Tax Reform in the 21st Century: A Volume in Memory of Richard Musgrave 9041128298, 9789041128294

Citation preview

John G. Head and Richard Krever (eds.)

Tax Reform in the 21st Century A Volume in Memory of Richard Musgrave

Series on International Taxation

34

Tax Reform in the 21st Century

Series on International Taxation VOLUME 34 Founding editors Richard Doernberg, Kees van Raad and Klaus Vogel y Senior editor Professor Dr. Kees van Raad, International Tax Center, University of Leiden Managing editors Professor Ruth Mason, University of Connecticut School of Law Professor Dr. Ekkehart Reimer, University of Heidelberg

The title published in this series are listed at the end of this volume.

Tax Reform in the 21st Century A Volume in Memory of Richard Musgrave

Edited by

John G. Head and Richard Krever

Law & Business AUSTIN

BOSTON

CHICAGO

NEW YORK

THE NETHERLANDS

Published by: Kluwer Law International P.O. Box 316 2400 AH Alphen aan den Rijn The Netherlands Website: www.kluwerlaw.com

Sold and distributed in North, Central and South America by: Aspen Publishers, Inc. 7201 McKinney Circle Frederick, MD 21704 United States of America Email: [email protected]

Sold and distributed in all other countries by: Turpin Distribution Services Ltd. Stratton Business Park Pegasus Drive, Biggleswade Bedfordshire SG18 8TQ United Kingdom Email: [email protected]

Printed on acid-free paper.

web-ISBN 978-90-411-5506-1 # 2009 Kluwer Law International BV, The Netherlands

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, photocopying, recording or otherwise, without prior written permission of the publisher.

Permission to use this content must be obtained from the copyright owner. Please apply to: Permissions Department, Wolters Kluwer Legal, 76 Ninth Avenue, 7th Floor, New York, NY 10011-5201, USA. Email: [email protected].

Printed in Great Britain.

Table of Contents Editors and Contributors

ix

Introduction and Dedication

xi

REFLECTION Remembering Richard Musgrave, 1910–2007 Peggy B. Musgrave

3

PURSUING TAX REFORM The Political Economy of Tax Reform: A Neo-Musgravian Perspective with Illustrations from Canadian, US, Australian and New Zealand Experience John G. Head

19

International Prospects for Consumption-Based Direct Taxes: A Guided Tour Charles E. McLure, Jr and George R. Zodrow

51

Taxing Corporations in the European Union: Towards a Common Base? Sijbren Cnossen

73

Income and Consumption Taxes in New Zealand: The Political Economy of Broad-Base, Low-Rate Reform in a Small Open Economy David White

95

v

vi

TAX REFORM IN THE 21ST CENTURY

PERSONAL TAX BASE: INCOME OR CONSUMPTION? Income or Consumption Taxes? Alan J. Auerbach

147

Consumption Taxes and Risk Revised Jane G. Gravelle

167

TAX RATE SCALE: EQUITY AND EFFICIENCY ASPECTS Taxation, Labour Supply and Saving Patricia Apps and Ray Rees

187

The Distributional Effect of Consumption Taxes in Tax Systems Neil Warren

217

A Restatement of the Case for a Progressive Income Tax Neil Brooks

277

BUSINESS TAX REFORM: STRUCTURAL AND DESIGN ISSUES Corporate Income Tax: Incidence, Economic Effects, and Structural Issues Jane G. Gravelle

355

The Deduction of Interest Payments in an Ideal Tax on Realized Business Profits Michael J. McIntyre

385

The Mirrlees Review: A Perspective on Fundamental Tax Reform Malcolm Gammie

407

Taxes or Tradable Permits to Reduce Greenhouse Gas Emissions John Freebairn

421

INTERJURISDICTIONAL ISSUES Tax Assignment Revisited Richard Bird

441

Inter-Nation Equity: The Development of an Important but Underappreciated International Tax Policy Objective Kim Brooks

471

TABLE OF CONTENTS INTRODUCTION AND DEDICATION

Taxation of Outbound Direct Investment: Economic Principles and Tax Policy Considerations Michael P. Devereux

vii

499

CONTROLLING TAX AVOIDANCE Containing Tax Avoidance: Anti-Avoidance Strategies Chris Evans

529

Editors and Contributors PATRICIA APPS University of Sydney ALAN J. AUERBACH University of California, Berkeley RICHARD BIRD University of Toronto NEIL BROOKS Osgoode Hall Law School, Toronto KIMBERLEY BROOKS McGill University SIJBREN CNOSSEN Erasmus University Rotterdam and University of Maastricht MICHAEL DEVEREUX Centre for Business Taxation, Oxford University CHRIS EVANS University of New South Wales JOHN FREEBAIRN University of Melbourne MALCOLM GAMMIE Practising barrister at One Essex Court, Temple, and Research Director, IFS’ Tax Law Review Committee JANE G. GRAVELLE Congressional Research Service, US Library of Congress ix

x

TAX REFORM IN THE 21ST CENTURY

JOHN G. HEAD Monash University RICHARD KREVER Monash University CHARLES E. McLURE, Jr Hoover Institution, Stanford University MICHAEL J. McINTYRE Wayne State University Law School PEGGY B. MUSGRAVE University of California at Santa Cruz RAY REES University of Munich NEIL WARREN University of New South Wales DAVID I. WHITE Victoria University of Wellington GEORGE R. ZODROW Baker Institute for Public Policy, Rice University

