have not yet been able to explain the direct link between lower rates and higher revenues. Base Broadening – New Targets for Tax Credits The broadening of the CIT base (and even enlarging the scope of the tax to include new taxpayers or new schemes to tax passive income for instance, or to hinder circular or conduit international situations) has been an important part of the recent changes made to tax systems. In this regard, differences between accounting results and tax base computation are fading through the elimination of exemptions and concessions, and accounting standards have become generally accepted for fiscal purposes. Tax depreciation rules are possibly now more in line with actual depreciation of assets; at least, they are less generous. Tax preferences for SMEs, as discussed below, however, counterbalance this view. There are other aspects of interest in relation to the CIT base. Several rules help to broaden the tax base; these include: the gradual strengthening of the fiscal regulation of ‘not at arm’s length’ transactions; the application of ‘advance price agreements’; or rules referring to ‘undercapitalization’. With regard to loss compensation, retroactive compensation is not common but carry-forward has been extended to a larger number of years (without a nominal after-tax rate of interest). With lower inflation in the latest few years, inflation adjustments are not normally applied except for the sale of fixed assets. Note that there are obviously differences between countries in this process of changing the CIT base and not all changes are broadening efforts. Considering the EU, tax depreciation rates and methods still differ, although less markedly. For inventory valuation, the LIFO (last in, first out) method is being gradually accepted for tax purposes but is

The challenges of corporate income taxes 145 not yet commonly admitted. Accounting-wise, inventories are valued at average or FIFO (first in, first out). The capital gains arising from assets, taxed at transfer and normally integrated in the base, show different tax preferences used in different countries, largely to encourage reinvestment, through deferral or exemption of the gain or the application of tax credits against tax payable so that the effective rate is reduced. Capital loss compensation may be subject to offsetting constraints, which vary between countries. Finally, accounting standards and practices also differ considerably in the EU. Even after the application of International Financial Reporting Standards (IFRS) in 2005, in accordance with EU Regulations, to publicly traded corporations with consolidated accounts, there are differences among European countries due to the options provided to firms by international standards. For example, capitalization of financial expenses is an option for corporations in different countries (and mandatory in Spain). Options between cost or reasonable value are open to firms for fixed and intangible assets; R&D expenses may be capitalized in Spain but not in other countries. Different consolidation methods are open to groups. Outside the corporations covered by the EU Regulation, national standards apply that may follow IFRS or not (Spain, for example, has recently adapted accounting legislation to IFRS). Where local standards have not been adapted to IFRS, the above accounting variations apply to more cases, such as lease agreements or the recording of income for long-term contracts or foreign currency transactions. Although this is not a case of CIT base broadening, but is also aimed at maintaining revenue with lower rates, the elimination of fixed assets investment tax credits has been another feature of corporate taxation during the last few decades. Investment tax credits may attenuate the cost of using capital if there are high financial costs. However, the short-term efficacy of investment incentives is also costly. Besides generating efficiency costs on the labour market, tax incentives may have undesirable effects on the composition of investment, compounded by other aspects of the CIT such as rates, loss compensation, depreciation of assets, capital gains treatment or, in general, sources of financing. This lack of neutrality concerning type of investment for different sorts of corporations explains policy-makers’ decisions regarding investment tax credits. The maintenance of revenue was another reason, but it has been counterbalanced by the tax concessions provided for new activities since the 1990s in OECD countries. The new areas mainly comprise tax credits for: ●

R&D, technological innovation activities and the promotion of informational technologies;

146 ● ●

The Elgar guide to tax systems investment aimed at environmental protection or the use of renewable sources of energy; occupational training costs, including the cost of providing employees with Internet connections or equipment.

It is clear that both policy-makers and the business community are largely in agreement concerning this new tax incentive system, at least in the European setting. At present, tax credits are chiefly limited to encouraging business activities in these areas in order to increase competitiveness, although we know little of their cost–benefit balance. Shareholder Taxation A traditional issue related to corporate taxation is the avoidance of multiple or cascading taxation of corporate profits, through chains of corporations, until the income ultimately reaches the shareholder. Even if this multiple taxation disappears, at the individual shareholder level corporate income attracts a double taxation due to the operation of the CIT coupled with personal income tax. The latter’s importance depends on the theoretical view of the combined effects of the taxation of shareholders and the corporation itself. In the traditional view of the effects of shareholder taxation, newly issued capital is the corporation’s marginal source of finance. It considers that dividend taxes matter because of the higher cost of dividends relative to the tax cost of financing investment with retained earnings and because double taxation of capital gains also increases the cost of capital, thus reducing investment. Therefore, it is appropriate to limit the double taxation of equity instruments in comparison to debt, which generates deductible returns in the CIT. This traditional view is probably the most commonly accepted among tax experts. Of course, the next question is: if dividends have a greater fiscal cost, why are profits distributed and new shares issued? In this respect, some shareholder groups are not seriously affected by the higher tax cost of dividends (clientele effect), showing a preference for holding shares. Dividend payments are also a way to prevent a firm’s managers from making excessive investments against shareholders’ wishes and to show how profitable a firm is. The greater tax cost of shareholder taxation is thus compensated by these advantages, even though the net returns are smaller. The ‘new view’ (see Zodrow, 1991 among others) is based on investment being financed with retained earnings if corporations defer the payment of dividends to avoid shareholder taxation. This approach considers that dividend taxes do not affect the marginal incentives for investment financed

The challenges of corporate income taxes 147 with retained earnings. This marginal investment will generate an after (corporation) tax return, which, when paid out to shareholders together with the previous retained earnings, will face a dividend tax for the total payout. The dividend tax not incurred initially, due to the retention of profits, is eventually paid. The final net rate of return for the shareholder will be the same, provided the dividend tax rate is constant.6 As the net rate of return does not vary, the effective marginal tax rate is zero, and the dividend tax is irrelevant in relation to the marginal incentives for investments financed with retained earnings (evidently, because the argument assumes that all returns are eventually distributed and taxed at the same rate). Any reduction in shareholder taxation will imply a net gain for the owners of existing capital, as the dividend tax simply reduces the value of ‘old’ capital. Empirical studies on the validity of these two views, ‘traditional’ or ‘new’, have unfortunately not had clear results regarding firm behaviour – see Auerbach (2002) or Auerbach and Hasset (2007). On a purely international scale,7 moreover, the situation is much more complicated, given different country CITs or the possible application of withholding taxes to income obtained by non-residents. Distortions related to cross-border investments or income and profit flows in corporate groups are also important. Indeed, in such an international context, with multinationals owned by numerous individuals living in different countries, who invest through complex corporate structures in different open economies, a third approach has arisen to compete with the ‘traditional’ or ‘new’ views. This ‘third view’ considers that, with perfectly mobile capital and open economies, the cost of equity finance will be internationally formed according to a world interest rate modified by appropriate risk premiums. The cost of capital for firms will be affected by CIT at the source and any further withholding, but not by shareholder taxation, which does not affect corporations’ financing and investment decisions. Shareholder taxation will affect the return on savings but not the return on investment, the cost of which will be increased by the CIT. Therefore, shareholder taxation is not significant for investment, according to this view. However, as Sørensen (2007) sustains, a country’s small and medium-size firms do not find it easy to obtain international financing, so for them this ‘third view’ is largely not applicable. Such firms are biased towards creating reserves (in which case, the ‘new view’ could be more appropriate) or obtaining bank loans (unless they attract risk capital). The parent–subsidiary case (participation exemption) Shareholder taxation, in the parent–subsidiary case, has been tending in many OECD countries towards the exemption method. Within the

148

The Elgar guide to tax systems

EU(15), all countries apply participation exemption of dividends received by residents from non-resident corporations, except Greece and Ireland, which have imputation-credit variations. Finland only allows the participation exemption regime for dividends received from another EU member state or a country with which Finland has a tax treaty. The imputationcredit method is offered as an alternative by different EU(15) member countries as well as participation exemption. Exemption may be complete or only applicable to 95 per cent of the dividends received (the 5 per cent taxed is a way to consider the deductible administrative costs of participations in subsidiaries, which are nondeductible in most countries with full exemption). The minimum (direct or indirect) participation required to obtain the exemption is either 10 per cent or 5 per cent, although exemption is provided irrespective of the size of the holding in other countries (in some cases the minimum participation is expressed as a minimum share acquisition price value). This minimum participation has to be maintained continuously for 1 or 2 years in most countries (this period can occur before or after profits are distributed). The participation exemption system is generally not applied if foreign income has been subject to low taxation (normally understood as a tax burden of less than 15 per cent) or a privileged tax regime. For dividends received from other resident corporations, the application of participation exemption is even more widespread. Only Spain follows an apparently similar system by granting (with minimum participations of 5 per cent in other resident corporations) a credit against the recipient’s tax liability of 100 per cent of the CIT attributable to the gross dividend. Unused deductions due to insufficient tax liability can be deducted from tax liabilities in the seven immediately following years. This deduction mechanism apparently acts like an exemption. This is not the case, however, if there are losses in the recipient corporation.8 With regard to capital gains arising from share transmissions, the participation exemption system is also widespread in the EU(15). In the case of transmission of shares from another resident firm, Spain applies a variation on the credit system explained in the previous paragraph. The credit is calculated (in the capital gains case) with reference to the reserves corresponding to the transmitted shares (including those already capitalized) that have been accumulated by the firm while the shares were held, with a limit comprising the capital gain. This system suffers from the problem described at the end of the previous paragraph – see note 8 –, and is only applied in relation to subsidiary reserves generated while the shares were held by the parent company and not to the total capital gain obtained9 (although this is accepted in many cases with the participation exemption system).

The challenges of corporate income taxes 149 Other particular aspects in the EU(15) are found in Austria, which taxes the gains from resident corporation share transmissions. Greece levies a 5 per cent tax on the gain from the sale of shares (domestic or foreign) not listed on a stock exchange. Taxation of gains from listed shares is deferred in Greece if they become part of a special reserve to offset future loss from the sale of shares (both listed and unlisted). Finally, Portugal only provides a partial (50 per cent) rollover relief scheme to gains from the sale of shares in any corporation (resident or non-resident), with the condition of reinvestment of the total consideration received and if the sold holding (retained for at least one year) represents at least 10 per cent of the participated corporation’s capital or an acquisition value of 20 million euros. Portfolio investments: corporate shareholder For portfolio investments by corporations (without reaching the arbitrary shareholding level defined for participation exemption), the exemption method is also used at the place of residence of the income’s recipient. This method is not applied as a general rule, however, and there is no clear trend in international tax legislation. Individual shareholders: personal income taxation Where one of the three approaches discussed at the beginning of this section is most required is in the taxation of individual shareholders. We should remember that the traditional approach is not unanimously accepted, but the ‘new view’ is not clearly supported by empirical evidence. Application of the ‘third’ approach would depend on how open the economy is and whether the firms generating shareholder income receive international financing. The lack of a widely accepted theoretical approach explains the different solutions applied in practice and the fact that the total exemption system, which could be supported by the traditional view, is not common in personal income taxation. Furthermore, if the CIT rate is lower than the highest marginal personal tax rates, as is common in practice, the total exemption of dividends and capital gains in the latter would foster the use of incorporation for performing economic activities in a country and a tax avoidance mechanism. Therefore, only if the maximum marginal personal income tax rate (or the only rate in a linear tax) is less than that of the CIT, would the total exemption method be feasible in personal taxes. Even in this case, with foreign dividends with an underlying CIT lower than the national rate, the exemption system would encourage international investment in countries with low corporate taxes (as we will see in Section 3, the system applicable to resident shareholders in the EU has to include national and other Union dividends).

150

The Elgar guide to tax systems

Given these problems with total exemption and the practical difficulties associated with the imputation method, different systems are understandably used in the case of individual shareholders. There is a tendency to apply modified classical systems with partial exemption or reduced rates to dividends and capital gains in personal income taxes. The systems used in Europe are basically (apart from marginal application of the imputation system): ● ● ●

dual personal income tax (which, in turn, can include either of the following two systems); partial exemption of dividends or capital gains generated by share transmissions; use of a single reduced rate applicable to shareholder income.

The idea of a dual personal income tax is both simple and practical. It is increasingly found that global income taxes are taxing different tax base components differently for several reasons, especially international capital mobility. Providing a simple, homogeneous tax on returns on capital, a dual tax applies a single rate (although more than one could be used) to capital income (normally excluding income from primary residence and pension funds) and a progressive rate scale to earned income. The capital income rate is close to the lowest rate applied to earned income (although not zero) and the CIT rate. Income from business, artistic or professional activities jointly generated by labour and capital is more or less arbitrarily distributed between labour (with a progressive rate) and capital components (taxed with other capital returns). This discretionary distribution is one of the most serious difficulties associated with dual taxation. There are three countries in Europe with dual taxation: Finland, Norway and Sweden. Partial exemption is also used in Finland and an ASE system in Norway. Partial exemption is relatively common in EU(15) for dividends. Examples are France, Germany, Italy (with substantial participation)10 or Portugal. Reduced rates are applied to dividends in Belgium, Italy (without substantial participation), the Netherlands or Spain. Taxation of capital gains from share transmissions at the shareholder’s place of residence is not always the same as dividend taxation in EU(15). Belgium, for example, applies total exemption to capital gains from share transmissions, if they are not derived from speculative operations, as does Portugal if the shares are held for over a year. France, however, applies a reduced rate to capital gains.

The challenges of corporate income taxes 151 Small and Medium-sized Enterprises (SMEs) Another important CIT trend in the last few years has been the tax advantages granted to small businesses, clearly supported by the EU. SMEs are numerous everywhere and provide a large number of jobs. Defined by the number of jobs involved (they all have to have less than 250 employees), these businesses include micro (up to nine employees), small (from ten to 49) or medium-sized firms (from 50 to 249). They combine corporate enterprises (CIT) with individual entrepreneurs or other entities, the treatment of which varies from one country to another (partnerships, for instance, or co-ownership businesses), which can be taxed through the individual income tax. The tax benefits granted to these businesses in CIT (according to size measured by turnover,11 taxable income, number of employees or amount of assets) tend also to be applied in individual taxation. The statistical information available from the Observatory of European SMEs shows that, over the years, these businesses have registered less labour productivity and value-added growth than large-scale enterprises (LSEs). Employment growth, however, tends to be negatively correlated with enterprise size over time. LSEs are more likely to export although, as SMEs supply goods and services to LSEs, their indirect contribution to exports may be significant. The large number of SMEs, their lower productivity, their contribution to employment and their more direct contact with the public make them in the eyes of policy-makers more ‘deserving’ and they are generally considered worthy of better tax treatment than large corporations. Special CIT regimes for SMEs vary, but they tend to apply lower rates to all or part of the tax base, with advantages that are not always very significant for each business but represent a considerable loss of revenue overall. Other tax incentives for small business may be accelerated depreciation, or even free depreciation in special cases, and rules to facilitate compliance, such as admitting global provisions for possible insolvent debtors for tax deductibility. Higher or special tax credits are also used, especially in relation to technological innovation, information and communication technology access (largely the Internet and e-commerce), staff training or good environmental performance. A range of problems are associated with these concessions to small businesses. In addition to lost revenue, they also represent an incentive to break up corporations, which leads to greater complexity when defining SMEs, and higher control costs. If SMEs are defined by turnover, for instance, it would have to be the turnover of all the members of a group of firms, in accordance with business legislation, and the same would apply (with regard to voting rights or the appointment and dismissal of board

152

The Elgar guide to tax systems

members) to an individual alone or with his or her family up to a certain degree of affinity. The interrelation between corporate and personal taxes on the small business level is also complicated, due to the incentive to incorporate given the possibility of converting income from labour into income from capital with better tax treatment. Furthermore, there are no advantages for companies in relation to their capacity for growth or innovation, but only to their size. In view of these problems, we are forced to question the reasons for providing special tax treatment for SMEs. The arguments for special tax treatment of SMEs The OECD (1994) maintains that special tax regimes for SMEs must be based on economic reasons and respond to problems specific to such businesses. The conclusions of another report – OECD (1997) – support tax systems playing a role favourable to small businesses. What economic rationale justifies tax concessions for small firms? An initial argument refers to the ‘market failures’ facing small businesses, such as lack of information or the barriers to entry generated by LSEs. These two problems are not of great tax significance, however, as they can be best solved with non-fiscal measures. The financial constraints making it difficult for small firms to obtain financing, however, are important. Their lack of access to national or international capital markets, insufficient collateral for credit institutions and limited information explain why SMEs depend on financing from banks, generally at a higher cost, the creation of reserves, provided there are sufficient profits, or their ability to attract risk capital. The tax system can therefore help business financing with legislation that is not only favourable for small businesses but also for ‘venture capital corporations’. At the same time, the application of lower tax rates to SMEs increases the profits available for reinvestment, although this cannot be guaranteed without the obligation to create specific reserves. Higher depreciation allowances represent free government financing, which also reduces the ‘market failures’ related to financial constraints. All the above assumes that such firms are earning a profit; appropriate regulation of loss compensation in new SMEs could also be important when initial investments are made. The lack of information found in small businesses is also an argument related to the use of new technologies and the R&D field in general. Indeed, tax legislation could facilitate technological innovation, personnel training or environmental compliance in small firms. Another aspect refers to reduced tax compliance costs for small businesses that have fewer possibilities (regarding personnel or expertise) than LSEs when facing tax legislation. This problem is probably greater for

The challenges of corporate income taxes 153 medium-sized enterprises, as small or individual firms often compensate directly for those costs with varying degrees of non-compliance. It is an important issue, nonetheless, and should be contemplated by the tax authorities not only in relation to income taxation but also to VAT. Finally, with regard to the intergenerational transmission of firms, the tax costs involved due to inheritance and donation tax (IDT) are also worthy of special attention – see European Commission (1994). This does not only affect SMEs but all family businesses in which the entrepreneur is not just the owner but represents a group of people with family ties. Individual entrepreneurs are indeed family businesses and many SMEs fit into this category, but a good number of family enterprises are large corporations. The basic problem of IDT in relation to family businesses is actually paying the tax because the property transferred between generations usually consists of productive assets, without the necessary cash being available for payment. IDT will be considered in Chapter 5 so my remarks will be limited. Possible solutions to this problem range from reductions applied to the IDT base for the transfer of family businesses to, and this probably makes more sense, deferral and fractioning tax payments for some years. Given these arguments in favour of providing tax concessions to SMEs, it is hardly surprising to find a large number of such concessions in current tax systems. This policy is somewhat traditional in the EU and can be expected to last. These arguments, however, are not fully convincing, especially because the tax system does not distinguish between SMEs with growth and innovation potential and others that will only last for a few years. But once a tax benefit is in place, it is not easily eliminated, especially when it enhances ‘political image’ and there are a large number of voters who are small and medium-sized business entrepreneurs.

3

INTERNATIONAL COORDINATION OF CIT: THE EU CASE

The reluctance of national governments to even slightly relinquish fiscal sovereignty in relation to direct taxes is well known and has been touched upon previously. However, a fair amount of international coordination is in place with regard to CIT. It has been largely developed through the OECD (from the 1963 Model Tax Convention to present-day transfer pricing or tax avoidance matters, including the Harmful Tax Practices Initiative) or, in a regional setting, by the EU. This part of the chapter will briefly discuss the current EU process of coordination and mutual

154

The Elgar guide to tax systems

approximation of CITs. It should be underlined from the beginning that international CIT coordination in the EU does not lead to any sort of systematic harmonization,12 and that the unanimity principle is maintained for all tax matters. The process satisfies the European principle of subsidiarity, though it establishes limited constraints for the tax policies of member states (MSs). EU Coordination Only the Directive on mutual assistance in direct tax matters was adopted prior to 1990, reinforcing the exchange of information content of most tax treaties that followed the OECD model. Anti-avoidance issues have always been of interest, chiefly in the transfer pricing field, with the European Commission working alongside the OECD. In 1990, the European Council adopted the Arbitration Convention (on the elimination of double taxation in connection with the adjustments of profits of associated enterprises, including permanent establishments). In 2006, the Council adopted a Code of Conduct on transfer pricing documentation. On 1 December 1997 the Ecofin Council agreed on different measures to reduce ‘harmful tax competition’ (EU Code of Conduct for Business Taxation), and in 1998 the Commission adopted a Communication on unacceptable state aid in the field of direct business taxation. In 1998 the OECD presented the Harmful Tax Practices Initiative. The EU Code and OECD Initiative (neither of them legally binding), together with the 1998 EC Communication, aim to discourage member countries (and tax havens) from using measures that could influence the location of investment and profits. In this respect, Keen (2001) has argued that preferential tax regimes help CIT revenues because countries tend to tax more immobile investments or activities when there is competition. Janeba and Smart (2003), however, considering the endogeneity of the total tax base, show that by eliminating tax concessions, EU countries will compete for less mobile tax bases and attract capital from other parts of the world, thus raising total tax revenue. Furthermore, the economic crisis of 2008 will probably lead to more pressure from the OECD, EU and individual countries against offshore centres. A large number of tax sanctuaries have already promised greater cooperation and the provision of more information. This may be seen with scepticism but, hypocrisy aside, there are sufficient means with which to curb non-cooperation. In any case, the results of the EU Code and the OECD Initiative cannot be described as successful so far (remember that some of the places where corporation anonymity is offered are within large economies).

The challenges of corporate income taxes 155 Three significant measures regarding corporations were adopted in July 1990: the Arbitration Convention, the Merger Directive and, most important for this chapter, the Parent–Subsidiary Directive. The Merger Directive provides for the deferral of taxation on capital gains (of corporations and their shareholders – either other corporations or individuals) on defined cross-border reorganizations within the EU (mergers, divisions or partial divisions of corporations, transfer of assets, exchange of shares, transfer of a permanent establishment or of the registered office of a Societas Europaea [European public limited company] or a European Cooperative Society). The central idea of the Directive is to defer taxation of hidden reserves of the transferring corporation until they are actually realized. It is a deferral of taxes on income accrued at source, but not realized. The historical book value of assets is maintained for the transferee’s tax purposes. MSs may, for the recipient corporation’s holding in the capital of the transferring corporation, require a minimum participation of up to 10 per cent for the former to gain the benefits of the Directive. The same scheme is also normally provided for domestic restructurings. The Parent–Subsidiary Directive deals with EU intercorporate dividends and profit distributions received by a permanent establishment (PE) of a corporation located in another MS, if profits are distributed by a subsidiary located in an MS. In many national legislations, the scope of the Directive has been extended to distributions from any country other than tax havens. MSs may decide between the following two shareholder taxation systems (or provide both, leaving the decision to the taxpayer): (1)

(2)

an imputation-credit system, covering not only the tax paid by the immediate subsidiary but also the taxes paid by any other lower-tie subsidiary; an exemption system (of no less than 95 per cent of distributed income if countries want to take into account the deductible management costs of the holding with the other 5 per cent).

A minimum holding of up to 10 per cent may be required in both systems, with an uninterrupted holding period of no more than two years (counted before and after distribution). The participation criterion may be replaced with one related to the holding of voting rights. As we have already discussed, and for the reasons given above, exemption has taken the lead as the most widely practised shareholder taxation system in the parent–subsidiary case in Europe and in many other OECD countries. This establishes a trend for source-based CITs and territorial taxation. However, the Directive contains another important element, which balances this trend by favouring a residence-based tax

156

The Elgar guide to tax systems

system: profits distributed by the subsidiary to the parent corporation are exempted from withholding tax (on non-resident shareholders). Within the EU, in the participation exemption situation, the source country cannot tax outbound distributions, only corporate profits. The residence country of the parent corporation will not tax distributed profits received, but may tax further payouts to resident or non-resident shareholders (individuals or corporations left out of the participation scheme). ECJ rulings have also influenced shareholder taxation. Following a decision of the European High Court in the Verkooijen case,13 on 19 December 2003 the European Commission – COM (2003) 810 – adopted a Communication providing guidance on dividend taxation of individuals, as portfolio investors, in the internal market. The aim of the Communication is to eliminate tax discrimination for individual shareholders so that capital markets are not fragmented in the Community. The main guideline established is that MSs cannot levy higher taxes on inbound dividends (from other MSs) than on domestic dividends. The guideline may be summarized as follows. MSs have to apply the same individual shareholder taxation systems to both inbound and domestic dividends. This provides a further explanation of why the imputation system (with credit usually restricted for domestic dividends, lest the foreign country’s high tax generates a large treasury transfer) has fallen into disuse. For outbound dividends (to other MSs), the withholding tax cannot exceed domestic taxation. Also note that the scope of this guideline includes juridical double taxation. MSs have to apply the same system to inbound dividends, giving credit for foreign withholding tax. Although the exemption from withholding tax on profit distributions by the subsidiary to the parent corporation and the main guideline of the 2003 Communication reinforce aspects of non-discriminatory residence-based taxation of dividends in the EU, this does not mean that CEN (capital export neutrality) is applied to individual shareholders. CEN would require the combined effects of CITs, withholdings and individual taxes to produce the same effective tax burden for domestic and foreign investments. This is extremely difficult in practice, if not impossible. Another step in the direction of retaining elements of residence-based taxation in the EU was provided by the 2003 Directive to eliminate withholding taxes on interest and royalty payments between related corporations from different MSs with cross-shareholdings of at least 25 per cent (including their permanent establishments). MSs may require shares to be held for two years uninterruptedly. In many cases, for both interest and mobile capital gains (obtained without a permanent establishment),

The challenges of corporate income taxes 157 exemption at source is granted to individuals or corporations resident in another MS without any further requirement. Finally, the EU Savings Tax Directive of 2005 provides for exchange of information between countries in the form of reporting interest paid to non-resident individuals to the tax authorities of the recipient’s residence. Alternatively, MSs must levy source-based withholding taxes on interest payments to non-residents (during a transitional period) – Austria, Belgium and Luxembourg have chosen to follow this alternative. These withholding rates increase with time (15 per cent for the first three years, 20 per cent for the following three years and 35 per cent thereafter), and 75 per cent of the revenue must be transferred to the recipient’s country of residence. The recipient is entitled to credit the withholding tax against his or her personal tax liability in his or her country of residence. The exchange of information to prevent tax evasion and enforce residence-based taxation of an individual’s capital income might be extended in the future to more countries if it proves effective (the Savings Directive’s transitional period will not expire until the US, and other countries, accept information exchange upon request). On the debit side, the Savings Directive generates a considerable compliance burden for EU financial institutions. It is also easy to avoid: individual investors can simply locate their wealth in tax havens outside the EU – again there is a clear need to hinder offshore operations and increase exchange of information on an international scale, if residence-based taxation of capital income is to be even partially applied. The most important loophole in the Savings Directive, however, is that it does not cover dividends, thus providing evident arbitrage opportunities between interest and dividends. The Directive’s efficacy in its present form is far from clear. Proposals for an EU Common Consolidated Corporate Tax Base (CCCTB) and Home State Taxation for SMEs Tax consolidation regimes are not widespread on an international scale, but they normally require high levels of participation in capital and their scope is reduced to resident corporations.14 Recent ECJ rulings on the Marks & Spencer and Papillon cases15 may change the latter, for reasons related to the free movement of capital, extending national tax groups to all corporations resident in any MS (it should be recalled that tax groups affect both CIT and VAT). Tax consolidation in CIT eliminates the problems of double taxation and loss offsets. It also provides a major advantage. Internationally, with strong economic integration as is found in the EU, consolidated taxation considerably reduces administration and compliance costs due

158

The Elgar guide to tax systems

to national CITs and other legislation. Albeit only for the EU group and not for transactions with affiliates outside the EU, it would eliminate the problems caused by transfer pricing, cross-border loss compensation, or cost allocation (financial, R&D or marketing) among group members. It would also facilitate business restructuring and the general application of tax and accounting standards. It makes no sense to foster a single business investment area in the EU, when firms face 27 different legislations and tax systems. All this has led the European Commission to suggest two lines of work leading to a CCCTB in the EU and Home Taxation for SMEs (HT-SMEs), in both cases applicable optionally to business activities performed in at least two member states by a group of corporations. This work-in-progress started in 2001 and there is as yet no date set for its completion. These lines of work have not considered the possibility of a harmonized tax base for CITs in the Union, and even less the establishment of a European CIT, which could be managed by the member states or an EU tax authority. A harmonized CIT base would create a situation similar to that of VAT, eliminating many of the present administrative and compliance costs. These alternatives, however, are technically difficult and politically impossible (Albi et al., 1997), so a less ambitious approach has been used. The CCCTB contemplates a common consolidated tax base system as an alternative option to national systems, according to which a group of European firms could establish tax consolidation resulting in a single tax base. This tax base would be allocated to each country (which would apply its national tax rate) according to an apportionment formula that would take different factors into consideration. The first step would therefore be to establish a common tax base. Some important aspects, among others, would refer to asset evaluation and depreciation policy, treatment of capital gains, provisions and reserves, or the consideration of international aspects with non-group members. International Accounting Standards could be of help in all these cases, once the differences between the accounting and tax approaches had been solved. The definition of tax group is also important, according to minimum degrees of participation, as is determination of the tax consolidation method to be applied. Finally, another significant factor is the choice of variables, and their weightings, to be included in the formula in order to allocate the CCCTB to each of the countries involved. The apportionment formula (AF) can include sales, payroll and/or capital. Other approaches consist of using value-added as the apportionment system, based on either source or destination (the underlying

The challenges of corporate income taxes 159 apportioning factors in this case are labour costs, interest costs and profits), or simply applying a ‘macro apportionment system’ based on the GDPs of the MSs where the group operates. These possibilities are not neutral in relation to the revenue received by each country or the group’s tax bill. They can also produce distortions, especially in relation to business location, or new factor-shifting incentives, as shown by Gordon and Wilson (1986) or Mintz (2004). The choice of apportionment system is therefore an essential part of the European tax consolidation project – see Agúndez (2006). The AF is also a focal point of the European Commission’s line of work related to SMEs, defined as companies with less than 250 employees, a turnover of 50 million euros or less and a balance sheet of 43 million euros or less. In this case, firms established in at least two member states can choose to determine their consolidated tax base according to the legislation of the parent company’s country of residence. This tax base will be allocated to the different states in which it is established, according to an AF, with each country applying its own tax rate. The AF for HT-SMEs will presumably be based on each country’s sales and payroll in proportion to the total, or just one of these variables. This home taxation system maintains the differences between European tax systems, based on the states’ mutual recognition of tax rules. Tax verification will be difficult for the tax authorities of other than the parents’ countries. Tax neutrality also diminishes, as firms could be subject to different tax rules (those of the parent’s country of residence). The system is not appropriate, therefore, other than as a first step towards the application of a single CCCTB in the European Union.

4

THE CHALLENGES OF CORPORATE TAXATION IN OPEN ECONOMIES – ALTERNATIVE BASES REVISITED

After 30 years of economic globalization, CIT now has new features in OECD countries, which can be characterized by the following stylized facts: ●

International tax competition for more mobile capital, governments’ reaction to income shifting by multinationals (lest the latter’s search for lower rates is too appealing) and the belief that limited rates (and base broadening) moderate the distortionary costs of corporate taxation, have resulted in a considerable reduction in statutory CIT rates in most OECD countries since 1985. The smaller

160

●

●

●

●

● ●

The Elgar guide to tax systems reduction in average and marginal effective rates may be explained by the simultaneous broadening of the CIT base in order to maintain revenue (which increased on average in the last 15 years of the twentieth century, in terms of both GDP and total tax revenue). CIT base broadening in the last 20 years resulted from the elimination of exemptions and tax preferences coupled with less generous depreciation allowances. Tax authorities’ performance in the field of transfer pricing, thin capitalization and tax avoidance, in general, provide another explanation of wider CIT bases. The tax concessions granted to SMEs have the opposite effect, although the overall result is a broader tax base that partially counteracts the effects of lower rates on revenue. Tax credits do not usually target capital investment, but attempt to influence corporations’ performance in favour of R&D and innovation, information technologies, environmental protection and occupational training. A careful revision of the ‘new’ CIT tax credits is still pending.16 With regard to shareholder taxation, the imputation system has been replaced in most countries by participation exemption in the intercorporate case. This reinforces source-based taxation of corporate income. The exemption system is also applied for corporations’ portfolio investments but not as generally as in the parent–subsidiary situation. Mainly for individual shareholders – dividends and capital gains – and, in some cases, for corporations’ portfolio investments, elements of residence-based taxation have been maintained by EU regulation and in practice in many tax systems. For individual taxation, pragmatic solutions such as the use of specific reduced rates or partial exemption both in dual income taxes or schedular personal income taxes are in place in European countries. International coordination of corporate taxation, though limited, is gaining strength at least in regional areas such as the EU. SMEs receive beneficial tax treatment in OECD countries and, given the political setting, this may continue for a long time.

The Challenges of Corporate Taxation in Open Economies Traditional CIT distorts many business decisions, financial or otherwise, causing neutrality deficiencies related to the different effective tax borne by assets and involving problems regarding the matching of income and expense (depreciation, interest expense, valuation of intangibles, and so on) or inflation accounting, when needed. New challenges appear with

The challenges of corporate income taxes 161 globalized economies. Tax competition and income shifting are clear examples. Compliance and enforcement costs are multiplied in an integrated economy, and shareholder taxation is deeply transformed. The most important distortions caused by CIT on business behaviour, within an international setting, can be summarized as follows: 1.

2.

3.

A traditional distortion regards corporate financial policy, given the tax deductibility of interest expense but not of the return to equity. In general, this tends to favour debt financing relative to equity, especially for international investors, with financial innovations making it increasingly difficult to distinguish between debt and equity instruments. However, non-tax reasons (risk of bankruptcy or financial distress and reduction of agency costs) balance the tax effect. In fact, CIT does not appear to have a great impact on debt/equity discrimination – see De Mooij and Ederveen (2008). Furthermore, the higher tax cost of distributions relative to creating reserves, generates a bias in favour of retaining profits to finance investments. Nevertheless, tax and non-tax benefits, mentioned earlier when discussing the ‘traditional view’ – see above on ‘Shareholder Taxation’ in Section 2, explain why profits are distributed and shares issued as a means of financing corporations (recall also that, under the ‘new view’, the higher tax cost of dividends is simply capitalized in ‘old’ share value, not affecting dividends or investment decisions). The effects of CIT and individual income taxation may encourage or discourage incorporation as a form of organizing business activity. Double taxation of corporate income biases the decision concerning organizational form in favour of acting as an entrepreneur subject to individual income tax. However, if, in addition to the non-tax benefits of incorporation (limited liability, scale economies, ability to attract capital), the source of finance is debt or retained earnings, the advantages of the non-corporate form of conducting business fade. If this is the case, the individual entrepreneur may become a corporate ownermanager under sole proprietorship. The latest empirical evidence on this topic is provided by De Mooij and Nicodème (2008), who report income shifting from the personal to the corporate tax base, with tax base semi-elasticity of around –1.0, using a panel of European countries from 1997 to 2003. The revenue effects of such income shifting are not considerable, since income is subject to either CIT or personal taxation. CIT may increase, decrease or even leave unaltered the cost of capital (depending on depreciation systems, types of finance, investment tax incentives or inflation). This effect, measured by the METR,

162

4.

5.

The Elgar guide to tax systems has an impact on investment, according to the two main different approaches used in empirical literature – Hasset and Hubbard (2002) and Chirinko (2002) – valued as an elasticity of investment to the cost of capital between –0.5 and –1.0. The influence of CIT on marginal investment provided by De Mooij and Ederveen (2008) – assuming, METR: 10 per cent and 50 per cent of the taxable base representing economic rents – is given by a semi-elasticity that lies between –0.3 and –0.6. Therefore, the effect of CIT on the scale of investment is neither large nor insignificant. With regard to FDI location decisions, the AETR is significant if discrete location choices are made. This happens in the case of ‘lumpy’ investments, where investors pledge a large amount of capital or nothing at all – see Devereux and Griffith (1998, 2002) – or when multinationals own mobile firm-specific assets. The meta-analysis performed by De Mooij and Ederveen (2008) suggests a large impact of CIT in this respect – semi-elasticity of –1.2, based on 20 per cent foreign share of capital as reported by Huizinga and Nicodème (2006). Finally, statutory CIT rates exert an important effect on international income shifting undertaken by multinationals – De Mooij and Ederveen (2008) report an effect size of –1.2, based on 60 per cent share of multinationals.

The result of the above list is that a source-based CIT, as applied in most tax systems, is highly distortionary, generating large compliance and enforcement costs and tax-induced managerial underperformance. Corporate income taxation, however, is alive and well and, for better or worse, playing a role in tax systems. There are different ways of reducing the distortionary effects and costs of CITs. One is to continue the process of base-broadening and ratemoderating that started in the second half of the 1980s. This process provides more neutrality, or at least more uniformity, in the tax treatment of business activity that, though not necessarily efficient, may reduce distortions and compliance or enforcement costs. In this respect, a better approximation of tax depreciation to the actual depreciation of assets will be of help, as will be to continue to eliminate tax preferences. With low inflation, it is better to avoid complex inflation adjustments, thus resulting in more revenue that can be used in rate reductions. With regard to the new areas where tax credits are presently applied (R&D, technological innovation, environment, training costs), so scant is the information about the efficiency of such incentives that the fore gone revenues might be more usefully employed in applying lower CIT rates. This field deserves further analysis, as economic arguments support public

The challenges of corporate income taxes 163 aid for these activities. As for the special tax treatment of SMEs, although it may be a lost cause given the political constraints at work, it might also be trimmed in some aspects. For instance, in relation to the favourable tax consideration provided to the intergenerational transmission of family businesses or to the application of lower tax rates without any obligation to create reserves for investment. This broader-based CIT, with a greater scope and limited rates and concessions, may also be less inviting for multinationals’ income-shifting schemes. In any case, a coordinated effort by different administrations with regard to transfer pricing, thin capitalization, interest allocation and tax avoidance in general are needed to prevent profit shifting. Good tax administration performance is a must for all taxes but provides special benefits with CITs in open economies. Of special significance in this respect, particularly in the area of transfer pricing, is the reduction of double taxation risks; it requires a good international system for settling disputes between countries (and an international database on arm’s length prices). A different sort of business risk reduction may be boosted through loss offsetting in inter-border operations, especially in business restructuring situations. With a fairly comprehensive source-based CIT, and with the possibility of applying moderated rates to obtain reasonable revenues, the policy-makers in open economies, competing for FDI, need not to be too unhappy even under the strain of tax competition and profit shifting. Obviously, source-based taxation is not efficient or simple to manage in a globalized world. For instance, it might be difficult to determine the source country of profits with integrated economies, and income needs to be allocated between source countries on the basis of arm’s length pricing. However, residence-based taxation – on an individual or corporate scale – appears to be unfeasible for a variety of reasons, including the fact that shareholders are mobile and can change their place of residence. The main reasons for using corporate taxation in a fiscal system, mentioned in the introduction, are attained by policy-makers through a source-based CIT, with good political support and without much confrontation with the corporate sector (domestic or international). Participation exemption, which globalization has given a leading role on an international scale, may generate more tax competition and profit shifting than the traditional imputation system with deferral (though perhaps not much more), as governments tend to reduce rates to attract investment and multinationals try to enjoy lower rates. Furthermore, with exemption at residence, the parent’s country’s revenue loss is evident (but not necessarily considerable) if the host country’s tax is lower than that of the parent corporation’s country of residence. The territorial system may

164

The Elgar guide to tax systems

therefore be convenient for capital importing countries but not for capital exporters such as the US – which maintains the imputation system – or the UK – which introduced tax exemption for foreign dividends in July 2009 (in the EU, Germany and France adopted the exemption system years ago). Territoriality, however, does not deter repatriation and it encourages multinationals to use the country as the base for international acquisitions, especially if the exemption extends to foreign capital gains. The latter, of course, requires some complication of the exemption system in the form of ‘restricted’ interest deduction rules, lest the funding costs of a foreign acquisition generate deductible expenses with capital gains of sales exempt. These restrictions may also apply to other expenses that support overseas investment (for 2011, the US, an imputation with deferral country, is proposing an increase in this sort of restriction, showing that they are not only needed when exemption is enforced). With regard to the shareholders of the parent corporation, the international exemption system introduces certain difficulties. For instance, domestic closely-held corporations may reincorporate abroad, establishing a home country’s holding corporation. Moreover, domestic exemption of dividends and capital gains would reinforce these challenges and encourage portfolio investment (by corporations or individuals) in countries with low corporate taxes (or even the incorporation of individual entrepreneurs, unless maximum marginal income tax rates are equal, or lower, than that of the CIT). All this explains why, at least for individual shareholders, a total exemption system is not used and a pragmatic approach based on the application of specific rates or partial exemption is leading the way, at least in the European setting. This pragmatic approach contains elements of residence-based taxation and borrows many ideas from Scandinavian dual income taxation. A dual income tax (DIT) may, or may not, become widespread in Europe, but its schedular nature with most capital income (and gains) taxed homogeneously (at lower rates than earned income) in the individual tax is a solid bet in a globalized world. More complete international exchange of information (including dividends) is required to foster residence-based taxation on an individual level. In this respect, fairly recent analytical literature, well represented by Keen and Ligthart (2006), provides ideas on information-sharing incentives. Moreover, better international coordination in tax matters seems a clear corollary to globalization. In the EU setting, coordination in the CIT field is fairly advanced. The completion of the proposals for an EU CCCTB will be an important test in this direction. All the above proposals for action and reform provide a more neutral approach to corporate taxation and tend to reduce compliance costs.

The challenges of corporate income taxes 165 Distortions, however, persist and international enforcement costs will increase. It is time to see whether more radical reforms, using alternatives to CIT, give a better benefit/cost balance, changing what is taxed (and where) as corporate income. Alternative Bases Revisited In Section 2 alternatives to the traditional CIT were presented together with a summary of their scant use in tax systems. We also showed that other combinations of these alternatives had been proposed. The different types of alternatives to CIT are: flow-of-funds corporation taxes (which may take a VAT-type form) or the possibility of bringing an ACE into a CIT, in both cases with the aim of taxing only rents above the normal return on investment; the CBIT (or a ‘pure’ DIT imposed at the corporate level) and the BVT affect all returns irrespective of the source of finance. In addition to these four main classes of taxes, there are different variations or combinations. For instance, ACE may be extended to corporate debt – allowance for corporate capital (ACC) – not admitting the tax deductibility of actual interest expense, as suggested initially by Boadway and Bruce (1984). The distinction between debt and equity for tax purposes is thus eliminated although, if capital gains tax is deferred until realization, without a surcharge, equity is favoured over debt. As mentioned in Section 2, in the individual shareholder tax, ACE may be transformed into ASE–Sørensen (2005a), also with its own variations. Other cases involve explicit two-tier systems such as the Hall and Rabushka (1995) flat tax or the X-tax variant series of the flat tax by Bradford (1986), where a corporate cash-flow tax is coupled with an individual tax on earned income and retirement benefits (pension contributions being deductible). CBIT may be combined, as another possibility, with immediate expensing of investments (ICBIT), only taxing economic rents on an accrual basis at the corporate level. All these alternatives to CIT may eliminate, or reduce, some or all of the distortions generated by CIT and listed in the previous subsection – see Sørensen (2007) or OECD (2007), among other references. On the choice between debt and equity to finance investments, cash-flow taxes or the use of ACE, with the ‘right’ imputation cost of equity, have a neutral impact, as both systems exempt normal return on capital (as interest or as a return to equity). Taxes applied to the whole return on capital regardless of the source of finance (CBIT, ‘pure’ DIT, or BVT) provide financial neutrality, taxing all returns to debt and equity at the corporate level with the same rate. Note, however, that CBIT and BVT highly increase the cost of debt financing for corporations – with risk of bankruptcy – and that, in

166

The Elgar guide to tax systems

an international setting, with shareholders subject to different individual income taxes, neutrality between debt and equity requires either exemption at residence or taxation, at the same rate, of interest, dividends and capital gains. Moreover, for all the alternatives considered, taxation of capital gains on a realization basis implies non-neutrality favouring equity finance via retention of profits unless, for instance, the fairly complex ‘retrospective’ capital gains tax on realization proposed by Auerbach (1991) is used.17 The decision regarding organizational form is not affected by cash-flow taxes, CBIT, BVT or even ACE if applied to corporate and uncorporate business. DIT, however, encourages the incorporation of active owners of small firms who try to shift labour income into lower taxed capital income (double taxation and ASE counterbalance this effect). Also on the debit side, DIT has to split income into labour and capital for individual businesses, which is always a difficult exercise. Scale of investment is unaffected by cash-flow taxes or by CIT with ACE as the marginal tax rate is zero in both cases, so the cost of capital remains unchanged. The rest of the alternatives do not score well in this respect, as they include the normal return to capital in the tax base (except ICBIT). With regard to multinationals’ decisions on location of investment and income shifting by way of transfer pricing, all source-based alternatives to CIT fail to eliminate the distortions generated by the wide international range of statutory or AETRs. These distortions may even worsen if statutory rates increase to maintain revenue in the cash-flow or ACE cases. However, since these source-based alternatives treat debt and equity finance alike, problems related to income shifting through financial transactions are removed. Only a destination-based cash-flow tax – as in the proposal by Auerbach et al. (2011) – is neutral towards location decisions and removes transfer pricing problems, also retaining the nondistortionary characteristics of cash-flow taxes. A destination-based cash-flow tax (R base) exempts exports and taxes imports. If applied like a VAT-type consumption tax (destination-based), its base would include sales to domestic consumers minus purchases from domestic suppliers and labour costs (the latter deduction makes the base unlike VAT, which quantifies value-added as economic rent plus labour income). It is, therefore, a tax on domestic consumption from non-labour income, which cannot be avoided by locating production in other countries. Transfer pricing problems are also eliminated as related export transactions are not taxed and the value of imports from international related parties is irrelevant because it is non-deductible. Nevertheless, destination-based cash-flow taxes suffer from other difficulties and transition problems.18 Among others, one primary difficulty is that it does not

The challenges of corporate income taxes 167 tax either the normal return or the economic rents consumed abroad, although one of the main objectives of corporate taxation is to tax rents that flow to foreign owners. As for ‘combined’ alternatives, their effects on the distortions inherent to a CIT follow the patterns already discussed with regard to the main types of alternatives to CIT. The series of flat tax initiatives include a cash-flow tax at the business level with the same flat rate applied in the individual income tax on labour income. Therefore, its good points and difficulties are those of consumption-based taxation – Zodrow (2006). The case of extending ACE to corporate debt (ACC) – interest expense being non-deductible – boosted by Kleinbard (2007a and 2007b) with his proposal of a business enterprise income tax (BEIT) basically has the merits of ACE. The elimination of any tax difference between debt and equity still leaves the problem that capital gains can be deferred and equity finance is therefore favoured. To deal with this difficulty, Kleinbard introduces a minimum distribution rule for corporations. Individual investors would also be taxed on a presumptive normal return on the investment (at the notional rate applied to corporate equity), which resembles the Dutch ‘deemed yield of asset’ system. Kleinbard’s system has many virtues – normal return is taxed at the individual level and rents at the corporate level – but it may be too complex for practical purposes. Finally, the already mentioned ICBIT – CBIT with immediate expensing of investment – attracts the benefits and costs of its tax of origin, with no deduction of interest expense but with instantaneous depreciation of assets, and might be used as a first step to gradually introduce cash-flow tax on corporations – OECD (2007). The above discussion on the worth of alternative corporate taxes to eliminate CIT distortions, shows that there are plenty of possibilities to choose from, each with its own pros and cons. Other difficulties arise, however, when moving from theory to possible implementation. Important transition problems are found with most of the alternatives, which have been widely discussed in the literature – OECD (2007). For example, cash-flow taxes impose a higher burden on existing capital that is inequitable and distorts competition, unless this ‘old’ capital obtains access to the new expense treatment. For the sake of brevity, we shall not be considering these well-known problems here. Loss of revenue is certain (unless rates rise, increasing the problems related to the location of FDI and profit shifting) with all the alternatives that relinquish the taxation of normal profits (mainly cash-flow systems and implementing ACE). However, the magnitude of such a loss is not clear, as revenue collections on the normal return to capital are reported by several studies as low – Gordon et al. (2004), Becker and Fuest (2005)

168

The Elgar guide to tax systems

for Germany, or Sørensen (2007). This topic requires further research on a per country basis if any of the alternatives to CIT can be even considered by governments. Policy-makers never want to lose revenue, but especially not during an economic crisis. When the current recession abates, reduction of public deficits and debt will mean increased taxation (and spending cuts). One may hope that greater tax revenue might also be used to implement ‘better’ taxes if the related collection losses are not high. To establish some boundaries for our discussion of the difficulties involved in applying any of the alternative corporate bases, and in order to help select the most appropriate, two pivotal factors must be considered. One is the international legal compatibility of fundamental corporate income tax reform. The other has to do with the possibilities of the agents involved accepting what would be a ‘radical’ tax reform. International legal compatibility In open economies, with important flows of investment or income across borders and foreign ownerships of assets, corporate tax reform has to seriously consider its international legal consequences. Even with the widespread use of participation exemption, some countries (such as the US) with huge in- and out-flows of capital, stick to the imputation system. Moreover, exemption is not the rule for portfolio investments and individual investors. Therefore, tax treaties or domestic legislation are required to prevent double taxation. If an isolated country, or group of countries, reforms its corporate tax base, this could give rise to a difficult situation, as other countries’ interpretation of treaties or internal legislation may rely on the legal definition of their CITs, which will more than likely be that of an income tax on the return to equity based on the profit and loss accounting balance to which fiscal adjustments are made. If this is so, some of the alternatives to CIT may not be creditable against these other countries’ taxes, certainly hindering capital and income flows. Zodrow (2006) reports that the US Internal Revenue Service may not be prepared to consider a cash-flow tax as a creditable tax for US multinationals, though it allows the creditability of a portion of the Italian IRAP (recall from Section 2 that it is an origin-based tax on the value-added of enterprises, net income type, close to BVT). The creditable portion of the IRAP in the US is, in general, the value-added base less interest payments and labour costs (both non-deductible for IRAP). This confirms what we said earlier about the appropriate definition of a CIT as an income tax. Cash-flow taxes (VAT-like or not) are not considered income taxes. All other alternatives to CIT that disallow the deductibility of interest expense or labour costs may have curtailed creditability. A CIT cum ACE, as in

The challenges of corporate income taxes 169 Belgium, should have no problems in relation to international legal compatibility even if it only taxes rents, as it is an income tax on part of the return to equity. The BEIT, CBIT or BVT may be doubtful candidates for total creditability against other countries’ taxes. Acceptance by other agents The agents involved in any tax reform are numerous and have their own interests that differ considerably among themselves and from those of academic public economists. We mentioned this in the introduction as being an important factor for this chapter. The agents with influence on tax reforms obviously include policy-makers and tax officials, mainly interested in revenue and administration and enforcement issues, though they may also appreciate the importance of the elimination of tax distortions for an economy. With regard to the ‘business community’, lower statutory or AETRs and the limitation of compliance costs are objectives to accomplish. Moreover, individuals responsible for tax matters in corporations, especially those who sign returns and incur legal responsibility (the CEO, board members, chief financial officers and so on) want to clearly understand the tax in question. In this regard, an important characteristic, which can be extended to most of the other agents involved, is that the training of these individuals in economics may not be very sophisticated but based more on accrual accounting, among other disciplines. The same applies to tax auditors and public accountants, or tax advisors, who have a direct or indirect say on tax reform matters. Different economic sectors, primarily the financial sector, may be very interested in a change of tax base for corporate taxation. Finally, Members of Parliament, politicians in general or trade unionists, with very different origins and training, will also be involved in tax reforms. A significant amount of information is needed for these agents to understand and accept a fundamental reform in corporate taxation (or in any other tax). This may be a messy job, but it is a necessary precondition for any radical reform. This is also of great help in selecting the appropriate alternative to CIT on which ‘educational’ efforts may concentrate during the coming years.19 Let’s start by discussing an important aspect of cash-flow taxes. They are normally seen as simpler than income taxes. This is true in so far as CIT problems with regard to depreciation or capital gains, the matching of revenue and expense or inflation accounting, for instance, simply vanish with the flow-of-funds concept. There are further difficulties, however. Accounting for business purposes is based internationally on accrual and not on cash-flows. This means that either a firm keeps two sets of accounts (increasing compliance costs) or transforms accrual accounting to, for

170

The Elgar guide to tax systems

instance, the R base of a flow-of-fund corporation tax, no easy task (of course, the S base would be far simpler to implement, although it may seem odd to many of the agents involved). International Accounting Standard 7 regulates cash-flow statements (C-FS), the application of which varies among countries. In the Spanish Accounting Plan, for instance, this statement is not mandatory for SMEs.20 C-FS classify cash-flows by operating, investing and financing activities. This activity-based classification is open to the criteria of businesses. Furthermore, a single transaction may include cash-flows that are classified differently. For example, when the cash repayment of a loan includes both interest and capital, the interest may be classified as an operating activity and capital as a financing activity. Additionally, two methods can be used to report cash-flow from operating activities: the direct method (major classes of gross cash receipts and payments are disclosed), which is the best, although fairly difficult to apply with the information in the accounts, or the indirect method, whereby profit or loss is adjusted for the effects of non-cash transactions, deferral or accruals of past or future operating cash receipts or payments and income or expense items associated with investing or financing cash-flows. In the Spanish case the indirect method is mandatory, possibly because of the difficulties involved in applying the direct method. Nevertheless, though C-FS show the net variation in cash-flows, the indirect method does not provide the information on receipts and payments of operating activities needed for a (R) cash-flow corporation base and other difficulties arise for investing activities. This requires further transformation of the C-FS mainly for operating activities through methodologies (not highly complex but with a fair amount of difficulty) based on information from the current profit or loss account and differences between the balance sheets of two consecutive years. The end result is an estimate of taxable income as receipts minus payments for operating and investing activities.21 Nevertheless, the main disadvantage of the cash-flow alternative is not just this complexity that increases compliance costs, or the need to choose a methodology and to describe it in written legal terms, plus the resulting tax audit difficulties. The big question mark is whether the lack of familiarity with and understanding of flow-of-funds concepts among tax officials and auditors, public accountants, individuals responsible in firms for tax issues or returns, and many other agents involved in reforming the corporate tax base, may be a definitive reason for not accepting the reform. I believe this to be the case and doubt that the situation will change even with a major information effort. We are all aware of the accounting and tax differences in CITs and the problems they generate. Both approaches

The challenges of corporate income taxes 171 – accounting and tax – however, aim to estimate profits. Cash-flow taxation requires a detachment from accrual accounting to compute rents. This simplifies the tax but may result in its non-acceptance. Different acceptability problems arise with regard to other than CIT bases. Consider the types of alternative taxes that disallow interest expense deductibility (CBIT, BVT and, to a lesser extent, BEIT). Apart from the considerable increase in the costs of debt financing, which creates difficult problems for firms, the financial sector’s opposition to these taxes is obvious. Other fundamental reforms based on combining taxes, such as proposals for a flat tax or a BEIT, require parallel reforms in individual income tax which make their general acceptance more difficult (and possibly that of the policy-maker, who may be in favour of changes in one tax and not in the other). A ‘pure’ DIT imposed at the corporate level, with interest withheld by the paying corporation – see Cnossen (2000) – would face the bitter German experience of the early 1990s trying to withhold interest flowing abroad. With current VAT (consumption sort) in all OECD tax systems, except the US, new VAT taxes (consumption or net income type) may be difficult to accept, even if labour costs are deducted or their taxation compensated in some way, simply because a similar tax is already in place. High increases in current VAT rates, to substitute the revenue of a repealed CIT, will not be acceptable for distributive and other economic reasons, or because the roles of a CIT in a tax system will not be fulfilled (the possibility of a moderate increase in ordinary VAT rates to enable CIT rate reductions is another issue). All the above leaves us with the alternative of using ACE within a traditional CIT.22 This alternative is internationally compatible (it is an income tax on the return to equity, though normal profits are exempt via the adjustment provided by the allowance) and it is easier to understand and accept by many of the agents involved in its selection, approval and implementation, simply because it is a very close alternative to the classical CIT. This does not mean that a CIT-ACE has to be the most suitable choice for everyone, but it is certainly the most convenient according to the two pivotal factors selected for our discussion. I hasten to point out that, obviously, ACE does not convert the CIT into a perfect tax. As already discussed, CIT-ACE scores high to obtain neutrality between debt and equity but, with capital gains taxed at realization, this tax favours equity finance and the creation of reserves. If ACE is not applied to unincorporated business, the decision on organizational form is affected (the use of ASE is administratively difficult since it requires a registry of the acquisition cost of shares).23 The level of investment may be unaffected with ACE, but, being a source-based tax, distortions with regard to the

172

The Elgar guide to tax systems

location of investment or income shifting via transfer pricing are not solved and may worsen if rates increase to maintain revenue (as ACE erodes the tax base). Moreover, other traditional problems related to CIT, mainly regarding the matching of income and expense, remain with ACE, except in the field of depreciation. One merit of ACE is investment neutrality. Applied asset depreciation systems are irrelevant as, if businesses use a more accelerated than a truly economic depreciation, it will be offset by a reduction in future ACE of equal present value. The allowance’s base computed either on the basis of the written-down value of assets for tax purposes or as the equity of the firm (capital and reserves)24 takes care of the differences between actual depreciation and ‘true’ economic depreciation. If, however, the equity method of computation is chosen and depreciation for tax purposes is accelerated with no effect on the accounting written-down value of assets (which is the usual situation with tax benefits), then equity has to be adjusted (negatively) by the respective base of the liability for (taxable) temporary differences: deferred tax (which is equal to the written-down value of the assets for accounting purposes at the beginning of each tax period). An example may help to explain this. Imagine that a corporation, with an accounting profit for year 1 of 2000 units, has invested 1000 in an asset working since the first day of the year, with an accounting life of four years (25 per cent accounting depreciation rate). The depreciation charged to profit and loss is, therefore, 250. This asset enjoys free depreciation for tax purposes. The CIT rate is 30 per cent and there are no other permanent or temporary differences. Assuming the total tax expensing of the asset in year 1, the taxable base will be 2000 – 750 = 1250. The current tax payable is 0.30 × 1250 = 375 and tax expense (current tax plus deferred tax) is 0.30 x 2000 = 600. It is irrelevant whether profits after tax expense (2000 – 600 = 1400) are distributed or retained. Assume, for simplicity, that the total amount is paid out as dividends, as in this case, if nothing else such as new shares issued or acquired (see note 24) has to be considered, equity value (capital and reserves) is not altered. The corporation, however, enjoys a higher level of liquid assets amounting to 225 (the tax not currently paid, which is in fact a non-interest loan from the government repayable in the following three years) and the allowance’s base has to incorporate this fact. This is accomplished via the deduction of the base of the related liability for temporary differences (deferred tax liability), so that the value of the allowance is reduced to compensate for the immediate depreciation for tax purposes that has not been accounted for. If the imputed rate of interest is, for instance, 5 per cent, equal to the

The challenges of corporate income taxes 173 discount rate, the net change in the present value of taxes paid by the corporation will be: 2 0.30 3 750 1

0.30 [ 250 1 0.05 3 500 ] 0.30 [ 250 1 0.05 3 750 ] 1 (1.05) 2 1.05 1

0.30 [ 250 1 0.05 3 250 ] 50 (1.05) 3

(750, 500, 250 are the values of the corresponding bases of the liabilities for temporary differences in years 2, 3 and 4), so that the initial tax benefit is offset in present value terms by the higher current tax of the following years due to the adjusted ACE and payment of the deferred tax (reversion of the temporary difference). In general, ACE compensates for the delay of a depreciation allowance instead of immediate expensing. This makes CIT-ACE similar to a (R+F) cash-flow tax. But, if a specific tax benefit provides, for instance, immediate depreciation, without accounting effects, the ACE base should be reduced, as there is no delay in depreciation for tax purposes. The selection of the suitable notional rate of return to equity for ACE is also a central element of this alternative to traditional CIT. After Bond and Devereux (1995), it is commonly accepted that the most convenient rate would be a risk-free rate of interest expressed in nominal terms (so that inflation may be disregarded). The risk-free nominal rate of interest is approximated by the average interest rates of short-term government bonds. In this respect, care should be taken to provide moving averages of past government bond issues, as a corporation’s financial years do not always coincide with the calendar year (recall from Section 2 that the Belgian ‘notional interest deduction’ is only applicable to corporations where the economic year is the calendar year, possibly for this reason). Furthermore, in order to obtain neutrality, full loss compensation has to be provided (so that the firm and its shareholders receive a safe cash-flow via the benefits of the allowance). This means admitting the indefinite carry-forward of losses, with the (notional) interest rate added – so that ACE is not reduced in present value terms – and that the shareholders will receive a tax credit for unutilized ACE in the case of bankruptcy. This represents a real problem for CIT-ACE as, in practice, tax legislations do not allow such a full loss offset, largely for revenue and administrative reasons. In spite of all problems of the ACE alternative, my bet for the future is on CIT-ACE, mainly because it scores high in the two pivotal factors selected for the above discussion: international legal compatibility and the possibility of acceptance by the agents involved in any major tax reform. Research efforts directed towards technical aspects (for instance, possible

174

The Elgar guide to tax systems

losses of revenue, definition of the allowance base, the best notional rate of return, the problem of loss compensation, etc.) and also intermediate steps in the direction of CIT base-broadening and rate reduction, will further clarify this debate.

5

CONCLUSIONS

Traditional CIT, with new features after globalization, will be ‘well and alive’ for a good number of years fulfilling its roles in tax systems. No finance minister will propose major reforms risking revenue that, on average for OECD countries, has amounted to around 3.5 per cent of the GDP (and around 10 per cent of total revenue) especially after the present, crisis-induced huge public deficits found in all economies. I am afraid that governments will continue to scramble for national gains through the CIT, and the tax system in general, for a long time to come. More base broadening, and perhaps rate reduction, will possibly be undertaken in the future in order to better control the distortions derived from the CIT. International coordination of capital income taxation poses a major question for the years to come, although it would appear to be inevitable if economic globalization continues to expand. For the more distant future, public economists have produced a large number of proposals for alternatives to the CIT. The option that probably scores better as a policy prescription is the implementation of ACE in a traditional CIT, although this will clearly not be the best choice for everyone. Further research on different aspects of the alternatives to CIT (primarily on the conditions for the success of a fundamental reform) is required if a professional consensus is ever to be attained.

NOTES * 1.

2.

3.

I wish to thank Julio López Laborda, Jorge Martinez-Vazquez and José Felix Sanz for helpful comments on previous drafts. In an international setting, if governments are not able to capture rents, Keen and  Piekkola (1997) suggest that, to obtain global optimality, a source-based corporate tax may be a good indirect form of taxing pure profits by a capital-importing country. The ideas on cash-flow taxes maintained by the Meade Committee (1978) have a precedent in Cary Brown (1948) and were discussed at that time, and previously, by others. Since about 1980, much literature has been produced on cash-flow taxation, with the latest major effort being The Mirrlees Review, Reforming the Tax System for the 21st Century (Mirrlees et al. 2010/11, two volumes). In the following paragraphs the main references are: European Tax Handbook (2008 and 2009), OECD (2007) and Bordignon et al. (2001). For the last sub-sections under

The challenges of corporate income taxes 175

4. 5.

6. 7.

8.

9.

10.

11.

12.

‘Shareholder Taxation’ in Section 2 we basically used the European Tax Handbook as a reference. Diario Oficial de la Federación (DOF), Mexico, 1 October 2007. For this information and other references, my thanks to Lourdes Jerez and her doctoral thesis on Cash-Flow Corporate Taxes, University of Extremadura (Spain). The text contains a simple explanation based on the experience of any multinational. An in-depth analysis can be found in Markusen (2002). Furthermore, tax competition (to an extent in reverse) can also act through tax exporting to foreigners, as the evidence in Huizinga and Nicodème (2006) appears to show, as they find a positive relationship between corporate tax burdens and foreign ownership. It is also worth mentioning that Sinn (1991), among others, has extended the ‘new view’ to a dynamic setting, using an analysis that is referred to as the ‘nucleus theory of the firm’. Mainly on a national scale, the effects of shareholder taxation extend to the choice of an individual or corporate form of business and the balanced treatment of labour and capital income in closely-held corporations. These issues might be best approached as part of an analysis of personal income. With losses in the recipient corporation, integration of the gross dividend received from other resident corporations (or capital gains produced by the transmission of shares in resident companies, in which case the same system is followed) in taxable income reduces the loss. The unused deduction has a seven-year carry-forward period providing the reduced loss is compensated during that time. Imagine that a corporation has a loss of 100 from its activity and 100 from dividends (or capital gains) from its participation in a Spanish subsidiary. With this credit against tax liability system, the loss of 100 becomes zero when the 100 units of dividends (or capital gains) are included in taxable income. In the year when the corporation obtains a profit of 100 (within the seven-year period), it can use the previously unused credit but the compensation of the initial loss has already been produced and its tax effect vanishes. On the other hand, with participation exemption, if a corporation has a loss of 100 from its activity and 100 of dividends (or capital gains) from its participation in a subsidiary, when it obtains a profit of 100 in subsequent years, it covers the loss with a tax saving equal to tax rate x 100. On the other hand, with participation of at least 5 per cent in the capital of the subsidiary, the reinvestment of capital gains tax credit allows a deduction from tax liability of 12 per cent of the gain, in the part that does not correspond to the reserves that are the source of the reduction discussed in the text, if all the proceeds are reinvested. This means that part of the capital gain may be exempted and the rest is taxed at an 18 per cent rate (considering the standard Spanish CIT rate of 30 per cent). Substantial participation in Italy is defined for this purpose as follows: (1) holding more than 5 per cent of the capital – or 2 per cent of the votes – in listed corporations; (2) holding more than 25 per cent of the capital – or 20 per cent of the votes – in unlisted corporations. In Spain, for example, the special regime for SMEs affects firms with a turnover of less than 8 million euros in the previous fiscal year. This is higher than the figure established in business legislation for presenting an abridged balance sheet and not being obliged to audit the annual accounts (4.75 million euros), so the definition of an SME is quite broad. It is very unclear whether full harmonization of the CIT may provide an optimal tax setting with sizeable growth or welfare gains for countries. For example, Brøchner et al. (2006), using the OECD TAX model to simulate the effects of harmonizing corporate tax bases and rates in EU(25), obtain a total GDP increase in the Union of 0.4 per cent and a rise in the average welfare of the representative consumer in each country of only 0.1 per cent of the GDP. These low gains are unequally distributed among countries; some have considerable gains and others losses in GDP and welfare. These results are in line with those obtained by Sørensen (2004a), and do not seem to convince governments to lose sovereignty.

176 13. 14.

15. 16. 17. 18.

19. 20. 21. 22. 23.

24.

The Elgar guide to tax systems Case C-35/98 Verkooijen (2000) ECR I-4071. This jurisprudence continued, for example, in the Lenz case (C-315/02) or the Manninen ruling (C-319/02). The Spanish tax consolidation regime requires that the dominant corporation has a direct or indirect participation of at least 75 per cent in the capital of other corporations in the tax group, maintained throughout the tax period. All corporations in the group have to be residents of Spain. Case C-446-03 Marks & Spencer, Ruling of 13 December 2005; Case C-418/07 Papillon, Ruling of 27 November 2008. With regard to R&D tax credits, accelerated depreciation allowances or financial concessions provided by governments, the results of the available review of the literature are inconclusive – see OECD (2002) and EC (2003b). Auerbach’s proposal borrows from Vickrey’s ideas on cumulative averaging – Vickrey (1939) and (1947). As pointed out by Sørensen (2007), a special transition problem has to do with exchange rates, or relative price levels, given the exemption of exports and taxation of imports. With flexible exchange rates this would lead to an appreciation of the currency, which would generate gains or losses to residents with net liabilities or claims denominated in foreign currency. In a currency area such as the euro, the domestic relative price level would increase to produce the revaluation effect. In general, I now believe, more than I did 20 years ago – Albi (1985) – that economists acting in an advisory role and providing tax policy prescriptions, must consider many more aspects of the policy implementation process than they usually take into account. An SME is defined for this purpose as an enterprise fulfilling two of the following three requisites: assets ≤ 2 850 000 euros, turnover ≤ 5 700 000 euros or number of employees ≤ 50. The aforementioned doctoral dissertation by Lourdes Jerez (Universidad de Extremadura, Spain) – note 5 – provides one methodology for adjustments. Two recent proposals in favour of ACE, combined with DIT at an individual level, are: Keuschnigg and Dietz (2007) and Griffith et al. (2011). This requirement is clear in order to administratively control ASE, at least if the number of taxpayers declaring dividends is large. In Spain, for 2006 (last year with numerical information), taxpayers declaring dividends were 3 056 308 – 17.13 per cent of the total number of returns: 17 840 783. With a negative adjustment for the net holding of new shares in other corporations to avoid domestic ‘double’ counting, as the acquisition cost of these shares will be included in the issuer’s equity. The same adjustment should be made for foreign shares so that (with exemption), if return on investment is not taxed, foreign investment does not erode the domestic base via the allowance.

REFERENCES Agúndez, A. (2006), ‘The Delineation and Apportionment of an EU Consolidated Tax Base for Multi-jurisdictional Corporate Income Taxation: A Review of Issues and Options’, European Commission, Working Paper No. 9. Albi, E. (1985), ‘Economic Advice and Tax Policy’, Western Tax Review, 6(1), 36–47. Albi, E., E. Corona and R. Paredes (1997), A Corporate Tax as a Possible Fifth Own Community Resource, XIX General Directorate, European Commission, Brussels. Alt, J., I. Preston and L. Sibieta (2011), ‘The Political Economy of Tax Policy’, in The Mirrlees  Review, Reforming the Tax Systems for the 21st Century, Vol. 2, Oxford University Press. Altshuler, R. and H. Grubert (2003), ‘Repatriation Taxes, Repatriation Strategies and Multinational Financial Policy’, Journal of Public Economics, 87(1), 73–107. Angelov, G. (2006), ‘Zero Tax on Reinvested Profit – the Example of Estonia’, Institute for

The challenges of corporate income taxes 177 Market Economics, Bulgaria, available at: http://www.irefeurope.org/en/sites/default/files/ Estonia_G_Angelov.pdf; accessed 19 April 2011. Auerbach, A. (1991), ‘Retrospective Capital Gains Taxation’, American Economic Review, 81(1), 167–78. Auerbach, A. (2002), ‘Taxation and Corporate Financial Policy’, in A. Auerbach and M. Feldstein (eds), Handbook of Public Economics, vol. 3, Amsterdam: North-Holland. Auerbach, A. (2005), ‘Who Bears the Corporate Tax? A Review of What We Know’, NBER Working Paper No. 11686. Auerbach, A. (2006), ‘The Future of Capital Income Taxation’, Fiscal Studies, 27(4), 399–420. Auerbach, A. (2007), ‘Why Have Corporate Tax Rates Declined? Another Look’, CESifo Economic Studies, 53(2), 153–71. Auerbach, A. and D. Bradford (2004), ‘Generalized Cash Flow Taxation’, Journal of Public Economics, 88(5), 957–80. Auerbach, A. and K. Hasset (2007), ‘The 2003 Dividend Tax Cuts and the Value of the Firm: An Event Study’, in A. Auerbach, J. Hines and T. Slemrod (eds) Taxing Corporate Income in the 21st Century, Oxford: OUP. Auerbach, A. and J.M. Poterba (1988), ‘Why Have Corporate Tax Revenues Declined?’, in E. Helpmann et al. (eds), Economic Effects of the Government Budget, Cambridge, MA: MIT Press, pp. 33–49. Auerbach, A., M.P. Devereux and H. Simpson (2011), ‘Taxing Corporate Income’, in The Mirrlees Review, Reforming the Tax System for the 21st Century, vol. 2, Oxford: Oxford University Press for IFS. Baldwin, R.E. and P. Krugman (2004), ‘Agglomeration, Integration and Tax Harmonization’, European Economic Review, 48(1), 1–23. Becker, J. and C. Fuest (2005), ‘Does Germany Collect Revenue from Taxing the Normal Return to Capital?’, Fiscal Studies, 26(4). Becker, J. and C. Fuest (2007a), ‘Corporate Tax Policy and International Mergers and Acquisitions – Is the Tax Exemption System Superior?’, CESifo Working Paper No. 1884. Becker, J. and C. Fuest (2007b), ‘Internationalization and Business Tax Revenue – Evidence from Germany’, ETPF Meeting, 23 April. Bertelsman, E.J. and R.M. Beetsma (2003), ‘Why Pay More? Corporate Tax Avoidance Through Transfer Pricing in OECD Countries’, Journal of Public Economics, 87(9–10), 2225–52. Bird, R. and J. Mintz (2001), ‘Tax Assignment in Canada: A Modest Proposal’, in H. Lazar (ed.), The State of Federation 1999/2000, Institute of Intergovernmental Relations, Queen’s University, Kingston, Ontario, pp. 262–92. Blumenthal, M. and J. Slemrod (1995), ‘The Compliance Costs of Taxing Foreign-Source Income: Its Magnitude, Determinants and Policy Implications’, International Tax and Public Finance, 2(1), 37–54. Brill, A. and K.A. Hasset (2007), ‘Revenue-Maximizing Corporate Income Taxes: The Laffer Curve in OECD Countries’, AEI, Working Paper No. 137, pp. 1–19. Boadway, R. and N. Bruce (1984), ‘A General Proposition on the Design of a Neutral Business Tax’, Journal of Public Economics, 24(2), 231–9. Bond, S. and M.P. Devereux (1995), ‘On the Design of a Neutral Business Tax Under Uncertainty’, Journal of Public Economics, 58(1), 57–71. Bond, S. and M.P. Devereux (2003), ‘Generalised R-based and S-based Taxes Under Uncertainty’, Journal of Public Economics, 87(5/6), 1291–1311. Bordignon, M., S. Giannini and P. Panteghini (2001), ‘Reforming Business Taxation: Lessons from Italy?’, International Tax and Public Finance, 8(2), 191–210. Bradford, D.F. (1986), Untangling the Income Tax, Cambridge, MA: Harvard University Press. Brøchner, J., J. Jensen, P. Svensson and P.B. Sørensen (2006), ‘The Dilemmas of Tax Coordination in the Enlarged EU’, CESifo Working Paper No. 1859. Brown, C.E. (1948), ‘Business-Income Taxation and Investment Incentives’, in Lloyd A.

178

The Elgar guide to tax systems

Metzler (ed.), Income, Employment and Public Policy, Essays in Honor of Alvin Hasen, New York: W.W. Norton. Chirinko, R.S. (2002), ‘Corporate Taxation, Capital Formation and the Substitution Elasticity between Labour and Capital’, CESIfo Working Paper No. 707. Clausing, K.A. (2007), ‘Corporate Tax Revenues in OECD Countries’, International Tax and Public Finance, 14(2), 115–33. Cnossen, S. (1996), ‘Company Taxes in the European Union: Criteria and Options for Reform’, Fiscal Studies, 17(4), 67–97. Cnossen, S. (2000), ‘Taxing Capital Income in the Nordic Countries: A Model for the European Union?’, in S. Cnossen (ed.), Taxing Capital Income in the European Union – Issues and Options for Reform, Oxford: Oxford University Press. Cnossen, S. (2002), ‘Tax Policy in the European Union. A Review of Issues and Options’, CESifo Working Paper No. 758. Crawford, C. and J. Freedman (2011), ‘Small Business Taxation’, in The Mirrlees Review: Reforming the Tax Systems for the 21st Century, Vol. 2, Oxford: Oxford University Press for IFS. De Mooij, R. and S. Ederveen (2003), ‘Taxation of Foreign Direct Investment: A Synthesis of Empirical Research’, International Tax and Public Finance, 10(6), 673–93. De Mooij, R. and S. Ederveen (2008), ‘Tax Margins of Business Behaviour’, paper presented at the 64th Congress of the IIPF. De Mooij, R. and G. Nicodème (2008), ‘Corporate Tax Policy and Incorporation in the EU’, International Tax and Public Finance, 15(4), 478–526. Desai, M. and J. Hines (2003), ‘Evaluating International Tax Reform’, National Tax Journal, 56(3), 487–502. Desai, M., C. Foley and J. Hines (2004), ‘Old Rules and New Realities: Corporate Tax Policy in a Global Setting’, National Tax Journal, 57(4), 937–60. Devereux, M.P. (1990), ‘Capital Export Neutrality, Capital Import Neutrality and Capital Ownership Neutrality and all that’, IFS Working Paper, London, Institute for Fiscal Studies, June. Devereux, M.P. (2004), ‘Debating Proposed Reforms of the Taxation of Corporate Income in the EU’, International Tax and Public Finance, 11(1), 71–89. Devereux, M.P. (2006), ‘Developments in the Taxation of Corporate Profit in the OECD since 1965: Rates, Bases and Revenues’, Conference of the Alliance for Competitive Taxation, 2 June. Devereux, M.P. (2007), ‘The Impact of Taxation on the Location of Capital, Firms and Profit: A Survey of Empirical Evidence’, Working Paper No. 02, Oxford University Centre for Business Taxation. Devereux, M.P. and R. Griffith (1998), ‘Taxes and the Location of Production; Evidence from a Panel of US Multinationals’, Journal of Public Economics, 68(3), 335–67. Devereux, M.P. and R. Griffith (2002), ‘Evaluating Tax Policy for Location Decisions’, International Tax and Public Finance, 10(2), 107–26. Devereux, M.P. and S. Loretz (2008), ‘The Effects of EU Formula Apportionment on Corporate Tax Revenues’, Fiscal Studies, 29(1), 1–33. Devereux, M.P. and P.B. Sørensen (2006), ‘The Corporate Income Tax: International Trends and Options for Fundamental Reform’, European Economy, Economic Papers No. 264, EC. Devereux, M.P., R. Griffith and A. Klemm (2002), ‘Corporate Income Tax Reforms and International Tax Competition’, Economic Policy, 17(35), 451–95. Devereux, M.P., R. Griffith and A. Klemm (2004), ‘Why has the UK Corporation Tax Raised so Much Revenue?’, Fiscal Studies, 25(4), 367–88. Devereux, M.P., B. Lockwood and M. Reodano (2006), ‘Do Countries Compete Over  Corporate Tax Rates?’, Working Paper, Oxford University Centre for Business Taxation. Diamond, P.A. and J.A. Mirrlees (1971a), ‘Optimal Taxation and Public Production, Part I: Production Efficiency’, American Economic Review, 61(1), 8–27.

The challenges of corporate income taxes 179 Diamond, P.A. and J.A. Mirrlees (1971b), ‘Optimal Taxation and Public Production, Part II: Tax Rules’, American Economic Review, 61(3), 261–78. Dischinger, M. and N. Riedel (2008), ‘Corporate Taxes and the Location of Intangible Assets within Multinational Firms’, Discussion Paper No. 2008-11, Department of Economics, University of Munich, available at: http://epub.ub.uni-muenchen.de; accessed 21 April 2011. European Commission (1994), ‘Recommendation of 7-12-94 on the Transfer of Small and Medium-sized Enterprises’ (94/1069/EC). European Commission (1997), ‘EU Code of Conduct for Business Taxation’, available at: http://www.ec.europa.eu/taxation_customs/taxation/company_tax/harmful_tax_practices/index_en.htm; accessed 21 April 2011. European Commission (2001), ‘Towards an Internal Market Without Tax Obstacles: A Strategy for Providing Companies with a Consolidated Corporate Tax Base for their EU-wide Activities’, Communication from the Commission to the Council, the European Parliament and the Economic and Social Committee, COM (2001), 582 final, 23 October (includes: Executive Summary of the Commission Services Study on ‘Company Taxation in the Internal Market’ [SEC (2001) 1681]). European Commission (2002), ‘Company Taxation in the Internal Market’, Commission Staff Working Paper, COM (2001) 582 final, Luxembourg – published previously as SEC (2001) 1681 23 October, available at: http://europa.eu.int/comm/taxation_customs/ resources/documents/company_tax_study_en.pdf; accessed 24 April 2011. European Commission (2003a), ‘The Application of International Accounting Standards (IAS) in 2005 and the Implication for the Introduction of a Consolidated Tax Base for Companies’ EU-wide Activities’, available at: http://europa.eu.int/comm/taxation_ customs/taxation/consultations/tax/article_385_en.htm; accessed 24 April 2011. European Commission (2003b), ‘Improving the Effectiveness of Fiscal Measures to Stimulate Private Investment Research’, document prepared by an independent expert working group for the EC. European Commission (2004), Commission Non-Papers to Informal Ecofin Council, 10–11 September, 2004: ‘Home State Taxation for Small and Medium-Sized Enterprises’, ‘A Common Consolidated EU Corporate Tax Base’, 7 July, available at: http://europa.eu.int/ comm/taxation_customs/index_eu.htm; accessed 21 April 2011. European Commission (2005), ‘Home State Taxation’ (COM/05/702). European Commission (2006), ‘Consolidated Corporate Tax Base’ (COM/06/157 final), 5 April. European Tax Handbook (2008, 2009), International Bureau of Fiscal Documentation. Eurostat (2008), Taxation Trends in the EU, Brussels. Feld, L. and J. Heckemeyer (2008), ‘FDI and Taxation. A Meta-Study’, paper presented at the 64th Congress of the International Institute of Public Finance. Fuest, C., T. Hemmelgarn and F. Ramb (2007), ‘How Would the Introduction of an EUwide Formula Apportionment Affect the Distribution and Size of the Corporate Tax Base? An Analysis Based on German Multinationals’, International Tax and Public Finance, 14(5), 605–26 (erratum: 627–9). Funke, M. (2005), ‘Taxation, Growth and Welfare: Dynamic Effects of Estonia’s 2000 Income Tax’, Hamburg University. Gammie, M. (2003), ‘The Role of the European Court of Justice in the Development of Direct Taxation in the EU’, Bulletin for International Fiscal Documentation, 57(3), 86–98. Garrett, G. and D. Mitchell (2001), ‘Globalization, Government Spending and Taxation in the OECD’, European Journal of Political Research, 39(2), 105–17. Gordon, R.H. (1986), ‘Taxation of Investment and Savings in a World Economy’, American Economic Review, 76(5), 1086–102. Gordon, R.H. and J.R. Hines Jr. (2002), ‘International Taxation’, in A. Auerbach and M. Feldstein (eds), Handbook of Public Economics, vol. 4, Amsterdam: North-Holland. Gordon, R.H., L. Kalambokidis and J. Slemrod (2004), ‘Do We Now Collect any Revenue from Taxing Capital Income?’, Journal of Public Economics, 88(5), 989–1009.

180

The Elgar guide to tax systems

Gordon, R.H. and J. Slemrod (2000), ‘Are Real Responses to Taxes Simply Income Shifting Between Corporate and Personal Tax Bases?’, in J. Slemrod (ed.), Does Atlas Shrug?, Cambridge, MA: Harvard University Press. Gordon, R.H. and J.D. Wilson (1986), ‘An Examination of Multijurisdictional Corporate Income Taxation under Formula Apportionment’, Econometrica, 54(6), 1357–73. Griffith, R. and Klemm, A. (2004), ‘What Has Been the Tax Competition Experience of the Last 20 Years?’, Tax Notes International, 34(13), 1299–1316. Griffith, R., J. Hines and P.B. Sørensen (2011), ‘International Capital Taxation’, in The Mirrlees Review, Reforming the Tax Systems for the 21st Century, vol. 2, Oxford: Oxford University Press for IFS. Grubert, H. (1998), ‘Taxes and the Division of Foreign Operating Income among Royalties, Interest, Dividends and Retained Earnings’, Journal of Public Economics, 68(2), 269–90. Grubert, H. (2001), ‘Tax Planning by Companies and Tax Competition by Governments: Is there Evidence of Changes in Behavior?’, in R. Hines, Jr. (ed.) International Taxation and Multinational Activity, Chicago: University of Chicago Press, pp. 113–39. Grubert, H. and J. Mutti (1995), ‘Taxing Multinationals in a World with Portfolio Flows and R&D: Is Capital Export Neutrality Obsolete?’, International Tax and Public Finance, 2(3). Hall, R.E and A. Rabushka (1995), The Flat Tax, 2nd edition, New York: McGraw-Hill. Harberger, A. (1962), ‘The Incidence of the Corporation Income Tax’, Journal of Political Economy, 70(3), 215–40. Hasset, K. and R.G. Hubbard (2002), ‘Tax Policy and Business Investment’, in A. Auerbach and M. Feldstein (eds), Handbook of Public Economics, vol. 3, Amsterdam: Elsevier North-Holland, pp. 1293–1343. Hellerstein, W. and C.E. McLure (2004), ‘The European Commission’s Report on Company Income Taxation: What the EU Can Learn from the Experience of the United States’, International Tax and Public Finance, 11, 199–220. Hines, J.R. (1996), ‘Altered States: Taxes and the Location of Foreign Direct Investment in America’, American Economic Review, 86(5), 1076–94. Hines, J.R. (1999), ‘Lessons from Behavioral Responses to International Taxation’, National Tax Journal, June, 52(2), 305–22. Huizinga, H. and G. Nicodème (2006), ‘Foreign Ownership and Corporate Income Taxation: An Empirical Evaluation’, European Economic Review, 50(5), 1223–44. IFS Capital Taxes Group (1991), Equity for Companies: A Corporation Tax for the 1990s, London: The Institute for Fiscal Studies. Janeba, E. and M. Smart (2003), ‘Is Targeted Tax Competition Less Harmful Than Its Remedies?’, International Tax and Public Finance, 10(3), 259–80. Jensen, M. (1986), ‘Agency Costs of Free Cash Flow, Corporate Finance and Takeovers’, American Economic Review, 7(2), 323–9. Keen, M. (2001), ‘Preferential Tax Regimes Can Make Tax Competition Less Harmful’, National Tax Journal, 54(4), 757–72. Keen, M. and J. King (2002), ‘The Croatian Profit Tax: An ACE in Practice’, Fiscal Studies, 23(3), 401–18. Keen, M. and J.E. Ligthart (2006), ‘Incentives and Information Exchange in International Taxation’, International Tax and Public Finance, 13(2/3), 163–80. Keen, M. and H. Piekkola (1997), ‘Simple Rules for the Optimal Taxation of International Capital Income’, Scandinavian Journal of Economics, 99(3), 447–61. Keen, M. and D. Wildasin (2004), ‘Pareto-efficient International Taxation’, American Economic Review, 94(1), 259–75. Keuschnigg, C. and M.D. Dietz (2007), ‘A Growth Oriented Dual Income Tax’, International Tax and Public Finance, 14(2), 191–221. King, M. and D. Fullerton (1984), The Taxation of Income from Capital: A Comparative Study of the United States, the United Kingdom, Sweden and West Germany, Chicago: Chicago University Press. Kleinbard, E. (2007a), ‘Rehabilitating the Business Income Tax’, Discussion Paper, The Hamilton Project, The Brookings Institution.

The challenges of corporate income taxes 181 Kleinbard, E. (2007b), ‘Designing an Income Tax on Capital’, in C.E. Steurele, L.E. Burman and H.J. Aaron (eds) Taxing Capital Income, Washington, DC: Urban Institute Press. Klemm, A. (2006), ‘Allowances for Corporate Equity in Practice’, IMF Working Paper, No. 259. Markusen, J. (2002), Multinational Firms and the Theory of International Trade, Cambridge, MA: MIT Press. McLure, C.E. and J.M. Weiner (2000), ‘Deciding Whether the European Union Should Adopt Formula Apportionment of Company Income’, in S. Cnossen (ed.), Taxing Capital Income in the European Union – Issues and Options for Reform, Oxford: Oxford University Press. Meade Committee (1978), The Structure and Reform of Direct Taxation, London: Allen & Unwin. Miller, M. (1977), ‘Debt and Taxes’, Journal of Finance, 32(2), 261–76. Mintz, J. (1994), ‘Is There a Future for Capital Income Taxation?’, Canadian Tax Journal, 42(6), 1469–1503. Mintz, J. (1995), ‘The Corporation Tax: A Survey’, Fiscal Studies, 16(4), 23–68. Mintz, J. (2004), ‘Corporate Tax Harmonization in Europe: It’s All About Compliance’, International Tax and Public Finance, 11(2), 221–34. Mintz, J. and J. Martens Weiner (2003), ‘Exploring Formula Allocation for the European Union’, International Tax and Public Finance, 10(6), 695–711. Mullins, P. (2006), ‘Moving to Territoriality? Implications for the U.S. and the Rest of the World’, Tax Notes International, 2 September. Musgrave, Peggy B. (1963), Taxation of Foreign Investment Income: An Economic Analysis, Baltimore: Johns Hopkins Press. Nicodème, G. (2006), ‘Corporate Tax Competition and Coordination in the EU: What Do We Know? Where Do We Stand?’, European Economy, Economic Papers, No. 250, EC. Nicodème, G. (2009), ‘Corporate Income Tax and Economic Distortions’, Taxation Papers, Directorate General for Taxation and Customs Union, EC, 1–19. OECD (1994), Taxation and Small Business, Paris: OECD. OECD (1997), Small Businesses, Job Creation and Growth: Facts, Obstacles and Best Practices, Paris: OECD. OECD (1998), ‘Harmful Tax Competition: An Emerging Global Issue’, OJC 002/1, January. OECD (2002), ‘Tax Incentives for Research and Development: Trends and Issues’, Science Technology Industry Review, Paris: OECD. OECD (2006), ‘The OECD’s Project on Harmful Tax Practices: 2006 Update on Progress in Member Countries’, OECD Centre for Tax Policy and Administration, Paris. OECD (2007), ‘Fundamental Reform of Corporate Income Tax’, Tax Policy Studies No. 16. Overesch, M. (2008), ‘The Effects of Multinational’s Profit Shifting Activities on Real Investments’, Mannheim: ZEW. Overesch, M. and J. Rinke (2008), ‘Tax Competition in Europe 1980–2007. Evidence from Dynamic Panel Data Estimation’, CESifo Conference, 26–27 April, Mannheim: ZEW. Redoano, M. (2007), ‘Fiscal Interactions among European Countries: Does the EU Matter?’, CSGR Working Paper Series No. 222. Sinn, H.-W. (1991), ‘The Vanishing Harberger Triangle’, Journal of Public Economics, 45(3), 271–300. Sørensen, P.B. (1994), ‘From the Global Income Tax to the Dual Income Tax: Recent Tax Reforms in the Nordic Countries’, International Tax and Public Finance, 1, 57–79. Sørensen, P.B. (2004a), ‘International Tax Coordination: Regionalism versus Globalism’, Journal of Public Economics, 88(6), 1187–214. Sørensen, P.B. (2004b), ‘Company Tax Reform in the European Union’, International Tax and Public Finance, 11(1), 91–115. Sørensen, P.B. (2005a), ‘Neutral Taxation of Shareholder Income’, International Tax and Public Finance, 12(6), 777–801. Sørensen, P.B. (2005b), ‘Dual Income Taxation – Why and How?’, FinanzArchiv/Public Finance Analysis, 61(4), 559–86.

182

The Elgar guide to tax systems

Sørensen, P.B. (2007), ‘Can Capital Income Taxes Survive? And Should They?’, CESifo Economic Studies, 53(2), 1–57. Spengel, C. and W. Wiegard (2004), ‘Dual Income Tax: A Pragmatic Reform Alternative for Germany’, CESifo DICE Report, 2(3), 15–22. Stiglitz, J.E. (1973), ‘Taxation, Corporate Financial Policy and the Cost of Capital’, Journal of Public Economics, 2(1), 1–34. Tanzi, V. (1995), Taxation in an Integrating World, Washington, DC: Brookings Institution. U.S. Treasury (1992), Integration of the Individual and Corporate Tax Systems – Taxing Business Income Once, Washington, DC. Vann, R.J. (2003), ‘General Report’, Cahiers de droit fiscal international, 88-a, Sidney Congress, IFA, 21–70. Vickrey, W. (1939), ‘Averaging of Income for Income Tax Purposes’, Journal of Political Economy, 47, 379–97. Vickrey, W. (1947), Agenda for Progressive Taxation, reprints of Economic Classics, New York: Augustus Kelly Publishers, pp. 179–95 and 417–27. Weichenrieder, A.J. (2009), ‘Profit Shifting in the EU: Evidence from Germany’, International Tax and Public Finance, 16(3), 281–97. Wilson, J.D. (1999), ‘Theories of Tax Competition’, National Tax Journal, 52(2), 269–304. Winner, H. (2005), ‘Has Tax Competition Emerged in OECD Countries? Evidence from Panel Data’, International Tax and Public Finance, 12(5), 667–87. Zodrow, G.R. (1991), ‘On the Traditional and New Views of Dividend Taxation’, National Tax Journal, 41(7), 109–21. Zodrow, G.R. (2003), ‘Tax Competition and Tax Coordination in the European Union’, International Tax and Public Finance, 10(6), 651–71. Zodrow, G.R. (2006), ‘Capital Mobility and Source-Based Taxation of Capital Income in Small Open Economies’, International Tax and Public Finance, 13(2/3), 269–94.

5

Wealth and wealth transfer taxation: a survey* Helmuth Cremer and Pierre Pestieau

1

INTRODUCTION

Taxes are rarely popular but those on wealth and wealth transfers are particularly controversial. Opponents claim that they are unfair and immoral. Because of many loopholes, people of equivalent wealth pay different amounts of tax depending on their acumen at tax avoidance. They penalize the frugal and the loving parents who accumulate wealth for their children, reducing incentive to save and to invest. Concerning the wealth transfer taxes per se, they allegedly hit families that are surprised by death (hence they are sometimes called a tax on sudden death or even a death tax). Supporters of these taxes, in contrast, retort that they are of all taxes the most efficient and the most equitable. They assert that they are highly progressive and counterweigh existing wealth concentration. They also argue that they have less disincentive effects than other taxes. For a number of social philosophers and classical economists, estate or inheritance taxation is even the ‘ideal tax’. Clearly, wealth taxation more than any other generates controversy at all levels: political philosophy, economic theory, political debate, and public opinion. The truth probably lies between these two opposite camps. For economists these taxes like all taxes should be judged against the two criteria of equity and efficiency to which one could add that of simplicity and compliance. In this survey we study the two major types of taxes levied on wealth: those applied sporadically or periodically on a person’s wealth (net wealth taxes), and those applied on a transfer of wealth (transfer taxes). In surveying these taxes, we focus on the criteria of equity and efficiency. Equity is hard to gauge. It has inter- and intragenerational aspects, which can only be measured by relying on some normative criteria. Efficiency implies minimizing distortions to economic activity with an important dynamic dimension. We shall stress two important features. First, these two types of taxation cannot be analyzed separately from other taxes; they have to be considered as part of the overall tax system. Second, their implications 183

184

The Elgar guide to tax systems

in terms of efficiency and equity depend on why people hold or transfer wealth. Besides the distinction between the two types of wealth taxation, it is interesting to distinguish among the motivations for accumulating wealth. Motivations are important because they determine the way individuals react to taxation. Among the traditional motivations are consumption smoothing (retirement, children’s education, precaution against life risks), prestige or status, and intergenerational transfers. Those transfers can be triggered by pure altruism, imperfect altruism (joy of giving) or by exchange motives (e.g., the so-called strategic bequests). As a benchmark, we consider a dynamic model in which the only motive for holding wealth is consumption smoothing through the individual’s life-cycle. In other words, wealth is neither held for the status it may bring nor for bequest reasons. Then, we introduce these additional motives for accumulating wealth. As it will appear, the resulting tax structure depends on the motive(s) underlying capital accumulation. The survey deliberately adopts a theoretical and normative view. It studies how wealth or wealth transfers ought to be taxed along with other tax tools and according to some welfare criteria. The tax structure thus obtained does not necessarily correspond to existing taxes. The rest of this chapter is organized as follows. Section 2 presents a brief overview of actual wealth taxation. Section 3 looks at the motives for capital accumulation. Section 4 develops the optimal tax structure when the only motive for saving is life-cycle consumption smoothing. We proceed in steps. We first assume that individuals’ horizon is finite and then that it is infinite. Section 5 looks at a number of additional motives for wealth accumulation and for each of them analyzes the optimal tax structure. Section 6 considers the possibility that the same individuals have a mix of motives for accumulating wealth and the case where society comprises individuals having different motives. A final section concludes.

2

FACTS ON WEALTH AND INHERITANCE TAXATION

Evolution and Importance In this survey we study the two major types of taxes levied on wealth: those applied sporadically or periodically on a person’s wealth (net wealth taxes), and those applied on a transfer of wealth (transfer taxes). They are presented in Table 5.1. We ignore the taxes on immovable property (called, in the US, property

Wealth and wealth transfer taxation 185 Table 5.1

Wealth taxes

Form

Examples

Net wealth tax

Periodic Sporadic (capital levy)

Transfer tax Transferor-based Recipient-based

Estate tax, gift tax, unified tax Inheritance tax, gift tax, accessions tax

taxation), which in most countries are local. They are often viewed as the price residents have to pay for local schools, as well as other local services and infrastructure of various kinds. Net wealth taxes are typically assessed on the net value of the taxpayer’s taxable assets (i.e., value of assets minus any related liability), either sporadically (often known as ‘capital levies’ ) or on an annual or other periodic basis. Transfer taxes, which are typically assessed on the net value of the taxable assets transferred, fall into two basic categories: those levied on the donor, more precisely on his or her estate (typical in common law countries), and those levied on the recipient, namely the heir. Donor-based taxes can be levied separately on inter vivos transfers (gift tax) and on transfers at death (estate tax), or together in a single integrated tax. Recipient-based taxes can also be levied on inter vivos transfers (gift tax), on transfers at death (inheritance tax), and on an integrated basis (accessions tax). OECD (2008) provides data on taxation for the period 1965–2006. We will restrict our presentation to EU15, the US, and Japan. Under the label of taxes on property we find both net wealth and transfer taxes but also many other taxes. This aggregate represented 5.6 per cent of total taxation and 2.2 per cent of GDP for EU15 in 2005. In 1965, these figures were respectively 6.7 and 1.8. See Tables 5A.1 and 5A.2 in Appendix A at the end of the chapter. As mentioned this aggregate is too heterogeneous to be informative. We thus focus on wealth and wealth transfer taxes. At the present time, in EU15, Japan, and the US, only a single country, namely France, has a wealth tax. Luxembourg, Greece, Finland, Sweden, and Spain have only recently abolished their wealth tax. (Note that Norway and Switzerland are the two other Western European countries with an annual wealth tax.) The French annual direct wealth tax, called ‘solidarity tax on wealth’ concerns those having assets in excess of 770 000 euros (as of 1 January 2008). It yielded in 2006 about 0.4 per cent of total tax revenue. One can really talk of an endangered tax species.

186

The Elgar guide to tax systems Spain

3

France 2.5

United Kingdom United States

2

Japan 1.5 1 0.5 0 1965

1970

1975

1980

1985

1990

1995

2000

2005

Source: OECD (2008).

Figure 5.1

Wealth transfer taxes as a percentage of GDP

0.90

Spain

0.80

France

0.70

United Kingdom

0.60

United States

0.50

Japan

0.40 0.30 0.20 0.10 0.00 1965 1970 1975 1980 1985 1990 1995 2000 2005 Source: OECD (2008).

Figure 5.2

Wealth transfer taxes as a percentage of total tax revenues

In contrast, of these 17 countries only Sweden and Italy do not have a wealth transfer tax. The UK and the US have an estate tax and all the other countries have some sort of inheritance tax. Figures 5.1 and 5.2 and Tables 5.2 and 5.3 give the size and the evolution of the wealth transfer tax over the period 1965–2006 in EU15, in the US, and in Japan.

Wealth and wealth transfer taxation 187 Table 5.2

Estate, inheritance, and gift taxes as a percentage of total taxation 1965 1970 1975 1980 1985 1990 1995 2000 2005 2006

Belgium Denmark Germany Ireland Greece Spain France Italy Luxembourg Netherlands Austria Portugal Finland Sweden United Kingdom United States Japan Source:

1.17 0.65 0.22 1.89 0.86 1.05 0.56 0.85 0.38 1.08 0.26 2.02 0.22 0.39 2.62

1.06 0.36 0.23 1.25 1.35 0.86 0.72 0.64 0.48 0.59 0.23 1.47 0.24 0.36 2.01

0.76 0.38 0.14 1.12 1.00 0.79 0.72 0.21 0.32 0.37 0.19 0.86 0.21 0.25 0.82

0.82 0.44 0.18 0.35 1.22 0.41 0.57 0.21 0.32 0.48 0.17 0.22 0.22 0.21 0.59

0.59 0.47 0.22 0.30 0.95 0.41 0.61 0.23 0.27 0.44 0.17 0.83 0.27 0.26 0.69

0.71 0.56 0.34 0.39 1.23 0.42 0.95 0.14 0.29 0.50 0.14 0.50 0.37 0.19 0.65

0.76 0.47 0.26 0.44 0.97 0.51 0.82 0.15 0.27 0.61 0.11 0.21 0.38 0.16 0.58

0.97 0.45 0.39 0.67 0.80 0.63 1.07 0.20 0.27 0.90 0.01 0.25 0.59 0.22 0.62

1.30 0.40 0.53 0.50 0.42 0.74 1.19 0.01 0.39 0.86 0.14 0.08 0.70 0.07 0.70

1.39 0.43 0.46 0.62 0.34 0.74 1.04 0.01 0.39 0.86 0.12 0.01 0.70 0.01 0.74

2.06 0.71

1.68 0.94

1.45 0.97

1.15 0.71

0.82 1.18

1.00 1.47

0.98 2.02

1.22 1.31

0.90 1.14

0.89 1.06

OECD (2008).

As it appears from these figures wealth transfer taxes play only a minor role in the total tax revenues of countries. Within our sample of OECD countries in 2006, Belgium, Japan, and France reach with respectively 1.39, 1.06, and 1.04 percent, the highest shares in total tax revenues (Table 5.2). In Portugal, by contrast, the share is less than 0.2 per cent and is the lowest. Sweden and Italy have abandoned it. As one sees from Figure 5.1, both the US and the UK experienced a huge decline, from 2.62 to 0.74 and from 2.06 to 0.89 respectively over the last four decades. France has experienced an increase with a peak in 1995. In Japan and Spain the evolution was less marked, increasing in the first and decreasing in the second. Tax Structure A comparison of the different national systems of wealth transfer taxation would be easy if the differences were only in the average yield. However, a given yield can be obtained with totally different systems. Take the example of the US and France. In the US, the tax is on the

188

The Elgar guide to tax systems

Table 5.3

Estate, inheritance, and gift taxes as a percentage of GDP 1965 1970 1975 1980 1985 1990 1995 2000 2005 2006

Belgium Denmark Germany Ireland Greece Spain France Italy Luxembourg Netherlands Austria Portugal Finland Sweden United Kingdom United States Japan

0.36 0.19 0.07 0.47 0.15 0.15 0.19 0.22 0.10 0.35 0.09 0.32 0.07 0.14 0.80

0.36 0.14 0.07 0.35 0.27 0.14 0.24 0.16 0.11 0.21 0.08 0.27 0.08 0.14 0.74

0.30 0.15 0.05 0.32 0.19 0.15 0.25 0.05 0.11 0.15 0.07 0.17 0.08 0.10 0.29

0.34 0.19 0.07 0.11 0.44 0.09 0.23 0.06 0.12 0.21 0.07 0.05 0.08 0.10 0.21

0.26 0.22 0.08 0.10 0.24 0.11 0.26 0.08 0.10 0.19 0.07 0.21 0.11 0.12 0.26

0.30 0.26 0.12 0.13 0.32 0.14 0.40 0.05 0.10 0.21 0.06 0.14 0.16 0.10 0.24

0.33 0.23 0.10 0.14 0.28 0.16 0.35 0.06 0.10 0.25 0.05 0.07 0.17 0.08 0.20

0.44 0.22 0.14 0.21 0.27 0.22 0.48 0.08 0.10 0.36 0.01 0.08 0.28 0.11 0.23

0.58 0.20 0.18 0.15 0.13 0.26 0.52 0.00 0.15 0.34 0.06 0.03 0.31 0.03 0.26

0.62 0.21 0.16 0.20 0.11 0.27 0.46 0.00 0.14 0.34 0.05 0.00 0.30 0.00 0.27

0.51 0.13

0.45 0.18

0.37 0.20

0.30 0.18

0.21 0.32

0.27 0.43

0.27 0.54

0.37 0.35

0.25 0.31

0.25 0.30

Source: OECD (2008).

whole estate, hence the name ‘estate taxation’. The rates range from 18 to 55 per cent over 17 brackets. They are independent ot the relation between the donor and the donees and on the number of donees. The threshold for exemption is close to 2 million dollars (2006–08). France has an inheritance tax, which means that the tax base is the amount received and thus decreases with the number of donees. The tax (euros) rates vary according to the relation between the donator and the recipients: 5–40 per cent for the descendants (46 000), 5–45 per cent for the spouse (7000) and 5–60 per cent for the siblings (15 000), with threshold for exemption in parentheses (in €). In other words the US estate tax concerns a smaller fraction of households than the French inheritance tax for about the same yield. Estate Versus Inheritance Tax In general, estate taxation gives one total freedom to bequeath one’s wealth to anyone or anything. Disinheritance is possible, as long as the decedent prepares an explicit will. Inheritance taxation, on the other hand,

Wealth and wealth transfer taxation 189 often comes with the legal obligation to bequeath one’s wealth to one’s children, if any, and with an equal sharing rule for most of the estate. Donors have some freedom to allocate a small fraction of the estate, but this fraction declines with the number of children. As the relation between recipient and donor gets more distant, the inheritance tax treatment becomes less and less generous. The relative merits of the estate-type and the inheritance-type taxation are clear. The first is simple and relatively easy to administer, leaving all discretion to donors to dispose of their wealth as they wish. This means that it is possible to compensate some children over others for differences in income or need, and that it is possible to disinherit one’s children. By contrast, the inheritance tax is more equitable than the estate tax in that it lightens the tax load of large families. Yet, it does not allow for compensatory treatment of children with uneven endowments.1 Basically, estate taxation reflects a concept of the family and of the state that is quite different from the one that governs inheritance taxation. If one trusts parents to be fair in disposing of their estate, and if one believes that intrafamily inequality is as important as interfamily inequality, then what is desirable is a combination of freedom of bequest and a very low estate tax. Unpopular Taxes There is something paradoxical about wealth and wealth transfer taxation. Both taxes are widely discussed in public finance and political circles, but at the same time they are not very popular. They are hardly mentioned in public finance textbooks, they are not very well documented, and their yield is negligible. Many countries have been more or less successfully struggling for the last decade with the question of whether to repeal their wealth or wealth transfer taxes (see above). As surveys show, they are not popular taxes, even though they typically hit only a minority of the population. The example of the US estate tax is typical. On the one hand, it affects fewer than 2 per cent of decedents and is therefore of no direct concern to most taxpayers. On the other hand it is so unpopular that Auerbach (2006, p. 19) writes, ‘it might make little sense at the moment to argue in favor of the estate tax in the United States’. Most recent public finance or public economics textbooks do not even mention those taxes. Data are scarce with the exception of those from the OECD (see e.g., OECD, 2008). On the issue of unpopularity, Frank (2005) argues that the way questions are phrased in opinion polls is of crucial importance. He shows that voters would not favor repealing the estate tax if they took into account the policy changes that such a reform would necessarily have to entail (raising other

190

The Elgar guide to tax systems

taxes, cutting government services, or increasing federal borrowing). When asked just about repealing the tax without mentioning its repercussions, respondents do favor the repeal by almost three to one. When respondents are reminded that the revenue shortfall would have unpleasant repercussions, these respondents opposed repeal by almost four to one.2

3

CAPITAL ACCUMULATION MOTIVES

It is now widely agreed that to understand the allocative and distributional effects of wealth and wealth transfer taxation one needs to have a better grasp of the saver’s motives. Among these motives, one has to distinguish those that are purely selfish and those that concern intergenerational transfers (gifts and inheritance). We examine briefly a number of motives that have been offered in the literature and sketch their implications. The first two motives are purely selfish. The last three concern bequests.3 Consumption Smoothing This is the most traditional motive for saving over one’s life-cycle, with or without uncertainty. It includes the need for replacement income after retirement, financing of children’s education, precautionary saving, and self-insurance. It is well known that this kind of saving decreases with social insurance and tends to be smaller when individuals think only short term. In case of imperfect annuity markets and ‘premature’ death, part of life-cycle saving is not consumed and leads to what is called accidental or unplanned bequests. This form of bequest is by its nature unaffected by estate or inheritance taxation. Preference for Wealth It is today widely agreed upon that neither life-cycle saving nor bequest motives can explain the top end of the wealth distribution. This brings us back to Max Weber’s theory of ‘the spirit of capitalism’ generalized by Kurz (1968): capitalists accumulate wealth for its own sake. To cite Weber (1958, p. 53): Man is dominated by making of money, by acquisition as the ultimate purpose of his life. Economic acquisition is no longer subordinated to man as the means for the satisfaction of his material needs. This reversal relationship, so irrational from a naive point of view, is evidently a leading principle of capitalism.

Wealth and wealth transfer taxation 191 As argued by Carroll (2000): the saving behavior of the [US] richest households cannot be explained by models in which the only purpose of wealth accumulation is to finance future consumption, either their own or that of heirs.’

Then, to explain such a behavior one has to assume that some consumers regard accumulation as an end in itself or as a channel leading to power, which is equivalent to assuming that wealth is intrinsically desirable, what we call here ‘preference for wealth’. Pure Dynastic Altruism: Altruistic Bequest4 Parents care about the likely lifetime utility of their children and hence about the welfare of future generations. Consequently, wealthier parents tend to make larger bequests. Conversely, holding parents’ wealth constant, children with higher labor earnings will receive smaller bequests. When there are no rules restricting freedom to testate, there is also a tendency for parents to leave different amounts to different children, in order to equalize their incomes. Finally, pure altruism typically leads to the Ricardian equivalence: parents compensate any intergenerational redistribution by the government through matching bequests. In consequence, debt and pay-as-you-go Social Security have no effect on capital accumulation. Joy of Giving: Paternalistic Bequest (Bequest-as-last-consumption)5 Parents here are motivated not by ‘pure’ altruism but by the direct utility they receive from the act of giving. This phenomenon is also referred to as ‘warm glow’ giving. It can be explained by some internal feeling of virtue arising from sacrifice in helping one’s children or by the desire to control their life. Formally, these bequests appear in the utility function as a consumption expenditure incurred in the last period of life. Ceteris paribus, they are subject to income and price effects but do not have any compensatory effect, namely they are not intended to smooth consumption across generations. A crucial element is whether what matters to the donor is the net or the gross of tax amount. Exchange-related Motives: Strategic Bequests6 In their canonical form, exchange-related models consider that children choose a level of ‘attention’ to provide to their parents. In exchange,

192

The Elgar guide to tax systems

parents ‘remunerate’ them through a prospective bequest. The exchanges can involve all sorts of non-pecuniary services and they can be part of a strategic game between parents and children. Strategic bequests, as they were originally presented, imply that parents extract all the surplus from their children by playing them against each other. Strategic or exchange bequests depend on the wealth and the needs of the donor; they are not compensatory between parents and children and they do not need to be equal across children.

4

OPTIMAL TAXATION OF CAPITAL INCOME

Our theoretical discussion of wealth taxation will be organized in two stages. First of all, in this section, we assume that there are neither bequest motives nor preference for wealth. In that case, within the standard overlapping generations model there is no distinction between wealth and capital income. In the next section, bequest motives and preference for wealth are introduced. In other words, it is recognized that saving is not motivated solely by retirement concerns. We then show that these other motives may have a significant impact on the rate and on the structure of taxation. In this section we examine two propositions that lead to zero taxation of capital income. The first one, called the Atkinson-Stiglitz proposition, is discussed within the overlapping generations model. The second one, known as the Chamley-Judd theorem, is presented in a model with infinitely lived individuals.7 The Overlapping Generations Model The overlapping generations (OLG) model is the conventional setting to discuss capital income taxation when saving is exclusively motivated by consumption smoothing. It considers finitely lived generations that overlap, along with an infinitely lived government. We use the two-period model, with labor supply in the first period and consumption in both the first and second periods. Saving from first-period earnings is used to finance second-period consumption, generating capital income that is taxable (in the second period). Since there is only a single period of work, the model can be viewed as shedding light on the taxation of saving for retirement. This model allows for introducing the Atkinson-Stiglitz (1976) (hereafter AS) proposition. It states that when the available tax tools include nonlinear labor income taxes, taxation of saving or of capital income is not optimal if two key conditions are satisfied: (1) preferences are (weakly)

Wealth and wealth transfer taxation 193 separable between consumption and labor and (2) all consumers have the same utility function. Originally the AS proposition relies on the idea that the income tax is optimal. Recently Laroque (2005) and Kaplow (2006) consider the desirability of capital income tax with any earned income tax function. They show that one can always move from taxing to not taxing capital income with an appropriate modification of the earned income tax. Such a move does generate more government revenue while leaving every consumer with the same utility and the same labor supply. The intuition of the original, as well as the revisited AS proposition, is that with separability and identical utilities taxing saving cannot relax the self-selection constraints but would have an added efficiency cost.8 To counter the AS result and its zero capital income taxation there are several angles of attack.9 The first one is clearly to question the assumption of separability and that of homogeneous preferences. Dropping the assumption of separability would not necessarily result in taxing capital income. Subsidizing is as likely. Introducing heterogeneity in preferences appears to be more promising. It has been done in different ways. There are at least three potential sources of heterogeneity that can lead to a tax on capital income: time discount rates, longevity, and initial endowments. Saez (2002) questions the Atkinson-Stiglitz theorem on the basis of differences in time preferences across individuals with different skills. He shows that capital income taxation becomes desirable under the plausible assumption that those with higher earnings abilities discount the future less (and thus save more out of any given income). Cremer et al. (2010) use another stylized fact, namely the positive correlation between income and longevity to reach the same conclusion. Cremer et al. (2003) introduce an endowment (inherited wealth) as a second unobservable characteristic. They show that if ability and endowment are positively correlated then it is efficient to tax capital income. If we discuss the AS proposition in the standard OLG setting, we have to keep in mind that there is no guarantee that the optimal accumulation of capital is achieved. If the government does not have direct control of capital, it can use tax policy to affect the capital–labor ratio. In that case, even with separability and identical utilities, a tax on capital income is needed. This in itself is quite intuitive. However, the design of the appropriate tax rule is more complex. For instance, a need for additional capital accumulation, because the capital stock is below the modified golden rule level, will not necessarily lead to less taxation of capital income and more taxation of labor income. What matters is aggregate saving and this may depend much more on net of tax earnings than on the rate of interest.10 Another variation of the standard model is to allow for uncertain

194

The Elgar guide to tax systems

earnings in the second period of life. Cremer and Gahvari (1995) have shown that if consumption decisions are to be taken before earnings uncertainties are resolved then the Atkinson-Stiglitz result fails to hold. Banks and Diamond (2010) discuss the implications of this result for capital income taxation. They argue that the case of uncertainty is similar to the situation (discussed above) where high-wage individuals discount the future less. In the latter case, a high-wage individual imitating someone with less skill saves more than a low-wage individual. Taxing capital income is then an effective way to release an otherwise binding incentive constraint. Under uncertainty, this argument goes through. An individual who plans to earn less than the government-planned amount in the event of high skill has a higher valuation of saving than the individual with the government-planned income level. Consequently, a tax on savings continues to relax an incentive constraint. To illustrate this argument, Banks and Diamond (2010) point out that retirement age tends to be lower for those with higher savings. Consequently, taxing savings discourages earlier retirement. Uncertain earnings are a central element of what is known as the ‘new dynamic public finance’.11 This literature is quite complex and leads to a number of interesting insights. However, the basic case for taxation of capital income is based on the same argument as in Cremer and Gahvari (1995). We can also look at simulation results to see how important and how large the capital income tax might be. Conesa et al. (2009) have done a complex simulation of the asymptotic position of an empirically calibrated OLG model with uncertain individual wages and lengths of life. They have a three-parameter earnings tax (the same for each age), a 100 per cent estate tax financing poll subsidies, a pay-as-you-go Social Security system, a linear tax on capital income, and no government debt or assets. They choose taxes to optimize the long-run position of the economy and find a capital income tax rate of 36 percent, while the tax on labor income is nearly linear at 23 percent. Erosa and Gervais (2002) have examined the most efficient taxation of a representative consumer within an OLG setting. If the utility discount rate differs from the real discount rate, individuals will choose non-constant age profiles in both consumption and earnings, even if annual utilities are additive and the same over time, while the wage rate is also constant over time. Consequently, the optimal age-dependent taxes on consumption and earnings are not uniform over time, resulting in non-zero implicit taxation of savings. Interestingly, taxation or subsidization of savings remains optimal when taxes are constrained to be uniform for workers of different ages.

Wealth and wealth transfer taxation 195 Infinite Horizon In the above models there is a contrast between finitely lived individuals, who are intergenerationally disconnected, and the government, which has an infinite horizon and a different time preference. Let us now look at another class of models wherein individuals are infinitely lived and have the same discount rate as the central planner. For the purpose at hand the central finding of this literature, due to and Judd (1985) and Chamley (1986), is the optimality of zero capital income taxation in the long term. The intuition behind this result can be understood by looking at the wedge that a capital income tax introduces between the intertemporal marginal rate of substitution (MRS) and the intertemporal marginal rate of transformation (MRT). Let us illustrate this through a simple example.12 Take a tax rate of 30 per cent and a rate of return of 10 percent. In a year, the wedge between MRS (5 1 1 0.1 (1 2 0.3) 5 1.07) and MRT (5 (1 1 0.1) 5 1.1) is small and the distortion on the saving choice is negligible. After 40 years, the capital income tax generates a 67 per cent wedge between consumption today and consumption in 40 years. As a matter of fact, as the time horizon T goes to infinity, the ratio between MRS (5 [ 1 1 0.1 (1 2 0.3) ] T) and MRT (5 [ 1 1 0.1 ] T) tends to zero. Consequently, when the investor has a very long time horizon the capital income tax becomes extremely inefficient. The Chamley-Judd no capital income taxation conclusion has become the standard rule for a number of public economists and particularly macroeconomists; see, for example, Chari and Kehoe (1999). It has also been challenged on various grounds. It relies on a set of strong assumptions. As with the Atkinson-Stiglitz result, a key question is how robust their theorem is to realistic changes in the model. There is first the steadystate assumption; we know that during the transition capital income is subject to taxation. There is also the assumed equality between the private and the social discount rate and the absence of liquidity constraints. If one departs from these assumptions the tax is not any more equal to zero even in the steady state. Their model assumes also that there are no constraints on the tax tools. As shown by Correia (1996) and Coleman (2000) as soon as some taxes are constrained the zero tax result ceases to hold. Uncertainty about earnings, along with borrowing constraints is shown to lead to a positive tax. See on this Chamley (2001) and Golosov et al. (2003). Finally, let us mention a paper by Saez (2002) who introduces a progressive tax on capital income into the Chamley-Judd model. Under some plausible assumptions, he shows that such a tax is desirable; it drives all the large estates down to a finite level, thus generating a truncated longrun wealth distribution.

196

The Elgar guide to tax systems

To conclude this section, it seems that the case for a zero-tax on capital income when the only motive for saving is life-cycle consumption smoothing is rather weak. While Atkinson-Stiglitz, on the one hand, and Chamley-Judd, on the other hand, are often invoked to advocate a tax exemption on capital income, there appears to be a striking discrepancy between common beliefs and actual results. Under closer scrutiny, it is clear that either of these zero tax results does not apply under ‘plausible’ circumstances. We shall now turn to the other motives for saving.

5

WEALTH TAXATION WITH ALTERNATIVE MOTIVES FOR SAVING

We now consider the case where besides life-cycle considerations individuals may have motives for accumulating wealth. We start by presenting the general model from which all the subsequent special cases are derived. The Canonical Model Identical individuals live for two periods, consuming in both while working only in the first one.13 Population is increasing at the rate n. The government has an exogenously given revenue requirement, to be financed through taxes on income from labor and capital and on estate transfers, if any. Individuals may derive some utility from transferring resources to their offspring. The problem of the representative consumer is to maximize utility subject to the budget constraint: bt 1 wt,t 5 ct 1

dt11 1 xt11 , 1 1 rt11

(5.1)

where bt is inherited wealth, xt11 is the amount of bequests, wt is the consumer wage (net of tax wage), rt11 the consumer rate of interest (after tax interest rate), ct, first period consumption, ,t, labor supply and dt11, second period consumption. Preferences are represented by the following utility function: ut 5 u (ct) 1 bu (dt11) 2 H (,t) 1 gBt11

(5.2)

where Bt11 is the utility derived from bequeathing if any, b and g are positive parameters, u ( # ) is strictly concave and H ( # ) strictly convex. The

Wealth and wealth transfer taxation 197 additive specification is used for the sake of simplicity. When there is just a life-cycle motive for saving as in the previous section, one has g 5 0, b 5 x 5 0. Altruistic Bequests In this subsection we consider the case where individuals save for their own retirement consumption needs and for making sure that their children’s welfare is sufficiently high. The standard way of dealing with this problem is to adopt the infinitely lived individuals model. Instead of considering an infinite series of years of one individual life we consider an infinite series of generations (a dynasty), which are linked by bequests. Using the above canonical model, we posit Bt11 5 ut11.To keep things relatively simple, we assume that b 5 0 so that d 5 0. In other words, people live only one period and only save for bequeathing. This assumption implies that the tax on saving is also the tax on wealth transfer.14 Then, by recursion, the problem of the individual at time 0, which is also that of the the social planner, it is to maximize: `

a g u (ct, ,t) , t

t50

with bt11 5 xt11 / (1 1 n) . We assume non-negative bequests, which corresponds to the liquidity constraint in the infinitely lived individuals model, and the equality between the social and the individual discount factor (g) . Then, one has the Ricardian equivalence implying the neutrality of the debt. One also has the Chamley-Judd result15 (1) initially a tax on both earnings and saving (that is, bequests); (2) in the long run the tax on saving tends to 0; but as we have seen it is not very robust. Paternalistic Bequests These bequests are also called ‘bequests as last consumption’ or ‘joy of giving’ bequests. We now have Bt11 5 v (xt11) and bt11 5 xt11 / (1 1 n) . Unlike in the case of pure altruism, the objective of individuals and  that  of the social planner may now diverge. Each individual maximizes: u (ct, dt11, ,t) 1 gv (xt11) . To obtain the social optimum, there is the issue of whether or not individual utilities should be ‘laundered’. Hammond (1988) and Harsanyi

198

The Elgar guide to tax systems

(1995) have advocated ‘excluding all external preferences, even benevolent ones, from our social utility function’. Advocates of a utilitarian approach, on the other hand, argue that the social planner cannot paternalistically modify individuals’ preferences. Bequests are potentially subject to a double tax: first, the tax on savings, tr, and then the specific tax on transfers tx. This latter tax depends on the extent of laundering. When there is laundering, bequests lose their direct social utility and are thus subject to a relatively higher tax. In the absence of laundering it is not impossible to have a negative marginal tax. For example, Fahri and Werning (2010), who do not launder their utilities, study efficient allocations in a model with altruistic parents and focus on the implications for estate taxation. They show that optimal estate taxes have two important features. First, taxation should be progressive, so that more productive parents face a lower net return on bequests. Second, marginal taxes should be negative, so that parents face a marginal subsidy on bequests. They show that these features can be implemented using a simple nonlinear estate tax schedule, independent of income taxation.16 Exchange-based Bequests To deal with exchange-based bequests, we modify the canonical model as follows: Bt11 5 h (at11) and ut 5 u [ ct 2 v (agt) , ,t, at11 ] 1 bu (dt11) , where at11 is attention received, agt is attention given representing a monetary cost of v (agt) that is paid by a bequest bt. In the strategic bequest vein, we assume that bt 5 v (agt) (parents extract all the surplus from their children who are just paid for the disutility of their effort). We now have three tax instruments: a proportional tax on earnings, interest income, and inherited wealth with rates tw,tr and tx. The overall tax on bequests, tr 1 tx (1 1 r) , may or may not be higher than that on future consumption. In other words, there is no particular reason to believe that the wealth transfer tax tx is positive. This will depend on the relative magnitude of the compensated derivatives that determine the overall tax on bequests and the tax on future consumption. For example, if the demand for attention is much more elastic than that for future consumption, the tax on inheritance, tx, is negative.

Wealth and wealth transfer taxation 199 Accidental Bequests | | To deal with accidental bequests, we posit g 5 0, b 5 b q, where b is the factor of time preference and q is the survival probability. There is a probability q that the individual will live until the end of the second period and (1 2 q) that he or she will die at the end of the first period. In the latter case, bt11 5 dt11 / (1 1 n) for a fraction (1 2 q) of children whose parents decease prematurely. The accidental bequest case is not much different from the case without bequest. Saving is affected by survival probabilities. Accidental transfers are taxed at 100 percent, without affecting the supply of saving. The part of public spending (if any) that exceeds the proceeds of the transfer tax is financed through labor and capital income taxes.17 Preference for Wealth The case with preferences for wealth is close to that of paternalistic bequests with one exception: here individuals obtain the same utility from saving for retirement and for bequests: Bt11 5 h (dt11 1 xt11) and bt11 5 xt11 / (1 1 n) . As in the case of paternalistic bequests, wealth can be viewed as a consumption good and be taxed accordingly. The issue of laundering does also play an important role here.

6

HETEROGENEITY

The theoretical literature on wealth transfer taxation tends to assume that individuals have only one type of bequest motive. This section examines through some examples how the results are affected when society consists of individuals with different motives. We first turn to a society consisting of individuals who combine different motives, namely who leave both altruistic and accidental bequests. Then we consider a society where individuals are either altruistic, pure ‘life-cyclers’ or with preference for wealth. Mixed Motives It is widely believed that bequests are in reality a hybrid of canonical types analyzed above, in particular of accidental bequests (related to imperfect annuity markets) and of paternalistic bequests (related to some joy of giving). In such a case, the estate consists of two components: a certain amount planned by altruistic parents and another part that results from the ‘premature’ death of parents. More specifically, this second part represents intended second-period consumption in an overlapping generations

200

The Elgar guide to tax systems

framework. We have seen that these two types of bequests have totally different implications. Determining the relative importance of either one is thus crucial to designing an optimal estate tax. To illustrate this, we use an additive utility function: | ut 5 u (ct) 1 b qu (dt11) 2 H (,t) 1 gh (xt11) | where b is the factor of time preference and q is the survival probability. Michel and Pestieau (2002) show that the optimal value of the estate tax rate represents a compromise between the equity objective and the desire of not discouraging wealth accumulation. With g 5 0, the tax is 1; with q 5 1, the tax can be low. In this very simple model the only source of inequality is longevity q. Introducing a second source of heterogeneity, for example, different productivities, is surely more realistic. Pestieau and Sato (2008) extend the Michel-Pestieau (2002) model to include wage differences and education. They reach the same though richer results. They use linear tax tools. With nonlinear taxation, as shown by Blumkin and Sadka (2004) and Cremer et  al. (2011), a 100 per cent tax is not necessarily desirable even when g 5 0. The reason is that accidental bequests can bring some information on non-observable characteristics. Altruists and Life-cyclers: Savers and Spenders For a long time economists have rejected the idea of heterogeneous preferences. Differences in behavior had to be explained by differences in ability, inherited wealth, or random shocks. The last few years have seen an increasing awareness that it is important to account for differences in preferences pertaining to altruism or time preference. In his celebrated paper, Ramsey (1928) already indicated that within a society consisting of individuals differing in time preferences, the most patient would end up with all the wealth in the long run. In this section we address the question of wealth transfer tax in a society with two types of individuals, pure life-cyclers and altruistic savers.18 Mankiw (2000) calls them spenders and savers. Formally, their utility function is: uit 5 u (cit, d it11) 1 giuit11 with i 5 L for life-cyclers and thus gL 5 0 and i 5 A for altruists and thus gA 5 g . 0. Preferences are dynastic and there is a fixed fraction p of altruistic dynasties and a fraction (1 2 p) of non-altruistic dynasties.

Wealth and wealth transfer taxation 201 Within this setting, government debt does not affect the steady-state capital stock and national income. As in Ramsey, the altruistic (the more patient) households hold the entire capital stock. Moreover, government debt though neutral in aggregate terms increases steady-state inequality. A higher level of debt means a higher level of taxation to pay for the interest payments. The taxes fall on both life-cyclers and altruists but the interest payments go entirely to the altruist. Consequently, a higher level of debt, or alternatively of pay-as-you-go Social Security, raises the steady-state consumption and income of the altruists and lowers the steady-state consumption and income of the life-cyclers. For the purpose at hand we are interested by the incidence of a wealth transfer tax, which in the present setting is only paid by altruistic dynasties. Assuming that the proceeds of the tax are redistributed uniformly to everyone, it can be shown that the tax may lower the utility of not only the altruists but also that of the life-cyclers. This paradoxical result was already obtained by Stiglitz (1978) in a slightly different setting.19 When capital is taxed its quantity falls, which, in turn, may depress the real wage. This effect may be large enough to make any tax on wealth transfer undesirable even from the standpoint of people who own no wealth, pay no tax, and indeed benefit from a transfer. One should recall that this result is obtained in the steady state. In the short run life-cyclers could be tempted to tax inheritance and enjoy a utility boost. If they have to vote they will vote for such a tax without being concerned by the fate of their descendants. The political economy of wealth transfer thus yields a result different from steady-state social welfare maximization. It explains why a tax that would be undesirable from the steady-state standpoint can be voted on when life-cyclers hold a majority. Savers, Spenders and Hoarders Pestieau and Thibault (2011) extend the above model by including a third type, the hoarders, those who like wealth for its own sake. Their paper proposes a simple OLG model that is consistent with the essential facts of consumer behavior, capital accumulation, and wealth distribution, and yields some new and surprising conclusions about fiscal policy. By considering a society in which individuals are distinguished according to two characteristics, altruism and wealth preference, they show that those who in the long run hold the bulk of private capital are not so much motivated by dynastic altruism as by preference for wealth. Two types of social segmentation can result with different wealth distribution. To a large extent their results seem to fit reality better than those obtained with standard

202

The Elgar guide to tax systems

optimal growth models in which dynastic altruism (or rate of impatience) is the only source of heterogeneity: overaccumulation can appear, public debt and unfunded pensions are not neutral, estate taxation can improve the welfare of the top wealthy and hurt that of those without wealth. In the equilibrium wherein spenders, savers (altruists), and hoarders (preference for wealth) coexist, they show that estate taxation worsens the welfare of both spenders and savers but increases (decreases) that of the hoarders if their preference for wealth is sufficiently high (low).

7

CONCLUSION

Let us now return to some practical questions under the light of the theory just surveyed. Recurrent questions in the debate are (1) do we need an annual wealth tax, (2) do we need a wealth transfer tax and if so of what type should it be, and (3), how serious are the threats of tax ‘avoision’ (evasion/ avoidance) and tax competition on the sustainability of wealth taxation? Most EU member countries do not have an annual wealth taxation. This is to some extent consistent with the theory that shows that it is redundant with a capital income tax, given that the tax base is the same and that realized capital gains are correctly taxed. As for a wealth transfer tax, whether or not it is desirable depends on the underlying motives of bequests. We have seen that accidental bequests can be heavily taxed without generating distortions, or having a clear regressive effect. At the same time empirical work indicates that a non negligible share of inheritance can be considered as accidental.20 Naturally there are other types of bequests that do not imply a 100 per cent tax. One can thus reasonably expect that on the basis of efficiency and equity considerations the taxes should neither be zero nor 100 percent. Should we have an estate or an inheritance tax with their implications in terms of tax rates and estate sharing? A good reason to advocate inheritance taxation with lower rates for children than for any other donees is that the likelihood of accidental bequests is lower for the first than for the second. Compared with estate taxation, inheritance taxation seems to favor large rather than small families. Keeping lower rates for inter vivos gifts particularly when they are early seems also desirable given that children often need resources at the start of their active life and not when their parents pass away, which generally occurs at the verge of their own retirement. In any event, gifts are by definition intended, which also pleads for a lower tax than on bequests (which are in part accidental). There are two non-normative arguments that are often used to justify the decrease or even the abandonment of any kind of wealth taxation.

Wealth and wealth transfer taxation 203 These are tax evasion/avoidance and tax competition. Accordingly even though wealth or wealth transfer tax would be shown to be highly desirable in a closed economy and in a setting of full compliance, some people maintain that they should be dismantled because of tax competition and tax avoision. These are threats that have to be taken seriously but within reasonable limits. International coordination and cooperation can moderate tax competition as well as limit ‘avoision’. Observe that avoidance and evasion not only lead to poor tax yields but also lead to strong departures from both vertical and horizontal equality. This may explain at least in part why wealth taxation is today so unpopular that in some countries the political system is considering abolishing it. Those issues have a real political impact and yet there is little evidence on how important is their effect. What is certain is that they can be dealt with by reforming the tax and not by repealing it.

NOTES * 1. 2.

3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15.

This paper was presented at the conference ‘Tax Systems: Whence and Whither?’, Malaga, 9–11 September, 2009. We thank Laura de Pablos Escobar for her insightful comments. Cremer and Pestieau (1988) argue that tax rates that decrease with the degree of consanguinity can be redistributive. It has been shown that the way questions are framed influences the outcome of surveys. A recent paper by Fatemi et al. (2008) studies the impact of framing and prior attitude on estate tax preferences. They show that the interaction between prior beliefs on estate taxation and framing of the questionnaire was contrary to intuition. Specifically, negative beliefs happen to be reinforced by positive framing. See also on this Cox (1987); Pestieau (2000). Among the classical references are Barro (1974), Becker and Tomes (1979, 1986). See also Altonji et al. (1992). Bevan and Stiglitz (1979); Kotlikoff and Spivak (1981); Andreoni (1990); Glomm and Ravikunar (1992). Bernheim et al. (1985); Cremer and Pestieau (1991, 1996, 1998); Cremer et al. (1993). See also Chari and Kehoe (1999). To relax an incentive constraint it is necessary to find a policy measure that affects the mimicker more than the mimicked individual. With identical and separable preferences, mimicker and mimicked have the same indifference curves in the produced good space. Consequently a distortion in relative prices affects them in the same way. For a survey, see Cremer (2003). With log-linear preferences saving does not depend on the interest rate and thus on capital income taxation. See, for example, Golosov et al. (2007). This example is borrowed from Banks and Diamond (2010) who provide a more detailed discussion of this intuition. Diamond (1965). We have the following equality between saving and bequest: st 5 xt 11 5 (1 1 n) kt 11. Another result of Chamley-Judd is that initial wealth should be taxed as much as possible. Such a ‘one-shot capital levy’ would not be distortionary.

204 16. 17.

18. 19. 20.

The Elgar guide to tax systems On the question of altruism in the design of estate tax see Kaplow (2000, 2008) and Boadway et al. (2010). Kaplow (2008, pp. 264–6) provides a more detailed discussion of the rationales for a confiscatory taxation of accidental bequests; see also Kopczuk (2003). More recently Cremer et al. (2011) have pointed out that this result will not hold in a second-best setting where even accidental bequest may have informational content. See Michel and Pestieau (1998, 1999, 2000, 2004). See also Stiglitz (1977). Arrondel et al. (1991).

REFERENCES Altonji, J.G., F. Hayashi and L.J. Kotlikoff (1992), ‘Is the Extended Family Altruistically  Linked? Direct Tests Using Micro Data’, American Economic Review, 105(6), 1121–66. Andreoni, J. (1990), ‘Impure Altruism and Donations to Public Goods: A Theory of Warmglow Giving?’, Economic Journal, 100(401), 464–77. Atkinson, A.B. (1971), Capital Taxes, the Redistribution of Wealth and Individual Savings’, Review of Economic Studies, 38(2), 209–27. Atkinson, A.B. and A. Sandmo (1980), ‘Welfare Implications of the Taxation of Savings’, Economic Journal, 90(359), 529–49. Atkinson, A.B. and J.E. Stiglitz (1976), ‘The Design of Tax Structure: Direct Versus Indirect Taxation’, Journal of Public Economics, 6(1–2), 55–75. Auerbach, A. (2006), ‘The Future of Capital Income Taxation’, Fiscal Studies, 27(4), 399–420. Banks, J. and P. Diamond (2010), ‘The Base for Direct Taxation’, Reforming the Tax System for the 21st Century: The Mirrlees Review, vol. I, Oxford: OUP for IFS. Barro, R. (1974) ‘Are Government Bonds Net Wealth?’, Journal of Political Economy, 82(6), 1095–117. Bernheim, B.D., A. Shleifer and L.H. Summers (1985), ‘The Strategic Bequest Motive’, Journal of Political Economy, 93(6), 1045–76. Bevan, D.L. and J.E. Stiglitz (1979), ‘Intergenerational Transfers and Inequality’, Greek Economic Review, 1(1), 8–26. Blumkin, T. and E. Sadka (2004), ‘Estate Taxation with Intended and Accidental Bequests’, Journal of Public Economics, 88(1–2), 1–21. Boadway, R., E. Chamberlain and C. Emmerson (2010), ‘Taxation of Wealth and Wealth Transfers’, Reforming the Tax System for the 21st Century: The Mirrlees Review, vol. I, Oxford: OUP for IFS. Carroll, C.D. (2000), ‘Why do the Rich Save so Much?’, in J. Slemrod (ed.), Does Atlas Shrug? The Economic Consequences of Taxing the Rich, Cambridge, MA: Harvard University Press, pp. 463–85. Chamley, Ch. (1986), ‘Optimal Taxation of Capital Income in General Equilibrium with Infinite Lives’, Economica, 54(3), 607–22. Chamley, Ch. (2001), ‘Capital Income Taxation, Wealth Distribution and Borrowing Constraints’, Journal of Public Economics, 79(1), 55–69. Chari, V.V. and P.J. Kehoe (1999), ‘Optimal Fiscal and Monetary Policy’, in John B. Taylor and Michael Woodford (eds) Handbook of Macroeconomics, vol. 1, Part 3. Coleman, W.J. (2000), ‘Welfare and Optimum Dynamic Taxation of Consumption and Income’, Journal of Public Economic, 76(1), 1–39. Conesa, J., S. Kitao and D. Krueger (2009), ‘Taxing Capital? Not a Bad Idea After All!’, American Economic Review, 99(1), 25–48. Correia, I. (1996), ‘Should Capital Income be Taxed in a Steady State?’, Journal of Public Economics, 60(1), 147–51.

Wealth and wealth transfer taxation 205 Cox, D. (1987), ‘Motives for Private Income Transfers’, Journal of Political Economy, 95(3), 508–46. Cremer, H. (2003), ‘Multi-dimensional Heterogeneity and the Design of Tax Policy’, The Baltic Journal of Economics, 4(1), 35–45. Cremer, H. and F. Gahvari (1995), ‘Uncertainty, Optimal Taxation and the Direct Versus Indirect Tax Controversy’, Economic Journal, 105(432), 1165–79. Cremer, H. and P. Pestieau (1988), ‘A Case for Differential Inheritance Taxation’, Annales d’Economie et de Statistiques, 10(9), 167–82. Cremer, H. and P. Pestieau (1991), ‘Bequest, Filial Attention and Fertility’, Economica, 58(231), 359–75. Cremer, H. and P. Pestieau (1996), ‘Bequests as a Heir “Discipline Device”’, Journal of Population Economics, 9(4), 405–14. Cremer, H. and P. Pestieau (1998), ‘Delaying Inter Vivos Transmission under Asymmetric Information’, Southern Economic Journal, 65(2), 322–31. Cremer, H., F. Gahvari and P. Pestieau (2011), ‘Accidental Bequests: A Curse for the Rich and a Boon for the Poor’, unpublished. Cremer, H., D. Kessler and P. Pestieau (1993), ‘Education for Attention: A Nash Bargaining Solution to the Bequest-as-exchange Model’, Public Finance, 48(supplement), 85–97. Cremer, H., J.-M. Lozachmeur and P. Pestieau (2010), ‘Collective Annuities and Redistribution’, Journal of Public Economic Theory, 12(1), 23–41. Cremer, H., P. Pestieau and J.-C. Rochet (2003), ‘Capital Income Taxation When Inherited Wealth is Not Observable’, Journal of Public Economics, 87, 2475–90. Diamond, P. (1965), ‘National Debt in a Neoclassical Growth Model’, American Economic Review, 58(1), 1126–50. Erosa, A. and M. Gervais (2002), ‘Optimal Taxation in Life Cycle Economies’, Journal of Economic Theory, 105(2), 338–69. Farhi, E. and I. Werning (2010), ‘Progressive Estate Taxation’, Quarterly Journal of Economics, 125(2), 635–73. Fatemi, D., J. Hasseldine and P. Hite (2008), ‘Resisting Framing Effects: The Importance of Prior Attitude on Estate Tax Preference’, Journal of the American Accounting Association, 30(1), 101–21. Frank, R. (2005), ‘The Estate Tax: Efficient, Fair and Misunderstood’, The New York Times, 12 May. Glomm, G. and R. Ravikunar (1992), ‘Public Versus Private Investment in Human Capital: Endogenous Growth and Income Inequality’, Journal of Political Economy, 100(4), 818–34. Golosov, M., N. Kocherlakota and A. Tsyvinski (2003), ‘Optimal Indirect and Capital Taxation’, Review of Economic Studies, 70(3), 569–87. Golosov, M., A. Tsyvinski and I. Werning (2007), ‘New Dynamic Public Finance: A User’s Guide’, NBER Macroeconomics Annual 2006, 317–63. Hammond, P. (1988), ‘Altruism’, in J. Eatwell, M. Milgate and P. Newman (eds) New Palgrave: A Dictionary of Economics, London: Macmillan Press, pp. 85–7. Harsanyi, J. (1995), ‘A Theory of Prudential Values and a Rule Utilitarian Theory of Morality’, Social Choice and Welfare, 12(4), 319–44. Jacobs, B. and L. Bovenberg (2004), ‘Optimal Dual Income Taxation with Endogenous Human Capital’, mimeo. Jones, Larry E., Rodolfo E. Manuelli and Peter E. Rossi (1997), ‘On the Optimal Taxation of Capital Income’, Journal of Economic Theory, 73(1) 93–117. Judd, K.L. (1985), ‘Redistributive Taxation in a Simple Perfect Foresight Model’, Journal of Public Economics, 28(1), 59–83. Kaplow, L. (2000), ‘A Framework for Assessing Estate and Gift Taxation’, in W.G. Gale, J.R. Hines and J. Slemrod (eds) Rethinking Estate and Gift Taxation, Washington DC: Brookings Institution. Kaplow, L. (2006), ‘On the Undesirability of Commodity Taxation Even When Income Taxation Is Not Optimal’, Journal of Public Economics, 90(6–7), 1235–50.

206

The Elgar guide to tax systems

Kaplow, L. (2008), The Theory of Taxation and Public Economics, Princeton, NJ: Princeton University Press. Kopczuk, W. (2003), ‘The Trick is to Live. Is the Estate Tax Social Security for the Rich?’, Journal of Political Economy, 111(6), 1318–41. Kotlikoff, L.J. and A. Spivak (1981), ‘The Family as an Incomplete Annuities Market’, Journal of Political Economy, 89(2), 372–91. Kurz, M. (1968), ‘Optimal Economic Growth and Wealth Effects’, International Economic Review, 9(3), 348–57. Laroque, G. (2005), ‘Indirect Taxation is Superfluous under Separability and Taste Homogeneity: A Simple Proof ’, Economics Letters, 87(1), 141–4. Mankiw, G. (2000), ‘The Savers-Spenders Theory of Fiscal Policy’, AEA Papers and Proceedings, 90(2), 120–25. Michel, Ph. and P. Pestieau (1998), ‘Fiscal Policy in a Growth Model with Both Altruistic and Non-altruistic Agents’, Southern Economic Journal, 64(3), 682–97. Michel, Ph. and P. Pestieau (1999), ‘Fiscal Policy in a Growth Model Where Individuals Differ with Regard to Altruism and Labor Supply’, Journal of Public Economic Theory, 1(2), 187–203. Michel, Ph. and P. Pestieau (2000), ‘Tax-transfer Policy with Altruists and Non-altruists’, in L.A. Gérard-Varet, S.C. Kolm and J. Mercier Ythier (eds) The Economics of Reciprocity, Giving and Altruism, London: Macmillan, pp. 275–84. Michel, Ph. and P. Pestieau (2002), ‘Wealth Transfer Taxation with Both Accidental and Desired Bequests’, unpublished. Michel, Ph. and P. Pestieau (2004), ‘Fiscal Policy with Agents Differing in Altruism and in Ability’, Economica, 72(285), 121–36. OECD (2008), Revenue Statistics 1965–2006, Paris: OECD. Pestieau, P. (1974), ‘Optimal Taxation and Discount Rate for Public Investment in a Growth Setting’, Journal of Public Economics, 3(3), 217–35. Pestieau, P. (2000), ‘Gifts, Wills and Inheritance Law’, in B. Bouckaert and G. De Geest (eds) Encyclopedia of Law and Economics, vol. 3, Cheltenham, UK and Northampton, MA, USA: Edward Edgar, pp. 888–906. Pestieau, P. and M. Sato (2008), ‘Estate Taxation with Both Accidental and Planned Bequests’, Asia Pacific Journal of Accounting and Economics, 15(3), 223–40. Pestieau, P. and E. Thibault (2007), The Spenders-Hoarders Theory of Capital Accumulation, Wealth Distribution and Fiscal Policy, CORE DP 2007/40. Pestieau, P. and E. Thibault (2011), ‘Love Thy Children or Money – Reflections on Debt Neutrality and Estate Taxation’, Economic Theory (forthcoming). Ramsey, F.P. (1928), ‘A Mathematical Theory of Saving’, Economic Journal, 38(152), 543–59. Saez, E. (2002), ‘Optimal Progressive Capital Income Taxes in the Infinite Horizon Model’, NBER Working Paper No. 9046. Stiglitz, J.E. (1977) ‘Equality, Taxation and Inheritance’, in W. Krelle and A. Shorrocks (eds) Personal Income Distribution, Amsterdam: North-Holland, pp. 271–303. Stiglitz, J.E. (1978), ‘Notes on Estate Taxes, Redistribution and the Concept of Balanced Growth Path Incidence’, Journal of Political Economy, 86(2), 137–50. Weber, M. (1958), The Protestant Ethic and the Spirit of Capitalism, New York: Charles Scribner’s Sons.

Wealth and wealth transfer taxation 207

APPENDIX A Table 5A.1

Taxes on property as a percentage of GDP 1965 1970 1975 1980 1985 1990 1995 2000 2005 2006

Belgium Denmark Germany Ireland Greece Spain France Italy Luxembourg Netherlands Austria Portugal Finland Sweden United Kingdom United States Japan

Table 5A.2

1.2 2.4 1.8 3.8 1.7 0.9 1.5 1.8 1.7 1.4 1.3 0.8 1.2 0.6 4.4

1.3 2.3 1.6 3.5 1.8 1 1.6 1.5 1.6 1.2 1.3 0.7 0.7 0.5 4.6

1.1 2.3 1.2 2.8 1.9 1.2 1.8 0.8 1.7 1 1.1 0.5 0.7 0.5 4.5

1.2 2.5 1.3 1.6 1 1 1.9 1.1 2 1.5 1.1 0.3 0.7 0.4 4.2

1.1 2 1.1 1.4 0.7 1.6 2.5 0.8 2.2 1.5 1 0.5 1.1 1.1 4.5

1.4 1.9 1.2 1.5 1.2 1.8 2.7 0.9 3 1.6 1.1 0.8 1.1 1.8 2.9

1.5 1.7 1 1.5 1.2 1.8 2.9 2.3 2.6 1.7 0.6 0.8 1 1.3 3.5

1.9 1.6 0.8 1.7 2.1 2.2 3.1 2 4.1 2.1 0.6 1.1 1.1 1.8 4.3

2.1 1.9 0.9 2.4 1.4 3.1 3.4 2 3.2 2.1 0.6 1.1 1.2 1.5 4.4

2.3 1.9 0.9 2.9 1.4 3.3 3.5 2.1 3.3 1.9 0.6 1.1 1.1 1.4 4.6

3.9 1.5

3.8 1.5

3.6 1.9

2.8 2.1

2.7 2.7

3.1 2.7

3.1 3.3

3 2.8

3.1 2.6

3.1 2.5

Taxes on property as a percentage of total taxation 1965 1970 1975 1980 1985 1990 1995 2000 2005 2006

Belgium Denmark Germany Ireland Greece Spain France Italy Luxembourg Netherlands Austria Portugal Finland Sweden United Kingdom United States Japan

3.7 8 5.8 15.1 9.7 6.4 4.3 7.2 6.2 4.4 4 5 4 1.8 14.5

3.8 6 4.9 12.2 9.3 6.5 4.8 6 6.7 3.3 3.7 4 2.2 1.5 12.5

2.9 6.1 3.9 9.7 9.7 6.3 5.1 3.3 5.2 2.4 3.1 2.5 1.9 1.1 12.7

2.9 5.8 3.3 5.3 4.6 4.6 4.8 3.7 5.7 3.6 2.9 1.4 1.9 0.9 12

2.5 4.3 3 4 2.7 5.9 5.8 2.5 5.6 3.5 2.4 1.9 2.7 2.3 12

3.4 4.2 3.4 4.7 4.6 5.5 6.3 2.3 8.3 3.7 2.7 2.7 2.4 3.5 8.2

3.4 3.5 2.8 4.5 4.1 5.5 6.7 5.6 7 4.1 1.5 2.5 2.2 2.7 10

4.2 3.2 2.3 5.5 6.2 6.5 7 4.6 10.6 5.3 1.3 3.2 2.4 3.4 11.6

4.7 3.7 2.5 7.9 4.3 8.7 7.8 5 8.5 5.3 1.3 3.1 2.7 3 12.1

5.1 3.8 2.5 9.1 4.4 9 8 5.1 9.3 4.7 1.4 3.1 2.5 3 12.4

15.9 8.1

14.2 7.6

13.9 9.1

10.7 8.2

10.7 9.7

11.5 9.4

11.1 12.2

10.1 10.5

11.3 9.7

11.1 9.1

208

The Elgar guide to tax systems

APPENDIX B DETAILS ON THE MODEL In this appendix we sketch the models that underly the discussion presented in this survey. We start by the basic overlapping generations model with production developed by Diamond (1965). Overlapping Generations Model In the Diamond (1965) model each generation lives for two periods, consuming in both and working in the first. There are no bequests and the lifetime budget constraint for the representative household born in period t may be written: ct 1

dt11 5 wt,t. 1 1 rt11

(5A.1)

It is clear that endowing the government with two instruments, taxes on labor income (tw 5 w 2 w) and capital income (tr 5 r 2 r) is equivalent to allowing the government to tax first- and second-period consumption at possibly different rates. A zero-tax on capital income – a labor income tax – would result in uniform taxation of consumption in the two periods.1 We now characterize the optimal steady-state taxes resulting from a utilitarian objective t

a d ut

(5A.2)

where 0 , d , 1 is the factor of social time preference and ut 5 u (ct, dt11, ,t)

(5A.3)

is the individual utility function. Two general results have been obtained. First, with the government able to redistribute resources across generations through debt policy, pay-as-you-go Social Security, or any other devices the marginal product of capital converges to the population growth rate divided by the factor of time preference ((1 1 n) /d) , namely the modified golden rule. Second, optimal taxes on labor and capital should follow the standard analysis of static optimal tax theory. Maximizing equation (5A.3) subject to equation (5.) (see Section 5 above) yields the demand function for c (wt, rt11) , d (wt, rt11) and , (wt, rt11) , which substituted back in the utility function yields the indirect utility function:

Wealth and wealth transfer taxation 209 vt 5 v (wt, rt11) , with 0vt atdt11 atst 0vt 5 at,t 5 5 , (1 1 rt11) 2 0wt 0rt11 1 1 rt11 where a is the marginal utility of income a 5 0u/0I and s is saving. We use I to denote nonlabor income, if any. There is a production sector represented by a CRS (constant rehins to scale) production function relating output Yt to capital Kt and labor Lt: Yt 5 F (Kt, Lt) , or yt 5 F a

Kt , 1b 5 f (kt) , Lt

with y 5 Y/L and k 5 K/L. With perfect competition factor payments equal to the value of marginal products: wt 5 FrL (Kt, Lt) and 1 1 rt 5 FrK (Kt, Lt) . We assume total depreciation after one period and Lt 5 ,tNt where Nt 5 Nt21 (1 1 n) is the size of generation t. In this simple economy, the dynamics is conducted by the capital accumulation equation: Kt11 5 Ntst, where st 5 s (wt, rt11) 5 wt 2 c (wt, rt11) . Under some assumptions, one can show that kt11 converges to a unique steady-state market equilibrium k*, which can be compared to the steadystate optimal value kˆd that is consistent with the modified golden rule and defined by: 11n fr (kˆd) 5 . d For the time being we assume that the economy is on the modified golden rule growth path through some appropriate intergenerational transfers by

210

The Elgar guide to tax systems

the government. So doing we focus on the optimal tax structure abstracting from dynamic efficiency considerations. The government’s budget constraint is simply: twt,t 1 trt

dt 5 R, (1 1 rt) (1 1 n)

(5A.4)

where R is given. The second term on the left is the revenue from capital income taxation that concerns the previous generation (st21 5 dt/ (1 1 rt)) . We solve this problem by differentiating the Lagrangean expression, dt (wt21, rt) 2 Rb f , 5 a dt e v (wt, rt11) 1 matwt,t (wt, rt11) 1 trt (1 1 r ) (1 1 n) t

with respect to wt and rt.This yields: 0,t 0dt11 0 d 5 dt aat,t 1 m c twt 2 ,t 1 trt db 0wt 0wt 0wt (1 1 n) (1 1 rt11)

(5A.5)

dt11 0 5 dt aat (1 1 rt11) 2 0rt11 1 m c twt

0,t 0dt11 1 dt11 (1 1 rt11) d 1 2 b d b. (5A.6) atrt11 (1 1 rt) 2 0rt11 11n 0rt11 1 1 rt

Evaluating (5A.5) and (5A.6) in the steady state, while adding and subtracting the income effect times , for 0/0w and times d/ (1 1 r) 2 for 0/0r yields: | | a 0, 0d d 5 0 (5A.7) 1 tr a 2 1 2 Db, 1 tw m 0w 0w (1 1 n) (1 1 r) a

a d 1 1 n 2 d (1 1 r) d 2 1 2 Db 2 (1 1 r) 2 (1 1 rt) 2 m 11n 1w

| | 0, 0d d 5 0, 1 tr 0r 0r (1 1 n) (1 1 r)

where D 5 tw

0, 0d d , 1 tr 0I 0I (1 1 n) (1 1 r)

(5A.8)

Wealth and wealth transfer taxation 211 and the | denotes the compensated effects. Given our assumption on the modified golden rule, this can be further simplified: | | 0, 0d d r 1t t 0w 0w (1 1 n) (1 1 r) w

| | , (1 1 r) 2 0, 0d d r 5 at 1t . b 0r 0r (1 1 n) (1 1 r) d w

(5A.9)

This equation characterizes the relative levels of the tax rates on earnings and capital income with the absolute levels being determined by the government’s revenue requirement R. As usual this characterization depends on compensated and not gross derivatives. Assume for simplicity of interpretation that the cross-effects are zero. Then we can have: | e dr 1 1 r tw /w 5| r t /r e ,w r (1 1 r)

(5A.10)

where the | e are the compensated elasticities. If labor is completely inelastic along the compensated supply curve, the optimal tax on interest income is zero because the tax on earnings is equivalent to a lump-sum tax. The argument is reversed when the demand for future consumption is inelastic. In general, however, there is no particular reason to believe that either tax will be zero nor that both taxes are the same. Let us come back to the assumption that the economy is on the modified golden rule path, that is, on the assumption that the government can control capital. From (5A.8) one can see that if 1 1 n 2 (1 1 r) d we have an additional term in either (5A.9) or (5A.10). In other words these taxes are not only used to finance R but also to foster or discourage capital accumulation depending on whether the rate of interest is higher or lower than the rate of population growth divided by the discount factor. As shown by Atkinson and Sandmo (1980) too little capital may call for a lower taxation of earnings and a higher tax on interest income than when the modified golden rule holds. This apparent paradox can be explained by noting that with a log-linear utility function saving depends only on earnings and not on the interest rate.2 We shall now introduce transfers into this model and successively consider the motives discussed in Section 3. Within each setting we study the design of factor income and wealth transfer taxes. To do so it is convenient to distinguish the case where the government has the instruments to secure the modified golden rule from the case where the government cannot fully control the capital stock.

212

The Elgar guide to tax systems

Accidental Bequest The accidental bequest case is not much different from the case without bequest. Saving is affected by survival probabilities. Accidental transfers are taxed at 100 percent, without affecting the supply of saving. The part of public spending (if any) that exceeds the proceeds of the transfer tax is financed through labor and capital income taxes designed à la Atkinson-Sandmo. Pure Altruism3 To keep things relatively simple, we assume that b 5 0 so that d 5 0. In other words, people live only one period and only save for bequeathing. This assumption implies that the tax on saving is also the tax on wealth  transfer.4 Then, the social planner’s problem at time 0 is to maximize: `

t a g u (ct, ,t) , t50

subject to the resource constraint: F (kt, ,t) 5 (1 1 n) kt11 1 ct 1 R, and to the revenue constraint: (1 1 n) zt11 5 (1 1 rt) zt 1 (1 1 rt) kt 1 wt,t 2 F (kt, ,t) 1 R, where z denotes per worker public debt. Recall that k is the per worker capital stock while R per worker public spending and that the production function exhibits constant returns to scale. Judd (1985), Chamley (1986) and Coleman (2000) show the following: ● ● ●

If one could tax initial wealth k0 as much as possible, one could do without using any distortionary tax. If this first-best solution is not accessible, one will have initially a tax on both earnings and saving (that is, bequests). In the long run the tax on saving tends to 0.

We restrict ourselves to proving the last point, which represents the main result. The government’s objective is the same as that of the representative individual (g 5 d) . It maximizes the Lagrangean:

Wealth and wealth transfer taxation 213 `

5 a gt [ u (ct, ,t) 1 lt (F (kt, ,t) 2 ct 2 (1 1 n) kt11 2 R) ] t50

1 mt [ (1 1 n) zt11 2 (1 1 rt) zt 2 (1 1 rt) kt 2 wt,t 1 F (kt, ,t) 2 R ] , where l and m are the Lagrange multipliers associated with the resource and the revenue constraint respectively. The FOC (first-order condition) with respect to z and k in the steady state are: (1 1 r) g 5 1 1 n,

(5A.11)

2 (1 1 n) l 1 gl (1 1 r) 1 mg (r 2 r) 5 0.

(5A.12)

and

Combining these two equations gives: 2 l (1 1 r) 1 l (1 1 r) 1 m (r 2 r) 5 0. This yields (l 1 m) (r 2 r) 5 0 and thus tr 5 0, so that (5A.11) implies (1 1 r) g 5 1 1 n. In words, we have the modified golden rule and most notably, a zero tax on savings that correspond to bequests in our setting. Consequently, wealth transfers are not taxed in the steady state.5 Chamley-Judd’s result has become the standard rule for a number of public economists and particularly macroeconomists. However, it has also been challenged on various grounds. It relies on a set of strong assumption that have been questioned. In any case the zero tax result only applies to the steady state; during the transition period, wealth transfers along with capital income are subject to taxation. Joy of Giving Unlike in the case of pure altruism, the objective of individuals and that of the social planner may now diverge. Each individual maximizes: u (ct, dt11, ,t) 1 gv (xt11) , subject to xt 1 wt,t 5 ct 1

dt11 1 (1 1 n) (1 1 tx) xt11 . 1 1 rt11

214

The Elgar guide to tax systems

In a laissez-faire equilibrium, each individual chooses ,t, ct, dt11 and xt11 given factor prices wt and rt and inherited wealth xt. As to the social optimum, one faces the issue of whether or not to launder individual utilities. Harsanyi (1995) and Hammond (1988) have advocated ‘excluding all external preferences, even benevolent ones, from our social utility function’ (Hammand, 1988, p. 87). Advocates of a utilitarian approach, on the other hand, argue that the social planner cannot paternalistically modify individuals’ preferences. We shall use a generalized objective that admits the two approaches as a special case. Denoting the social factor of time preference by d, social welfare is given by: `

Ut 5 a ds [ u (cs,ds11,,s) 1 egv (xs11) ] , s51

where 0 # e # 1 with e 5 0 for the non utilitarian and e 5 1 for the utilitarian case. With this setting, the steady-state rule of optimal capital accumulation is the modified golden rule. The key issue is the treatment of xt. For e 5 1 the first-best optimal value of x is that for which vr (x) 5 0. In other words without laundering out utilities the social planner will push for a very high value of x (that could be infinity). In a first-best world, such a solution could be implemented through a subsidy on x financed by public debt. It is clearly not reasonable and such a pathological outcome provides an argument in favor of laundering out the joy of giving from the donors’ welfare. In the second-best world, with linear taxes on earnings, capital income, and bequests, the revenue constraint is given by: R 5 twt,t 1 trtst21 1 txt (1 1 n) xt, which can also be written as: R 5 twt,t 1 tt

dt 1 qxt (1 1 n) xt, 1 1 ,t

where qxt 5

trt (1 1 txt) 1 txt 1 1 rt

is the total (or effective) tax on transfers. Observe that bequests are subject to a double tax: first, the tax on savings, tr, and then the specific tax on transfers tx. The total tax on bequest is higher than that on second period consumption if qx . tr/ (1 1 rt) , which occurs when tx . 0.

Wealth and wealth transfer taxation 215 Michel and Pestieau (2002) show that with no laundering the tax structure is not much different from (5A.9). Taxes on earnings, on second-period consumption and on bequests only depend on compensated elasticities and on the revenue requirement when the capital stock is directly controlled. In the case of zero cross-elasticities, the tax on secondperiod consumption (tr) may be higher than the estate tax (qx) if the own compensated elasticity of second-period consumption is lower than that of bequests. When there is laundering, bequest loses its direct social utility and is thus subject to a relatively higher tax. Preference for Wealth Formally as seen above the case of preference for wealth and that of joy of giving are very similar. The only difference is that individuals who have a preference for wealth are interested by their entire life-cycle saving: dt11 1 xt11 Exchange We will use an exchange model of the strategic type in which parents obtain attention from their children in exchange for some bequests. By playing their children against each other they control the exchange to their full benefit.6 The utility function of an individual belonging to generation t is given by: u (ct 2 v (agt) , dt11, ,t, at11) ,

(5A.13)

where at11 denotes attention received and agt attention given that requires some effort. The disutility of attention given is expressed in monetary terms. First- and second-period budget constraints are: wt,t 1 bt 5 ct 1 st,

(5A.14)

(1 1 rt11) st 5 (1 1 txt11) xt11 1 dt11.

(5A.15)

In addition, we have xt11 5 (1 1 n) bt11

(5A.16)

v (agt) 5 bt.

(5A.17)

and

216

The Elgar guide to tax systems

Equation (5A.16) gives the straightforward relation between bequest and inherited wealth. Equation (5A.17) results from our strategic bequest assumption: parents extract all the surplus from their children who are just paid for the disutility of their effort. Substituting 5A.14–5A.17 into (5A.13) shows that each member of generation t maximizes the following expression: uawt,t 2

(dt11) v (at11) (1 2 txt11) 2 , dt11, ,t, at11 b. 1 1 rt11 1 1 rt11

The indirect utility is given by: Vt 5 V (wt, rt11, txt11) . The problem for the social planner is to maximize the discounted sum of utilities, g dtVt, subject to the revenue constraint: R 5 tw, 1

trt 1 txt (1 1 rt) trdt 1 v (at) . (1 1 rt) (1 1 n) (1 1 rt) (1 1 n)

We continue to assume that capital accumulation is socially optimal (i.e., 1 1 r 5 (1 1 n) /d). The FOC in the steady state can be written as: | 0, t 0 tw w

1

| | 0d 0a tr tr 1 tx (1 1 r) vr (a) w 1 w (1 1 r) (1 1 r) 0 t (1 1 r) (1 1 r) 0t 1a

tw

a 2 1 2 Db, 5 0 m

| | | 0a 0d 0, tr tr 1 tx (1 1 r) ( ) vr a 1 1 (1 1 r) (1 1 r) 0 tr (1 1 r) (1 1 r) 0 tr 0 tr 1a

a d 2 1 2 Db 50 (1 1 r) 2 m

| | | 0d 0a 0, tr tr 1 tx (1 1 r) vr (a) x t 1 1 x x (1 1 r) (1 1 r) 0 t (1 1 r) (1 1 r) 0t 0t w

1a

a v (a) 2 1 2 Db 50 m 11r

Wealth and wealth transfer taxation 217 For the same reasons as developed above (‘Joy of Giving’), the overall tax on bequests, tr 1 tx (1 1 r) , may or may not be higher than that on future consumption. In other words, there is no particular reason to believe that the wealth transfer tax tx is positive. This will depend on the relative magnitude of the compensated derivatives that determine the overall tax on bequests and the tax on future consumption through Atkinson and Sandmo-type rules. To illustrate this point in the simplest possible way, assume again that the cross-elasticities are zero. Then, we have: tr 1 tx (1 1 r) 5 tr

| 0d (1 1 r) 0tr . | 0a vr (a) x d 0t

v (a)

Clearly if the demand for attention is much more elastic than that for future consumption, the tax on inheritance, tx, is negative.

NOTES 1. 2. 3. 4. 5.

See Pestieau (1974), Atkinson and Sandmo (1980). Naturally, their argument applies also to other utility functions. The classical papers on this are Judd (1985) and Chamley (1986). We have the following equality between saving and bequest: st 5 xt 11 5 (1 1 n) kt 11. This result generalizes to the case where b . 0 and d . 0. However, the proof becomes much more complicated. 6. We exclude collusion between children whereby they would agree to supply a minimal amount of attention and share the inheritance.

6

Value-added tax: onward and upward? Jorge Martinez-Vazquez and Richard M. Bird*

1

INTRODUCTION

The most important tax development of the last half-century has undoubtedly been the rise to prominence of the value-added tax (VAT).1 This tax has taken center stage almost everywhere (with the significant exception of the United States) and has become a revenue mainstay for many countries. The success of the VAT reflects a variety of factors: its high revenue potential, its relative simplicity and logic from an administrative perspective, its impact on economic efficiency, trade, and growth, the ease with which its relatively mild consequences on income distribution and equity may be mitigated, and the fact that fewer and relatively less complex political economy issues than often arise with respect to other potential revenue-producing taxes seem to afflict its introduction and development. After taking place for decades almost out of sight of the academic world, more recently these properties of the VAT – and of course the dramatic expansion of its use in rich and poor countries alike over the years – have begun to attract considerable attention in the tax policy and economics literature.2 We do not attempt to cover all developments in this chapter. Our objective is considerably more modest. Following a brief review of the spread of VAT in the next section, the major new contribution of this chapter is the detailed analysis in Sections 3 and 4 of how well the VAT has actually performed from several perspectives. In the balance of the chapter, we turn to a much briefer consideration of several important policy issues that arise with VAT, concluding with a closer look at some administrative challenges, in particular the issue of fraud. Although some in Europe have argued that this may prove to be the Achilles’ heel of the VAT, we conclude that, on the whole, those fears are as overstated as the more fundamental objections to VAT frequently voiced in the United States and increasingly beginning to appear, in more sophisticated language, in some recent academic literature.3

218

Value-added tax 219 140 120 100 80 60 40 20

Source:

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

1966

Ebril et al (2001) and various sources.

Figure 6.1

2

1964

1962

1960

0

Number of countries with the VAT

THE RISE OF THE VAT

Figure 6.1 tells the story: beginning with a few pioneers in the 1960s, the VAT began to spread around the world. By 1992 about 80 countries had adopted it; by 2008, at least 136 countries had a VAT. Since then, still more countries have joined the VAT ranks: in 2009, for example, Laos adopted a VAT and Pakistan is now considering doing the same. Interestingly, in some countries VATs are also now becoming important sub-national taxes, notably in India and Canada. At the national level, however, with the exception of some countries in the Middle East – where the Gulf Cooperation Council has for some time been considering adopting a VAT – and a few small islands, almost every country, with the notable exception of the United States, has now introduced some form of VAT. Norregaard and Khan (2007) provide a useful review of the main factors often said to have led to the popularity of the VAT. For example, from both an administrative and an economic perspective, VAT may be considered a particularly efficient way to tax consumption because revenues are collected throughout the chain of production. As a result, the revenues are likely to be more secure than in the case of another form of tax that reaches approximately the same base – the retail sales tax (RST). RSTs – now found only in most US states and a few Canadian provinces – have the unfortunate characteristic that all revenue is lost if there is evasion at the final retail stage, and it is precisely this stage that is most difficult to

220

The Elgar guide to tax systems

control. Moreover, in principle VAT is considerably more economically efficient than any form of pre-retail stage sales taxes (such as turnover taxes or taxes imposed at the manufacturing or wholesale level) because unlike those taxes VAT does not result in cascading and distortions of production choices. Yet another positive attribute of adopting a VAT in many countries, particularly poorer countries, is that doing so has often proved to be a catalyst for reforming tax administration more generally, for example by extending the self-assessment method usually considered integral to VAT to income and perhaps other taxes. Of course, each of the ‘silver linings’ associated with VAT is itself accompanied by a possible cloudy outcome. For example, if the crediting chain of the VAT is broken, not only may production decisions be distorted but the final incidence of the tax becomes uncertain. Similarly the prospective efficiency gains associated with VAT may be achieved in many cases only by incurring substantial administrative and compliance costs and may sometimes be accompanied by possibly adverse distributional consequences. Finally, since refunding input taxes borne by outputs is an integral feature of a VAT not only are any weaknesses in the administrative system put under more strain but failures become much more obvious. In a rigorous recent analysis, Keen and Lockwood (2010) find that a country is less likely to adopt a VAT the larger its agricultural sector and the more open its economy. These results are not unexpected: agriculture is notoriously ‘hard-to-tax’ under any regime, while on the other hand international trade provides an attractive and accessible alternative tax base. Interestingly, countries with lower past revenue to GDP ratios – although they might be presumed to have a greater ‘need’ for revenues – were also found to be less likely to adopt a VAT. On the other hand, given the role IMF advice and assistance has played in spreading the VAT gospel around the world, it is not surprising that Keen and Lockwood (2010) find that a country’s participation in a non-crisis IMF program also increases the likelihood of VAT adoption, as does the ‘yardstick competition’ effect of having neighbors that have already done so. Perhaps the most striking result of this recent study, however, is that income per capita is no longer a significant determinant for VAT take-up: rich or poor, almost everyone has now taken the VAT leap. Not only do VATs now exist in most countries, but they provide a sizable share of revenues in many of them. The picture is different in developed and developing countries, however. As Figure 6.2 shows, for example, consumption taxes have long been the second key pillar of the tax systems of developed countries, following income taxes and in turn being followed by Social Security taxes as the third pillar.4 This figure masks the growing importance of VAT, however, because the composition of consumption taxes in developed countries changed drastically over this period. In 1965,

Value-added tax 221 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05

Income Taxes Property Taxes

Payroll Tax Customs

Social Security Contributions

2005

2004

2003

2002

2001

2000

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

0

Consumption Taxes

Other Taxes

Note: a. General government data. Based on a sample of 32 developed countries (number of countries in the sample varies across years); (note: changes in GFS methodology after 1990): for 1990–99 tax data not available. Sources:

IMF GFS Database and World Bank World Development indicators.

Figure 6.2

Average annual tax structure as a share of total taxes in developed countries,a 1972–2005

for example, almost two-thirds of consumption tax revenue in OECD countries came from excises and other specific consumption taxes; by 2003, this share had fallen to little more than one-third.5 The growing importance of the VAT is even more obvious in developing countries, as shown in Figure 6.3. In recent decades, trade liberalization has resulted in a substantial decline in the once dominant customs duties, which are now more or less tied with payroll taxes for third place. However, the increase in VAT revenues has more than compensated for this decline to the point where, as Figure 6.3 shows, consumption taxes are now significantly more important in revenue terms in many developing countries than income taxes, and in most the growth is almost entirely attributable to VAT. One reason revenues may increase is because rates do. In some countries, VAT rates have indeed risen since VAT was first introduced. On the whole, however, as Table 6.1 shows, although the average standard rate has increased slightly over the last two decades it has not moved much from the range of 16 to 17 percent. Of course, there are more noticeable trends among different groups of countries. For example, while the average standard rate has decreased sharply in transitional countries, it has trended upward, especially over the last decade, in both developing and developed countries, especially the latter, as shown in Figure 6.4.

222

The Elgar guide to tax systems

0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05

Income Taxes Property Taxes

Payroll Tax Customs

Social Security Contributions

2005

2004

2003

2002

2001

2000

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

0

Consumption Taxes

Other Taxes

Note: a. General government data. Based on a sample of 75 developing countries (number of countries in the sample varies across years); (note: change in GFS methodology after 1990); for 1990–99 tax data not available. Sources: IMF GF5 Database and World Bank World Development indicators.

Figure 6.3

Average annual tax structure as a share of total taxes in developing countries,a 1972–2005

Tax theory definitely lends little or no support to a uniform consumption tax rate as ideal. Nonetheless, in order to reduce compliance and administrative costs and improve economic efficiency, most expert advice has long recommended a single standard VAT rate, with zero rating only for exports. Perhaps reflecting greater susceptibility to outside advice, many developing countries have in fact adopted such systems. However, on the whole, most developed countries impose more than one rate, often largely on equity grounds (Norregaard and Khan, 2007). This outcome is a bit curious, given that the equity case for favorable rates for certain items is presumably stronger in developing countries. Nonetheless, such arguments continue to resonate in many developed countries also. For example, a recent report from Copenhagen Economics (2007), although arguing for uniform VAT as a superior approach, notes four possibly persuasive arguments for reduced VAT rates in certain circumstances. Two of these arguments – that a reduced VAT may increase efficiency by increasing productivity or by reducing structural unemployment – are efficiency-based; the other two – that a reduced VAT may enhance equity by improving the income distribution or by making particular products more accessible to the entire population – take the more traditional equity route. Yet another argument for differential rates that has recently found

Value-added tax 223 Table 6.1

Average standard VAT rates in developed, developing, and countries in transition (based on IMF classification) Advanced Countries

Developing Countries

Countries in Transition

Year

Mean Min

Max Mean Min

Max Mean Min

Max

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

16.93 16.81 16.91 16.96 16.70 16.30 16.44 16.56 16.77 16.73 16.49 16.62 16.62 16.80 17.02 17.14 17.07 17.24 17.24

24.50 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00

22.00 35.00 35.00 23.00 23.00 23.00 23.00 23.00 23.00 23.00 23.00 23.00 23.00 23.00 25.00 23.00 23.00 22.00 22.00

25.00 25.00 30.00 30.00 28.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 25.00 20.00 20.00 20.00 20.00 20.00

3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 4.00 5.00 5.00 5.00 5.00 5.00

12.26 13.36 13.30 13.46 13.37 13.80 14.07 14.26 14.38 14.02 14.46 14.46 14.38 14.65 14.93 14.97 14.89 14.84 14.83

5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00 3.00 2.00 2.00 2.00 2.00 5.00 5.00 5.00 5.00 5.00

25.00 25.00 24.39 20.71 19.74 19.35 19.10 19.15 19.08 18.78 18.46 18.48 18.48 18.53 17.95 17.61 17.54 17.29 17.29

25.00 25.00 5.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00

Unweighted Average

14.79 15.01 16.88 16.21 15.79 15.76 15.87 15.96 16.05 15.72 15.82 15.84 15.78 15.96 15.89 15.89 15.84 15.79 15.79

favor with some is the desirability of shifting taxes, including VAT, off ‘green’ goods and onto ‘environmentally damaging’ ones.