Introduction and Dedication

This volume is dedicated to the memory of our dear friend and revered colleague Richard Musgrave who died at the age of 96 on January 15, 2007. The papers in the volume were presented at the Musgrave Memorial Conference and Colloquium held in Sydney from June 2–6, 2008. We were fortunate indeed to have Peggy Musgrave, his wife and close professional collaborator for 43 years, to open proceedings. Other presenters included leading public finance economists and tax lawyers from universities in Australia, Canada, the United Kingdom and the United States. We were fortunate also in having a particularly well informed audience with heavy representation from the Australian and New Zealand Treasury. The result was a very high quality discussion with lively interaction and considerable diversity of opinion between the different groups represented. Musgrave himself was feisty, but also fair and probing in debate; and as Peggy observes, he would no doubt have much enjoyed the occasion. Richard Musgrave was without question the greatest public finance economist of his generation – among a fine group of scholars and policy advisers that included Richard Goode, Joseph Pechman, Carl Shoup and William Vickrey. Over a 33-year period he was a professor of economics successively at the University of Michigan, Johns Hopkins University, Princeton University and Harvard University. After retiring from Harvard, he ended his career as an adjunct professor at the University of California, Santa Cruz. His most celebrated work The Theory of Public Finance, published in 1959, remains a landmark in the field. It offers an impressive synthesis of the most important theoretical contributions to the discipline – many of them his own – over the previous 20 years; and it served as the leading graduate text for a further 20 years. Musgrave was a lively and effective graduate teacher and an excellent PhD supervisor, as can be seen from his many gifted and independent-minded students. Among the latter, perhaps Alan Auerbach and Charles McLure, Jr. are particularly outstanding. His most enduring scholarly achievements were in the xi

xii

TAX REFORM IN THE 21ST CENTURY

area of taxation theory and policy, though he also contributed seminal insights in the area of public expenditure analysis, notably the concepts of social wants and merit wants. His influence on younger public finance scholars has been profound and enduring. I was myself already a committed Musgrave disciple when I began teaching public finance in Canberra in 1957. And I vividly remember the excitement when his great book landed on my desk from McGraw-Hill in late 1959. I have never in my long career learned half so much from between a single pair of hard covers. Others of my generation had no doubt a similar experience. We became friends when he invited me to take up a visiting lecturer position at Princeton for the Spring semester in 1965. When I arrived in mid-January he had not long been married to Peggy with whom he shared a long and happy personal and professional relationship. Richard had been an advisor on her PhD committee at Johns Hopkins, and her revised dissertation on The Taxation of Foreign Investment Income (1963) soon became an influential monograph, indeed a standard reference, on international tax. The undergraduate textbook Public Finance in Theory and Practice, which he published jointly with Peggy in 1973, with four subsequent editions and translations into many other languages, transformed the teaching of public finance with its unique blend of theory, fiscal institutions and the findings of recent empirical research. Reflecting no doubt his early experience of the rise of Nazism in Germany, Musgrave was a passionate believer in the role of democratic institutions. Tax policy objectives had to be achieved within the framework of a democratic society. He believed strongly that academic economists should contribute to policy making by periodically working or consulting with government and also by actively participating in serious public debate on taxation policy. In this respect he set a truly inspiring example. He worked for some years at the Federal Reserve during World War II, and he advised Congressional Committees and government bodies such as the US Treasury, the Council of Economic Advisors, the Federal Reserve Board and the World Bank. He led important tax missions to developing countries, notably Colombia and Bolivia. The published reports of these missions are models of their type, and have greatly influenced the course of tax reform in these and other jurisdictions. His contribution to scholarly debate on taxation policy in Australia was of immense significance, beginning with his active participation in my inaugural conference on Australian taxation policy organised for the Australian Tax Research Foundation in August 1982. The conference volume Taxation Issues of the 1980s (1983) provided a wealth of ideas which fed into the tax reform debate and, with the succeeding conference on Changing the Tax Mix (February 1985), strongly influenced subsequent policy initiatives of the National Taxation Summit and post-Summit period. Musgrave returned in 1986 for our Retrospect and Prospect