3

VAT PERFORMANCE

VATs are now found almost everywhere and have become important components of revenue systems in most countries. But how well have VATs actually performed? We consider several aspects of this question in this and the next section. Although some fundamental features (use of the invoice-credit method, self-assessment, the destination principle, etc.) are common to most VATs, other important features of VAT structure (rates, exemptions, use of zero rating, etc.) differ sharply from country to country as of course does the efficiency and effectiveness with which the tax is implemented. Unsurprisingly, therefore, VATs differ greatly across countries with respect to how productive they are in raising revenues.

224

The Elgar guide to tax systems

26.00 24.00 22.00 20.00 18.00 16.00 14.00 12.00

All countries

Developed

Developing

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

10.00

Transition

Source: Various sources for the VAT rates.

Figure 6.4

Average standard VAT rates in developed, developing, and countries in transition, 1990–2008

Three related measures of VAT productivity may be found in the literature: (1) the VAT efficiency ratio; (2) the C-efficiency measure; and (3) what has sometimes been called the VAT gross collection measure. All three measures calculate efficiency as the ratio of actual VAT collections in the country to the potential revenues that would be derived from applying the standard VAT rate to, respectively, three potential tax bases: GDP, total consumption expenditure, and private consumption expenditure. The formulas for these three measures can be written as: VAT Efficiency Ratio 5

VAT Revenue Collection Standard VAT Rate * GDP

C 2 Efficiency Ratio 5

VAT Revenue Collection Standard VAT Rate * Total Consumption Expenditure

VAT Gross Collection Ratio 5

VAT Revenue Collection Standard VAT Rate * Private Consumption Expenditure

In principle, a VAT with no exemptions, a single rate, and full compliance should result in efficiency ratios of close to 100 percent (IMF, 2010). In practice, of course, many VATs are very far from this possible goal. In fact, none of these measures gets close to the actual tax base on which the

Value-added tax 225 VAT falls, even though given the exclusion of investment and most public consumption from the VAT base in most countries, the VAT gross collection measure should be closer than the C-efficiency ratio and definitely closer than the VAT efficiency ratio, based as it is on GDP. In addition the use of the standard VAT rate to compute potential revenues ignores the existence of multiple rates, some lower than the standard rate and some higher; exemptions and zero rating provisions are also ignored. The potential revenue of a VAT in any country thus depends to a considerable extent on political decisions made in determining the VAT base. The actual revenues then depend further upon how fully that potential base is actually reached, which in turns depends upon two interrelated factors – the level of tax compliance (tax morale) and the effectiveness of the tax administration. Table 6.2 shows these three ‘efficiency’ measures for 74 countries using three-year averages for two periods, for the early 1990s and the three most recent years available in the 2000s.6 The change in the three efficiency measures between these two periods is also reported in Table 6.2. Despite the use of three-year averages, these data remain noisy and the various measures show considerable jumps for some countries. Sometimes, of course, these changes may reflect changes in legislation, as, for example, seems to be the case at least to some extent in the case of the Netherlands and Belarus. In other cases, as in Cyprus and Jamaica, it is hard to think of any simple explanatory factor. The range between ‘high’ and ‘low’ performers on the various measures is astounding. Considering only the gross compliance ratio in the most recent period, for instance, the range is from a low of 10 in Brazil to a high of 114 in Luxembourg (and an incredible 112 in Côte d’Ivoire). Countries that fail to rebate input taxes (and hence tax production as well as consumption) as well as countries that reap substantial revenues from increased rates on, for example, fuel or have a large tourist or transit consumption base to tap may of course score high on this measure, which may explain Luxembourg’s performance. Brazil’s apparent poor showing is equally explicable since the VAT data clearly refer only to the very limited national VAT and not the much more important state VATs. Sometimes countries such as Estonia and Singapore appear to be consistent high performers, as might be expected: but then so is Côte d’Ivoire, which makes one wonder what these numbers are really measuring. Overall, one thing is clear: all the figures in Table 6.2 need to be interpreted with considerable caution given the noise in the data. One way to facilitate the interpretation of the trends in Table 6.2 is to graph the average evolution of the VAT efficiency measures by region. Figure 6.5 does this for the gross collection ratio. Regardless of which of the three measures we examine, broadly similar

226

37.3 12.7 36.8 37.8 29.4 19.5 53.2 32.9 43.7 6.3 57.8 37.9 41.6 28.9 15.1

42.1 67.7 47.3 30.5

26.2 64.3 14.4 89.2

Last three years*

32.8 10.0 29.9 31.3 26.1 18.9 28.1 32.8 32.7 10.4 37.3 34.6 40.7 33.4 21.5

Initial three years*

15.9 3.4 32.9 −58.7

4.5 2.7 6.9 6.6 3.3 0.6 25.1 0.1 11.0 4.1 20.5 3.4 0.9 −4.5 −6.4

Difference

VAT Efficiency (%)

75.8 17.7 123.7

36.5 12.0 33.5 43.2 34.3 23.2 38.9 43.1 35.7 13.2 42.4 42.0 55.8 43.8 33.4

Initial three years

85.4 56.8 43.8

43.0 17.2 44.7 54.7 40.7 24.5 74.2 43.7 55.8 7.9 67.5 50.9 61.8 33.5 33.0

Last three years

9.6 39.2 −79.9

6.5 5.3 11.1 11.5 6.4 1.4 35.3 0.6 20.1 −5.3 25.2 8.9 6.0 −10.3 −0.4

Difference

C-Efficiency (%)

36.1 10.4 22.2 177.8

41.3 12.4 37.3 56.4 46.8 24.7 51.4 59.1 41.5 17.0 51.2 59.0 64.9 49.9 48.9

Initial three years

63.0 111.6 72.7 63.0

48.6 20.5 51.5 72.0 54.4 26.3 102.9 62.5 68.7 10.5 83.4 68.5 74.1 44.1 49.7 26.9 7.2 50.5 −114.8

7.4 8.0 14.2 15.6 7.6 1.7 51.4 3.3 27.3 −6.5 32.3 9.5 9.2 −5.8 0.7

Last Difference three years

VAT Gross Compliance Ratio (%)

Levels and changes in VAT efficiency ratio (three-year average)

Albania Argentina Armenia Australia Austria Bangladesh Belarus Belgium Bolivia Brazil Bulgaria Canada Chile Colombia Congo, Republic of Costa Rica Côte d’Ivoire Croatia Cyprus

Table 6.2

1990–92 1990–92 1998–00 1992–94

1996–98 1990–92 2003–05 2000–02 1995–97 2001–03 1992–94 1990–92 1990–92 1990–92 1994–96 1991–93 1990–92 1990–92 1997–99

2005–07 2005–07 2005–07 2005–07

2002–04 2002–04 2006–08 2005–07 2005–07 2006–08 2005–07 2005–07 2005–07 2006–08 2006–08 2005–07 2006–08 1998–00 2003–05

Initial Last three three years years

227

Czech Republic Denmark Egypt El Salvador Estonia Finland France Georgia Germany Greece Honduras Hungary Iceland Indonesia Ireland Israel Italy Jamaica Kazakhstan Korea Latvia Lithuania Luxembourg Madagascar Malta Mauritius

7.8

41.1 34.2 53.2 49.1 38.3 36.1 55.1 19.8 37.8 49.4 42.4 42.2 39.0 35.8 62.2 28.2 44.1 31.8 41.9 44.3 43.6 38.2 10.1 43.1 44.4

10.4

39.3 32.0 37.8 71.5 34.6 36.7 19.7 23.9 33.6 42.7 26.0 40.3 36.3 21.9 66.3 25.3 69.1 18.3 37.6 47.7 35.6 35.3 7.9 44.6 38.9

1.8 2.2 15.4 −22.4 3.6 −0.5 35.5 −4.1 4.1 6.7 16.5 1.9 2.7 13.9 −4.0 2.9 −25.1 13.5 4.2 −3.4 8.0 2.9 2.1 −1.5 5.5

−2.6 51.9 37.5 39.2 94.8 45.7 46.2 20.0 31.2 36.7 47.8 33.3 49.8 55.9 28.5 73.6 33.0 85.5 21.5 60.0 55.9 40.3 62.1 8.3 53.7 51.5

13.0 55.0 41.1 51.2 67.9 52.5 45.2 60.6 25.9 42.9 51.8 55.6 52.0 51.8 58.4 76.0 35.7 47.5 57.2 61.1 55.1 52.2 77.4 20.3 52.0 52.8

9.8 3.0 3.7 11.9 −26.8 6.8 −1.0 40.5 −5.3 6.3 4.0 22.3 2.2 −4.1 29.9 2.4 2.7 −38.0 37.3 1.1 −0.8 9.4 15.3 11.9 −1.7 1.2

−3.3 77.9 44.0 43.7 125.0 65.7 64.2 22.1 41.6 43.9 57.6 38.7 66.7 64.4 36.7 109.3 43.3 99.8 25.4 73.9 75.2 52.2 85.6 9.0 71.6 63.4

15.5 84.6 47.7 56.4 88.7 74.7 63.8 76.6 34.0 52.8 62.2 64.1 74.2 57.9 78.2 109.6 47.9 55.9 74.8 77.3 70.1 67.1 113.7 22.6 68.3 63.5

10.9 6.7 3.8 12.7 −36.3 9.0 −0.4 54.4 −7.6 8.9 4.6 25.4 7.5 −6.5 41.5 0.3 4.6 −44.0 49.5 3.4 −5.1 14.9 28.1 13.6 −3.3 0.0

−4.6 1990–92 2002–04 1998–00 1992–94 1994–96 1990–92 1997–99 1990–92 1995–97 2003–05 1990–92 1990–92 1990–92 1990–92 1990–92 1995–97 1992–94 1997–99 1990–92 1994–96 1993–95 1999–01 1994–96 1995–97 1998–00

1993–95 2005–07 2006–08 2005–07 2005–07 2005–07 2005–07 2005–07 2005–07 2005–07 2006–08 2005–07 2006–08 2002–04 2005–07 2006–08 2005–07 2005–07 2006–08 2005–07 2005–07 2006–08 2005–07 2005–07 2005–07 2005–07

2005–07

228

28.5 46.8 43.3 30.7 24.0 84.8 48.9 22.7 36.6 13.9 32.9 25.0 29.7 39.4 32.6 24.3 49.0 42.4

Initial three years*

21.9 70.5 46.3 39.6 29.5 41.3 51.1 38.6 31.5 24.6 12.9 33.4 36.6 39.5 21.8 30.2 36.3 53.9

Last three years* −6.6 23.7 3.0 8.9 5.5 –43.5 2.1 15.9 −5.1 10.7 −20.0 8.4 6.9 0.1 −10.8 5.8 −12.7 11.5

Difference

VAT Efficiency (%)

(continued)

Mexico Moldova Mongolia Morocco Nepal Netherlands New Zealand Nicaragua Norway Pakistan Panama Peru Poland Portugal Romania Russia Singapore South Africa

Table 6.2

28.1 62.3 72.9 51.7 32.8 57.2 66.0 52.6 28.9 46.1 45.7 45.9 26.1 45.8 72.6 66.0

50.8 16.4 29.8 37.2 48.0 41.8 34.3 95.3 52.0

Last three years

35.7 48.5 49.5 37.1 27.9 115.7 64.5

Initial three years

12.0 8.5 −0.6 −15.7 11.5 −13.1 −0.6

1.8 12.5

−7.6 13.8 17.3 14.6 2.1 −58.5 1.5

Difference

C-Efficiency (%)

40.3 57.2 61.4 46.5 31.2 169.9 83.8 24.8 72.9 19.7 57.1 32.7 48.6 61.8 50.2 47.1 114.8 68.1

Initial three years 32.7 76.4 90.6 67.4 37.3 87.1 86.6 45.9 78.0 32.4 21.0 53.1 59.9 60.5 30.6 61.7 91.7 86.7

−7.7 19.2 29.2 20.9 6.1 −82.9 2.9 21.1 5.1 12.7 −36.1 20.3 11.3 −1.3 −19.6 14.7 −23.1 18.6

Last Difference three years

VAT Gross Compliance Ratio (%)

1990–92 1997–99 1998–00 1990–92 1997–99 1990–92 2001–03 1991–93 1990–92 1990–92 1990–92 1990–92 1994–96 1997–99 1993–95 1994–96 1994–96 1991–93

1998–00 2006–08 2005–07 2006–08 2006–08 2005–07 2005–07 1999–01 2006–08 2005–07 1999–01 2005–07 2005–07 2005–07 2005–07 2006–08 2005–07 2005–07

Initial Last three three years years

229

21.3 25.5 51.5 32.2 52.3 23.7

36.9 45.3 36.9

45.2 38.3

37.3 70.5 6.3

34.7 30.7 50.2 27.6 45.1 32.0

15.6 32.4 37.4

26.1 22.7

34.6 89.2 7.9

2.7 35.5 −58.7

19.1 15.5

21.3 12.9 −0.5

−13.5 −5.2 1.2 4.6 7.2 −8.2

44.8 123.7 8.3

31.5 29.5

19.8 42.4 44.9

44.5 39.2 69.7 33.2 70.1 45.7

49.5 85.4 7.9

56.0 62.4

47.6 63.6 43.1

28.1 33.3 72.1 37.7 78.1 41.4

4.3 40.5 −79.9

24.5 25.8

27.8 21.2 −1.8

−16.4 −5.9 2.4 4.5 8.0 −4.3

56.7 177.8 9.0

36.8 32.5

25.0 57.8 59.6

57.3 60.7 83.4 37.5 82.9 55.5

62.9 113.7 10.5

65.0 77.2

58.4 84.9 57.4

37.0 51.4 86.0 41.8 95.5 52.6

6.2 54.4 −114.8

28.2 44.8

33.4 27.2 −2.2

−20.3 −9.2 2.7 4.4 12.6 −2.9

1990–92 1993–95

1990–92 1999–01 1990–92

1990–92 1990–92 1995–97 1998–00 1992–94 1993–95

2005–07 2003–05

2006–08 2006–08 2005–07

2005–07 1997–99 2005–07 2002–04 2006–08 2005–07

Sources:

IMF GFS 2010; World Development Indicators, various sources for the standard VAT rates.

Note: * ‘Initial three years’ and ‘Last three years’ refer to the first and last three years in the sample for which we have data available for a specific country.

Spain Sweden Switzerland Tajikistan Thailand Trinidad and Tobago Tunisia Ukraine United Kingdom Uruguay Venezuela, Rep. Bol. Average Max Min

230

The Elgar guide to tax systems

0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35

Americas Asia and Pacific

EU15+ Sub Saharan Africa

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

0.30

Eastern Europe North Africa and Middle East

Sources: IMF GFS 2010; World Development Indicators; Ebrill et al. (2001); various sources for standard VAT rates.

Figure 6.5

Gross collection ratio by region, 1990–2008

patterns emerge across these regions. The general trend is toward a consistent improvement over the last two decades, particularly for Eastern Europe and North Africa and the Middle East. The most striking exception is Sub-Saharan Africa, where a significant drop in VAT efficiency (however measured) in the first half of the 1990s has not since been reversed. With respect to the average VAT efficiency, for example, it is between 35 and 40 percent in most regions with the striking exceptions of Eastern Europe at 45 percent and Sub-Saharan Africa at 25 percent. While the ratios are generally higher for the C-efficiency ratio (about 5 percentage points higher) and the VAT gross collection ratio (about 10 percentage points higher), the pattern is the same in all cases. As Figure 6.5 shows, for example, for the VAT gross collection ratio the average figure for most regions of the world is between 60 and 65 percent, with the outliers being again Eastern Europe at over 70 percent and Sub-Saharan Africa at 40 percent. Despite the defects of these measures, one message thus comes through clearly. In almost every country, the revenue productivity of VAT could be improved. This could be done in two quite different ways. First, VAT structure could be strengthened by increasing the base; second, VAT administration and compliance could also be strengthened. IMF (2010)

Value-added tax 231 suggests that the main way VAT can be strengthened in developed countries is through changing VAT structure, while in most cases developing countries need to emphasize the longer and harder road of improving tax administration and compliance.7 As Bird and Gendron (2007) point out, it seems likely that some less developed countries may have adopted a VAT before they were ready to administer it: as suggested by Ebrill et al. (2001) the VAT, like the income tax, is best administrated through ‘selfassessment’, and it is far from clear that countries that have had great difficulty in administering income taxes are likely to do much better in administering a VAT. On the other hand, as we argue below, difficulty is not impossibility and even a bad VAT is often likely to be better than the possible alternatives in even the poorest countries. Despite the ‘noisy’ nature of the various VAT efficiency measures, the existence of a new set of numbers inevitably leads to attempts to identify the main determinants of the variations across countries and over time. A number of recent papers have examined this question. Aizenman and Jinjarak (2008) use data from a panel of 44 countries over 1970–99. They focus on the roles of political economy (greater polarization and political instability) and structural factors (such as urbanization, agriculture share, openness) to explain differences in C-efficiency of the VAT. They find that increases in the durability of the political regime and the ease and fluidity of political participation lead to higher VAT collection efficiency as do increases in urbanization and trade openness. On the other hand, a larger share of agriculture significantly decreases VAT efficiency, as Buettner et al. (2006) also find. More recently, De Mello (2009), using cross-section data for OECD and non-OECD countries, finds that C-efficiency increases with lower VAT rates, as well as with lower shares of administrative costs in tax revenues (as a proxy for the efficiency of the tax administration),8 more pro-competition regulatory frameworks in product markets (measuring non-tax incentives for non-compliance), trade openness and better country governance indicators (regulatory quality, rule of law, and government effectiveness).9 Overall, De Mello (2009) finds no discernible difference in VAT productivity between OECD and non-OECD countries in his sample. These studies are interesting but far from conclusive given their inevitable reliance on proxies and the unclear interpretation of the various VAT efficiency measures. In particular, there are no good measures of tax administrative efficiency, and the various measures of VAT efficiency do not discriminate between differences in tax structure (rates and base) and changes in tax evasion – for example, resulting from changes in tax morale or in the probability of evaders being caught or the size of penalties if they are. While we cannot deal with these problems fully, we can build to some

232

The Elgar guide to tax systems

extent on previous research both by expanding the sample of countries and time period covered and by introducing refinements in the estimation methodology. In particular, we use data for 107 countries,10 covering the period 1990 to 2008.11 As the dependent variable we use alternatively the three different measures of VAT efficiency: (1) VAT efficiency ratio; (2) C-efficiency ratio; and (3) VAT gross collection ratio. In terms of estimation methodology, applying the fixed effects estimator to the panel data would produce inefficient estimates due to presence of autocorrelation and heteroskedasticity. By using the lagged dependent variable as an additional explanatory variable – given that this year’s efficiency ratio is likely to depend on last year’s – we can address the problem of first-order autocorrelation; however, the problems of second-order autocorrelation and heteroskedasticity remain. Therefore, we apply the Arellano-Bond (1991) difference generalized method of moments (GMM) estimator.12 Basically, to cope with the time-invariant country characteristics (μi), the difference GMM uses first differences to transform equation Yit 5 a1Yi,t21 1 b1Xit 1 mi 1 mit

(6.1)

DYit 5 a1DYi,t21 1 b1DXit 1 Duit

(6.2)

into

where Yit is one of the three efficiency ratios, Xit includes potential determinants of the efficiency ratio, such as, GDP per capita, population, share of agriculture in GDP, share of imports in GDP, share of urban population in total population, number of years since the introduction of the VAT, domestic consumption of crude petrol, domestic consumption of alcohol, education index, business concentration (measured as the average number of employees per registered company in manufacturing), a dummy variable for developed countries, and tax morale. These variables and their sources are described in Table 6A.1 and the descriptive statistics are provided in Table 6A.2, both in Appendix B. Finally, uit is the error term. Note that the first-difference lagged dependent variable DYi,t21 is instrumented with its past levels. The results of the estimation are provided in Table 6.3. The estimated coefficients for population, used as a proxy for the size of a country, suggest that country size is positively associated with the collection efficiency. This effect, however, does not seem to be economically important – one percentage point increase in population leads to between 0.013 and 0.031 percentage points increase in the VAT collection efficiency. The

233

Tax morale * Developed

Tax morale

Education

Petrol consumption

Urban

VAT experience

Ln(imports)

Agriculture

0.356** (0.150) −0.016 (0.032) 1.875*** (0.566) −3.266*** (0.905) 0.108 (0.113) −0.002 (0.005) −6.289** (2.913) 6.349 (6.225) 0.268 (1.229) −1.094*** (0.411) 1.467*** (0.509)

(1) 0.526*** (0.089) −0.021 (0.024) 1.351** (0.672) −2.386** (0.988) 0.062 (0.070) −0.003 (0.008) −3.440** (1.427) 5.617 (5.610) −0.606 (0.844) −0.759 (0.480) 1.263*** (0.487)

(2)

Y = VAT Efficiency

Determinants of VAT efficiency ratios

Ln(population)

Ln(GDP pc)

Y−1

Table 6.3

0.392*** (0.139) −0.010 (0.040) 2.236*** (0.720) −3.737*** (1.145) 0.163 (0.141) 0.000 (0.007) −7.748** (3.482) 8.332 (7.795) −0.119 (1.507) −1.517*** (0.547) 1.856*** (0.656)

(3) 0.534*** (0.093) −0.022 (0.034) 1.769* (0.907) −3.069** (1.343) 0.112 (0.097) 0.001 (0.011) −4.830** (1.926) 7.042 (7.658) −0.968 (1.199) −1.112 (0.701) 1.762*** (0.652)

(4)

Y = C-Efficiency

0.158 (0.143) 0.041 (0.062) 3.114*** (1.035) −5.118*** (1.612) 0.305 (0.191) −0.011 (0.010) −7.143 (4.962) 13.527 (10.311) −0.509 (2.055) −1.510* (0.780) 1.440 (0.922)

(5)

0.483*** (0.094) −0.036 (0.046) 2.688** (1.305) −4.518** (1.929) 0.171 (0.135) 0.003 (0.015) −7.182*** (2.700) 10.115 (10.676) −1.838 (1.623) −1.691* (0.946) 2.379*** (0.903)

(6)

Y = VAT Gross Compliance

234

(continued)

(2)

Source:

(3)

(4)

Y = C-Efficiency (5)

(6)

Y = VAT Gross Compliance

0.001 −0.004 −0.003 (0.022) (0.031) (0.042) −0.195** −0.096 −0.200* −0.107 −0.278* −0.178 (0.088) (0.063) (0.107) (0.086) (0.148) (0.123) 0.087** 0.025 0.096** 0.031 0.146** 0.049 (0.038) (0.024) (0.048) (0.032) (0.064) (0.046) 368 368 368 368 368 368 39 39 39 39 39 39 0.058 0.052 0.033 0.03 0.044 0.033 F(14, 354)= 5.86 F(15, 353)= 6.68 F(14, 354)= 6.01 F(15, 353)= 7.96 F(14, 354)= 5.04 F(15, 353)= 7.02 Prob>F=0.000 Prob>F=0.000 Prob>F=0.000 Prob>F=0.000 Prob>F=0.000 Prob>F=0.000 0.111 0.499 0.221 0.291 0.262 0.257 0.248 0.978 0.309 0.804 0.394 0.786

(1)

Y = VAT Efficiency

Authors’ computations.

Notes: Standard errors in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%.

Sargan Test (p-value) AR(2) Test (p-value)

Observations Number of id Corr(Y, Yhat) squared F-statistics

Alcohol consumption

Business concentration

Tax morale * Experience

Table 6.3

Value-added tax 235 negative significant coefficient for the share of agriculture in GDP confirms previous findings – in line with the hypothesis that a larger part of the agricultural activities might belong to the informal sector or otherwise be exempt – on the negative effect of the size of this sector on collection efficiency. However, in contrast to earlier studies, our estimates suggest that a higher degree of urbanization is associated with less efficient VAT collection. Urbanization may affect collection efficiency through two distinct channels. On the one hand, higher urbanization may reduce the cost of monitoring tax compliance, implying overall higher tax compliance.13 On the other hand, because people live close to their neighbors in urban agglomerations, informal transactions can become more feasible, reducing tax collections of both indirect and direct taxes. Our results on urbanization may perhaps be interpreted as suggesting the latter effect is stronger.14 One percentage point increase in share of urban population leads to between 3.4 and 7.7 percentage points lower VAT collection efficiency. In principle, higher tax morale should be positively correlated with the VAT collection efficiency. However, this effect may be complex and dependent on such factors as the country’s level of economic development and the number of years since the VAT was introduced. Even though in some specifications we do not obtain statistically significant coefficients on tax morale and its interaction terms, we find them jointly significant. In particular, our results suggest that developed countries with the same number of years of experience with the VAT experience stronger positive effects from tax morale on the VAT collection efficiency; one percentage point increase in tax morale leads to between 1.3 and 2.4 percentage points higher increase in VAT collection in developed than in developing countries.

4

VAT AND REVENUE

Although many different objectives such equity and even efficiency may weigh in the minds of tax policy-makers, revenue considerations are usually to the fore. The rapid rise to success of the VAT suggests that it is likely to be a good revenue-raiser. Indeed, the VAT has sometimes even been dubbed a ‘money machine’ – a tax that might make it all too easy to expand revenues and the size of government.15 However, it is not that easy to say just what has happened to overall tax revenues in countries that have introduced a VAT. In many, for example, VAT introduction was accompanied by other changes in tax policy and administration, making it difficult to isolate the effects exclusively due to the VAT. A good point to start is taking a look at the raw data. The great increase

236

The Elgar guide to tax systems

0.40

0.35

0.30

0.25

0.20

Developing - No VAT

Developing - VAT

Developed - No VAT

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

0.15

Developed - VAT

Note: * Central government revenues. Source: IMF GFS 2010; World Development Indicators.

Figure 6.6

Total revenue collection* as a share of GDP, countries with versus countries without VAT, developed versus developing countries, 1990–2008

in the number of countries worldwide that have adopted a VAT (Figure 6.1) has been accompanied by a mild growth in the tax revenues to GDP ratio but, as Figure 6.6 shows, if we compare tax ratios among developed countries that have and have not adopted a VAT and developing countries that have and have not adopted a VAT, the story seems a bit different. The much higher ratio for developed countries with a VAT than for developed countries without a VAT seems compatible with the VAT ‘money machine’ story. However, the tax ratio is lower for developing countries with a VAT than for developing countries without a VAT. Moreover, while the ratio has increased over the last decade for developing countries without a VAT, it has declined for developing countries with a VAT. This is not quite compatible with the ‘money machine’ story. Part of the explanation for these divergent trends may lie in the fact that the adoption of the VAT in developing countries in recent years has often been accompanied not only by the elimination of other general consumption taxes such as turnover and general sales taxes but, most importantly, by significant reforms in trade policies including major reductions in average tariffs. As Figure 6.7 shows, customs duties have declined significantly in developing countries. However, there is considerable diversity across different regions, and Sub-Saharan Africa stands out both because

Value-added tax 237 0.07 0.06 0.05 0.04 0.03 0.02 0.01

GST Developing

GST Developed

Customs Developing

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

0.00

Customs Developed

Note: * Central government revenues. Source:

IMF GFS 2010; World Development Indicators; Ebrill et al. (2001).

Figure 6.7

General sales tax (GST) and customs duties collection* as a share of GDP, developed versus developing countries, 1990–2008

customs duties are not only double the average level in other regions but also remain more important than VAT in revenue terms. A number of studies have examined the impact of VAT on revenues. For example, Aizenman and Jinjarak (2006) studied the impact of globalization (measured by trade openness and financial openness) on different taxes including the VAT and conjectured that if globalization works to cut fiscal revenue from traditional taxes (tariffs and seigniorage), this would set in motion forces to increase revenues from such ‘hard to collect’ taxes as VAT and income taxes. They concluded that globalization factors were responsible for about two-thirds of the 12 percent drop of the ‘easy to collect’ tax revenues to GDP in developing countries and about a fifth of the 16 percent increase of the ratio of the ‘hard to collect’ taxes to GDP. Keen and Lockwood (2006) focus specifically on the ‘money machine’ argument. On the basis of a panel data set for OECD countries (excluding the US) they find evidence that the VAT does indeed appear to have been a ‘money machine’ in the sense that countries with a VAT tend to raise more revenue, all else equal, than do those without it, although they also found that VAT did not appear to have resulted in any increase in government size in a statistical sense in part because the revenue that it raises has to some degree been offset by reduced revenues from other taxes. They

238

The Elgar guide to tax systems

interpret these results as suggesting that VAT use has been more driven by its greater effectiveness as a tax instrument than by a desire to finance bigger government. In a more recent paper, Keen and Lockwood (2010) use an unbalanced panel of 143 countries over 26 years to explore much the same issues. Their point of departure is that access to a more efficient tax instrument should lead an optimizing government to increase the ratio of total revenues to GDP. Empirically, their approach consists in adding a dummy variable for the presence of a VAT (and several interaction terms) to a conventional ‘tax effort’ regression framework to examine whether those variables have a statistically significant impact on the tax revenue to GDP ratio.16 They find that the adoption of a VAT is associated with a long-run increase in the overall revenue to GDP ratio of 4.5 percent. However, interpreting this result is complicated because the interaction terms of the VAT take-up dummy with income levels and openness are positive and significant while the dummy variable for take-up itself now takes a negative sign. Nonetheless, when all these effects are taken into account, numerical simulations suggest that in most countries adopting a VAT increased revenues; however, for Sub-Saharan Africa the evidence is mixed. Using our panel data for 107 countries over the period 1990 to 2008, we build on Keen and Lockwood (2010) to examine the question of whether the introduction of the VAT has on average led to an increase in revenue collections, with special attention to differences in the experience of developing and developed countries. Because this type of estimation gives rise to an array of econometric challenges, our approach is to proceed with different steps involving different estimation methodologies. This incremental approach not only allows for more transparency in the handling of the econometric issues but also provides a series of robustness tests for our estimates. We will first discuss the results for the whole sample in Table 6.4 and next the results obtained when the sample is split between developed and developing countries. As a first step, columns 1 and 2 in Table 6.4 present the results obtained by applying the fixed effects Estimator to the entire data panel. However, these results are likely to be inconsistent due to three main problems: (1) the potential endogeneity of the VAT dummy variable (since the decision to introduce a VAT may be driven by the desire to increase revenue collections); (2) autocorrelation; and (3) heteroskedasticity. To resolve the endogeneity problem, we apply the instrumental variable fixed effects estimator in columns 3 and 4 of Table 6.4, where column 3 presents the results obtained without the interaction terms involving the VAT dummy. To address the endogeneity of the VAT dummy, we instrument it with three instrumental variables: (1) the percentage of countries

239

Ln(GDP pc)* DVAT

Education

Urban

Age dependency

Federal

Ln(population)

Agriculture

Openness

Ln(GDP pc)

DVAT

0.004 (0.007) 0.045*** (0.009) −0.027*** (0.010) −0.041 (0.060) 0.043 (0.038) 0.228 (0.200) −0.146*** (0.041) −0.074 (0.113) −0.162*** (0.043)

(1)

(2)

0.152*** (0.044) 0.051*** (0.010) −0.026*** (0.010) 0.085 (0.067) 0.051 (0.038) 0.235 (0.199) −0.138*** (0.042) −0.051 (0.110) −0.141*** (0.043) −0.013*** (0.004)

FE (4)

(5)

(6)

IV FE II (7)

(8)

IV + Regional Dummies

0.727*** 0.726*** 0.942*** 0.933*** (0.043) (0.046) (0.017) (0.016) −0.032* −0.189 0.000 0.000 0.030* 0.114** (0.019) (0.271) (0.012) (0.138) (0.016) (0.051) 0.048*** 0.022 0.004 0.002 0.001 0.001 (0.010) (0.019) (0.007) (0.011) (0.003) (0.003) −0.026*** −0.023** 0.007 0.007 0.003 0.001 (0.010) (0.011) (0.007) (0.007) (0.002) (0.002) −0.111* −0.186 −0.088* −0.072 0.000 0.021 (0.061) (0.197) (0.046) (0.104) (0.030) (0.035) 0.093** 0.069 −0.021 −0.017 −0.001 0.005*** (0.046) (0.046) (0.029) (0.025) (0.003) (0.002) 0.416* −0.183 −0.014 0.056 0.002 −0.004** (0.228) (0.148) (0.027) (0.119) (0.003) (0.002) −0.110** −0.137*** −0.051 −0.049 0.015 −0.002 (0.046) (0.047) (0.034) (0.031) (0.014) (0.012) −0.134 −0.069 0.078 0.080 0.004 0.015** (0.121) (0.132) (0.069) (0.072) (0.006) (0.007) −0.163*** −0.177*** −0.025 −0.022 0.011 0.014 (0.045) (0.058) (0.030) (0.033) (0.015) (0.014) 0.018 0.000 −0.009** (0.024) (0.012) (0.004)

(3)

IV FE I

Effect of the VAT introduction on revenue collection, dependent variable: revenues/GDP

Revenues/GDP−1

Table 6.4

(10)

0.940*** 0.932*** (0.016) (0.016) 0.028* 0.120** (0.015) (0.051) 0.001 0.001 (0.003) (0.003) 0.003 0.001 (0.002) (0.002) −0.001 0.029 (0.030) (0.034) −0.000 0.005*** (0.003) (0.002) 0.002 −0.004** (0.003) (0.002) 0.013 −0.004 (0.014) (0.011) 0.004 0.015** (0.006) (0.007) 0.013 0.017 (0.014) (0.013) −0.009** (0.004)

(9)

GMM

240

(1)

(continued)

(2)

(3)

IV FE I (4)

0.000 0.000

0.000 0.002

0.010

0.000

−0.177*** 0.051 (0.052) (0.208) −1.216* −1.915** −0.366 (0.718) (0.871) (0.500) YES YES YES NO NO NO 1157 1157 1113 0.92 0.91 0.91 0.067 0.092

FE

0.026

0.881

0.396 (0.535) YES NO 1075 0.96 0.984

(5)

0.054

0.905

−0.032 (0.108) 0.279 (0.306) YES NO 1075 0.96 0.982

(6)

IV FE II

Source:

Authors’ computations.

(7)

0.258

0.049

(9)

0.092

0.392

0.215

0.163

(10)

0.105

0.351

−0.128** (0.051) −0.097** (0.042) NO YES 1075 0.95 0.616

GMM

−0.118** (0.053) −0.089** −0.027 (0.043) (0.025) NO NO YES YES 1075 1075 0.95 0.94 0.354 0.571

(8)

IV + Regional Dummies

−0.027 (0.026) NO YES 1075 0.94 0.499

Notes: Robust standard errors in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%. Base category = ‘Small Islands’, including Cyprus, Iceland, Malta, Seychelles, and St. Kitts and Nevis.

−0.996 (0.714) Country dummies YES Regional dummies NO Observations 1157 R-squared 0.91 Sargan-Hansen Test (p-value) AR(2) Test 0.000 (p-value) Heteroskedasticity 0.000 Test (p-value)

Agriculture* DVAT Constant

Table 6.4

Value-added tax 241 in region k that have implemented a VAT in year t (regions are defined on the basis of Ebrill et al., 2001); (2) dummy variable = 1 if country i was in a non-crisis IMF program (SAF, PRGF) in year t; and (3) dummy = 1 if country i was in a crisis IMF program (SBA, EFF) in year t.17 Column 4 shows the results obtained by applying the instrumental variable fixed effects estimator when the interaction terms with the VAT variable are included. Because the VAT dummy is an endogenous variable, we also have to obtain instrumental variables for the interaction terms. Because it is difficult to find good instruments for the interaction terms, we first estimate the reduced form of the VAT equation and then predict the VAT variable and interact it with the exogenous variables. We then use these ‘predicted’ interaction terms as instrumental variables for the original interaction terms in the equation.18 Basically, we want to estimate the following structural equation: Total Revenue 5 a1DVAT 1 b1Z1 1 u1

(6.3)

DVAT 5 a2 Total Revenue 1 b2Z2 1 u2

(6.4)

where

After substituting (6.3) in (6.4) for total revenues, equation (6.4) becomes: DVAT 5 p1Z1 1 p2Z2 1 v2

(6.4)

which we estimate by applying the probit estimator and predict DVAT. This predicted variable is the one interacted with the exogenous variables to instrument for the interaction terms.19 Because the autocorrelation and heteroskedasticity problems are not resolved by applying the instrumental variable fixed effects estimator, we add the lagged dependent variable as an explanatory variable to specifications and estimate them using the same procedure. These results are shown in columns 5 and 6 of Table 6.4. This resolves the problem of autocorrelation but not the one of heteroskedasticity. To deal with this problem, we apply the GMM estimator, which is efficient in the presence of the arbitrary heteroskedasticity to address for this issue. However, when the GMM estimator is applied in specifications 5 and 6 of Table 6.4, it does not produce full rank estimated covariance matrix of moment conditions, preventing the calculation of the optimal weighting matrix for the GMM estimation. We, therefore, re-estimate these equations with the GMM estimator but including regional dummies instead of country dummies (columns 9 and 10 of Table 6.4). For comparison purposes, we also

242

The Elgar guide to tax systems

present the results obtained by re-estimating specifications 5 and 6 using the instrumental variable estimator when regional dummies are included instead of country dummies (columns 7 and 8). As the results in column 10 of Table 6.4 suggest, countries having a VAT in their tax structure have, on average, a higher share of total revenue in GDP of 12 percentage points than those countries that do not have it. This point estimate is almost three times larger than that in Keen and Lockwood (2010). However, this estimate assumes that countries have equal levels of economic development and size of their agricultural sector. To capture the possible different effects of the presence of a VAT in countries with different levels of economic development, Table 6.5 shows the results for developed and developing countries separately. These results (columns 2 and 4 in Table 6.5) suggest that the positive effect of the VAT on total revenue collections is significant only in the developing countries. Developing countries having a VAT in their tax structure have, on average, a higher share of total revenue in GDP of 11 percentage points than those countries that do not have it. In developed countries, on the other hand, this effect does not seem to be significant, which differs from Keen and Lockwood (2010). As already mentioned, whether or not the introduction of a VAT leads to an overall increase in revenues, it may result in some substitution for other perhaps less efficient revenue instruments such as customs duties. A commonly recommended policy package in recent decades has been to introduce a VAT to compensate for the revenue reduction owing to tariff reform and has received some attention in the literature. Keen and Ligthart (2002) provide a theoretical rationale for this strategy.20 Buettner et al. (2006) studied 21 developing and transition countries where WTO accession with trade liberalization policies was accompanied by an introduction of a VAT system over the period 1990 to 2004 and found that both developing and transition countries were indeed able to restructure their revenue sources along these lines, with developing countries in general being able to raise from VAT more than they lost from tariff reform. More recently, however, Baunsgaard and Keen (2010) looked at the same question for 117 countries over the last 32 years. While they also found that revenue substitution (VAT for customs duties) was complete for highincome countries and also, in the long run, for middle-income countries, for low-income countries the evidence is more mixed, with considerable variation in country experience and full replacement being problematic in many cases. Using our sample, we re-examine this question – to what extent may reduction in customs duties be offset through the introduction of a VAT? – as well as looking at how other important tax sources have adjusted to

Value-added tax 243 Table 6.5

Effect of the VAT introduction on revenue collection, GMM, developed versus developing countries, dependent variable: revenues/GDP Developed Countries

Revenues/GDP−1 DVAT Ln(GDP pc) Openness Agriculture Ln(population) Federal Age dependency Urban Education

(1)

(2)

(3)

(4)

0.885*** (0.259) −0.517 (3.371) −0.049 (0.313) 0.005 (0.035) 0.912 (6.526) 0.038 (0.244) −0.086 (0.513) −1.269 (8.463) 0.102 (0.763) 1.124 (6.840)

0.903*** (0.047) −0.114 (0.495) −0.010 (0.024) −0.000 (0.002) −2.157 (4.926) 0.004 (0.004) −0.006** (0.003) 0.064 (0.094) −0.017 (0.032) 0.102 (0.145) 0.005 (0.027) 2.019 (5.001) −0.021* (0.012) −0.008 (0.009) −0.020** (0.009) −0.004 (0.006)

0.934*** (0.020) 0.022 (0.015) −0.002 (0.004) 0.006* (0.004) −0.014 (0.034) 0.002 (0.004) 0.004 (0.003) 0.002 (0.015) 0.010 (0.009) 0.007 (0.014)

−0.011** (0.005)

0.929*** (0.020) 0.106* (0.058) 0.001 (0.006) 0.004 (0.004) 0.015 (0.040) 0.005* (0.003) −0.001 (0.003) −0.006 (0.013) 0.013 (0.008) 0.012 (0.013) −0.009* (0.005) −0.095** (0.046) −0.013*** (0.005)

−0.005 (0.004) −0.006 (0.007) 0.008 (0.009)

−0.010** (0.004) −0.011** (0.005) −0.000 (0.005)

−0.023 (0.028)

−0.084* (0.045)

Ln(GDP pc)* DVAT Agriculture* DVAT Americas EU15+2a Eastern Europe Asia and Pacific

−0.154 (0.898) 0.072 (0.517) −0.060 (0.305) −0.111 (0.667)

Sub-Saharan Africa North Africa and Middle East Constant

Developing Countries

−0.051 (0.576)

0.044 (0.336)

244

The Elgar guide to tax systems

Table 6.5

(continued) Developed Countries

Observations R-squared Sargan-Hansen Test (p-value) AR(2) Test (p-value) Heteroskedasticity Test (p-value)

Developing Countries

(1)

(2)

(3)

(4)

372 0.89 0.966

372 0.95 0.479

703 0.929 0.815

703 0.932 0.722

0.880 0.935

0.820 0.383

0.325 0.652

0.472 0.235

Notes: Robust standard errors in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%. a. Original 15 EU countries plus Norway and Switzerland. Source: Authors’ computations.

the introduction of a VAT. An appropriate approach to doing so is to use Zellner’s (1962) seemingly unrelated regressions (SURs) to capture the efficiency due to the correlation of the disturbances across equations (Baltagi, 2005). The SUR approach both gains efficiency in estimation by combining information from different equations and also allows us to impose and/or test restrictions on the parameters in different equations.21 Using our panel data for 107 countries over the period 1990 to 2008, a first set of empirical results is presented in Table 6.6. The results in columns 4, 5, and 6 in Table 6.6 suggest that among countries with equal level of trade openness, share of agriculture in GDP, urbanization, quality of bureaucracy and political durability, countries that have implemented a VAT have higher income tax yields (5.8 percentage points) as well as higher general sales collection (around 2.4 percentage points) and lower customs duties collection (2.9 percentage points). However, the coefficient for general sales tax (GST) is only statistically significant when the VAT interaction terms are excluded from the estimating equation (column 2). In the case of income taxes, the effect of the VAT is reduced with higher shares of agricultural sector and urbanization – one percentage point increase in agriculture and urbanization reduces the effect of the VAT on income tax collection by 0.13 and 0.08 percentage points, respectively. In the case of customs duties, one percentage point increase in agriculture and urbanization shares reduces the effect of the VAT by 0.11 and 0.02 percentage points, respectively. Higher political durability decreases the effect of VAT on customs duties collection by 0.01 percentage points

Value-added tax 245 Table 6.6

DVAT

The effects of the introduction of the VAT on customs duties, income taxes, and general consumption taxes (narrowly defined) (seemingly unrelated regressions) (1)

(2)

(3)

(4)

(5)

(6)

Income

GST

Customs

Income

GST

Customs

−0.238 (0.314) −0.213 (0.216) −12.363*** (1.551) −3.548*** (0.685) 1.399*** (0.133) 0.012*** (0.003)

2.932*** (0.237) 0.477*** (0.163) −2.270* (1.171) 1.472*** (0.518) 0.051 (0.101) −0.012*** (0.003)

−0.547*** (0.114) 0.093 (0.079) 0.838 (0.565) −1.971*** (0.250) −0.229*** (0.048) −0.006*** (0.001)

5.838*** 2.390 −2.885*** (2.049) (1.567) (0.687) Openness 0.525 0.286 0.255 (0.598) (0.457) (0.201) Agriculture −4.224 −1.930 −4.182*** (2.888) (2.209) (0.969) Urban 2.750 0.742 −3.750*** (1.820) (1.392) (0.610) Bureaucracy 1.082*** −0.062 0.120 (0.414) (0.316) (0.139) Political durability 0.011 −0.013** −0.012*** (0.007) (0.005) (0.002) Openness*VAT −0.987 0.229 −0.265 (0.642) (0.491) (0.215) Agriculture*VAT −13.017*** −2.546 11.258*** (3.566) (2.727) (1.196) Urban*VAT −7.714*** 0.753 2.439*** (1.963) (1.501) (0.658) Bureaucracy*VAT 0.254 0.047 −0.275* (0.438) (0.335) (0.147) Political 0.004 0.003 0.010*** durability*VAT (0.008) (0.006) (0.003) Constant 6.057*** 1.683*** 3.578*** 1.593 2.493* 4.739*** (0.780) (0.589) (0.284) (1.903) (1.455) (0.638) Observations 914 914 914 914 914 914 R-squared 0.377 0.280 0.293 0.397 0.286 0.419 chi2(3) = 31.356, Pr = 0.0000 Breusch-Pagan chi2(3) = 39.754, Pr = 0.0000 test of independence Notes: Standard errors in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%. Source:

Authors’ computations.

and the quality of bureaucracy increases the effect of the VAT on custom duties collection by 0.27 percentage points. As Table 6.7 shows, the results are not as consistent when we adopt a more expansive definition of income taxes, to include payroll and Social

246

The Elgar guide to tax systems

Table 6.7

DVAT

The effects of the introduction of the VAT on customs duties, income taxes, and general consumption taxes (broadly defined, including social security revenues and excises) (seemingly unrelated regressions) (1)

(2)

(3)

(4)

(5)

(6)

Income +SS

GST + Excises

Customs

Income +SS

GST + Excises

Customs

2.246*** (0.573) 2.373*** (0.533) −19.218*** (3.411) 0.186 (1.376) 2.767*** (0.236) 0.007 (0.006)

9.484 (7.553) −8.220 (7.027) 5.976 (44.953) 18.182 (18.129) 7.419** (3.109) −0.168** (0.080)

−0.841*** −0.562 −4.698 −2.695*** (0.123) (4.253) (57.569) (0.914) Openness 0.166 1.002 2.267 0.602** (0.115) (1.236) (16.737) (0.266) Agriculture 2.134*** −7.403 8.709 −3.893*** (0.733) (6.728) (91.070) (1.445) Urban −1.360*** 5.286 −1.641 −3.219*** (0.296) (3.394) (45.939) (0.729) Bureaucracy −0.287*** −0.704 0.876 −0.153 (0.051) (0.793) (10.740) (0.170) Political −0.005*** 0.046*** −0.009 −0.010*** durability (0.001) (0.013) (0.171) (0.003) Openness*VAT 1.657 −11.381 −0.625** (1.379) (18.670) (0.296) Agriculture*VAT −18.654** −8.748 8.667*** (7.859) (106.373) (1.688) Urban*VAT −6.266* 20.738 1.936** (3.723) (50.393) (0.800) Bureaucracy*VAT 3.660*** 7.104 −0.106 (0.833) (11.270) (0.179) Political −0.044*** −0.198 0.007** durability*VAT (0.015) (0.201) (0.003) Constant 3.134* −13.965 3.412*** 6.731* 0.475 5.018*** (1.601) (21.095) (0.344) (3.943) (53.378) (0.847) Observations 797 797 797 797 797 797 R-squared 0.387 0.016 0.289 0.420 0.017 0.327 Breuschchi2(3) = 67.462, Pr = 0.0000 chi2(3) = 56.164, Pr = 0.0000 Pagan test of independence Notes: Standard errors in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%. Source: Authors’ computations.

Value-added tax 247 Security contributions, as well as a more expansive definition of general sales taxes to include excise tax revenues. Using the full specification when the interaction dummy variables for the presence of VAT (columns 4–6 in Table 6.7), the only statistically significant coefficient is for customs duties. That is, including Social Security contributions and excise duties into the analysis results in the introduction of a VAT having no different effects on income taxes (or taxes on goods and services) in countries with similar levels of trade openness, share of agriculture in GDP, urbanization, quality of bureaucracy and political durability. However, the introduction of the VAT continues to have a positive significant effect on income tax plus Social Security revenues collections when the dummy interaction variables are excluded (column 1 in Table 6.7). Finally, as Baunsgaard and Keen (2010) show, the evidence on the potential revenue consequences of the introduction of a VAT is especially problematic in the poorer countries. We therefore ran separate sets of regressions for the sub-sample of developing countries using both the narrow and broad definitions of income taxes and general consumption taxes, although only the latter is shown here, in Table 6.8. In terms of direct effects, the presence of a VAT, as captured by the coefficient for the VAT dummy variable, tends, as expected, to decrease revenues from the customs tariff and to increase revenues from general consumption taxes. However, the positive effect of the presence of VAT for income tax collections is now very weak, with the coefficient for the VAT dummy being positive only for the broad definition of income taxes (Table 6.8) and even then significant only at the 10 percent level. There is thus little empirical evidence in support of the common assertion that the introduction of VAT does much to reinforce compliance with income taxes in developing countries. This conclusion does not change much when we take into account the interaction terms between VAT presence and such country features as openness, shares of agriculture and urban population, quality of bureaucracy, and durability of political regime.

5

IMPROVING VAT

Despite its spread and success throughout the world the VAT is by no means a perfect tax. As IMF (2010) recently noted, for example, the ‘gap’ in VAT coverage displayed by the various VAT efficiency measures discussed earlier may be usefully divided into two components: the coverage gap (the extent to which some of the potential tax base is excluded by design features of the tax) and the administration gap (the extent to which the tax is not effectively administered). Like any tax, the VAT now

248

The Elgar guide to tax systems

Table 6.8

Developing countries. The effects of the introduction of the VAT on customs duties, income taxes, and general consumption taxes (broadly defined, including Social Security revenues and excises) (seemingly unrelated regressions)

DVAT Openness Agriculture Urban Bureaucracy Political durability Openness* DVAT

(1)

(2)

(3)

(4)

(5)

(6)

Income +SS

GST + Excises

Customs

Income +SS

GST + Excises

Customs

1.081* (0.573) 3.070*** (0.523) −11.063*** (3.141) −0.102 (1.362) 1.223*** (0.267) −0.022* (0.012)

4.286*** (0.397) 1.892*** (0.363) 6.630*** (2.177) 1.837* (0.944) 0.364** (0.185) −0.003 (0.008)

2.022 (3.001) 2.456*** (0.814) 6.883 (4.213) −2.276 (2.587) 0.129 (0.562) 0.032* (0.019) −0.083 (0.923) −2.622 (4.951) 5.589** (2.800) 0.149 (0.595) −0.046** (0.021)

−2.642** (1.203) 0.632* (0.326) −3.816** (1.689) −3.225*** (1.037) −0.084 (0.225) −0.013* (0.008) −0.623* (0.370) 7.146*** (1.985) 1.798 (1.123) 0.025 (0.239) 0.012 (0.009)

Agriculture* DVAT Urban* DVAT Bureaucracy* DVAT Political durability* DVAT Constant Observations R-squared BreuschPagan test of independence

−0.720*** −8.349* (0.159) (4.310) 0.222 2.445** (0.146) (1.169) 0.978 −10.104* (0.875) (6.050) −1.398*** −3.215 (0.379) (3.715) −0.070 −1.374* (0.074) (0.807) −0.003 −0.031 (0.003) (0.028) 1.254 (1.326) −4.741 (7.110) 3.607 (4.020) 2.960*** (0.854) 0.018 (0.031)

5.453*** 0.312 3.093*** 13.657*** 1.848 4.898*** (1.651) (1.144) (0.460) (3.979) (2.770) (1.111) 554 554 554 554 554 554 0.152 0.227 0.11 0.188 0.252 0.143 Chi2(3) = 160.195, Pr = 0.0000 chi2(3) = 151.881, Pr = 0.0000

Notes: Standard errors in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%. Source: Authors’ computations.

Value-added tax 249 in place in most countries can almost always be improved, sometimes substantially. In this section we discuss briefly and selectively some aspects of three paths to VAT improvement – structural change, administrative change, and what may perhaps be called political change. To begin with, however, it seems worth taking a brief look at what may be called ‘VAT economics’ in part because – despite the virtually world-wide triumph of VAT – as yet we have surprisingly little solid knowledge of some critical factors with respect to the economic effects of VAT. VAT Economics Growth As noted earlier, the VAT is usually considered to cause fewer distortions in the economy and thus be more efficient than income taxes or other forms of general consumption tax. Since one reason for this greater efficiency is because savings and investments decisions are in particular unaffected by the presence of a VAT, it follows that choosing a VAT to finance part of public expenditures can be considered a pro-growth choice (Keen and Ligthart, 2002). As Martinez-Vazquez et al. (Chapter 2 this volume) show, the empirical evidence appears to support the belief that wider reliance on the VAT (and other taxes on consumption) as opposed to direct taxes may indeed result in faster economic growth. As Smart and Bird (2009) show in a study of several Canadian provinces, investment and growth also respond positively when a VAT – a better consumption tax – replaces a less efficient tax such as a retail sales tax (RST) that imposes (as most RSTs do) significant fiscal burdens on capital goods.22 The beneficial effects of VAT on economic growth have not gone unquestioned in the literature. As Piggott and Whalley (2001) noted, both self-supply and informal sector activities become tax preferred as VAT base broadening occurs. Using 1994 Canadian data from the substitution of the manufacturer’s sales tax (MST) by the goods and services tax (GST), they found that the consumption of substitutable market production (such as restaurant meals) was reduced in favor of less efficient household production (consumption of own food), thus reducing welfare.23 Emran and Stiglitz (2005) in effect extend this argument in the form of a theoretical model of the effect of replacing trade taxes by VAT in developing countries with a large informal sector. Owing to the incomplete coverage of VAT due to the existence of a large informal sector, their results suggest that, under conditions they consider plausible, replacing an import tariff by a VAT that falls disproportionately on the formal sector may reduce welfare. Taking a different tack, Hines (2004) suggests that increasing consumption taxes will foster the expansion of the (presumably less

250

The Elgar guide to tax systems

productive) informal sector provided that, as seems plausible, the laborintensity of production in that sector is greater than in the formal sector. On the other hand, such an increase may obviously expand employment, as was noted recently in Copenhagen Economics (2007).24 Two offsetting arguments have been put forward in the literature. First, as Keen (2008) has observed, the evidence is that VAT actually functions, in part, as a tax on the informal sector because it falls on their purchases from formal sector businesses as well as on imports.25 Second, as Auriol and Warlters (2005) note, even governments aware of such problems may nonetheless choose to impose higher taxes, including VAT, on the formal sector of the economy. This is simply because, with their relatively weak tax administrations, the best way for them to raise revenue may be to increase barriers to entry to the formal sector, thus creating ‘rents’ that may then be taxed. We return briefly to the question of informality later. Trade More prominent than either growth or welfare in the original adoption of VAT in Europe was the argument that trade would be facilitated by turning the various sales taxes then existing in the member states of the nascent European Union into true destination-based consumption taxes both by ‘untaxing’ exports (and removing hidden subsidies) and by placing the taxation of imports and domestic production on a level playing field. While the theoretical necessity for this step in a world of at least imperfectly flexible exchange rates is still debated by some, on the whole VAT’s effects on trade have been considered to be largely beneficial, with economists applauding the level playing field for imports and governments generally paying more attention to the removal of barriers to exports.26 As with growth, however, the superiority of VAT in promoting trade has not gone uncontested. For example, as Ebrill et al. (2001) note, the model underlying the argument in the previous paragraph (e.g., Feldstein and Krugman, 1990) is based on such strong assumptions as uniform taxation and the absence of both revenue and intergenerational wealth effects – assumptions that are invariably violated in practice. The effect on trade of replacing other taxes by a VAT inevitably depends heavily upon the relative size of various elasticities and marginal reactions. Unfortunately, the empirical literature reaches no clear conclusions on these matters. Using a sample of 136 countries in 2000, Desai and Hines (2005), for example, conclude not only that countries relying on VAT have fewer exports and imports (relative to GDP) than countries without a VAT but also that the negative correlation between VAT and trade (the sum of exports and imports) is stronger for low-income countries. A similar pattern appears in an unbalanced panel of 168 countries from 1950–2000, in which VAT

Value-added tax 251 use is associated with 12 percent fewer exports and persists with the inclusion of income and geographic controls. Desai and Hines (2005) conjecture that these effects may arise, first, because VAT tends to be imposed at higher rates on traded goods than on non-traded goods, and, second because many countries collect VAT promptly on imports, but fail to provide VAT export rebates in a timely and complete fashion. The result is thus that in some ways a VAT continues to have tariff-like effects, reducing trade, as Edmiston and Fox (2006) have also argued. In contrast, Keen and Syed (2006), using a panel data set for OECD countries from 1967 to 2003 to examine the effect of the VAT on export performance, support the view that the VAT is inherently trade-neutral. Ligthart and van der Meijden (2010) go further in their study of the revenue, efficiency, and distributional implications of offsetting tariff reductions by increases in destination-based consumption taxes so as to leave consumer prices unchanged: they find that such reforms would increase government revenue, imports, and exports. Base broadening The ‘economic’ component of improving VAT in practice generally comes down to some form of structural reform to broaden the tax base. As noted earlier, it is common (IMF, 2010) to argue that many countries can and should expand VAT revenues – and, not so incidentally, simultaneously improve the economic efficiency of their tax systems – by broadening the tax base. Once again, however, most statements to this effect rest more on assertion than on evidence. Consider, for example, the three major sectors currently ‘favorably’ treated under most VATs: finance, housing, and the public sector. Is the case really so clear for including these sectors within the scope of the VAT? And if they are so included, would the effects be both economically beneficial and also yield revenue? We consider each of these sectors briefly in turn.27 Typically, the financial sector is exempt, apart from a few financial services. On average, a quarter of GDP in developed countries originates in the financial sector, while the corresponding figure is around 10 percent for developing countries (Zee, 2004). Excluding this large part of the potential base must, it would seem, both reduce revenues and also distort resource allocation. Various approaches have been proposed to allow the taxation of financial transactions under the VAT and some countries do so.28 To sum up this complex issue rather cavalierly: it can be done; the logic of VAT suggests that it should be done; and if it is done, theory suggests that the result should be a reduction in tax-induced distortion. On the other hand, since most financial transactions take place between businesses, and no one wants to tax ‘pure’ interest under a consumption

252

The Elgar guide to tax systems

tax, there is unlikely to be much net revenue collected from extending the VAT to encompass all financial services. While developed countries like those in the EU should undoubtedly continue to explore this question and to expand the extent to which financial services are now taxed – as New Zealand and a few other countries (South Africa, Australia, Singapore) have already done in various ways – it thus seems unlikely, as Bird and Gendron (2007) argue, that many developing countries should venture much further into these deep and difficult waters beyond taxing explicit fees charged for financial services. The story is somewhat different with respect to taxing services provided by the public sector, not-for-profit activities, and charities (the PNC sector).29 As Wassenaar and Gradus (2004) note, for example, failing to tax public provision of services like water and refuse removal while taxing such services when provided by the private sector biases market decisions. In principle, governments engaged in providing private services (whether through direct activities or public enterprises) should presumably both be subject to output tax and be able to claim input tax credits, as should other non-profit providers. Indeed, it makes allocative sense to require public bodies to pay VAT on the inputs they buy from the private sector even if their own outputs are not taxed so that budget decisions are made on a full (VAT-included) opportunity cost basis. Of course, governments may for policy reasons choose explicitly to offset such costs in part by ‘rebating’ VAT on such inputs; Canada does this to varying extents for local governments, providers of health and education services, and charities (Bird, 2009). In the end, as this example perhaps suggests, the case for extending the VAT base into the PNC sector is not so much to increase or even to improve resource allocation within the private sector as to improve public sector budgeting decisions. Perhaps the major component of consumption now not taxed in many countries that might result in a potential gain in VAT revenue – and that would also do much to reduce a particularly prevalent distortion in resource allocation – would be to extend the tax base to encompass real property more adequately. At present, housing, and more generally real property, is taxed to very different extents and in very different ways even within the EU, let alone across the VAT universe as a whole.30 Most commonly, not only are the ‘consumption services’ received by owner-occupants exempt (as they are also from income tax) but so also are residential rentals. In many developed countries new buildings (and improvements) are subject to tax – in effect, a ‘prepaid’ VAT. But in many developing countries real property as a whole often goes virtually untaxed. There are, of course, administrative and political problems in extending VAT further into this sector, not least those arising from the

Value-added tax 253 existence of various special taxes (often at sub-central levels of government) on property transfers as well as property values. Nonetheless, both revenue concerns and significant administrative and economic arguments suggest that many countries could usefully consider how and to what extent they might improve the treatment of real property under their VATs.31 Sub-central VATs A final topic that may perhaps be mentioned briefly under the heading of ‘VAT economics’ is that the case for VAT as a replacement for other forms of sales tax is not limited to the central government.32 In a number of countries, notably such federal countries as the United States, Canada, Brazil, India, and Argentina, sales taxes constitute a major source of revenue for regional (and in some cases local) governments. Indeed, in the US case, state dependence on sales taxation has been a major reason that the national government has been reluctant to enter into serious consideration of a national VAT. In neighboring Canada, however, not only have VATs existed at both levels for some years, but the replacement in 2010 of long-standing RSTs by provincial VATs in two additional large provinces (Ontario and British Columbia), once again underlined how generally successful Canada’s ‘dual VAT’ experience has been.33 Two other federal countries with (some form of) VAT at both levels of government, Brazil and India, have encountered many more problems, however. In both cases, one problem is that the national VAT is very limited in scope and hence unable to provide the ‘sheltering’ effect that Bird and Gendron (1998) argue helps make the regional VATs in Canada, which are imposed at rates set by the provinces on essentially the same base as the national VAT, work so well. Of course, as developing countries, both Brazil and India face much more difficult economic circumstances and have relatively lesser administrative capacity than Canada. Brazil, the pioneer (in 1967) of regional VATs, got around some potential problems by imposing the taxes on an origin basis; however, by doing so it paid a price in the form of both economic distortions and fraud. India, which launched its state VATs, is still to reform its central VAT and to determine how its ‘dual VAT’ system is to work – for example, how interstate trade is to be treated.34 Argentina, as yet, remains with its longstanding system of a central government VAT and state gross receipts tax, just as the United States remains with no national VAT and most states (and many localities) with RSTs. In these and other countries, however, it seems likely that more will be heard in the future about sub-central VATs.35

254

The Elgar guide to tax systems

VAT Administration Is VAT easier or more difficult to administer than other general consumption taxes? In principle, not only should a perfectly functioning VAT collect the same revenue from the same base as a perfectly functioning RST but the two taxes would also have the same taxpayer incidence and economic effects. However, in practice, the two taxes are likely to differ substantially in both coverage and administration. Few RSTs have succeeded in either taxing many service activities or excluding many business inputs from taxes (Smart and Bird, 2009). The number of firms taxed may be larger or smaller under the VAT depending on the threshold level set for the VAT and the effectiveness with which the RST is applied, but on the whole, VATs, while cheaper to administer than income taxes (GAO, 2008), appear to be more expensive to administer than RSTs. However, since VATs generally not only encompass a broader base than most RSTs but also as a rule tend to be enforced more tightly, such comparisons tell us little.36 On the other hand, some administrative advantages commonly alleged for VAT are unpersuasive. For example, VAT is sometimes said to increase revenue security since the VAT is collected at each stage of the production–distribution chain while the RST is collected only at the point of the final sale of the commodity. This argument is invalid because under a properly functioning VAT the government only receives revenue once the final sale is made.37 VAT collected at stages before the final sale – that is, the first sale to a non-registrant, whether a consumer or an informal producer – is creditable against subsequent sales and hence generates no net revenue for government, except for a possible small ‘cash flow’ gain depending on the turnover period between various stages in the chain of transactions and the risk-free interest rate. VAT has also been touted as being ‘self-enforcing’ because, under the transaction base invoice-credit system commonly employed, in principle each intermediate buyer has an incentive to make sure that the VAT has been properly charged by the seller so that a refund credit can be applied for. This advantage too is usually greatly overstated. In reality, VAT is definitely not self-enforcing for a number of reasons. Since VAT is not in fact enforced on a transaction basis but on an accounts basis, the real advantage of the ‘chain’ of VAT transactions is to provide an evidentiary trail that the tax administration may follow both to check the validity of the output and input taxes reported by the business in question and also to detect prior or subsequent gaps that may exist in the chain. Of course, to reap any gains from this process actual cross-checking and auditing is required: nothing happens automatically.

Value-added tax 255 Moreover, since in many countries purchasers are not legally liable if they fail to verify that those who sell to them actually paid the VAT shown on the input invoices for which they claim credit, there is not much incentive for them to do so. Recently, web-based technology has for the first time, made instantaneous verification of the VAT status of sellers (and purchasers) possible, as we discuss further below. Prior to this development, in most countries it was highly unlikely that firms accepting ‘false’ VAT invoices on purchases and claiming credits as a result would be prosecuted. Finally, the existence of a chain of VAT transactions has recently been argued by some to be more of a danger than a gain to the revenue. Under a VAT as usually administered, when a firm issues a VAT invoice it in effect creates a potential claim on the public budget. No country is short of those who can and will devise schemes to exploit this potential weakness by issuing false invoices, claiming false credits, and even (when, as with exports, there is zero-rating) claiming false refunds, as we discuss next. VAT evasion GAO (2008), drawing on the experiences of five developed countries (Australia, Canada, France, New Zealand, and the United Kingdom), recently emphasized that in the real world any and all VAT designs have compliance risks. All five countries are concerned about illegitimate businesses submitting fraudulent refund claims based on false paperwork. Particularly difficult to deal with are so-called ‘missing traders’ who set up businesses for the sole purpose of collecting VAT on sales or sometimes simply to claim (often excessive) rebates on ‘investments’ in equipment and inventories that entitle them to a refund – and then disappear with the proceeds. Such cases have proved particularly troublesome when crossborder transactions are involved. Attempting to deal with such problems either add complexity to tax design or requires substantial administrative resources; in either case, the result is likely to be increased administrative and compliance costs. Keen and Smith (2006) describe the main types of fraud and evasion to which the VAT is liable, giving special attention to the risks that arise from the zero-rating of exports. Frauds such as non-registered taxpayers, misclassification of commodities, sales underreporting, omission of selfdelivered commodities, and the non-remittance of already charged tax are of course common to all sales taxes, especially those applied at the retail level. The main frauds peculiar to VAT are fraudulent claims for credit or refund such as credits claimed for VAT paid on purchases that are not creditable and especially credits generated by bogus (missing) traders – ‘invoice mills’ – set up solely to generate invoices for VAT credits or

256

The Elgar guide to tax systems

refunds. The so-called ‘carousel frauds’ that have attracted so much attention in the EU are an example. In these frauds, firms combine two features of the VAT – zero-rating of exports and the deferred payment mechanism adopted in the EU (in effect removing fiscal frontiers for imports also by collecting VAT otherwise due at the border on the next taxable transaction) with false invoices to defraud the revenue. In an early study, Agha and Haughton (1996) examined VAT compliance for a group of 17 OECD countries for 1987 and found that both multiple rates and a higher average rate were associated with lower compliance. On the basis of this slim database, they went even further, suggesting that the longer a VAT had been in place the better compliance was, that an extra dollar spent on VAT administration would on average raise revenue by $12, and that the revenue-maximizing VAT rate was less than 25 percent. Few studies since have been quite as daring, as we noted earlier when reporting the much less conclusive results found in more recent and much more extensive studies with respect to the effect of the time of VAT adoption on VAT performance.38 Estimates vary concerning the actual level of VAT fraud in the EU (IVA, 2007). Reckon (2009), in a study of the VAT gap in EU member states over the period 2000–06 using national accounts data, found a fairly stable pattern in the size of the gaps between 2000 and 2004, followed by a decline between 2004 and 2006, with considerable variation from country to country. With the financial crisis, however, matters took a turn for the worse (IMF, 2010): for example, in the United Kingdom, the VAT gap increased by 3 percentage points between 2007 – 08 and 2008 – 09. IMF (2010) estimates VAT evasion to be on average 0.7 percent of GDP in OECD countries and suggests that extra revenue equivalent to 0.8 percent of GDP could be collected in G-20 countries by reducing the VAT gap. Again, however, there is considerable inter-country variation. For example, the estimated VAT gap (the difference between actual and potential VAT revenues) was 20 percent in some (e.g., Mexico and Italy) but closer to 10 percent in others (e.g., France and Germany). Even larger gaps – over 40 percent – have been estimated in some developing and  transitional countries.39 In a recent EU study, Reckon (2009) identified the perceived level of corruption as the most significant factor explaining the variation in the size of the gap. There are many ways to cheat on any sales tax, and some of the frauds mentioned above are, in principle, more difficult to get away with when there is a VAT than is the case with other forms of (non-cascading) sales taxes. Although VAT fraud may be more obvious because it takes the form of explicit payments rather than simply lower revenues as with other forms of evasion, the net impact on the budget is the same

Value-added tax 257 in the end. The best way to deal with VAT refund fraud is similar to the best way to deal with most tax evasion: administer the tax better. From this perspective, as IVA (2007) recently noted, there are a range of possible solutions to controlling VAT evasion in the EU such as the increased use of technology to allow real-time declarations of transactions via VAT returns as well as improved administrative cooperation across borders.40 The correct treatment for VAT refunds is simply to pay legitimate claims promptly and not to pay fraudulent claims at all. The problem, of course, is how to distinguish the good from the bad. The answer is to be found not so much in special treatment of refund claims as in better administration of all aspects of the VAT system. Nonetheless, even in countries with well-established and experienced tax administrations like Germany, so much fraud has been uncovered in the form of illegitimate invoices that some have actually proposed that VAT refunds should not be paid unless satisfactory proof is shown that the input taxes claimed have been received by government (Sinn et al., 2004). Since 2002, Germany has made the buyer legally liable for tax not paid by the seller. However, this provision had little effect because it is virtually impossible to prove that a buyer had any knowledge of a seller’s intention not to pay the tax for which the buyer claims input tax credit. Another common form of VAT fraud is, as noted earlier, to establish a bogus new firm that claims credits for inputs it does not actually buy and then disappears before it can be audited. Germany tried to deal with this fraud by demanding some form of guarantee from new firms. But this measure too proved ineffective largely because firms were still able to make claims and go bankrupt before the authorities got around to acting. To avoid such problems, the scheme mentioned in the previous paragraph would have changed the law to require clear proof that all taxes claimed as input credits must actually have been paid. The scheme would require banks to remit the tax directly to the government at the time of sale through the device of an intermediate ‘trust’ account. At the same time the bank would issue a receipt to the seller for VAT paid, with this receipt serving as proof of the input tax claim. Alternatively, for cash payments, sellers would be required to issue a tax receipt that demonstrates that the tax has been paid (either in the form of a verified credit-card-like transfer to the government at the time of sale or by a prepaid ‘tax stamp’). The complexity of such schemes probably makes them inadvisable in the conditions prevailing in most developing countries. Several alternative approaches have, therefore, been proposed to curb VAT refund fraud in such countries (Harrison and Krelove, 2005). For example:

258 ●

●

● ●

The Elgar guide to tax systems The law might be changed, for example, to restrict zero-rating solely to exports to limit the potential range of legitimate refund claims. Similarly, any reduced rates should be high enough not to generate such claims except in highly unusual circumstances. Administrative procedures might be changed to ensure that export sales against which input tax claims are made are adequately supported by verified export entry forms. Refunds to new registrants might be made only after a mandatory six-month carry-forward of unused credits. Refunds might be limited only to firms in certain industries (as in China) or of a certain size (as is done in Quebec with respect to credits for capital goods).

Such methods may indeed make fraud less likely or less attractive. Unfortunately, not only do they increase the degree of cascading in the tax – thus negating some of its alleged virtues – but they create yet more barriers to the creation of new formal-sector businesses (not least because interest is seldom paid on carried-forward or deferred credits). In the end, as usual when it comes to tax evasion, it is not really possible to ‘design’ most of the problem away. One must also be willing and able to uncover and punish those who defraud the revenue. Closing the VAT gap through better design Evasion is of course only one part and by no means necessarily always the largest part of the ‘VAT gap’ shown in the performance measures discussed earlier. The other part of the gap is that attributable to inadequate base coverage. One aspect of this ‘base gap’ – sectoral exclusions or exemptions – was discussed above. Another important component may sometimes be that VAT exemptions (or zero-rating) are being used to reward political supporters or perhaps as an instrument of industrial or regional policy. Both factors appear to be at play, for example, in Ukraine (Bird, 2008) and in Pakistan (Martinez-Vazquez and Richter, 2009), and to some extent in Jamaica (Edmiston and Bird, 2006). No doubt other countries could be added to this list, but a much more important question in most countries is the extent to which the informal sector is, or is not, captured in the VAT net and how this is done. As Gordon and Li (2009) have recently emphasized, the level of financial development is a crucial factor in determining the extent to which countries can establish a secure revenue basis for public sector expansion. The other side of this coin is that the extent to which economic activity takes place in the largely unrecorded ‘informal’ sector is inversely related to the achievement of this goal.41 In most countries a small number of

Value-added tax 259 VAT registrants, usually less than 10 percent, account for 80–90 percent of VAT collections. Obviously, it is critical to keep a very close idea on these fiscal ‘whales’, as much recent literature on tax administration has argued (Baer et al., 2002). What has proved much more troublesome in VATs around the world is the question of how best to deal with the ‘minnows’ of the system – small firms that are at least potential VAT taxpayers. At least three distinct design issues arise in this respect. The first issue is where to set the threshold at which those engaged in business activities should be required to register for VAT. The second issue is, what, if anything, should be done to ‘simplify’ VAT procedures for small registrants. The third issue is how to make sure that those who are treated as ‘small’ for VAT purposes – that is, those excluded from the system or treated favorably within it – really are small. The threshold issue has been well discussed by Keen and Mintz (2004), who note that many countries have, somewhat oddly, set such low thresholds for VAT registration that their administrations end up being encumbered with a large amount of essentially useless work. Of course, it is not hard to speculate why countries may do this. For instance, since good tax administration rests on information – and for no tax is this truer than VAT – it is obviously advantageous in principle to include as large a share of economic activity in the tax base as possible in order to be sure to capture the necessary information. Such an explanation would be more convincing, however, if there were more evidence that countries put such information to good use. In the real world, however, the very countries that set unduly low thresholds often provide many of those thus caught in the VAT net with escape routes through various simplified systems or, in some cases, simple neglect. On the whole, as Keen and Mintz (2004) conclude, there is much to be said, particularly in the context of most developing countries, for deliberately excluding at least some of the potential tax base by setting a fairly high threshold and then concentrating administrative efforts on ensuring that those larger firms within the VAT net comply. A quite different approach to the perceived and real problems of dealing with small taxpayers is the so-called ‘VAT withholding’ found in some countries (e.g., Argentina). Under this system, the underlying assumption is that VAT will not be reported properly by small firms, regardless of whether they are registered or not. Hence larger firms (and public agencies) selling to or buying from such firms are required to ‘withhold’ an additional VAT on such sales or purchases in order to make up for the VAT those firms are supposed to collect (but are expected not to remit even if they do collect) on their own sales. Such ‘dual price’ systems usually appear to be imposed at rates that are both arbitrary and make no logical or administrative sense. No studies appear to exist of the extent to which

260

The Elgar guide to tax systems

such presumptively ‘withheld’ VAT is ever credited against VAT actually reported by the firms from whom they have been withheld, or of the net effect on revenue of such systems. On the whole, where this approach is in place it appears to amount to little more than a set of arbitrary presumptive levies that may be administered as part of the VAT but in reality have little or nothing to do with the proper administration and operation of the tax. Of course, it has long been recognized that compliance costs are relatively more burdensome for smaller firms and hence constitute a potentially important barrier to entering the formal sector of the economy. In an attempt to alleviate these problems, some countries have attempted to alleviate the blow in various ways, not always with good consequences. Perhaps the most common approach is in effect to take out of VAT most of the very firms that the unduly low threshold has brought in by applying some form of turnover or presumptive levy to firms below a (usually selfreported) threshold. The extreme version of this approach is the ‘simplified’ or ‘unique’ tax for small businesses found in some countries (Bird and Wallace, 2004). This approach may sometimes have the not inconsiderable virtue of allowing new and potentially growing firms to escape from often arbitrary tax administrative practices (Engelschalk, 2004). On the whole, however, not nearly enough attention seems to have been paid to the (lack of) dynamics in such systems and the barrier their existence may erect to both revenue expansion and to the real expansion of the formal economy. In the nature of the business of tax administration, not surprisingly, the customers are not very willing and often try to opt out of the system. Those who do so may include not only genuinely small businesses but also profitable large- or medium-sized businesses that only look small, as well as firms that are losing money but continuing to function by not paying over taxes such as VAT, for which in effect they are supposed to be withholding agents. Tax administrations often have to choose whether to go after the larger firms who are already in the tax net, where potential tax revenue payback may be higher, or to pursue instead the less lucrative smaller taxpayers who are largely outside that net. Many have chosen, perhaps rationally, to spend little time on the small but instead to attempt to cope both with them and to some extent with the whole shadow economy issue by adopting some form of specific presumptive tax regime in lieu of VAT, often accompanied by other taxes as well. Most discussion of the appropriate treatment of small firms (e.g., World Bank, 2007) appears to assume that there is no difficulty in telling which firms are small. As with giraffes, it appears that one is supposed to know one when one sees one. This assumption may be dangerously wrong. For example, some recent studies suggest that in at least some countries not

Value-added tax 261 only has the informal sector been becoming more, not less, important but also that persons and enterprises at all income (and size) levels are engaged to varying extents in the informal sector (de Ferranti et al., 2004). In fact, many businesses seem to operate in both the formal and informal sectors at the same time. As mentioned earlier, however, although firms operating in the shadow economy may escape VAT liability on their sales they are also not able to reclaim credit for any VAT paid on inputs, which, to the extent such activities buy from the formal sector, implies that one way to tax them is through a VAT. On the other hand, as Emran and Stiglitz (2005) have argued, increasing taxation of the formal sector may expand, not reduce, the amount of hidden economic activity taking place to the extent that market-based activities may be able to disappear into the shadow sector. In response to these problems, countries such as India have often engaged in various activities intended to chase ‘shadows’ back into the fiscal light. At one level, tax officials may simply walk along the street, sweeping hawkers and peddlers into the tax net, entering premises, checking records, and imposing penalties. A more sophisticated approach is to follow the audit trail down the productive–distributive value chain in each line of business, starting with those who are in the tax net and working outward on the assumption that it is difficult for even the most sophisticated evader never to have traceable contact with someone who is already known to the tax authorities. The existence of presumptions of various sorts in the VATs of many countries, whether in the form of presumptive gross receipts taxes in lieu of VAT, VAT withholding (as discussed above) or even simplified accounting systems, in effect constitutes formal recognition of the inability or unwillingness of the administration to apply the letter of the law to a large part of the potential taxpayer population. These measures – often touted as ways not only of raising some revenue from this hard-to-tax population but also encouraging the growth of small informal businesses into larger regular taxpaying businesses – may backfire badly. In a detailed exploration of a unique micro-database in Brazil, for instance, De Paula and Scheinkman (2009) show the key role of VAT enforcement in fostering ‘chains of informality’. The more suppliers (and purchasers) in the chain are subject to an effectively enforced VAT, the less likely are others in the chain to operate informally. On the other hand, where informality is tolerated in one stage, it is likely to spread both upstream and downstream. Importantly from the perspective of the argument made here, the application of presumptive rates to links in a value chain has the same result as tolerating informality: it reduces both tax compliance and formalization. Every effort should be made to avoid

262

The Elgar guide to tax systems

breaking the information chain, rather than encouraging firms to do so, as simplified systems in effect do. In practice, it is often difficult in many countries to distinguish between small firms who do not keep good books and records but are potentially (and legally) taxable and firms whose activities are clearly large enough to fall within the tax system but are tax evaders. Some in the latter group may be completely off the fiscal radar – the so-called ‘ghosts’ – while others are more like ‘icebergs’, in that the portion of their activities visible to the authorities may be miniscule compared with the hidden reality (Bird and Wallace, 2004). The introduction of ‘special’ regimes in VAT (or other taxes) in an attempt to cope with these circumstances inevitably results in the fragmentation of the tax system and is generally inherently inconsistent with good tax administration. Any time such a ‘disconnect’ is created between a special tax regime and the general tax system, problems are likely to emerge. It is, in the long run, no easier to have two national tax regimes than two national currencies because each regime constitutes an integral part of the other and affects the entire system. Even the best ‘special’ tax regime, whether the intent is to supplement a normal VAT by replacing its complexities with a simplified regime for small business or to extend the reach of the tax further out into the shadow economy, must therefore include explicit transition arrangements to link the special regime to the more general tax system. In practice, however, such arrangements seem seldom to exist. A key problem is how to keep out of the (simplified) system large and medium enterprises that try to look like small enterprises and hide themselves from the tax collector’s eye – and almost invariably from the attention of any auditors.42 The temptation to shelter from the fiscal blast within such systems is likely to be especially strong when, as is usually the case, the effective tax rates applied to those who make it to the ‘safe harbor’ of the simplified system are considerably lower than those in the normal tax system. Attempts to supplement a VAT by introducing a simplified system intended to help the small or to catch shadows may thus end up making matters worse particularly when, as is too often the case, firms once safely hidden in the ‘small’ sector can stay there almost indefinitely with little or no risk of audit or exposure.43 VAT Politics Taxation is always in the end more a matter of politics and economics, and politics is more often concerned with fairness and distribution than it is with economic efficiency.44 Certainly equity is always and everywhere a central issue in taxation. Equity issues may be approached at two different levels. First, one may consider the details of exactly how different

Value-added tax 263 taxes impose burdens on taxpayers who are in the same and different economic circumstances. Second, one may instead focus on the overall effects of taxation on the income and level of well-being of different people. Economists often tend to take the second approach. However, popular discussion of taxation usually takes the first approach, as shown by the many proposals to alter the rates and structures of particular taxes such as VAT. Sometimes such proposals may indeed improve horizontal and vertical equity within the limited group subject to the full legal burden of the tax. However, sometimes the same changes may actually exacerbate inequity more broadly considered. Because the poor consume a higher proportion of their income than the rich, consumption taxes are generally considered regressive. Consumption taxes are less regressive on a lifetime rather than annual perspective, but given the relatively short life expectancies in many developing countries and the subsistence level at which many people in such countries live such refined arguments carry little weight. It is thus not surprising to find that many VATs provide for reduced rates or exemptions for ‘basic’ items such as some foods, passenger transport, medical services, and cooking fuel. The common response of the expert is that whatever small degree of progressivity such deviations from uniform treatment may achieve could be more effectively and fairly attained through small changes in the income tax or by adjustments in transfer payments. Again, however, the relevance of this response is arguable, if not largely irrelevant, in countries in which the poor do not as a rule suffer from income tax or benefit from transfers. Although the equally conventional argument that there is unlikely to be much gain in imposing differential ‘luxury’ rates under a VAT given the efficiency and administrative costs to which such differentiation gives rise, seems more convincing – if desired, more can be done with less collateral damage through excise taxes on such commodities (Cnossen, 2004). But, on the whole the conventional case for imposing VAT at a uniform standard rate and on as broad a base as possible in poor developing countries seems less than completely convincing. A uniform VAT is likely to increase the price of many goods essential to the poor (Ahmad and Stern, 1987). Because the poor may consume a relatively small amount of such products, it is undoubtedly true that much of the benefit of such exemptions will go to the non-poor. Nonetheless, in view of both the relatively heavy tax burden from such taxes on the poor in some countries and the general inability of governments in those countries to provide offsets to increased tax burdens through other fiscal adjustments, some form of offset measures with the VAT may often seem warranted. However, going to the logical conclusion of zero-rating for distributive reasons is unlikely to be advisable in countries already facing many

264

The Elgar guide to tax systems

difficulties with VAT refunds. Exemptions are similarly bad: they may both increase cascading and, by breaking the VAT chain, make effective enforcement more difficult. In the end, perhaps a reduced rate might be the best – or the least worst – approach; at least some form of this argument may be part of the rationale behind the common reduced rates found in many EU countries. On the whole, however, there are too many instances in which the items taxed (or not taxed) in different ways appear more to have been chosen arbitrarily rather than in any reasoned fashion to make one comfortable with this conclusion.45 As with all taxes, these issues must ultimately be decided through the political process, whether in developing or developed countries. Nonetheless, the fact remains that concerns about ‘fairness’ and the regressivity of the VAT have proved critical in both the adoption and the shaping of the VAT in many countries and that such matters cannot and will not be relegated to the whims (or reasoned conclusions) of tax policy designers. Although this issue has certainly been prominent in VAT discussions in many poor countries46 the fact that the very poor are likely to remain largely outside the formal sector where the VAT operates, actually means that at least in some such countries VAT may be progressive (Jenkins et al., 2006). As in developed countries, however, where not only did ‘fairness’ play an important role in the initial adoption of VATs (Eccleston, 2007) but the issue has again come to the fore as a result of proposals to raise VAT rates to deal with the deficits arising from the 2008 financial crisis, the distributional concerns aroused by VAT continue to require close attention from those concerned with tax policy in developing countries.

6

ONWARD AND UPWARD?

Where does VAT go from here? Onward even into every little island and even oil-rich country, or so it seems. As Bird and Gendron (2006) note with respect to developing countries, all countries need taxes, most need a general consumption tax, and VAT, while not perfect, is the best such tax available. Indeed, as noted earlier, in at least some countries – Australia, Spain, Mexico, Argentina – regional governments too may gradually find their way onto the VAT bandwagon. Of course, whether ‘American exceptionalism’ can continue to withstand the worldwide rush to VAT remains to be seen, but this story is far too complex to tackle here. Recently, the IMF (2010) has stressed VAT’s revenue potential, noting, for instance, that introducing a VAT in the United States, and doubling the low VAT rate in Japan, could substantially raise revenues in those

Value-added tax 265 countries, just as strengthening tax compliance could do in many developing countries. It does not follow, however, that VAT in all or even most countries will over time come to account for a larger and larger share of revenues. As noted earlier, the evidence that VAT is a ‘money machine’ is far from clear or persuasive. As was also discussed earlier, in economic and administrative terms a case can indeed be made in many countries for increasing VAT rates, VAT bases, the effectiveness of VAT administration or perhaps even all three. However, it does not follow that countries will or indeed necessarily should do so. One reason lies in VAT politics in the broadest sense: in countries in which popular support for the state often needs shoring up rather than shaking up, it is often far from clear that increasing VAT is always the way to go. A case can be made in many poorer countries, for example, that a critical ingredient in the creation of a sustainable and legitimate governance structure may lie more in strengthening the income tax than the VAT (Bird and Zolt, 2010). Equally, a case can be made that many countries may have more to gain by strengthening and improving sub-central governments and that this too may often be best accomplished in ways other than simply bestowing VATs – or at least VAT revenues – on them (Bahl and Bird, 2008). IMF (2010) is right in the sense that, provided it can be administered adequately, in most countries VAT remains the most economically desirable and administratively effective way to collect an additional share of national income through a general consumption tax. However, in many countries the reality is that the VAT is not all that well designed and certainly not always well administered and, more importantly, that the political economy environment may be such as to make serious VAT reform, let alone VAT revenue expansion, much more difficult than some of the rather casual assertions in IMF (2010) appear to suggest. In the United States, for example, a VAT at 13 percent might indeed raise 6 percent of GDP (Graetz, 2005) but it is probably close to impossible that the net revenue gain from such a major reform would come even close to this level.47 While much can certainly be done to improve VAT, and indeed to introduce it in the few countries, like the US, where it does not now exist, what happens in reality in any country is inevitably more dependent on political developments than on technical improvements in VAT design or administration. In the end, with VAT as with all taxes, the ‘NOSFA principle’ (no one size fits all) is thus likely to prevail. Although the questions that must be answered in designing VAT and the problems faced in implementing it are in principle similar everywhere, the contexts (political, human resources, technological, etc.) within which these questions necessarily have to be answered differ significantly from country to country and indeed change

266

The Elgar guide to tax systems

over time in any one country. Countries may of course usefully learn much from the experience of others, and so they have. In the final analysis, however, as the development of VAT over the last half-century has clearly demonstrated, the best answers and solutions for any country – those that will be accepted, sustained over time, and produce beneficial effects – must always and everywhere in the end be homegrown to fit their particular circumstances.

NOTES * 1. 2. 3. 4. 5.

6.

7.

8. 9.

10. 11. 12.

We are indebted to Violeta Vulovic for very helpful research assistance and to Yongzheng Liu, Nicholas Warner, and Tetyana Zelenska for data collection. VATs in different countries may of course have many different names, among which the most popular in recent years is the more easily comprehensible label goods and services tax (GST). See, for example, Chelliah et al. (2001), Ebrill et al. (2001), and Bird and Gendron (2007). ‘American exceptionalism’ is mentioned briefly in Section 6 below but not discussed in detail here. As an academic critique, see, for example, Emran and Stiglitz (2005). The break in the data between 1990 and 2000 is due to reporting irregularities and a sizable drop in the number of observations. The same observation applies to Figure 6.3. The unweighted average share of general consumption taxes – essentially the VAT – in OECD countries rose from only 35.9 per cent of all consumption taxes (13.6 per cent of tax revenues and 3.8 per cent of GDP) in 1965 to 62.2 per cent of consumption taxes (18.9 per cent of tax revenues and 6.8 per cent of GDP) in 2006 (OECD, 2008). The years used for the three averages are indicated in the last two columns of Table 6.2; the actual years used vary by country depending on data availability. To a large extent, the three measures move together: the simple correlation between the VAT efficiency and C-efficiency ratios is 0.92 and that between the VAT efficiency and the VAT gross compliance ratios is 0.89. Similar distinctions have been made within groups of countries. For example, Glenday and Hollinrake (2005) in analyzing VAT performance among the Southern African Development Community (SADC) member states, observe that that the lowest per capita income countries tend to have the weakest tax capacities and lowest VAT efficiencies. As OECD (2009) demonstrates, however, administrative cost ratios are at best very questionable measures of tax administrative efficiency: for a fuller exploration of this question, see Vazquez-Caro and Bird (2010). In an early study, Agha and Haughton (1996) found the number of years since VAT to be a significant determinant of VAT efficiency, but a more recent study by Aizenman and Jinjarak (2008) found this variable did not turn out to be significant. In contrast to Aizenman and Jinjarak (2008), De Mello (2009) did not find urbanization to be significant, perhaps because most countries in the sample were already highly urbanized. In another study, McCarten (2006) found indexes of government’s capacity to control corruption and the cost of registering a new business to have some explanatory value with respect to C-efficiency, while (using a smaller sample of transitional countries) the prevalence of bribery was found to be negatively related to the same measure. The list of countries is in Appendix A. The panel is unbalanced since we were not able to collect data on all variables for all countries and periods in the sample. The Arellano-Bond estimator is designed for small-T large-N panels, as is a case with

Value-added tax 267

13. 14.

15. 16. 17. 18. 19.

20.

21.

22. 23. 24. 25. 26.

27. 28. 29. 30.

our sample (ideally, T=19, N=92 to 107). This estimator was first proposed by HoltzEakin et al. (1988). The same phenomenon, associated with financial development, rather than urbanization – though the two are highly correlated – is seen in, for example, Gordon and Li (2009). Given the common association with higher degrees of informality and lesser development, which itself is usually considered to be associated with lesser urbanization, this may seem odd. On the other hand, as De Ferranti et al. (2004), Chen (2005), and others have pointed out, there is increasing evidence in a number of developing countries that the level of informal activities in urban areas is large and increasing. This phenomenon may also provide a possible explanation for the differential effects of tax morale in developed and developing countries discussed next. Obviously, however, there is still much to be sorted out here. President’s Advisory Panel (2005, p. 192). To control for a potential endogeneity bias (since the adoption of a VAT may also be related with the desire to raise overall tax effort), the VAT dummy is itself explicitly estimated in a separate take-up regression equation. The information on the IMF programs is obtained from Dreher (2006, updated February 2010). See, for example, Wooldridge (2002, pp. 236–7). In STATA, equation Total Revenue 5 a1DVAT 1 b1Z1 1 g2DVAT * Z1 1 v1 is estimated using command ivreg rather than estimating first-stage and second-stage regression ˆ b2SLS) , manually, because the second stage leads to incorrect residuals (vˆ2 5 Y 2 X rather than correct ones (vˆ2 5 Y 2 Xb2SLS) . Using command ivreg avoids this problem (Davidson and MacKinnon, 2004; Baum, 2006). However, using a Cournot duopoly model the same authors also showed that the strategy of coordinated tariff and domestic tax reform that is welfare-improving under perfect competition may be undesirable in a setting of imperfect competition (Keen and Ligthart, 2005). Applying the OLS (ordinary least squares) would yield unbiased and consistent estimates for each separate equation only. The seemingly unrelated regression approach takes into account the variability across equations and yields BLU (best, linear, unbiased) estimates. Studies demonstrating the importance of this common characteristic of real-world RSTs have been carried out in both Canada (Kuo et al., 1988) and the United States (Ring, 1999). Other papers on household production reach similar results: see, for example, Boadway and Gavhari (2006). Sorensen (1997) found that this almost Gandhian argument favoring informality on employment grounds may have welfare benefits in some circumstances. Some evidence in support of this argument may be found, for example, in Glenday and Hollinrake (2005). The state of the theory is rather neatly shown in the contrasting results reported in Keen and Ligthart (2002, 2005). There is nothing wrong with either of the models examined in these papers. The problem is that no one has a good handle on which, if either, best models reality. As we discuss next, so far the empirical examination of this issue has not advanced matters much. Additional ‘base broadening’ issues – VAT thresholds and the treatment of small business and ‘equity-based’ rate favoritism, zero-rating, and exemption – are discussed briefly later. See, for example, the discussions in Boadway and Keen (2003), Gendron (2008), and Kerrigan (2010). For a detailed exploration of this subject, see Gendron (2005). For example, OECD (2008) lists at least ten different ways in which ‘supply of land and buildings’ is treated in EU member states.

268 31. 32. 33.

34.

35.

36. 37. 38.

39. 40.

41. 42. 43.

44.

The Elgar guide to tax systems The importance of the revenue base provided by the housing sector has recently been emphasized in Canada by Smart and Bird (2009). This topic may of course also be treated under both ‘VAT administration’ and ‘VAT politics’: for a discussion of all aspects of sub-national VATs in Canada, see Bird and Gendron (2010). A detailed examination of the Canadian experience may be found in Bird and Gendron (2010). While as GAO (2008) notes, this system is not costless for either governments or business, the offsetting reduction in compliance costs owing to the single administration of the two taxes does not appear to have perceptibly increased costs either. For further discussion of these issues in both Brazil and India, see Bird (2009). Canada deals with the problems mentioned by having one administration deal with both VATs, by using much the same ‘deferred payment’ system as in the EU for dealing with interstate trade, and, finally, by using what are in essence origin ‘place of supply’ rules for certain trade, notably in services, while avoiding the distorting public sector competition found in Brazil through an agreed revenue allocation system that allocates revenues to provinces on a statistically-determined destination basis. For further details, see Bird and Gendron (2010). For further discussion, see Bird (2010a). We do not discuss here such local taxes as Italy’s IRAP, Japan’s local enterprise tax, or the new French replacement for the taxe professionelle – all of which in economic terms resemble VATs imposed on an origin base to some extent: for an earlier treatment of such taxes, see Bird (2003). See the extended discussion of this point in Bird and Gendron (2007, Chap. 3). This point was demonstrated clearly in OECD (1988). Neither the issue of the revenue-maximizing rate nor the size of the revenue ‘pay-off ’ of additional administrative outlays appears as yet to have been rigorously examined empirically, although see Edmiston and Bird (2006) for a rather casual estimate of the former in a particular country (Jamaica). See Gebauer et al. (2005) as well as the Latin American studies summarized in Bird and Gendron (2007, p. 60). On the use of technology to improve tax administration in general, see Bird and Zolt (2008). Early experiments with invoice-matching to control fraud in VAT were not effective (Choi, 1990) but over time such a system became quite effective in Taiwan (Jenkins et al., 2003) and recently China has launched a full-fledged e-invoicing system to facilitate control (Winn and Zhang, 2010). The importance of international cooperation in the EU context has long been obvious: for example, Bird and Gendron (1998) suggested that the EU needed a ‘virtual VAT’ that, like the federal VAT in Canada, would cover essentially the same tax base as that of the member states and hence facilitate joint control of cross-border transactions. With the recent strengthening of the VAT Information Exchange System (VIES) system and increased online cross-country access to VAT registration data, the EU is to some extent moving in this direction. For example, Alm et al. (2004) estimate that on average the ‘shadow economy’ in both Africa and Latin America is 41 percent of GDP or more than twice the level found in OECD countries. For some (weak) evidence of such ‘migration’ into a simplified system, see Bird (2008). Since purchases from these taxpayers by regular VAT sellers cannot be used to claim input credits, an incentive exists for them to voluntarily (or under pressure from their customers) enter the VAT system. How effective this incentive is likely to be, however, is far from clear given the general difficulties in policing the fringes of the VAT system. For example, other registered sellers – some of whom themselves may be conducting significant ‘shadow’ business – may agree to issue VAT receipts in their own name, a practice that seems not unlikely in the context of countries with large shadow economies and generally weak tax auditing capacity. Of course there is much more to VAT politics than the issue of fairness or regressivity discussed here. See, for instance, the discussion of the importance of tax visibility in

Value-added tax 269

45. 46. 47.

Bird (2010b) as well as the broader discussion of the ‘state-building’ role of taxation (Brautigam et al., 2007) – the importance of designing and implementing taxes in such a way as to make different groups in society feel, correctly, that it is ‘their’ tax system, thus fostering a ‘high compliance’ rather than a ‘low compliance’ system (Bergman, 2003). Such interesting subjects must, however, be left for another day. Crawford et al. (2010) show clearly how complex it can be to decide such matters rationally even in countries with capable administrations and ample data. See, for example, the cases cited in Bird and Gendron (2007, p. 24). See, for example, the discussion of the US case in Gale and Harris (2010) as well as the state-level implications discussed by Duncan and Sedon (2010).

REFERENCES Agha, A. and J. Haughton (1996), ‘Designing Vat Systems: Some Efficiency Considerations’, The Review of Economics and Statistics, 78(2), 303–8. Ahmad, E. and N. Stern (1987), ‘Alternative Sources of Government Revenue: Illustrations from India, 1979–80’, in D. Newbery and N. Stern (eds), The Theory of Taxation for Developing Countries, New York: published for the World Bank by Oxford University Press, pp. 281–332. Aizenman, J. and Y. Jinjarak (2006), ‘Globalization and Developing Countries – A Shrinking Tax Base?’, NBER Working Paper No. 11933. Aizenman, J. and Y. Jinjarak (2008), ‘The Collection Efficiency of the Value Added Tax: Theory and International Evidence’, The Journal of International Trade and Economic Development, 17(3), 391–410. Alm, J., J. Martinez-Vazquez and F. Schneider (2004), ‘“Sizing” the Problem of the Hardto-Tax’, in J. Alm, J. Martinez-Vazquez and S. Wallace (eds), Taxing the Hard-to-Tax, Amsterdam: Elsevier. Arellano, M. and S. Bond (1991), ‘Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations’, Review of Economic Studies, 58(2), 277–97. Auriol, E. and M. Warlters (2005), ‘Taxation Base in Developing Countries’, Journal of Public Economics, 89(4), 625–46. Baer, K., O.P. Benon and J.A. Toro Rivera (2002), ‘Improving Large Taxpayers’ Compliance: A Review of Country Experience’, Occasional Paper No. 215, Washington, DC: International Monetary Fund. Bahl, R. and R.M. Bird (2008), ‘Subnational Taxes in Developing Countries: The Way Forward’, Public Budgeting and Finance, 28(4), 1–25. Baltagi, B.H. (2005), Econometric Analysis of Panel Data, New York: John Wiley and Sons. Baum, C.F. (2006), An Introduction to Modern Econometrics Using Stata, College Station, TX: Stata Press. Baunsgaard, T. and M. Keen (2010), ‘Tax Revenue and (or?) Trade Liberalization’, Journal of Public Economics, 94(9/10), 563–77. Bergman, M. (2003), ‘Tax Reforms and Tax Compliance: The Divergent Paths of Chile and Argentina’, Journal of Latin American Studies, 35(3), 593–624. Bird, R.M. (2003), ‘A New Look at Local Business Taxes’, Tax Notes International, 30(7), 19 May, 695–711. Bird, R.M. (2008), ‘VAT in Ukraine: An Interim Report’, in Robert W. McGee (ed.), Taxation and Public Finance in Transition and Developing Countries, New York: Springer, pp. 621–38. Bird, R.M. (2009), ‘Tax Assignment Revisited’, in J. Head and R. Krever (eds), Tax Reform in the 21st Century, New York: Wolters Kluwer, pp. 441–70. Bird, R.M. (2010a), ‘Subnational Taxation in Emerging Countries: Guidance from the Literature’, World Bank Working Paper.

270

The Elgar guide to tax systems

Bird, R.M. (2010b), ‘Visibility and Accountability – Is Tax-Inclusive Pricing a Good Thing?’, Canadian Tax Journal, 58(1), 63–76. Bird, R.M. and P.-P. Gendron (1998), ‘Dual VATs and Cross-Border Trade: Two Problems, One Solution?’, International Tax and Public Finance, 5(3), 429–42. Bird, R.M. and P.-P. Gendron (2006), ‘Is VAT the Best Way to Impose a General Consumption Tax in Developing Countries?’, Bulletin for International Taxation, 60(7), 287–96. Bird, R.M. and P-P. Gendron (2007), The VAT in Developing and Transitional Countries, Cambridge, UK and New York: Cambridge University Press. Bird, R.M. and P.-P. Gendron (2010), ‘The Canadian Example: The GST-HST-QST-RST “System”’, Tax Law Review, 63(3), 517–82. Bird, R.M. and S. Wallace (2004), ‘Is it Really so Hard to Tax the Hard-to-Tax? The Context and Role of Presumptive Taxes’, in J. Alm, J. Martinez-Vazquez and S. Wallace (eds), Taxing the Hard-to-Tax: Lessons from Theory and Practice, Amsterdam: North-Holland, pp. 121–58. Bird, R. and E. Zolt (2008), ‘Technology and Taxation in Developing Countries: From Hand to Mouse’, National Tax Journal, 61(4, Part 2), 791–821. Bird, R. and E. Zolt (2010), ‘Dual Income Taxation and Developing Countries’, Columbia Journal of Tax Law, 1(2), 174–217. Boadway, R. and F. Gahvari (2006), ‘Optimal Taxation with Consumption Taxes on Leisure or Labor as Substitutes’, Journal of Public Economics, 90(10–11), 1851–78. Boadway, R. and M. Keen (2003), ‘Theoretical Perspectives on the Taxation of Capital Income and Financial Services’, in P. Honohan (ed.), Taxation of Financial Intermediation: Theory and Practice for Emerging Economies, New York: published for the World Bank by Oxford University Press. Brautigam, D., O. Fjeldstad and M. Moore (eds) (2007), Taxation and State Building in Developing Countries: Capacity and Consent, New York: Cambridge University Press. Buettner, T., R. Hertz and C.W. Nam (2006), ‘VAT Introduction and WTO Membership: The Experience in Developing and Transition Countries’, mimeo, Ifo Institute. Chelliah, Raja J. et al. (2001), Primer on Value Added Tax, New Delhi: Har-Anand Publications Pvt. Ltd. Chen, M. (2005), ‘Rethinking the Informal Economy: Linkages with the Formal Economy and the Formal Regulatory Environment’, WIDER Research Paper No. 2005/10, Helsinki, April. Choi, K. (1990), ‘Value-added Taxation: Experiences and Lessons of Korea’, in R.M. Bird and O. Oldman (eds), Taxation in Developing Countries, Baltimore, MD: Johns Hopkins University Press. Cnossen, S. (2004), ‘VAT in South Africa: What Kind of Rate Structure?’, International VAT Monitor, 15(1), 19–24. Copenhagen Economics (2007), ‘Study on Reduced VAT Applied to Goods and Services in the Member States of the European Union Directorate General Taxation and Customs Union’, European Commission Taxation Papers No. 13. Crawford, I., M. Keen and S. Smith (2010), ‘Value-Added Tax and Excises’ , in J. Mirrlees, Reforming the Tax System for the 21st Century, vol. I, The Institute for Fiscal Studies. Davidson, R. and J.G. MacKinnon (2004), Econometric Theory and Methods, Oxford: Oxford University Press. De Ferranti, D. et al. (2004), Inequality in Latin America: Breaking with History?, World Bank Latin American and Caribbean Studies, Washington, DC: World Bank. De Mello, L. (2009), ‘Avoiding the Value Added Tax: Theory and Cross-Country Evidence’, Public Finance Review, 37(1), 27–46. De Paula, A. and J.A. Scheinkman (2009), ‘Value Added Taxes, Chain Effects and Informality’, PIER Working Paper No. 09-041, Penn Institute for Economic Research, Dept. of Economics, University of Pennsylvania, September. Desai, M.A. and J.R. Hines (2005), ‘Value Added Taxes and International Trade: The Evidence’, mimeo, University of Michigan.

Value-added tax 271 Dreher, A. (2006) (updated February 2010), ‘IMF and Economic Growth: The Effects of Programs, Loans and Compliance with Conditionality’, World Development, 34(5), 769–88. Duncan, H. and J. Sedon (2010), ‘Coordinating a Federal VAT with State and Local Sales Taxes’, Tax Notes, 31 May, 1029–38. Ebrill, L.P., M. Keen, J.-P. Bodin and V. Summers (2001), The Modern VAT, Washington, DC: International Monetary Fund. Eccleston, R. (2007), Taxing Reforms: The Politics of the Consumption Tax in Japan, the United States, Canada and Australia, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Edmiston, K. and R.M. Bird (2006), ‘Taxing Consumption in Jamaica’, Public Finance Review, 20(10), 1–31. Edmiston, K.D. and W.F. Fox (2006), ‘A Fresh Look at the VAT’, in J. Alm, J. MartinezVazquez and M. Rider, The Challenges of Tax Reform in a Global Economy, New York: Springer, pp. 249–66. Emran, S.M. and J.E. Stiglitz (2005), ‘On Selective Indirect Tax Reform in Developing Countries’, Journal of Public Economics, 89(4), 599–623. Engelschalk, M. (2004), ‘Creating a Favorable Tax Environment for Small Business’, in J. Alm, J. Martinez-Vazquez and S. Wallace (eds), Taxing the Hard-to-Tax: Lessons from Theory and Practice, Amsterdam: Elsevier. Feldstein, M. and P. Krugman (1990), ‘International Trade Effects of Value-Added Taxation’, in A. Razin and J. Slemrod (eds), Taxation in the Global Economy, Chicago: University of Chicago Press. Gale, W.G. and B.H. Harris (2010), ‘A Value-Added Tax for the United States: Part of the Solution’, Brookings Institution and Tax Policy Center, May. GAO (2008), Value Added Taxes, Lessons Learned from Other Countries on Compliance Risks, Administrative Costs, Compliance Burden and Transition, Washington, DC: United States Government Accountability Office. Gebauer, Andrea, Chang Woon Nam and Rudiger Parsche (2005), ‘VAT Evasion and Its Consequence for Macroeconomic Clearing in the EU’, Ifo Institute of Economic Research (November). Gendron, P.-P. (2005), ‘Value-Added Tax Treatment of Public Bodies and Non-Profit Organizations’, Bulletin for International Fiscal Documentation, 59(11), 514–26. Gendron, P-P. (2008), ‘VAT Treatment of Financial Services: Assessment and Policy Proposal for Developing Countries’, Bulletin for International Taxation, November, 494–507. Glenday, G. and D. Hollinrake (2005), Assessment of the Current State of VAT Implementation in SADC Member States, report prepared for the SADC Trade, Industry, Finance and Investment (TIFI) Directorate. Gordon, R. and W. Li (2009), ‘Tax Structures in Developing Countries: Many Puzzles and a Possible Explanation’, Journal of Public Economics, 93(7), 855–66. Graetz, M.J. (2005), A Fair and Balanced Tax System for the 21st Century, Washington, DC: American Enterprise Institute. Harrison, G. and R. Krelove (2005), ‘VAT Refunds: A Review of Country Experience’, IMF Working Paper No. WP/05/218. Hines, J.R., Jr. (2004), ‘Might Fundamental Tax Reform Increase Criminal Activity?’, Economica, 71(283), 483–92. Holtz-Eakin, D., W. Newey and H.S. Rosen (1988), ‘Estimating Vector Autoregressions with Panel Data’, Econometrica, 56(6), 1371–95. IMF (2010), From Stimulus to Consolidation: Revenue and Expenditure Policies in Advanced and Emerging Economies, Washington, DC: Fiscal Affairs Department, International Monetary Fund. IVA (2007), Combating Vat Fraud in the EU: The Way Forward, Brussels: International VAT Association. Jenkins, G.P., H. Jenkins and C.-Y. Kuo (2006), ‘Is the Value Added Tax Naturally Progressive?’, Queen’s Economics Department Working Paper No. 1059.

272

The Elgar guide to tax systems

Jenkins, G.P., C.-Y. Kuo and K.-N. Sun (2003), Taxation and Economic Development in Taiwan, Cambridge, MA: John F. Kennedy School of Government. Keen, M. (2008), ‘VAT, Tariffs and Withholding: Border Taxes and Informality in Developing Countries’, Journal of Public Economics, 92(10–11), 1892–1906. Keen, M. and J.E. Ligthart (2002), ‘Coordinating Tariff Reduction and Domestic Tariff Reform’, Journal of International Economics, 56(2), 489–507. Keen, M. and J.E. Ligthart (2005), ‘Coordinating Tariff Reduction and Domestic Tax Reform under Imperfect Competition’, Review of International Economics, 13(2), 385–90. Keen, M. and B. Lockwood (2006), ‘Is the VAT a Money Machine?’, National Tax Journal, 59(4), 905–28. Keen, M. and B. Lockwood (2010), ‘The Value Added Tax: Its Causes and Consequences’, Journal of Development Economics, 92(2), 138–51. Keen, M. and J. Mintz (2004), ‘The Optimal Threshold for a Value-added Tax’, Journal of Public Economics, 88(3–4), 559–76. Keen, M. and S. Smith (2006), ‘VAT Fraud and Evasion: What Do We Know and What Can Be Done?’, National Tax Journal, 59(4). Keen, M. and M. Syed (2006), ‘Domestic Taxes and International Trade: Some Evidence’, IMF Working Paper No. WP/06/47. Kerrigan, A. (2010), ‘The Elusiveness of Neutrality – Why is it so Difficult to Apply VAT to Financial Services?’, International VAT Monitor, 21(2), 103–12. Kuo, C.-Y., T. McGirr and S. Poddar (1988), ‘Measuring the Non-Neutralities of Sales and Excise Tax in Canada’, Canadian Tax Journal, 36(3), 655–70. Ligthart, J.E. and G. van der Meijden (2010), ‘Coordinated Tax-Tariff Reforms, Informality and Welfare Distribution’, mimeo, Tilburg University. Martinez-Vazquez, J. and K. Richter (2009), Pakistan Tax Policy Report: Tapping Tax Bases for Development, World Bank Report No. 50078-PK, July. McCarten, W. (2006), ‘The Role of Organizational Design in the Revenue Strategies of Developing Countries: Benchmarking with VAT Performance’, in J. Alm, J. MartinezVazquez and M. Rider (eds), The Challenges of Tax Reform in a Global Economy, New York: Springer, pp. 13–39. Norregaard, J. and T.S. Khan (2007), ‘Tax Policy: Recent Trends and Coming Challenges’, IMF Working Paper No. WP/07/274. OECD (1988), Taxing Consumption, Paris: OECD. OECD (2008), Consumption Tax Trends, Paris: OECD. OECD (2009), Tax Administration in OECD Countries; Comparative Information Series (2008), Paris: OECD. Piggott, J. and J. Whalley (2001), ‘VAT Base Broadening, Self Supply and the Informal Sector’, American Economic Review, 91(4), 1084–94. President’s Advisory Panel on Federal Tax Reform (2005), Simple, Fair and Pro-Growth: Proposals to Fix America’s Tax System, Washington, DC. Reckon (2009), ‘Study to Quantify and Analyze the VAT Gap in the EU-25 Member States’, Directorate General Taxation and Customs Union, European Commission Taxation Studies No. 0029. Ring, R.J., Jr. (1999), ‘Consumers’ Share and Producers’ Share of the General Sales Tax’, National Tax Journal, 52(1), 79–90. Sinn, H.-W., A. Gebauer and R. Parsche (2004), ‘The IFO Institute’s Model for Reducing VAT Fraud: Payment First, Refund Later’, CESifo Forum, 5(2), 30–34. Smart, M. and R.M. Bird (2009), ‘The Impact on Investment of Replacing a Retail Sales Tax by Value-Added Tax: Evidence from Canadian Experience’, National Tax Journal, 62(4), 591–609. Sørensen, P.B. (1997), ‘Public Finance Solutions to the European Unemployment Problem?’, Economic Policy, 25, 223–64. Vazquez-Caro, J. and R.M. Bird (2010), ‘Benchmarking Tax Administration in Developing Countries: A Systemic Approach’, paper presented to 9th International Conference on Tax Administration, Sydney, April.

Value-added tax 273 Wassenaar, M.C. and R.H.J.M. Gradus (2004), ‘Contracting Out: The Importance of a Solution for the VAT Distortion’, CESifo Economic Studies, 50(2), 377–96. Winn, J.K. and A. Zhang (2010), ‘China’s Golden Tax Project: A Technological Strategy for Reducing VAT Fraud’, available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_ id=1641379; accessed 25 April 2011. Wooldridge, J.M. (2002), Econometric Analysis of Cross Section and Panel Data, Cambridge, MA: MIT Press. World Bank (2007), Designing a Tax System for Micro and Small Businesses: Guide for Practitioners, Washington, DC: World Bank. Zee, Howell H. (2004), ‘A New Approach to Taxing Financial Intermediation Services under a Value-Added Tax’, IMF Working Paper No. WP/04/119. Zellner, A. (1962), ‘An Efficient Method of Estimating Seemingly Unrelated Regression Equations and Tests for Aggregation Bias’, Journal of the American Statistical Association, 57(298), 348–68.

274

The Elgar guide to tax systems

APPENDIX A DATA Our sample contains data for 107 countries between 1990 and 2008. However, we were not able to collect data on all variables for all countries in the sample over the 19 years period. List of countries Afghanistan, Albania, Argentina, Armenia, Australia, Austria, Kingdom of Bahrain, Bangladesh, Barbados, Belarus, Belgium, Bhutan, Bolivia, Bosnia and Herzegovina, Brazil, Bulgaria, Burundi, Cambodia, Cameroon, Canada, Cape Verde, Chile, China, Colombia, Republic of Congo, Costa Rica, Côte d’Ivoire, Croatia, Cyprus, Czech Republic, Denmark, Dominican Republic, Egypt, El Salvador, Estonia, Finland, France, Georgia, Germany, Greece, Guinea, Honduras, Hungary, Iceland, India, Indonesia, Iran, Ireland, Israel, Italy, Jamaica, Japan, Kazakhstan, Republic of Korea, Kuwait, Latvia, Lebanon, Lesotho, Lithuania, Luxembourg, Macedonia, Madagascar, Malaysia, Maldives, Malta, Mauritius, Mexico, Moldova, Mongolia, Morocco, Myanmar, Nepal, Netherlands, New Zealand, Nicaragua, Norway, Pakistan, Panama, Paraguay, Peru, Poland, Portugal, Romania, Russia, Rwanda, Serbia, Seychelles, Singapore, South Africa, Spain, St. Kitts and Nevis, Sweden, Switzerland, Syrian Arab Republic, Tajikistan, Thailand, Trinidad and Tobago, Tunisia, Turkey, Ukraine, United Arab Emirates, United Kingdom, United States, Uruguay, Vanuatu, Venezuela, Zimbabwe.

275

Imports Openness Petrol consumption Alcohol consumption GDP pc

Business concentration

Total (central government) revenue collection to GDP Income taxes (personal income tax, corporate income tax) to GDP Social Security contributions to GDP

Revenues Income taxes Social security contributions Taxes on goods and services General sales tax Customs duties VAT efficiency ratio C-efficiency ratio VAT gross collection ratio DVAT VAT experience Tax morale

=1 if the VAT is implemented Number of years since the VAT introduction Share of population that declares cheating on taxes to be never justifiable Average employment (in 000) in manufacturing (employment/number of establishments) Imports share in GDP (Imports + exports)/GDP Crude petrol per capita consumption (in 000 metric tons) Alcohol per capita consumption (in liters) Current GDP per capita (in USD)

General sales taxes to GDP Customs duties collection to GDP VAT collection/(standard VAT rate*GDP) VAT collection/(standard VAT rate*total consumption expenditure) VAT collection/(standard VAT rate*private consumption expenditure)

Sum of the general sales taxes and excises duties to GDP

Description

Variables description and sources

Variable

Table 6A.1

APPENDIX B

UNIDO Industrial Statistics Databases, authors’ calculations World Development Indicators World Development Indicators The UN Energy Statistics Database World Health Organization World Development Indicators

World Development Indicators, IMF GFS Database 2010, various sources for the VAT rates, authors’ calculations Ebrill et al. (2001) and various other sources World Value Survey

IMF GFS Database 2010, authors’ calculations

Source

276

Share of agriculture in GDP Share of urban in total population Age dependency ratio (dependents to working-age population) Education index (measures a country’s relative achievement in both adult literacy and combined primary, secondary and tertiary gross enrollment) Bureaucracy quality (between 0 and 4 points, higher points being given to countries where the bureaucracy has the strength and expertise to govern without drastic changes in policy or interruptions in government services) Regime durability: The number of years since the most recent regime change. = 1 if federal country = 1 if developed country Number of countries in the region having the VAT implemented (regions defined based on Ebrill et al., 2001) The IMF non-crises program (SAF, PRGF) The IMF crises program (SBA, EFF)

Agriculture Urbanization Age dependency Education

The World Bank The World Bank Ebrill et al. (2001) and various other sources Dreher (2006, updated February 2010)

The Polity IV

International Country Risk Guide (ICRG) , The PRS Group

United Nations World Development Indicators

World Development Indicators

Source

Note: IMF NCR = the IMF non-crises program (SAF, PRGF); IMF CR = the IMF crises program (SBA, EFF); SAF = the Structural Adjustment Facility; PRGF = the Poverty Reduction and Growth Facility; SBA = Stand-By Arrangement; EFF = the Extended Fund Facility.

IMF NCR IMF CR

Federal Developed Neighbors

Political durability

Bureaucracy

Description

(continued)

Variable

Table 6A.1

Value-added tax 277 Table 6A.2

Variables descriptive statistics

Variable

Obs

Mean

Std. Dev.

Min

Max

Revenues Income taxes Social Security contributions Taxes on goods and services General sales tax Customs duties VAT efficiency ratio C-efficiency ratio VAT gross collection ratio DVAT VAT experience Tax morale Business concentration Imports Openness Petrol consumption Alcohol consumption GDP pc Agriculture Urbanization Age dependency Education Bureaucracy Political durability Federal Developed Neighbors IMF NCR IMF CR

1469 1369 1127

0.276 0.060 0.062

0.101 0.042 0.050

0.046 0.000 0.000

0.662 0.224 0.209

1407

0.107

0.528

0.000

0.198

1345

0.051

0.030

0.000

0.156

1368 1084

0.022 0.357

0.039 0.142

0.000 0.002

0.384 0.962

1040

0.461

0.189

0.003

1.343

1084

0.589

0.261

0.003

1.959

2033 1248 930 1491

0.724 15.057 0.600 0.089

0.447 11.084 0.154 0.243

0.000 0.000 0.205 0.001

1.000 48.000 0.979 5.786

1958 1958 1490

0.445 0.845 0.002

0.253 0.493 0.005

0.001 0.000 0.000

2.156 4.570 0.038

1942

5.787

4.212

0.000

21.600

1975 1862 2033 1992 1734 1599 1833

9 847.8 0.125 0.587 0.602 0.819 2.540 28.1

13 496.0 0.117 0.224 0.163 0.165 1.101 34.8

85.5 0.000 0.054 0.252 0.000 0.000 0.000

109 903.1 0.659 1.000 1.100 0.993 4.000 199.0

2033 2033 2033 2033 2033

0.187 0.280 0.677 0.032 0.090

0.390 0.449 0.244 0.176 0.286

0.000 0.000 0.000 0.000 0.000

1.000 1.000 0.960 1.000 1.000

7

The economics of excise taxation Sijbren Cnossen

1

INTRODUCTION

Excise duties used to be called the orphans of tax policy, because they received relatively little attention in the tax literature. This has changed greatly in recent years, due to growing awareness of the detrimental health effects of smoking and excessive drinking, as well as the social costs associated with the phenomenal increase in traffic. Perhaps more importantly, the environmental problems caused by the burning of fossil fuels have led to a burgeoning literature on the use of ‘corrective’ excises to restrain harmful emissions. Attention has also been given to the use of higher-than-average taxes on goods and services regarded as items of luxury consumption in order to enhance the progressivity of the overall tax burden distribution. As implied by these examples, excise systems can be defined to comprise all selective taxes or duties, related levies, and charges on tobacco, alcohol, gambling, petroleum products, motor vehicles, and other specific goods, services, and activities.1 Broadly speaking, the distinguishing features of excise taxation are selectivity in coverage, discrimination in intent, and often some form of quantitative measurement in determining the tax liability, along with the application of specific rates and physical controls over production for enforcement purposes. This contrasts with general consumption taxes, such as value-added taxes (VATs) – sometimes referred to as goods and services taxes (GSTs) – whose bases are typically defined to include all goods and services (hereinafter called commodities) other than those specifically exempted. VATs, moreover, are levied primarily to raise revenue, while their liability (determined by imposing the statutory rate on the actual value of goods and services) is generally verified through checks on books of account and other documentary evidence, similar to business income taxes. This chapter reviews the economics of taxation.2 It falls into three parts. The first part examines the rationale of excise taxation by reference to the non-revenue objectives that are pursued through the imposition of the various duties. The second part discusses the instruments that can be applied, that is, duties, regulations, and permits. The third part reviews some issues – discrimination, coordination, and earmarking – that often arise in connection with excise taxation. 278

The economics of excise taxation 279

2

RATIONALE OF EXCISE TAXATION

The economic analysis of excise taxation starts with Atkinson and Stiglitz (1976) who prove that if the income tax schedule is chosen optimally, then under fairly reasonable conditions, social welfare cannot be improved by levying excise duties on commodities.3 But if the income tax is not optimal, excises have a role to play, because they are relatively efficient sources of revenue, improve resource allocation by internalizing the external costs associated with the consumption or production of excisable products, discourage the consumption of products considered harmful, serve as a proxy for charging road users for the cost of government-provided services, or promote progressivity in taxation. Revenue-raising Efficiency Aspects In practice, most excises have probably been enacted for revenue purposes, the main consideration being that they could be administered more easily than other taxes. Excise duties on tobacco, alcohol, petrol, and motor vehicles are good potential sources of revenue, because the products are easy to identify, the volume of sales is high, and the fact that there are few producers – wine excepted – simplifies collection. Also, there are few substitutes that consumers would find equally satisfactory, so that consumption, and by extension revenue, remain high despite excise-induced price rises. In the OECD area, excise receipts (item 5120 only) account for on average some 11 percent of total tax revenue, defined to include social security contributions (OECD, 2009).4 The differentially higher taxation of excisable goods for revenue purposes also has an economic rationale. The absence of close substitutes for addictive or indispensable products, such as tobacco, alcohol, and energy, implies that the demand for them is inelastic. In turn, this means that the potential for distortion of economic decisions by the imposition of excise duties is relatively small. In economic jargon, the non-distortionary income effect outweighs the distortionary substitution effect. More generally, Ramsey (1927) has shown that, under very restrictive conditions, the total excess burden or deadweight loss of product taxation can be minimized by setting tax rates so that the percentage reduction in the quantity demanded of goods, which results from taxation, is the same (taking into account cross-price effects on demand). The Ramsey rule can be reinterpreted in two ways depending upon one’s assumption about utility functions. If it is assumed that demand is independent for each good (i.e., cross-elasticities are zero) and that each good represents a trivial fraction of expenditures (so that income effects

280

The Elgar guide to tax systems

can be ignored), the result is the so-called inverse elasticity rule. This rule states that the optimal tax rate on each good is proportional to the inverse of the price elasticity of demand for that good. The intuition is that the least distortionary tax system hits harder those goods for which demand is invariant to its own price. Crawford et al. (2010) point out that the implications of the inverseelasticity rule can be dangerously misleading. For one thing, if the assumption that each good represents a trivial fraction of expenditures is relaxed (as it must be for tobacco, alcohol, and energy), an increase in tax on that good has effects on the demand for other goods, particularly if related, such as beer vs. wine and spirits. By extension, there would be effects on distortions and revenue collections from other goods, while it would not be a priori clear whether or not the effects would be welfare improving. On the other hand, if it is assumed that utility is weakly separable between consumption and leisure (so that changes in the relative prices of goods do not affect labor supply), then the optimal tax should impose the same tax rate on all goods. The intuition is that if changes in the overall price of aggregate consumption relative to the price of labor do not distort labor supplied, then a tax system that does not distort relative prices between products is optimal. Here, the implication is that broadbased, uniform-rate commodity taxes, for example, VATs, are superior to the extent that labor supplied is invariant to changes in the price of consumption.5 Crawford et al. (2010) firmly reject weak separability between market consumption of goods and time in paid work. Some goods are more complementary to leisure than are other goods. Accordingly, as Corlett and Hague (1953) have proved, the second-best situation in which leisure cannot be taxed, can be moved closer to the first-best situation in which leisure would be taxed, by taxing the complementary goods at a relatively high rate (or, conversely, subsidizing goods that are complementary to paid work). As a result, the consumption tax approximates a lump-sum tax without excess burden, that is, there is no loss of welfare above and beyond the tax revenue collected. To see which goods would be eligible for differential tax treatment, Crawford et al. (2010) calculate estimates of commodity complementarities with leisure in the form of the impact of an additional hour worked on the budget (percentage) of a large number of commodity groups in household spending. Products found to be complements with leisure (in the sense of time not in paid work) and hence candidates for additional (excise) taxes include foodstuffs, domestic fuels, tobacco, children’s clothing, and public transport. Complements with work, on the other hand, include alcoholic beverages, food eaten out, motor fuels, and leisure items

The economics of excise taxation 281 (perhaps reflecting the use of such items as substitutes for time in producing relaxation). Accordingly, these products are candidates for lower-thanaverage taxes (presumably lower-than-standard VAT rates) or, possibly, subsidies. A similar reasoning applies to the relatively low taxation of market supplies of goods and services, for example, work in and around the home, that are close substitutes for self-supply. Nevertheless, the effects on product demands of hours worked are small and, as the authors point out, consideration should also be given to the evident practical costs of implementing differential tax rate structures. Externality-correcting Issues Furthermore, excises are often rationalized as charges for the cost that consumers or producers of excisable products impose on others, but that is not reflected in price. This effect on the utility or production possibilities of some other consumer or producer is called a negative externality. It means that the marginal cost of an individual consumer or producer’s action is less than the marginal cost of his or her action to society and, as a result, the individual engages in more of the activity than is socially optimal. Charging consumers or producers6 for external costs, which should induce them to reduce their activities to the socially optimal level, is known as the Pigouvian prescription (Pigou, 1920). The prescription is that efficient consumption or production can be achieved through the tax system by imposing an excise on the activity equal to the marginal cost of the damage caused to other people.7 The literature is replete with examples of smokers, boozers, and polluters who impose financial, physical, and psychological costs on others without being charged for them directly or indirectly (for example, through higher insurance premiums). Marginal costs are often difficult to identify and measure, however, because they depend on who does what, where, and under what circumstances. In practice, therefore, average external costs are estimated and a ‘pooling’ approach (akin to insurance) is adopted in charging for these costs. Perpetrators as a group meet the costs by paying a uniform excise calculated as the total external costs divided by, say, the number of packs of cigarettes, drinks consumed, or liters of gasoline used. This average-cost approach seems acceptable if damage – for example, through smoking or pollution – is approximately proportional to cost. But measurement problems come back in full force if there are threshold levels of consumption below which adverse effects are absent or attenuated – one or two glasses of wine per day are good for you. In this situation, ideally, Pigouvian taxes should be non-linear in the level of consumption and become exceedingly complex to design. Nevertheless, as shown by

282

The Elgar guide to tax systems

BOX 7.1 TAXING TO CONTROL EXTERNAL COSTS The external cost of consuming an extra unit of alcohol varies from consumer to consumer. A person who consumes two glasses of wine at home may not impose costs on others. An inebriated person who consumes a glass of beer in a public bar and then gets behind the wheel of his or her car may impose significant additional external costs. Excise duties on wine and beer do not distinguish between cases where the marginal external cost is high and cases where it is low. Nevertheless, overall welfare may still improve if the reduction in the external cost caused by the high-risk person is greater than the loss in consumer welfare of the low-risk drinker. This is illustrated in Figure 7.1. Curve DA reflects the demand for alcohol from a high-risk drinker (who imposes increasing costs on others per unit of drink), while curve DB reflects the demand from a low-risk drinker, where the external costs are zero. Imposing an excise on alcohol reduces the demand in both markets (from XA to XA1, and from XB to XB1). This would reduce Alcohol price (euro)

Marginal social cost

Price + Tax

a b

Externality

c

Marginal private cost

Price DB

XB1

Source:

XB

DA

XA1

XA

Alcohol consumption (units/person/period)

Pogue and Sgontz (1989) and Commonwealth of Australia (2008).

Figure 7.1

Dilemmas in taxing alcohol to control external cost

The economics of excise taxation 283

the satisfaction of the high-risk drinker net of taxes (area c), but this is less than the reduction in harm caused to others (area a + c), in which case society as a whole benefits from the reduction in high-risk consumption. However, the duty would also reduce the satisfaction of those whose consumption is low-risk (area b), reducing the benefit from the duty. Accordingly, the net welfare gain is a minus b. Much depends, of course, on the (relative) elasticity of demand for alcohol, which tends to be greater for low-risk than for high-risk drinkers. Pogue and Sgontz (1989) and replicated in Box 7.1, even uniform taxation may still improve overall welfare if the reduction in external cost caused by heavy drinkers is greater than the loss in consumer welfare of moderate drinkers.8 Sandmo (1976) has shown that Ramsey and Pigouvian optimal taxes can be applied sequentially: the least distortionary tax is levied on each good according to the Ramsey rule and then additional Pigouvian taxes are imposed on those goods that generate negative externalities. In many instances, moreover, goods eligible for Ramsey and Pigouvian optimal taxes are the same, for example, tobacco products, alcoholic beverages, gambling activities, and petroleum products. There is a presumption, therefore, that the level of taxation on these goods should be rather high. Information Failure and Internality-correcting Arguments Information failures are other instances that justify government intervention, even in the absence of explicit external costs. If the young are not fully cognizant of the detrimental health effects of smoking and drinking, then the excise could be used to raise the price of tobacco and alcohol for them and thus reduce their consumption. Apparently, this approach can be successful since research has indicated that the price elasticity of demand for cigarettes and alcoholic beverages among the young is, on average, twice the price elasticity among adults (see, e.g., Chaloupka et al., 2000).9 On the other hand, an increase in the excise to reduce consumption by the young would appear to be an unwarranted additional tax on habitual users. Alternatively, therefore, other instruments could be applied, for example, better dissemination of information on the health hazards of smoking and drinking, coupled perhaps with legislation restricting supply or (place of) consumption. More generally, the public health or, more broadly, public goods model

284

The Elgar guide to tax systems

views the main role of excise duties as simply to discourage the consumption of these products in aggregate (Crooks, 1989). The less people smoke and drink, or, for that matter, gamble, pollute, and drive, the argument goes, the better. And it is pointed out that a drop in the overall level of consumption also reduces the level of abusive consumption. The economic approach adopted in this chapter takes a narrower view on the social costs of (excessive) consumption by focusing on the externalities rather than the total costs borne by society and the consumer him or herself.10 The economic approach also pays more attention to the revenue-raising efficiency of excise taxation and its distributional consequences. Nevertheless, public health rather than economic considerations generally dominate the policymaking decision process. Thus far, it has been widely accepted among economists that the irrationality of the young (information failures) and externalities are the only reasons for government intervention. Beyond this, government intervention should be rejected as a form of paternalism. After all, the principle of consumer sovereignty implies that a rational person who weighs up all the costs and benefits of his actions should be free to smoke, drink, gamble, and pollute as long as he is fully informed about the consequences of his choice and does not impose costs on other people. The classic reference is to the Becker-Murphy (1988) ‘rational addiction’ model. Recently, however, the assumptions of this model have been questioned by Gruber and Köszegi (2001) and others, who argue that rational and fully informed adults can still be time-inconsistent in their behavior in the sense that they discount the short-term costs and benefits of their actions to a greater extent than the long-term effects. Instead of discounting the future exponentially, as in the Becker-Murphy model, Gruber and Köszegi argue that the discounted utility of a sophisticated hyperbolic consumer can rise if a tax is imposed.11 The reason is that the tax serves as a self-commitment device, which the private sector cannot perfectly supply. Without the device, consumers miss the incentive to control their short-term desires for their own longer-term well-being.12 Gruber (2010) suggests that the ‘internal’ costs from smoking a pack of cigarettes are US$35 per pack. The rejection of the a priori assumption of 100 percent rationality has received support in the economics literature. O’Donoghue and Rabin (2003), for instance, express fear that paternalism will involve regulatory capture or transactions cost in implementation, but nevertheless recommend further policy analysis. As they write (p. 191), economists should be more realistic about the nature of errors people make: The possibilities that 15-year-olds err in becoming tobacco addicts or that 25-year-olds err in borrowing heavily on their credit cards or that 35-year-olds

The economics of excise taxation 285 err in too wildly playing the stock market with their retirement savings all strike us as profoundly plausible and of great policy relevance.

In the same vein, Thaler and Sunstein (2003) stress that the antipaternalistic fervor expressed by many economists falsely assumes that people always make decisions that are in their best interest. In addition, they point out that most economists wrongly believe that there are viable alternatives to paternalism or that paternalism always involves coercion. They believe that people have self-control problems and that the goal should be to devise policies that help some people who are making mistakes, while minimizing the costs imposed on others. They dub this approach libertarian paternalism, which avoids random, arbitrary, or harmful effects, and steers people in directions that will promote their welfare.13 Benefit-charging Features Road (and similar transport) services resemble goods produced in the private sector that are used optimally when their price, commonly referred to as the economic user charge, equals the total social costs of operating the road network. Accordingly, as outlined by Smith (2006), road user charges should contain charges for the following main categories of uncharged external costs, viz.: ●

●

● ●

Consumption of road infrastructure, in the form of marginal road damage costs, that is, the physical wear and tear caused by vehicles using the road system. Environmental costs, including global and local air pollution (greenhouse gases, nitrogen oxides, which contribute to acid rain, and particulates, which can cause health problems). Noise pollution and landscape degradation would also fall under this heading. Congestion costs, that is, the extra journey time that road users impose on each other. Accident costs, that is, the costs of injuries and fatalities caused to pedestrians and other road users (Jones-Lee, 1990).

In the event, more than one tax and regulatory instrument will have to be used to address the various external costs.14 Taxing instruments include excise duties on motor fuels differentiated by type of fuel (gasoline vs. diesel, leaded vs. unleaded fuel), vehicle license fees differentiated by type of vehicle (cars vs. trucks) and vehicle characteristics (weight, engine capacity), tolls and congestion charges, taxes on the purchase of

286

The Elgar guide to tax systems

new vehicles, and perhaps duties on insurance premiums to account for accident costs. Regulatory measures include compulsory check-ups and catalytic converters. Road user charges can be set to cover the total costs of operating the road network or the difference between the marginal social cost and the average private cost of road use. The duties, charges, and regulatory provisions apply to consumers as well as producers. This does not violate the Diamond-Mirrlees (1971a, 1971b) theorem, which prescribes that, subject to certain general conditions, intermediate goods should not be subject to revenue-raising taxes (if they were, producers would incur excess burdens in trying to pass the tax on to consumers in addition to the tax itself). For example, the fuel excise imposed on production inputs is a proxy for the cost of government-provided road services and internalizes the pollution costs borne by other people. A charge, even if indirect, is therefore appropriate. Box 7.2 illustrates the workings of road user charges. Progressivity-enhancing aspects In addition to the four main rationales of excise taxation outlined above, excises (including higher-than-standard VAT rates) on (luxury) goods and services, whose income-elasticity of demand exceeds unity, have been justified as instruments to improve the progressivity of the tax system. For the promotion of progressivity to be appreciable, consumption by higher-income classes should be significant. In addition, it should be possible to break income-elastic products down into sub-groups, permitting the application of graduated rates that differ on the basis of the price of taxable products, on the assumption that consumption patterns vary accordingly between rich and poor. On these grounds, passenger cars have been singled out for higher-than-average taxes in developing countries. It is widely agreed, however, that the case for the use of excise duties to enhance progressivity in taxation is weak if a government’s administrative capacity is strong. In the event, other instruments, such as the income tax and the benefit system would be better targeted to achieve distributional objectives.15 This does not mean, of course, that no attention should be paid to the distributional equity aspects of traditional and other excises. Indeed, excises on Ramsey-type products that are price inelastic are usually also income inelastic, that is, disproportionately consumed by the poor.16 In other words, the excises are regressive, that is, as a proportion of consumption or income, they bear more heavily on the poor than on the rich. However, note that the Gruber and Köszegi (2001) argument changes the perspective on the burden distribution of the traditional excise duties.

The economics of excise taxation 287

BOX 7.2

CHARGING FOR GOVERNMENTPROVIDED ROAD SERVICES

Figure 7.2 illustrates the workings of a road user charge to improve efficiency in resource allocation. If government provides road services free of charge, demand equals Qfree and the price (PF) only reflects the private marginal cost (e) of the use of the vehicle (fuel, depreciation). Moving from free provision to market pricing (Pp) would see revenue of A+B flowing to government, which would cover the cost of provision, and a social gain (TC1) in terms of improved resource allocation. Social marginal cost (1) now includes the user charge. If there were no external cost associated with the consumption of road services, but the government chose to charge Ps, area E would be considered a tax, while A would be a user charge. Government fees therefore have both tax and user charge components. However, where there is an external cost (E) associated Price

Marginal private benefit Social gains

Social marginal cost (2)

PS C2 E

Social marginal cost (1)

PP

C1

PF

A

B

QS

Source:

Private marginal cost (1)

QP

Qfree

Quantity

Commonwealth of Australia (2008).

Figure 7.2

Charging for government-provided road services

288

The Elgar guide to tax systems

with the use of the good, the market price (Pp) would allow too much consumption. In this case consumers do not take sufficient account of the cost they impose on others. The government could therefore charge a higher and more efficient price (Ps), at which there is a further social gain (C2). Tobacco taxes, for instance, have often been vilified for falling much more heavily on the poor than on the rich. But in Gruber and Köszegi’s view the goal of tax incidence is not only to measure who pays more of the tax, but rather who is ‘hurt’ most by it. Those who are hurt most by the tax are also those who are most price sensitive to tobacco, meaning that their selfcontrol value of taxation is larger. By this reasoning they find that tobacco taxes may actually be progressive. And much the same reasoning applies to alcohol taxes.

3

INSTRUMENTS OF EXCISE TAXATION

Excise duties differ from VATs, among others, because they are often imposed at specific rates, while VATs are levied at ad valorem rates. These rates differ in effect, depending on the market situation, revenue requirements, and non-revenue objectives. The latter objectives can also be achieved through regulations, which can often be targeted more specifically than excise duties. Indeed, the search should be for an optimal combination of taxation and regulation. Specific vs. Ad Valorem Duties17 In a perfectly competitive market for a homogeneous good, the choice between specific and ad valorem taxation is irrelevant: any specific tax could be replaced by its percentage equivalent with no effect on consumer and producer prices or on government revenue. However, in an imperfectly competitive market – a much more common phenomenon – quality levels between similar excisable products, such as cigarettes, differ widely: someone who smokes knows that there are large differences in quality between a Virginia and two sticks of sawdust. In such a market, a common specific tax rate reduces relative price differences between low-quality and high-quality brands, while a common ad valorem rate does not. Standard optimal tax considerations would therefore seem to argue for ad valorem taxation – relative prices would be unchanged, and consumers

The economics of excise taxation 289 would continue to choose a brand on the basis of cost rather than tax differences.18 These arguments apply to competitive markets in which the set of quality levels on offer is given exogenously. With imperfect competition, however, firms’ incentives to raise price and to distort quality may be quite different under specific and ad valorem taxation. In the case of a monopolist, for example, specific taxation increases marginal costs by a fixed amount, whereas ad valorem taxation acts as a proportional tax on costs, together with a proportional (lump-sum) tax on monopoly profits. By taxing marginal revenue, ad valorem taxation, ta, increases the firm’s perceived demand elasticity by the multiplier 1/(1–ta) and so diminishes incentives for the firm to raise price above marginal cost. Thus, one might expect consumer prices to be lower under ad valorem than under specific taxation. Indeed, it is possible to show, in the monopoly case, that replacing a specific tax, ts, by its ad valorem equivalent, ta = ts/p, causes consumer prices to fall and tax revenue and monopoly profits to rise (Skeath and Trandel, 1994). So everyone gains from ad valorem taxation – except the public health advocate.19 Just as ad valorem taxation seems to induce firms to cut prices, it also creates a clear incentive to downgrade product quality (Barzel, 1976), because the multiplier effect of ad valorem taxation makes improvements in product quality more expensive for the firm. The cost of carbon filters, for example, which purify the tobacco of tar and other harmful substances, is subject to the multiplier effect. Likewise, ad valorem taxation reduces incentives to invest in advertising, promotion, and other demandenhancing fixed costs of production. In contrast, specific taxation does not directly distort manufacturers’ decisions to invest in product quality.20 In short, the choice between specific and ad valorem taxation depends on whether the primary aim of the policy is to discourage consumption or to raise revenue and on whether improvements in product quality are deemed desirable or not. Furthermore, if the goal of policy is to reduce consumption, there is some tension between the tendency of specific taxes to lead to higher consumer prices and the tendency of ad valorem taxes to discourage investments in quality that keep consumers ‘hooked’. On the other hand, if the goal is to reduce consumption damage, ad valorem rates have the drawback that they discourage, say, expensive filters on cigarettes. On balance, the solution is likely to be ad valorem taxation at a higher equivalent rate to achieve the desired level of consumer prices, and with concomitant gains for government treasuries. The Pigouvian perspective leads to a very different conclusion, however – the damage caused by smoking is, at any point in time, independent of the price at which cigarettes are sold, so that correction of externalities favors specific over ad valorem taxation.21 Furthermore, other, more immediate,

290

The Elgar guide to tax systems

considerations might govern the choice of tax structure. Thus, a specific tax can be imposed at the manufacturer’s or importer’s stage where it is easiest to collect, whereas, under a system of free trade prices, an ad valorem levy must be collected at the retail stage if trade distortions and tax avoidance are to be avoided. In the EU, of course, most member states circumvent this issue by determining the ad valorem excise by reference to agreed retail prices, making the excise a specific tax as long as cigarette producers do not negotiate new retail prices with the excise tax authorities.22 Where it is clear that the excise duty should be specific, further choices have to be made about the precise form of the duty. Thus, specific tobacco excises can be designed by reference to the weight of tobacco, the number of cigarettes, or their nicotine or tar content, while the specific alcohol excise can be based on volume, alcohol content, or some combination of these attributes – varying depending on the type of beverage. Taxes and Regulations Excise duties, specific or ad valorem, are not the only and often not the best instrument to influence the behavior of smokers, drinkers, gamblers, polluters, and drivers. Depending on circumstances, regulations are an appropriate alternative. High taxes on tobacco and drink reduce average and usually also excessive consumption. But a tobacco tax cannot deal in a cost-effective way with the effects of passive smoking; (inflexible) bans on smoking in public places are necessary to deal with this externality. Similarly, the alcohol excise is an inadequate instrument to restrain people from getting behind the wheel of their car after they have had a drink. Drink-driving breath tests are better targeted to deal with this situation. Yet another example is the regulation of the age at which people are allowed to gamble, or the circumstances under which the discharge of pollutants is permitted. Whereas specific and ad valorem duties would be equivalent in a perfect market, tax and regulatory instruments would be equivalent under conditions of full information, costless implementation, and certainty. But, as Christiansen and Smith (2009) point out, these conditions are not likely to be found in the real world. Poor targeting, imperfect differentiation (e.g., because the externalities are non-linear in consumption) and unintended side-effects of taxes constrain their efficiency. Asymmetric information and costly monitoring and enforcement are central factors behind this situation. They imply that there is a role for regulations alongside taxes in alleviating externalities. In fact, there is a range of situations in which the most optimal approach is a combination of taxes and regulations.23 Addressing the variations on the type of tax-regulatory combination that should be imposed, Christiansen and Smith (2009) draw the

The economics of excise taxation 291 conclusion that where the externality-correcting tax cannot be adequately differentiated, the outcome can be improved by direct regulation of consumption generating the larger marginal cost. The optimal externality tax rate in this context takes the same form as the well-known weighted average formula derived by Diamond (1973). How it is affected by the addition of regulation will depend on how marginal external costs and price responsiveness of demand vary with consumption. Where the marginal external effect is increasing (non-decreasing) in consumption and regulations make demand no more (no less) price responsive, the effect of stricter regulation is to lower the tax rate. In sum, imperfect and inadequate tax and regulatory instruments can and should be used in combination to achieve a superior result in discouraging behavior that gives rise to externalities. Accordingly, a review of excise taxation would be incomplete without an analysis of the efficacy and interaction with accompanying regulatory measures. Pollution: An Illustration The choice between duties and regulations (including permits) can be usefully illustrated with respect to environmental problems, where these instruments are considered alternative, often supplemental, although not necessarily equivalent, ways of achieving a given standard of environmental protection, or, alternatively, achieving a greater environmental impact for a given economic cost. The case for the use of excise duties over conventional regulatory policies based on technology or emission standards is now well established. If firms are faced with different marginal costs of abatement, excise duties can achieve a given level of abatement at lower total abatement costs. And even if the available instruments can take account of differences in abatement costs, excise duties can sidestep the need for the regulatory authority to acquire detailed information on the abatement costs of individual sources. In addition, excise duties provide a continuing incentive for polluters to seek ways to reduce emissions, are more robust to negotiated erosion (‘regulatory capture’), and insulate polluters from the risk that regulatory requirements might involve excessive abatement costs. A drawback to environmental excise duties is that they cannot guarantee that a particular environmental impact will be achieved; polluters’ behavioral responses may be less, or more, than expected. By contrast, quantitative instruments guarantee a particular impact on pollution, but at uncertain abatement costs. In making the choice, the risk of environmental quality must be balanced against the risk of the costs of environmental policy. Environmental excise duties are likely to be particularly

292

The Elgar guide to tax systems

valuable where wide-ranging changes in behavior are needed across a large number of production and consumption activities. Source-by-source regulation, on the other hand, may achieve a more efficient outcome if pollution damage varies depending on the source of the emissions. Another point worth making is that minor environmental duties may be ignored by business; hence, if enforcement is not problematical, environmental duties should be high or not imposed at all. This leaves the choice between pollution duties and tradable pollution permits. In theory, duties and permits are very similar. After all, in an efficient, competitive auction market the market-determined price for each permit would be expected to equal the rate of environmental duty per unit of emissions that would otherwise achieve the same emissions reduction. However, excise duties and permits differ regarding the impact of uncertainty. A system of tradable permits guarantees the envisaged quantitative reduction in pollution but at an uncertain cost, while an environmental duty has an uncertain impact on the quantity of emissions but fixes the marginal cost of emission controls for polluters. A drawback of pollution permits is that they tend to deter new firms from entering a market dominated by large firms that are able to buy up pollution licenses in excess of the firms’ cost-minimizing requirements. An advantage is that pollution permits can be freely distributed to existing firms (grandfathered) and thus do not significantly increase the average financial burden on existing polluters. Grandfathering, however, forgoes the chance to raise revenue that can be recycled through cuts in the marginal rates of other more distortionary taxes.

4

DISCRIMINATION, COORDINATION, AND EARMARKING

Unlike VATs, which are broad-based, excise duties can easily be designed to discriminate against imported products and between products within the same commodity group, for example, wine vs. beer or spirits. Some coordination seems essential, particularly in common markets that purport to maintain equal competitive conditions. Furthermore, the acceptability of excise duties has often been promoted by earmarking them for designated expenditure programs. Discrimination In international trade, excise duties follow the destination principle, that is, exports are freed of tax and imports are taxed on par with domestically

The economics of excise taxation 293 produced commodities. Exceptionally, taxes on environmentally harmful emissions by production units should be levied on an origin basis, that is, in the country where the emissions take place, although this may harm the competitive position of domestic energy-intensive industries if their foreign counterparts do not face the same levy. Destination-based excise duties can discriminate by origin, while originbased taxes can harm a country’s competitive position. Discrimination by origin occurs when a country levies an excise on a product, which is mainly imported, higher than the excise on a similar domestically produced product. An example is the high excise on grain-based spirits that France used to levy while grape-based spirits were subject to a lower duty (Cnossen, 2007). More subtly, expensive imported tobaccos can be discriminated against by levying high ad valorem instead of specific excise duties on cigarettes. The ad valorem duty would favor cheap home-grown tobaccos (Cnossen and Smart, 2005). Similar forms of discrimination are found within countries between various forms of tobacco, alcoholic beverages, and petroleum products. Roll-your-own tobacco and cigars tend to be favored in the EU over cigarettes by a margin of one to three. As a result, low-taxed cigars are on sale that closely resemble high-taxed cigarettes. As another example, the excise on wine has been harmonized at 0 percent and half of all EU member states do not levy a wine excise. In terms of alcohol content – arguably the most appropriate index to account for externalities – the zero rate on wine discriminates against beer and spirits. By the same token, beer is very lowly taxed in beer-producing countries, such as Germany, Austria, and the Czech Republic. On the other hand, spirits are relatively lowly taxed in the UK, the world’s leading exporter of spirits. Coordination Excises, import duties, and VATs should be properly coordinated. The non-revenue function of import duties is to protect domestic industry, while excises are imposed to account for external costs. By contrast, the VAT’s only function is to raise revenue. Logically, in terms of coordination, the protective import duty should be imposed first on the cost, insurance, and freight (CIF) value of imports. This puts imports competitively on par with similar domestically produced goods if these need to be protected. Subsequently, externality-correcting excises should be levied on the import duty inclusive value of goods. Again, this ensures equal treatment vis-à-vis domestically produced excisable goods. Finally, the VAT should be imposed on the import and excise duty inclusive value of goods to put these goods on par with other goods not subject to import and excise duties.

294

The Elgar guide to tax systems

This prescription assumes, of course, that the protective import duty and the corrective excises reflect correctly determined external costs and as such should be subject to the VAT. If the amount of the excise duty exceeds external costs, there would be additional VAT on the excess of these excises over the corrective component, and that part of the total VAT should be considered to be part of the residual tax system rather than the VAT as such. While the sequential application of import duties, excise taxes, and VATs is well established, the level of excise taxation still differs widely between countries, even in the EU’s internal market. In fact, more than 50 years after the Treaty of Rome was signed, excise harmonization is largely limited to common definitions of excisable products and agreement on minimum rates (generally the lowest common rate). Not only are the wide differences in duty rates between the member states at variance with the Treaty’s harmonization goal, they also violate exchange efficiency (under which all consumers face the same product prices) and increase incentives for bootlegging (the purchase of taxed products in a low-excise duty state for consumption in the higher-duty home state) and tax-base snatching (setting low-excise duty rates to attract consumers from other high-duty states). Indeed, cross-border shopping incentives may harm incentives to cut costs and to compete. Earmarking It is often asserted that excise revenues should be earmarked to finance health expenditure projects that stimulate the production and consumption of clean energy or repair the degradation of the environment, or to pay for the building and maintenance of the road transport system. Bird and Jun (2007) distinguish eight types of earmarking by the degree of specificity of the expenditures involved, the strength and nature of the linkage between the earmarked revenues and the expenditure, and whether or not there is an identifiable benefit rationale for the linkage. Tobacco or alcohol excise revenues might be used for instance to finance health expenditure for the treatment of the ailments that they cause or campaigns against smoking or abusive drinking, although the linkage would be quite loose. Similarly, fuel excises might be earmarked for road maintenance purposes, or environmental taxes to finance clean-up programs. They call these forms of earmarking symbolic earmarking, because the various revenue amounts finance only some of the expenditures, although there may be a loose benefit rationale. In economic terms, the earmarking is ‘irrelevant’, because the marginal expenditure decision remains firmly in the hands of the budgetary authorities.

The economics of excise taxation 295 Earmarking is particularly suspect in the case of tobacco and alcohol excise revenues. It would be difficult to isolate health expenditures on smoking-related diseases and finance them by tobacco duties. There is evidence, moreover, that tobacco tax revenues exceed the external cost associated with smoking, primarily because smokers tend to die earlier than non-smokers and hence may not attract age-related diseases that require expensive treatment (Cnossen and Smart, 2005). With alcohol the case for earmarking is just as dubious, because moderate drinkers would be asked to pay for the health and other social costs attributable to abusive drinkers. The case is stronger for earmarking the proceeds from taxes on road transport for infrastructure purposes, as argued by Gwilliam and Shalizi (1999). Although overall fiscal control and allocational efficiency may suffer, operational efficiency might improve because with a stable source of revenue road infrastructure management can make better use of more efficient private sector contracting arrangements for road maintenance. These authors also argue that a user-managed fund, financed from taxes that are reasonable proxies for benefits received, can properly reflect the interests of road users in better-quality services, while reducing the interests of non-users who have little interest in the service. Although the case for road funds is somewhat stronger therefore, a lock-in effect remains because past arrangements weigh heavily on current realities, which may indicate that it would be better to shift resources to alternative modes of transportation, such as public transport. In conclusion, the case of earmarking, even if the benefit rationale is quite strong, remains tenuous at best.

NOTES 1.

2. 3.

4.

In the terminology of the Organisation for Economic Co-operation and Development (OECD, 2009), excise systems therefore comprise all selective taxes on the production, sale, transfer, leasing, and delivery of goods, and the rendering of services (item 5120 in the OECD classification), as well as all selective taxes on the use of goods, or on the permission to use goods or perform activities (item 5200), other than general taxes on goods and services (item 5110). By this definition, import and export duties also can be considered excise-type levies, but they are not dealt with in this chapter. For a wide-ranging review of the actual taxation of smoking, drinking, gambling, polluting, and driving, see the contributions in Cnossen (2005) and (2008). Suppose that the utility function of each individual is a function of his or her consumption of leisure and a set of other commodities. Then as long as the marginal rate of substitution between any two commodities is independent of the amount of leisure, differential commodity taxation cannot improve social welfare in the presence of an optimal income tax. See Atkinson and Stiglitz (1976) and Salanié (2003). It is often asserted that excise receipts have declined in recent decades, but much of

296

5.

6.

7.

8. 9.

10.

11.

12.

13.

14. 15. 16. 17. 18.

The Elgar guide to tax systems this decline should be attributed to the imposition of VAT on excise-inclusive product prices, whereas the turnover and other sales taxes levied previously were not generally imposed on excisable products. Accordingly, upon introduction of the VAT, excise duties were lowered to maintain the same total tax burden. In line with this argument, Bovenberg and De Mooij (1994) have shown that revenue considerations should lead taxes on ‘dirty’ goods away from Pigouvian levels. As the overall level of taxation increases, the marginal excess burden of a Pigouvian tax rises relative to its external benefits. Hence, differential taxation of polluting goods should fall as the overall level of taxation rises. In the event, charging producers would not violate the Diamond-Mirrlees (1971a, 1971b ) theorem, which holds that the pursuit of production efficiency (all firms face the same input and output prices) as a policy objective takes precedence over the pursuit of exchange efficiency (all consumers face the same product prices). Coase (1960) has pointed out that government intervention to deal with externalities would not be required if property rights to, say, air and water were established. But the application of this important finding is limited if resource owners cannot identify the source of damages to their property (and legally prevent the damages) or if the cost of bargaining deters the parties involved from finding their way to an efficient solution. (Even if these conditions were met, the assignment of property rights would still, of course, affect income distribution.) For a survey and evaluation of alcohol taxation in the European Union (EU), see Cnossen (2007). Using repeat cross-sections for the period from 1991 through 2005, Carpenter and Cook (2008) report that the large tobacco tax increases in US States during the past 15 years were associated with significant reductions in smoking participation by youths. For a review of tobacco taxation in the EU, see Smith (2008). See Bird and Wallace (2006) who discuss the differences between the public health approach and the economic approach. As they say, if someone drinks too much and dies sooner than he or she otherwise would have done – for example, by crashing a motor vehicle while drunk – it may be a tragedy, but it is not an externality. If, however, the drunken driver kills a passer-by or a passenger, then it is both. By contrast, Khwaja et al. (2009) find no empirical support for hyperbolic discounting as an explanation for continued smoking. They find that people who smoke have a lower non-pecuniary internal cost of getting a major smoking-related disease than those who do not smoke. This weakens the case for intervention based on helping smokers internalize their internal costs. In support of this view, Gruber and Mullainathan (2005) provide evidence that higher cigarette taxes increase smokers’ self-reported happiness. However, Bernheim and Rangel (2005) provide an alternative model of partially rational addictive behavior, which they call ‘characterization’ failure, in which taxation of addictive substances can never generate efficiency gains. In addition to the ‘optimal paternalism’ examined by O’Donoghue and Rabin (2003) and the ‘libertarian paternalism’, explored by Thaler and Sunstein (2003), earlier, O’Donoghue and Rabin (2001) discussed ‘cautious paternalism’, while Choi et al. (2003) investigate ‘benign paternalism’. For a later treatment, see also O’Donoghue and Rabin (2006). For a full treatment, see Newbery (2006). The case for and against levying excise taxes on luxury goods in developing countries is examined by Cnossen (2006). Strictly speaking, price inelasticity of demand need not imply income inelasticity (unless utility is additive), but it is usually assumed to be the case. For a review of the theoretical arguments, see especially Keen (1998). This section has been adapted from Cnossen and Smart (2005). Of course, the tax would still have income effects that might induce consumers to choose lower-quality brands, but so would a non-distortionary lump-sum tax. The

The economics of excise taxation 297

19.

20.

21. 22. 23.

theory of optimal taxation implies that a uniform percentage tax on a subset of commodities is desirable only under restrictive conditions on preferences (Atkinson and Stiglitz, 1980), but in this context those restrictions seem plausible. In the Cournot model of an oligopoly industry, the story is largely the same: a shift to ad valorem taxation will reduce prices and increase government revenues. In this case, however, industry profits may fall, as competition among firms intensifies. A further, testable, implication of the theory is that the pass-through of tax increases to consumer prices should be greater under specific than under ad valorem taxation (Delipalla and Keen, 1992). Indeed, there is some evidence that specific taxes in the EU are more likely to be ‘over-shifted’ (consumer prices rise by more than the tax) than ad valorem taxes (Delipalla and O’Donnell, 2001). However, specific taxation may induce consumers to opt for higher-quality brands, if the degree of tax shifting is independent of product quality. In support of this view, Sobel and Garrett (1997) find that specific tax increases in US states are associated with significant declines in the market share of generic brands. An economic counter-argument is that the share of specific in total taxation should be smaller when the marginal cost of public funds is higher and the importance of excise duties for generating revenue correspondingly greater. An incidental, if welcome, side-effect of this practice is that it weakens the argument that the value of a specific excise erodes with inflation. After all, inflation would compel producers to approach the excise authorities with a proposal for a new retail price. As the authors point out, regulations, like excises, can have two effects: they may reduce harmful consumption per se and they may induce changes in the circumstances of consumption. Drinking alcohol at home is less harmful than drinking alcohol outside the home. The excise duty would reduce overall consumption regardless of where the drink is consumed (called the consumption response), while the prohibition on drink driving restrains the place of consumption (the abatement response).

REFERENCES Atkinson, A.B. and J.E. Stiglitz (1976), ‘The Design of Tax Structure: Direct versus Indirect Taxation’, Journal of Public Economics, 6(1–2), July–August, 55–75. Atkinson, A.B. and J.E. Stiglitz (1980), Lectures on Public Economics, New York: McGraw-Hill. Barzel, Y. (1976), ‘An Alternative Approach to the Analysis of Taxation’, Journal of Political Economy, 84(6), 1177–97. Becker, G.S. and K.M. Murphy (1988), ‘A Theory of Rational Addiction’, Journal of Political Economy, 96(4), 675–700. Bernheim, B.D. and A. Rangel (2005), ‘From Neuroscience to Public Policy: A New Economic View of Addiction’, Swedish Economic Policy Review, 12, 99–144. Bird, R.M. and J. Jun (2007), ‘Earmarking in Theory and Korean Practice’, in S.L.H. Phua (ed.), Excise Taxation in Asia, Singapore: National University of Singapore. Bird, R.M. and S. Wallace (2006), ‘Taxing Alcohol: Reflections from International Experience’, in S. Cnossen (ed.), Excise Tax Policy and Administration in Southern African Countries, Pretoria: University of South Africa Press. Bovenberg, A.L. and R. De Mooij (1994), ‘Environmental Levies and Distortionary Taxation’, American Economic Review, 84(4), 1085–9. Carpenter, C. and P.J. Cook (2008), ‘Cigarette Taxes and Youth Smoking: New Evidence from National, State and Local Risk Behavior Surveys’, Journal of Health Economics, 27(2), 287–99. Chaloupka, F.J., T-W. Hu, K.E. Warner, R. Jacobs and A. Yurekli (2000), ‘The Taxation of Tobacco Products’, in P. Jha and F.J. Chaloupka (eds), Tobacco Control in Developing Countries, Oxford: Oxford University Press.

298

The Elgar guide to tax systems

Choi, J., D. Laibson, B. Madrian and A. Metrick (2003), ‘Optimal Defaults’, American Economic Review, 93(2), 180–85. Christiansen, V. and S. Smith (2009), ‘Externality-correcting Taxes and Regulation’ (11 February 2009 version), paper presented at the 2008 IIPF conference in Maastricht. Cnossen, S. (ed.) (2005), Theory and Practice of Excise Taxation: Smoking, Drinking, Gambling, Polluting, and Driving, Oxford: Oxford University Press. Cnossen, S. (2006), ‘The Impact of Consumption Taxes’, in S. Cnossen (ed.), Excise Tax Policy and Administration in Southern African Countries, Pretoria: University of South Africa Press. Cnossen, S. (2007), ‘Alcohol Taxation and Regulation in the European Union’, International Tax and Public Finance, 14(6), 699–732. Cnossen, S. (ed.) (2008), ‘Fiscal Policy in Action’, FinanzArchiv, 64(December). Cnossen, S. and M. Smart (2005), ‘Taxation of Tobacco’, in S. Cnossen (ed.), Theory and Practice of Excise Taxation: Smoking, Drinking, Gambling, Polluting and Driving, Oxford: Oxford University Press. Coase, R.H. (1960), ‘The Problem of Social Cost’, Journal of Law and Economics, 3(1), 1–44. Commonwealth of Australia (2008), ‘Australia’s Future Tax System’, available at: http:// www.taxreview.treasury.gov.au/content/downloads/consultation_paper/Consultation_ Paper.pdf; accessed 26 April 2011. Corlett, W.J. and D.C. Hague (1953), ‘Complementarity and the Excess Burden of Taxation’, Review of Economic Studies, 21(1) 21–30. Crawford, I., M. Keen and S. Smith (2010), ‘VAT and Excises’, in The Mirrlees Review, Reforming the Tax System for the 21st Century, vol. I, Oxford: Oxford University Press for the Institute for Fiscal Studies. Crooks, E. (1989), Alcohol Consumption and Taxation, Report No. 34, London: Institute for Fiscal Studies. Delipalla, S. and M. Keen (1992), ‘The Comparison between Ad Valorem and Specific Taxation under Imperfect Competition’, Journal of Public Economics, 49(3), 351–67. Delipalla, S. and O. O’Donnell (2001), ‘Estimating Tax Incidence, Market Power and Market Conduct: The European Cigarette Industry’, International Journal of Industrial Organization, 19(6), 885–908. Diamond, P. (1973), ‘Consumption Externalities and Imperfect Corrective Pricing’, Bell Journal of Economics, 4(2), 526–38. Diamond, P.A. and J.A. Mirrlees (1971a), ‘Optimal Taxation and Public Production, Part I: Production Efficiency’, American Economic Review, 61(1), 8–27. Diamond, P.A. and J.A. Mirrlees (1971b), ‘Optimal Taxation and Public Production, Part II: Tax Rules’, American Economic Review, 61(3), 261–78. Gruber, J. (2010), Commentary on ‘VAT and Excises’, paper by I. Crawford, M. Keen and S. Smith in The Mirrlees Review, Reforming the Tax System for the 21st Century, vol I, Oxford: Oxford University Press for the Institute for Fiscal Studies. Gruber, J. and B. Köszegi (2001), ‘Is Addiction “Rational”? Theory and Evidence’, Quarterly Journal of Economics, 116(4), 1261–1303. Gruber, J. and S. Mullainathan (2005), ‘Do Cigarette Taxes Make Smokers Happier?’, Advances in Economic Analysis and Policy, 5(1). Gwilliam, K. and Z. Shalizi (1999), ‘Road Funds, User Charges and Taxes’, The World Bank Research Observer, 14(2), 159–85. Jones-Lee, M. (1990), ‘The Value of Transport Safety’, Oxford Review of Economic Policy, 6(2), 39–60. Keen, M. (1998), ‘The Balance between Specific and Ad Valorem Taxation’, Fiscal Studies, 19(1), 1–37. Khwaja, A., F. Sloan and Y. Wang (2009), ‘Do Smokers Value Their Health and Longevity Less?’, Journal of Law and Economics, 52(1), 171–96. Newbery, D.M. (2006), ‘Road User and Congestion Charges’, in S. Cnossen (ed.), Excise Tax Policy and Administration in Southern African Countries, Pretoria: University of South Africa Press.

The economics of excise taxation 299 O’Donoghue, T. and M. Rabin (2001), ‘Choice and Procrastination’, Quarterly Journal of Economics, 116(1), 121–60. O’Donoghue, T. and M. Rabin (2003), ‘Studying Optimal Paternalism, Illustrated by a Model of Sin Taxes’, American Economic Review, 93(2), 186–91. O’Donoghue, T. and M. Rabin (2006), ‘Optimal Sin Taxes’, Journal of Public Economics, 90(10–11), 1825–49. OECD (2009), Revenue Statistics 1965–2008, Paris: Organisation for Economic Co-operation and Development. Pigou, A.C. (1920), The Economics of Welfare, New York: Macmillan. Pogue, T.F. and L.G. Sgontz (1989), ‘Taxing to Control Social Costs: The Case of Alcohol’, American Economic Review, 79(1), 235–43. Ramsey, F. (1927), ‘A Contribution to the Theory of Taxation’, Economic Journal, 37(145), 47–61. Salanié, B. (2003), The Economics of Taxation, Cambridge, MA: The MIT Press. Sandmo, A. (1976), ‘Direct versus Indirect Pigouvian Taxation’, European Economic Review, 7(4), 337–49. Skeath, S.E. and G.E. Trandel (1994), ‘A Pareto Comparison of Ad Valorem and Specific Taxation in Noncompetitive Environments’, Journal of Public Economics, 53(1), 53–71. Smith, S. (2006), ‘Taxes on Road Transport’, in S. Cnossen (ed.), Excise Tax Policy and Administration in Southern African Countries, Pretoria: University of South Africa Press. Smith, S. (2008), ‘Restraining the Golden Weed: Taxation and Regulation of Tobacco’, FinanzArchiv, 64(December). Sobel, R.S. and T.A. Garrett (1997), ‘Taxation and Product Quality: New Evidence from Generic Cigarettes’, Journal of Political Economy, 105(4), 880–87. Thaler, R.H. and C.R. Sunstein (2003), ‘Libertarian Paternalism’, American Economic Review, 93(2), 175–9.

8

The scale and scope of environmental taxation* Agnar Sandmo

1

INTRODUCTION

One of the first lessons that new students of economics learn about the principles of public finance is that indirect or commodity taxes are harmful to the efficiency of the economy. The partial equilibrium analysis that forms the basis for this conclusion is a simple and compelling one: a commodity tax drives a wedge between the marginal cost of production (as represented by the supply curve) and the marginal consumer benefit (as represented by the demand curve). The tax therefore prevents the market mechanism from reaching the efficient equilibrium solution where marginal cost is equal to marginal benefit. In a later lesson the student may learn that there are exceptions to this rule. If there are negative externalities associated with the production or consumption of a particular commodity – the typical example being adverse effects on the quality of the environment – efficiency may in fact be improved by taxation. Suppose, for the sake of the argument, that the externality in question generates a positive difference between social and private marginal cost, while private and social marginal benefits coincide. The requirement for efficiency is that the social marginal cost of production should be equal to the social marginal benefit, which again is equal to the private marginal benefit: SMC 5 SMB 5 PMB. Suppose further that this market operates according to the principles of perfect competition except that the market prices with which producers and consumers are faced are allowed to differ. Producers, who are assumed to maximize profits, set their private marginal cost equal to the producer price (p), while consumers, who maximize utility, equate their marginal benefit to the consumer price (P): PMC 5 p, PMB 5 P.

300

The scale and scope of environmental taxation 301 Now, if the tax rate is defined by the equation P = p + t, we can substitute in the first equation to obtain the socially optimal deviation between the consumer and producer prices, SMC 2 PMC 5 P 2 p, or the socially optimal tax rate: t 5 SMC 2 PMC. A tax rate that is equal to the difference between social and private marginal costs perfectly internalizes the externality and restores efficiency to a market that would otherwise have found itself at an inefficient equilibrium. Thus, the case for an environmental or green or Pigouvian tax basically rests on its incentive effects; it ensures that decisions about production and consumption that are rational from a private point of view are consistent with efficiency for society as a whole. In addition, since t is the same for all polluters, the taxation scheme ensures that the marginal cost of pollution cutbacks is the same for all polluters, implying that the aggregate social cost of pollution reduction is minimized. A further benefit of environmental taxes is that they generate revenue for the government, and they do so without creating inefficiencies in an otherwise efficient market system. Substituting environmental taxes for ordinary commodity and income taxes should accordingly be able to raise the same amount of revenue as before but with less social cost. This ‘double dividend’ of environmental taxes could be expected to make them extremely popular taxes; in fact, one would expect them to be the first building block of an efficient tax system. But this is rather far from being the case in real life. There could be several reasons for this, and some of them will be explored in a little more detail in the following.

2

HISTORY OF ENVIRONMENTAL TAXES

The history of environmental taxation can be regarded from two different angles. One is the perspective of the history of economic thought, while the other is the history of the taxes and tax systems that have actually been used. The first to set out the basic idea of environmental taxation was A.C. Pigou who first introduced it in his 1920 book The Economics of Welfare with further discussion in his later A Study in Public Finance (1928).1 He did not use the concept of environmental taxation, although he discussed

302

The Elgar guide to tax systems

a number of examples that we would now classify as environmental problems. Instead, he set his discussion in the more general analytical framework of deviations between private and social net product, as he called it; this corresponds roughly to the concepts of private and social marginal costs that were used above.2 His analysis covered not only negative but also positive externalities that called for subsidies (‘bounties’) to the commodities or activities that generated them. He sketched a situation where he appeared to have in mind a tax system whose only task was to introduce corrections to situations where private incentives needed to be supplemented by interventions to overcome these deviations: it is always possible, on the assumption that no administrative costs are involved, to correct them [the deviations] by imposing appropriate rates of tax on resources employed in uses that tend to be pushed too far and employing the proceeds to provide bounties, at appropriate rates, on uses of the opposite class. There will necessarily exist a certain determinate scheme of taxes and bounties, which, in given conditions, distributional considerations being ignored, would lead to the optimum result. (Pigou [1928] 1947, p. 99; original emphasis)

Today, public finance and environmental economists recognize this idea as a fundamental and important one. But in the first few decades following Pigou’s analysis, it received relatively little attention in the academic literature. A prominent example of this neglect of the idea is Musgrave’s famous treatise on public finance (Musgrave, 1959), which devotes little more than a paragraph to Pigouvian taxation and this in a chapter entitled ‘The ability-to-pay approach’. The environmental perspective is not mentioned in this paragraph and indeed is absent from the book as a whole. The interest in Pigouvian taxation as a tool of environmental policy started to take off around 1970. Since then the idea has been explored in depth by a number of researchers. Among the most significant extensions of Pigou’s analysis have been the generalization of the theory from a partial to a general equilibrium framework, the attention to issues of second best where environmental taxes must be introduced in a distorted market system, and the broadening of the perspective from local and national environmental issues to those that arise in the global environment. To what extent has environmental taxation actually been used? There is no doubt that these taxes are on the policy agenda to a larger extent than ever before, but summary measures of their importance are hard to obtain. How is importance to be measured? One suggestion has been to use the share of green taxes in government revenue or GDP. Data from the OECD for the period 1994–2006 show that the share of ‘environmentally related taxes’3 in GDP varied from 4.5–5 per cent (Denmark) to less than 1 per cent (United States); the weighted average for all OECD countries was

The scale and scope of environmental taxation 303 a little less than 2 per cent, while the unweighted average for all 30 countries included in the survey was about 2.5 per cent. In most of the countries the share was fairly stable over this period, although it dropped slightly in 2006. In terms of the share of environmental taxes in total tax revenue, the weighted average was a little less than 6 per cent, while the unweighted average was about 7 per cent. It is worth noting that these figures exclude fees and charges (e.g., for road use and parking).4 These figures should naturally be interpreted with some care. There is an inevitable element of judgement involved in selecting the taxes that are environmentally related. Also, the period considered is too short to allow us to draw conclusions about long-term trends. Nevertheless, we may conclude that we are concerned with taxes that are of substantial importance for government revenue and the economy as a whole, although they have fallen considerably short of satisfying the revenue needs of modern governments.

3

FINANCIAL VS. ENVIRONMENTAL IMPACT

To calculate the financial importance of environmental taxes is clearly an interesting exercise. However, if we are concerned with the effects of taxes on the quality of the environment this cannot be the main focus of our interest. Instead, this should be the effects that these taxes might have on the activities or commodities whose consumption and production we wish to reduce. The crucial concepts to use in this connection are the elasticities of demand and supply. I focus here on demand elasticities, taking the view that the composition of output is essentially driven by consumer demand. It is easy to see that in terms of environmental impact, the share of a particular tax in government revenue or GDP is a poor measure of its environmental effect. In the case of a commodity with a very high elasticity of demand, a substantial effect on the environment may be achieved at a very low rate of tax, while conversely when the elasticity of demand is low the effect on the environment may be very small although the tax could generate substantial revenue. In that case, Pigouvian tax enthusiasts will have to face the common criticism that this is just another way for the government to satisfy its demand for increased tax revenue. In general, economists ought perhaps to be more explicit about the time frame that they have in mind when analysing green tax effects on behaviour. The introduction of taxes on cars and car use may be a case in point.5 In the short run, the effects of taxes on cars and petrol may be small, given the existing stock of cars and the supply of collective transport alternatives. But in a longer perspective this may change as a result

304

The Elgar guide to tax systems

of the long-run tax incentives. Households may decide not to replace their second car or modify their habits as regards travelling to work. In the even longer run, increased taxes on private transport may reverse the trend in city development that has been so characteristic of the post-war period. Instead of the continuation of urban sprawl, we may come to see a movement towards more compact cities and greater reliance on less energy-intensive and collective means of transportation. The Marshallian insight that elasticities are higher in the long run than the short is of crucial importance for environmental policy. The magnitude of the price elasticities are also likely to depend on other elements of public policy. According to Leape (2006), the success of the London congestion charge in reducing traffic in the central areas of the city was to a large extent due to the presence of a substitute for car use in the form of a well-developed system of public transport. The availability of this substitute was probably also the main reason why the congestion charge could be introduced with only minor effects on local business. Thus, the design of environmental tax policy should be seen in conjunction with that of publicly provided goods and services.

4

THE IDENTIFICATION OF ENVIRONMENTAL TAXES

How do we decide which taxes are green taxes? It is natural, as in the OECD procedure, to define green taxes as those that are levied on tax bases that are correlated with adverse environmental effects. But this correlation may take many forms, and the classification of taxes may to a large extent depend on the judgement of those who carry it out. As an example of the difficulties that arise we may once more take taxes related to car use. In many countries, including my own, there are three classes of such taxes or charges. The first is related to car ownership: in Norway you pay an annual tax on cars and lorries (in addition to what you may pay on their asset value through the wealth tax) that is related to size, although there is no differentiation between different types of passenger cars.6 The second type is the tax on petrol, which is clearly directly related to car use. The third class consists of charges related to road use and includes tolls and parking fees. All three types have their obvious weaknesses as instruments of environmental policy. In the short run, the annual car tax does not discourage car use since it does not vary with the use of the car. However, in the long run it may affect the number of cars owned by families and thereby the intensity of car use. The petrol tax payments do vary with car use and the

The scale and scope of environmental taxation 305 petrol tax can therefore with greater justification be called an environmental tax. But the petrol tax still suffers from some weaknesses in terms of environmental policy; it does not discriminate between various uses of the car according to location and time of day, which may have very different environmental effects. Tolls and road user charges clearly have the potential to function as Pigouvian taxes in that the use of congested roads could be taxed at higher rates than other uses.7 Parking fees can also be designed so as to vary with time and location and thereby improve environmental quality in congested cities. The tax rate should reflect the difference between social and private marginal cost, but how do we know whether a particular tax rate satisfies this requirement or whether it is too high or too low? This is a difficult question to answer and will have to be decided on empirical grounds in each particular case, but it also raises complex and interesting issues of a theoretical nature.8

5

THE ACTIVITIES PERSPECTIVE AND THE TAX BASE

Environmental pollution may be generated both on the production and consumption sides of the economy. To simplify the theoretical treatment, the focus in the following will be on environmental externalities generated by consumers – in their capacities as drivers, energy users and so on. It is by focusing on consumers that we most easily see some of the most crucial problems of environmental policy, such as the measurement of costs and benefits, issues of optimal taxation and distributional considerations. In much of the theoretical literature, the assumption is made that environmental effects depend on the aggregate consumption of particular commodities, and that there is a one-to-one correspondence between externalities and goods. Although this assumption is often sufficient to analyse the central theoretical issues it also gives a highly simplified picture of the difficulties of policy design. As was illustrated by the example of the taxation of cars, environmental effects are basically generated by activities that may depend on the consumption of goods in rather complicated ways. An activity like driving during the rush hour uses the car, petrol and time as important inputs, while the activity of central heating may depend on inputs like house insulation and electricity. Obviously, consumer goods like electricity, petrol and time are used in additional activities that may have no or very much smaller environmental effects. The general perspective is that the total quantity that the consumer buys of any particular commodity is allocated between a number of different activities, some of

306

The Elgar guide to tax systems

which have negative effects on the quality of the environment and some have not. Suppose now that one particular activity generates environmental externalities and that we wish to internalize these effects by means of Pigouvian taxation. The first question that arises is what the tax base should be. Obviously, if it is the activity that generates the pollution, the tax should ideally be levied at the amount of the activity undertaken by consumers. This is an application of the general principle of policy targeting. But in practice this advice may be difficult to follow: the level at which the activity is carried out may be difficult to measure, because the activity itself is not a subject of market transactions. A more feasible alternative is to tax goods, but the taxation of goods encounters difficulties when it comes to the discrimination between the uses of the goods in different activities. We may ideally wish to tax the use of a particular commodity in a particular activity, but administrative feasibility may limit us to the taxation of the total quantity bought by the consumer, whereby we tax the use of the commodity in all activities, whether they have adverse environmental effects or not. The car example affords a good illustration of the problems involved. For environmental reasons we might wish to tax the activity of driving in particular areas and at particular times. But if this is not feasible, the petrol tax might seem like a more practical alternative. However, for administrative reasons petrol has to be taxed at the same rate, whatever the activity is for which it is being used. The Pigouvian taxation of commodity i in activity j therefore has to be linked with the distortionary taxation of commodity i as it is used in environmentally neutral activities. This formulation provides a stylized explanation of why the ideal base for Pigouvian taxation may be difficult to achieve. Since there are high informational and administrative costs involved in taxing polluting activities, one has in practice to tax goods rather than activities. But because of the additional difficulty of linking environmental damage to the use of one particular commodity in the context of the harmful activity, the ideal of perfect targeting may be hard to achieve. With broad-based commodity taxes – in the sense of taxes that are levied at the same rate on all uses of the commodity – environmental taxes are likely to introduce distortions jointly with the corrections of imperfections in the market mechanism. This leads to the conclusion that several commodity taxes – or even several commodity taxes supplemented by quantitative regulations – may have to be used in order to achieve an optimal result.9 It might perhaps be tempting to draw pessimistic policy conclusions from this analysis: if corrective Pigouvian taxes are likely to introduce new distortions, is it not to be expected that the net welfare gain from their use might easily become negative? In my view, this pessimism is unfounded.

The scale and scope of environmental taxation 307 It is important to distinguish between achieving on the one hand a social optimum and on the other hand a welfare improvement. Imperfect taxes, if used with good judgement and empirical knowledge, can clearly result in substantial welfare improvement in spite of falling short of the first-best welfare ideal.

6

PRINCIPLES OF OPTIMAL TAXATION

Needless to say, taxes are not only imposed for the purpose of improving the environment; their primary purpose is to raise revenue for publicly provided goods. Now since environmental taxes are not the only taxes used by the government to raise revenue, the question arises of how Pigouvian taxes should be determined given the existence of these other taxes. The simplest case is that where the government’s other source of revenue is lump sum taxes: taxes that are levied on individuals as a fixed sum with a marginal tax rate of zero. In that case the Pigouvian taxes on commodities that generate adverse environmental effects should simply be equal to the difference between social and private marginal costs – the marginal social damage – all other commodity taxes being set equal to zero. This is often referred to as the first-best case of environmental taxation. Unfortunately, for all practical purposes, ideal lump sum taxes do not exist outside of economic textbooks. To move the theory closer to the concerns of the problems faced by policy-makers we have to consider environmental taxes in the context of a more realistic environment in which the government uses both ordinary – that is, non-environmental – commodity taxes as well as direct taxes on income and wealth. These taxes introduce distortions in the economy by violating the conditions for an optimal allocation of resources or Pareto optimality. In the standard theory of optimal taxation, which goes back to the work of the English economist and mathematician Frank Ramsey (1927), the problem of externalities is neglected, and the focus is on the issue of how to raise a given amount of revenue from commodity taxes with a minimum loss of efficiency for the economy as a whole. Ramsey showed that under some simplifying assumptions tax rates should be inversely proportional to the elasticities of demand: commodities with inelastic demand should be taxed at high rates, while those whose demand is more elastic should be taxed at lower rates. This is in contrast to the idea, which used to be prevalent in the public finance literature, that the least distortionary system of commodity taxation would be one where taxes were uniform, that is, levied at the same percentage rate on all goods. The Ramsey elasticity rule indicates that what we should worry about is not the distortions of relative

308

The Elgar guide to tax systems

prices but of relative quantities, and this is precisely what is achieved by the inverse elasticity rule.10 Similarly, the marginal rate of income tax, to  the extent that it applies mainly to income from labour, should be higher, the lower the elasticity of labour supply. How are environmental taxes to be fitted into this framework of optimal taxation? For simplicity of exposition, let us assume that there is one commodity, the consumption or production of which creates a negative environmental externality (in the jargon of environmental economists this is often referred to as a ‘dirty good’). The remaining commodities have no such effects (they are ‘clean goods’). Then it can be shown (Sandmo, 1975, 2000) that for all clean goods the optimal tax rates satisfy the Ramsey rule: the rates should be inversely proportional by a factor a to the elasticities of demand. In the case of the dirty good, the tax rate should be a weighted average of that computed under the Ramsey inverse elasticity rule and the Pigouvian marginal social damage term, the weights being a and (1–a). The magnitude of a should reflect the tightness of the government’s budget constraint or the marginal cost of public funds. If this is very high, the parameter a becomes a number close to one, and the tax structure becomes similar to that which maximizes the revenue to the government. In this case the whole tax structure becomes very similar to that of Ramsey. If, on the other hand, a is very small the structure of the optimal tax system moves closer to that of the first-best Pigouvian structure: taxes on clean goods are approximately zero, while the taxes on dirty goods reflect their respective marginal social damage, just as in the first-best case. One simplifying assumption that underlies this analysis should be emphasized: the version of the Ramsey rule that has been used here makes optimal tax rates proportional to the own elasticities of demand, while cross-elasticities are disregarded. However, the Ramsey rule can be extended to the case of cross-elasticities that differ from zero, and it can be shown that a similar rule, as well as the weighted average formula for dirty goods, holds for the more general case also.11 A short semi-mathematical digression may illuminate this. Let us denote by uR the ad valorem tax rate that maximizes government revenue (i.e., the inverse elasticity) and by uD the rate that reflects the marginal social damage. Then, if u is the overall tax rate, the optimal tax structure can be shown to satisfy the conditions u = auR for all clean goods, and u = auR + (1–a)uD for all dirty goods. A remarkable feature of this solution is that the tax rates on none of the clean goods should contain a component reflecting the marginal social damage, whether the clean goods are substitutes or complements in regard to the dirty good. This is an example of the principle of targeting: the policy instrument should be targeted as precisely as possible on the goal that it aims to influence.

The scale and scope of environmental taxation 309 There are a number of special cases and extensions of this line of reasoning. Suppose that one of the clean goods is interpreted as leisure; the optimal tax rate could in this case be interpreted as the marginal rate of income tax under a linear tax system. Assume now that the demand for leisure – or, equivalently, the supply of labour – is completely inelastic. In this case an arbitrary large amount of revenue can be raised by the income tax without adverse distortionary effects. The income tax therefore becomes the equivalent of a lump sum tax, and the optimal solution is to raise the required revenue through the income tax while letting commodity taxes serve the sole purpose of correcting the market failure that follows from the environmental externalities. Commodity taxes should be zero on clean commodities and positive on dirty commodities. A difficulty about this whole line of reasoning is that it disregards distributional effects. To see the shortcomings of this restriction, assume that the benefits of environmental improvement accrue mainly to the rich while the ‘dirty goods’ are chiefly consumed by the poor. In this example the effects of a green tax reform would be regressive, redistributing real income from the poor to the rich. With even a mildly egalitarian social welfare function such an outcome would be unattractive and would need to be modified in the direction of more social justice. This could be done along several lines. One solution is to modify the tax formulas above: in the calculation of the aggregate marginal social damage, the damage suffered by the rich could receive a lower weight than the damage experienced by the poor. Moreover, in calculating the Ramsey efficiency terms of the marginal tax formulas, the inverse elasticity terms could be weighted according to the distributive profile of the consumption patterns of the various goods. By so doing, there would be a bias towards taxing necessities (which tend to have low price elasticities of demand) at lower rates than luxury goods (whose elasticities are higher). The overall tax structure would therefore emerge as a compromise between the regard for social efficiency and distributive justice. Another response to the possible conflict between the concerns for efficiency and distributive justice is to develop policy instruments that are better targeted on the redistributive goal. We have already noted the administrative impossibility of a system of individualized lump sum taxes, but there are clearly interesting cases in between this extreme case and the other extreme where all taxes are proportional. One possibility that has been extensively discussed in the literature is that of a general non-linear income tax, an idea that receives its inspiration from the work of Mirrlees (1971). In the theoretical formulation of Mirrlees the marginal rate of income tax is allowed to vary continuously with income, and it has been shown that the optimal marginal tax rate at the top of the income scale

310

The Elgar guide to tax systems

should in fact be zero. In an aggregate perspective, this is just the outcome that we would have under lump sum taxation, and under some particular assumptions it has been demonstrated that with a general non-linear income tax the redistributive concern should in its entirety be taken care of by the income tax, while Pigouvian environmental taxes should be calculated according to the first-best rule; see Kaplow (2008). The discussion above has proceeded on the assumption that environmental damage is generated by the consumption and production of particular commodities that are feasible objects of taxation. This is not always a realistic assumption. As pointed out in Section 5 above, environmental externalities are frequently generated by activities, not goods. But the taxation of activities may be difficult since activities as such are not traded in the market and therefore are awkward bases for taxation. To achieve an effective taxation of activities one has to have a broad-based system in which taxes are levied on several commodities that go into the production of the activity in question. This, of course, is a type of policy with which we have practical experience: car use is taxed by a number of different tax types, and the same is true for other kinds of environmental problems.

7

THE DOUBLE DIVIDEND ISSUE

One issue that has received a lot of attention in policy debates and also in the academic literature is that of the so-called double dividend. We have already touched upon the main point: green taxes improve the environment, but they also raise revenue. Consider a tax reform where environmental taxes are increased while other taxes, like income taxes and the VAT, are reduced in an amount that keeps total tax revenue constant. Since these other taxes impose efficiency losses on the economy, this reform succeeds in substituting efficiency-enhancing for efficiencyreducing taxes. So the result is apparently that we are able to raise the same amount of tax revenue with less dead-weight loss at the same time as we improve the quality of the environment. This implies that we get two benefits or dividends from a green tax reform: (1) the state of the environment is improved and (2) we adopt a tax system with less distortion of commodity and factor markets. The argument is an attractive one with a strong appeal to economic intuition. But a closer analysis shows that we need to think carefully about it before making serious recommendations about tax reform.12 One problem with this idea is that it fails to be precise about the assumption that is made about the state of the tax system at the time of the reform. To see the importance of this, let us assume that the tax system has

The scale and scope of environmental taxation 311 been optimized according to the principles laid out in the previous section. Then, by the nature of an optimum, a small shift in the tax system towards more green taxation will have no effect on social welfare; the change in the environmental dividend is exactly cancelled by the associated change of the tax efficiency dividend. There is no double dividend at all. But the assumption that the tax system is optimal to begin with is an extreme one. The theory of optimal taxation was never designed to provide a realistic description of the actual tax system. An alternative approach is therefore to start from some description of the actual tax system that is closer to what the proponents of the double dividend idea have probably had in mind. One such assumption is that we start from a situation with no green taxes at all. The reason for this state of affairs might be that individuals and politicians have been ignorant of the environmental problems that exist in the economy or have not been aware that taxation could provide the solution to the problems. Then environmental taxes are introduced as a new element of the overall tax system with a reduction of the level of distortionary taxation in commodity and factor markets. Can we be sure, abstracting from distributional effects, that there is a gain in economic welfare? The answer, in general, is no, and it is not difficult to see why. The demand both for clean and dirty goods depends on the whole set of taxinclusive prices in the economy. Thus, a reduction in the consumer price of a particular clean good could increase the demand for a dirty good, thereby counteracting the initial effect of an increase in green taxes, making the environmental dividend doubtful. On the other hand, a price increase for dirty goods could by a similar argument affect the structure of demand for clean goods in such a way as to make the effects of price distortions worse than they were before. Our conclusion regarding the existence of a double dividend from a green tax reform must therefore be a cautious one: there could be such an effect, but it is not assured. Whether it will occur or not depends both on the initial state of the tax system and on the structure of demand, especially as regards the cross-price effects between markets for clean and dirty goods. Empirical studies are required in order to determine the outcome in any particular reform situation. A special version of the double dividend hypothesis concerns its effects on unemployment. It has been argued by several economists and policymakers that a tax reform that increases the level of green taxes and combines this with a cut in the payroll tax would lead both to an environmental gain and to a higher level of employment. The underlying idea is that in a situation of some kind of Keynesian unemployment equilibrium actual employment is determined by the demand for labour in private firms, and

312

The Elgar guide to tax systems

by lowering the payroll tax one decreases the price of labour to employers, thereby generating an increase of labour demand and employment.13 It should be noted that in order to evaluate this claim we need to use a model that is different from that of perfect competition. If the wage rate does not adjust so as to equalize demand and supply for labour we need an alternative theory of wage formation. One possibility is to use a model of trade union behaviour in which the union sets the wage rate while firms, taking this wage rate as given, determine the amount of employment – the so-called monopoly union model.14 In this case, the union will set the wage above the level that leads to elimination of unemployment. Although the union aims to maximize the welfare of its members, it has to weigh the welfare effects of higher wages for the employed against the welfare of the unemployed, and this will result in a wage rate that is above the market-clearing level. Cutting the payroll tax, it could be argued, leads to a reduction of the wage rate and therefore to less unemployment. Like the previous argument in favour of the double dividend, this also needs careful evaluation. A notable weakness of it is that it completely disregards the problem of tax incidence. If firms are confronted with lower gross wages (wages plus payroll tax) so that they earn higher profits, it is reasonable to believe that unions will claim some of these gains for themselves by increasing their wage demands; if so, this would lead to a smaller gain for society in terms of unemployment reduction. A further complication is that the increase of green taxes on goods like energy and transportation will reduce the real wage of the workers, giving them cause to argue for compensatory nominal wage increases. Taking account of both elements of tax incidence, one has to face the possibility that the net effect of a green tax reform on wages and unemployment may in fact be quite small. Although our discussion of the double dividend has to end on an agnostic note, this is not to say that a green tax reform may not be able to produce both a better environment and a more efficient tax system. But whether it will actually succeed in this is a question that cannot be settled on theoretical grounds alone; each tax reform proposal has to be evaluated on the basis of the empirical context in which it is proposed.

8

TAXES, REGULATIONS AND THE COSTS OF ADMINISTRATION

Pigouvian taxes are clearly not the only instruments whereby governments can attempt to internalize environmental externalities. An alternative to taxes in many cases is to use transferable quotas. This instrument,

The scale and scope of environmental taxation 313 however, has so many points of similarity with taxes that it will not be discussed further here.15 But in practice governments use a number of other instruments, and although economists have traditionally had a strong preference for the market-based instruments of taxes and transferable quotas, the use of command-and-control instruments exemplified by non-transferable quotas and direct regulation of production technology or product quality is in fact widespread. The standard criticism of command-and-control instruments is first that they are likely to have higher administrative costs and second that they fail to achieve a given reduction of pollution at minimum cost to society. One may naturally ask, therefore, whether the use of command-and-control instruments is simply a case of misguided policy or whether it can be justified by more careful theoretical analysis. A crucial point regarding the efficiency properties of Pigouvian taxes is that the government is capable of choosing the right tax base. But, as pointed out in Section 5 above this is not always a reasonable assumption. Quite often it is the consumption or production of a particular good in a particular social situation or activity that generates the externality, whereas taxes typically have to be general taxes that are paid on every unit of the good that is produced or consumed. A case in point is the consumption of cigarettes. The taxation of smoking is often justified by the argument that ‘smoking is bad for you’, which is one that goes beyond the externalities framework. Although paternalistic arguments deserve more serious consideration than they often receive, the focus here will be on the externalities of cigarette smoking, which have also been used as arguments for high excise taxes on cigarettes. But in many social contexts or situations it can be argued that there are hardly any externalities from smoking, or at least that the externalities generated in some contexts are much less than in other social settings. This would call for differentiated taxation of cigarettes depending on the social context or activity in which they are used; as pointed out in Section 5, one would ideally like to tax activities rather than commodities. But this is not a feasible strategy of taxation, since it would involve prohibitive administrative costs. A simpler alternative would be to impose a ban on cigarette smoking in particular places like lecture halls, football stadiums and restaurants, where the negative externalities from passive smoking are particularly serious. A combination of taxes and regulations might therefore in many cases be a more efficient policy – both in the sense of minimizing distortions and having lower costs of administration – than that of exclusive reliance on taxes. This raises interesting questions of the choice of an optimum combination of taxes and regulations, and a number of these have been analysed in a recent paper by Christiansen and Smith (2009). One of their conclusions is that if a Pigouvian commodity tax is

314

The Elgar guide to tax systems

supplemented by regulation it is the use of the commodity that causes the larger externality that should be regulated. While the cost of differentiation is an important element of the overall costs of administration of the tax system, another element is the control of tax evasion. Tax evasion has received most attention by economists in the context of income tax compliance, but there is also considerable evidence of the importance of indirect tax evasion. An interesting question is now whether the efficiency properties of Pigouvian taxes carry over to the case where there is tax evasion. This problem has been analysed in Sandmo (2002), where it is shown that to a large extent they do, since the theoretical analysis shows that decisions about output and emissions are often separable from the decision to evade taxes; the efficiency properties of taxes therefore remain the same as in the standard analysis. Using the same theoretical framework for the analysis of quota violations it can be shown that the combination of the fine or penalty rate and the probability of detection may be designed so that the expected fine has many of the efficiency properties of Pigouvian taxes. The gap between taxes and quotas on the one hand and regulations on the other may therefore not be as wide as often alleged in the literature.

9

DO TAXES CROWD OUT INTRINSIC INCENTIVES?

Most of the literature on the public good type of externalities – prophetically referred to as ‘the creation of atmosphere’ by Meade (1952) – assumes that the individual consumer and firm do not care about the amount of their own emissions of pollution. The agent may care about the public good that is negatively affected by their activities, such as clean air and the absence of noise, but in their behaviour they do not take into account their own (negative) contribution to the public good. The reason is simply than in a setting where many agents contribute to the quality of the public good, incentives are such that it is rational for the agent to neglect the link between their own decisions about consumption and production and the public good, although they may well realize that in the aggregate such a link does in fact exist. The purpose of environmental taxation is to create private incentives to act in a socially rational way by creating private incentives that would otherwise be non-existent. However, the realism of this assumption is questionable. If it were literally true, we would certainly see much more littering of public places than we actually do, and there would be fewer cases of people who buy environmentally friendly products when cheaper alternatives are available.

The scale and scope of environmental taxation 315 In order to capture this aspect of behaviour we need to extend our model of individual motivation. We must assume that the individual agents not only care about their private interest in the sense of their utility of consumption or their profit, but that their own individual emissions contribute negatively to their utility or perceived profit. In order to fix ideas it may be useful to consider the simple case of a competitive firm that emits a harmful substance as a by-product of its activities. The firm maximizes revenue minus costs, where the cost function varies positively with output and negatively with emissions (this implies that costs increase when emissions are reduced). Another deduction from revenue is a tax per unit of emissions. Under standard profit-maximizing assumptions, the firm will set its marginal cost of production equal to the market price and the marginal cost of pollution reduction equal to the tax rate on pollution (the Pigouvian tax). Suppose now that there is an additional element of cost that reflects the firm’s negative evaluation of its pollution activities.16 This subjective element of cost is increasing in the amount of pollution, and it is natural to assume that the marginal cost is also increasing. The conditions for profit maximization are now, first, that the marginal cost of production equals the market price as before, and that the marginal cost of pollution reduction is equal to the sum of the Pigouvian tax and the ‘conscience tax’. The latter concept is simply the additional element of cost that is incurred with a small increase in the amount of pollution. An interesting question is what happens to the conscience tax with an increase in the Pigouvian tax rate. This implies that ‘extrinsic incentives’ are introduced in a situation where ‘intrinsic incentives’ already exist and an important question is now whether a strengthening of the former will lead to a weakening of the latter.17 It has been claimed by several authors, for example, by Weck-Hannemann and Frey (1995) and Frey (1997), that an increased use of extrinsic or financial incentives will so weaken intrinsic or moral incentives that the net effects of policy instruments like Pigouvian taxation may be doubtful. In the model outlined above this hypothesis is supported in the sense that an increase in the rate of Pigouvian tax leads to a reduction in the conscience tax, but by less than the increase in the formal tax rate. In other words, the conscience tax is partially but not completely crowded out by the Pigouvian tax. This tax still has the power to curb pollution but by less than one would be led to think on the basis of the pure theory of profit maximization. How destructive are these considerations in relation to the economic case for green taxes? In the model of the example, an increase of the Pigouvian tax rate still leads to reduced emissions and is therefore an important policy instrument if policy-makers consider that intrinsic

316

The Elgar guide to tax systems

incentives are too weak to reach the socially optimal reduction of emissions. Moreover, on the natural assumption that the conscience tax varies among individual agents, the Pigouvian tax helps to make the total tax on pollution – that is, the formal tax rate plus the conscience tax – more uniform among polluters. This would be a step towards equalizing the marginal cost of pollution cutbacks across economic agents and therefore towards production efficiency in environmental policy. There is more to be said about this issue, however. It is clearly possible to argue that the preservation of intrinsic incentives in itself is good for the well-being of society. An incentive scheme that drives out moral attitudes may in this view create a society of individuals and firms that feel less responsible towards the environment, and the social cost of such a development of moral attitudes may be serious. On the other hand, one could also argue that the adoption of a system of extrinsic incentives for environmental improvement could increase people’s awareness of these problems and thus contribute to more environmentally friendly attitudes.

10 INTERNATIONAL ASPECTS OF ENVIRONMENTAL TAXATION Some of the most significant environmental problems of our time are of a global nature.18 In particular, the issue of global warming has caused great concern among natural scientists and politicians as well as among economists as witnessed by the debate following the publication of the Stern Review (Stern, 2006). The perspective taken by Stern is that of environmental policy as a form of social insurance: we may not know with perfect certainty the extent and effects of global warming, but the risks involved are certainly large enough to justify spending about 1 per cent of global GDP annually to reduce greenhouse gas emissions. The policy alternatives that are available to achieve the goal of a substantial decrease in global emissions include both taxes, quotas and regulations, so that much of the analysis of international environmental policy, including the use of Pigouvian taxes, can benefit from the insights derived from the study of domestic policies. However, the international setting raises some new problems that require careful consideration. The state of the atmosphere is an example of a global public good. A global public good is characterized by being publicly available to all of the world’s population in equal amount. This does not imply that all the world’s inhabitants value the good in equal measure, but simply that we cannot improve the quality of the atmosphere for one group of people

The scale and scope of environmental taxation 317 without at the same time improving it for everyone in the world.19 The optimal state of the atmosphere has been achieved when the marginal benefit of preventing global warming is equal to the cost of preventing it. One of the ways to achieve the reduction of emissions into the atmosphere is the use of Pigouvian taxes on a global scale. The problem that arises at this stage is that while the benefits of environmental policy are global, the costs may have to be borne locally. Confronted with the problem of the determination of global public goods supply, all individual countries of the world find themselves in the situation of the single individual in a conventional public goods setting: in the process of arriving at a global solution, each country has an incentive to misrepresent its preferences and costs and to free ride on the benefits provided by the policies of other countries. However, if all countries reason in the same way, the result will be a serious underinvestment in the prevention of global warming, taking the form, for example, of suboptimal use of green taxes from the point of view of global economic welfare. One of the advantages of Pigouvian taxes in a domestic setting is that they tend to produce a reduction of pollution emissions that satisfies the demand for production efficiency. Because all polluters face the same tax rate, they will equate the marginal cost of pollution cutbacks to the tax rate. This marginal cost accordingly becomes equalized among producers and the aggregate reduction of emissions is minimized for society as a whole. Applying this line of reasoning to global externalities, one would like to ensure that the marginal cost of reducing pollution is the same for all polluters, and this can be achieved by imposing the same tax rate on emissions from all countries. This is the chief justification for the much discussed proposal of a global carbon tax, that is, a tax on CO2 emissions. There are two issues that arise in connection with this proposal that are significantly different from the analysis of domestic taxes. One is the lack of an international authority that has the power to impose global taxes, that is, taxes that all the world’s nations would be obliged to adopt. In the absence of such an authority, the adoption of a global carbon tax requires the voluntary consent of all countries affected by the measure. This is clearly a difficult agreement to bring about in view of the international free-rider problem. But even if this problem could somehow be overcome there is an additional although related problem regarding the distributional impact of the tax in a world with enormous disparities in terms of per capita income and consumption. Is it fair that the poor areas of the world should pay the same rate of tax per unit of CO2 emissions as the rich industrialized countries? There are basically two ways out of this dilemma. One is to differentiate

318

The Elgar guide to tax systems

the tax between rich and poor countries, letting the developing countries pay at a lower rate than the industrialized countries. Obviously, the differentiation scheme would have to be designed with some care, for example, so as to avoid major jumps in the tax rate as countries move into higher tax brackets with increasing GDP per capita.20 Whatever the degree of differentiation, this system would entail a loss of global production efficiency in that the marginal cost of pollution reduction would no longer be equalized among countries, but this could be regarded as a price to be paid for a more equitable distribution of world income. The other solution to the dilemma is to combine a uniform CO2 tax with an international scheme of redistribution whereby rich nations make transfers of income to the poor countries in return for their agreement to adopt a globally uniform emissions tax. In this scheme, the international uniformity of the tax would ensure global production efficiency without adverse consequences for the degree of international inequality.

11 THE POLITICAL ECONOMY OF ENVIRONMENTAL POLICY Given the efficiency properties of green taxes one might expect them to be among the more attractive forms of taxation to be considered by governments and their electorates, but this is in fact far from being the case. Many politicians are difficult to convince about the social benefits of environmental taxation and among the general public proposals to introduce such taxes are frequently met with a good deal of scepticism if not outright hostility. There may be a number of reasons for this state of affairs, and the possible explanations may vary in importance from one country to another and also over time. Below I consider briefly some of these. An important motivation for many of those on the right of the political spectrum who resist the introduction of environmental taxes is undoubtedly that they have a negative attitude towards attempts at what they see as unwarranted expansion of the public sector. The thought experiment of economists who analyse tax reform under the assumption of constant tax revenue is not well understood; instead, the arguments are interpreted as just another set of proposals to increase tax revenue and thereby the size of the public sector.21 Another explanation for sceptical attitudes towards green taxes is that the incentive arguments may in fact not be easy to understand. The arguments of economists regarding private incentives, the importance of elasticities, the efficiency costs of taxes and so on, are much less evident to the general public than many economists tend to believe. Perhaps the

The scale and scope of environmental taxation 319 emphasis in some of the policy literature on the double dividend from a green tax reform – especially in the form of its effects on unemployment – stem from the desire of some of its supporters to provide a sales argument that is of more concrete substance than the benefits to the natural and social environment. A related explanation is that green taxes may be regarded by many as ‘lifestyle taxes’. Higher taxes on petrol and other sources of energy will make it more expensive to drive private cars, live in the suburbs, have a second home in the countryside as well as carrying out a number of other energy-intensive leisure activities. To the extent that lifestyles are fixed in the short run, the argument that energy taxes may have longrun benefits, for example, in the form of smaller houses that require less energy and less commuting may carry little weight with much of today’s electorate. Taxes that threaten the financial viability of a lifestyle that one has become accustomed to are hardly likely to be very popular among voters. For some aspects of environmental taxation there may also be considerable scepticism and ignorance regarding the real effects of environmental policy. This applies with special force to the crucial issue of global warming and may explain the popular appeal of the arguments advanced by those who claim that there is no global warming problem that requires painful measures like increased energy taxes. Obviously, it is also difficult to get popular support for policies that impose sacrifices on today’s consumers and producers for what is seen as the doubtful benefits to those who will live in a distant future.

12 FURTHER ISSUES AND CONCLUDING REMARKS This chapter has covered a number of problems related to the theory and practice of environmental taxation. Nevertheless, there are clearly a number of issues that have been left out and would have deserved more careful attention, and some of these may require brief mention. First, as in almost the whole of the literature in this area, our discussion of environmental externalities has taken the competitive equilibrium as its point of departure. Firms and consumers take market prices as given; similarly, because of their individual inability to affect market outcomes they also take tax rates as given. Analytically, there is much to be said for this procedure because it abstracts from the complexities involved in taking account of other imperfections of the market system. However, real economies are in fact characterized to a large extent by

320

The Elgar guide to tax systems

imperfect competition, which raises two interesting issues for the analysis of environmental taxes. One is the design of optimal taxes in a setting where market prices, even in the absence of externalities, are not such as to lead to an efficient allocation. The other is that with only a few firms on the production side of the market, a tax that is levied on these is likely to be seen by the firms concerned as subject to negotiation with the tax authorities, and the competitive model with price and tax takers no longer applies. I have only briefly touched on the argument of production efficiency in environmental policy: when all agents face the same tax rate on their polluting activity the marginal cost of pollution cutbacks will be equalized between them. (This argument applies obviously to firms, but it can also easily be adapted to the case of consumers, especially if we adopt the activities perspective of Section 5 above.) In this way, environmental taxation achieves static production efficiency.22 But there is a further argument that relates to dynamic efficiency. The static argument applies to a situation where the technology of production is given. However, higher environmental taxes also provide incentives to change the technology of production so as to lower costs, thereby encouraging firms to develop new and more environmentally friendly production technologies. The foregoing discussion has set out the benefits of environmental taxation but it has also emphasized the difficulties that arise regarding their practical design. Although it is important to be aware of the complexities involved in a process of tax reform, it should also be pointed out that the adoption of green taxes that are less than perfect when measured in terms of idealized theoretical models may still involve substantial benefits for society. To revert once more to the car example: the fact that a tax on petrol is unable to differentiate between different types of car use should not be taken as a decisive argument against regarding the petrol tax as a green tax. Green taxation is an interesting example of the cases where economists can be regarded as the inventors of policy instruments. While both the income tax and ordinary indirect taxes have been developed over a long period of time, having been refined and extended by generations of politicians and bureaucrats, environmental taxation is a relatively recent innovation in the tax system that owes its existence to a large extent to the creative efforts of academic researchers. Since the time of Pigou, a number of economists have been involved in establishing both better theoretical insights and more extensive empirical knowledge of the effects of these taxes. More effort should also be spent in making the theory of green taxation better understood among the general public.

The scale and scope of environmental taxation 321

NOTES * 1. 2.

3.

4. 5. 6. 7. 8.

9. 10 11. 12. 13. 14. 15. 16.

17. 18.

This paper was prepared for the conference ‘Tax Systems: Whence and Whither. Recent Evolution, Current Problems and Future Challenges’, Malaga, 9–12 September 2009. I am indebted to Xavier Labandeira for his comments on a previous version of the paper. Externalities had also been discussed by Pigou’s teacher, Alfred Marshall, but his interest was mainly motivated by his desire to explain how the industry supply curve could be downward-sloping under competitive conditions. The distinction between on the one hand deviations between private and social marginal costs and on the other hand between private and social marginal benefits is often arbitrary and of little real significance. One of the costs that car congestion imposes on society is an increase in the time use required to commute, but whether this should be classified as an extra cost or as a reduction in the benefits of driving is mostly a matter of analytical convenience. These are defined as unrequited payments to governments on tax bases that are deemed to be of particular environmental relevance. Tax bases of this kind are mainly to be found in the fields of energy, transportation, emissions to air and water, waste management and noise. For a more detailed presentation of the OECD data and an outline of recent environmental tax policy in various countries see Barde and Braathen (2005). The various externalities connected with car ownership and use have been discussed by Parry et al. (2007). The exception to this rule is electric cars, which receive preferential treatment in being allowed to drive in the bus lanes and park at meters without charge. A sophisticated scheme for road pricing along these lines was proposed by Vickrey (1963). For a recent discussion of road user and congestion charges see Newbery (2005). The question has been discussed in Bruvoll (2009) who compares the OECD assessment of the revenues from environmental taxes in Norway with an estimate based on a more narrow definition used by a government commission on excise taxes. The latter shows the revenue from environmental taxes in Norway in 2007 to be less than 20 per cent of the OECD estimate. This issue was discussed in the context of optimal taxation theory in Sandmo (1976). Indeed, one interpretation of the inverse elasticity rule is that the percentage reduction of private consumption should be the same for all goods. For more details, the reader is referred to the Appendix to this chapter. An anthology of recent contributions and a survey of the double dividend debate is Goulder (2002). Among the contributions to the analysis of this issue are Bovenberg and van der Ploeg (1996) and Koskela et al. (1998). An excellent survey of such models, although now somewhat dated in its coverage of the literature, is Oswald (1985). The equivalence of taxes and quotas pertains especially to their efficiency properties. Their distributional consequences might, however, be quite different, dependent on the manner in which the government chooses to distribute quotas to polluters. It may be objected that this assumption is inconsistent with the assumption of competitive markets with free entry: if the firm imposes such a cost on itself it will be squeezed out of the market by new entrants that do not ‘suffer’ from concerns about the environment. On the other hand, it is open to debate whether a realistic model of competitive markets really requires this knife-edge view of the consequences of deviations from pure profit maximization. Anyway, the case that is discussed here should be taken as a simple example of a more general approach that needs more justification than can be provided in the present context. A general analysis of the various issues that arise in the interaction between intrinsic and extrinsic incentives is Bénabou and Tirole (2006). Although the emphasis here is on global warming, there are other important examples

322

19.

20.

21. 22.

The Elgar guide to tax systems of global environmental problems, such as pollution of the oceans and the reduction of global biodiversity. This statement needs some modification in that groups of the world’s population might evaluate global warming positively because their own local environment might benefit from enjoying a more pleasant climate and higher agricultural productivity. For an analytical treatment of the theory of global public goods and externalities see Sandmo (2003, 2005). Such jumps would have unfortunate effects especially with regard to tax compliance; they would give countries strong incentives to underreport their per capita GDP when it approaches the lower limit of a new tax bracket. Reinforcing this problem is the considerable degree of uncertainty attached to the computation of the exact magnitude of GDP. Letting small differences in the principles of GDP estimation have significant effects for a country’s tax liability would clearly be undesirable. In fairness, it should also be pointed out that some of the adherents of environmental taxes on the left may put too much weight on these taxes as sources of revenue for a larger public sector. This argument received early emphasis in the article by Baumol and Oates (1971).

REFERENCES Barde, Jean-Phillippe and Nils Axel Braathen (2005), ‘Environmentally Related Levies’, in Sijbren Cnossen (ed.), Theory and Practice of Excise Taxation, Oxford: Oxford University Press. Baumol, William J. and Wallace E. Oates (1971), ‘The Use of Standards and Prices for Protection of the Environment’, Swedish Journal of Economics, 73(1), 42–54. Bénabou, Roland and Jean Tirole (2006), ‘Incentives and Prosocial Behavior’, American Economic Review, 96(5), 1652–78. Bovenberg, A. Lans and Frederick van der Ploeg (1996), ‘Optimal Taxation, Public Goods and Environmental Policy with Involuntary Unemployment’, Journal of Public Economics, 62(1), 59–83. Bruvoll, Annegrete (2009), ‘On the Correct Measurement of Environmental Taxes’, Discussion Paper No. 599, Oslo: Statistics Norway. Christiansen, Vidar and Stephen Smith (2009), ‘Externality-correcting Taxes and Regulation’, Working Paper No. W09/16, London: Institute for Fiscal Studies. Diamond, P.A. and J.A. Mirrlees (1971a), ‘Optimal Taxation and Public Production, Part I: Production Efficiency’, American Economic Review, 61(1), 8–27. Diamond, P.A. and J.A. Mirrlees (1971b), ‘Optimal Taxation and Public Production, Part II: Tax Rules’, American Economic Review, 61(3), 261–78. Frey, Bruno S. (1997), ‘A Constitution for Knaves Drives out Public Virtues’, Economic Journal, 107(443), 1043–53. Goulder, Lawrence H. (ed.) (2002), Environmental Policy Making in Economies with Prior Tax Distortions, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Kaplow, Louis (2008), The Theory of Taxation and Public Economics, Princeton: Princeton University Press. Koskela, Erkki, Ronny Schöb and Hans-Werner Sinn (1998), ‘Pollution, Factor Taxes and Unemployment’, International Tax and Public Finance, 5(3), 379–96. Leape, Jonathan (2006), ‘The London Congestion Charge’, Journal of Economic Perspectives, 20(4), 157–76. Meade, James E. (1952), ‘External Economies and Diseconomies in a Competitive Situation’, Economic Journal, 62(245), 54–67. Mirrlees, James A. (1971), ‘An Exploration in the Theory of Optimum Income Taxation’, Review of Economic Studies, 38(2), 175–208. Musgrave, Richard A. (1959), The Theory of Public Finance, New York: McGraw-Hill.

The scale and scope of environmental taxation 323 Newbery, David M. (2005), ‘Road User and Congestion Charges’, Chapter 7 in Sijbren Cnossen (ed.) (2005), Theory and Practice of Excise Taxation, Oxford: Oxford University Press. Oswald, Andrew J. (1985), ‘The Economic Theory of Trade Unions: An Introductory Survey’, Scandinavian Journal of Economics, 87(2), 160–93. Parry, Ian W.H., Margaret Walls and Winston Harrington (2007), ‘Automobile Externalities and Policies’, Journal of Economic Literature, 45(3), 373–99. Pigou, Arthur C. (1920), The Economics of Welfare, London: Macmillan. Pigou, Arthur C. ([1928] 1947), A Study in Public Finance, London: Macmillan. Ramsey, Frank P. (1927), ‘A Contribution to the Theory of Taxation’, Economic Journal, 37(145), 47–61. Samuelson, Paul A. ([1952] 1986), ‘Theory of Optimal Taxation’, unpublished version, 1952; Journal of Public Economics, 30(2), 137–43. Sandmo, Agnar (1975), ‘Optimal Taxation in the Presence of Externalities’, Swedish Journal of Economics, 77(1), 86–98. Sandmo, Agnar (1976), ‘Direct versus Indirect Pigouvian Taxation’, European Economic Review, 7(4), 337–49. Sandmo, Agnar (2000), The Public Economics of the Environment, Oxford: Oxford University Press. Sandmo, Agnar (2002), ‘Efficient Environmental Policy with Imperfect Compliance’, Environmental and Resource Economics, 23(1), 85–103. Sandmo, Agnar (2003), ‘International Aspects of Public Goods Provision’, in Inge Kaul et al. (eds), Providing Global Public Goods, Oxford: Oxford University Press. Sandmo, Agnar (2005), ‘Environmental Taxation and Revenue for Development’, in A.B. Atkinson (ed.), New Sources of Development Finance, Oxford: Oxford University Press. Stern, Nicholas (2006), The Economics of Climate Change: The Stern Review, Cambridge: Cambridge University Press. Vickrey, William (1963), ‘Pricing in Urban and Suburban Transport’, American Economic Review, 53(2), 168–81. Reprinted in R. Arnott, K. Arrow, A.B. Atkinson and J.H. Drèze (eds), Public Economics. Selected Papers by William Vickrey, Cambridge: Cambridge University Press. Weck-Hannemann, Hannelore and Bruno S. Frey (1995), ‘Are Incentive Instruments as Good as Economists Believe?’, Chapter 8 in Lans Bovenberg and Sijbren Cnossen (eds), Public Economics and the Environment in an Imperfect World, Dordrecht: Kluwer Academic Publishers.

324

The Elgar guide to tax systems

APPENDIX This appendix sets out in more (although not complete) detail the steps leading to the conclusions about the optimal tax structure in an economy with environmental externalities that are discussed less formally in Section 6. Assume that the economy consists of n identical individual consumers. Each of them has a utility function that depends on the consumption of three private goods, numbered 0, 1, 2. In addition, utility depends on the amount of environmental pollution, e. Assuming that the utility function is weakly separable between consumption and the quality of the environment,1 the utility of a representative consumer can be written as: U 5 U (u (x0, x1, x2) ,e) .

(8A.1)

The marginal utilities of the three consumption goods are positive,2 while the derivate with respect to the environmental variable, e, is negative. As another simplification, and abstracting from the complexities discussed in Section 5, the amount of environmental pollution is assumed to be proportional to the aggregate consumption of commodity 2: e 5 nx2.

(8A.2)

The budget constraint of the consumer is: x0 1 P1x1 1 P2x2 5 y.

(8A.3)

Thus, commodity 0 is taken to be the numeraire, which is untaxed, while P1 and P2 are the consumer prices. Specific taxes are levied on the two taxable goods, so that: P1 5 p1 1 t1 and P2 5 p2 1 t2.

(8A.4)

Producer prices, denoted by p1 and p2, are assumed to be given from the production side of the economy. We first consider the demand behaviour of the representative consumer. Maximizing the utility function (8A.1) subject to the budget constraint (8A.3) enables us to derive the individual demand functions as: x0 5 x0 (P1, P2, y) , x1 5 x1 (P1, P2, y) , x2 5 x2 (P1, P2, y) .

(8A.5)

Substituting the demand functions into the utility function (8A.1) we get the indirect utility function:

The scale and scope of environmental taxation 325 V 5 V (v (P1, P2, y) , e) .

(8A.6)

At this point it is natural to remark on the assumption of weak separability that was introduced above. It is easy to see that because environmental pollution, e, does not enter into the sub-utility function u, or into the budget constraint, the demand functions will be independent of the quality of the environment. If this assumption had not been made, we would have had to take account of the environmental feedback on demand: a price increase would have a direct effect on demand for each of the two goods, but since the price would also affect the quality of the environment via the consumption of commodity 2, there would be an additional effect that can be neglected in the present formulation.3 However, this simplification does not affect the substantive economic contents of the analysis. Note also that consumers take the state of the environment as exogenously given so that each consumer disregards the fact that his or her own consumption affects the state of the environment, which is a natural assumption in the context of a large economy. The government chooses optimal tax rates t1 and t2 with the objective of maximizing the utility of the representative individual. This objective is obviously equivalent, in the case of identical individuals, to the utilitarian sum of utilities: W 5 U (u (x0, x1, x2) , e) 5 V (v (P1, P2, y) , e) .

(8A.7)

The budget constraint of the government requires that a given amount of per capita revenue, T, be raised through commodity taxation, so that: t1x1 1 t2x2 5 T.

(8A.8)

It is assumed that lump sum taxation, that is, a tax levied directly on y, is not an available instrument of tax policy. The impossibility of lump sum taxation is motivated by more complex models with heterogeneous individuals and informational problems; here we use this insight in a simplified context, which in itself admittedly does not explain the impossibility of first-best optimal taxation. The government maximizes social welfare subject to its revenue constraint. Forming the Lagrangian function:

L 5 V(v (P1, P2, y), e) 1 m[t1x1 1 t2x2 2 T],

(8A.9)

we can take the derivatives with respect to t1 and t2 and use the first-order conditions to characterize the optimal tax system. Instead of presenting

326

The Elgar guide to tax systems

the results in their most general form, given the limitations of the threecommodity model, we focus on the special case where demand functions are independent, so that the cross-price derivatives are zero (for more details of the mathematical derivation as well as an analysis of the more general case, see Sandmo, 2000). Defining the percentage tax rates as u1 = t1/P1 and u2 = t2/P2 we can write the optimal tax rates as: u1 5 a (21/e11)

(8A.10)

u2 5 a (21/e22) 1 (1 2 a) (2nVe /lP2)

(8A.11)

Here we have defined a = (m 2 l)/m as the percentage gap between the marginal utilities of income in the public and private sectors. In order to interpret the optimality conditions, it is useful to begin by disregarding the second term in equation (8A.11). We see then that optimal tax rates are characterized as being proportional to the inverse of the price elasticities of demand; a result first derived by Ramsey (1927) and later restated and extended by Samuelson ([1952] 1986), Diamond and Mirrlees (1971a, 1971b) and many others. Intuitively, taxes ought, for efficiency reasons to distort relative quantities as little as possible, and this is precisely what is achieved by taxing the more price elastic commodities at the highest rates. The second term in equation (8A.11) expresses the marginal social damage. The damage at the individual level, expressed in unities of the polluting commodity, is Ve /lP2, and this must be multiplied by n to arrive at the social damage. Finally, in order to connect this analysis with the presentation in the main text of this chapter, we may define: u1R 5 (21/e11) , u2R 5 (21/e22) and uD 5 (2nVe /lP2) . Substituting these expressions into (8A.10) and (8A.11), the equations can be rewritten as: u1 5 a u1R.

(8A.12)

u2 5 a u2R 1 (1 2 a) uD.

(8A.13)

These equations correspond to the expressions in the text of Section 6 above. As also noted in Section 6, a notable feature of this solution is that the marginal social damage only enters into the tax formula for commodity 2,

The scale and scope of environmental taxation 327 while the tax on commodity 1 is characterized by its Ramsey term only. This is an example of the principle of targeting.

NOTES 1. The implications of this assumption will be discussed below. 2. If one of the three commodities is interpreted as leisure, the model could also be redesigned to accommodate a linear income tax. 3. For a fuller discussion of the environmental feedback effect see Sandmo (2000, Ch. 5).

9

Financing subnational governments with decentralized taxes Roy Bahl*

This chapter is about the case for assigning taxing powers to subnational governments, and about the structure of this revenue assignment. As Musgrave (1983) put it in perhaps the seminal paper on this subject, ‘Who Should Tax, Where and What’? This review reconsiders the Musgrave questions after 25 years, asks whether the international trend in tax assignment is in step with what economists have prescribed, and concludes with some thoughts about the most likely future for the decentralization of tax systems. We begin with a discussion of the concept of revenue assignment, and with the theoretical justifications and a priori reasoning usually given for revenue assignment to subnational governments. A basic issue taken up in this discussion is whether the practice matches up with the theory. In the following section, we turn to a consideration of the case for decentralizing each of the major tax bases. Throughout this discussion, we distinguish among industrial, transition, and developing countries.

1

CURRENT MODELS OF SUBNATIONAL FINANCE

Perhaps as much as by any other factor, subnational government financing systems are differentiated according to how much taxing power is decentralized. The assignment of taxing powers gives revenue-raising autonomy to subnational governments and is consistent with the goals of fiscal decentralization. Most students of fiscal decentralization would argue that revenue-raising power at the subnational government level is a necessary condition for successful fiscal decentralization. Simply assigning larger shares of central government tax collections may fill the vertical financing gap for subnational governments, but it will not give them autonomy to determine the size of their budget. It is no easy matter to measure and compare the degree of tax decentralization across countries. Even the concept of a ‘subnational government tax’ is confused. In theory, the subnational government could be given a range of powers to use in determining the level of taxes it will raise: 328

Financing subnational governments with decentralized taxes 329 ● ● ● ●

determine the tax rate; determine the tax base; collect the tax; receive all of the revenues.

A complete separation of central and subnational taxing authority would imply that subnational governments have the power to control all four dimensions of tax policy and administration. However, the only necessary condition for subnational government taxation, many would argue, is the power to set the tax rate (Bahl and Linn, 1992; Bahl and Bird, 2008a). The gray area in this definition is derivation-based intergovernmental transfers, that is, the case where a share of the revenues from a central tax is allocated to the regional government on a basis of origin of collection. The Chinese follow such a revenue-sharing system, as do most Eastern European countries. It might be argued in this case that the subnational governments can have an impact on revenues received through their promotion of local area economic development (growth in the tax base). To the extent they also have responsibility for tax administration, they may have an even greater influence on the growth in the tax base.1 Still, subnational governments have no power to directly affect revenues by changing the tax rate or the tax base. As to current models of tax decentralization, four general approaches may be identified. Each of these revenue assignment models leads to a different outcome in terms of the autonomy given to state and local governments. First, and most consistent with the goals of fiscal decentralization, is the assignment of taxing powers and tax administration responsibility to subnational governments, for example, as in the US and Canada. A second arrangement is for the subnational governments to piggyback on a central government tax base. Under this regime, the subnational government sets the tax rate, but the central government defines the base and administers the tax. Third, the tax rate and base may be set by the central government, but responsibility for collection may be assigned to the lower-level governments. The subnational governments then are given a prescribed share of collections. This is more or less the system used in Germany, and for certain taxes in many transition countries including China. Finally, the central government may set the tax rate and base, and collect the tax, but assign a portion of collections to the subnational government. The return may be on a derivation basis or on a formula basis. This is the system that operates in many developing countries, and is least in step with the concept of tax autonomy. It is best viewed as a form of intergovernmental transfer.

330

The Elgar guide to tax systems

One cannot rely on secondary data to classify countries according to which of these four models they have adopted. The available data sets that allow international comparison do not adequately take account of the discretion that subnational governments have in setting tax rates and determining tax bases.2 In particular, the IMF Government Finance Statistics (arguably the only comparative data available) do not distinguish among the first three approaches to subnational government finance.3

2

REVENUE ASSIGNMENT: THEORY AND PRACTICE

The traditional starting point for thinking about tax assignment is Musgrave’s (1983) multi-level budget framework that would assign the stabilization and distribution functions to the central government and allocation responsibility to the local governments. This division of responsibility leads to guidance about the placement of various instruments of taxation at the central, ‘middle’ and local government levels.4 Progressive taxes with a distributional goal and taxes containing automatic stabilizers would be left with the central government. Subnational governments would rely mostly on those taxes levied against relatively immobile bases. Most economists still hold to the Musgrave rules as the basic organizing principles for a fiscal federalism.5 But the theory and certainly the practice have moved on. Later work focused on the importance of competition among subnational governments as a way of controlling the size of governments (Brennan and Buchanan, 1980), on taxing non-residents for benefits received and on the political economy of decentralized tax assignment (Hettich and Winer, 1999; Lockwood, 2006). This thinking has explained the case for some claim on mobile tax bases by subnational governments. In terms of the contemporary practice, subnational governments now have taken on a substantial redistribution role, regularly use tax policy to stimulate their economies, and have been assigned major expenditure responsibilities. Moreover, ‘things have changed’. Economic integration in Europe and the reality of elected government in the developing countries have stimulated the demand for fiscal decentralization (Bahl, 2008; Stegarescu, 2009). The present practice of tax assignment is guided by many more considerations than economic efficiency. It would seem to rely more on a set of disparate arguments for and against passing more taxing powers to subnational governments, and on assessing the probability that any particular tax decentralization strategy could be implemented. As Oates (2008,

Financing subnational governments with decentralized taxes 331 pp. 329–30) has put it in commenting on the newer political economy approach: while the form of the analysis has new elements, the nature of the problem remains essentially the same: the issue is one of a tradeoff between the capacity of a centralized solution to provide ‘coordination’ of local outputs (i.e., internalize spillover effects) and the ability of a decentralized system to tailor outcomes to the preferences . . . of the local jurisdiction.

‘Why subnational government taxes, how much taxing power should be decentralized, and which taxes should be subnational?’ are questions that are now answered many ways. The answers are likely to be very different for countries at different stages of economic development. In the next three sections of this chapter, we revisit three old questions: ‘Are there net economic gains to be had from the imposition of independent subnational government taxes?’ ‘Under what conditions is a subnational government tax preferable to a user charge or an intergovernmental transfer?’, and ‘What are the norms for a good subnational government tax?’ The Benefits and Costs of Subnational Government Taxes The decision to decentralize more vs. less taxing power to state and local governments implies costs and benefits. Some of these costs and benefits have to do with economics but others are about bureaucratic politics, technical issues of tax administration, social equity, a paternalistic central attitude toward local government, and the degree to which the central government sees itself as the guardian of tax policy. On the expenditure side, the decentralization theorem is straightforward: push all expenditure responsibility down to the lowest level that is consistent with efficiency considerations. On the revenue side, the rules are much less easy to state. Potentially, the core benefit from increased revenue assignment is the welfare gain that results from financing decentralized services with decentralized taxes. This gain will be greater or smaller depending on the extent to which local government taxation is necessary to ensure full accountability of elected local officials. And, there are other considerations: ● ●

Under what conditions can tax decentralization be key to effectively imposing a hard budget constraint on subnational governments? Will tax decentralization lead to revenue competition between central and local governments and a dampening of overall tax effort, or will it lead to an enhancement in overall revenue mobilization?

332 ● ●

The Elgar guide to tax systems Will subnational government taxation lead to higher overall costs of tax administration? How can the fiscal disparities that come with the decentralization of taxing powers be dealt with?

These issues are taken up in the discussion below. Accountability6 Most students of fiscal federalism link revenue assignment to the welfare gains from the decentralization of service responsibilities. Locally imposed taxes make locally elected officials more accountable to their voting population for the public services that they deliver, that is, more accountable than if the services were financed by intergovernmental transfers. More accountability, ceteris paribus, translates into ‘better public services’, which in this case means that people will get more of what they want. It should follow that they will be more willing (or less resistant) to paying for it. This in turn would drive down the cost of revenue mobilization. The benefits of accountability (local voters get what they want from their elected officials) might be missed in many developing and transition countries. The electoral process might not be open and contested (China and Vietnam) or elections may have been suspended (Nepal). Voters are less mobile and in some cases have not learned how to use the vote to hold their officials accountable. Moreover, the central government may have held on to taxing powers so that the local government revenue structure is dominated by intergovernmental transfers. To the extent taxpayers perceive a linkage between what they pay and the services they get, they will tie service benefits more to the level of grants than to the level of local taxes. Even more likely, the benefits from increased subnational government taxes could be so negligible as to go unnoticed. The same will be true in industrialized countries with highly centralized fiscal systems (Spahn and Fottinger, p. 245). Finally, under such a centralized system, the process of making a decision to increase spending (and taxes) is not a transparent one and the voters may not know who to hold accountable. Hard budget constraint The conditions for fiscal decentralization with fiscal responsibility are that subnational governments should have autonomy at the margin in determining revenue and expenditure levels and that they should be required to balance their budgets. An interesting question is how much discretion at the margin is necessary to achieve this outcome. It is not possible to put a firm number to this margin, but clearly it will vary with the amount

Financing subnational governments with decentralized taxes 333 Table 9.1

OECD Transition Developing

Subnational government tax as a percentage of total taxa 1970s

1980s

1990s

2000s

19.8 (21) 22.4 (1) 10.1 (14)

19.6 (21) 16.4 (5) 12.0 (15)

24.1 (20) 21.6 (20) 13.4 (18)

25.3 (19) 20.9 (21) 9.6 (19)

Note: a. The average value for the decade is reported in each cell. The number of countries reporting is shown in parentheses. Source:

Computed from IMF (various years).

of expenditure assignment to the subnational government. For example, a subnational government with responsibility for no more than basic housekeeping functions (e.g., most rural local governments) will need less revenue discretion than will a provincial government with extensive Social Service responsibility (e.g., Spanish provinces). In many industrialized countries the answer is that a considerable amount of taxing power has been devolved – the OECD average is 25 percent of all taxes (Table 9.1). In developing countries, the answer seems to be that very little should be devolved – the average is less than 10 percent.7 With this concept of marginal discretion in mind, we might think of the following taxonomy of local governments. If they have no expenditure or revenue discretion and balance their budgets, they are little more than spending units of the higher-level government. If they have important expenditure responsibilities and expenditure discretion, but inadequate authority to raise revenues, they may choose to incur deficits. If the subnational governments are given discretion on both sides of the budget, but if the central government stands ready to cover overspending with ad hoc grants and other forms of subsidy, deficit spending by subnational governments is all but guaranteed. If the subnational governments are given independent taxing powers and expenditure discretion, and if the central government is committed to fiscal discipline (a hard budget constraint rule), balanced budgets can be guaranteed. If there is to be tax decentralization and fiscal responsibility, the local government must have authority to increase tax revenues in line with the expected range of natural and discretionary increase in assigned expenditure responsibility. Moreover, the subnational government should have the power to increase revenues for the fiscal year in which the deficit might occur, for example, via a rate increase or the adoption of a new tax. An ‘adequate’ margin is one condition for fiscal discipline. The other is a well-structured grant system where the amount that each local government

334

The Elgar guide to tax systems

may expect in any given year is fixed before the budget is formed. If grants are negotiable, that is, if the senior-level government is soft on the size of entitlements, tax discretion will not lead to fiscal discipline because local governments will overbudget with the expectation that their deficits will be covered. Japanese local governments have incurred an increasing financing gap, which has been covered by central government transfers and subsidies in one form or another. A so-called ‘trinity reform’ will address this problem by replacing a significant amount of grant funding with local government taxation authority (Ikawa, 2008). The German Länder regularly run deficits and finance these by short- and long-term borrowing (Rodden, 2003). Because of the equalizing nature of intergovernmental fiscal transfers, the poorer Länder have little incentive to restructure their fiscal system to find sustainable budget balance (Spahn and Fottinger, 1997, pp. 244–5). Subnational governments in most low-income and transition countries often have quite a limited margin, even by comparison with their assigned expenditure responsibility. And where they do have revenue discretion, it may be limited by a ceiling rate, or it may be over the more contentious levies such as the property tax, agricultural taxes or user charges. Of course, subnational government revenue discretion is not a sufficient condition for fiscal responsibility, as the cases of Brazil and Argentina demonstrate. Increased revenues? Will increased subnational government taxing powers lead to a higher or lower overall (central plus local) level of revenue mobilization? One could argue that there would be a dampening effect, for two reasons. First, the voting population at the subnational government level might not buy into the idea that higher taxes will result in better services.8 They might be more persuaded that these new revenues will lead to over-bloated payrolls and perquisites for local officials, or will be spent to satisfy the whims of bureaucrats and politicians rather than voters. Moreover, better-off families in some industrial countries have good substitutes for many local government services (e.g., private schools, security, and refuse collection), so they might be hesitant to vote a tax increase on themselves. The result could be that the newfound taxing powers of subnational governments would go unused. The tax revolt in the US, which was initiated by California’s Proposition 13, was implemented by referenda and by voterimposed, permanent limits on tax levels.9 A second reason to expect a dampening effect on aggregate revenues is that increased subnational government taxing power may encroach on the taxing space of higher-level governments. The result of higher local

Financing subnational governments with decentralized taxes 335 government taxes may be that voters will resist future increases in central government tax rates. A related problem is that higher joint tax rates and dual administration might lead to reduced compliance rates. For example, Plamondon and Zussman (1998) estimate that a single administration of the Canadian federal and provincial business taxes would reduce compliance costs by 1.3 percent of collections. The contrary argument is that decentralization on the tax side of the budget would enhance overall revenue mobilization. This could be due in part to a greater willingness to pay for services delivered under a decentralized fiscal system. It also could be due to some overall base-broadening that might result from dividing tax bases according to comparative advantages in assessment and collection. Subnational governments have comparative advantages in reaching some of those who are ‘hard to tax’ under central government regimes. The result of tax decentralization in such cases may be a net revenue gain. These ‘informational advantages’ of subnational governments that are most relevant to the case of developing and transition countries (Bahl and Bird, 2008b), can take many forms. Often, for instance, state and local governments oversee a variety of licensing and regulatory activities, and they track property ownership and land-based transactions. They thus have ample opportunity to identify local businesses and to gain some knowledge about their assets and scale of operation. Because the potential revenue gain is much more important for them in relative terms, subnational governments have more incentive to carry out such activities than do national governments. This provincial and local government knowledge of the tax base may allow them to capture some of those who presently do not fully comply with national taxes, or evade taxes altogether. This would include the self-employed – including small businesses – who often underdeclare taxable income and consumption. There is another factor that suggests a revenue enhancement effect from decentralization. ‘New taxation’ might lead to an overall revenue increase. In many countries, provincial and local governments have broadened their tax net with a variety of special tax instruments and administrative measures such as levies on the sales of assets of firms, licenses to operate, betterment charges, and various forms of property and land taxation (Bird and Wallace, 2004; Bird, 2006). What is the verdict? Does increased subnational government taxation crowd out central revenues and reduce overall revenue mobilization, or is it revenue-enhancing? There is not much empirical evidence on this question. Lotz (2006) has observed that there is no clear conclusion as to whether decentralized taxation power leads to an increase in the overall level of taxation. He points out that the Nordic countries have high overall

336

The Elgar guide to tax systems

levels of taxation and high levels of tax assignment to subnational governments, but Austria and Switzerland have high levels of assignment and lower levels of taxation. Some might see the pattern as more related to history and culture than to economic factors. Bahl and Cyan (2009) have carried out an econometric analysis of ‘crowding out’ for 58 developed and developing countries in the 1990s. They find evidence that increased local taxes do lead to reduced levels of central taxes for higher-income countries with more open economies, but not for developing countries. Tax administration An argument often made against proposals to decentralize taxing powers is that it would not take advantage of the superior tax administration capabilities of the higher-level governments. The result of shifting collection responsibility from more to less efficient administrative regimes will be an increased cost of collection for any given amount of tax revenues. While this proposition is almost always discussed in the case of developing countries, it could hold true to some extent for industrialized countries as well: ●

●

●

●

●

Central government tax administration can capture economies of scale. This might include centralized data processing services and record-keeping, uniform approaches to assessment and audit, the development of centralized training programs, and so on (Vehorn and Ahmad, 1997). Large taxpayers (companies) often operate on a country-wide basis, and also account for a significant share of total national revenues. Taxes on their operations can be most effectively administered (and allocated) by the national tax administration. Some sectors are harder to tax under a state or local government administration. For example, services and internet sales are not easily reached under the state sales tax in the US because of crossborder complications. The enforcement of tax collection requires the administering of penalties and possibly court actions that are beyond the reach of many subnational governments. Moreover, local governments in particular are close to the taxpayers who might be penalized, whereas provincial and central governments are one or several steps removed. The objectives of national policy may force central control. For example, the distribution objective of the income tax might require a central design and administration of the tax, industrial policy might dictate central administration of the tax on company income, and

Financing subnational governments with decentralized taxes 337

●

taxes on international trade are too locked in to trade agreements and valuation complications to be effectively administered by subnational governments. Fairness in taxation requires a uniform implementation of the tax code, and this is best done by a single (national) tax administration.

All this said, there are some areas of tax administration where central governments do not have a comparative advantage. In terms of all three of the basic functions of tax administration – identification of the taxpayer, assessment of tax liability, and collections – subnational governments might have a comparative advantage for certain taxes or for certain segments of the tax base.10 There is an especially strong administrative case for the property tax to be a local government levy. Local governments have a comparative advantage in identifying the tax base, because of their familiarity with local land use patterns. The methods of building a tax roll and valuing properties require site visitations, identification of ownership, and the tracking of improvements to properties. Central government administration of the land tax, where this is practiced, might work in a small country (e.g., Jamaica) but has not been very effective in large countries (e.g., Indonesia). The other comparative advantage for subnational government taxation relates to small taxpayers, who are usually less easily reached by central tax systems. Tax authorities in many developing countries bemoan the fact that the self-employed and the informal sector mostly escape taxation. In developing and transition countries, most revenues from major central government taxes can be traced to a relatively small number of taxpayers. The individual income tax is paid mostly by the formal PAYE sector of the economy and the VAT by larger firms. Not only are these failures at capturing the low end of the tax base recognized, they are institutionalized in that many developing countries set relatively high thresholds for the broad tax bases: ●

●

●

It is common for the VAT threshold to be set relatively high in terms of gross sales, precisely to exclude small taxpayers that are not costeffective to reach.11 The exemption level for the individual income tax sometimes is above the average income level. This is done partly to guarantee subsistence levels of after – tax income, but partly to recognize the collection difficulties for income earned outside the formal sector of the economy. Corporate income taxes often cover only large firms.

338

The Elgar guide to tax systems

Tax Assignment and Expenditure Assignment The correct revenue assignment to subnational governments depends on the expenditure responsibilities assigned to that government. (Bahl and Linn, 1983, 1992). Only when expenditure assignment is taken into account can the efficient level of local taxes be determined. For publicly provided goods and services, where the benefits accrue to individuals within a jurisdiction and where the exclusion principle can be applied in pricing, user charges are the most efficient financing instrument. This is a particularly relevant argument for public utilities such as water supply, sewerage, power, and telephones, but also for public transit and housing. These services may involve some external benefits, but most of the benefits are likely to be local in nature. Other local government services, such as general local administration, traffic control, road maintenance, street lighting, security, primary schools, local clinics and parks, and recreation are local public goods whose primary benefits accrue to the local population. The same may be said, at the provincial or state government level, of secondary schools, universities, mental hospitals, trunk roads, bridges, and the like. However, here the exclusion principle in pricing cannot be applied. These services are most appropriately financed by taxes whose burden is local (provincial). For services in which costs or benefits spill over local boundaries – such as health, higher education and certain types of infrastructure expenditures – provincial or national intergovernmental transfers should contribute to financing. Full local financing would lead to under-provision of these services from a regional or national perspective, and full financing from transfers would not recognize local benefits. This efficiency justification for subnational government taxation, and the link to expenditure assignment is important. It enables us to restate a basic principle of intergovernmental fiscal reform: determine expenditure assignment before working out the revenue assignment and revenuesharing provisions. Otherwise, it is not possible to develop an efficient mix of revenues for the local authorities. What Is a Good Subnational Government Tax? The (updated) traditional view on this question is well summarized by Oates (1996, p. 36): (1) Lower levels of government . . . should, as much as possible, rely on benefit taxation of mobile economic units, including households and mobile factors of production. (2) To the extent that nonbenefit taxes need to be employed on

Financing subnational governments with decentralized taxes 339 mobile economic units, perhaps for redistributive purposes, this should be done at higher levels of .  .  . government. (3) To the extent that local governments make use of nonbenefit taxes, they should employ them on tax bases that are relatively immobile across local jurisdictions.

This traditional view, narrowly interpreted, leads to a recommendation that (third-tier) local governments should choose immobile tax bases that are easily administrable. In many countries this means narrow-based taxes that are not revenue-productive (and are of not much interest to higherlevel governments). But Oates’s argument for subnational government benefit charges also makes the case for local taxation of mobile factors. An important footnote to this ‘rule’ is that factor mobility and thus efficient tax choices will depend on whether the subnational government is a municipality, a metropolitan region, or a province. In general, factor mobility will be less of a constraint in larger taxing regions. Of course, resource allocation is not the only criterion that is used in revenue assignment. Various countries have made the assignment decision based on revenue needs and administration criteria. For example, developing countries rarely devolve personal income or payroll taxes to subnational governments because the central government feels too pressed to give up any access to the income base. The same may be said of general consumption taxes and the more revenue-productive excises. Most industrial countries, by contrast, seem quite willing to decentralize broad-based taxes. Where does this leave it in terms of some basic principles for choosing a ‘good’ subnational government tax? Public finance economists and practitioners often recite revenue adequacy, economic efficiency, vertical equity, administrative feasibility, and political feasibility as the goals, but are not always in agreement about which taxes should be assigned to which level (Bahl and Linn, 1992; McLure, 1998; Martinez-Vazquez, 2008: Bird, 2008). Revenue adequacy The primary goal for subnational government taxation is to raise revenues to meet assigned expenditure needs. If tax bases that cannot yield adequate revenues at reasonable rates are assigned, the result is likely to be both deficient public services and a haphazard gap-filling strategy that includes the imposition of a variety of undesirable and distortionary fees, levies, and informal charges.12 The revenue adequacy criteria suggests that any potential revenue source should be evaluated in terms of two ‘yield’ characteristics. First, the base should be broad enough so that adequate revenues can be raised

340

The Elgar guide to tax systems

Table 9.2

OECD Transition Developing

Subnational government tax revenues as a percentage of GDPa 1970s

1980s

1990s

2000s

8.59 (15)

9.49 (15) 7.55 (3) 1.36 (18)

8.61 (20) 4.62 (19) 1.98 (25)

7.8 (15) 4.55 (21) 1.57 (23)

1.42 (18)

Note: a. The average value for the decade is reported in each cell. The number of countries reporting is shown in parentheses. Source: Computed from IMF (various years).

at a reasonable tax rate. An ‘adequate’ level of revenues will depend on the expenditure responsibilities that have been assigned to the subnational government. If the government has adopted fiscal decentralization as a development strategy, and has assigned important functions to subnational governments, a subnational government tax base that is cobbled together from the ‘minor’ taxes – property, entertainment, restaurants, signboards, and so forth – will not be adequate to cover the service delivery costs. Neither will reliance on politically and administratively difficult taxes (e.g., agricultural income tax, property tax) allow provincial or urban local governments to move measurably closer to becoming selfsufficient in terms of financing. In the case of many industrialized countries with a tradition of decentralization, broad-based taxes have been assigned to the subnational governments. For example, subnational governments are empowered to tax income in the Nordic countries, Spain, and Switzerland, and to impose general sales and income taxes in the United States and Canada. But there are some industrialized countries where the subnational governments do not levy broad-based taxes. These include Australia, the Netherlands, and Japan. In the case of developing countries, it is much more common to find that the power to levy broad-based taxes is reserved for the central government. There are a few exceptions such as the state-level value-added taxes in Brazil and India, and the general sales tax in Argentina, but mostly the broadbased taxes are denied the subnational governments. A striking finding reported in Table 9.2 is that relative to GDP, subnational government taxes are five times greater in OECD than in developing countries.13 However, as Stegarescu (2005) points out, the IMF data overstate the actual taxing autonomy of the subnational governments in OECD countries. Second, and contrary to the above, it is frequently argued that subnational government tax bases should be stable rather than cyclical,

Financing subnational governments with decentralized taxes 341 because subnational governments often provide essential public services. Moreover, subnational governments (usually) cannot budget for deficits, so they do not have a borrowing route to smooth out erratic revenue flows. By this reasoning, subnational government taxes on minerals or agricultural products that sell in world markets, for example, are not a good policy idea.14 In terms of the practice, some industrial countries do allow subnational governments to tax national resources but in developing and transition countries this is rare. The same caution might be raised about taxes on profits. Income and consumption tax revenues are also subject to economic downturns, but the general view is that this can be dealt with by tax rate increases or expenditure retrenchment. The property tax, the mainstay of local government finances, is probably the least sensitive to economic downturns.15 There is a pecking order among subnational governments with respect to the need for a broad-based and elastic tax structure. Certainly the greatest need would be for the larger state and provincial governments in federations, and in some unitary countries, because they have the greatest revenue-raising potential, and because they are often assigned more important expenditure responsibilities. The revenue base of the large urban governments also should include some access to broad-based taxes, in part because their service delivery needs are usually greater, and in part because it is often a goal of public policy to make these local governments more self-financing than others. Rural local governments can develop their fiscal systems with the property tax and with user charges, fees, and intergovernmental transfers. In most countries, there is not the expectation that rural local governments will become financially self-sufficient. Administration A good subnational government tax will allow collection of a target level of revenues at a reasonable administrative cost. This is an appealing objective, but not one that is easily translated into policy advice. The question of a ‘reasonable’ administrative cost for subnational governments is not easily pinned down. Even the concept of collection cost is not generally agreed. The common practice is to measure administrative cost against collections, but the thinking here is flawed. Collection costs can be low by such measures precisely because of a poor assessment or collection effort. If the goal is to choose a tax based in part on its collection cost, the better measure would be the cost of collecting some normative, target amount. Even if one could find an answer to the question, ‘What is the least cost method of raising $x in revenue?’ this may not be the best guideline for policy because there are so many other important considerations. For example, the property tax is notoriously costly to administer well, but almost everyone

342

The Elgar guide to tax systems

advocates it as a local government tax. It might be argued that the higher cost of local government property taxation (versus some other tax) might be justified by the potential accountability gains that come with it. Or, a local government excise tax on a cigarette factory might be administered at a low cost per dollar of collections, but because of tax exporting it might impose welfare costs that offset these administrative savings. In the industrial countries, administration is not usually a binding constraint in the tax assignment decision. The state/province-level tax administration is usually efficient, and where it is not, the central government serves as the collection agent. The industrialized countries are able to take advantage of the formality of their economies to mix and match tax administration styles to find an administrative regime that works.16 By contrast, administration is usually the most binding constraint on the tax assignment decision in developing countries. For some of the broad-based taxes that are levied on businesses (e.g., income tax, VAT), provincial or local government assignment is possible only for the largest provinces/states because of the book audit required in the tax administration. Most subnational governments in developing countries tend not to have the personnel or the budget to carry out such tasks. Moreover, taxes on business income or value-added are levied on a company basis rather than a plant basis and the tax administration cooperation required among subnational governments usually is beyond the present level of administrative capacity. There are several ways for subnational governments in developing countries to overcome the tax administration constraint: 1.

2.

Subnational governments might be restricted to choose only those tax bases that are easily administered. Provinces and some of the larger cities might concentrate on individual income taxes, certain selective sales taxes, operating licenses for large businesses, and taxes on the use of motor vehicles. To date, this has not been common practice. Smaller local governments would concentrate on formal and presumptive levies on shops, motor vehicles, factories, registered professionals, and real property. The practice is more or less in step with this advice. Administrative costs may be kept in check by limiting the number of taxes included in the system. In a sometimes desperate search for revenues, subnational governments will pile on many specific local levies, for example, entertainment and advertising taxes. These tend to be costly to collect and sometimes arbitrary in their administration. Moreover, they siphon off administrative resources from the more (potentially) revenue-productive levies.

Financing subnational governments with decentralized taxes 343 3.

4.

5.

Subnational governments might find ways to compensate for what otherwise would be an expensive tax administration cost. An example is shortcuts to administering the property tax, which might include substituting tax mapping based on site visits for more detailed and expensive aerial mapping, or using area-based assessments instead of more sophisticated parcel-by-parcel approaches based on comparative values. Subnational governments could piggyback on the tax base of the higher-level governments thereby holding rate setting powers while avoiding any tax administration responsibility. In practice, this is rarely done. Another way to lessen the administrative problem is to invest more in the administrative infrastructure of subnational governments. Too often central governments back away from significant revenue assignment to subnational governments on grounds that the administration is too weak to absorb significant taxes. To make the prophesy come true, they hold off on financing administrative improvements such as advanced data processing systems, appropriate training programs for subnational government officers, and the passage of tougher enforcement laws to support collection efforts.

Sometimes high administrative costs do not deter the assignment of a tax to subnational governments. For example, the administration of the property tax – particularly the job of valuing parcels – is notoriously hard to do and expensive if done right. The taxation of small business is also costly because it is often not based on self-reporting systems and requires considerable field work in the assessment and collection process.17 Yet both types of tax are levied by local governments in many developing countries. Correspondence between benefits and burdens If a subnational government is able to export the burden of its taxes to (non-benefitting) residents in other jurisdictions, then its residents will vote to spend more than they would have if there had been no tax exporting. While this situation is the dream of local politicians, it leads to an efficiency loss that is usually not in the public interest. The correspondence principle would limit subnational government tax choices to those taxes whose burden cannot be exported (to nonbeneficaries). Those who receive the benefits of expenditures made by a subnational government should bear the burden of the taxes levied by that subnational government. This version of the traditional ‘no-exporting’ rule for subnational government taxation does provide for the taxation

344

The Elgar guide to tax systems

of mobile factors in the case of non-resident beneficiaries. A non-resident who commutes into the province to work could be charged (a portion of) the provincial wage tax, or non-residents who shop in city markets could pay the local consumption tax, in both cases on grounds that they use city services. For example, city and county government retail sales taxes are widely used in the US. The correspondence principle would seem to rule out the assignment to subnational governments of origin-based consumption taxes and corporate income taxes. But, there is probably some exporting of burdens for nearly all taxes. For example, a significant portion of the property tax – that levied on firms who sell in national markets or are owned by non-residents – is probably exported. This is a small error, by comparison, say, with allowing a local government to levy an excise tax on a soft drink bottler who sells in a national market. The acceptability of a tax whose burden is exported will depend on degree. Vertical equity Subnational governments are probably ill-advised to design their tax system around vertical equity considerations. There are a number of reasons for this advice. First, income redistribution is a component of national social policy, with targets set by the central government. If the tax system is to be progressive, let this be part of a central design. Second, subnational government pursuit of redistribution through the tax system is likely to be thwarted by migration and tax compliance problems. Third, local government taxes are usually blunt instruments and do not do a good job with income redistribution. Fourth, the expenditure side of the budget might be the better route for subnational governments to deal with the quality of life for low-income families. On this matter, there is a difference between developing and industrial countries. In this former case, the approach to recognizing vertical equity in subnational tax structures is more often ad hoc than it is based on a careful estimate of the impact on income distribution. It usually takes the form of adopting some reform that is thought to minimize payment burdens on those with income below the poverty line. Examples would be a decision to exempt the lowest-valued premises from the property tax, to not impose taxes or charges that would deny the very poor access to basic services, and to consider cross-subsidization in setting user charge rates so as to protect those families with incomes below the poverty level. Industrial countries often do more fine-tuning to achieve distribution effects. In the US, for example, most state governments exempt food from their retail sales tax, largely out of a concern for vertical equity. Fletcher and Murray (2008) report that in 2005, 34 US states used progressive

Financing subnational governments with decentralized taxes 345 personal income tax rate structures. Spain is an interesting example in this regard. Regional governments in the autonomous regions may adjust their individual income tax rate schedule, but with the proviso that the rate schedule must be progressive and have the same number of brackets as the central government income tax (Lopez-Laborda et al., 2007). The dangers of subnational government taxation The rules for good tax choices seem clear enough, but politics and selfinterest often get in the way of good economics. Central governments would like to ensure that it is not possible for a province to export the tax burden to non-residents and that the provincial or local tax base is visible enough to ensure accountability. Subnational government officials are likely to see things quite differently, that is, tax exporting would be politically popular and a lack of transparency in the tax system could provide political cover. Depending on how the bargaining over tax assignment turns out, the danger is a welfare cost of resource misallocation. A second danger is that the overall success of a fiscal decentralization program could be compromised if the tax assignment decision is badly made. Unless subnational governments have some significant degree of freedom to alter the level and composition of their revenues, full accountability cannot be achieved. At a minimum, some degree of rate flexibility with respect to a significant component of local revenues is essential if the budget is to be responsive to local needs and if local leaders are to be accountable to their citizens. Some argue that a potential danger in permitting provincial and local governments even limited freedom to tax is that they will not utilize fully all the revenue sources that are open to them. This could allow the level and quality of public services in their region to deteriorate – a version of the infamous ‘race to the bottom.’ By another reckoning, perhaps this is not so great a danger. If expenditure assignments are properly organized, this should not be a problem (Bird and Slack, 2008). If the service in question is one of national importance or one in which there is a strong national interest in maintaining standards, it should presumably be funded and monitored by the central government. If it is not a matter of national interest, why should the central government be concerned if one region chooses more or less government than another? A decentralized tax system will feature fiscal disparities because subnational governments will have very different capacities to raise revenues. Some will choose higher tax rates and some will choose lower tax rates. Central governments bent on keeping fiscal outcomes uniform should not embrace fiscal decentralization. Decentralists, on the other hand, will argue that provincial and local government taxes are best viewed as

346

The Elgar guide to tax systems

benefit charges for locally provided services and they should vary from place to place depending on voter preferences. An equalizing grant system can go some way toward reducing the fiscal disparities that will arise with tax decentralization, but significant gaps in public service levels inevitably remain. The danger is the political unrest that can come with nonuniformity and fiscal disparities. Finally, there are issues of political control and sovereignty. In some industrialized countries, where traditions of decentralized governance are strong, tax decentralization is seen as an entitlement by subnational governments. In the US, for example, there is formidable state-level resistance to a national sales tax on grounds that it would encroach on state sovereignty. In developing countries, central ministries of finance guard against decentralizing taxing powers on grounds that this could compromise macroeconomic policy. Central governments in low-income countries seem more willing to dedicate a share of national taxes for financing decentralized services than to decentralize rate-setting powers.

3

TAX CHOICES FOR DECENTRALIZATION

With all of these considerations taken into account, do we conclude that there are tax choices available to subnational governments that will generate adequate revenue? Is this set of choices more limited for developing and transition versus industrial countries? Do the choices that have been made conform with the principles for good revenue assignment? The answers given here to these three questions are ‘yes’, ‘yes, in the short run’ and ‘not always’. Individual Income Tax The individual income tax can meet most of the tests for a good subnational government tax. It can generate significant revenue from an elastic tax base, and it is roughly consistent with the correspondence principle in that the burden falls mostly on those who benefit from the services provided. Correspondence problems do arise with respect to those who cross provincial borders to reach their place of work, and there is an incentive for jobs and for those who receive capital income to migrate to lower-taxing jurisdictions. The former problem might be resolved with a residence-based income tax, where non-residents file returns and pay an amount that would serve as a benefit charge for local services received (McLure, 1998). Income taxes work better if levied by jurisdictions that cover most of

Financing subnational governments with decentralized taxes 347 the commuting range. For example, states/provinces work better than metropolitan areas, metropolitan areas work better than cities, and so on. In the industrial countries, subnational government income tax administration has been shown to be feasible. The PAYE portion can be assessed and collected at the place of work with relatively little difficulty. The ‘hard to tax’ informal sector should be no more difficult a task for provincial/state governments than it is for the central government, and the collection rate should be no lower under a decentralized than under a centralized tax system.18 Income taxes on capital sources (dividends, interest, rents) can be easily collected by states and provinces on a residence basis if there are information-sharing agreements with the federal government. Moreover, administrative difficulties can be circumvented by a piggyback arrangement where the central government defines the base and takes responsibility for collection while allowing local governments to set sur-rates. State and local government income taxes are levied in many industrialized countries. Slack (2006, pp. 106–7) reports that income taxes represent the most important source of subnational government tax revenues in 13 of the 27 OECD countries. The approach to subnational government income taxation, however, varies among countries. The US and Canadian model is for a state government tax, with state discretion to define the tax rate and the tax base. The Swiss model is similar to that of the United States. Cantons levy an individual income tax and also permit local governments (communes) to levy surcharges at locally-established rates on the cantonal income taxes (Dafflon, 2008). Like some US state government income taxes, the Swiss subnational government income taxes are not fully harmonized with the central income tax. The case of Spain is different. The regions are divided into an autonomous group and a charter group. The former may levy a personal income tax and have some discretion over the tax rate. But, the tax base is common with the central government and collection is centralized. In the case of the charter regions, the subnational governments have full control over income tax policy and administration, much as is the case for US states (Lopez-Laborda et al., 2007). Under the Nordic model (Lotz, 2006), the central government collects the tax, but local governments have discretion to set the tax rate. In Norway, however, local governments may not exceed a centrally-set maximum rate and most local governments are at the ceiling. These local income taxes are levied at a flat rate on the same tax base as the national income tax. Provincial personal income taxes in Canada are also collected by the federal government. The German arrangement is unique. The Länder are responsible for collection, but have no authority to set the tax

348

The Elgar guide to tax systems

rate or determine the tax base. In effect, there are no subnational income taxes in Germany. Subnational government income taxes are rarely found in developing countries. Very few subnational governments have the ability to maintain a tax roll, or to do the necessary audit to police the compliance. Even under a piggyback arrangement – which gets around these administrative constraints – the tax base would be concentrated in a very few local governments and most would be shut out. There are two other forces pushing toward centralization of the individual income tax. One is that central governments may rely heavily on this source of revenue, and even central governments often have trouble collecting much from the personal income tax (Bird and Zolt, 2005). The second reason is that income taxes often carry income distribution goals and these are perceived to be the exclusive responsibility of the central government. In a number of transition economies, subnational governments have been assigned significant shares of income tax revenue. At present in Russia, for example, 100 percent of individual income tax revenues are shared with subnational governments. A 60 percent share is allocated to provincial governments in China. Dillinger’s (2007) review of the practice in eight Eastern European countries indicates that the individual income tax-sharing rate varies from a 94 percent allocation to local governments in Slovakia to 30 percent in the Czech Republic and Poland. In none of these countries, however, do local governments have any significant freedom in establishing the tax rate. Payroll Taxes A closely-related question is whether payroll taxes are a viable source of financing for subnational governments. Payroll taxes are levied on both employees and employers. Such taxes have several merits. They are easily administrable and revenue-productive at relatively low nominal rates. The problems are that payroll taxes act as a tax barrier to employment in the modern sector, they encourage firms to substitute capital for labor, and they penalize formal sector versus informal sector employment.19 Moreover, in many countries the payroll tax base is already heavily exploited to finance (central) Social Security and unemployment compensation systems. Payroll taxes are not frequently used to finance general services in industrial countries. But the practice is not unheard of, for example, state governments in Australia levy a tax on employer payrolls. Only rarely are subnational governments in developing countries given direct access to the payroll tax base. When they are, it takes the form of a wage tax that is withheld by employers. One of the few examples is Mexico

Financing subnational governments with decentralized taxes 349 where state governments and the federal district have the power to levy and administer a payroll tax. Municipal governments in South Africa were allowed to levy against payrolls and turnover, but this will be abolished in 2011. As in Mexico, the payroll tax is levied on an origin basis, and it accounts for only half as much revenue as the turnover component (Bahl and Solomon, 2003). Company Income Tax Most who study revenue assignment would argue that the company income tax is not a good choice for subnational governments (McLure, 1998; Martinez-Vazquez, 2008; Bahl and Bird, 2008a). To the extent that the incidence of the tax is shifted backwards to owners of the firms, or forward to consumers of the product, it is not fully borne by local residents. Moreover, if levied according to the point of payment, there arises the headquarters problem – the tax revenue will accrue to the headquarters city or province. Countries usually try to trick the system out of this undesirable tax exporting and headquarters effect by devising a formula to give every province where the company produces or sells its product a claim on some part of the tax base. But these allocation formulae give rough justice at best and few would argue that they are a reasonable way to convert the company income tax to a destination-based levy. The United States has learned well the great problems that come with trying to pro-rate the net income of national companies across state boundaries (McLure, 1980).20 Other problems with a subnational government enterprise income tax are no less worrisome: the tax base (profits) is cyclically unstable, and provincial and local government revenues can be significantly affected by changes in central government tax or industrial policy. In many countries, there is also the question of whether the subnational government tax administration machinery is up to handling something so complicated as the corporate income tax. And, finally, there is the question of exporting the burden. These problems notwithstanding, some industrial countries do assign the corporate income tax to subnational governments. It accounts for about 4 percent of state and local government tax revenues in the United States. In Switzerland, Spain, Japan, and Canada, subnational governments tax business income. Developing countries make little if any use of company income taxes at the subnational government level. Subnational governments in transition countries, however, do share in enterprise income taxes. In China the central government determines the tax rate and the tax base, but in

350

The Elgar guide to tax systems

Russia the subnational governments have some rate-setting discretion on a federally-determined base. Consumption Taxes21 Consumption taxes can be revenue-productive at low nominal rates, and oftentimes are less objectionable to taxpayers than income or property taxes. These may take the form of either general sales taxes or selective sales taxes. With the notable exceptions of the United States and Canada, general sales taxes are not often used by subnational governments in industrial countries. The more decentralized European countries have settled on assigning general consumption taxes to the center and allowing the subnational governments to share the individual income tax base. General sales tax (VAT) The general sales tax that dominates the revenue structure in most countries is a VAT. There would seem little reason to deny provincial-level governments in industrial countries the authority to levy a VAT. Nevertheless, the only well-functioning, destination-base subnational VATs now in existence are those in Canada (Bird et al., 2006; Bird and Gendron, 2007, 2009). Two different systems are in operation: one in Quebec and the other in three small provinces. Canadian experience shows that with good tax administration it is perfectly feasible to operate a VAT at the subnational level on a destination basis, with each taxing government having the power to determine its own VAT rate. In most developing countries, there is no realistic prospect that the tax administration will be able to support a subnational government VAT.22 Subnational VATs were considered to be either infeasible or undesirable for a variety of reasons: high administrative and compliance costs, the possible loss of macroeconomic control, the general reluctance of central governments to share VAT space, the tax treatment of international trade, and the problems arising from cross-border (interstate) trade.23 Brazil has long levied a state VAT, and it generates an impressive 8 percent of GDP. The state governments have considerable discretion to adjust the rate and base, with the result that the VAT has been used by states as an instrument of industrial policy (Alfonso and Barroso, 2007). India has implemented a state-level value-added tax, but is still working out the details of how it will operate, particularly with respect to interstate trade. Retail sales and gross receipts taxes The retail sales tax is a major source of revenue in most US states. The tax rates and bases are determined by the individual states. The biggest

Financing subnational governments with decentralized taxes 351 difference among the states is in the extent to which they tax services and purchases of inputs by businesses. There is no national sales tax in the US. The major problem with the retail sales tax as a subnational government levy are the difficulties with taxing services and Internet Purchases. With the shift in spending patterns toward the consumption of services, and with Internet ordering on the rise, there is an erosion of the state revenue base. The taxation of electronic purchases is held back by administrative constraints and by legal rulings (the need to show nexus for a company as a prerequisite for taxation).24 The taxation of services is limited by administrative considerations and by a historical tradition of not taxing services. Subnational governments in some developing countries do make use of consumption taxes. Buenos Aires, both city and province, levy a gross receipts tax. The tax rate varies widely by type of product. The major own source revenue of Brazilian municipalities comes from a gross receipts tax on services (ISS), almost all of which is collected by the largest municipalities (Rezende and Garson, 2006). In Nicaragua, the local gross receipts tax is levied at a rate of 1 percent. Colombian municipalities also derive much of their revenue from a gross receipts tax. South African cities have in the past derived a significant amount of revenue from a combination payroll and turnover tax (Bahl and Solomon, 2003). All of these taxes are levied on an origin basis, so all are guilty of tax exporting. Moreover, there is the ‘headquarters problem’, which arises because national firms tend to pay tax for all branches at the headquarters location. Some urban local governments in South Asia have raised significant revenue from the Octroi, which is a combination of sales tax and terminal tax. It is levied on goods entering the local government area, according to a complicated rate schedule, and is collected at Octroi stations. It imposes compliance costs on sellers (the waiting times at the Octroi posts can be long) and is collected in a way that invites corruption. Nevertheless, the Octroi was revenue-productive, for example, 50 percent of revenues for the Bombay Municipal Corporation. The Octroi has now been abolished in Pakistan and in all but three Indian states. Provincial governments in Pakistan are empowered to levy a sales tax on the consumption of services, but this yields little revenue. Excise taxes Excise taxes can be easily administered by regional governments and lend themselves to regionally-differentiated rate determination. Moreover, in terms of efficiency, subnational government excise taxes, if applied on a destination basis, should meet the correspondence test for subnational assignment. It is sometimes argued that there is a benefit argument for subnational government excises – for example, on motor fuels – to the

352

The Elgar guide to tax systems

extent that regional governments are responsible for road maintenance. In practice, subnational government excise taxes (selective sales taxes) work well in industrial countries and are widely used. The case for decentralizing excises in developing countries is not so strong, for two reasons. First, special excises on petroleum, liquor, beer, and tobacco are of significant revenue importance to central governments, and are not likely to be surrendered to decentralization. Second, administrative constraints limit the degree to which a destination-based excise could be implemented in most developing countries.25 Moreover, local governments themselves are hesitant to take on the political cost that might come with heavier taxation of motor vehicle use. The strongest economic and administrative case for regional (and perhaps even local) excises is with respect to vehicle-related taxes (Bahl and Linn, 1992; Bahl and Bird, 2008b). The most important tax on automobiles from a revenue perspective is the fuel tax, which is also the simplest and cheapest form of automotive taxation from an administrative perspective. It is imposed with success by US state governments. The case for provincial-level motor fuel taxes in developing countries is less easily made, and they are imposed in only a few countries. Collection at the pump is usually difficult. Differential provincial fuel taxes can be imposed at the refinery or wholesale level, with the refiner or wholesaler acting as a collection agent for the states, remitting taxes in accordance with fuel shipments. Fuel taxes are related in a general way to both to road usage and to such external effects of vehicles as accidents, pollution, and congestion. However, to the extent that automotive taxation is intended to price either the utilization of publicly-provided services or externalities, fuel taxes are a very crude instrument. Toll roads and an appropriate set of annual automobile (and driver) license fees can in principle serve this benefit tax function much better. Property and Land Taxes Virtually all countries assign the property tax to local governments, and in the industrialized countries these local governments are given rate-setting powers. Administration of the tax is often divided between the central (or state) government and the local government, but there is no one dominant pattern on the division of administrative duties. In some cases higher-level governments develop the cadastre and even do the valuation work, while local governments focus on collections. In other countries, valuation is a local function, especially in the larger cities. Most countries assign responsibility for collections to the local level.

Financing subnational governments with decentralized taxes 353 Table 9.3

Property tax as a percentage of GDPa

OECD countries (number of countries) Developing countries (number of countries) Transition countries (number of countries) All countries (number of countries)

1970s

1980s

1990s

2000sb

1.24 16 0.42 20 0.34 1 0.77 37

1.31 18 0.36 27 0.59 4 0.73 49

1.44 16 0.42 23 0.54 20 0.75 59

2.12 18 0.60 29 0.68 18 1.04 65

Notes: a. The average value for the decade is reported in each cell. The number of countries reporting is shown in parentheses. b. The average for the 2000s is for the years 2000 and 2001. Sources: Bahl and Martinez-Vazquez (2008), calculations from IMF (various issues).

The property tax passes many of the tests of a good subnational government tax. The base is broad and the tax can be revenue-productive at reasonable levels of the statutory rate. Typically, revenues are relatively stable over the business cycle. There is a rough jurisdictional correspondence between the benefits received from services financed by the tax and the burden distribution.26 It fails the tests for a good subnational government tax in terms of its high administrative cost and its unpopularity with voters. There is always controversy about the revenue yield of the property tax, that is, about whether its burden is too high or about whether it raises enough revenue to finance local public services. The revenue yield from the property tax in OECD countries is equivalent to about 2 percent of GDP, more than three times higher than the yield in developing countries (Table 9.3). The discussion about property tax practices in low-income countries is mostly pointed toward its almost uniformly weak revenue performance. A number of hypotheses have been offered about why property tax revenues are so low in developing countries. Arguably the most important reason is that the property tax works best as a local government tax, and fiscal decentralization has not been as embraced in developing as in industrialized countries. Bahl and Martinez-Vazquez (2008) use data from a panel of 70 countries for 1990, 1995, and 2000 to show a significant positive effect of both expenditure decentralization and the level of per capita GDP on the level of the effective property tax rate.27 Lower-income countries are less decentralized and therefore use the property tax less.

354

The Elgar guide to tax systems

The property tax offers several advantages as a good local tax in developing countries. Real property is visible and can be reached by local government administrations, and with effort, effective administration is possible (DeCesare, 2004). The distribution of the tax burden is progressive, and the revenue potential is well above amounts now collected in most developing countries. A major problem is that delays in general revaluation are commonplace, and significantly lower the revenue income elasticity.

4

CONCLUSIONS

Subnational government tax revenues in OECD countries average about 8 percent of GDP, and account for around one-fourth of all taxes raised. The level in developing countries is less than one-half as much. When one factors in the actual control over local taxes, the decentralized shares are lower in both industrialized and developing countries. The variation across countries, both within the OECD and among less developed countries, is very large. Countries tend to choose more decentralized tax systems largely because they are more decentralized on the expenditure side of the budget. Finance seems to follow function. For example, Denmark and Canada are very decentralized on the tax side, but only 5 percent of taxes are raised by subnational governments in the UK, and only 2 percent in Ireland. Culture and history seem to play a major role, but federal structure does not seem to be a defining factor. While tax assignment has gone quite far in many industrial countries, there is no one approach that can be pointed to as most successful. In fact, tax decentralization is implemented in many ways. Some countries use very decentralized administrative regimes (Germany), some use central collection machinery (Nordic countries) and some rely heavily on cooperation between governments (United States and Canada). The United States gives almost complete discretion to the state governments to make tax policy, Spain has legislated an asymmetric system, and Australia prescribes a very limited role for its states in setting tax policy. One cannot make the same statement about the less developed and transition countries. The preferred model seems to be one of central determination of the tax rate and base for all major taxes and revenue-sharing programs to sort out the problems with the vertical gap. There are some notable exceptions (e.g., Brazil, Argentina, and India) but for the most part, subnational governments in low-income and transition countries have little independent taxing power. The choice of decentralized taxes is much more limited in developing

Financing subnational governments with decentralized taxes 355 than in industrialized countries. The individual income tax, which theory tells us is a preferred subnational government levy, is used in more than half of the OECD countries. By contrast, it is not a decentralized tax in the developing economies or the transition countries, because of administrative considerations or because central governments are unwilling to give up control. Neither industrial nor developing countries have made much use of payroll taxes for subnational government financing, though Australia, France, and South Africa have in the past been notable exceptions. The most egregious violation of the basic principles for good subnational government taxation would appear to be the decentralization of the corporation income tax in several industrial countries. The use of indirect taxes by subnational governments is more of a mixed bag. Provincial and state governments in Canada and the US do levy general sales taxes but European countries have not assigned general sales taxes to subnational governments. There is little use of decentralized general sales taxes in developing countries. Neither can they make use of destination-based excises, because of administrative constraints. Whither tax decentralization? Tax effort in the OECD countries does not appear to be rising, but there appears to be a drift toward more reliance on subnational government taxes. However, there is reason to speculate that this trend might be stuck. The value-added tax is not a likely candidate for increased autonomous taxation by regional governments, in Europe or in the US; the corporation income tax is a bad choice, and there is only so much revenue to be had from destination-based excises and from property taxes. OECD countries may not be able to continue working the individual income tax so hard. On the side of developing countries and transition countries, a groundswell of support for tax decentralization has not yet materialized. Rather, there is continued heavy reliance on intergovernmental transfers to finance locally provided services, and therefore some compromise to the accountability requirement for successful fiscal decentralization.

NOTES * 1. 2. 3.

In writing this chapter, I benefitted from discussion with Richard Bird and from a reading of his paper ‘Tax Assignment Revisited’ (2008). I acknowledge the valuable research assistance of Yared Seid in the preparation of this chapter. The dual subordination issue in tax administration is discussed for the Russia case in Martinez-Vazquez et al. (2008), and for China in Bahl (1999). For a good discussion of problems with the measurement of revenue decentralization, see Ebel and Yilmaz (2003). The World Bank has recognized the need for differentiating revenue assignment from revenue autonomy and is developing a set of Qualitative Decentralization

356

4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

The Elgar guide to tax systems Indicators, available at: http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/ EXTPUBLICSECTORANDGOVERNANCE/EXTDSRE/0,,contentMDK:20 259516~isCURL:Y~menuPK:2086395~pagePK:210058~piPK:210062~theSit ePK:390243,00.html accessed 30 April 2010. So far, these data are only available for 31 countries. The OECD (1999) also is developing a set of indicators of tax decentralization which takes account of the control over various taxes by subnational governments. However, these data are only available for OECD countries. Paradoxically, most discussions of fiscal federalism are more suited to the situation that might exist in a unitary country. In fact, most studies that give theoretical guidance are relatively vague about how taxes would be divided between the state (provincial) and local government levels. For a good discussion of Musgrave’s lasting contributions to the tax assignment question, see Bird (2008). See Bahl and Bird (2008). Though the IMF data are used here, the shortcomings, as discussed above, should be kept in mind. They may have the same view about higher central government taxes, but feel that central government decisions are too far removed from them to have an influence. For a good discussion of the tax limitation movement in the US, see Mullins (2010). For good discussions of comparative advantages of local and central governments in tax administration, see Mikesell (2007), and Martinez-Vazquez and Timofeev (2004). For a discussion of VAT practice, see Bird and Gendron (2007). In the 1990s, local governments in China dealt with revenue shortfalls by imposing a number of ad hoc, extra budgetary levies. The central government took action to abolish these levies (Bahl, 1999). For the sample of countries reported in Table 9.2, which varies from period to period, there is no evidence that the revenue buoyancy with respect to GDP has exceeded unity. Musgrave (1983, p. 11) also argued that in principle, ‘tax bases which are distributed highly unequally among sub-jurisdictions should be used centrally’. At least this was the conventional wisdom before the dramatic decline in US housing prices in 2008. For a good discussion of the factors underlying the choice between centralized and decentralized administrative regimes, see Martinez-Vazquez and Timofeev (2004). For a good discussion of the complications that arise from presumptive taxes on small businesses, see Engelschalk (2004). However, if there is a dual administration of the tax, compliance costs may rise and drive down collection rates. For a discussion of the advantages and disadvantages of a payroll tax, see Alm and Wallace (2007). Similar allocation problems are observed in Switzerland (Spahn, 1997). This section draws from Bahl and Bird (2008b). See Bahl and Bird (2008b). Broadly, the argument with respect to such trade was that subnational VATs were, if levied on an origin basis, distortionary, and if levied on a destination basis, unworkable. For a discussion of technology and the revenue capture of remote sales, see Fox et al. (2009). For a good discussion of the excise tax practice in developing countries, see Cnossen (2006). The exception to this rule is the burden of taxes on some non-residential properties that are not owned by local residents, and that sell products outside the local area. The effective rate of property tax is measured as the ratio of property tax collections to GDP.

Financing subnational governments with decentralized taxes 357

REFERENCES Alfonso, Jose Roberto and Rafael Barroso (2007), ‘Brazilian Tax Affairs’, Berkeley Program in Law and Economics, Paper No. 043007-8, University of California. Alm, James and Sally Wallace (2007), ‘Are Jamaica’s Direct Taxes on Labor “Fair”?’, Public Finance Review, 35(1), 83–102. Bahl, Roy (1999), Fiscal Policy in China, San Francisco: 1990 Institute and University of Michigan Press. Bahl, Roy (2008), ‘Opportunities and Risks of Fiscal Decentralization: A Developing Country Perspective’, in Gregory Ingram (ed.) Achieving Decentralization Objectives, Cambridge, MA: Lincoln Institute of Land Policy, pp. 19–37. Bahl, Roy and Richard Bird (2008a), ‘Subnational Taxes in Developing Countries: The Way Forward’, Public Budgeting & Finance, 28(4), pp. 1–25. Bahl, Roy and Richard Bird (2008b), ‘Tax Policy in Developing Countries: Looking Back – and Forward’, National Tax Journal, LXI(2), 279–302. Bahl, Roy and Musharraf Cyan (2009), ‘Fiscal Decentralization and Tax Assignment’, paper presented at a conference ‘The Political and Economic Consequences of Decentralization’, Santiago de Compostela, Spain, 5–6 November, 2009. Bahl, Roy and Johannes Linn (1983), ‘The Assignment of Local Government Revenues in Developing Countries’, in Charles E. McLure (ed.) Tax Assignment in Federal Countries, Canberra: Australian National University, Centre for Research on Federal Financial Relations. Bahl, Roy and Johannes Linn (1992), Urban Public Finance in Developing Countries, New York: Oxford University Press. Bahl, Roy and Jorge Martinez-Vazquez (2008), ‘The Property Tax in Developing Countries: Current Practice and Prospects’, in Roy Bahl, Jorge Martinez-Vazquez and Joan Youngman (eds) Making the Property Tax Work: Experiences in Developing and Transitional Countries, Cambridge, MA: Lincoln Institute of Land Policy. Bahl, Roy and David Solomon (2003), ‘The Regional Service Council Levy’, in Roy Bahl and Paul Smoke (eds) Restructuring Local Government Finance in Developing Countries: Lessons from South Africa, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 127–72. Bird, Richard (2006), ‘Local and Regional Revenues: Realities and Prospects’, in Richard Bird and Francois Vaillancourt (eds) Perspectives on Fiscal Federalism, WBI Learning Resources Series, Washington, DC: World Bank, pp. 177–96. Bird, Richard (2008), ‘Tax Assignment Revisited’, Working Paper No. 08-05, Atlanta: International Studies Program, Georgia State University. Bird, Richard and Pierre-Pascal Gendron (2007), The VAT in Developing and Transition Countries, New York: Cambridge University Press. Bird, Richard and Pierre-Pascal Gendron (2009), ‘Sales Taxes in Canada: The GST-HST-QSTRST System’, paper presented at American Tax Policy Institute Conference on ‘Structuring a Federal VAT: Design and Coordination Issues’, Washington, DC, 18–19 February 2009. Bird, Richard and Enid Slack (2008), ‘Property Tax and Rural Local Finance’, in Roy Bahl, Jorge Martinez-Vazquez and Joan Youngman (eds) Making the Property Tax Work: Experiences in Developing and Transitional Countries, Cambridge, MA: Institute of Land Policy, pp. 103–24. Bird, Richard and Sally Wallace (2004), ‘Is It Really So Hard to Tax the Hard-to-Tax? The Context and Role of Presumptive Taxes’, in James Alm, Jorge Martinez-Vazquez and Sally Wallace (eds) Taxing the Hard-to-Tax: Lessons from Theory and Practice, Amsterdam: Elsevier, pp. 121–58. Bird, Richard and Eric Zolt (2005), ‘Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries’, UCLA Law Review, 52(6), 1627–95. Bird, Richard, Jack Mintz and Thomas Wilson (2006), ‘Coordinating Federal and Provincial Sales Taxes: Lessons from the Canadian Experience’, National Tax Journal, 49(4), 889–903.

358

The Elgar guide to tax systems

Brennan, G. and J.M. Buchanan (1980), The Power to Tax: Analytical Foundations of a Fiscal Constitution, Cambridge: Cambridge University Press. Cnossen, Sjibren (ed.) (2006), Excise Tax Policy and Administration in Southern African Countries, Pretoria: University of South Africa. Dafflon, Bernard (2008), ‘Accommodating Asymmetry through Pragmatism: An Overview of Swiss Fiscal Federalism’, International Studies Program, Georgia State University. De Cesare, Claudia M. (2004), ‘General Characteristics of Property Tax Systems in Latin America’, paper presented at 7th International Conference on ‘Optimizing Property Tax Systems in Latin America’, Guadalajara, Jalisco, Mexico. Dillinger, William (2007), ‘Intergovernmental Fiscal Relations in the New EU Member States: Consolidating Reforms’, Working Paper No. 111, Washington, DC: World Bank. Ebel, Robert and Serdar Yilmaz (2003), ‘On the Measurement and Impact of Fiscal Decentralization’, in Jorge Martinez-Vazquez and James Alm (eds) Public Finance in Developing and Transitional Countries: Essays in Honor of Richard Bird, Cheltenham, UK and Northampton, USA: Edward Elgar, pp. 101–20. Engelschalk, Michael (2004), ‘Creating a Favorable Tax Environment for Small Business’, in James Alm, Jorge Martinez-Vazquez and Sally Wallace (eds), Taxing the Hard-to-tax, Amsterdam: Elsevier. Fox, William, LeAnn Luna and Matthew Murray (2009), ‘The SSTP and Technology: Implications for the Future of the Sales Tax’, National Tax Journal, LXI(4), Part 2, 823–42. Fletcher, Jason and Matthew Murray (2008), ‘What Factors Influence the Structure of the State Income Tax?’, Public Finance Review, 36(4), 475–96. Hettich, W. and S. Winer (1999), Democratic Choice and Taxation: A Theoretical and Empirical Analysis, New York: Cambridge University Press. Ikawa, Hiroshi (2008), ‘Trinity Reform of Local Fiscal System in Japan’, in Shinichi Ichimura and Roy Bahl (eds) Decentralization Policies in Asian Development, Singapore: World Scientific Press, Chapter II. International Monetary Fund (various issues), Government Finance Statistics, Washington, DC: International Monetary Fund. Lockwood, Ben (2006), ‘The Political Economy of Decentralization’, in Ehtisham Ahmad and Giorgio Brosio (eds) Handbook of Fiscal Federalism, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 33–60. Lopez-Laborda, Julio, Jorge Martinez-Vazquez and Carlos Monasterio (2007), ‘Kingdom of Spain’, in Anwar Shah (ed.) The Practice of Fiscal Federalism: Comparative Perspectives, Montreal: McGill-Queens University Press. Lotz, Jorgen (2006), ‘Local Government Organization and Finance: Nordic Countries’, in Anwar Shah (ed.) Local Governance in Industrial Countries, Washington, DC: World Bank, pp. 223–64. Martinez-Vazquez, Jorge (2008), ‘Revenue Assignments in the Practice of Fiscal Decentralization’, in Nuria Bosch and Jose Duran (eds) Fiscal Federalism and Political Decentralization: Lessons from Spain, Germany and Canada, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Martinez-Vazquez and Andrey Timofeev (2004), ‘Choosing between Centralized and Decentralized Models of Tax Administration’, in La Financiacion de las comunidades autonomas: Politicas tributarias y solidaridad interterritorial. Martinez-Vazquez, Jorge, Mark Rider and Sally Wallace (2008), Tax Reform in Russia, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. McLure, C.E., Jr. (1980), ‘The State Corporate Income Tax: Lambs in Wolves’ Clothing’, in H.J. Aaron and M.J. Boskin (eds) The Economics of Taxation, Washington, DC: The Brookings Institution, pp. 327–46. McLure, C.E. (1998), ‘The Revenue Assignment Problem: Ends, Means and Constraints’, Public Budgeting, Accounting and Financial Management, 9(4), 652–83. Mikesell, John (2007), ‘Developing Options for the Administration of Local Taxes: An International Review’, Public Budgeting & Finance, 27(1), 41–68. Mullins, Daniel R. (2010), ‘Fiscal Limitations on Local Choice: The Imposition and Effects

Financing subnational governments with decentralized taxes 359 of Local Government Tax and Expenditure Limitations’, in Sally Wallace (ed.) State and Local Fiscal Policy: Thinking Outside the Box?, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 201–65. Musgrave, Richard A. (1983), ‘Who Should Tax, Where and What?’, in Charles McLure (ed.), Tax Assignment in Federal Countries, Canberra: Centre for Research on Federal Financial Relations, Australian National University, pp. 2–19. Oates, Wallace E. (1996), ‘Taxation in a Federal System: The Tax-assignment Problem’, Public Economics Review, 1(1), 35–60. Oates, Wallace (2008), ‘On the Evolution of Fiscal Federalism: Theory and Institutions’, National Tax Journal, LXI(2), 313–34. OECD (1999), Taxing Powers of State and Local Government, OECD Tax Policy Studies No. 1, Paris. Plamondon, R. and D. Zussman (1998), ‘The Compliance Costs of Canada’s Major Tax Systems and the Impact of Single Administration’, Canadian Tax Journal, 46(4), 761–85. Rezende, F. and S. Garson (2006), ‘Financing Metropolitan Areas in Brazil’, Revista de economia contemporanea, 10(1), 5–34. Rodden, Jonathan (2003), ‘Soft Budget Constraints and German Federalism’, in Jonathan Rodden, Gunnar S. Eskeland and Jennie Ilene Litvack (eds) Fiscal Decentralization and the Challenge of Hard Budget Constraints, Cambridge, MA: MIT Press, Chapter 5. Slack, Enid (2006), ‘Alternative Approaches to Taxing Land and Property’, in R.M. Bird and F. Vaillancourt (eds) Perspectives on Fiscal Federalism, WBI Learning Resources Series, Washington, DC: World Bank, pp. 197–223. Spahn, Paul Bernd (1997), ‘Switzerland’, in Teresa Ter-Minassian (ed.) Fiscal Federalism in Theory and Practice, Washington, DC: International Monetary Fund. Spahn, Paul Bernd and Wolfgang Fottinger (1997), ‘Germany’, in Teresa Ter-Minassian (ed.) Fiscal Federalism in Theory and Practice, Washington, DC: International Monetary Fund. Stegarescu, Dan (2005), ‘Public Sector Decentralization: Measurement Concepts and Recent International Trends’, Fiscal Studies, 26(3), 301–33. Stegarescu, Dan (2009), ‘The Effects of Economic and Political Integration on Fiscal Decentralization: Evidence from OECD Countries’, Canadian Journal of Economics, 42(2), 694–718. Vehorn, Charles L. and Ehtisham Ahmad (1997), ‘Tax Administration’, in Teresa TerMinassian (ed.) Fiscal Federalism in Theory and Practice, Washington, DC: International Monetary Fund. World Bank ‘Qualitative Decentralization Indicators Data Base’, World Bank website.

10 The administration of tax systems John Hasseldine*

1

INTRODUCTION

Adam Smith established that there were four primary principles in relation to tax systems. The world in general, and that of taxation, has become far more complex since, yet tax administrations still need to pay heed to these principles. Naturally, in both tax reform and administration, this is easier said than done. There continue to be budget pressures on tax administrations with an increased focus on performance measurement and evaluation, efficiency audits and enforced cost savings, public performance reporting and more scrutiny to governance and oversight mechanisms (in terms of governance or new governance structures. Effectively, ‘new public management’, as outlined in Flynn (2007), has ratcheted up the level of accountability from tax administrations and their leaders. These essentially external pressures coupled with features of the environment for taxpayer (non)compliance create huge challenges for senior management. So, while senior management in tax agencies are no doubt well aware of these trends, in the scholarly tax literature, there is relatively little recognition of these forces by tax researchers, although there are some obvious exceptions especially with respect to developing countries (Bird and Casanegra de Jantscher, 1992; Bird and Zolt, 2008) and by those calling for further cross-disciplinary research (Alley and Bentley, 2008). The purpose of this chapter is to address the recent evolution of tax administration, evaluate the current situation and conjecture on future best practice. In the next section, Gill’s (2003) framework is used to classify possible areas within which best practice can be examined and then this is used as a tool to present relevant issues in tax administration and some exemplars. Section 3 then develops this internal focus into one incorporating external factors, particularly within a European context. Section 4 documents the role and efforts of international organizations and others in measuring and sharing best practice. Section 5 offers some concluding remarks.

360

The administration of tax systems 361

2

THE TASKS OF TAX ADMINISTRATION

Gill’s (2003) study provides a useful initial framework in addressing tax administration issues because it outlines the more generic internal management issues involved in running a tax administration alongside the more widely recognized operational tasks (see Table 10.1). The former are now briefly discussed in turn. Strategy and Policy Formulation Clearly, senior management must provide strategic direction and assert their influence over organization form. In the US, Rainey and Thompson (2006, 2007) discuss how Charles Rossotti pioneered change, for example, the shift to new operating divisions rather than the geography-based structure in place prior to the IRS Reform and Restructuring Act of 1998. Not everyone agreed with dissenting contributions from Mikesell and Birskyte (2007) and Cyr and Swanson (2007), pointing to the worsening US tax gap as evidence that Rossotti’s tenure was ‘not quite the triumph they describe’ (Cyr and Swanson, 2007, p. 576). In the UK, a 2007 Capability Review of HM Revenue & Customs (HMRC) concluded that the department was complex, both in terms of its many constituent parts and in terms of its matrix management structure that did not relate roles and responsibilities amongst its senior management to accountability.1 Consequently in 2008 in a shift from prior practice, a non-executive chairman was appointed – part of a trend of greater accountability in tax agencies. Policy formulation and strategy is best considered within an organizational framework. Authors such as Vehorn and Brondolo (1999) and McCarten (2005) stress the importance of clear organizational structure in operational tasks. However, Tomkins et al. (2001) stress that organization structure is also important in internal management issues. Planning, Budgeting, Resource Allocation Tax agencies face budget constraints and are often under-funded (Owens and Hamilton, 2004; Slemrod et al., 2010). Their position within the government budgeting framework that exists in each country can pose problems for long-term investment projects (e.g., modernizing organizational structure and information systems etc.) and factors into the selection of cases for audit (i.e., there is seen to be little point in investigating lossmaking firms). Another difficulty is that in terms of reducing taxpayer compliance costs, there is a trade-off in providing customer service. The more

362

The Elgar guide to tax systems

Table 10.1

Main tasks of revenue administration requiring analysis

Organization and Management Tasks

Operational Tasks

Strategy and policy formulation Planning, budgeting, resource allocation

Registration of taxpayers Taxpayer services: Taxpayer education Taxpayer assistance Facilitation of voluntary compliance Processing of declarations and payments Monitoring of tax withholders and collection agents Collection of information about taxable transactions: Collection of information from third parties Intelligence operations Search and seizure and survey operations to obtain incriminating evidence Risk analysis and selection of cases for audit and investigation Audit and investigation

Monitoring and evaluation Coordination Financial management

Personnel management Information technology management Asset management Internal control

Anti-corruption External relations

Recovery of tax arrears Legal and judicial matters: Legislation Appeals Prosecution Fiscal studies

Source: Adapted from Gill (2003, Table 1).

service provided, the less the costs to the taxpayer, but the higher the costs to the tax agency. These trade-offs may even have compliance implications. Monitoring and Evaluation Coupled with the push for a cooperative approach, there has been a shift to increased transparency in public performance reporting. Thus, HMRC in the UK publishes the results of its staff surveys as well as its progress on its key Public Service Agreement targets with the government, and Australia publishes its Compliance Program annually.

The administration of tax systems 363 Governance and accountability in tax agencies is obviously important and there is a trend for additional external oversight mechanisms. For example, in Australia there is the Office of the Inspector-General of Taxation, and in the US there is the Treasury Inspector General of Tax Administration, the IRS Oversight Board, as well as the office of the National Taxpayer Advocate. In the UK however, there appear to be no moves to extend the oversight provided by the Public Accounts Committee of Parliament.2 The progress of tax agencies is also monitored by supreme audit institutions such as the General Accountability Office in the US and respective institutions elsewhere. Coordination Tax agencies need to coordinate with each other (through regional and international alliances), with other government departments, with taxpayers and their representatives (e.g., industry bodies and professional accounting associations) and with politicians. Of course they also need to coordinate internally, and there is a clear case for the effective implementation of knowledge management systems in tax agencies, which hitherto has received scant attention (Hasseldine et al., 2009). A useful example of coordination is the CIAT Award for Innovation in Tax Administration, which seeks to identify and recognize innovative good practices that promote transparency and integrity, efficiency and effectiveness in tax administration, disseminate successful innovation and aid synergic exchanges of successful experiences. Management of Finance, Personnel, Information Technology and Assets In the UK media, there has been criticism of the lack of suitably qualified finance directors within government departments that has arisen due to the difficulties in recruiting public sector accountants. In terms of effective human resource management, there is a dearth of studies on the management of tax agency staff (one exception is Proctor and Currie, 2004). While the management of information technology projects can be a disaster area or ‘escalating commitment to a failing course of action’ (see Bird and Zolt, 2008) it can be done well. Countries with a record of successful implementations include Singapore (Tan and Pan, 2003), Australia and the Nordic approach of pre-populated tax returns – which can be considered best practice (Section 4). Ireland recognizes the importance of electronic service provision with e-filing and other services provided to customers, but it is also crucial to have technology available internally, for example, for risk management, data capture, management information and so on.

364

The Elgar guide to tax systems

Internal Control and Anti-corruption An in-depth review on the literature on corruption is outside the scope of this chapter and readers are referred to Martinez-Vazquez et al. (2007). Suffice to say, for some European countries, corruption is not a serious problem, however, there must be strong internal controls and tax administration must be seen as a professional career. External Relations In the author’s experience (based on casual observation from sitting on a government consultative committee), it seems that tax agencies come in for a lot of flak and while they are often criticized in the media and by politicians, they find it difficult to defend themselves in the public domain. The importance of being seen positively in public opinion and in providing high-quality customer service is thus crucial for a tax agency to function well, let alone for it to be able to collect taxes due. If the Commissioner of a tax agency has a high media profile, this can be useful in maintaining good public relations.3 One example of transparency and this increasing focus on external relations is the Canada Revenue Agency, which publishes executive summaries of its public opinion research on its website and full reports are available on request. Like other tax agencies, these surveys are used to gauge public opinion and measure service delivery.4

3

THE CONTEXT OF TAX ADMINISTRATION

In looking at the context of tax administration, it is instructive to highlight Bird (2004) who sketches the broad outlines of administrative reform, with a focus on the environmental context (or ‘cultural’ factors), viewing tax administration as a production process, the importance of political goodwill and an emphasis on honesty (for both taxpayers and tax office staff). Owens and Hamilton (2004, p. 348) note that while there is significant diversity in the OECD, there are also many similarities. They suggest that it is not so much the behaviour of the tax administration, rather what they have to administer viz: In looking at the root causes of problems in tax administration, what needs to be considered is what is being administered: the tax law and how it is interpreted. And problems caused by the law cannot be considered until one reflects on the efficacy and practicality of the tax policy that the law is meant to implement. The entire system, all of its players, their behaviours, and

The administration of tax systems 365 drivers of those behaviours need to be considered in an objective, holistic, and systemic manner if countries are going to tackle successfully their crises in tax administration.

Owens and Hamilton contribute to the debate through their useful observations that: ● ● ● ● ● ● ●

just simplifying the law does not work; policy simplification needs a stronger voice; the complexity of policy and law may need to be reduced; small business needs special consideration; new compliance approaches are needed; a new compact is needed; tax administrations are underfunded.

Europe In addressing the case of the UK, and tax reform/administration more generally, it is fortunate that the Institute of Fiscal Studies has recently published the Mirrlees Review (2010). In this report, the Slemrod et al. (2010) chapter on administration and compliance analyses in detail issues of tax system design, determinants of evasion, avoidance and administrative and compliance costs, enforcement systems, and the various taxes in the UK tax system. However, in the author’s opinion, the UK needs to spend more time in understanding the role of professional advisors5 and more work needs to be completed in the compliance area, including collection of arrears, recognition of a more behavioural approach to compliance, areas of risk management and so on. Sweden has a long-term objective for its administration (Skatteverket) that by 2012 it should be ‘the best tax administration within the OECD countries’.6 Such a definition is of course difficult to measure and the main purpose of the objective is that it should learn from others, not to make an exact rating.7 Skatteverket has started to work with this by looking at different areas and making comparisons with other countries. It chose areas where there are other long-term objectives or where there is information that is available. Thus it may be a more benchmarking best performance as opposed to a best practice exercise. Skatteverket has used different sources of information, such as the OECD Comparative Information Series and it has also studied annual reports from different tax administrations and other sources. Onrubia (2006) highlights the importance of examining tax administration in light of the underlying political system. He shows that until at least

366

The Elgar guide to tax systems

1977, Spain had an ‘obsolete tax system’ that prevented it from moving in parallel with developments in tax policy and administration in other OECD countries. Now, in the post-Franco era the Agencia Tributaria has, and continues to focus on the fight against tax fraud, the reduction of administrative and compliance costs, and improving taxpayer advisory services in an attempt to improve tax morale in the country (MartinezVazquez and Torgler, 2009). Increasing Emphasis on Compliance and Risk Management One notable trend is that many countries now emphasize cooperative compliance, such as through the use of a compliance model (Braithwaite, 2003). This can be seen as best practice as shown by the Australian Tax Office report on Managing Compliance Effectiveness (2007) and the Swedish Tax Agency’s Right from the Start research report (2005). Compliance models works better with a system of consultation and responsive regulation (Braithwaite, 2007). The benefits are better ‘buy-in’ and one consequence may be that taxpayers (and their representatives) may take government/tax administration efforts to lower taxpayer compliance costs at face value, rather than with cynicism. Tax agencies are thus paying more attention to compliance and consequently the voluminous literature is starting to be summarized in some useful books (e.g., Braithwaite, 2003; Kirchler, 2007; Torgler, 2007). Also tax agencies realize that their penalty regimes should be geared to changing poor compliance behaviour in order to improve future tax compliance (known as the behavioural approach to penalties). In managing large business compliance, there is growing concern at large companies’ ability to successfully avoid huge amounts of taxes (e.g., Desai, 2005). Extremely high-profile concerns are expressed by lobbyists who are agitating for greater corporate social responsibility in the tax area (e.g., in the Guardian newspaper and elsewhere). Tax agencies themselves have responded by calling on corporates to improve their tax governance and by focusing on the amount of tax at risk – particularly with very large companies, who might account for a relatively sizeable contribution to a country’s tax take. The response of these large taxpayers is not widely known although there are some qualitative studies that are beginning to explore companies’ views on tax planning and avoidance (e.g., Freedman et al. 2009; Toumi, 2009). The Big 4 professional accounting firms (PWC, Deloitte, Ernst & Young, KPMG) have responded to this change in the corporate tax landscape by offering their consultancy services in terms of risk management advice.

The administration of tax systems 367

4

MEASURING AND SHARING BEST PRACTICE

The concept of systematically measuring best practice over a range of countries and publishing the results for a general audience is relatively recent to tax administration. As this section documents, it is only in the last few years that surveys of experience have been conducted. Previously, researchers have tended to focus on the management of actual tax reforms in developing and western economies (Sandford, 1993; Alm et al., 2006) with this work revolving around case studies, infrastructure (e.g., banking practices), and a large amount of ‘lesson-drawing’ from other countries’ experiences. Role of International Organizations Much work has occurred through regional agencies and technical assistance provided by the IMF.8 Examples based upon this work include Ebrill et al. (2001) documenting countries’ experiences with VAT and van Kommer and Alink’s documentation of a collaboration between several Latin and North American countries and the Netherlands.9 Over the last decade, the OECD has been pivotal in knowledge sharing in tax administration, as has the launch of the International Tax Dialogue.10 In addition, the Forum on Tax Administration (FTA), a subgroup of the OECD’s Committee on Fiscal Affairs, allows tax administrators to share information and experience, and identify effective strategies/ measures for various areas of tax administration. The FTA has released a number of reports and the Comparative Information Series released in February 2007 is of direct relevance to this chapter. This report provides internationally comparative data on various aspects of tax systems and their administration in OECD and selected non-OECD countries and consists of the following:11 1. 2. 3. 4. 5. 6. 7.

international and organizational arrangements for tax administration; aspects of management approaches and practices; return filing, payment and assessment regimes for the major taxes; selected administrative powers of revenue bodies; tax revenue collections; operational performance information; administrative practice.

Other reports include topics such as: compliance risk management,12 strengthening tax audit capabilities,13 survey of trends in taxpayer service delivery using new technologies,14 using third party information reports to

368

The Elgar guide to tax systems

assist taxpayers meet their return filing obligations15 and recent studies on tax intermediaries16 and banks.17 In addition to the work of the major international organizations, several associations operate at a regional level. These include the InterAmerican Center of Tax Administrations18 and the Commonwealth Association of Tax Administrators19 who are active in training throughout the Commonwealth. There is also the Study Group on Asian Tax Administration and Research (known as SGATAR), and the IntraEuropean Organisation of Tax Administration.20 Consultancy Firms Some non-governmental work is taking place with many of the large consultancy firms providing input into the area, for example, KPMG’s Tax Business School has published a series of research papers on tax and corporate social responsibility, risk management and environmental taxes. The approach of PwC is quite different with its Paying Taxes 2009 report, joint with the World Bank, involving the gathering of reporting information on the tax affairs of a standard case study company (a ceramic flowerpot manufacturer/retailer) in 181 countries. The case study financials were flexed to reflect the relative wealth of the country within which it operates and tables were produced on (1) ease of paying taxes, (2) number of tax payments and (3) the time required to comply. While the report provides some interesting commentary, it is naturally limited as a benchmarking of best practice.21 Within Europe, EUROSAI (see National Audit Office, 2008) conducted a consulting exercise on benchmarking tax administrations. They note that comparing costs and performance is only the first step in a benchmarking process. The next step is identify the reasons for differences in performance – so that best practice can be shared. In addition, it is not a trivial task to specify an optimal set of key performance indicators (KPIs). The project by EUROSAI involved benchmarking data from Australia, Canada, Germany, the Netherlands, Spain, Sweden, UK and the US (see Table 10.2). While Table 10.3 uses data from the OECD Comparative Information Series, benchmarking efforts such as these would seem best suited to countries with similar tax administrations or when comparing specific contexts. In a different approach, Hasseldine (2008) conducted a non-random survey of tax agencies and academics, using Gill’s framework per Table 10.1 to identify various exemplars of best practice around the world. The purpose was to provide a possible pointer for tax agencies wishing to study the actions of others, rather than simply making use of statistical data. The results are reproduced in Table 10.3.

369 9

48

80

4.04

856

8.1

2.27

974

1.17

1.05

7

4.8

4.46

699

1.8

Germany

69

3

1.99

627

1.30

The Netherlands

23

7.8

4.13

1,557

0.82

Spain

15

2.7

0.27

829

0.59

Sweden

17

n/a

0.69

810

0.97

United Kingdom

Source:

47

6.2

1.16

2,974

0.56

United States

Van Stolk and Holmes (2007), ‘A Study on Reforms in International Tax Administrations’, Personal communication to author.

Note: a. Note that different countries have different methods of calculating administrative costs and that the mandates of some tax administrations are more extensive than others. This reduces the comparability of administrative costs indicators.

Aggregate administrative costs to net revenue collections in 2004a Citizens to full-time staff in 2004 Value of tax audit assessments to total net revenue collections (%) in 2004 Annual gross debt to total annual revenue collections (%) in 2004 Electronic filing take-up rates for personal income taxes (%) in 2004

Canada

Australia

Performance statistics for selected countries measured by the OECD

Performance Indicator

Table 10.2

370

The Elgar guide to tax systems

Table 10.3

Summary from 2007 survey on best practice

Area of Administration Operational areas Registration of taxpayers

Processing of customer information

Monitoring of withholders/ agents Information collection about taxable transactions – and audit/ investigation work

Risk analysis

Recovery of tax arrears and debt management

Legal and judicial matters

External relations and customer focus

References Developing Internet search tools – Netherlands (and UK), Germany, Hong Kong On the spot checks – Netherlands and others Business registration system – Australia Pre-filled tax returns – Denmark, Norway, Sweden E-filing – Australia and Singapore Business activity statements – Australia Cumulative PAYE – Ireland Blue Return Taxpayers system and bookkeeping Classification system – Japan Cash receipt system introduced by Korea Information returns program – USA Knowledge groups taxpayer segments – Netherlands Customer insight – New Zealand VAT: Automated invoice service – Chile Joint audits – Scandinavian countries/NL/UK ATO Compliance Risk model adopted by OECD Sweden: See Swedish Tax Agency (2005) Netherlands similar approach Risk management in construction sector – Ireland Canadian reform of Debt Collection function Debt Collection Reforms in the Netherlands (business-driven IT solutions, process improvements) Risk-based approach to debt collection in Norway Debt collection practices – Ireland Simplified tax regimes for small business – France, Austria, Australia Tax rulings program – Australia Channel strategies – Norway Horizontal supervision – Netherlands Visiting starting business – Netherlands, Canada Tax Education programs – Sweden, Chile, Japan Enforcement communication – Sweden, Australia Taxpayer Assistance Blueprint (phase 1, 2) – USA

The administration of tax systems 371 Table 10.3

(continued)

Area of Administration

References

Organization and management tasks Including: Strategy Planning Monitoring Personnel management IT systems Internal control Asset management

Growing importance of research on compliance and non-compliance of citizens and companies, that includes cultural and behaviour aspects – Australia, Canada, Netherlands Compliance costs – New Zealand XBRL/Standard business reporting – Netherlands CIAT Integrity project led by Canada Compliance Measurement Framework – Canada Tax gap research by US IT systems – Estonia, Chile, Singapore, Australia Modern phone telephony system – New Zealand

Source:

Hasseldine (2008).

Lack of Statistical Intra-country Benchmarking As discussed above, there has been an increase in the amount of benchmarking between countries, yet there are few publicly available studies that measure (and publish!) intra-country performance. Obviously, it is customary for public sector performance targets to be set, measured and published (as in the UK), but there are few studies examining productivity and efficiency between tax offices in a single country. The exceptions are three studies that use data envelopment analysis (DEA), being Moesen and Persoon (2002), Forsund et al. (2006) and Barros (2007). As newer econometric techniques such as stochastic frontier analysis can be considered superior to DEA, the time would seem ripe for further work into tax office efficiency. Of course, and this may explain the dearth of research, one difficult issue for would-be researchers is access to reliable input/ output data for analysis.

5

CONCLUDING REMARKS

Given the resources placed into tax administration around the world, it might be expected that tax administration would be regarded as a science, however I believe it is more of an art. Running a tax agency is like running a large (and in some cases very large) company. As such, the management

372

The Elgar guide to tax systems

literature is relevant although clearly public sector organizations are a different creature from for-profit organizations. Based on the literature outlined in the prior section, it is certainly possible to benchmark on key performance measures such as cost-effectiveness, success with e-services, telephone helplines and so on. This can be achieved using data from international organizations and in addition, regional cooperation and data collection can allow for further benchmarking (e.g. as is the case for the Nordic countries). Benchmarking can be conducted using quantitative performance metrics or using a qualitative approach if the activity being compared reflects the way work is done (e.g., planning, risk analysis, strategic management, etc.). In terms of performance measurement, caution is recommended in specifying the correct key performance indicators and the entire administrative system must be managed. Otherwise, whatever is measured gets done, and while more in-depth benchmarking and best practice measures seem highly likely in tax administration, and a desirable course of action, care needs to be taken. Hood (2006) analysed the targets approach to managing British public services including the various types of gaming response and found that the central managers of the target regime did not put substantial resources into checking performance data, took reported performance gains at face value and had no coherent anti-gaming strategy. There seems no a priori reason that this type of gaming response would not occur in a tax agency.22 In sum, considering the scale of tax administration worldwide, there is a relative small evidence base on best practice in tax administration and a dearth of scholarly literature. However, as outlined earlier, improvements can be expected due to an improved culture of knowledge sharing between tax agencies and other stakeholders. Although the following risks being an oversimplification, in conclusion, future best practice is likely to be associated with tax administrations that can demonstrate the following seven characteristics considered indicative of effective tax administration: 1. 2. 3. 4. 5. 6. 7.

a professional approach to internal management issues (HR; strategic planning); attention to cost-efficiency and effectiveness; responsive engagement with all stakeholders; successful introduction of technology applications; understanding what drives taxpayer and tax agent behaviour; sophisticated risk profiling and informed responses to taxpayer behaviour, including the areas of enforcement and service provision; transparency of governance and detailed performance reporting.

The administration of tax systems 373

NOTES * 1. 2. 3.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22.

I wish to thank Jorge Onrubia (Discussant), Jose Manuel Gonzalez-Paramo (Session Chair) and conference participants for their helpful comments on an earlier version of this chapter. Capability Review of HM Revenue & Customs, London: Cabinet Office, December 2007. K. Poynter (letter to Rt Hon. Alistair Darling MP), available at: www.hm-treasury.gov. uk/media/E/E/poynter_review171207.pdf; accessed 30 April 2011. For example, See HM Revenue and Customs: Management of Large Business Corporation Tax, report by the Comptroller and Auditor General, London: NAO, 25 July 2007. See for example, M. D’Ascenzo, ‘Sustaining Good Practice Tax Administration’, speech to the Australasian Tax Teachers’ Association Conference, Christchurch, New Zealand, January 2009, available at: www.ato.gov.au/corporate/content.asp?doc=/ content/00176063.htm&page=1; accessed 30 April 2011. See www.cra-arc.gc.ca/gncy/pr/menu-eng.html; accessed 30 April 2011. Although a recent consultative document goes part way in addressing this, see HM Government’s Consultative Document, Working with Tax Agents, available at: www. hmrc.gov.uk/budget2009/tax-agent-6440.pdf; accessed 30 April 2011. Personal communication, Lennart Wittberg to author, 14 September, 2007. ‘L. Wittberg’, Best Tax Agencies of the OECD Countries, Swedish Tax Agency (Skatteverket), personal communication to the author, 2007. See www.imf.org/external/pubs/ft/exrp/techass/techass.htm; accessed 30 April 2011. Van Kommer and Alink, Handbook for Tax Administrations, CD provided to the author. See www.itdweb.org; accessed 30 April 2011. OECD (2007), Tax Administration in OECD and Selected Non-OECD Countries: Comparative Information Series, Paris: OECD available at: www.oecd.org/dataoecd/37/56/38093382.pdf; accessed 1 May 2011. For example OECD (2004), Compliance Risk Management: Managing and Improving Tax Compliance, Paris: OECD, available at: www.oecd.org/dataoecd/44/19/33818656. pdf; accessed 1 May 2011. For example OECD (2006), Strengthening Tax Audit Capabilities: General Principles and Approaches, Paris: OECD, available at: www.oecd.org/dataoecd/46/18/37589900. pdf; accessed 1 May 2011. OECD (2005), Survey of Trends in Taxpayer Service Delivery Using New Technologies, Paris: OECD, available at: www.oecd.org/dataoecd/56/41/34904237.pdf; accessed 1 May 2011. OECD (2006), Using Third Party Information Reports to Assist Taxpayers Meet Their Return Filing Obligations: Country Experiences with the Use of Pre-populated Personal Tax Returns, Paris: OECD, available at: www.oecd.org/dataoecd/42/14/36280368.pdf; accessed 1 May 2011. OECD (2008), Study into the Role of Tax Intermediaries, Paris: OECD, available at: www.oecd.org/dataoecd/28/34/39882938.pdf; accessed 1 May 2011. OECD (2009), Building Transparent Tax Compliance by Banks, Paris: OECD, available at: www.oecd.org/dataoecd/3/43/42827589_pdf; accessed 1 May 2011. See www.ciat.org; accessed 1 May 2011. See www.catatax.org; accessed 1 May 2011. See www.iota-tax.org; accessed 1 May 2011. PricewaterhouseCoopers, Paying Taxes 2009: The Global Picture, Washington: PwC. In fact, several senior tax officials in several Western countries have shared with the author how gaming does in fact take place in their own tax agency.

374

The Elgar guide to tax systems

REFERENCES Alley, C. and D. Bentley (2008), ‘The Increasing Imperative of Cross-Disciplinary Research in Tax Administration’, e-Journal of Tax Research, 6(2), 122–44. Alm, J., J. Martinez-Vazquez and M. Rider (2005), The Challenges of Tax Reform in a Global Economy, New York: Springer. Australian Tax Office (2007), Literature Review: Measuring Compliance Effectiveness, Canberra: ATO. Barros, C. (2007), ‘Technical and Allocative Efficiency of Tax Offices: A Case Study’, International Journal of Public Sector Performance Management, 1(1), 41–61. Bird, R. (2004), ‘Administrative Dimensions of Tax Reform’, Asia-Pacific Tax Bulletin, 10(3), 134–50. Bird, R. and M. Casanegra de Jantscher (1992), Improving Tax Administration in Developing Countries, Washington, DC: International Monetary Fund. Bird, R. and E. Zolt (2008), ‘Technology and Taxation in Developing Countries: From Hand to Mouse’, National Tax Journal, 61(4), 791–821. Braithwaite, V. (2003), Taxing Democracy: Understanding Tax Avoidance and Evasion, Aldershot: Ashgate. Braithwaite, V. (2007), ‘Responsive Regulation and Taxation: Introduction’, Law and Policy, 29(1), 3–10. Cyr, D. and G. Swanson (2007), ‘Not Quite the Triumph They Describe: A Response to Rainey and Thompson’, Public Administration Review, 67(3), 576–8. Desai, M. (2005), ‘The Degradation of Reported Corporate Profits’, Journal of Economic Perspectives, 19(4), 171–92. Ebrill, L., M. Keen, J-P. Bodin and V. Summers (2001), The Modern VAT, Washington, DC: International Monetary Fund. Flynn, N. (2007), Public Sector Management. London: Sage. Forsund, F., S. Kittelsen, F. Linseth and D Edvardsen (2006), ‘The Tax Man Cometh – But Is He Efficient?’, National Institute Economic Review, 197, 106–19. Freedman, J., G. Loomer and J. Vella (2009), ‘Corporate Tax Risk and Tax Avoidance: New Approaches’, British Tax Review, 1, 74–116. Gill, J. (2003), ‘The Nuts and Bolts of Revenue Administration Reform’, Washington, DC: World Bank, mimeo. Hasseldine, J. (2008), ‘The Search for Best Practice in Tax Administration’, in C. Evans and M. Walpole (eds), Tax Administration: Safe Harbours and New Horizons Birmingham: Fiscal Publications, pp. 7–18. Hasseldine, J., K. Holland and P. Van der Rijt (2009) The Management of Tax Knowledge, London: Association of Chartered Certified Accountants. Hood, C. (2006), ‘Gaming in Targetworld: The Targets Approach to Managing British Public Services’, Public Administration Review, 66(4), 515–21. Kirchler, E. (2007), The Economic Psychology of Tax Behaviour, Cambridge: Cambridge University Press. Martinez-Vazquez, J. and B. Torgler (2009), ‘The Evolution of Tax Morale in Modern Spain’, Journal of Economic Issues, 43(1), 1–28. Martinez-Vazquez, J., J. Arze del Granado and J. Boex (2007), Fighting Corruption in the Public Sector, Oxford: Elsevier. McCarten, W. (2005), ‘The Role of Organizational Design in the Revenue Strategies of Developing Countries’, paper presented at the Global Conference on Value Added Tax, Rome. Mikesell, J. and L. Birskyte (2007), ‘Another View of IRS Results: A Comment on Rainey and Thompson’, Public Administration Review, 67(3), 574–6. Moesen, D. and A. Persoon (2002), ‘Measuring and Explaining the Productive Efficiency of Tax Offices: A Non-parametric Best Practice Frontier Approach’, Tijsdchrift Voor Economie en Management, 47(3), 399–416. National Audit Office (2008), ‘Benchmarking of Tax Administrations: Report of the EUROSAI Study Group’, unpublished report.

The administration of tax systems 375 Onrubia, J. (2006), ‘The Reform of Tax Administration in Spain’, International Studies Program Working Paper No. 06-12, Andrew Young School of Policy Studies, Georgia State University. Owens, J. and S. Hamilton (2004), ‘Experience and Innovations in Other Countries’, in H.  Aaron and J. Slemrod (eds) The Crisis in Tax Administration, Washington, DC: Brookings. Proctor, S. and G. Currie (2004), ‘Target-based Teamworking: Groups, Work and Interdependence in the UK Civil Service’, Human Relations, 57(12), 1547–72. Rainey, H. and J. Thompson (2006), ‘Leadership and the Transformation of a Major Institution: Charles Rossotti and the Internal Revenue Service’, Public Administration Review, 66(4), 596–604. Rainey, H. and J. Thompson (2007), ‘Response to Mikesell and Birskyte, and to Cyr and Swanson’, Public Administration Review, 67(3), 579–83. Sandford, C. (1993), Successful Tax Reform: Lessons from an Analysis of Tax Reform in Six Countries, Bath: Fiscal Publications. Slemrod, J., J. Whiting and J. Shaw (2010), ‘Administration and Compliance’, Chapter 12 in The Mirrlees Review: Reforming the Tax System for the 21st Century, vol. I, Oxford: OUP for the IFS. Swedish Tax Agency (2005), Right from the Start, Stockholm: Swedish Tax Agency. Tan, C. and S. Pan (2003), ‘Managing e-Transformation in the Public Sector: An e-Government Study of the Inland Revenue Authority of Singapore’, European Journal of Information Systems 12(4), 269–81. Tomkins, C., C. Packman, S. Russell and I. Colville, (2001), ‘Managing Tax Regimes: A Call for Research’, Public Administration 79(3), 751–8. Torgler, B. (2007), Tax Compliance and Tax Morale: A Theoretical and Empirical Analysis, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Toumi, M. (2009), ‘Cultures of Compliance: British and French Tax Enforcement Compared’, unpublished, University of Lancaster PhD dissertation. Vehorn, C. and J. Brondolo (1999), ‘Organizational Options for Tax Administration’, Bulletin for International Fiscal Documentation 53(11), 499–512.

11 Political regimes, institutions, and the nature of tax systems* Stanley L. Winer, Lawrence W. Kenny, and Walter Hettich

1

INTRODUCTION

To the casual observer who does a quick survey of the data, tax systems present a confusing array of structures and forms, particularly if international comparisons are included. To the political economist, the variety of observed forms of taxation presents an interesting research challenge and raises a number of important questions. Can the seemingly confusing array of data be classified in a meaningful way? Is it possible to explain both differences and similarities in tax regimes? How do political factors and institutions influence the nature of observed tax systems? How do such factors interact with the underlying economy in determining the use of different revenue sources? Can the comparison of international tax regimes help in formulating better policy? In this chapter, we assess the contributions of current research in political economy to provide answers to these questions, while also presenting some new statistical results on the relation between tax structure and political regimes. Our discussion of the literature is selective and is empirically oriented. Our primary goal is to give a sense of some of the empirical research possibilities that lie ahead.1 It is now widely recognized in the literature that observed tax systems reflect the interplay among political forces together with the influence of current and past economic factors. In democratic regimes, they can be modeled as direct outcomes of political competition, tempered by economic factors. In non-democratic countries, a similar logic applies. As pointed out by McGuire and Olson (1996) more than a decade ago, both types of regimes have similar incentives to capture the efficiency gains in designing policy structures for large segments of the population, leading to broad similarities in public sector policies. In conducting research, it is useful to view tax systems as policy structures with a few essential features that are both identifiable and measurable. We shall refer to the collection of these characteristics, consisting of tax bases, rate structures and special provisions, as the ‘tax skeleton’. 376

Political regimes, institutions, and the nature of tax systems 377 Depending on the level of aggregation, comparisons of different fiscal regimes may focus mainly on tax mix (the relative importance of revenues from different bases), they may include data on both revenue bases and tax rates, or may just deal with rates. Most of the work discussed below deals either with the tax mix, or with tax rates. In rare cases, comparative studies will cover all elements of the skeleton, although the complexity of special provisions found in most political regimes generally forces researchers to use a highly stylized or selective representation of this component of the tax skeleton. Our discussion proceeds in several steps. In the second section, we present an overview of major questions asked and of the nature of the comparisons that will be examined. This is followed by a brief discussion of the underlying model of political economy. The analysis in the following sections moves from the general to the more detailed, starting with the comparative analysis of the tax mix in democratic and non-democratic political regimes. The focus then turns to comparisons between democratic states that have different electoral systems, followed by consideration of the role of the microstructure of democratic political institutions. In a further section, we comment on the logic of equilibrium revenue structures in mature democratic states, and draw attention to research that aims to uncover the direct influence on fiscal outcomes of the distributions of relevant characteristics over the electorate. A final substantive section relates the chapter’s approach to globalization and tax competition.

2

POLITICAL INSTITUTIONS AND TAX STRUCTURE – AN OVERVIEW

To understand the logic of the analysis in this chapter more easily, it is useful to begin with a schematic presentation of related challenges and issues that we shall address in order to contribute to an understanding of the relationship between political institutions and observed tax structures. This schematic is presented in Figure 11.1. As in all investigations in empirical research, we need to start with an underlying theoretical model. In this case, we require a model of how tax structure is generated as an outcome of the interaction of political forces that is general enough to suggest hypotheses for all types of tax systems. Such a model will be presented in some detail in Section 3. At this point, we merely identify it as a necessary beginning of the analysis at the top of Figure 11.1. Given this starting point, we can identify the relevant comparisons

378

The Elgar guide to tax systems All Tax Systems (Model of Tax Structure) Democratic

Proportional

Non-Democratic

Majoritarian

Microstructure of Electoral Systems Examples: Historical and descriptive Transaction costs in Canada vs US Veto players in different systems

Logic of Equilibrium Tax Structure in Stable Political System

Source: Authors.

Figure 11.1

Effects of political institutions on tax systems: comparative analysis

among different types of political regimes or among institutional features that are of most interest. As the figure shows, we shall move from broader to more narrowly defined questions. First we concern ourselves with comparisons of tax structure in democratic versus non-democratic countries. The main question here is this: after controlling for the usual determinants of tax structure suggested by the theoretical model, what effect on fiscal structure can be attributed directly to the degree of democracy? Although the underlying model yields hypotheses for all types of tax systems, it is most directly applicable to democratic regimes that exhibit a sustained degree of political competition. There are interesting differences in institutions across the democratic countries. A major distinction that has drawn much attention among researchers relates to the nature of electoral systems, namely whether parliamentary decisions are made under proportional or majoritarian rules. In this connection, we want to ask what observable differences in tax structure can be traced to such differences in the nature of democracy, while controlling for other well-known determinants. Further disaggregation gives rise to additional questions. If we look at the microstructure of electoral systems, we can focus on more specific

Political regimes, institutions, and the nature of tax systems 379 elements, such as differences in political transaction costs and the role and importance of veto players in a particular political setting. It is also possible to approach the analysis in a more historical or descriptive way, with particular attention being paid to political history and to the influence of pressure groups previously identified in the political science literature. Research at the micro-level also involves the interactions of economic and political characteristics in a general equilibrium analysis of tax structure in a stable democratic system. Because of the necessary complexity of a fully developed general equilibrium model suitable to deal with such interactions, analysis of this type uses simulation techniques rather than the statistical analysis of observed data. Although only limited work has been done in this area, research suggests that, in addition to the average levels, the higher moments of the distribution of skills, tastes for public goods, and of political influence also have a significant impact on the determination of the tax mix.

3

A BASIC MODEL OF TAX STRUCTURE

Tax structure in competitive political systems is part of a broader political and economic equilibrium. In the underlying model, political parties compete for support from a population of heterogeneous voters by offering different policies to the electorate. With regard to tax policy, such competition results in a revenue structure that minimizes the loss in support, or the political costs, associated with the use of different tax bases (Hettich and Winer, 1999). The analysis represents an application and extension of the theory of probabilistic voting, a theory that is now employed widely in the literature dealing with equilibrium outcomes in the public sector (see, for example, Coughlin, 1992; Hinich and Munger, 1997; Hettich and Winer, 1999; Persson and Tabellini, 2000; Adams et al., 2005; and Schofield and Sened, 2006, among others). To aid the discussion, we present a brief graphical exposition of a model of tax mix in Figure 11.2, based on Hettich and Winer (1988, 1999) and developed further in an empirical context in Kenny and Winer (2006). A similar model has been used by several authors including, for example, by Aidt and Jensen (2009), Canegrati (2009), Galli and Profeta (2009), Geys and Revelli (2009), and Profeta and Scrambrosetti (2010). Let us assume that the government pursues a sole objective, namely to get re-elected, and that it tries to reach this goal by choosing policies that maximize total expected support from a heterogeneous population of voters. The probability of receiving support is affected negatively by the imposition of taxes needed to finance the public budget, but it is also

380

The Elgar guide to tax systems

Marginal Political Costs and Benefits Per Dollar

Tax Base A

Tax Base B

Marginal political costs per $

Total marginal political benefits per $

Marginal political costs per $ 3

3

1

Total marginal political costs per $

2

3

1

1

2

2

0 tA

Rb

Ra ta

Laffer curves net of admin costs

R1

Ra + Rb

Legend: Laffer curves net of admin costs

Equilibrium ‘1’: Initial equilibrium Equilibrium ‘2’: Tax base A expands Equilibrium ‘3’: Administration costs of base B increase

Tax Rates

Note: Except by mistake, or in the strenuous pursuit of objectives that necessarily compromise its ability to raise revenue, no government would choose a point on the backward bending part of any rate–revenue (Laffer) curve shown, since this would lead to a marginal political cost that is higher than that implied by the lower tax rate that raises the same revenue. Source: Hettich and Winer (1999).

Figure 11.2

Political equilibria in a competitive political system

influenced positively by the provision of public goods. In reacting to the government’s policies, voters regard their own tax payments as being independent of the benefits of the public goods and services that they receive, a condition that reflects the difficulty that voters face in the real world in trying to relate marginal individual tax payments to marginal adjustments in public output. Taxation reduces the disposable income of voters and, in addition, it creates a welfare loss because of economic adjustments made by voters in response to tax laws. The resulting consequences for the relationship between tax rates, the size of tax bases and tax revenues are summarized in Figure 11.2 by the rate–revenue relationships or Laffer curves in the bottom part of the first two panels. The loss in disposable income combined with the welfare burden of taxation create a loss in full income for taxpayers that motivates their opposition to taxation. The marginal political cost functions in the upper part of the panels corresponding to the two tax bases show how the loss

Political regimes, institutions, and the nature of tax systems 381 in full income at various tax rates is translated into political opposition or into reduced support, with marginal political cost being expressed per dollar of revenue raised. Individuals are affected differently by taxation, depending on the taxable activities in which they engage and their willingness and abilities to make adjustments in economic behavior in order to escape tax liabilities. In addition, they face unequal costs in organizing political opposition. Regarding public output, voters have different tastes for and evaluations of public goods and services. As a result of these factors, individual marginal political cost curves, as well as marginal political benefit curves, will differ among them. In the figure, these individual marginal curves have been added vertically into total marginal political cost functions associated with each revenue source and into a total marginal benefit function. As shown by the curves passing through the points labeled as ‘1’, the resulting marginal political cost per dollar functions differ across bases, reflecting the fact that voters as a group have different evaluations of the economic effects of taxation levied on a particular activity, and different ways of translating their evaluations into political opposition. The third panel of Figure 11.2 shows the determination of budget size. The desired initial budget is at point ‘1’ where total marginal benefit per dollar of expenditure equals total marginal cost per dollar raised. The implications for fiscal structure are then readily understood if we realize that any government, in its attempt to maximize political support, will minimize total political costs for any given level of revenue collected. To achieve this, marginal costs per dollar raised must be equalized across bases. Given two bases, the government will tax both and will do so at different rates. One such equilibrium is represented in Figure 11.2 by the point labeled ‘1’ with associated tax rates tA and tB and a total budget size of R1. It is important to point out that the rate–revenue curves are defined net of administration curves. By doing so, we are able to incorporate the resource costs of taxation, consisting of the costs of administering, monitoring, and enforcing tax collections as well as compliance costs, directly into the graphic analysis, thus making them part of the budgetary equilibrium. Different versions of Figure 11.2 can be used to investigate the effects of exogenous shocks on budget equilibrium and to derive testable hypotheses in this manner. Three types of shocks have proven particularly important in the empirical literature, namely those related to the size of tax bases, to changes in administration costs and to differences in the scale of the public sector.

382

The Elgar guide to tax systems

Base effects Figure 11.2 illustrates the effects of an exogenous increase or growth in the size of Base A on tax mix and on government size. Such a change will lead to an outward shift of the associated Laffer curve and, as a result, a downward shift in the marginal political cost function (now represented by the dotted line). This is so because any level of revenue for Base A can now be collected at a lower rate than before, thus reducing the loss in full income per dollar raised for affected taxpayers. In addition, given the fixed costs of organizing opposition, the incentive to generate such opposition per dollar is now reduced. The new equilibrium in the panel for Base A is represented by point ‘2’, indicating a higher amount of revenue collected from this source. The shift also has implications for total budget size (see point ‘2’ in the third panel) and for collections from Base B. As shown in the figure, total budget has increased, while revenue from Base B is now lower than before. Administration cost effects An exogenous increase in the costs of administering taxation on Base B is depicted in Figure 11.2 by a leftward shift in the Laffer curve for this base (defined net of such costs). The new equilibrium is labeled as ‘3’, indicating less total collection from Base B. As before, adjustments will also occur in the other two panels. In this case, collections from Base A will increase in relation to the starting point of the analysis, while total budget or government size will decrease. Scale effects While not shown explicitly to make the figure easier to view, an upward shift in the political benefit curve in the third panel, which obviously leads to an increase in the equilibrium size of the public sector, will also normally lead to increased reliance on all tax sources. (To see this, draw an imaginary line backwards from this new intersection in panel 3 to the marginal cost functions in the first two panels.) The graphic analysis in Figure 11.2 is limited to the tax mix. Although this model can in principle be expanded to include the other components of the tax skeleton (see Hettich and Winer, 1999), we shall not do so here since the empirical work reported later focuses mainly on variations in the mix of tax sources. As mentioned in the introduction, there are good arguments for applying the approach represented by Figure 11.2 also to non-democratic regimes, even though the model was initially developed for competitive democracies. As McGuire and Olson (1996) point out, all types of regimes, and especially those that expect to survive for some time, will attempt to capture efficiency gains and thus will experience similar scale, base, and administration cost effects. In addition, work by Wintrobe on dictatorship

Political regimes, institutions, and the nature of tax systems 383 (1998) suggests that at least some types of authoritarian regimes (such as those that want to maximize their power) may have incentives to minimize the costs of raising given levels of revenue, even though he does not specifically analyze the tax mix. In any case, it is clear that all types of regimes will have to face rate– revenue relationships for particular tax bases, although these may assume different shapes for dictatorial and democratic societies since dictatorships may of necessity emphasize enforced rather than voluntary compliance. In fact, empirical research does lend support to the proposition that democratic regimes rely more heavily on taxes requiring voluntary compliance, thus implying that the degree of democracy itself is a relevant variable in explaining observed differences in tax structure across democratic and non-democratic regimes.

4

TAXATION IN DEMOCRATIC AND AUTHORITARIAN REGIMES: DOES CONSENT MATTER?

We return to the question posed earlier: after controlling for the usual determinants of tax structure, what effect on the tax mix can be attributed to the degree of democracy? By the degree of democracy, we are referring to the extent to which there is effective political competition for governmental power. Table 11.1 presents a record of the average tax revenue raised from various sources for 100 democratic and non-democratic regimes between 1975 and 1992, for each of three ranges of the combined Gastil index of liberal-democracy (Freedom House, various years). This combined index adds together the Gastil indexes of civil liberties and of political rights. Each component goes from 1 for the most liberal-democratic regime to 7 for the least, so that 2 is the best score and 14 the worst. The countries of the OECD tend to be in the group with a combined Gastil index of less than 5. The table deliberately excludes the period after the collapse of the Soviet Union and the resulting political and economic transition. The table shows clearly that in the advanced democracies, where the combined Gastil index of civil liberty and political rights has a value of less than 5, relative reliance on personal income taxation (including Social Security and payroll taxes) is higher. More precisely, the share of revenue coming from personal income taxation in advanced democracies is more than two and a half times the share found in less democratic regimes. The advanced democracies – where rule of law applies to everyone including

384

The Elgar guide to tax systems

Table 11.1

Tax structure across 100 democratic and non-democratic regimes, 1975–92

Revenue Sources

Combined Gastil Index 2.0–4.9

5.0–9.9

10–14

Revenue shares Corporate Individual income Social security & payroll Goods & services (domestic) Trade Property Non-tax sources Rate of inflation Number of observations = 269

0.105 0.218 0.195 0.241 0.109 0.020 0.109 16.64 97

0.126 0.092 0.069 0.265 0.228 0.017 0.195 33.83 74

0.130 0.080 0.060 0.216 0.271 0.017 0.216 21.32 98

Note: Observation numbers refer to averages over three, five-to-seven-year periods from 1975 to 1992, used in the empirical work reported below. Combined Gastil = sum of indexes of civil liberties (1 to 7) and political rights (1 to 7): Most democratic = 2; least democratic = 14. Sources: See Kenny and Winer (2006, Table 2) for exact definitions and sources of data. Tax data are from the IMF: see Kenny and Winer (2006) and Woldemariam (1995).

elites, and competitive forces in both politics and economics are impersonal and robust – raise just over 40 percent of total tax revenues from individual income and from Social Security and payroll taxes. On the other hand, authoritarian regimes – those with a combined Gastil of more than 5 – rely less than half as much on this tax source. One should note that as far as the authoritarian regimes are concerned, it does not appear to matter whether the regime is of a type that Wintrobe (1998) would say just attempts to stay in power (5–9.9), or is the sort of totalitarian regime that wants to maximize power (10–14).2 All types of non-democratic regimes appear to be similar in this respect. The personal income tax is hard to collect compared with forms of taxation that do not tax income directly, and especially so from people who own small businesses. Income tax withholding greatly simplifies collection from individuals who earn a wage. But in the mature, democratic societies represented in the table, political competition over the decades would surely have led to the abolition of withholding if this practice did not have the widespread and continuing support of the majority of the electorate. So it is tempting to hypothesize that this hard-to-collect tax, which is

Political regimes, institutions, and the nature of tax systems 385 administratively and economically more efficient for the taxation of the income base than (indirect) sales taxes, can be relied upon to a greater extent in advanced democracies than in autocratic regimes because their citizens consent to this form of taxation. Since the index of democracy in Table 11.1 is positively correlated with the level of economic development or per capita GDP, one might easily object to drawing inferences about the role of consent without controlling for the determinants of tax structure that may also be correlated with the nature of economic and political development. This is done to the extent allowed by available data in Kenny and Winer (2006), where the basic model is applied to empirically model the tax mix as a whole across the 100 countries represented in Table 11.1 for 1975 to 1992. Here all data are averaged over the five to seven years that comprise each of three time periods considered to reduce the influence of macroeconomic fluctuations on the results. The set of equations from Kenny and Winer’s work representing the share of revenues raised from each source by central governments, reported in Table 11.2, is a full tax system. That is, a rise in the share of revenue taken from one base is offset by a fall in the share of revenue obtained from all the other bases.3 Revenue shares reflect choices about tax bases, rate structures, and about the precise definition of taxable activity and effective rates of tax through special provisions such as exemptions and deductions. Thus all three major elements of the tax skeleton are implicitly included in a model of the tax mix. Modeling tax shares is more interesting than modeling nominal tax rates, which are just one simple part of the tax skeleton and for which implementation of a full system would be extremely difficult. A model of effective tax rates does incorporate special provisions, but also carries with it the problem of representing the system as a whole. An ideal approach would be to model a tax system in which each of the three major elements of the tax skeleton is separately represented. But to the best of our knowledge, this remains to be accomplished in empirical research.4 The model in Table 11.2 excludes the scale effect represented by total revenues relative to GDP that is included in Table 3 of Kenny and Winer (2006). One may regard the equations in this table as reduced forms in which the scale of the public sector, which is endogenous and depends on the included right-hand-side variables, has been solved out of the system. Regressions based on this specification have not been previously published. The results provide strong support for our predictions that a rise in the tax base or a fall in administrative costs leads to greater reliance on this revenue source. The results remain very similar if the scale variable is included. A full discussion of the results will not be given here so that

386

GDP coef. var

Log pop. density

Urbanization

Second enroll

LF: % female

GDP per worker

Trade

Crude petrol

–0.0149 (0.25) −.0254 (1.45) 17.532 (7.76) 0.0237 (1.83) 0.51X10−5 (2.51) −0.00389 (6.27) −0.00027 (0.73) 0.24X10−4 (0.06) 0.00458 (0.90) 0.273 (1.17)

Non-tax 0.197 (4.11) 0.070 (2.89) 8.635 (2.20) −.00271 (0.37) −0.49X10−5 (2.84) 0.00028 (0.47) −0.00074 (1.64) 0.00078 (1.38) −0.00301 (0.79) 0.235 (1.26)

Corporate 0.295 (4.40) −0.00845 (0.48) −7.330 (3.82) −0.0330 (5.64) 0.34X10−5 (2.05) 0.00091 (1.76) −0.23X10−4 (0.07) 0.00010 (0.26) −0.00946 (2.02) −0.254 (1.87)

Individ. Income –0.0102 (0.23) 0.0390 (1.63) −9.595 (5.32) −0.0306 (3.35) 0.98X10−5 (6.19) 0.00065 (1.06) 0.00022 (0.46) −0.00076 (1.72) 0.0117 (2.08) −0.205 (1.09)

Soc. Sec. & Payroll 0.0978 (1.57) −0.0382 (1.89) −5.077 (2.32) −0.0531 (3.80) −0.38X10−5 (2.02) 0.00268 (3.64) 0.00087 (1.69) 0.00134 (2.46) 0.00342 (0.73) −0.168 (0.82)

Goods & Services 0.417 (6.78) 0.00836 (0.54) −2.094 (0.72) 0.0850 (8.45) −0.80X10−5 (4.17) −0.00058 (0.74) −0.11X10−4 (0.02) −0.00227 (4.28) −0.00937 (1.90) 0.0275 (0.11)

Trade

– 0.00120 (0.11) −0.00977 (3.41) −1.161 (3.92) .00040 (0.22) 0.25X10−6 (0.79) 0.00012 (0.87) −0.00011 (1.32) 0.00019 (2.14) 0.00343 (3.21) 0.0210 (0.48)

Property

Tax share regressions for central governments, 1975–92 analogous to Table 3 in Kenny and Winer (2006), but with total revenue/GDP (the scale effect) excluded (absolute t-statistics in parentheses, based on Huber-White standard errors)

Federal structure

Intercept

Table 11.2

387

0.00265 (0.25) 0.0246 (0.88) 0.0358 (3.56) 0.0100 (2.70) 0.0229 (1.50) 0.0228 (1.53) 0.5938 0.1004 269 5

−0.0114 (0.81) 0.0499 (2.09) −0.00796 (2.33) 0.00672 (0.49) 0.00122 (0.08) −0.00321 (0.23) 0.1550 0.0961 248 11

−0.00654 (0.73) −0.0108 (0.48) −0.0324 (3.12) −0.00182 (0.71) −0.00967 (0.66) −0.0138 (0.94) 0.4244 0.0910 246 5 −0.00805 (0.40) −0.0195 (0.98) 0.3971 0.1104 198

0.00531 (0.43) 0.0880 (2.53) −0.00110 (0.41)

−0.00434 (0.34) 0.0708 (2.26) 0.0130 (1.60) −0.0163 (3.09) −0.00734 (0.43) −0.00754 (0.45) 0.2733 0.1106 269 7

0.0154 (1.19) −0.128 (5.21) −0.00246 (0.44) 0.00655 (0.89) −0.00753 (0.43) 0.00432 (0.23) 0.5518 0.1173 269 9

−0.00160 (0.82) −0.0110 (2.58) −0.00135 (0.82) 0.00120 (2.02) 0.00011 (0.04) 0.00050 (0.15) 0.1288 0.0198 258 5

Source:

Authors.

Note: Intercepts omitted. Results are essentially the same if the scale of the public sector (total revenue/GDP) is included as in the Kenny-Winer paper. In that case, the Gastil 1st segment in the individual income tax equation has a coefficient of –0.0327 (t = 3.13). Other coefficients are also similar in sign and size. The full list of countries used in this table is found in Kenny and Winer (2006).

Adj. R–square Root MSE Number of obs. Spline Break

1986–92

Combined Gastil: 1st segment Combined Gastil: 2nd segment 1981–85

Socialist

Coups

388

The Elgar guide to tax systems

we may focus on the role of the degree of democracy as measured by the Gastil index. The combined Gastil index is entered in the estimating equations using a two-segment spline, which is a piece-wise linear function that changes slope at some combined Gastil value called the break point. The break point that provides the best fit of the data is labeled in Table 11.2 as ‘Spline break’. In the individual income tax regression, the break point occurs when the combined Gastil equals 5, just as the descriptive data suggests. Moreover, the coefficient on the first segment of the spline is negative and statistically significant, with a coefficient that indicates that the relative reliance on the income tax would drop by about 0.1, that is, from 0.2 or 20 percent to about 10 percent when the Gastil index rises (and democracy declines) from 2 to 5. The coefficient on the second spline segment is not statistically significant, implying that additional increases in the Gastil do not lead to a further drop in the reliance on income taxation. The results in Table 11.2 show that the first impression one gets from the descriptive data is not misleading. Democratic countries do rely substantially more on personal income taxation than do non-democratic regimes by a factor of about two, even after controlling for various factors that shape tax structures and that are correlated with the level of development.5 Such a finding could be caused by higher citizen consent in democracies as we suggested earlier or, perhaps, to democracies redistributing more by using progressive income taxation, or both. Mulligan et al.’s (2004) demonstration that income tax rate structures are somewhat flatter in democratic countries than in non-democratic ones suggests that our result is due not only to redistribution being relatively more important in advanced democracies.6 So too do the results on the second segment of the spline, which show that once the combined Gastil index rises above the value indicated in the row labeled ‘Spline break’, non-democratic regimes rely more heavily on property and non-tax revenues (revenues from sales, interest, and the like), where self-reporting in not usually involved.7 The importance in Table 11.2 of what appears to be popular consent to taxation of a particular form is striking. The role of consent is not usually dealt with in public finance texts, despite its importance in the well-known Wicksell-Lindahl framework.8 Further research to confirm, amend, or refute these results is warranted because of what they imply for the way we think about the evolution and reform of tax structures both in advanced and in less democratic regimes.

Political regimes, institutions, and the nature of tax systems 389

5

TAX STRUCTURE AND ELECTORAL SYSTEMS: DO MAJORITARIAN SYSTEMS RELY MORE ON PERSONAL INCOME TAXATION?

We now narrow the discussion to consider the role of electoral systems in the more advanced democracies. There has been a lot of empirical work on the size of government in a comparative perspective; see, for example, Persson and Tabellini (2003) and Bueno de Mesquita et al. (2005). But there has not been much comparative empirical work in medium to large N situations dealing with the effects of governance on tax structure, with Cusack and Beramendi ‘s (2006) study of labor taxation in OECD countries being a notable exception.9 Iversen and Soskice (2006) begin a comparative analysis of majoritarian and proportional electoral systems with the observation that in a sample of 17 advanced democracies between 1945 and 1998, those with proportional representation had a far higher incidence of left-wing governments. According to Iversen and Soskice (2006, Table 1), about 75 percent of governments in PR systems were left of center, meaning that they favored redistribution to lower-income groups. In contrast, in countries with majoritarian systems, about 75 percent of governments were right of center and less favorably inclined to support progressive redistribution. The obvious hypothesis is that there will be more redistribution in countries with a proportional electoral system. Iversen and Soskice provide an excellent theoretical model of why this pattern is predictable, which we only briefly summarize in a footnote in order to focus on their empirical work and our new results.10 Iverson and Soskice test their model by regressing the percentage reduction in the Gini from before to after taxes and transfers on an indicator of the type of electoral system, along with several control variables, using data for 14 countries in the Luxembourg Income Study (for 1967 to 1997). Their regressions confirm that PR is significantly associated with a greater reduction in inequality compared with countries with a majoritarian electoral system. It is difficult to know the pre-fisc Gini, since a complicated general equilibrium calculation would be required to determine it. For this reason, but also because it is an interesting issue in its own right, we investigate the direct impact of majoritarian and proportional electoral systems on the tax mix in the context of the model presented in Table 11.2. In view of the correlation of electoral systems with party ideology, and since personal income taxation tends to be the most progressive element in the tax structures of advanced democracies, it seems reasonable to expect that our investigation will show that countries with a majoritarian electoral system rely relatively less heavily on the individual income tax.

390

The Elgar guide to tax systems

Table 11.3

Tax share regressions for central governments, 1975–92, Majoritarian included (absolute t-statistics in parentheses, based on Huber-White standard errors) Non-tax Corporate Individ. Soc. Sec. Goods & Trade Property Income & Payroll Services

Majoritarian 0.063 (vs. PR or (3.36) mixed) Adj. R-square 0.7819 Root MSE 0.0833 Number of 132 obs.

0.027 (2.53) 0.1023 0.0700 128

0.034 (1.70)

−0.089 (4.74)

−0.058 (3.16)

0.041 −0.0066 (1.91) (2.00)

0.4513 0.5605 0.3487 0.6575 0.2557 0.0986 0.0987 0.1057 0.1026 0.0184 127 116 132 132 132

Note: Full results from model in Table 11.2 not reported. Results are similar if a scale variable (total revenue/GDP) is included. Countries with a Presidential system are excluded from the sample. Results are similar if these countries are included. Source: Based on re-estimation of model presented in Table 11.2 amended as indicated.

A graph of the personal tax (including Social Security and payroll taxes) to corporate tax ratio in OECD countries by type of electoral system is given in Figure 3 below. We see that countries with PR do appear to rely relatively more heavily on the taxation of personal incomes. But this graph is only a representation of simple correlations. Estimation of a full regression model may reveal something quite different. The results of adding an indicator representing the type of electoral system into the model of tax structure in Table 11.2 are recorded in Tables 11.3 and 11.4. Here only the estimated coefficients of the indicator variable Majoritarian (= 1 if the electoral system for the lower elected chamber is majoritarian, = 0 otherwise) are presented. Presidential systems have been removed from the sample, so that the alternative to a majoritarian system is PR or mixed PR. For the sample of countries used in Table 11.2, we see in Table 11.3 that majoritarian electoral systems make significantly more use of the individual income tax, and place less reliance on somewhat more regressive Social Security and payroll taxes. They also rely less on domestic consumption and property taxes, and more on corporate taxation and non-tax revenues. In Table 11.4 the sample is restricted to countries for which the combined Gastil is 5 or less, as is suggested by the results for the spline break in Table 11.2. The coefficient value for Majoritarian in the individual income tax equation is 0.07, more than twice its value in Table 11.3. This

Political regimes, institutions, and the nature of tax systems 391 Table 11.4

Tax share regressions for central governments, 1975–92, Majoritarian included, sample restricted to countries for which sum of GASTIL indexes = 2 to 5 (absolute t-statistics in parentheses, based on Huber-White standard errors) Non-tax Corporate Individ. Soc. Sec. Goods & Trade Property Income & Payroll Services

Majoritarian (vs. PR or mixed) Adj. R-square Root MSE Number of obs.

0.0046 (0.57)

0.027 (2.16)

0.070 (2.32)

−0.082 (2.88)

0.3994

0.1612

0.3597

0.4473

0.0548 84

0.0653 83

0.1022 83

0.1109 79

−0.088 (5.08) 0.4922

0.091 (3.90) 0.6483

−0.0029 (0.99) 0.3615

0.0849 0.0951 0.0125 84 84 84

Note: Full results from model in Table 11.2 not reported. Results are similar if a scale variable (total revenue/GDP) is included. Countries with a presidential system are excluded from the sample. Results are similar if these countries are included. Source:

Based on re-estimation of model presented in Table 11.2 amended as indicated.

coefficient indicates that the share of revenue coming from individual income taxes is 0.07 higher in the more advanced democratic countries using a majoritarian system in the lower chamber than if PR (or a mixed system) is employed. Since the share of taxes raised by the personal income tax ranges from 0 to 0.563 and has a standard deviation of 0.135 in the sample with a combined Gastil index of 5 or less (see Appendix Table 11A.1), the estimated effect is substantial. These estimation results refute the simple prediction made earlier. Majoritarianism is associated with greater reliance on personal income taxation than is PR, not less. Instead majoritarian countries rely less on Social Security and payroll taxes, the usual structure of which limits their progressivity, less on consumption taxes and more on corporate taxation. This conclusion is further bolstered by the work of Aidt and Jensen (2009), who find a similar pattern for a sample of ten Western European countries between 1860 and 1938! They show that the presence of PR is significantly associated with less reliance on direct taxes including corporate taxation, and also greater reliance on domestic trade taxes. Cusack and Beramendi (2006) point out that the correlation of PR with the existence of a coordinated market, in which employers and unions bargain at a central level over wages, is high, at over 0.7 (Gourevitch and Haves, 2002). PR and the existence of coalition government is also highly

392

The Elgar guide to tax systems

correlated. Their analysis of wage taxation in these countries in the light of these facts helps to explain the pattern of results in Tables 11.3 and 11.4.11 The Cusack and Beramendi analysis, on our reading, suggests that (the usually) left of center coalition governments in PR countries provide a high level of public benefits, including generous pensions, and pay for them with relatively high taxes on the labor incomes of those who benefit, especially where the legislature is strong relative to the executive. It is important in their view that wage bargaining reduces the extent to which these taxes are passed on to employers by unions, so that unemployment does not rise and investment is maintained. Moreover, since governments in PR systems are usually coalitions, the credibility of wage bargains is enhanced because it is difficult for any particular government to meddle with them. There is a complementary story based on Lindert (2004). The evidence assembled by Lindert suggests that the relatively large public sectors of the EU leads them to impose tax structures that are less harmful of economic efficiency and growth than would be the tax systems of countries with smaller public sectors. A related argument appears in Steinmo (1993). One wonders, though, why vigorous political competition fails to insure the same degree of efficiency in countries like the UK, the US, and Canada. In any case, the new results in Tables 11.3 and 11.4 indicate that in PR, as opposed to majoritarian electoral systems, there is greater reliance on Social Security and payroll taxes and also on domestic trade taxes. Exactly how these facts can be reconciled with the Iversen-Soskice finding that PR countries reduce inequality more than do majoritarian ones remains to be seen. No doubt more than tax structure is involved.

6

THE COMPARATIVE MICROSTRUCTURE OF ELECTORAL SYSTEMS AND TAX STRUCTURE

Sven Steinmo’s book Taxation and Democracy (1993) represents an early attempt to link tax systems to the microstructure of their specific institutional setting. He argues that political institutions are not neutral. Going one step further, he proposes that such institutions not only affect the relative distribution of power among participants in the political process, but that ‘the structure of a polity’s decision-making institutions also profoundly affects how interest groups, politicians and bureaucrats develop their policy preferences’ (p. 7). Steinmo covers the development of modern tax structures for Sweden, the US and Great Britain, breaking the discussion for each country into three parallel historical periods. The analysis is given in narrative form and

Political regimes, institutions, and the nature of tax systems 393 comparisons are incorporated into the separate treatment of each country. There is much material in the book that elucidates the different cultures and histories of tax policy formulation in the three nations. However, Steinmo’s approach does not lend itself to the derivation of specific testable hypotheses. This is not surprising if we consider his statement above that implies a need for three distinct groups of variables: those relevant to preferences and preference formation of policy-makers, those describing actual institutions, and those characterizing observed fiscal outcomes. A rather complex formal model would be required to implement a research program of this nature in a quantitative manner. As it stands, Steinmo’s work serves as a stimulating introduction to the political economy and history of taxation in the three countries. In their comparison of fiscal decision-making in Canada and the United States, Hettich and Winer (1991) adopt a more narrowly focused approach by relating differences in the transaction costs associated with policy-making to specific features in the tax systems of the two countries. As they point out, decision-making in the parliamentary system of Canada involves fewer actors and requires less time than the making of significant national tax choices in the US with its political system of checks and balances. While proposed new tax laws in Canada are prepared in secret in the Department of Finance, proposed by the Minister of Finance to Cabinet, and then introduced to Parliament, a set of steps that can be accomplished fairly quickly, major tax legislation in the US involves participation by the President, the House of Representatives, and the Senate in a process that is much lengthier, more open to negotiation by different centers of power and to public discussion, and more directly influenced by specialized interest groups. Based on the discussion of national political institutions, the authors derive several specific hypotheses about expected differences in the tax structures of the two countries. Among them are the predictions that Canada will use more discretionary fiscal policy than the US, and that its tax code will be less complex. The authors test their hypotheses with considerable success, even though the small number of relevant observations prevents them from carrying out formal statistical tests. Results are thus suggestive only, but they clearly support the notion that specific institutional features that affect the nature of policy-making will lead to identifiable differences in fiscal structure. Tsebelis (1995, 2002) uses the concept of a veto player to formalize the meaning of key actors in a political system. Veto players are individual or collective actors whose agreement is required for a change of the status quo (2002, p. 19). The number of veto players is likely an important determinant of transactions costs referred to in the analysis of Hettich

394

The Elgar guide to tax systems

and Winer.12 Even when there is a minority government, the Canadian Westminster-style parliamentary government has fewer veto players than the US congressional system. Empirical studies show that the number of veto players does affect the number and size of tax changes. Of special interest from our perspective is the work of Hallerberg and Basinger (1998), Iversen and Soskice (2006), and Ganghof (2006).13 Iversen and Soskice show that in their sample of advanced democracies, while PR leads to more redistribution, an increase in the number of veto points reduces the extent of it. Hallerberg and Basinger study tax changes in OECD countries that followed the 1986 tax reforms in the United States, when tax rates on personal income and on corporations fell. They find that an increase in veto points from one to two reduced the fall in tax rates over the period studied by 18 to 20 percent. Ganghof (2006, Chap. 8) replicates and extends the Hallerberg-Basinger study to distinguish between corporate and personal taxation. He argues that their result for corporate taxation is brittle, and that veto points had in fact no effect on the course of business taxation. On the other hand, he presents evidence showing that an increase in the number of veto points reduced the decline in top personal income tax rates. Ganghof attributes these results to the difference in the strength of international versus domestic constraints on public policy. He argues that the number of veto players is irrelevant to the choice of corporate tax structure, as all political actors’ preferences are forced to accommodate the same harsh international realities. On the other hand, the domestic political system plays a significant role in determining the personal tax structure. He does not discuss in detail why international constraints bite more deeply, and we shall return to this matter in a later section.14 The veto point framework permits empirical research to be conducted using medium-sized samples that include countries with different electoral and legislative systems. Most work of this kind deals only with how the number or size of tax policy changes over time are affected. Such a focus is a natural product of a framework that emphasizes the ability of key actors to block changes from the status quo if they are not adequately compensated. There does not appear to be any research that considers the effect of veto points on the equilibrium tax mix. Accordingly, in Table 11.5 we show the results of adding an index of veto points constructed by Tsebelis15 to the model in Table 11.2. The sample is a set of 21 advanced democracies, again for 1975–92 (see Table 11A.2 in the Appendix for a list of included countries). In Table 11.6 this is augmented with the Majoritarian indicator of electoral systems.

Political regimes, institutions, and the nature of tax systems 395 Table 11.5

Tax share regressions for central governments in 21 advanced democracies, 1975–92, # Veto Players included (absolute t-statistics in parentheses, based on Huber-White standard errors) Non-tax Corporate

# Veto −0.013 players (6.04) Adj. 0.6796 R-square Root MSE 0.0203 Number of 62 obs.

Individ. Soc. Sec. Goods & Trade Property Income & Payroll Services

−0.0069 (1.39) 0.0537

−0.00037 −0.0077 (0.03) (0.72) 0.1519 0.5334

0.019 0.0052 0.00062 (3.17) (1.50) (0.48) 0.6135 0.4330 0.2395

0.0609 62

0.1167 62

0.0582 0.0322 0.0132 62 62 62

0.1134 62

Note: Full results from model in Table 11.2 not reported. Results are similar if a scale effect (total revenue/GDP) is included. List of countries included is found in Table 11A.2. Source:

Based on re-estimation of model presented in Table 11.2 amended as indicated.

Table 11.6

Tax share regressions for central governments in 21 advanced democracies, 1975–92, # veto players and Majoritarian included (absolute t-statistics in parentheses, based on HuberWhite standard errors) Non-tax Corporate Individ. Soc. Sec. Goods & Trade Property Income & Payroll Services

# Veto −0.014 players (5.73) Majoritarian −0.0037 (0.57) Adj. 0.6739 R-Square Root MSE 0.0205 Number of 62 obs.

−0.0073 (1.30) −0.0033 (0.16) 0.0335

0.016 (1.39) 0.126 (2.77) 0.2554

−0.016 (1.21) −0.061 (1.26) 0.5401

0.015 0.0032 (2.32) (1.00) −0.033 −0.015 (1.54) (1.34) 0.6208 0.4363

−0.0019 (1.51) −0.019 (4.14) 0.4165

0.0615 62

0.1093 62

0.1125 62

0.0577 0.0321 0.0115 62 62 62

Note: Full results from model in Table 11.2 not reported. Results are similar if a scale effect (total revenue/GDP) is included. List of countries included is found in Table 11A.2. Source:

Based on re-estimation of model in Table 11.2 amended as indicated.

396

The Elgar guide to tax systems

In both tables we see that as the number of veto players increases, central governments turn towards domestic consumption taxes and away from non-tax revenues (including fees, property income, interest on investments, sales of goods and services, fines, and other items). An explanation for this pattern is not immediately obvious. On the other hand, the insignificant coefficients on the number of veto points in the corporate tax regressions is consistent with Ganghof ’s (2006) assertion that veto players are largely irrelevant to the choice of corporate tax policies. Other major elements of institutional microstructure not so far discussed include federalism, presidentialism, and bicameralism. Federalism is included in Table 11.2, where we see that central governments in federal systems (compared with unitary states) tend to rely more on corporate taxation and less on domestic trade taxes and property taxes. This seems understandable: lower levels of government find it easier to tax domestic consumption and property than mobile capital, and the opposite is the case for corporate taxation by the central government, while competition between jurisdictions leads to an efficient assignment of policy instruments in the federation.16 We must leave investigation of presidentialism and bicameralism for another occasion.17

7

EQUILIBRIUM TAX STRUCTURE IN A MATURE DEMOCRACY: THREE MOMENTS IN THE POLITICAL ECONOMY OF THE CONSUMPTION– INCOME TAX MIX

We now turn to the determinants of tax structure in mature democratic states with a stable electoral system, a simple model of which is illustrated in Figure 11.2. If we look around the world, a wide variation in the tax systems of such countries is apparent. Variation in the personal to corporate tax ratios of OECD countries, shown by the vertical lines in Figure 11.3, is just one example. Just as varied is the choice between consumption and income as sources of public finance. European Union (15) countries raised on average about 29 percent of their revenues from taxes on general consumption in 2006, while the United States and Canada derived from 14 to 22 percent of revenues respectively from this source, with much variation occurring among countries falling in between (see OECD, 2008, p. 108). Reliance on personal income taxation is similarly varied, and generally moves in the ‘opposite’ direction: excluding Social Security contributions, the US and Canada raised approximately 36 percent of total revenues in this manner,

Political regimes, institutions, and the nature of tax systems 397 16.00 14.00

Proportional

Majoritarian

2002

2004

12.00 10.00 8.00 6.00 4.00 2.00 0.00 1970

1975

1980

1985

1990

1995

2000

2001

2003

2005

Note: Bars represent averages within the categories. Vertical lines represent standard deviations. Majoritarian (t) = 1 if the lower chamber is elected by a majoritarian electoral system in year t, = 0 otherwise (proportional or mixed system, or other, in year t). Ireland and Japan (before 1994) had, respectively, transferable and non-transferable voting systems. Sources: Revenue statistics from OECD databases, www.sourceoecd.org, electoral system data from Persson and Tabellini (2003) before 1999, IDEA website, and Colomer (2005).

Figure 11.3

Evolution of the ratio of personal taxes (including Social Security and payroll) to corporate taxes, by electoral system, OECD, 1970–2005

a figure falling substantially above the average for EU 15 members, which was close to 25 percent. There is a well-known tradition of fiscal analysis relating to the choice between consumption and income taxation. (Reviews of the extensive literature are provided by Bradford, 1996 and Zodrow and McClure, 2007.) This literature is mostly normative in nature, and work on the choice between the two revenue sources when both are determined endogenously as part of a political process is sparse. The question we address in this section is this: differences in the structure of electoral institutions aside, what factors may lead to variation in the income–consumption tax ratio across democratic countries?18 In a framework appropriate to the study of this question, several key elements of both the private and public sectors must be present. Some of these factors may act directly on tax structure, while others may only exert an indirect influence. The empirical work reported in Table 11.2 and similar work by others has shown the importance of the relative size of tax bases, as measured by the per capita size of the corresponding taxable activity. Previous research has also demonstrated that the variance of incomes plays a significant role in determining income tax structure at least because the taxpaying public is heterogeneous in its tastes for leisure (Cukierman and Meltzer, 1991). The factors underlying the size of government are also involved, as tax

398

The Elgar guide to tax systems

structure may change with the overall role of government in the economy. We referred to this earlier as the scale effect. Work on the size of government has identified per capita income as a significant determinant (see the literature on Wagner’s Law, reviewed by Mueller, 2003), so here is another reason to add this factor to the list. The median voter theorem of Black (1958) points to the skewness of the income distribution as a key determinant of government size (Meltzer and Richard, 1981, 1983). Thus we can point to the first three moments of the distribution of income as potential determinants of the tax mix in mature democracies. In a different but related context, Usher (1977) has drawn attention to the heterogeneity of preferences for public goods, including both second and third moments of the distribution of tastes, as a determinant of whether a commodity is provided publically or not. Lower-income citizens especially are faced with a trade-off between the benefits of redistributive financing of publically provided goods (as in the Meltzer-Richards framework), and the loss of welfare that may occur when their tastes differ substantially from others who are involved in making collective decisions. Preferences for public goods relative to private goods and leisure may also be implicated directly, as well as indirectly via the scale effect, in the evolution of tax structure. The nature of such preferences may affect the way in which individual taxpayers at different places in the income distribution react to changes in different types of taxation. Thus, in moving beyond the existing literature, it seems sensible to explore the first three moments – the mean, variance, and skewness – of the distributions of income and tastes for public goods, as factors that may shape the consumption–income tax mix.19 Studying the Role of the Three Moments in the Consumption–Income Tax Mix Using Simulation Investigating the role of the three moments identified above is a complex task. Winer et al. (2009) use simulation of a spatial voting model to examine the role of these factors. As is now well known, such a model can be solved for its equilibrium by optimizing a synthetic (political support) function, which is weighted sum of indirect utilities, with weights reflecting the effective political influence of each group of citizens (see, for example, Coughlin, 1992). In the Winer et al. model, utility is Cobb-Douglas in one pure public good, one private good, and leisure. Heterogeneity of tastes for the public good involves assumptions about the mean, variance, and skewness of parameters in the utility functions. A distribution of skills and thus of wages and incomes is introduced via the budget constraints of citizens, which also include an exogenous capital income for each person.

Political regimes, institutions, and the nature of tax systems 399 Table 11.7

A homogeneous (in tastes) society, model mean skill = 850, change in variance of skills

Model Variance 10 000 20 000 30 000 40 000 50 000 Source:

tl

tc

tl/tc

gsize

0.22 0.23 0.24 0.25 0.27

0.88 0.85 0.84 0.82 0.79

0.25 0.27 0.29 0.30 0.34

0.54 0.54 0.54 0.54 0.54

Winer et al. (2009).

A distribution of political influence is introduced through assumptions about the weights on indirect utilities in the political support function. There are five groups of voters with population weights symmetrical about the mean, and taxes on the wage income and consumption of these groups are assumed to be strictly proportional. The model is calibrated so that the relative size of the public sector is 0.5 and the tax on wages and on consumption are both non-negative. Further details are provided in their paper. Simulations of this model for societies with symmetric and asymmetric distributions of income and tastes show that the tax mix may vary substantially even when the distribution of political influence is held fixed. Here we consider as an illustrative case the effects of changes in the variance of skills, since this is not usually investigated. In this case, mean skill is held constant and tastes are homogeneous across the income groups. Table 11.7 records the simulated values of tax rates on labor and consumption tl and tc , the tax ratio tl/tc , and the relative size of government, gsize. (In this model, tax rates and revenues move together and only rates are reported here.) We see that as the model equivalent of the variance of skills falls, tl falls and tc rises, with government size relative to aggregate income remaining more or less the same. There is, thus, an increase in relative reliance on consumption taxation as the distribution of skills becomes more concentrated. These results are foreshadowed by Cukierman and Meltzer (1991) to whose work we referred earlier. In their median voter model, a decrease in income inequality reduces the size of government because total taxable income, the only base in their model, shrinks as a result of the declining importance of differences in the elasticities of labor supply of the rich and the poor as the distribution of incomes is reduced. Here, in a model with a multi-dimensional issue space (two tax rates and one public good), we see changes in the tax mix even though the size of the public sector does not change.

400

The Elgar guide to tax systems

1 mean = 500

850

1200

0.5

0 tl

tc

Source: Winer et al. (2009).

Figure 11.4

A homogeneous society with minimal skill variance

It is useful to examine the change in tax structure further by now imposing a near zero dispersion in skills and hence in income, so that there are no rich or poor relative to the mean.20 The results, illustrated graphically in Figure 11.4, shows three cases in which the (model) mean of skills declines from 1200 to 500 while the distribution of skills remains very highly concentrated. The population with low mean skills has relatively low income, both earned and from its endowments. Then the elasticity of leisure with respect to income is relatively high (see equation 2.2b in Winer et al., 2009), and a tax on labor income represents a strong disincentive to work for everyone. It is therefore optimal when maximizing expected votes to have a lower tl. In the case of low mean skills, this leads to a corner solution where there is virtually no labor income taxation at all (tl = 0). What is happening at this corner is that labor income is taxed as little as possible to make gross incomes as high as possible, while desired public services are then financed relatively efficiently by diverting private expenditure into the public purse, a use of the consumption tax that is not commonly considered. Finally, we note that when the mean taste for public goods is increased in any of the experiments outlined above, consumption taxation tends to grow in importance in the tax mix for any given skill variance. A general lesson from the model is that the distribution of tastes for public and private goods interacting with the distribution of income may influence tax structure by affecting the elasticity of tax bases and the efficient means of financing public services. While the specific nature of this interaction depends on the particular model structure adopted, it seems reasonable to expect some adjustments of this sort when there is substantial variation in the distributions of income and of tastes. It is always a good idea to remain cognizant of how assumptions drive conclusions in the sort of simulation model we have briefly explored.21 Still, in the light of our experience with the simulation model, we may

Political regimes, institutions, and the nature of tax systems 401 speculate that the relatively high reliance on consumption taxation in the European Union compared with North America may be partly due to greater income equality coupled with a stronger preference for public goods. Whether or not, and the extent to which this may be so has not been considered in the empirical literature.

8

GLOBALIZATION AND TAX COMPETITION

Tax competition exists and is a factor reshaping tax systems – see the reviews of work on this matter by Devereux (2008), Feld and Heckmeyer (2008), and Zodrow (2008) – even if there is no obvious withering away of the state so far (Garrett, 1998; Swank and Steinmo, 2002; Ha, 2007; Bergh, 2008; Bernholz, 2008).22 In this last substantive section we consider how one might study the consequences of globalization and international tax competition in the context of the basic model outlined earlier. At one level, modeling the effects of globalization is straightforward. To make use of a model like the one illustrated in Figure 11.2, one needs to know which tax bases become relatively more elastic, and how the economic consequences of this for a specific rate–revenue relationship are translated into political costs of raising revenue from that base. Then the consequences for the tax mix follow in a straightforward manner. There are several complications to deal with before one arrives at that point, however. First, the elasticities involved are equilibrium quantities being, to some extent, the result of strategic interaction between national governments and even of states or provinces that act on their own to attract foreign investment. Economists have been on the track of these tax elasticities empirically (see Feld and Heckmeyer, 2008 for references). Second, how changes in elasticities translate into political support is even harder to study. Though they may not refer to their enterprise in these terms, political scientists have been working on this part of the story (see, for example, Garrett 1998 and Ha, 2007) by investigating how domestic political institutions attenuate the consequences of the increasing international integration of national economies. Third, acknowledging that tax structures are systems of related elements in this context remains a substantial challenge in its own right. This third aspect of the study of globalization leads to several thorny issues and problems confronting researchers working on the empirical political economy of taxation. One of the most interesting of these is that even though tax structures are systems of related elements, strategic interaction tends to be studied tax by tax, such as with respect to corporate taxation. It is necessary to allow for the fact that shocks anywhere

402

The Elgar guide to tax systems

in the system may propagate everywhere. Indeed, there is no reason to limit international competition to that over taxation, since there are other types of policy instruments that can often serve as good substitutes (Sinn, 2003).23 A second problem stems from the fact that since tax structures are systems, big shocks that are common across countries influence much of the tax structure of all the affected countries, and often in a similar manner. For this reason, it is difficult to distinguish a situation in which there are similar, autonomously arrived at reactions in several countries to a common shock, from a situation in which countries strategically interact (on this point, see Revelli, 2005). As a final comment, we return to the question of why there is a difference between domestic strategic interaction among taxpayers and the government, on the one hand, and strategic interaction between countries when international migration as opposed to ‘internal migration’ between political parties is possible, on the other. To recall, this difference played a key role in some of the research on veto points we considered, and surely is an important part of any study of international competition. One sensible (and often provided) answer is that except for highly skilled people who supply internationally traded services, companies can move across national borders more easily than can labor in response to favorable policy differentials. But while this answer may appear adequate, It is important to keep in mind that the difference in mobility of factors is to a considerable extent a product of public policy. Governments can and do regulate the mobility of labor and capital, both independently and through bilateral and multilateral agreements. Labor mobility is lower and that of capital increasing because of deliberate policy choices. It seems sensible to suspect, therefore, that further work on the difference between domestic and international constraints on policy choices may lead to new perspectives on both the positive and normative study of fiscal systems.

9

CONCLUDING REMARKS

The comparative analysis of fiscal systems raises many significant questions for the study of political economy. Foremost among them are those concerned with the influence of institutions. In this chapter, we pay particular attention to the impact of political institutions on tax structure, though taxation is not determined in isolation from other aspects of public policy. After briefly presenting a model suitable for explaining fiscal outcomes as a political equilibrium, we examine some of the relevant literature, while

Political regimes, institutions, and the nature of tax systems 403 also reporting on new statistical results. Our survey is selective rather than comprehensive, focusing on what we consider to be some of the important issues, results, and outstanding questions in a growing body of literature. At the broadest level, international comparisons across all types of regimes show that the degree of democracy is an important influence on tax structure, with more democratic countries appearing to rely more heavily on personal income taxes and Social Security and payroll taxes even after controlling for the determinants of tax structure that may be correlated with the level of development. This suggests that the degree of consent may play a vital role in the nature of observed revenue structures. When we turn to comparative analysis among modern democracies, a major institutional feature of the electoral system – whether they use proportional or majoritarian electoral systems – is of particular relevance. Again we find that the institutional setup matters. Our statistical work shows that countries using PR make heavier use of Social Security, payroll, and domestic trade taxes and rely less on individual income, corporate, and international trade taxation than nations using a majoritarian electoral system. While it is possible to suggest explanations for some of these effects, further research will be needed to confirm them. When we examine the comparative microstructure of revenue systems, further questions arise. Particularly relevant at this level is the influence of transaction costs, or in a related formulation, the role of veto players who can block or delay relevant policy decisions. We present some new results on the impact of veto points on the tax mix, a topic not so far explored in the literature. Research on fiscal structure is complex since the analyst is faced with equilibrium systems where all significant policy instruments are related. It is not surprising therefore that most empirical research is confined to data that consists of average values, since there are multiple challenges to estimation even without the consideration of higher moments of the relevant variables. New work using simulation techniques points, however, to influences that can be captured only with higher moments. It suggests, for example, that changes in the variance of incomes interacting with the distribution of tastes for public goods may result in significant differences over time or across countries in the use of revenue sources. The quantitative importance of second and third moments relative to that of the institutional factors we have considered remains to be investigated. Comparative analysis offers a rich field for research on tax systems. It also presents significant challenges because observed data reflect equilibrium outcomes, requiring estimation techniques taking account of the simultaneity of different policy features. Various parts of revenue systems

404

The Elgar guide to tax systems

– what we have called the tax skeleton – are all related and cannot be studied properly in isolation. We started the chapter by asking, among other things, whether the literature in this field offers some lessons for current tax design. As we noted earlier, the analysis clearly suggests that institutions matter. Reform proposals that do not take account of the institutional context, should they be implemented, will have significant and unforeseen side-effects in other parts of the fiscal system. Further work of the type reviewed here will allow us to better predict such effects, as well as lead us to a better understanding of why we observe such wide variation in revenue systems across countries in both the developed and the developing parts of the world.

NOTES *

1.

2.

3. 4. 5. 6.

Authors are listed in reverse alphabetical order. Prepared for ‘Tax Systems: Whence and Whither (Recent Evolution, Current Problems and Future Challenges), a conference sponsored by the Savings Banks Foundation of Spain (FUNCAS) and UNICAJA, Malaga, Spain, 9–11 September, 2009. We are grateful to Sophie Bernard for research assistance. We thank our discussant at the conference, Julio Lopez Laborda, for his thoughtful remarks, as well as conference participants. The paper also served as the basis for a general address by Winer at the XXI Riunione Scientifica Siep (Italian Public Finance Society), Pavia, 24 September, 2009, with Silvia Giannini as discussant. In addition to the challenge of explaining why contemporary tax structures vary across space and time, which is our concern in this chapter, other major questions in the political economy of taxation include: (i) what determines the power to tax, (ii) how and why do observed tax structures differ from optimal tax systems, and (iii) how can we understand the process of tax reform? For recent contributions concerning (i), see Besley and Persson (2009) and North et al. (2009). On (ii) see, for example, some of the background papers and comments for the Commission on Reforming the Tax System (the Mirrlees Review). On (iii) see, Winer and Hettich (1999) and Hettich and Winer (1999, Chap. 6) and the references therein. The practical division between regime types using the Gastil indexes is based on Islam and Winer’s (2004) test of Wintrobe’s model. North et al. (2009) consider both of these types to be what they call natural states, which are ruled by elites and may be stable, but in which the rule of law does not extend to elites and where interference by elites in the ‘democratic’ process often occurs. This is the case if all regressions contain exactly the same set of right-hand side variables, and the left-hand side variables exactly sum to 1 (the sum of revenue shares). See Bodkin (1964) for explanation of this property of the estimation. For a study of the interaction of two elements – an income tax rate and a special provision in the US state income tax system – see Winer and Hettich (1993). Profeta et al. (2009) find a similar result – more freedom leads to greater reliance on personal income taxation – but for a sample (only) of developing countries, and using only the civil liberty component of the Gastil index. In his comment on this chapter, Julio Laborda points out that the degree of redistribution will depend on the size of the public sector and the relative importance of income taxation, as well as the steepness of the income tax schedule. Substantial redistribution could occur by passing a suitably large part of the national income flow through a tax

Political regimes, institutions, and the nature of tax systems 405

7. 8. 9. 10.

11. 12.

13. 14.

15. 16. 17.

18.

system with an income tax schedule that is not very steep, especially if the income tax plays a relatively important role in the tax system as a whole. The puzzling insignificance of the second segment spline coefficient on trade taxes is, perhaps, due to corruption in customs collection. For a discussion of the role of consent, and its opposite – coercion – in taxation, see for example Levi (1981), Breton (1996), and Winer et al. (2008). A useful survey of research up to 2002 is provided by Baker and Gould (2002). See also Hettich and Winer (1999). In the Iversen-Soskice model, there are three groups of voters – lower-income, middle- and higher-income citizens – and there is a constraint on all governments that prevents the introduction of regressive policies. (This constraint may originate in the threat of political instability, and is not explained.) Tax policy is multi-dimensional and the median voter theorem does not apply. Under PR, there are three parties – L, M, and H – each of which is a perfect representative of the corresponding income group. Governments consist of a coalition of two of these parties. Given the nonregressivity constraint, the decisive middle-income party, M, will likely join with L rather than with H. If M joins L, it can benefit with L from progressive taxation of H. If it joins with H, the non-regressivity constraint prevents the exploitation of L. In other words, M has to share with both L and H in an MH coalition, and only with L in an LM coalition. In a Majoritarian system, the governing party is a coalition of two groups and will have a centrist platform. Since the government is a coalition of different groups, its leadership may not be a good representative of its constituents. In that case, the decisive M voters prefer to join an MH party rather than an LM party. Although both parties will promise a centrist platform, the leadership of the LM party may waffle to the left, in which case M voters are at risk of being taxed to pay for redistribution to lower-income voters. On the other hand, in view of the nonregressivity constraint, if the MH party waffles to the right, M can be made no worse off, and the best option for an H-dominated MH party will be to reduce the size of the public sector leaving H (and everyone else) with more disposable income. For an alternative model with a similar focus on why post-fisc inequality is less under PR, see Austen-Smith (2000). The analysis relies in part on the varieties of capitalism framework of Hall and Soskice (2001). The ideological distance between veto players may also play a role, as a greater distance generally makes it more difficult to agree on changes from the status quo. Thus increasing elite polarization in the United States may also be important in explaining the Hettich-Winer findings. On polarization in the US see, for example, the recent survey by Fiorina and Abrams (2008). Stewart (1991) is an early paper on tax US reform in the 1981 to 1987 period that can be seen, with hindsight at least, as being consistent with, or involving elements of, a veto player analysis, as may the analysis of Hettich and Winer. In a complementary study, Ha (2007) argues that empirical evidence from a similar data set of industrialized countries from 1960 to 2000 indicates that the pressure of globalization to increase domestic welfare spending is reduced if there are more partisan or institutional veto players. See http://sitemaker.umich.edu/tsebelis/veto_players_data; accessed 2 May 2011. On the assignment of policy instruments in a federal state, see, for example, Breton and Scott (1978). When a Presidential indicator replaces Majoritarian in Tables 11.3 and 11.4, presidentialism leads to less reliance on personal income taxation and more reliance on Social Security and payroll taxation, as under PR. We have no obvious explanation for this pattern. For work of differing sorts on partisanship and taxation see, for example, Stewart (1991) on US national tax policy in the 1980s and Reed (2006) on tax policy in US states.

406 19.

20. 21.

22.

23.

The Elgar guide to tax systems The second and third moments of the distribution of political influence will also be important, since tax policy is a result of collective choice. (The first moment or average degree of political influence has no meaning since influence is a relative concept.) We note that partisan party politics, including differences in the ideology of redistribution, is not involved here, and is left for further research. The presence of some distribution is essential, however small the dispersion, as otherwise there is no equilibrium in the model for a small countable number of voters. See, for example, Renstrom (1996) who adopts a quasi-linear specification of preferences instead of Cobb-Douglas, producing a model in which an increase in tastes for public goods leads to more labor income taxation rather than less. A general message of the same kind is found in Deaton (1987). Garrett (1998) makes a good case that the importance of the state has increased as a result of globalization, particularly in countries with left of center coalition governments (and proportional representation) and encompassing labor market institutions. Rutherford and Winer (1990) and Hettich and Winer with Rutherford (1999, Ch. 7) illustrate the connection between domestic politics in the US and globalization in the 70s in a unique fashion. They construct a spatial voting model of the sort employed in this chapter, with the GEMTAP tax model embedded in it. The influence weights of each of three income groups in the political support function (the optimization of which replicates the equilibrium) are calibrated so that the model reproduces the GEMTAP benchmark data sets in 1973 and 1983. Then, applying the 1973 effective influence weights of the poor (without capital income), middle- and high-income groups to the 1983 data set leads to an equilibrium tax on capital that is substantially lower and a tax on labor income that is higher than actually observed in 1983. This shows that to replicate the 1983 data set, what is required is an increase (compared with 1973) in the relative domestic political influence of voters who favor lower taxation of labor income. The Trebilcock-Prichard-Hartle-Dewees (1982) study of substitutability of governing instruments (expenditures, taxes, regulation, law and so on) is, perhaps, the seminal work on systems of governing instruments. Sinn (2003) analyzes competition between national systems of policy instruments. A few authors have recently explored aspects of the connection between different types of policy instruments. Among them are Fredriksson et al. (2004), Hettich and Winer (2006) who include regulation as well as taxation, and Hauptmeir et al. (2008). Gordon (2010) explores the importance of the connection between regulation of the informal sector and taxation of the formal sector in developing countries.

BIBLIOGRAPHY Adams, James, Samuel Merrill and Bernard Grofman (2005), A Unified Theory of Party Competition, New York: Cambridge University Press. Aidt, Toke S. and Peter S. Jensen (2009), ‘Tax Structure, Size of Government, and the Extension of the Voting Franchise in Western Europe, 1860–1938’, Journal of Public Economics, 93, 160–75. Austen-Smith, David (2000), ‘Redistributing Income under Proportional Representation’, Journal of Political Economy, 108(6), 1235–69. Baker, Andrew C. and Peter J. Gould (2002), ‘Democracy and Taxation’, Annual Review of Political Science, 5(1), 87–110. Bergh, Andreas (2008), ‘A Race to the Bottom for the Big Welfare States?’, in A. Bergh and R. Hoeijer, Institutional Competition, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, 182–201. Bernholz, Peter (2008), ‘Institutional Competition: International Environment, Levels and Consequences’, in A. Bergh and R. Hoeijer, Institutional Competition, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp 90–111.

Political regimes, institutions, and the nature of tax systems 407 Besley, Timothy and Torsten Persson (2009), ‘The Origins of State Capacity: Property Rights, Taxation, and Politics’, American Economic Review, 99(4), 1218–44. Black, Duncan (1958), The Theory of Committees and Elections, New York: Cambridge University Press. Bodkin, Ronald G. (1964), ‘Additively Consistent Relationship for Personal Savings and the Categories of Consumption Expenditures, U.S.A., 1949–1963’, Cowles Foundation Paper No. 403, Yale University. Bradford, David F. (1996), Fundamental Issues in Consumption Taxation, Washington, DC: American Enterprise Institute. Breton, Albert (1996), Competitive Governments: An Economic Theory of Politics and Public Finance, Cambridge: Cambridge University Press. Breton, Albert and Anthony Scott (1978), The Economic Constitution of Federal States, Toronto: University of Toronto Press. Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson and James D. Morrow (2005), The Logic of Political Survival, Cambridge, MA: MIT Press. Canegrati, Emanuele (2009), ‘A Probabilistic Voting Model of Indirect Taxation’, Catholic University of Milan, unpublished. Castanheira, Michele, Gaetan Nicodème, Paola Profeta and Viktor Trasberg (2009), ‘On the Political Economics of Tax Reform’, paper presented to the Societa Italiana di Economia Pubblica, Pavia September. Colomer, Josep M. (2005), ‘It’s Parties That Choose Electoral Systems (or, Duverger’s Laws Upside Down)’, Political Studies, 53, 1–21. Coughlin, Peter (1992), Probabilistic Voting Theory, Cambridge, UK: Cambridge University Press. Cukierman, Alex and Allan H. Meltzer (1991), ‘A Political Theory of Progressive Income Taxation’, in A. Meltzer, A. Cukierman and S.F. Richard (1991), Political Economy, Oxford: Oxford University Press, pp. 76–108. Cusack, Thomas, R. and Pablo Beramendi (2006), ‘Taxing Work’, European Journal of Political Research, 45(1), 43–73. Deaton, Angus (1987), ‘Econometric Issues for Tax Design in Developing Countries’, in D. Newbery and N. Stern (eds), The Theory of Taxation for Developing Countries, Oxford: Oxford University Press, pp. 92-113. Devereux, Michael (2008), ‘Business Taxation in a Globalized World’, Oxford Review of Economic Policy, 24(4), 625–38. Feld, Lars P. and Jost H. Heckemeyer (2008), ‘FDI and Taxation: A Meta-Study’, Discussion Paper No. 08-128, ZEW Center for European Economic Research. Fiorina, Morris P. and Samuel J. Abrams (2008), ‘Political Polarization in the American Public’, Annual Review of Political Science, 11(1), 563–88. Fredriksson, Per G., John List, and Daniel L. Millimet (2004), ‘Chasing the Smokestack: Strategic Policymaking with Multiple Instruments’, Regional Science and Urban Economics, 34(4), 387–410. Freedom House (various years), Freedom in the World, also at www.freedomhouse.org; accessed 2 May 2011. Galli, Emma and Paola Profeta (2009), ‘Tax Complexity with Heterogeneous Voters’, Heterogeneity and Public Policy, a Special Issue of Public Finance and Management. Ganghof, Steffen (2006), The Politics of Income Taxation. A Comparative Analysis, ECPR Monographs Series, Colchester: ECPR Press. Geys, Benny and Federico Revelli (2009), ‘Decentralization, Competition and the Local Tax Mix: Evidence from Flanders’, Department of Economics Working Papers No. 200902, University of Turin. Garrett, Geoffrey (1998), Partisan Politics in the Global Economy, Cambridge, UK: Cambridge University Press. Gordon, Roger (2010), ‘Public Finance and Economic Development: Reflections based on Experience in China’, Journal of Globalization and Development, 1(1), article 7.

408

The Elgar guide to tax systems

Gourevitch P. and M. Haves (2002), ‘The Politics of Choice Among National Production Systems’, L’Année de la Régulation, 6, 241–70. Ha, Eunyoung (2008), ‘Globalization, Veto Players and Welfare Spending’, Comparative Political Studies, 41(6), 783–812. Hall, Peter A. and David Soskice (2001), ‘Introduction’, in P. Hall and D. Soskice (eds) Varieties of Capitalism, Oxford: Oxford University Press. Hallerberg, Mark and Scott Basinger (1998), ‘Internationalization and Changes in Tax Policy in OECD Countries: The Importance of Domestic Veto Players’, Comparative Political Studies, 31(3), 321–53. Hauptmeir, Sebastien, Ferdinand Mittermaier and Johannes Rincke (2008), ‘Fiscal Competition over Taxes and Public Inputs: Theory and Evidence’, CESifo Working Paper No. 2499, December. Hettich, Walter and Stanley L. Winer (1988), ‘Economic and Political Foundations of Tax Structure.’ American Economic Review, 78(4), 701–12. Hettich, Walter and Stanley L. Winer (1991), ‘Checks and Balances and the Structure of Taxation in the United States & Canada’, in A. Breton, G. Galeotti, P. Salmon and R. Wintrobe (eds), The Competitive State, Hingham, MA: Kluwer Academic Publishers, pp. 39–56. Hettich, Walter and Stanley L. Winer (1999), Democratic Choice and Taxation: A Theoretical and Empirical Analysis, New York: Cambridge University Press. Hettich, Walter and Stanley L. Winer (2006), ‘Analyzing the Interdependence of Regulation and Taxation’, Public Finance Review, 34(4), 1–26. Hinich, Melvin J. and Michael C. Munger (1997), Analytical Politics, New York: Cambridge University Press. IDEA (International Institute for Democracy and Electoral Assistance), available at: http:// www.idea.int; accessed 3 May 2011. Islam, Muhammed and Stanley L. Winer (2004), ‘Tinpots, Totalitarians (and Democrats): An Empirical Investigation of the Effects of Economic Growth on Civil Liberties and Political Rights’, Public Choice, 108, 289–323. Iversen, Torben and David Soskice (2006), ‘Electoral Institutions and the Politics of Coalitions: Why Some Democracies Redistribute More than Others’, American Political Science Review, 100(2), 165–81. Kenny, Lawrence and Stanley L. Winer (2006), ‘Tax Systems in the World: An Empirical Investigation into the Importance of Tax Bases, Administration Costs, Scale and Political Regime’, International Tax and Public Finance, 13(2), 181–215. Levi, Margaret (1981), Of Rule and Revenue, Berkeley and LA: University of California Press. Lindert, Peter (1994), ‘The Rise of Social Spending: 1880–1930’, Explorations in Economic History, 3(1), 1–37. Lindert, Peter H. (2004), Growing Public: Volume 1, The Story: Social Spending and Economic Growth since the Eighteenth Century, Cambridge: Cambridge University Press. Lindert, Peter H. (2009), Growing Public: Volume 2, Further Evidence: Social Spending and Economic Growth since the Eighteenth Century, Cambridge: Cambridge University Press. McGuire, Martin C. and Mancur Olson Jr. (1996), ‘The Economics of Autocracy and Majority Rule: The Invisible Hand and the Use of Force’, Journal of Economic Literature, 34(1), 72–96. Meltzer, Allan H. and Scott F. Richard (1981), ‘A Rational Theory of the Size of Government’, Journal of Political Economy, 89(5), 914–27. Meltzer, Allan H. and Scott F. Richard (1983), ‘Tests of a Rational Theory of the Size of Government’, Public Choice, 41(3), 403–18. Mueller, Dennis (2003), Public Choice III, Cambridge: Cambridge University Press. Mulligan, Casey B., Ricard Gil and Xavier Sala-i-Martin (2004), ‘Do Democracies Have Different Public Policies than Nondemocracies?, Journal of Economic Perspectives, 18(1), 51–74. Musgrave, Richard A. (1969), Fiscal Systems, New Haven, CT: Yale University Press.

Political regimes, institutions, and the nature of tax systems 409 North, Douglas C., John Wallis and Barry R. Weingast (2009), Violence and Social Order: A Conceptual Framework for Interpreting Recorded Human History, New York: Cambridge University Press. OECD (2008), Revenue Statistics, 1965–2007. Paris: OECD. OECD (2006a), Consumption Tax Trends: VAT/GST and Excise Rates, Trends and Administration Issues. Paris: OECD. OECD (2006b), Fundamental Reform of Personal Income Tax, Tax Policy Studies No. 13. Paris: OECD. Persson, Torsten and Guido Tabellini (2000), Political Economics: Explaining Economic Policy, Cambridge, MA: MIT Press. Persson, Torsten and Guido Tabellini (2003), The Economic Effects of Constitutions, Cambridge, MA: MIT Press. Profeta, Paola and Simona Scambrosetti (2010), The Political Economy of Taxation: Lessons from Developing Countries, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Profeta, Paola, Riccardo Puglisi and Simona Scabrosetti (2009), ‘Does Democracy Affect Taxation? Evidence from Developing Countries’, paper presented to the Societa Italiana di Economia Pubblica, Pavia, September. Renstrom, Thomas (1996), ‘Endogenous Taxation: An Overlapping Generations Approach’, Economic Journal, 106(435), 471–82. Report of a Commission on Reforming the Tax System for the 21st Century (the Mirrlees Review) (2010/11), at http://www.ifs.org.uk/mirrleesReview/publications; accessed 3 May 2011. Revelli, Frederico (2005), ‘On Spatial Public Finance Empirics’, International Tax and Public Finance, 12(4), 475–92. Rutherford, Thomas and Stanley L. Winer (1990), ‘Endogenous Policy in a Computational General Equilibrium Framework’, reprinted in S.L. Winer, Political Economy in Federal States: Selected Essays of Stanley L. Winer, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing, 2002, pp. 285–324. Rutherford, Thomas and Stanley L. Winer, with Walter Hettich (1995), ‘Tax Policy in a Computable Model of Economic and Political Equilibrium’, Chapter 7 in Hettich and Winer (1999). Schofield, Norman and Itai Sened (2006), Multiparty Democracy: Elections And Legislative Politics, Cambridge, UK: Cambridge University Press. Sinn, Hans-Werner (2003), The New Systems Competition, New Haven, CT and Oxford: Blackwell. Steinmo, Sven (1993), Taxation and Democracy: Swedish, British and American Approaches to Financing the Modern State, New Haven, CT: Yale University Press. Stewart III, Charles H. (1991), ‘The Politics of Tax Reform in the 1980’s’, in A. Alessina and G. Carliner, Politics and Economics in the Eighties, Chicago: University of Chicago Press, pp. 143–73. Swank, Duane and Sven Steinmo (2002), ‘The New Political Economy of Taxation in Advanced Capitalist Democracies’, American Journal of Political Science, 46(3), 642–55. Trebilcock, M.J., R.S. Prichard, D.G. Hartle and D.N. Dewees (1982), The Choice of Governing Instrument, Ottawa: Queen’s Printer for the Economic Council of Canada. Tsebelis, George (1995), ‘Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarianism, Multicameralism, and Multipartyism’, British Journal of Political Science, 25(33), 289–326. Tsebelis, George (2002), Veto Players: How Political Institutions Work, Princeton, NJ: Princeton University Press. Tsebelis, George, http://sitemaker.umich.edu/tsebelis/veto_players_data; accessed 13 May 2011. Usher, Dan (1977), ‘The Welfare Economics of the Socialization of Commodities’, Journal of Public Economics, 8(2), 1977, 151–68. Winer, Stanley L. and Walter Hettich (1993), ‘Explaining the Use of Related Tax Instruments’,

410

The Elgar guide to tax systems

in S.L. Winer, Political Economy in Federal States: Selected Essays of Stanley L. Winer, Cheltenham, UK and Northampton, MA, USA: Edward Elgar 2002, pp. 19–39. Winer, Stanley L. and Walter Hettich (1999), ‘Information, Co-ordination and Tax Policy Making’, in R. Bird and T. Wilson (eds), Rationality in Public Policy: Retrospect and Prospect; A Tribute to Douglas Hartle, Toronto: Canadian Tax Foundation, pp. 19-42. Also at www.carleton.ca/~winers; accessed 1 May 2011. Winer, Stanley L., George Tridimas and Walter Hettich (2008), ‘Social Welfare and Coercion in Public Finance’, CESifo Working Paper No. 2482, revised June 2009. Winer, Stanley L., George Warskett and Walter Hettich (2009), ‘Consumption versus Income Taxation: Three Moments in the Political Economy of Taxation’, in S.L. Winer (ed.), Heterogeneity and Public Policy. A Special Issue of Public Finance and Management, Part II, 9(3). Wintrobe, Ronald (1998), The Political Economy of Dictatorship, New York: Cambridge University Press. Woldemariam, Asegedech (1995), ‘Summary Tax Structure Tables, 1975–92’, in Parthasarathi Shome (ed.), Tax Policy Handbook, Washington, DC: International Monetary Fund. Zodrow, George (2008), ‘Capital Mobility and Capital tax Competition’, paper presented to the conference on ‘Mobility and Tax Policy’, University of Tennessee, 22–23 October, 2008. Zodrow, George and Charles E. McClure, Jr (2007), ‘Consumption-Based Direct Taxes: A Guided Tour of the Amusement Park’, Finanzarchiv, 63(2), 285–307.

Political regimes, institutions, and the nature of tax systems 411

APPENDIX Table 11A.1

Descriptive statistics, countries with a combined Gastil index of 5 or less, 1975–92

Majoritarian Presidential Non-tax Corporate Individual income Social Security & Payroll Goods & Services Trade Property

Mean

Standard Deviation

Minimum

Maximum

0.388 0.252 0.122 0.106 0.193

0.487 0.436 0.075 0.108 0.135

0 0 0.021 0 0

1 1 0.460 0.577 0.563

0.185

0.156

0

0.540

0.240

0.124

0.0096

0.500

0.142 0.018

0.177 0.016

0 0

0.681 0.067

Note: Majoritarian (t) = 1 if lower chamber is majoritarian in year t; = 0 otherwise (PR, Mixed PR or Other in a few cases). Presidential (t) = 1 if there is a presidential system in year t; = 0 otherwise. Classification follows Persson and Tabellini (2003). Majoritarian in the regressions is defined as the average value of the dummy variable over each sample period employed. Gastil index is an average of the Gastil values for the years in a time period in which we have values for it. See Table 11A.2 for list of countries. Source:

Authors.

412

The Elgar guide to tax systems

Table 11A.2

List of countries with a combined Gastil index of 5 or less, 1975–92

Country Argentina Australia Austria Bahamas Barbados Belgium Bolivia Botswana Canada Colombia Denmark Dominican Republic Ecuador Fiji Finland France Gambia Greece Iceland India Ireland Israel Italy Jamaica Japan Luxembourg Mauritius Netherlands New Zealand Norway Papua New Guinea Portugal Solomon Islands Spain Sri Lanka Sweden Switzerland United Kingdom USA Venezuela West Germany

In OECD?

Included with Veto Players Data?

OECD OECD

Veto players data Veto players data

OECD

Veto players data

OECD

Veto players data

OECD

Veto players data

OECD OECD

Veto players data Veto players data

OECD OECD

Veto players data

OECD

Veto players data

OECD

Veto players data

OECD OECD

Veto players data Veto players data

OECD OECD OECD

Veto players data Veto players data Veto players data

OECD

Veto players data

OECD

Veto players data

OECD OECD OECD OECD

Veto players data Veto players data Veto players data

OECD

Veto players data

Note: OECD countries not in subsample: Czech Republic, Hungary, Korea, Mexico, Poland, Slovak Republic, Turkey. Full list of countries used in Table 11.2 is found in Kenny and Winer (2006). Source: Authors.

12 Tax system change and the impact of tax research* Richard M. Bird

The three decades between the end of World War II and the first oil crisis of the mid-1970s have sometimes been called les trentes glorieuses, the thirty glorious years – years that, at least when viewed with the rosy nostalgia of hindsight, produced growth and increasing prosperity for all. In reality the 1945–75 period was more nuanced and differentiated than this label suggests, and it was doubtless far from ‘glorious’ for many, not least in the very heterogeneous group of ‘non-OECD’ (or ‘developing’) countries. It is no surprise that the subsequent three decades that are the focus of this chapter have also produced mixed results, not least in the fiscal sphere. Of course what one sees always depends not only on where one sits but also on precisely where (and when) one looks – as well as on what one is looking for. What I am looking for in this chapter is essentially evidence that recent tax policy has been influenced by tax research. Or, to put it another way, what evidence is there that tax researchers have been focusing on the problems that really shape tax policy? Most tax research is carried out in the OECD countries that are the principal focus of this volume, and the issues discussed here are of course relevant to those countries. For the most part, however, I focus on developing (non-OECD) countries. I do so for three reasons. First, since in the era of globalization we are all in a real sense in the same boat, it may be useful even for those whose primary concerns are for a particular developed country to reflect in more general terms on how tax issues arise and are dealt with around the world. Second, since both the fiscal problems and the concerns about the relevance of much economic research on taxation are greater in developing countries, the question raised above comes out most sharply in this context.1 Finally, the dominance of ‘first world’ economists both in shaping tax research and in giving advice on tax policy to developing countries, suggests that experience in the non-OECD world may provide a particularly clear test of whether that advice makes sense. Have we been giving good advice? Has anyone been listening? Are there important issues to which we should pay more attention if we want tax research to contribute more to tax system improvement? How may tax 413

414

The Elgar guide to tax systems

researchers in any country, OECD or developing, be more successful than experience suggests they have been to date not only in ‘speaking truth to power’ (Wildavsky, 1979) – but also in being heard?

1

TRENDS IN TAXATION

How countries tax themselves changes continuously, as Heady (2010) has recently discussed in detail with respect to OECD countries and as Norregaard and Khan (2007) review in broader perspective. The world changes, and so taxes. New research may also change our understanding and ideas about what constitutes a good tax system. At first glance, it may thus be surprising to learn that neither tax levels nor (in broad perspective) tax structures in developing countries look all that different than they did 30 years ago. This outcome seems especially odd since most developing countries face substantial fiscal challenges both from their changing environments and from the international development community, which constantly delivers often conflicting messages to spend more, spend better, tax more, and tax better. A recent IMF (2005) assessment, for example, set out a revenue-to-GDP ratio of 15–20 percent as a reasonable minimum ‘threshold’ for developing countries.2 Similarly, the UN Millennium Project (2005) informed developing countries that on average they need to mobilize an additional 4 percent of GDP in tax revenue to achieve the minimal targets set by the project (the Millennium Development Goals) – that is, to increase from their current average tax level of 17–18 percent to something closer to 22 percent. Tax Levels As Table 12.1 shows, on average the share of taxes (excluding Social Security) increased from 30 to about 35 percent in recent decades in the developed (OECD) countries.3 In developing countries, however, the tax share of output increased only slightly: indeed, since the 1980s their tax shares have been almost constant.4 In contrast, an earlier study found that the average tax ratio for central governments in (a smaller sample of) developing countries had increased by about 24 percent over the previous two decades, suggesting that ‘convergence’ in tax levels between developed and developing countries appeared to be well on its way (Chelliah, 1971). The last few decades have changed the picture. In fact, at the beginning of the present century, the tax ratio in developed countries was roughly twice that in developing countries – a much greater difference than in the 1970s.5 As Easterly (2009) has recently reminded us, there has never been a

Tax system change and the impact of tax research 415 Table 12.1

Tax revenues as a percentage of GDPa

Country Groups Industrialized Developing Totalc

1970s

1980s

1990s

2000sb

30.1 16.2 19.8

33.7 17.3 21.6

35.5 17.0 22.6

33.4 17.0 21.8

Notes: a. Decade averages. b. Only limited data are available. c. Total includes ‘transitional’ countries not included in the other two groups. Source:

Bahl (2006).

shortage of outside experts urging developing countries to make a ‘big push’ to break out of the low-income trap. This observation is certainly true with respect to taxation. In the early 1960s, for example, Nicholas Kaldor (1963), fresh from recent exposure to India’s tax system, argued that for a country to become ‘developed’ it needed to collect in taxes something of the order of 25–30 percent of GDP.6 Even the more modest recent targets mentioned above seem unduly ambitious in terms of the historical record. A few fast-growing Asian countries such as India managed to reach and even exceed the UN-prescribed 4 percent of GDP increase in tax ratio in the early years of this century but it is by no means clear that such levels are sustainable.7 On the whole, fiscal inertia rather than fiscal growth appears to prevail, with many countries remaining for relatively long periods at more or less the same tax–GDP level.8 Of course, there is considerable variation within the diverse group of developing countries.9 While this is not the place to go into details, a recent analysis of the determinants of tax ratios suggests, among other things, that (1) developing countries that increased taxes did so largely in response to an increase in per capita GDP and (2) that there is at least some support for the argument that corruption and taxation are substitutes (Bahl, 2006).10 While there are some striking exceptions – such as the case of Nicaragua mentioned later in this chapter – when it comes to tax levels in developing countries, on the whole less has been going on than may be apparent to eyes dazzled by the seemingly endless changes of tax rates and tax legislation in many countries. To paraphrase a remark Galbraith (1964) once made, few of the irreversible transformative changes so often predicted as a result of this or that tax reform seem to have occurred. ‘Business as usual’ is a better description than ‘tax reform’ when it comes to tax reality in many countries. Although tax ratios vary by income levels, even the poorest countries,

416

The Elgar guide to tax systems

although obviously more constrained than richer countries, have considerable discretion as to how much they raise in taxation. Both opportunity and choice affect tax levels. For example, countries with access to rich natural resource revenues, like Venezuela and Azerbaijan, tend to have higher tax ratios than otherwise comparable countries, though such revenues may also be highly volatile in response to commodity price changes. Many factors other than the level of per capita GDP affect tax ratios. Tridimas and Winer (2004) usefully divide the possible explanatory factors into demand factors, supply factors, and what they call ‘political’ factors that affect the way in which changes in demand and supply variables enter into and shape policy decisions; they also set out an interesting integrative model incorporating all groups of factors. Less ambitiously, Bird et al. (2006), review a number of previous empirical studies of the traditional supply-side (‘tax handle’) variables, and then make new estimates with broadly similar results to most earlier studies along similar lines: per capita GDP and the non-agricultural share of GDP appear to be major factors explaining the size of public revenues in different countries. Like Baunsgaard and Keen (2005), Bird et al. (2006) find that openness is no longer as significant an explanatory factor as in most earlier studies, presumably as a result of the substantial trade liberalization of recent years. It is hardly a surprise to learn that the availability of an oil sector is important in explaining how much a country raises in revenue. However, telling a country that wants to raise its tax to GDP ratio to find oil is not very helpful. Supply-side studies make the problem facing most developing countries look more like a dilemma than a challenge: (1) poor countries tax less because they have less to tax; (2) to develop their economy (and tax base), poor countries need to spend more on public infrastructure, education, and so on so; (3) therefore they need to tax more. One way out of this dilemma is to argue (as did Kaldor, 1963) that the real reason countries do not tax more is not so much because nature makes it impossible but because it is not in the interest of those who dominate their political institutions to increase taxes even to the extent ‘nature’ (and the world economy) permits. If this were the story, economists – who as a group seem somewhat professionally reluctant to leap to the revolutionary barricades – would appear to find it difficult to suggest an alternative solution. Bird et al. (2006) offer a somewhat more hopeful version of this story. Using several new ‘demand-side’ variables (such as quality of governance, inequality, size of informal sector, and tax morale), their estimates suggest that to a significant extent tax levels reflect people’s perception of the quality and responsiveness of the state. Kaldor (1963) was thus right in the important sense that countries that wish to tax more need to ensure their

Tax system change and the impact of tax research 417 Table 12.2

Tax categories as percentage of total taxes

Income taxes

1970

1980

1990

2000

Industrialized Developing Total

35.5 29.6 30.7

37.8 28.6 30.2

38.6 27.6 29.7

53.8 28.3 28.5

Consumption taxes Industrialized Developing Total

27.2 25.2 25.3

29.4 29.3 28.9

30.5 34.9 34.2

19.8 40.1 39.0

Trade taxes Industrialized Developing Total

4.6 32.4 25.2

2.8 30.7 23.8

1.0 25.6 18.2

1.0 19.0 14.1

Source and notes: As for Table 12.1.

governing institutions facilitate the achievement of this goal. Doing so by such oft-suggested means as enhancing the rule of law, reducing corruption and the shadow economy, and improving tax morale, is neither simple nor easy. Nonetheless, progress along these lines may be more feasible than attempting to, as it were, ‘engineer’ fiscal gains by altering the relative share of the non-agriculture sector in the economy or the weight of imports and exports in GDP.11 Tax Structures The manner in which countries raise taxes differs as widely as the amounts they raise. The pattern of taxes in any country depends upon economic structure, history, the tax structures found in neighboring countries, administrative capacity, and political institutions. To illustrate, tax design is, for instance, strongly influenced by economic structure. Many developing countries have a large traditional agricultural sector that is not easily taxed. Many also have a significant informal economy that is largely outside the formal tax structure. As a result, the potentially reachable tax base usually constitutes a smaller portion of total economic activity than in developed countries. As Table 12.2 shows, consumption taxes are much more important in developing countries. On the other hand, income taxes are much more important in developed countries. For the whole sample studied by Fox and Gurley (2005), personal income taxes were a bit more important

418

The Elgar guide to tax systems

than corporate taxes (including extractive taxes) and VATs accounted for about 40 percent of consumption taxes, with excises being almost as important. In developing countries, however, the personal income tax plays a very limited role (Bird and Zolt, 2005). Such countries have been hesitant to go too far in taxing labor in the formal sector, and labor in the informal sector is largely beyond the taxman’s reach. The result is that, although personal income tax revenues are frequently three to four times corporate tax revenues in developed countries, in developing countries corporate tax revenues often exceed personal income tax revenues, sometimes by substantial amounts (Tanzi and Zee, 2000). Even company income taxes have shown little growth in many developing countries as a result in part of continued and even increasing recourse to tax incentives as an instrument of growth policy (Keen and Simone, 2004).12 An obvious reason why most developing countries reap little from either income or property taxes is their continued inability to administer such taxes effectively. The differences in the relative use of income taxes are even more pronounced when examined on a regional basis. For instance, personal income taxes account for only about 1 percent of GDP in Latin America compared with (almost) 3 percent in Africa (Fox and Gurley, 2005). Variations between countries within regions are even greater. In small island countries such as Barbados, for instance, international trade taxes may play an unusually important role. Trade taxes tend on the whole to be more important in poor countries, where they account for 24 percent of tax revenues compared with only 1 percent in rich countries. Trade taxes (mainly customs duties) decline steadily as countries become more developed.13 It is the poorest countries that have faced the greatest challenge in replacing such revenues in recent years as a result of trade liberalization. As Baunsgaard and Keen (2005) demonstrate, many of these countries have not been able to rise to this challenge. On the other hand, the higher per capita income the more countries tend to rely on direct taxes (especially those on personal income) owing both to differences in economic structure and differences in collection capacity. Low-income countries raise much more of their revenue at the border where relatively few collection points need to be controlled. For the same reason, such countries are more likely to rely on excise taxes, such as those on tobacco and alcohol, for significant shares of their revenue. VATs, like direct taxes, tend to require not only more effective tax administration but also taxpayers who are more sophisticated – both conditions more likely to exist in more developed countries. Nonetheless, undoubtedly the most striking trend around the world in recent years has been the increase in the share of revenues generated by taxes on domestic consumption taxes.

Tax system change and the impact of tax research 419 Particularly notable has been the widespread adoption of broad-based VATs (Bird and Gendron, 2007). However, the change in the composition of consumption taxes may mean less than at first appears for two reasons. First, although the manner in which consumption taxes are collected has changed, their relative importance has not. For developing countries as a whole, the rise in VAT has been almost exactly offset by the decline of trade taxes, while excise taxes have more or less maintained their position. Second, since a high proportion of VAT revenues are still collected at the customs house in many developing countries (Keen, 2007), not even the composition of consumption taxation has changed all that much in reality. How countries structure their tax systems depends also upon such factors as the need and desire for increased public services and the capacity to levy taxes effectively as well as the strength of preferences for such public policy goals as attaining a desired distribution of income and wealth and increasing the rate of growth. In a recent study based on observations for 100 countries over the 1975–92 period Kenny and Winer (2006) show that countries tend to utilize all tax bases (including some not normally included in ‘tax ratio’ studies such as seignorage) more as tax levels rise. To illustrate, if one compares OECD countries with Latin American countries, the latter collect less as a share of GDP from every tax source (Barreix and Roca, 2006). More interestingly, Kenny and Winer (2006) also show that the degree of reliance on different tax bases over time increases more on bases that become relatively more important. For example, as oil production and prices increase, oil countries get more revenue from this source.14 Further, much as argued in the traditional ‘tax handles’ approach (Musgrave, 1969) taxes on particular bases tend to increase when the administrative costs of imposing those taxes decline. For example, rising education levels lower the cost of imposing personal income taxes and are hence associated with more reliance on such taxes. Finally, Kenny and Winer (2006) suggest that a critical factor influencing tax structure choices is the extent to which reliance on particular tax sources can be translated into effective political opposition. For example, as Prichard (2009) details with respect to Ghana, fuel taxes may be derailed by strong opposition from well-organized taxi and truck operators. As Kenny and Winer (2006, p. 209) conclude, it seems clear that ‘the onus should be on tax reformers . . . to justify why any particular country’s tax mix should be substantially altered in relation to its existing political equilibrium’. I return to this important point later. Developing countries obviously face difficult challenges in designing and implementing suitable tax systems. Many countries have large traditional agriculture sectors that everyone finds difficult to tax (Bird,

420

The Elgar guide to tax systems

1974). Other significant components of the potential tax base lurk in other equally ‘hard-to-tax’ sectors ranging from small business and the informal economy to cross-border investments (Bird and Wallace, 2004). However, tax yields should grow as openness and incomes increase since economic growth is both encouraged by and usually results in closer involvement with the international economy. As countries develop and become more open, the expanding mass modern production and consumption activities on which the tax systems of developed countries rest – taxes on wages and personal income, on corporate profits, on value-added – need to be captured in the tax base without either overstraining administrative capacity or unduly discouraging the expansion of such activities. This can be a fine line to draw and one that may be beyond the reach of some countries, not least because globalization may also exacerbate fiscal problems. The leading edge of growth – outward-oriented development – may all too easily become the bleeding edge of the fiscal system as it becomes more and more difficult to levy taxes effectively on capital income, thus potentially exacerbating internal inequalities and political pressures on the tax system. Similarly, although developing countries in the past have often relied heavily on border taxes on international trade, this tax base too is becoming increasingly hard to exploit in the face of pressures for trade liberalization. Life is not easy for those concerned with taxation issues in developing countries. Over the past 50 years, academic research and international institutions have issued many policy prescriptions for economic growth (Easterly, 2002). In rough chronological order, the advice has included increased capital investment, improvements in education, population control, trade and capital market liberalization, and reduction of government controls on market activities. Each has in turn been marketed by some as a ‘silver bullet’ that would result in improved economic performance. Unfortunately, none of these cures has worked as advertised. Nor is there any magic tax strategy to encourage economic growth. On the whole, empirical work on the impact of tax levels on growth in developing countries has come to no firm conclusions. Even endogenous growth models that allow for the effects of tax policy on growth do not give a consensus answer about whether higher taxes crowd out faster rates of economic growth (Mintz, 2003). It is difficult to separate the effect of the level of taxes from the level of expenditures and the budget balance. Not surprisingly, different model specifications produce different results. The effect of tax structure on economic growth is an equally unresolved issue (OECD, 2008). In theory, of course, non-neutralities in tax structure impose a drag on the economy. Using computable general equilibrium models, the welfare cost of some taxes in some developing countries have

Tax system change and the impact of tax research 421 been estimated to be more than 100 percent of the amount of tax raised (Rutherford et al., 2005). Others point to the stimulus effects of tax rate reductions. The evidence here is also not clear. Ivanova et al. (2005), for example, find no evidence of a supply-side effect from Russia’s rate reduction and adoption of a flat rate income tax, but Martinez-Vazquez et al. (2008) do find evidence of a labor supply effect.15 As Lindert (2003) shows in historical context, the effects of taxes in particular country settings often depend on very detailed characteristics of tax design and implementation that are not easily captured in econometric models. Some countries with high tax burdens have high growth rates. Some countries with low tax burdens have low growth rates. Looking at the relationship between growth rates and tax rates in the United States over the last 50 years, for instance, shows that the US has had its greatest periods of economic growth during those years where the tax rates were the highest (Slemrod and Bakija, 1996). This does not mean that high tax rates are the key to economic growth: growth rates might have been even higher in years with high tax rates if the rates had been lower. But it does provide yet one more indication that there is still much that we do not understand about the relation between taxes and growth. Indeed, no matter how the data are manipulated, it usually turns out to be difficult to detect any economically significant relationship between variations in either tax levels and tax structures over time and growth rates either in OECD countries or in developing countries. Tax Systems Of course, many important aspects of tax systems are not directly visible in the recorded revenue figures. Tax laws emerge from a political process and produce revenue only when implemented. What can be done to a considerable extent determines what is done in any country. In many developing countries, as already mentioned, there is a large traditional agricultural sector that is not easily taxed.16 Often there is also a significant informal (shadow) economy that is largely outside the formal tax structure (Alm et al., 2004). To some extent the size of the ‘untaxed’ economy may itself be a function of the design and implementation of the tax system. For example, the high social insurance tax rates levied in some countries may discourage employers from reporting the extent of employment, encourage the under-reporting of wage levels, and foster the development of the informal economy. If the resulting lower tax revenues lead governments to raise tax rates still further, incentives to evade taxes will be exacerbated. Such problems are more difficult to cope with when a country’s administrative capacity is limited, as it is in most low-income countries.

422

The Elgar guide to tax systems

The importance of good administration has long been as obvious to those concerned with tax policy in developing countries as has its absence in practice. The real tax system people and businesses face reflects not just tax law but also how that law is actually implemented in practice. How a tax system is administered affects its yield, its incidence, and its efficiency (Tanzi, 1991). Tax administration is too important to policy outcomes to be neglected by tax policy reformers.17 Unfortunately, tax administration is a difficult task even at the best of times and in the best of places, and conditions in few developing countries match these specifications. Moreover, administration is inherently country-specific and surprisingly hard to quantify in terms of both outputs and inputs. The best tax administration is not simply that which collects the most revenues; facilitating tax compliance is not simply a matter of adequately penalizing non-compliance; tax administration depends as much or more on private as on public actions (and reactions); and there is a complex interaction between various environmental factors, the specifics of substantive and procedural tax law, and the outcome of a given administrative effort. All this makes tax administration a complex matter. Nonetheless, in a very real sense, ‘tax administration is tax policy’ (Casanegra de Jantscher, 1990, p. 179: original emphasis). How revenue is raised – the effect of revenue-generation effort on equity, the political fortunes of the government, and the level of economic welfare – may be equally (or more) important as how much revenue is raised. Private as well as public costs of tax administration must be taken into account. Attention must be paid to the extent to which revenue is attributable to enforcement (the active intervention of the administration) rather than compliance (the relatively passive role of the administration as the recipient of revenues generated by other features of the system).18 Assessing the relation between administrative effort and revenue outcome is by no means a simple task. Work on this subject has barely begun in developed countries (OECD, 2009), and despite a few pioneering efforts (Gallagher, 2005) we know almost nothing about this critical dimension of taxation in developing countries.19 Much the same can be said of an even more fundamental determinant of tax system change – the political economy of taxation. Those who design and implement tax systems, like those who try to escape them, probably consider themselves to be eminently ‘practical’ people responding to the world around them as they see it. Keynes (1936, pp. 384–5) once said that ‘practical men, who believe themselves to be quite free from any intellectual influences, are usually the slaves of some defunct economist . . . soon or late, it is ideas, not vested interests, which are dangerous for good or evil’. This dictum both unduly flatters economists and gives too

Tax system change and the impact of tax research 423 little weight to interests and other factors. Tax policy is shaped not only by ideas and vested interests but also by changing economic conditions, by administrative constraints and technological possibilities, and, especially, by the political institutions within which these factors affect policy decisions.20 Developing countries are no different than others: ideas, interests, and institutions determine tax policy. The best tax system for any country is presumably one that reflects its economic structure, its capacity to administer taxes, its public service needs, and its access to such other sources of revenue as aid or oil. In addition, it must also take into account such nebulous but important factors as ‘tax morale’, ‘tax culture’, and, perhaps above all, the level of ‘trust’ existing between people and their government.21 Tax policy decisions are not made in a vacuum. Nor are tax systems implemented in one. The taxes that are adopted in a country and how they are administered are always and everywhere both pathdependent and context-specific. They reflect the outcome of complex social and political interactions between different groups in society in an institutional context established by history and state administrative capacity. Like tax administration, tax politics thus deserves close attention by those interested in improving tax policy.

2

TRENDS IN TAX RESEARCH

Until recently, however, few economists devoted much thought to either the administrative or the political dimensions of taxation. Understandably, they instead approached research on taxation almost entirely from the disciplinary base of mainstream economics. What lessons has this impressive body of research suggested in recent decades for those concerned with improving tax policy in any country? Income vs. Consumption As Auerbach (2010) notes, the accepted academic view of good tax policy, circa 1970, was approximately as follows: 1. 2.

3.

The ideal tax was a broad-based income tax with progressive rates. Capital gains were properly taxed under such a tax. (Practically, however, gains could only be taxed on a realization basis, with the resulting lock-in effects often being mitigated by a reduced tax rate.) Integrating corporate and personal income taxes was seen as a way of making both the decision to incorporate and the debt–equity choice

424

4.

The Elgar guide to tax systems more neutral. (Usefully, dividend relief would also dampen the distortion from lower capital gains tax rates.) The dominant approach to taxing cross-border income flows was capital export neutrality although a limited case could also be made for capital import neutrality.

This concise summary seems to me to tell the story of the 1960s; it was all about the income tax, at least in English-speaking countries. Indeed, the report of the Canadian Royal Commission on Taxation (1966) at the time was called ‘a landmark in the annals of taxation’ by Harberger (1968) precisely because it was considered to be the most detailed attempt to turn these ideas into practical policy recommendations. Over the last few decades, however, many of the policy verities listed above were overturned.22 New and different guidelines for tax policymakers emerged from tax research. Most importantly, taxing consumption became intellectually respectable. Indeed, to many current economists – if as yet few policy-makers – a good tax system is likely to be one based on taxing consumption. Indeed, much of the (largely US) theoretical literature on this subject in the 1980s and 1990s focused on how to design and implement an ‘optimal’ progressive consumption tax to replace the income tax, although no one has actually done this, anywhere.23 Recently, the earlier enthusiasm for such a wholesale tax substitution seems to have died down even in theoretical circles. However, an important residual effect remains; few economists are now keen on taxing capital income at high rates, if at all. Two main policy-relevant outcomes of recent tax research stand out. First, the world-wide expansion of VAT probably benefited to some extent by the removal of (most) professional economists from the list of opponents of general consumption taxation. Curiously, the main practical result of the prolonged theoretical discussion about ‘new’ forms of progressive direct consumption taxes may thus have been to foster a better – but still not very (if at all) progressive – indirect consumption tax.24 Second, in the eyes of an increasing number of economists, it is becoming more difficult to make a good case – other than the practical difficulty of distinguishing one from the other – for taxing income from labor and income from capital identically. What does this mean for developing countries? Replacing income taxes by progressive consumption taxation in such countries was – as efforts inspired by Kaldor in India and what was then Ceylon (now Sri Lanka) showed long ago (Goode, 1961) – never a real possibility. In the end, in addition to making VAT more acceptable, the main practical impact of all the fuss about potential new forms of direct consumption taxes may turn

Tax system change and the impact of tax research 425 out to be to revive income taxes in developing countries by turning them into something closer to dual taxes, with mildly progressive rates on wages but largely flat taxes on capital (Bird and Zolt, 2010). An additional important conclusion from recent research is that, as Auerbach (2010) makes clear, countries can no longer treat the ‘international’ dimension of tax policy as some sort of simple ‘add-on’ to domestic tax policy. In the more open international capital market of recent decades, all countries are to some extent in competition for capital and need to factor this reality into their income tax systems. Unfortunately, as Auerbach (2010, p. 69) notes, tax research has as yet produced ‘. . . no simple norms that tell us what a system of international taxation should look like’. While this is not a subject that can be explored in depth here, it seems unlikely that answers to the intractable problems facing international tax designers lie in either the traditional source or residence principles (Bird and Wilkie, 2000) or for that matter in such new ‘principles’ as ‘capital ownership neutrality’ (Desai and Hines, 2003). Some (Tanzi, 1995) have suggested that what the world may need to grapple with the tax problems of globalization is some kind of formal ‘International Tax Organization’ with at least some enforcement powers. However, this seems to be several giant steps closer to ‘world government’ than anyone is likely to go in the near future – or perhaps should go at all. Instead, those looking for solutions to the international tax dilemma should pay more attention in the international context to the institutional setting from which, if we are persistent, reasonable (that is, acceptable) solutions may eventually emerge (Bird and Mintz, 2003). When principle provides no answer, practice will eventually do so in one way or another. As Babcock and Loewenstein (1997, p. 122) put it: ‘. . . there are many problems that people are unable to solve in the abstract, but are able to solve when placed in a real-world context’. International taxation may be one such problem Optimal Tax Theory An additional characteristic of the ‘1970’ tax setting noted by Auerbach (2010) is that second-best theory made it difficult to decide whether one could or should correct certain distortions or to carry out partial reforms. Fortunately those thus baffled were soon saved, or so it seemed, by the debut of optimal tax theory (OT). A recent paper by Mankiw et al. (2009) helpfully sets out a number of lessons that policy-makers may derive from OT, including the following: 1. 2.

Marginal tax rates on higher incomes should be lowered. More linear (flatter) tax rate schedules might be better.

426 3. 4. 5.

The Elgar guide to tax systems Taxes should vary with personal characteristics as well as income. Only final goods should be taxed, and usually they should be taxed uniformly. Capital income should be untaxed, at least ex ante.

Mankiw et al. (2009) conclude that policy-makers in the OECD countries seem to have heard at least part of the OT siren call. With respect to the personal income tax, for example, although both the theory and the evidence on the gains from doing so remain ambiguous, many countries, developed and developing alike, have lowered marginal tax rates on high income recipients.25 In Latin America, for example, the average top personal income tax rate fell from 51 percent in 1985 to 28 percent in 2003 and the average corporate rate from 41 to 28 percent (Lora and Cardenas, 2006). High marginal tax rates (MTRs) clearly induce a variety of changes in the behavior of taxpayers, with resulting economic costs. Tax-induced changes may include changes in hours worked and in labor force participation, the substitution of non-taxable for taxable consumption, changes in the timing of income realization, changes in the form of compensation (including incorporation), use of deferred compensation and other tax shelters, and increased evasion.26 These statements may be uncontroversial to economists. Unfortunately, they are also not particularly useful to policy-makers. OT makes it clear that a high MTR is more costly when applied to a tax base that is more responsive to tax rates – when, for example, affected taxpayers may easily substitute from paid work to unpaid family care, or from conventional employment to activities in the less-taxed informal sector, or they may even move to another country. Such behavioral responses to taxation can be usefully summarized in the ‘elasticity of taxable income’ (Feldstein, 1995) – the average percentage decrease in a taxpayer’s taxable income due to all behavioral responses when the taxpayer’s marginal share (one minus the marginal tax rate) is decreased by 1 percent. Examining the effects of the 1986 US tax reform on a sample of taxpayers, Feldstein (1995) estimated the elasticity of taxable income to be quite large, with preferred estimates ranging from 1.0 to 1.5. To put these estimates in perspective, note that unitary elasticity implies that government revenues would reach their maximum level at a tax rate of 50 per cent; further tax increases would actually decrease revenues. This approach provides a useful way to help make old arguments about efficiency more meaningful and palatable to policy-makers. As Mankiw et al. (2009, p. 120) put it, ‘though the optimal tax literature has not conclusively answered the question of how far from flat is the optimal tax policy, policymakers seem to have decided

Tax system change and the impact of tax research 427 that flatter is better’. However, although income tax rates have come down around the world, as Mankiw et al. (2009) themselves conclude there is little evidence that the structure of MTRs accords with what analysis suggests. Moreover, although most OECD countries have varied tax schedules (including direct income-related transfers as well as refundable tax credits) that to some extent take into account such non-income characteristics as age, parenthood, and disability, most have been reluctant to pursue ‘tagging’ any further. Mankiw et al. (2009) suggest that one reason for this reluctance is perhaps because criteria other than ‘ability’ such as benefits received and horizontal equity are considered relevant. Not only are they probably correct in this suggestion but, as is discussed further below, policy-makers are equally correct to take such criteria into account – regardless of whether they fit neatly within the OT framework. The inability of tax researchers to provide clear guidance on such basic questions of income tax design as the appropriate tax unit (Auerbach, 2010) is just one more example of the limitations of the OT approach as a guide to good tax policy. ‘There are more things in heaven and earth, Horatio, than are dreamt of in your philosophy’ as Shakespeare once put it. Turning to consumption taxes, it is not surprising that authors who have diligently toiled in the OT vineyard suggest that ‘the large and growing importance of value-added taxes suggests that policymakers have internalized certain lessons of optimal tax theory with regard to commodity taxation’ (Mankiw et al., 2009, p. 20). It seems more plausible, however, to suggest that VAT has swept the board for a number of reasons that have little, if anything, to do with the ‘internalization’ of OT lessons. In what is now the European Union, for example, the rationale for adopting the VAT was primarily to deal with troublesome cross-border trade issues. In most of the rest of the world, the rationale was in part the reputed administrative advantages of VAT compared with other forms of sales taxes, in part the desire of the IMF and other external advisers to increase the income-elasticity of consumption taxes, and most recently the need to accommodate WTO-linked trade liberalization. Similarly, while Mankiw et al. (2009, p. 22) seem a bit disappointed when they conclude with respect to OECD countries that ‘both statutory tax rates on capital and measures of effective tax rates remain far from zero, the level recommended by standard optimal tax models’, a few pages later they note more cheerfully that ‘. . . some trends in tax policy look like at least partial victories for optimal tax theory. Perhaps the most important is the worldwide trend toward reduced taxation of capital income’ (p. 25). Again, one may doubt the extent to which this trend is attributable to the slow inculcation of OT into the minds of policy-makers. A much

428

The Elgar guide to tax systems

more important factor has surely been the much discussed ‘globalization’ of international capital markets and the accompanying regional and international competition for capital that was such a marked factor shaping tax policy prior to the current crisis. Some years ago, in a review of Newbery and Stern (1987), the first major attempt to apply OT to developing country taxation, McLure (1990, p. 431) noted that ‘. . . graduate students, who are notably inept at seeing the forests of reality while surrounded by the tall trees of economic theory . . . should be aware that there is much more to the analysis of taxation in developing countries than the sophisticated economic theory developed in this impressive volume’. Mankiw et al. (2009, p. 2) correctly recognize that ‘where large gaps between theory and policy remain, the . . . question is whether policymakers need to learn more from theorists, or the other way round’. With respect to developing countries, the choice seems clear: tax researchers need to understand the real constraints and objectives facing policy-makers before offering them pre-cut solutions to what researchers think are their problems. As Slemrod (1990) and others have argued, for example, it is critical to incorporate administrative constraints into OT analysis. Moreover, as Moore (2007) and others have said, we also need to take explicitly into account the extent to which such concerns as horizontal equity may reflect, and affect, the basic ‘state-building’ function of taxation in developing countries. This point comes out particularly strongly with respect to the increasingly sophisticated OT theory that has recently been developed in a stochastic and dynamic analytical framework. Work along these lines is complex and, as Mankiw et al. (2009) note, as yet yields few clear messages for policy-makers. Nor does it seem likely to do so in the near future, if ever, for policy-makers in developing countries. Such countries are populated not only by the homo economicus of mainstream neoclassical economics – graduates, as it were, of ‘Max U’ (the school of utility maximization) – but also by the very different people whom Cullis et al. (2010) call homo realitus. This being – the object of much recent study in behavioral economics – satisfices within bounded rationality and is sensitive to framing effects, reference points, preference interdependence, and other characteristics that are troublesome from a Max U perspective.27 To further complicate the story, developing countries presumably live in the hope that lurking about the scene somewhere there may even be a few specimens of the recently rediscovered creature that Bronk (2009) calls homo romanticus – essentially the imaginative and entrepreneurial change agent central to Schumpeter’s (1934, 1942) development model.28 Not only is the world within which tax policy decisions are made complex, so are the people who make, and react to, such decisions.

Tax system change and the impact of tax research 429

3

WHAT HAVE WE LEARNED?

The two most important changes in tax systems in recent decades have been the introduction of the VAT and the general lowering and flattening of statutory income tax rates. Economic research has undoubtedly contributed significantly to both these important changes in taxation. However, countries have not done these things because economists produced persuasive theories or empirical evidence that it would be good to do them. On the contrary, fiscal history suggests that much tax research takes place precisely because countries change tax policies for their own reasons.29 Often, tax researchers are not so much leading the reform elephant as mopping up behind it. VAT, for example, was developed in France and then adopted in Europe as a better way to administer a general consumption tax, particularly for cross-border trade. Countries lowered and flattened income taxes both because more people rose into tax brackets formerly occupied, if at all, by a few rich people and also because of perceived international competition for capital and to some extent even for highly-skilled workers.30 Economists then came along and tidied up and rationalized in terms of accepted theory developments that had to some extent already occurred in the real world. Even when policy-makers buy economic arguments and evidence, they often do so for their own reasons, not simply because the economic doctor tells them it would be good for them. To illustrate how little research usually has to do with policy consider, for instance, property taxes in the United States. Research on this tax has long been a major industry in that country.31 Nonetheless, it is hard to detect much, if any, effect on actual tax policy as a result of all this research. None of the many and sometimes major changes that have taken place in property tax policy in recent years seem to reflect much, if any, concern for what researchers have found. For example, it is impossible to find any support in the literature for the increasingly common policy of imposing arbitrary caps and limits on property taxes. Property tax research and property tax policy appear to be activities carried out by different people in different rooms who do not communicate well with each other. Excise taxes are another ancient tax that still generates significant revenues in many countries. Countless studies have considered the efficiency effects of taxes on alcoholic beverages and tobacco. Again, however, such studies have had little perceptible effect on either the level or structure of these taxes.32 Much the same is true with respect to taxes on vehicles and fuel, despite the substantial economic literature suggesting alternative designs of these levies on efficiency grounds. Perceived effects on equity and politics almost always trump efficiency analysis even if – unusually

430

The Elgar guide to tax systems

– the latter is presented to policy-makers in both language and a context to which they can relate. Does this rather negative appraisal of the effects of tax research on tax policy mean we should give up? Of course not. First, it is essential, to paraphrase Harberger (1971), that economists continue to carry the flag of efficiency proudly into battle. If we do not, no one else is likely to do so, and efficiency really does matter . . . a lot, even though, as suggested earlier, more should be done in expanding the normative perspective as suggested in some of the recent behavioral economics literature (Congdon et al., 2009) as well as, in some respects the positive perspective – for example, in terms of the probably much higher marginal cost of public funds (MCPF) in the context of developing than of developed countries (Estache and Wren-Lewis, 2009). Second, the research glass is at least half full from a policy perspective. The downward pressure on personal and corporate income tax rates has certainly been supported, if not initiated, by the increasing research measuring the distortions caused by high marginal tax rates. Equally, the widespread adoption of VAT is probably due at least in part to acceptance of the economic argument that this form of sales tax is less economically distorting.33 However, economists should not flatter themselves too much: equally convincing (or unconvincing) economic studies of the damage done by poorly-designed corporate, excise, property, and payroll taxes have not met with a similarly receptive audience. Both the broad-based low-rate (BBLR) approach to income taxes and the VAT as a superior sales tax and the best way to replace trade taxes were extensively marketed in the developing world by fiscal missionaries from the IMF and elsewhere.34 Not all that such missionaries had to sell, however, found such a receptive market. The relative success of the two ‘big’ changes mentioned above may well lie at least to some extent in the fact that, arguably, they coincide with elite interests. By definition, elites pay most of the income taxes in developing countries. Consequently they benefit directly from reduced and flatter income taxes. Elites do not pay most VAT. However, they control the companies that act as collection agents for this tax, and it is generally in their interests to extend the tax base as far as possible in order to draw into the tax net as much of their ‘informal sector’ competition as possible. Distribution Matters A number of less bleak lessons may also be drawn from the broader postwar experience. One is that tax economists who wish to influence public policy must pay more attention to the issues that motivate policy-makers

Tax system change and the impact of tax research 431 – both officials and politicians – as well as people more generally, even if this means that they may pay somewhat less attention to producing publishable academic papers. This is not a plea for ‘dumbing down’ policy analysis. On the contrary, in many ways – certainly from the perspective of improving tax policy outcomes – it is a plea to smarten it up. To illustrate, analysis that assumes distributional considerations are either unimportant or can easily be accommodated by (usually unspecified) adjustments somewhere else in the tax-transfer system simply does not resonate in the policy context of most countries. Distributional issues matter in tax policy. Indeed, they often dominate in the minds of those who shape that policy. However, it is unlikely that the principal concern is the sort of broad poverty-alleviation goal beloved by many academic writers on development. Policy-makers may talk about poverty-alleviation and more egalitarian income distribution. In reality, however, their concerns about policy outcomes are usually much more narrowly focused. How will this region or locality be affected relative to that one? Will home-owners be disadvantaged or benefited? How will children be affected? Older people? What will be the effects on farmers and the rural sector? Will this industry be better or worse off? Such questions are seldom considered in mainline academic research, but they are frequently the meat and bones of actual policy decisions. To resonate in policy circles, research thus needs to speak to such questions, whether or not they fit easily or at all within the accepted disciplinary research model. As the fascinating story of the (transitory) death of death taxes in the US a few years ago shows, one good story – even if false – can easily trump 100 sound econometric studies (Graetz and Shapiro, 2005). If economists want their research to be taken seriously, they too need to be able to tell good stories that relate to the concerns that people have – regardless of whether those concerns are seen as central, or even acceptable, in academic circles. Fairness is always a key issue in designing any tax regime. Of course, what is considered equitable or fair by one person may differ from the conceptions held by others. Some may stress horizontal over vertical equity, for example, as OECD (2006) suggests is increasingly true in developed countries and as Boylan (1996) and Bergman (2003) have suggested is also the case in Latin America. Others may tilt the balance the other way, as ‘progressive’ thinkers have long done. Broadly, horizontal equity requires those in similar circumstances to pay the same amount of taxes, while vertical equity requires appropriate differences among taxpayers in different economic circumstances. Those who have the same ability to pay should bear the same tax liability; equally, fairness would seem to require those taxpayers with greater ability to pay to pay relatively higher taxes. Both

432

The Elgar guide to tax systems

concepts have considerable appeal. Unfortunately, neither concept is very useful in actually setting up a tax system. Chu et al. (2000) surveyed 36 studies of tax incidence in 19 different developing countries: 13 found the tax system progressive, and seven each found the tax system proportional and regressive; the others had mixed findings or insignificant effects. Most of the reported progressivity came from income taxes. Income taxes in most developing countries are likely progressive to some extent. Indeed, to the extent income taxes do not impinge on the poorest people – those outside the market sector – they are bound to be progressive. Even within the taxed sector, progressive rates mean that the impact of the tax is progressive at least within the group of those who must pay tax on most of their income – for example, because they receive it in the form of wages from a public sector employer. In most developing countries, however, little if any tax is collected from capital income or self-employment (mixed) income and that little is most unlikely to be distributed very progressively. Does tax progressivity matter for distributional policy? Harberger (2006) argues it does not matter much: insofar as government policy affects the distribution of income and wealth, providing education (even when financed from regressive taxes) is much more powerful than taxing the rich. Even in a relatively advanced country like Chile with an unusually well-developed and effective tax administration the progressivity of the income tax has no significant influence on distributional outcomes (Engel et al., 1999). Although one may argue that progressive taxes share the burden of government more fairly, there are few poor countries in which even the most progressive taxes on the rich can, even ignoring incentive effects, make much difference to the poor (Ravaillon, 2009). It was not by taxing the rich but by taxing the growing middle class that developed countries ‘grew’ large states (Lindert, 2003). In notoriously unequal Latin America, for example, income taxes only began to be relatively efficient and effective revenue raisers as (some) countries have begun to bite into the middle class, essentially by combining reduced top rates with lowering the level at which those reduced rates come into play (Lora and Cardenas, 2006). The general failure of tax economists to say anything that policy-makers consider useful about distributional issues has often relegated them, and their evidence, to the sidelines in policy discussion. Distributional studies are out of fashion in academic circles. Such studies are often both conceptually and empirically difficult and very model-specific. Nonetheless, since equity continues to lie at the heart of public economics, unless and until economists can deal more explicitly and satisfactorily with this issue, their success record in influencing tax policy seems unlikely to improve much.35

Tax system change and the impact of tax research 433 In addition, if one wishes to affect policy, one must write in a way, and in a forum, that will come to the notice of policy-makers. A few prominent academic economists have not hesitated to engage in public discussion of key policy issues, attempting to convince wider audiences of the cogency and importance of the economic analysis of taxation. Such involvement in policy is usually off limits for those who aspire to tenure in an academic economics department, however. Nonetheless, economists concerned with the persistence of ‘bad’ policies unsupported by any coherent reasoning or cogent empirical evidence need both to communicate with policy-makers and to engage with the (small) informed public with which policy-makers interact. Tax policy is not just about economics but about politics. To understand how economic analysis of tax issues is perceived in any country and to affect policy outcomes one must understand the political as well as the economic factors that shape policy decisions and policy outcomes in that country. The level and the structure of taxation reflect deepseated institutional factors that, in the absence of severe shocks, do not change quickly. Tax policy decisions are not made by a benevolent government. Taxation is not just a means of financing government; it is also a very visible component of the social contract underlying the state. Citizens are more likely to comply with tax laws if they accept the state as legitimate and credible and are thus to some extent both willing to support it and afraid of what will happen to them if they don’t. Tax policy changes thus depend largely on how different political groups perceive proposed changes and how they react to these perceptions. As Lledo et al. (2003) put it, any major tax reform is thus always and everywhere ‘an exercise in political legitimation’. Those who will have to pay more must be convinced that they will get something worthwhile for their money. Those who do not want to pay more must not be able to block reform and, in the end, must be willing to go along without taking to the hills in revolt or fleeing the country.36 Those within government and in the private sector who have to implement reform must support it or at least not actively sabotage it. And of course politicians have to see sufficient support to warrant putting reform not only on the agenda but on the ground. Political scientists have recently rediscovered that taxes matter and are now doing some of the most interesting and promising work on tax policy in developing countries.37 Economists too need to pay more attention to the question of how tax policy is (and might be) formulated in democratic and non-democratic settings alike if they expect policy-makers in developing countries to make better use of economic evidence in developing and implementing tax policy.

434

The Elgar guide to tax systems

Accountability and Visibility As an example of the importance of getting the politics of taxation right, consider an apparently minor question about VAT. Should VAT be quoted separately from prices? This question does not have a simple answer.38 Retail sales taxes (RSTs) like those in most US states are invariably stated as a separate explicit charge imposed on the posted price when the consumer arrives at the cash register. While this process is cumbersome and unwelcome – no one ever has the correct change ready! – the very fact that it is annoying may be considered good for democracy if one believes that citizens should be fully aware of the cost of government. However, such transparency also makes it more difficult to increase taxes because everyone is instantly aware of increases. In contrast, in most countries VAT is included in posted prices and hence is not immediately obvious to the consumer at the point of sale (although it may be stated explicitly on the cash register statement or invoice). Visible, or invisible: does it matter? While there have been few serious studies of this issue,39 it is intuitively plausible that it should be easier to introduce (or increase) a tax if people are not painfully reminded that the tax exists by having to add it separately to the quoted price every time they buy something.40 Moreover, if VAT is relatively invisible it may be easier to have a better, broader tax base, for instance, by taxing a wider range of services. A broader base is undoubtedly more desirable on both administrative and economic grounds than one eroded by popular resistance to taxes on food, on medicines, on school books, and so on. On the other hand, if people should know what they are paying for what they are getting from government, then even if the price of more transparency is a less perfect VAT, the price might be considered worth paying.41 Given these conflicting arguments, it is striking that almost every country with a VAT has chosen to hide the tax from the public.42 This outcome is understandable from the standpoint of those who make the decisions. However, hiding the tax bill does not seem, on its face, to be a particularly attractive way to build a sustainable democratic consensus in support of fiscal equilibrium. Such issues need to be considered carefully in formulating tax policy. Marketing matters. Those who want serious reform in tax policy or administration in any country need to understand not only the details of tax theory and practice but also the dark art of political salesmanship. This warning needs to be taken particularly seriously by those are concerned not only with how tax revenues may build up a more sustainable

Tax system change and the impact of tax research 435 state but also with how the way in which such revenues are collected may contribute to, or detract from, the long-term development of state legitimacy. Is ‘fiscal illusion’ a way to achieve sustainability? Consider earmarking, and what Breton (1996) calls the ‘Wicksellian connection’ between the two sides of the budget. Establishing such a linkage in the minds of people is critical to the whole issue of state legitimation.43 Earmarking – linking specific revenue sources to specific expenditures – has existed since the earliest recorded fiscal practices (Webber and Wildavsky, 1986). Both politicians and taxpayers often find earmarking an attractive and feasible way to finance Social Security, road works, education, environmental programs, and other good things. Politicians like earmarking as a means of reducing taxpayer resistance to higher taxes. Taxpayers like the greater accountability they perceive with respect to how their tax dollars are spent. Economists, however, have come only lately to the table when it comes to understanding and analyzing this common fiscal practice.44 Those concerned with budgeting almost unanimously conclude that earmarking is a bad thing – often, it seems, implicitly assuming that budgetary decisions would otherwise be made by a benevolent dictator whose sole objective is to maximize social welfare. In more practical terms, budgetary experts, observing the mess that rampant earmarking has created in some countries, tend understandably to argue that no rational budgetary process is conceivable unless the practice is essentially banned.45 This Gladstonian approach remained essentially unchallenged in theory until Buchanan (1963) revived an important efficiency argument (made earlier by Wicksell) in favor of establishing as tight a linkage as possible between taxing and spending decisions.46 In this approach, earmarking is not simply a way to secure political consent for a tax increase. It may actually be the best way we have to deal with the fundamental normative problem of public economics – how to provide people with the public services they really want – where ‘want’ is interpreted in the only economically relevant sense of what they (collectively, as determined through their political institutions) are willing to pay for. Like all fiscal institutions, earmarking comes in many variants. The economic case for earmarking is strongest when there is a close benefit link between taxes and spending and the revenues received are spent on the service in question. Benefit-related earmarking (like user charges) – provided it is properly designed and implemented – reveals taxpayer preferences for public services and sends a clear demand signal to the public sector about how much of a service should be supplied. Since revenues received (and only such revenues) are spent on the service in question, supply automatically adjusts to demand and economic efficiency is

436

The Elgar guide to tax systems

achieved. Such earmarking may also be considered equitable in the sense that no one receives a service without paying for it or pays without receiving service. Provided that the public service thus financed is like a privately supplied service in the sense that both an individual’s consumption of the service and the marginal cost of providing the service can be satisfactorily measured, most people would probably consider such levies fair on the whole.47 Since taxpayers are aware that when payments are extracted from them the funds will be used to pay for certain expenditures, they will presumably support the taxes or charges if, and only if, they support the expansion in the supply of government services for which the earmarked revenues are targeted. Both tax and expenditure decisions may thus be made more rationally than under general fund financing.48 In practice, such perfection seems seldom to have been achieved. The relatively disappointing experiences with earmarking in many countries reflect many factors – the cost and difficulty of controlling many separately budgeted funds (Fullerton, 1996), the inappropriateness of many of the linkages that have been established for political reasons between particular revenues and expenditures (McCleary, 1991), and the understandable resistance of citizens to attempts to charge properly for services that initially were provided ‘free’ or at highly subsidized prices (Bird and Tsiopoulos, 1997). Much earmarking found around the world makes little economic sense. In many instances, not only is there no logical link between the tax imposed and the spending for which it is designated but there is also no solid budgetary link. The amount collected from the designated source and the amount spent on the designated activity are decided independently. What is the rationale for such non-logical non-linkages? Do they affect budgetary outcomes? These are not trivial questions. Many countries have many earmarked revenues, and new ones are likely created (or at least suggested) every day, not least in the environmental field. One important rationale for imposing taxes earmarked to support some popular activity is simply revenue enhancement. Politicians justify the imposition of a new tax by capitalizing on the presumed ‘halo effect’ of something popular like education or health or ‘greenness’. In some instances, earmarking may also be motivated by rent-seeking behavior. For example, an increase in tobacco taxes with the proceeds earmarked for increased health spending may receive support on all sides: from nonsmokers, who would not pay the tax but want more spending on health; from smokers who feel guilty about smoking and are worried about the health consequences of smoking; and also, of course, from the large number of people who are engaged in the health business and who (presumably) not only believe that higher taxes on tobacco and more spending on health are good but also realize that they would be clear gainers

Tax system change and the impact of tax research 437 because health spending would increase as a result – even though this need not necessarily happen. Some studies (Cnossen and Smart, 2005) suggest that tobacco taxes are both highly regressive and already higher in many countries than can be rationalized on any externality arguments. Still, how can politicians resist such a win–win combination?49 Even when people – or the relevant elites – feel overtaxed in general or are resistant to increases in taxation, new earmarking initiatives may sometimes be welcomed for reasons like those just mentioned. If the earmarking makes sense, then such initiatives may be economically sensible. However, except when there is a clear benefit rationale, it is hard for an economist to defend any form of non-benefit earmarking that actually alters expenditures (in the sense that the level of expenditure on any particular activity is determined at the margin by the amount of revenue collected from any particular tax or charge). But is earmarking that is not economically sensible necessarily bad? Those who think that more taxes are needed for good purposes (whether people agree or not) so that revenues obtained by fiscal sleight-of-hand are better than nothing will not agree. Those who agree with Musgrave (2000, p. 101) ‘. . . that society is capable of resolving its common concerns in a reasonably efficient, just and democratic manner’ must hope that, in the long run, as Abraham Lincoln once said, if one cannot fool all of the people all of the time it is unlikely that higher revenue levels than would otherwise be acceptable can long be sustained by manipulating fiscal illusions through symbolic earmarking.50 A corollary of this argument about earmarking is that sensible benefitrelated taxes or fees should be levied when feasible with the proceeds earmarked appropriately. Unfortunately, this proposition too is seldom satisfied in reality. The fees, charges, and public prices found in most countries are seldom sensible. Indeed, an important question needing further study is why the impressive economic evidence supporting the case for marginal-cost pricing has had so little impact on policy. In many countries, budgets are littered with fees, charges, and minor taxes that could be abolished with little loss in revenue and some gain in efficiency. Even the few ‘benefit’ and ‘earmarked’ levies that in principle might have some potential economic rationale invariably need substantial revision to be conducive to efficient resource use. There are many reasons for this sad state of affairs. For example, even in the best circumstances it is often surprisingly difficult to design and implement good user charges, and as a rule even good charges are not very popular with either administrators or citizens (Bird and Tsiopoulos, 1997). Still, it is surprising that so little attention has been paid to these issues. Much can be done to improve revenue systems in most developing countries in this respect. Those who look to ‘green’ taxes as potential sources of

438

The Elgar guide to tax systems

future revenue should be particularly concerned to explore such questions. There is much to be gained in terms of efficiency (if not always revenue) by improving our knowledge of ‘extra-budgetary’ budgeting, earmarking, user charges, and the like. Such subjects are often considered peripheral to the core interests of public economics. This is a mistake: not only are such ‘fringe’ areas often central to improving tax and other policy outcomes in many countries but there is much to be learned by exploring them more carefully.

4

PLAY THE RIGHT GAME

In the end, if tax research is to impact more effectively on tax policy, researchers must play the right game. Major changes in taxation generally require major changes either in the political reality of a country or in its economic circumstances. In normal times a ‘good’ tax reform – one intended to raise more revenue in a more efficient and equitable fashion – may be like a ‘good’ seat belt law: if everything else stays the same, lives will be saved (the tax ratio will increase). But things do not stay the same: some people drive faster when they are belted in (administrative efforts slacken off, concessions to favored groups multiply) with the result that death rates (tax ratios) show little change. Countries appear to achieve a sort of equilibrium position with respect to the size and nature of their fiscal systems that reflects the balance of political forces and institutions and then to stay there, fluctuating around some relatively stable ratio, until ‘shocked’ into a new equilibrium (Bird, 2003). Two alternative explanations may lie behind this process. Either – somewhat improbably – ‘supply’ (‘capacity’) factors may alter over time in such a way as to offset all attempts to raise tax ratios. Or, more plausibly, ideas as to the ‘proper’ tax level – ‘demand’ factors – may change over time or may be articulated differently through political institutions. As noted earlier, the tax policy world is very different in many respects now than it was 50 or even 20 years ago. Both the economic and the intellectual environment has changed. Ideas matter: as Blyth (2002, p. 274) puts it, ‘. . . neither material resources nor the self-interest of agents can dictate . . . ends or tell agents what future to construct. Ideas do this’. Of course, institutions and interests also matter in shaping tax policy. To understand the tax policy game, all these aspects need to be considered. An example may help make the point. Some years ago an analysis of Central American tax policy argued that in principle changes in tax level structure (e.g., the degree of emphasis on income taxation) essentially reflected the changing political balance of power between landlords,

Tax system change and the impact of tax research 439 capitalists, workers, and peasants (Best, 1976). Shortly after this article appeared, an explicitly ‘leftist’ regime (the Sandinistas) took over in Nicaragua. What happened to taxes? Three things: 1.

2.

3.

As Best (1976) would likely have predicted, the tax ratio rose very quickly, from 18 to 32 percent of GDP within the first five years of the Sandinista regime. However, almost all the increase in tax revenue came from (probably) regressive indirect taxes and not from the (at least nominally) progressive income taxes. In many ways most interestingly, once Nicaragua’s tax ratio was increased, it stayed up there even a decade (and three subsequent governments) after the defeat of the Sandinistas.51

As this example suggests, political ideas definitely matter in taxation; but they do not necessarily dominate. Economic and administrative realities also matter. Few real-world tax structures have been designed with any particular objective in mind. Often, they have been shaped by both the changing local environment and the changing external context. In the case of the United States, for example, over the 1860 to 1920 period ‘economic crises and wars helped create a consensus for an income tax that falls most heavily on the wealthiest taxpayers’ (Weisman, 2002, p. 366). The lengthy debate about taxes that took place over this period was not really about taxes at all but rather about what kind of society Americans wanted. Much the same can be said about other countries. Ideas on the relevant balance between taxes and society forged over the first half of the twentieth century have changed in recent years, as evidenced, for instance, by both the death of death taxes in developed countries52 and the limited success of developing countries in achieving the high levels of income taxation to which many of them aspired in the postcolonial period (Bird and Zolt, 2005). As noted earlier, however, both tax levels and the distribution of tax burdens have changed less than have ideas about the nature of a good tax system. The question raised earlier of how best to make the ‘Wicksellian connection’ operational is not simply as yet unanswered by tax research, it has not really been asked. It is clear, however, that the extent to which good decisions – that is, decisions that reflect people’s real preferences as closely as practically feasible – are made on either side of the budget cannot be separated from the question of the institutional structure within which political decisions are reached and expressed. In the end, it seems, for tax research to help tax policy-makers make the right decisions is to ensure that they – as well as, ideally, all those

440

The Elgar guide to tax systems

affected – are as aware as possible of all the relevant consequences. If a country is to have a better tax system – better in the sense of giving people what they want (and are willing to pay for) – it must first have a better political system that transmutes citizen preferences into policy decisions as efficiently as possible. ‘Democracy’, as Churchill once said, ‘is the worst form of Government except all those other forms that have been tried from time to time’.53 Bahl and Bird (2008b) suggest that ‘representation without taxation’ is not the way to produce good outcomes in decentralizing countries. Similarly ‘taxation without representation’ is not the way to create a fiscal system that is likely to be either economically or politically sustainable in the long run. Taxation is always and everywhere a contested concept. Some pay and some do not. Some pay more than others. Some receive compensating services, some do not. Such matters are in democratic states resolved only through political channels. Indeed, history suggests that the need to secure an adequate degree of consensus from the taxed is one of the principal ways in which, over the centuries, democratic institutions have spread (Sokoloff and Zolt, 2005). No non-dictatorial government in this age of information and mobility can long stay in power without securing a certain degree of consent from the populace, not least in the area of taxation. State legitimacy thus rests to a considerable extent on the ‘quasivoluntary compliance’ of citizens with respect to taxation (Levi, 1988). To secure such compliance in a sustainable way tax systems must, over time, represent in some real sense the basic values of at least a minimum supporting coalition of the population.54 The central problem in many Latin American countries, for instance, is clearly inequality (De Ferranti et al., 2004).55 The key, and related, governance problem in most of the same countries is lack of accountability. A better tax system is critical to the solution of both problems. Reforms that link taxes and benefits more tightly such as decentralization and more reliance on user charges may help accountability – though not necessarily reduce inequality.56 On the other hand, reforms that replace highly regressive and inelastic excises by a less regressive and more elastic VAT may reduce inequality – especially, of course, if the increased revenues are invested in growth-facilitating activities such as education and infrastructure.57 From a developmental perspective, the key function of taxes in developing countries is to provide (non-inflationary) funding for pro-poor and pro-growth spending programs particularly on developing human capital. The best way to achieve this goal in most countries is likely through a broad-based non-distortionary consumption tax like VAT, as has long been recognized (Heady, 2004). What countries actually do, however, is always and inevitably deter-

Tax system change and the impact of tax research 441 mined in the first instance by political and not economic calculations.58 Countries vary enormously in the effectiveness and nature of their political systems. Some may be close to ‘failed states’ in which institutions are so ineffective that it does not matter much what they attempt to do, it will not work. Others may be ‘developmentalist’ and wish to use their fiscal systems as part of a relatively dirigiste interventionist policy. Still others may be of a more laissez-faire disposition. Some may be more populist, some more elitist, some more predatory. The dominant policy ideas in any country (about equity and fairness, efficiency, and growth), like the dominant economic and social interests (capital, labor, regional, ethnic, rich, poor), and the key institutions, both political (democracy, decentralization, budgetary) and economic (protectionism, macroeconomic policy, market structure), interact in the formulation and implementation of tax (and budgetary) policy. Uniform results are unlikely to emerge from the always boiling policy cauldron, with its different mixes of ingredients in each country. The changing interplay of ideas, interests, and institutions affects both the level of taxation and its structure. Indeed, as Schumpeter (1954) emphasized long ago, taxation is one of the clearest arenas in which to witness the working out of these complex forces. Viewed in long-term perspective, few developing countries seem to have completed even the earlier parts of the long cycle that produced the (more or less) redistributive and (more or less) growth-facilitating fiscal states now found in most developed countries – the long preparatory period during which the idea of the desirability and even necessity of a larger state and a more or less progressive fiscal system became established to different degrees in different countries.59 Some countries in Latin America, for example, might be argued to have moved more from the colonial inequality of land-based maldistribution to the modern inequality of capital-based maldistribution.60 The point here is not that all countries must follow the same evolutionary path. However, unless countries do so to at least some extent they are unlikely to develop tax systems resembling those now found in most OECD countries. Many governments in developing countries – not just those in Latin America – are in dire straits. Even countries that have reached relatively safe harbors politically, achieving a certain degree of legitimacy and stability, almost always feel – often correctly – that they are in an economically precarious situation. The budget is politically and economically constrained. Life is difficult. Nothing can be done. All this may be true to some extent, but it is also both too much a counsel of despair and too easy a way out. Even in the most hopeless situations something usually can be done to improve matters. No doubt there will continue in most countries to be considerable dispute over what should be done to improve

442

The Elgar guide to tax systems

tax systems. Unless and until an adequate degree of political consensus on what should be done is achieved, no significant tax changes are likely to be made. In short, to a considerable extent the main tax challenge facing many developing countries is simply that there is as yet no implicit ‘social contract between governments and the general population of the kind that is embedded in taxation and fiscal principles and practices in politically more stable parts of the world’ (Lledo et al., 2004, p. 39). History tells us that such principles do not become embedded either painlessly or quickly. The specific substantive suggestions that Lledo et al. (2004) make to improve matters – such as better VAT administration on a broader base – are already the stuff of countless existing reports. I agree. Most countries should do all (or most of) the good things that experts tell them. The real question is: why have so many done so little? Lledo et al. (2004, p. 40) suggest, perhaps rather wistfully, that if Latin American countries wish to improve their tax systems they should ‘.  .  . improve political institutions in ways that broaden and deepen social contracts. For example, create more responsive and less clientelistic political parties, more cohesive and less polarised party systems, and improved capacity of civil society to monitor government and participate in tax debates’. In other words, what they say is that there can be no good taxation without good representation. They are right. But how useful is it to advise countries they should be something other than what they are? Reforming taxes in any country is always a one-off operation in the sense that it occurs in the unique circumstances of that place at that time.61 Most studies of tax reform focus on the substance of what should be considered. A more fundamental question, however, is not what should be taken into account in developing a tax reform proposal but rather how the issues, whatever they are, should be approached (IDB, 2006). McIntyre and Oldman (1975), for example, argue that more careful and comprehensive attention to institutional arrangements for tax reform will both improve the quality of the reforms proposed and increase the likelihood of their adoption and successful implementation. To be successful, for instance, reforms must be both comprehensible and fit with existing institutions. Even the best system for studying and developing reform proposals will never suffice to bring about good policy changes in the absence of a coherent strategy, continuing support from above, and an acceptable level of administration.62 Good policy formulation is never enough; but presumably it is always better than its absence, or its opposite. Some countries have tried to finesse the system by appointing some kind of special tax reform commission, whether foreign, domestic, or mixed. The track record of such efforts is not good. Appointing an outside group is often simply a way to postpone dealing with a problem. However good

Tax system change and the impact of tax research 443 the final output of such efforts may be, the results are seldom ‘owned’ by those who must first sell them and then make them work. Ownership matters. So does leadership. So does a coherent strategy, and of course so do adequate resources. Good tax policy planning involves economists, lawyers, administrators, and – importantly – adequate discussion with taxpayers and ‘third party’ tax collectors like banks and companies. Successful tax reform involves all this plus solid and continuing political support and adequate administrative follow-up. It is not easy anywhere. But it can be done. Good planning and policy formulation focuses on what matters and what can be done and pays close attention to detail and implementation (Gillis, 1989; Thirsk, 1997). In particular, building up adequate institutional capacity in the tax field, both inside and outside government, is critical to being able to adapt policies to changing circumstances and needs, thus ensuring some degree of robustness and resiliency (IDB, 2006). The role of outsiders such as academics and aid agencies in this process is more to be supportive when countries want to reform their systems than to tell them when and how to do it. The most important long-term contribution outsiders have probably made to improving tax policy through research in many countries has been by encouraging, and even funding, informed non-government ‘think tanks’ and university studies in relevant fields (including foreign training) as well as by fostering data gathering and analysis both within and outside government. Such activities are not glamorous and seldom produce clear short-run pay-offs. Nonetheless, they are critical to establishing institutions that, over time, will be able to provide ideas and information that may inform (and perhaps even make more articulate) the interests that shape tax systems. In the end, if a country needs or wants better tax policy or administration, it can have it: the answer largely lies in its own hands. Even those who want to do the right thing, however, can often use help in finding out just what is right and how it can best be done. It is always easy and seldom effective for those not in a game to give advice to those who are trying to play it. In general, however, academic researchers and outside agencies interested in fostering better sustainable tax systems in developing countries will employ their efforts and resources most usefully if they play in the right game. Fifty years of experience tells us that the right game for tax researchers is not the short-term political game in which policy decisions are made. Instead, it is the long-term game of building up the institutional capacity both within and outside governments to articulate relevant ideas for change, to collect and analyze relevant data, and of course to assess and criticize the effects of such changes as are made. Tax researchers can and should play an active role in all these activities. Such long-term

444

The Elgar guide to tax systems

‘institution-building’ activities are seldom immediately rewarding. They appear to be out of fashion with international agencies concerned with development. It seems much more appealing and immediately productive to establish performance benchmarks for success, to support this particular organizational change reform here (revenue authority) and that new technology (computerization) there, in the apparent belief that such simple ‘one-size-fits-all’ approaches can provide quick (but sustainable!) answers to the many complex problems inherent in policy reform in difficult environments. It may be appealing. But it is wrong.

NOTES * 1. 2.

3. 4.

5.

6.

7.

8.

9.

10.

Revised version of a paper originally presented at the conference on ‘Tax Systems: Whence and Whither? Recent Evolution, Current Problems and Future Challenges’, Malaga, Spain, September 2009. Some portions of the chapter draw heavily on an earlier treatment of some of these issues in Bird (2007). ‘Revenue’ (even ‘current revenue’) and taxes are not identical, although IMF (2005) seems to use the two words interchangeably. However, Bird et al. (2006) analyzed both revenue and tax ratios in developing countries and found no great differences in most instances. The two brief tables included in this section as well as some of the discussion follow Bahl and Bird (2008a). In the transitional countries (the former Soviet bloc countries as well as China and Vietnam, which are seen as being in transition from centrally planned to more marketdominated systems) tax shares actually declined in the 1990s, presumably reflecting the continuing realignment of public–private expenditure responsibilities. I do not discuss these countries further here: see Martinez-Vazquez and McNab (2000). For a more detailed review of tax levels and structures in countries grouped by income level, see, for example, Fox and Gurley (2005). (Population-weighted averages behave somewhat differently since larger countries like Brazil, India, and China dominate the results, but this is not relevant in the present context since tax policy decisions are made at the level of the nation-state.) Kaldor obviously did not agree with Colin Clark (1945) or, for that matter with Keynes, whom Clark (1977, p. 28) quotes as having, in private correspondence, agreed that ‘as a practical proposition . . . 25 per cent taxation is about the limit of what is easily borne’. As Poirson (2006) shows, general government revenues as a share of GDP have been surprisingly constant over time in India. Much of the increase in revenues in India in the early part of this decade reflected increased corporate taxes – always a rather volatile revenue source, as many countries have recently again learned. Often, even if the introduction of new technology or increased administrative effort initially appears to expand revenues, a variety of factors seems soon to dampen or even fully offset any resulting net increase in tax ratios: for an interesting discussion of the Mexican case, for example, see Martinez-Vazquez (2001). The term ‘developing countries’ encompasses such a wide spectrum – from small, fragile, and fragmented post-conflict states like Liberia and Afghanistan to large, wellestablished, and rapidly growing countries like Brazil, China, and India – that any generalization is likely to have many exceptions. The last point also receives some support from the analysis of Bird et al. (2008). Of

Tax system change and the impact of tax research 445

11.

12.

13.

14. 15. 16. 17. 18.

19. 20. 21. 22.

23. 24. 25.

course, one must always view cross-country regression-based interpretations with some skepticism (Easterly, 2009). Nonetheless, as Bird (2007) emphasizes, tax bases are not simply ‘given’; to some extent they may be ‘grown’ – or destroyed – through the manner in which a given tax burden is collected. For example, taxes may discourage, or encourage, the ‘formalization’ of the economy, they may foster or discourage the growth of such ‘tax handles’ as imports, or they may be used to shape and direct economic growth into particular channels in a variety of ways and for a variety of purposes. Of course, as mentioned earlier, all generalizations about so heterogeneous a group are suspect. On one hand, as mentioned in note 7 above, corporate taxes accounted for much of the recent rise in India’s tax ratio. On the other hand, in smaller countries the loss of as few as one or two firms may decimate the tax base. Fox and Gurley (2005) report that the coefficient in a regression of per capita GDP on international trade taxes as a share of GDP is negative and statistically significant. An exception are the transitional countries of Central and Eastern Europe, which – although many of them fall within the low-income group – have traditionally relied little on trade taxes (Martinez-Vazquez and McNab, 2000). In contrast, expanding trade in general has in recent years not been associated with increasing dependence on trade taxes (Baunsgaard and Keen, 2005). For a recent review of the Eastern European experience with flat taxes, see Gray et al. (2007). The special problems of taxing agriculture are not discussed here: a useful recent review is Khan (2001); the administrative aspect of agricultural taxation is discussed in some detail in Bird (1974). As McLaren (2003, p. v), puts the point: ‘.  .  . optimal policy requires simultaneous consideration of the design of the tax code and of the administrative structure created to enforce it’. In what is still one of the few studies of how tax administrations actually function in developing countries, Radian (1980) emphasizes the extent to which such administrations tend to be passive recipients of funds rather than active collectors. Radian labels this important aspect of tax administration ‘tellering’ as opposed to ‘collecting’. Rather than go out and look for tax revenues, such administrations tend to sit behind a counter and wait for people to bring money to them. As I note later, the facilitating and monitoring of such ‘quasi-voluntary’ compliance (Levi, 1988) is critically important for more than revenue reasons. For an earlier detailed review of the critical administrative dimension of tax policy, see Bird (2004a). For an interesting recent study along these lines for Latin America, see IDB (2006). Tax technology is not discussed here, but see Bird and Zolt (2008). For introductory discussions of the three factors listed in the text, see, respectively, Frey (2002), Edling and Nguyen-Thanh (2005), and Bergman (2002) as well as the case studies in Brautigam et al. (2007). For example, Auerbach (2010) notes that recent research raises questions about the strength of the traditional case for corporate–personal tax integration and also shows that taxing capital gains on an accrual rather than realization basis appears to be technically feasible. However, no one has yet managed to tax accrual gains and the few partial attempts made to do so (e.g., in Canada in the 1980s) soon vanished in the face of popular indifference and interested opposition. One cannot, it seems, ignore the strong perceptual barriers to accrual taxation that arise from liquidity concerns and especially from the way in which people ‘tie’ taxes to the specific object taxed. For a useful recent review of the literature on direct consumption taxes, see McLure and Zodrow (2007). For a review of the issue of VAT progressivity in developing countries, see Bird and Gendron (2007). For a useful recent compilation of the evidence, see Peter et al. (2009).

446 26. 27. 28.

29.

30. 31.

32.

33.

34. 35.

36. 37.

38. 39.

The Elgar guide to tax systems The next paragraph draws on Bird and Smart (2001). See the useful recent surveys by DellaVigna (2009) and Congdon et al. (2009). If such beings manifest themselves, countries must also pray that they appear in their ‘creative destruction’ mode, tearing down the old to build up the new, rather than in the ‘destruction for the sake of destruction’ mode of ‘romantic revolution’ that has played havoc in the past in some countries. For a revealing contrast of the two ‘romantic’ modes, see Bronk (2009); see also the interesting intellectual history of the impact of the left in Latin America in Castañeda (1994). At least this is my reading of the tax history set out in such sources as the magisterial survey by Webber and Wildavsky (1986), the country studies by Daunton (2001, 2002), Smith (2004), Gillespie (1991), Steinmo (1993), Brownlee (2007), and Lindert (2003), and the recent overview of Latin America in IDB (2006). The latter argument was, for example, the dominant rationale given for lowering income tax rates in the Carter Report (Royal Commission, 1966). Witness the major survey papers by Zodrow (2001) and Fischel (2001): the rate of production has not slowed down, as the references in Augustine et al. (2009) demonstrate. The recent kind words about the growth effects of reliance on property taxes in OECD (2008) are unlikely to result in more use of this tax. Cook (2007) provides a recent review of both research and policy on alcohol in the US. For a more general overview of excise tax research and policy, see Cnossen (2005). The important interaction between tax and regulatory policy is one aspect that clearly needs more attention: the theoretical framework is set out in Christiansen and Smith (2009). For example, Bird and Smart (2001) suggest that in the case of Canada economists did play an important role in persuading the government to move from the archaic manufacturers sales tax to the new VAT (the goods and services tax, GST). Indeed, since reducing economic distortion was the only clear gain from introducing the GST, it may be argued that the government that did so was the first – and probably the last – in Canadian history willing to sacrifice itself on the altar of economic efficiency. Of course, much of the public discussion of the GST in Canada as in other countries focused not on efficiency issues, but on equity issues. Two-handed economists argued on both sides of this issue; again, however, one might argue that their contribution in showing, on the basis of plausible assumptions, that the tax was not seriously regressive helped carry the day. All in all, regardless of what one may think of the GST, its adoption stands as probably the single most important example to date of the influence of economists on Canadian tax policy. For an interesting discussion of the political dimension of the VAT choice in a number of developed countries, see Eccleston 2011. On VAT, see Bird and Gendron (2007) and on the BBLR reforms, see Bird (2011). For example, although surprisingly few tax policy economists seem even to be familiar with Rawls (1971), they would do well to think more deeply about some of the important, and difficult, issues raised in Sen’s (2009) recent impressive reconsideration of the idea of ‘justice’. Or perhaps creating their own country: for example, Rabushka (2008, p. 868) concludes his recent exhaustive recent review of taxation in colonial America with the unequivocal statement that ‘The American Revolution . . . was a tax revolt, first and foremost’. For a good sampler, see Brautigam et al. (2007). Economists have of course also recently done excellent work on the political economy of taxation, as summarized a few years ago by Persson and Tabellini (2000, 2003) and since developed further by them and others. For the most part, however, this work understandably plays to our disciplinary expertise in model-building and econometrics. Apart from a few economic historians (such as Lindert, 2003), however, so far neither development nor public economists have contributed relatively much to the detailed and nuanced country-specific analysis that seems most needed to make progress in this area despite the promising theoretical frameworks set out by such authors as Hettich and Winer (1999) and Wintrobe (1998). This discussion draws heavily on Bird and Gendron (2007). For two early studies, see Tanzi (1970) and Bird (1982). Recently, under the name of

Tax system change and the impact of tax research 447

40.

41.

42. 43. 44. 45. 46.

47.

48.

49.

50. 51.

‘saliency’ theoretical and empirical literature has finally begun to take this important characteristic explicitly into account: see especially Chetty et al. (2007), although it should be noted that my argument here is, in a sense, the reverse side of theirs. One reason some conservatives in the US have resisted VAT and argued for replacing the income tax by a separately-quoted national RST is that they think a ‘hidden’ VAT would lead to a larger government – a belief that evidence from developed countries supports to some extent (Keen and Lockwood, 2006) – while on the other hand they think that replacing the (largely withheld) income tax by a more visible RST may result in smaller government. Such arguments, on both sides, rest more on faith than on evidence. Even thorough-going democrats might prefer some taxes to be hidden for the same reason people support forced saving schemes – as a way of forcing themselves to behave more rationally in the long run (Elster, 1984). Or they may prefer such taxes for paternalistic (or, perhaps, ‘merit good’) reasons similar to those of some supporters of Social Security systems – to force others to behave more rationally (and thus ultimately in their own best interests) in the long run (Musgrave, 1981). The major exception is Canada: see Bird (2010) for a recent examination of the Canadian case. This argument is, for example, one of the key elements in explaining the popularity  (and  mixed results) of fiscal decentralization around the world (Bahl and Bird, 2008b). This discussion draws on Bird and Jun (2007). For example, the IMF, observing how badly public finances are run in many developing countries, commonly insists that an essential element of sound public financial management is a single Treasury account. Whether earmarking increases public trust in government or reduces it depends on the  context. For example, earmarking (hypothecation) was widespread in Britain at the turn of the nineteenth century but was then rejected and replaced in mid-century by the ‘Gladstonian’ approach to public finance, an important feature of which was a consolidated budget with no earmarking (Daunton, 2001). Interestingly, the stated reason for the turn away from earmarking was to restrict the growth of the state in order to restore public trust in the neutrality of the public finances in the face of the then-common perception that hypothecated revenues were being (mis)used by the political elite to expand the ‘fiscal-military’ state in their own interests. As this example suggests, as times change, what is ‘sound’ policy may also change. Partial earmarking may be appropriate even if consumption of a particular public service generates external benefits for other households. In the limiting case in which the service is a pure public good and the marginal cost of extending service to another household is zero there is clearly no efficiency case for either user pricing or earmarking. In theory, properly designed earmarking may in certain circumstances be an effective way to enforce an inter-temporal agreement (Teja, 1988). It may also provide a way to overcome the incomplete contracting problem that would otherwise exist, in effect by increasing the political costs to future governments of breaching the contract (Cremer et al., 1995). In some instances earmarking may thus both provide voters with a way to pin down politicians about whom they are uncertain and politicians with a means of ‘signaling’ their concerns to voters in a credible way (Brett and Keen, 2000). Earmarking tobacco taxes to health programs can be rationalized to some extent on ‘benefit’ grounds but is unlikely to have much effect on health spending. Earmarking such taxes to, for example, ‘anti-tobacco’ advertising campaigns may significantly increase the flow of funds to such activities (Jha and Chaloupka, 2000) but this does not mean it is the best use of such funds. In some contexts, however, even symbolic earmarking may be understandable and justified as a way of signaling commitment and ensuring greater political accountability (Smart and Bird, 2010). Peacock and Wiseman (1961) had earlier explained a similar discrete jump in tax effort

448

The Elgar guide to tax systems

and public expenditure in post-war Britain as a ‘displacement effect’: general perceptions about what is a tolerable level of taxation tend to be stable until shocked by a social upheaval so that levels of taxation that would have been previously intolerable become acceptable and remain at the new higher level after the social perturbations have disappeared. 52. For a neat explanation of this trend, encompassing changes in both economic structure and income inequality, see Bertocchi (2007). 53. This quotation actually had a somewhat different implication in its original context, but nonetheless seems largely right even if one’s main concern is growth: as Lindert (2004, p. 344), concludes, history tells us that ‘the average democracy has been better for economic growth than the average autocracy . . .’. 54. Daunton (2001, 2002) shows that a great deal of attention was paid to precisely this task in Britain, with quite different tax levels and tax mixes being found most suitable to the ‘consensus-maintaining’ objective over the years. Gillespie (1991) tells a similar (more economics-focused) tale for Canada. Lieberman (2003) to some extent tells similar stories with respect to Brazil and South Africa. For a stimulating general model of the balancing of political and economic concerns in formulating and implementing tax policy in a democratic setting, see Hettich and Winer (1999). Tridimas and Winer (2004) in effect extend this framework to non-democratic settings. 55. Some of the present discussion draws on a fuller exploration of the Latin American case in Bird (2003). 56. As discussed earlier, another such reform is earmarking (Bird and Jun, 2007), but this too may instead of improving matters worsen them if it is captured by a particular interest as may happen all too easily even in developed and democratic countries. There is no such thing as a free lunch when it comes to institutional design. 57. Note that one implication is that relevant analysis of the incidence of policy changes must often consider both sides of the budget: as Break (1974) noted, the ‘differential’ tax analysis beloved of economists (and followed in most incidence studies) is seldom the most relevant approach to incidence for policy purposes. 58. Of course, as Hettich and Winer (1999) develop in detail, political and economic factors are often interdependent. 59. Compare the different, but parallel, stories told by Lindert (2003) and Alesina and Angeletos (2003) about how different developed countries have reached quite different fiscal equilibria. Why should uniform outcomes be expected in the much more heterogeneous developing world? 60. Warriner (1969) once noted, despairingly, that many Latin Americans did not seem to know what a good land reform means – probably because they had never seen one. Equally, in most countries of the region (as Engerman and Sokoloff, 2002 almost – but not quite – say) most people may not know what either moderate or justifiable inequality might mean since they have never seen it. 61. Much of what follows relates only to ‘major’ tax reforms (Bird, 2004b). Many countries constantly ‘reform’ their tax systems by altering rates, redefining bases, and adding and clarifying interpretations to existing law. It is not always a simple matter to tell when such ‘technical changes’ constitute a major reform but this issue is not discussed further here. 62. For example, as Yoingco (1976) shows, for many years the Philippines had by far the best developed and institutionalized tax planning process in Asia and in some respects perhaps in the developing world. The results in terms of good policy, however, are not very evident.

Tax system change and the impact of tax research 449

REFERENCES Alesina, A. and G. Angeletos (2003), ‘Fairness and Redistribution: US versus Europe’, Working Paper No. 9502, National Bureau of Economic Research, MA, February. Alm, J., J. Martinez-Vazquez and S. Wallace (eds), Taxing the Hard-to-Tax, Amsterdam: Elsevier. Auerbach, A. (2010), ‘Directions in Tax and Transfer Theory’, paper at Melbourne Institute – Australia’s Future Tax and Transfer Policy Conference (Melbourne: Melbourne Institute of Applied Economic and Social Research), pp. 63–71. Augustine, N. et al. (eds) (2009), Erosion of the Property Tax Base: Trends, Causes, and Consequences, Cambridge, MA: Lincoln Institute of Land Policy. Babcock, L. and G. Loewenstein (1997), ‘Explaining Bargaining Impasse: The Role of SelfServing Biases’, Journal of Economic Perspectives, 11(1), 109–26. Bahl, R. (2006), ‘How to Approach Comprehensive Tax Reform: Have the Rules of the Game Changed?’, in J. Alm, J. Martinez-Vazquez and M.Rider (eds), The Challenges of Tax Reform in a Global Economy, New York: Springer. Bahl, R. and R. Bird (2008a), ‘Tax Policy in Developing Countries: Looking Back – and Forward’, National Tax Journal, 61(2), 279–301. Bahl, R. and R. Bird (2008b), ‘Subnational Taxes in Developing Countries: The Way Forward’, Public Budgeting and Finance, 28(4), 1–25. Barreix, A. and J. Roca (2006), ‘Arquitectura de una propuesta de reforma tributaria’, Universidad Católica de Uruguay, April. Baunsgaard, T. and M. Keen (2005), ‘Tax Revenue and (or?) Trade Liberalization’, Working Paper No. WP/05/112, International Monetary Fund, June. Bergman, M. (2002), ‘Who Pays for Social Policy? A Study on Taxes and Trust’, Journal of Social Policy, 31(2), 289–305. Bergman, M. (2003), ‘Tax Reforms and Tax Compliance: The Divergent Paths of Chile and Argentina’, Journal of Latin American Studies, 35(3), 593–624. Bertocchi, G. (2007), ‘The Vanishing Bequest Tax: The Comparative Evolution of Bequest Taxation in Historical Perspective’, IAZ DP No. 2578, Institute for the Study of Labor, Bonn, January. Best, M. (1976), ‘Political Power and Tax Revenues in Central America’, Journal of Development Economics, 3(1), 49–82. Bird, R. (1974), Taxing Agricultural Land in Developing Countries, Cambridge, MA: Harvard University Press. Bird, R. (1982), ‘Closing the Scissors, or The Real Public Sector has Two Sides’, National Tax Journal, 35(4), 477–81. Bird, R. (2003), ‘Taxation in Latin America: Reflections on Sustainability and the Balance between Efficiency and Equity’, available at http://papers.ssrn.com/sol3/papers. cfm?abstract_id=1393962; accessed 3 May 2011. Bird, R. (2004a), ‘Administrative Dimensions of Tax Reform’, Asia-Pacific Tax Bulletin, 10(3), 134–50. Bird, R. (2004b), ‘Managing Tax Reform’, Bulletin for International Fiscal Documentation, 58(2), 42–55. Bird, R. (2007), ‘Tax Challenges Facing Developing Countries: A Perspective from Outside the Policy Arena’, available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_ id=1393991; accessed 3 May 2011. Bird, R. (2010), ‘Visibility and Accountability – Is Tax-Inclusive Pricing a Good Thing?’ Canadian Tax Journal, 58(1), 63–76. Bird, R. (2011), ‘The BBLR Approach to Tax Reform in Emerging Countries’, in M. G. Rao and M. Rakshit (eds) , Public Economics: Theory and Policy, New Delhi: Sage Publishers, pp. 37–63. Bird, R. and P. Gendron (2007), The VAT in Developing and Transitional Countries, New York: Cambridge University Press.

450

The Elgar guide to tax systems

Bird, R. and J. Jun (2007), ‘Earmarking in Theory and Korean Practice’, in S. Phua (ed), Excise Taxation in Asia, Singapore: National University of Singapore. Bird, R. and J. Mintz (2003), ‘Sharing the International Tax Base in a Changing World’, in S. Cnossen and H. Sinn (eds), Public Finance and Public Policy in the New Century, Cambridge, MA: MIT Press. Bird, R. and M. Smart (2001), ‘Tax Policy and Tax Research in Canada’, in P. Grady and A. Sharpe (eds), The State of Economics in Canada, Montreal: McGill-Queen’s University Press. Bird, R. and T. Tsiopoulos (1997), ‘User Charges for Public Services: Potentials and Problems’, Canadian Tax Journal, 45(1), 25–86. Bird, R. and S. Wallace (2004), ‘Is it Really so Hard to Tax the Hard-to-Tax? The Context and Role of Presumptive Taxes’, in J. Alm, J. Martinez-Vazquez and S. Wallace (eds), Taxing the Hard-to-Tax: Lessons from Theory and Practice, Amsterdam: North-Holland. Bird, R. and S. Wilkie (2000), ‘Source vs. Residence Taxation in the European Union: The Wrong Question?’, in S. Cnossen (ed.), Taxing Capital Income in the European Union, Oxford: Oxford University Press. Bird, R. and E. Zolt (2005), ‘Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries’, UCLA Law Review, 52(6), 1627–95. Bird, R. and E. Zolt (2008), ‘Technology and Taxation in Developing Countries: From Hand to Mouse’, National Tax Journal, 61(4, Part 2), 791–821. Bird, R. and E. Zolt (2010), ‘Dual Income Taxation and Developing Countries’, Columbia Journal of Tax Law, 1(2), 174–217. Bird, R., J. Martinez-Vazquez and B. Torgler (2006), ‘Societal Institutions and Tax Effort in Developing Countries’, in J. Alm, J. Martinez-Vazquez and M. Rider (eds), The Challenges of Tax Reform in a Global Economy, New York: Springer. Bird, R., J. Martinez-Vazquez and B. Torgler (2008), ‘Tax Effort in Developing Countries and High Income Countries: The Impact of Corruption, Voice and Accountability’, Economic Analysis & Policy, 38(1), 55–71. Blyth, M. (2002), Great Transformations: Economic Ideas and Institutional Change in the Twentieth Century, Cambridge: Cambridge University Press. Boylan, Delia M. (1996), ‘Taxation and Transition: The Politics of the 1990 Chilean Tax Reform’, Latin American Research Review, 31(1), 7–31. Brautigam, D., O. Fjeldstad and M. Moore (eds) (2007), Taxation and State Building in Developing Countries: Capacity and Consent, New York: Cambridge University Press. Break, G.F. (1974), ‘The Incidence and Economic Effects of Taxation’, in A. Blinder et al., Economics of Public Finance, Washington, DC: Brookings Institution. Breton, A. (1996), Competitive Governments, Cambridge: Cambridge University Press. Brett, C. and M. Keen (2000), ‘Political Uncertainty and the Earmarking of Environmental Taxes’, Journal of Public Economics, 75(3), 315–40. Bronk, R. (2009), The Romantic Economist: Imagination in Economics, Cambridge: Cambridge University Press. Brownlee, W. (2007), Federal Taxation in America: A Short History, 2nd edn, New York: Cambridge University Press. Buchanan, J. (1963), ‘The Economics of Earmarked Taxes’, Journal of Political Economy, 71(5), 457–69. Casanegra de Jantscher, M. (1990), ‘Administering the VAT’, in M. Gillis, C. Shoup and G.P. Sicat (eds), Value Added Taxation in Developing Countries, Washington, DC: World Bank. Castañeda, J. (1994), Utopia Unarmed: The Latin American Left after the Cold War, New York: Random House. Chelliah, R. (1971), ‘Trends in Taxation in Developing Countries’, International Monetary Fund Staff Papers, 18(1971), 254–331. Chetty, R., A. Looney and K. Kroft (2007), ‘Salience and Taxation: Theory and Evidence’,  Working Paper No. 13330, National Bureau of Economic Research, Cambridge, August.

Tax system change and the impact of tax research 451 Christiansen, V. and S. Smith (2009), ‘Externality-correcting Taxes and Regulation’, WP09/16, Institute of Fiscal Studies. Chu, K., H. Davoodi and S. Gupta (2000), ‘Income Distribution and Tax and Government Social Spending Policies in Developing Countries’, Working Papers No. 214, UNU/ WIDER, December. Clark, C. (1945), ‘Public Finance and Changes in the Value of Money’, Economic Journal, 55(220), 371–89. Clark, C. (1977), ‘The Scope for, and the Limits of, Taxation’, in Institute of Economic Affairs, The State of Taxation, London. Cnossen, S. (ed) (2005), Theory and Practice of Excise Taxation: Smoking, Drinking, Gambling, Polluting and Driving, Oxford: Oxford University Press. Cnossen, S. and M. Smart (2005), ‘Tobacco Taxation’, in S. Cnossen (ed), Theory and Practice of Excise Taxation, Oxford: Oxford University Press. Congdon, W.J., J.R. Kling and S. Mullainathan (2009), ‘Behavioral Economics and Tax Policy’, National Tax Journal, 52(3), 375–86. Cook, P. (2007), Paying the Tab: The Costs and Benefits of Alcohol Control, Princeton, NJ: Princeton University Press. Cremer, J., A. Estache and P. Seabright (1995), ‘The Decentralization of Public Services: Lessons from the Theory of the Firm’, in A. Estache (ed), Decentralizing Infrastructure: Advantages and Limitations, World Bank Discussion Paper No. 290, Washington, DC. Cullis, J., P. Jones and A. Lewis (2010), ‘Tax Compliance: Social Norms, Culture and Endogeneity’, in J. Alm, J. Martinez-Vazquez and B. Torgler (eds), Developing Alternative Frameworks for Explaining Tax Compliance, New York: Routledge. Daunton, M. (2001), Trusting Leviathan: The Politics of Taxation in Britain 1799–1914, Cambridge: Cambridge University Press. Daunton, M. (2002), Just Taxes: The Politics of Taxation in Britain 1914–1979, Cambridge: Cambridge University Press. De Ferranti, D. et al. (2004), Inequality in Latin America: Breaking with History?, Washington, DC: The World Bank. DellaVigna, S. (2009), ‘Psychology and Economics: Evidence from the Field’, Journal of Economic Literature, 47(2), 315–72. Desai, M. and J. Hines (2003), ‘Old Rules and New Realities: Corporate Tax Policy in a Global Setting’, National Tax Journal, 57(4), 937–60. Easterly, W. (2002), The Elusive Quest for Growth, Cambridge MA: MIT Press. Easterly, W. (2009), ‘Can the West Save Africa?’, Journal of Economic Literature, 48(2), 373–447. Eccleston, R. (2007), Taxing Reforms: The Politics of the Consumption Tax in Japan, the United States, Canada and Australia, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Edling, Herbert and David Nguyen-Thanh (2006), ‘Taxes and Culture – Tax Reforms for Sustainable Development’, Fiscal Studies No. 5, Public Finance and Administrative Reform Studies, GTZ, Eschborn. Elster, J. (1984), Ulysses and the Sirens: Studies in Rationality and Irrationality, Cambridge UK: Cambridge University Press. Engel, E., A. Galetovic and C.E. Raddatz (1999), ‘Taxes and Income Distribution in Chile: Some Unpleasant Distributional Arithmetic’, Journal of Development Economics, 59(1), 155–92. Engerman, S. and K. Sokoloff (2002), ‘Factor Endowments, Inequality, and Paths of Development among New World Economies’, Working Paper No. 9259, National Bureau of Economic Research, October. Estache, A. and L. Wren-Lewis (2009), ‘Toward a Theory of Regulation for Developing Countries: Following Jean-Jacques Laffont’s Lead’, Journal of Economic Literature, 47(3), 729–70. Feldstein, M. (1995), ‘The Effects of Marginal Tax Rates on Taxable Income: A Panel Study of the 1986 Tax Reform Act’, Journal of Political Economy, 103(3), 551–72.

452

The Elgar guide to tax systems

Fischel, W. (2001), ‘Homevoters, Municipal Corporate Governance, and the Benefit View of the Property Tax’, National Tax Journal, 54(1), 157–73. Fox, W. and T. Gurley (2005), ‘An Exploration of Tax Patterns around the World’, Tax Notes International, 28 February, 793–807. Frey, B. (2002), The Role of Deterrence and Tax Morale in Taxation in the European Union, Jelle Zijlstra Lecture 1, Netherlands Institute for Advanced Study in the Humanities and Social Sciences. Fullerton, D. (1996), ‘Why Have Separate Environmental Taxes?’, in J. Poterba (ed), Tax Policy and the Economy, Vol. 10, Cambridge, MA: MIT Press. Galbraith, J.K. (1964), Economic Development, Cambridge, MA: Harvard University Press. Gallagher, M. (2005), ‘Benchmarking Tax Systems’, Public Administration and Development, 25(2), 125–44. Gillespie, W.I. (1991), Tax, Borrow, and Spend: Financing Federal Spending in Canada 1867–1990, Ottawa: Carleton University Press. Gillis, M. (ed.) (1989), Tax Reform in Developing Countries, Durham, NC: Duke University Press. Goode, R. (1961), ‘Taxation of Savings and Consumption in Underdeveloped Countries’, National Tax Journal, 14(4), 305–21. Graetz, M.J. and I. Shapiro (2005), Death by a Thousand Cuts: The Fight over Taxing Inherited Wealth, Princeton: Princeton University Press. Gray, C., T. Lane and A. Varoudakis (eds) (2007), Fiscal Policy and Economic Growth: Lessons for Eastern Europe and Central Asia, Washington, DC: World Bank. Harberger, A. (1968), ‘A Landmark in the Annals of Taxation’, Canadian Journal of Economics, 1(1, Supplement), 183–94. Harberger, A. (1971), ‘Three Basic Postulates for Applied Welfare Economics’, Journal of Economic Literature, 9(3), 785–91. Harberger, A. (2006), ‘Taxation and Income Distribution: Myths and Realities’, in J. Alm, J. Martinez-Vazquez and M. Rider (eds), The Challenges of Tax Reform in a Global Economy, New York: Springer. Heady, C. (2004), ‘Taxation Policy in Low-income Countries’, in T. Addison and A. Roe (eds), Fiscal Policy for Development: Poverty, Reconstruction and Growth, London: Palgrave. Heady, C. (2010), ‘Directions in Overseas Tax Policy’, paper at Melbourne Institute – Australia’s Future Tax and Transfer Policy Conference (Melbourne; Melbourne Institute of Applied Economic and Social Research), pp. 4–19. Hettich, W. and S. Winer (1999), Democratic Choice and Taxation, Cambridge: Cambridge University Press. Inter-American Development Bank (IDB) (2006), The Politics of Policies, Washington, DC. International Monetary Fund (IMF) (2005), ‘Monetary and Fiscal Policy Design Issues in Low-Income Countries’, August. Ivanova, A., M. Keen and A. Klemm (2005), ‘The Russian Flat Tax Reform’, IMF Working Paper No. 05/16, Washington, DC: International Monetary Fund. Jha, P. and F. Chaloupka (eds) (2000), Tobacco Control in Developing Countries, Oxford: Oxford University Press. Kaldor, N. (1963), ‘Will Underdeveloped Countries Learn to Tax?’, Foreign Affairs, 41(2), 410–19. Keen, M. (2007), ‘VAT Attacks! Second-best Perspectives on the Value Added Tax’, International Tax and Public Finance, 14(4), 35–81. Keen, M. and B. Lockwood (2006), ‘Is the VAT a Money Machine?’, National Tax Journal, 49(4), 905–28. Keen, M. and A. Simone (2004), ‘Tax Policy in Developing Countries: Some Lessons from the 1990s and Some Challenges Ahead’, in S. Gupta, B. Clements and G. Inchauste (eds), Helping Countries Develop – The Role of Fiscal Policy, Washington, DC: International Monetary Fund. Kenny, L.W. and S.L. Winer (2006), ‘Tax Systems in the World: An Empirical Investigation

Tax system change and the impact of tax research 453 into the Importance of Tax Bases, Administration Costs, Scale and Political Regime’, International Tax and Public Finance, 13(2/3), 181–216. Keynes, J. (1936), The General Theory of Employment, Interest and Money, New York: Harcourt, Brace. Khan, M. (2001), ‘Agricultural Taxation in Developing Countries: A Survey of Issues and Policy’, Agricultural Economics, 24(3), 315–28. Levi, M. (1988), Of Rule and Revenue, Berkeley: University of California Press. Lieberman, E. (2003), Race and Regionalism in the Politics of Taxation in Brazil and South Africa, Cambridge: Cambridge University Press. Lindert, P. (2003), Growing Public: Social Spending and Economic Growth Since the Eighteenth Century, New York: Cambridge University Press. Lindert, P. (2004), ‘Voice and Growth; Was Churchill Right?’, Journal of Economic History, 63(2), 315–50. Lledo, V., A. Schneider and M. Moore (2003), ‘Pro-poor Tax Reform in Latin America: A Critical Survey and Policy Recommendations’, Institute of Development Studies, Brighton, March. Lledo, V., A. Schneider and M. Moore (2004), ‘Governance, Taxes, and Tax Reform in Latin America’, IDS Working Paper No. 221, Institute of Development Studies, Brighton. Lora, E. and M. Cardenas (2006), ‘La reforma de las instituciones fiscales en América Latina’, Inter-American Development Bank, Research Department, Documento de Trabajo No. 559, April Mankiw, G., M. Weinzierl and D. Yagan (2009), ‘Optimal Taxation in Theory and Practice’, NBER Working Paper No. 15071, Cambridge, MA, June. Martinez-Vazquez, J. (2001), ‘Mexico: An Evaluation of the Main Features of the Tax System’, Working Paper No. 01–12, International Studies Program, Andrew Young School of Policy Studies, Georgia State University, November. Martinez-Vazquez, J. and R. McNab (2000), ‘The Tax Reform Experiment in Transitional Countries’, National Tax Journal, 53(2), 273–98. Martinez-Vazquez, J., M. Rider and S. Wallace (2008), Tax Reform in Russia, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. McCleary, W. (1991), ‘The Earmarking of Government Revenue: A Review of Some World Bank Experience’, World Bank Research Observer, 6, 81–104. McIntyre, M. and O. Oldman (1975), Institutionalizing the Process of Tax Reform: A Comparative Analysis, Amsterdam: International Bureau of Fiscal Documentation. McLaren, J. (ed) (2003), Institutional Elements of Tax Design, Washington, DC: World Bank. McLure, C. (1990), ‘Review of Newbery and Stern’, Economic Development and Cultural Change, 38(2), 426–31. McLure, C. and G. Zodrow (2007), ‘Consumption-based Direct Taxes: A Guided Tour of the Amusement Park’, FinanzArchiv/Public Finance Analysis, 63(2), 285–307. Mintz, J. (2003), ‘Growth and Taxes: Some Implications for Developing Economies’, in J. Martinez-Vazquez and J. Alm (eds), Public Finance in Developing and Transition Countries, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Moore, M. (2007), ‘How Does Taxation Affect the Quality of Governance?’, Tax Notes International, 47(July 2), 79–98. Musgrave, R. (1969), Fiscal Systems, New Haven: Yale University Press. Musgrave, R. (1981), ‘A Reappraisal of Financing Social Security’, in F. Skidmore (ed), Social Security Financing, Cambridge MA: MIT Press. Musgrave, R. (2000), Public Finance in a Democratic Society: Volume III. The Foundations  of  Taxation and Expenditure, Cheltenham UK, Northampton, MA, USA: Edward Elgar. Newbery, D. and N. Stern (eds) (1987), The Theory of Taxation in Developing Countries, published for the World Bank by Oxford University Press. Norregaard, J. and T. Khan (2007), ‘Tax Policy: Recent Trends and Coming Challenges’, WP/07/274, International Monetary Fund, Washington, DC, December

454

The Elgar guide to tax systems

OECD (Organisation for Economic Co-operation and Development) (2006), ‘Reforming Personal Income Tax’, Policy Brief, March. OECD (2008), ‘Tax and Economic Growth’, Economic Department Working Paper No. 620, July. OECD (2009), Tax Administration in OECD and Selected Non-OECD Countries: Comparative Information Series (2008), Paris: OECD. Peacock, A. and J. Wiseman (1961), The Growth of Government Expenditures in the United Kingdom, Princeton: Princeton University Press. Persson, T. and G. Tabellini (2000), Political Economics: Explaining Economic Policy, Cambridge, MA: MIT Press. Persson, T. and G. Tabellini (2003), The Economic Effects of Constitutions: What do the Data Say?, Cambridge, MA: MIT Press. Peter, K. S. Buttrick and D. Duncan (2009), ‘Global Reform of Personal Income Taxation, 1981–2005: Evidence from 189 Countries’, IZA Discussion Paper No. 42228. Poirson, H. (2006), ‘The Tax System in India: Could Reforms Spur Growth?’, IMF Working Paper No. WP/06/93, Washington, DC, April. Pritchard, W. (2009), ‘The Politics of Taxation and Implications for Accountability in Ghana 1981–2008’, Working Paper No. 330, Institute of Development Studies, July. Rabushka, A. (2008), Taxation in Colonial America, Princeton, NJ: Princeton University Press. Radian, A. (1980), Resource Mobilization in Poor Countries: Implementing Tax Reform, New Brunswick, NJ: Transaction Books. Ravaillon, M. (2009), ‘Do Poorer Countries Have Less Capacity for Redistribution?’, Policy Research Working Paper No. 5046, World Bank, September. Rawls, John (1971), A Theory of Justice, Cambridge, MA: Harvard University Press. Royal Commission on Taxation (1966), Report, six vols, Ottawa. Rutherford, T., M. Light and F. Barrera (2005), ‘Equity and Efficiency Costs of Raising Tax Revenue in Colombia’, in R. Bird, J. Poterba and J. Slemrod (eds), Fiscal Reform in Colombia: Problems and Prospects, Cambridge, MA: The MIT Press. Schumpeter, J. (1934), Theory of Economic Development, Cambridge, MA: Harvard University Press. Schumpeter, J. (1942), Capitalism, Socialism and Democracy, New York: Harper & Brothers. Schumpeter, J.A. (1954), History of Economic Analysis, New York: Oxford University Press. Sen, Amartya (2009), The Idea of Justice, Cambridge MA: The Belknap Press of Harvard University Press. Slemrod, J. (1990), ‘Optimal Taxation and Optimal Tax Systems’, Journal of Economic Perspectives, 4(1), 1257–78. Slemrod, J. and E. Bakija (1996), Taxing Ourselves: A Citizen’s Guide to the Great Debate on Tax Reform, Cambridge MA: MIT Press. Smart, M. and R. Bird (2010), ‘Earmarked Grants and Accountability in Government’, in J. Kim, J. Lotz and N.J. Mau (eds), General Grants Versus Earmarked Grants – Theory and Practice, Copenhagen and Seoul: Korea Institute of Public Finance and Danish Ministry of Interior and Health, pp. 40–73. Smith, J. (2004), Taxing Popularity: The Story of Taxation in Australia, Sydney: Australian Tax Research Foundation. Sokoloff, K.L. and E.M. Zolt (2005), ‘Inequality and Taxation: Evidence from the Americas on How Inequality may Influence Tax Institutions’, Tax Law Review, 59(2), 167–241. Steinmo, S. (1993), Taxation and Democracy: Swedish, British, and American Approaches to Financing the Modern State, New Haven: Yale University Press. Tanzi, V. (1970), ‘International Tax Burdens: A Study of Tax Ratios in the OECD Countries’, in Taxation: A Radical Approach, London: Institute for Economic Affairs. Tanzi, V. (1991), Public Finance in Developing Countries, Aldershot, UK and Brookfield, VT, USA: Edward Elgar. Tanzi, V. (1995), Taxation in an Integrating World, Washington, DC: Brookings Institution.

Tax system change and the impact of tax research 455 Tanzi, V. and H. Zee (2000), ‘Tax Policy for Emerging Markets: Developing Countries’, National Tax Journal, 53(2), 299–328. Teja, R. (1988), ‘The Case for Earmarked Taxes’, International Monetary Fund Staff Papers, 35, 523–33. Thirsk, W. (1997), Tax Reform in Developing Countries, Washington, DC: World Bank. Tridimas, G. and S.L. Winer (2004), ‘The Political Economy of Government Size’, European Journal of Political Economy, 21(3), 643–66. UN Millennium Project (2005), Investing in Development, New York: OUP and World Bank. Warriner, D. (1969), Land Reform in Theory and Practice, Oxford: Clarendon Press. Webber, C. and A. Wildavsky (1986), A History of Taxation and Expenditure in the Western World, New York: Simon and Schuster. Weisman, S. (2002), The Great Tax Wars, New York: Simon and Schuster. Wildavsky, A. (1979), Speaking Truth to Power: The Art and Craft of Policy Analysis, Boston: Little, Brown. Wintrobe, R. (1998), The Political Economy of Dictatorship, Cambridge: Cambridge University Press. Yoingco, A.Q. (1976), ‘Tax Research: The Philippine Experience’, Graduate School, Centro Escolar University, Manila, March. Zodrow, G. (2001), ‘The Property Tax as a Capital Tax: A Room with Three Views’, National Tax Journal, 54(1), 139–56.

Index

accessions tax 185 accidental bequests 199–200, 202 ad valorem tax 5, 288–90, 293, 297, 308 Agencia Tributaria 366 agglomeration external effects 139 allowance for corporate capital (ACC) 165, 167 allowance for corporate equity (ACE) 133–6, 165–8, 171–4, 176 altruistic bequests 197 apportionment formula (AF) 158 Arbitration Convention 154–5 Atkinson–Stiglitz theorem 45, 99, 119, 192–6 automatic stabilizers 3, 62–3, 74, 330 average effective tax rate (AETR) 71–3, 84, 87, 140–41, 162 base broadening 4, 132, 144–5, 159–60, 162, 174, 249, 251, 267, 335 British Institute for Fiscal Studies (IFS) 133 broad-based low-rate (BBLR) 430, 446 Bureau of Economic Analysis 71, 73, 84–5 business cycle 4, 62, 353 business value tax (BVT) 134, 137, 165–6, 168–9, 171 Canada Revenue Agency 364 canonical model 196–8 capital accumulation 46, 48, 97, 120, 123, 184, 190–91, 193, 201, 209, 211, 214, 216 capital export neutrality (CEN) 156, 424 capital flight 25 capital gains 19, 25, 94, 98, 101, 107, 115, 117, 142, 145–6, 148–9, 150, 155–6, 158, 160, 164–7, 169, 171, 175, 202, 423–4, 445

realized 25 unrealized 25 capital income 1, 3, 8, 25, 94–101, 103, 107–8, 110–11, 117–23, 127, 150, 157, 164, 166, 174–5, 192–6, 199, 202–3, 208, 210–14, 346, 398, 406, 420, 424, 426–7, 432 capital levies 139, 185 capital taxation 100 cash-flow statements (C-FS) 170 cash-flow tax 112, 134–5, 165–8, 173 chains of informality 261 Chamley-Judd theorem 192, 195–7, 203 CIAT Award for Innovation in Tax Administration 363 Clark, Colin 16, 444 clean good 311 commodity tax 46, 98, 300, 313 common consolidated corporate tax base (CCCTB) 157–9, 164 company income tax 349 comprehensive business income tax (CBIT) 134, 165–7, 169, 171 comprehensive income tax 25, 111, 117, 127 conscience tax 315–16 consumption smoothing 5, 184, 190, 192, 196 consumption tax 43, 75, 100, 104, 108, 111–12, 119–21, 166, 221, 249, 264–5, 280, 341, 344, 397, 400, 424, 429, 440 domestic consumption tax 43 general consumption tax 249, 264–5, 429 contraction phase 62 Corlett and Hague Theorem 98, 280 corporate income tax (CIT) 4, 7, 17, 26–31, 50, 52, 84, 131–9, 141–2, 144–51, 153–4, 157– 69, 171–5, 275, 349, 430 credit for corporate equity (CCE) 136

457

458

The Elgar guide to tax systems

data envelopment analysis (DEA) 371 Deloitte 366 demonstration effect 20 dirty good 308, 311 dividend tax 94, 107–8, 147 double dividend 6, 301, 310–12, 319, 321 dual income tax (DIT) 3, 94, 101, 110, 121–2, 134, 136, 164–6, 171, 176 dynamic efficiency 48, 210, 320 Ecofin Council 154 economic growth rate 62 economic rents 131, 133, 136, 162, 165, 167 effective tax rate 31, 44, 71, 73, 84, 87, 130, 139–40, 262, 385, 427 environmental tax 1, 6, 8, 294, 300–27, 368 equity 1, 3–4, 20, 32, 38, 45, 58, 66, 93, 95–6, 109, 114, 116–20, 122–5, 131, 133–7, 141, 146–7, 161, 165–9, 171–3, 176, 183–4, 200, 202, 218, 222, 235, 262–3, 267, 286, 331, 339, 344, 422–3, 427–9, 431–2, 441, 446 horizontal equity 45, 109, 120, 427–8, 431 vertical equity 66, 263, 339, 344, 431 Ernst & Young 366 estate tax 185–6, 188–9, 194, 198, 200–204, 215 EU Savings Tax Directive of 2005 157 EURO SAI 368 European Court of Justice (ECJ) 130, 132, 156–7 European Union (EU) 4, 19, 31–2, 70, 130, 132, 134–8, 142–5, 148–51, 153–60, 164, 175, 186, 202, 244, 250, 252, 256–7, 264, 267–8, 290, 293, 296–7, 392, 396–7, 401, 427 exchange-based bequests 198 excise tax 247, 290, 342, 344, 356, 446 expansion phase 62 expenditure programs 105, 292 fiscal decentralization 2, 6, 30, 52–5, 84, 87, 91, 127, 328–33, 335, 340, 346, 352–6, 440–41, 447

fiscal federalism 30, 127, 330, 332, 356 flat rate tax (FRT) 8, 25–6, 121, 167, 347, 421 flow-of-funds corporation tax 133–5, 165, 169–70 foreign direct investment (FDI) 3, 37, 39, 57, 70–74, 78, 84–5, 87, 139, 141, 162–3, 167 forms of taxation 5, 37–9, 45, 74, 318, 378, 384 direct tax 3, 50, 58, 63, 92, 101, 117, 124, 154 indirect tax 1–3, 38–41, 46–52, 54–5, 57–60, 62–3, 66–74, 76, 92, 314 Forum on Tax Administration (FTA) 367 Friedman, Milton 20 Gastil index 86, 383–4, 387–8, 390–91, 404, 411–12 General Accountability Office (GAO) 255, 268, 363 general consumption tax 249, 264–5, 429 general equilibrium analysis 379 general sales tax 37, 237, 244, 275, 277, 340, 350 generalized method of moments (GMM) 232, 239–41, 243 gift tax 185 globalization 2, 4, 19–20, 25, 34, 51, 53, 56–7, 70, 76, 85, 87, 92, 130, 132, 138, 159, 163–4, 174, 237, 377, 401, 405–6, 413, 420, 425, 428 goods and services tax (GST) 237, 244–6, 248–9, 266, 446 Government Finance Statistics (GFS) 40–44, 50, 75, 84, 86, 89–90, 221–2, 229–30, 236–7, 275, 330 gross domestic product (GDP) 3, 5–6, 11–17, 22, 27–30, 35, 47–55, 57–8, 60, 63–9, 71–3, 85–8, 91–2, 138, 142–3, 160, 174–5, 185–6, 188, 207, 224–5, 232–3, 235–9, 242–4, 247, 250–51, 256, 265–6, 268, 275–7, 302–3, 316, 318, 322, 350, 353–4, 356, 385–7, 390–91, 395, 414–19, 439, 444–5

Index Harmful Tax Practices Initiative 153–4 horizontal equity 45, 109, 120, 427–8, 431 human capital investment 4, 98, 112–13, 119, 121, 123 immediate expensing of investments (ICBIT) 165–7 Impuesto Empresarial aTasa Unica 135 indirect tax 1–3, 38–41, 46–52, 54–5, 57–60, 62–3, 66–74, 76, 92, 314 inflation 30, 58, 61–3, 77, 85, 87, 98, 120, 142, 144, 160–62, 169, 173, 297, 384 inheritance and donation tax (IDT) 153 inter vivos transfer 100–101, 124, 185, 202 Inter-American Center of Tax Administrations 368 Internal Revenue Service (IRS) 168 International Accounting Standard 7, 33, 361, 363 International Financial Reporting Standards (IFRS) 145 International Monetary Fund (IMF) 7, 40, 50, 84, 86, 89, 221, 223–4, 230, 241, 247, 251, 256, 264–5, 267, 275–7, 330, 356, 367, 414, 427, 430, 444, 447 international tax competition 2, 20, 130, 132, 137–8, 401 International Tax Dialogue 367 Intra-European Organisation of Tax Administration 368 inventory valuation 144 last in, first out (LIFO) 144 first in, first out (FIFO) 145 IRS Oversight Board 363 IRS Reform and Restructuring Act of 1998 361 Italian regional tax on productive activities (IRAP) 136–7, 168, 268 key performance indicators (KPIs) 368, 372 KPMG 366

459

Laffer curve 20, 380, 382 land tax 337 large-scale enterprises (LSEs) 151–2 lifestyle tax 319 Luxembourg Income Study 389 macro apportionment system 159 macroeconomic stability 37, 39, 49, 62 marginal cost of public funds (MCPF) 297, 308, 430 marginal effective tax rate (METR) 140–41, 161–2 marginal rate of substitution (MRS) 195, 295 marginal rate of transformation (MRT) 195 marginal tax rate 20, 22, 25, 35, 94, 123, 147, 166, 307, 309, 426 market failure 309 Meade Committee 133, 135, 174 Meade Report 96 median voter theorem 398, 405 Merger Directive 155 multi-level budget framework 330 National Taxpayer Advocate 363 net wealth tax 185 new public management 360 notional interest deduction 135, 173 OECD Committee on Fiscal Affairs 367 OECD Comparative Information Series 365, 367–8, 373 Office of the Inspector-General of Taxation 363 opportunity cost 252 optimal tax 38–9, 45–6, 74, 98–9, 118, 120, 122, 175, 184, 208, 210–11, 280, 288, 301, 308–9, 325–6, 404, 425–7 Organisation for Economic Co-operation and Development (OECD) 2, 4, 6–8, 11–35, 43, 47–8, 77, 94, 106, 116, 130, 132, 135, 137–8, 142–5, 147, 152–5, 159–60, 165, 167, 171, 174–6, 185, 221, 231, 237, 251, 256, 266–8, 279, 295, 302, 304, 321, 333, 340, 347, 353–6, 364–70, 373, 389–90,

460

The Elgar guide to tax systems

396–7, 412–14, 419–22, 426–7, 431, 441, 446 overhead cost 33 overlapping generations model (OLG) 192–4, 201, 208 Parent–Subsidiary Directive 135, 155 Pareto optimality 307 Pareto-efficient 46 partial equilibrium analysis 300 paternalistic bequests 197, 199 payroll tax 41–4, 50, 86, 96, 221–2, 311–12, 348–9, 356 personal income tax 3–4, 17, 19–22, 30, 52, 57, 75, 136, 141, 146, 149–50, 275, 345, 347–8, 384, 391, 394, 418, 426 Pigou, A.C. 301, 320 Pigouvian prescription 281 Pigouvian tax 296, 301, 303, 315–16 political economy 2–3, 5–7, 100, 126, 201, 218, 231, 265, 318, 330–31, 376–7, 393, 396, 401–2, 404, 422, 446 political equilibria 380 pre-populated tax returns 363 principle of compensation 104 principle of responsibility 104 progressive tax 19, 25, 66, 94, 104, 195 property tax 7, 50, 63, 86, 334, 337, 340–41, 343–4, 352–4, 356, 429 Proposition 13 334 protective interest 136 Public Accounts Committee of Parliament 363 PWC 366, 368, 373 race to the bottom 138, 345 Ramsey elasticity rule 307 Ramsey rules 46 Ramsey, Frank 307 rate–revenue paradox 132, 137, 142–4 refundable tax credit 95, 102, 105–6, 115–16, 123, 125, 427 regressive tax 2, 33–4, 103 retail sales tax (RST) 219, 249, 254, 344, 350–51, 447 revenue adequacy 339

risk management 7, 363, 365–8, 370, 373 Rossotti, Charles 361 scale effects 51–2, 382 schedular income tax 3, 25, 94, 96, 103, 118, 121, 160, 164 seemingly unrelated regressions (SURs) 244–6, 248 shareholder taxation 130, 132–3, 141, 146–7, 155–6, 160–61, 175 small and medium-sized enterprises (SMEs) 132, 135, 144, 151–3, 157–60, 163, 170 Smith, Adam 360 social engineering 5, 18 social insurance programs 105 Social Security tax 1–3, 8, 17, 22, 38, 40–44, 50, 73, 86, 89, 137–8, 191, 194, 201, 208, 220–22, 246–8, 275, 277, 279, 348, 383–4, 390–91, 392, 396–7, 403–5, 411, 414, 435, 447 solidarity tax on wealth 185 spirit of capitalism theory 190 stabilization policy 18 static production efficiency 320 Steinmo, Sven 392 stochastic frontier analysis 371 Study Group on Asian Tax Administration and Research (SGATAR) 368 supply-side revolution 20, 26 tax accessions tax 185 ad valorem tax 5, 288–90, 293, 297, 308 business value tax (BVT) 134, 137, 165–6, 168–9, 171 capital taxation 100 cash-flow tax 112, 134–5, 165–8, 173 commodity tax 46, 98, 300, 313 company income tax 349 comprehensive business income tax (CBIT) 134, 165–7, 169, 171 comprehensive income tax 25, 111, 117, 127 conscience tax 315–16 consumption tax 43, 75, 100, 104, 108, 111–12, 119–21, 166, 221,

Index 249, 264–5, 280, 341, 344, 397, 400, 424, 429, 440 corporate income tax (CIT) 4, 7, 17, 26–31, 50, 52, 84, 131–9, 141–2, 144–51, 153–4, 157–69, 171–5, 275, 349, 430 dividend tax 94, 107–8, 147 domestic consumption tax 43 dual income tax (DIT) 3, 94, 101, 110, 121–2, 134, 136, 164–6, 171, 176 environmental tax 1, 6, 8, 294, 300–27, 368 estate tax 185–6, 188–9, 194, 198, 200–204, 215 excise tax 247, 290, 342, 344, 356, 446 flat rate tax (FRT) 8, 25–6, 121, 167, 347, 421 flow-of-funds corporation tax 133–5, 165, 169–70 general consumption tax 249, 264–5, 429 general sales tax 37, 237, 244, 275, 277, 340, 350 gift tax 185 goods and services tax (GST) 237, 244–6, 248–9, 266, 446 indirect tax 1–3, 38–41, 46–52, 54–5, 57–60, 62–3, 66–74, 76, 92, 314 inheritance and donation tax (IDT) 153 land tax 337 lifestyle tax 319 net wealth tax 185 optimal tax 38–9, 45–6, 74, 98–9, 118, 120, 122, 175, 184, 208, 210–11, 280, 288, 301, 308–9, 325–6, 404, 425–7 payroll tax 41–4, 50, 86, 96, 221–2, 311–12, 348–9, 356 personal income tax 3–4, 17, 19–22, 30, 52, 57, 75, 136, 141, 146, 149–50, 275, 345, 347–8, 384, 391, 394, 418, 426 Pigouvian tax 296, 301, 303, 315–16 property tax 7, 50, 63, 86, 334, 337, 340–41, 343–4, 352–4, 356, 429

461

regressive tax 2, 33–4, 103 retail sales tax (RST) 219, 249, 254, 344, 350–51, 447 schedular income tax 3, 25, 94, 96, 103, 118, 121, 160, 164 Social Security tax 1–3, 8, 17, 22, 38, 40–44, 50, 73, 86, 89, 137–8, 191, 194, 201, 208, 220–22, 246–8, 275, 277, 279, 348, 383–4, 390–91, 392, 396–7, 403–5, 411, 414, 435, 447 transfer tax 185–6, 198–203, 212, 217 unified tax 185 wealth transfer tax 94, 183–203, 207–17 withholding tax 94, 135, 147, 156–7 tax administration 2, 7, 15, 25, 45, 51, 100, 163, 220, 225, 231, 250, 254, 257, 259–60, 262, 268, 329, 331–2, 336–7, 342–3, 347, 349–50, 355–6, 360–61, 363–9, 371–3, 418, 422–3, 432, 445 tax base 6, 8, 20, 22, 26, 30–31, 34–5, 44, 51–2, 54–7, 91–8, 100, 102, 104, 107, 109–11, 114–21, 123, 126–7, 132, 136–7, 141–4, 150–51, 154, 157–61, 166, 168–70, 175, 188, 202, 220, 224, 247, 251–2, 259, 268, 294, 304–6, 313, 321, 328–30, 335, 337, 339–43, 345–50, 356, 376, 379–81, 383, 385, 397, 400–401, 416–17, 419–20, 426, 430, 434, 445 global tax base 30 national tax base 30 tax competition 2, 4, 6, 20, 25, 27, 30–31, 34, 130–32, 137–9, 141–2, 144, 154, 159, 161, 163, 175, 202–3, 377, 401 tax culture 423 tax evasion 2, 19–20, 31, 35, 38, 45, 75, 96, 103, 117, 157, 203, 231, 257–8, 314 tax expenditures 18–20, 26, 32 tax fraud 19, 366 tax handles 51, 419, 445 tax haven 19, 25, 31, 143, 154–5, 157 tax incentives 20, 32, 93, 111–13, 145, 151, 161, 304, 418

462

The Elgar guide to tax systems

tax morale 5, 225, 231–5, 267, 275, 277, 366, 416–17, 423 tax ratio 12, 39–41, 46–7, 50–52, 54–5, 57–60, 62–74, 76–8, 86, 88, 91, 236, 390, 396–7, 399, 414–16, 419, 438–9, 444–5 Tax Reform Act of 1986 (TRA–86) 48 tax skeleton 376–7, 382, 385, 404 top marginal tax rates (TMTR) 23–5 transfer tax 185–6, 198–203, 212, 217 transferor-based 185 recipient based 185 transferable quotas 312–13 Treasury Inspector General of Tax Administration 363 trinity reform 334 UN Millennium Project 414 unified tax 185 value added tax (VAT) 1, 18–19, 28, 37, 75, 93, 218–69, 277–8, 340, 350, 355, 427 sub-central VAT 253

administration 254 evasion 255 vertical equity 66, 263, 339, 344, 431 veto player 378–9, 393–6, 403, 405, 412 wealth and inheritance taxation 184 net wealth tax 183–5 transfer tax 94, 183–203, 207–17 wealth transfer 4, 94, 114–16, 121, 183–203, 207–17 involuntary wealth transfer 114 voluntary wealth transfer 114 wealth transfer tax 94, 183–203, 207–17 Weber, Max 190 Wicksell–Lindahl framework 388, 435, 439 withholding tax 94, 135, 147, 156–7 World Bank 42–4, 110, 221–2, 260, 276, 355, 368 World Tax Organization (WTO) 32, 242 World Trade Organization (WTO) 427