INTRODUCTION AND DEDICATION

xiii

conference on the Keating Tax Reform Package of September 1985 which by that time had been substantially implemented. The high quality and scholarly impact of the conferences and associated volumes published by the Australian Tax Research Foundation over the period 1983–97 owed much to Musgrave’s early encouragement and support. His active participation, along with that of a succession of eminent overseas colleagues and former students – including Carl Shoup, Richard Bird, Sijbren Cnossen, Charles McLure, Peter Mieszkowski and Jonathan Kesselman – served as an inspiration to a younger generation of Australian public finance economists who presented papers at these conferences. Musgrave was a strongly committed supporter of collaborative research by economists and lawyers in the tax policy area, nicely exemplified by the tax policy seminar he conducted jointly with Stanley Surrey in the Law School at Harvard. Indeed we are indebted to Harvard Law School for their contribution towards the funding of the Memorial Conference. Leading tax lawyers from Australia and overseas, including Chris Evans, Malcolm Gammie, Neil Brooks and Michael McIntyre, contributed papers at the conference, and the involvement of Kluwer Law International as publisher of the volume clearly reflects a balance of legal as well as economic analysis represented at the conference. Musgrave actively encouraged our early efforts in 1983 to establish a new interdisciplinary journal Australian Tax Forum under the aegis of the Monash Law School with sponsorship from the Monash Centre of Policy Studies. Under the energetic and enterprising ten-year editorship of Rick Krever, the Journal successfully fostered scholarly debate through regular seminars on major issues in Australian tax policy featuring local and overseas tax lawyers and economists. Papers from these seminars were subsequently published as issues of the journal. This tradition continues with the publication of papers from the April, 2007 Personal Income Tax Reform Symposium (organised for Atax, University of New South Wales by Chris Evans) as a double issue of Australian Tax Forum dedicated to Musgrave. Papers from the Colloquium on Taxation and Risk-Taking, held in conjunction with the Memorial Conference, will also be published as a special issue of the journal in recognition of Musgrave’s pioneering contribution to this subject. We honour therefore the memory of this great tax scholar and policy advisor who went far out of his way at an advanced age to pursue the cause of tax reform with his friends in the antipodes. For those who knew and admired him, he will be greatly missed, though his influence, even in Australia will long endure. John Head May, 2009

Reflection

Remembering Richard Musgrave, 1910–2007 Peggy B. Musgrave* Thank you one and all for contributing to this conference in Richard’s honor, and special thanks to John Head and Rick Krever for arranging it. The topics chosen certainly reflect some of Richard’s own deep interests, and I wish he were here to join in the discussion. I find it very heart warming that he should be remembered with such respect and affection in Australian public finance circles. Richard and I always cherished our association with Australia and had vivid memories of our last visit to this country some 25 years ago. As a bird enthusiast, I still remember those weird, wonderful birds and the little penguins scuttling up the beach to their burrows each evening. Earlier on, it was our pleasure to establish a lasting friendship with John Head when he joined the Princeton economics faculty as a visiting lecturer, and I greatly appreciate his strong support over the years of Richard’s work and values. Forgive me those who have heard this story before. But many years ago in response to a comment by a discussant of my paper that I was too much influenced by Richard’s ideas, Richard intervened with the observation that he didn’t know whether we thought alike because we were married or that we were married because we thought alike. Whichever way it went, I am privileged to have shared 43 years of his life and to have been a junior partner in some of his endeavours. It is not possible in this brief introduction to comment comprehensively on his abundant

*

Professor Emeritus, University of California at Santa Cruz; Senior Fellow, Taxation Law and Policy Research Institute, Monash University.

4

TAX REFORM IN THE 21ST CENTURY

work which includes some 180 published articles1 and numerous books. Moreover, since I am not exactly a dispassionate observer, I leave it to others to write tributes to Richard and his work, which I appreciate and of course consider well-deserved.2 But perhaps I can offer a few special insights into his philosophy and background. I will also briefly refer to some of his principal areas of interest and the extent to which his work has stood the test of time. Richard’s interest in the public sector continued throughout his professional life, and his vision of its importance in a democratic society grew out of his early education in Germany when in the early 1930s he attended the Universities of Munich and Heidelberg. At first he was attracted to the field of literature and even had an interest in becoming a stage director, but when as a student he was exposed to social philosophers and economists such as Max and Alfred Weber, von Zwiedineck-Südenhorst and Jacob Marschak, his deepest interest quickly turned to the philosophy and economics of the public sector. Indeed his philosophical background pervaded his vision of the ‘good society’, a term to which he frequently referred and in which he saw the public sector performing a crucial role. The content of his bookshelves attests to that; works from Kant to Hegel and Weber and other great German philosophers stand side-by-side with those of Adam Smith, Hume, Locke and others. They are well-worn books with many scraps of paper marking passages that most interested him. Worth special mention is the enduring influence that the work of the social philosopher Max Weber had on Richard. First, he derived a commitment to the basic standards which should apply to legitimate scholarship and the place of value judgments therein – that value judgments may determine the choice of problems to analyze, but the analysis itself should not be influenced by the analyst’s own subjective values. But, secondly, it gave him an overarching vision of the social relationships and aspirations which, in a democratic society, may contribute to the quality of life of all its members. I think that it is true to say that many of his students derived from Richard, as did I, the inspiration that what they were studying was connected to this larger vision of a good society. He was nevertheless insistent that the collective decisions which such a society makes should reflect the individual preferences of its members as democratically determined through the voting process. His experiences with the early emergence of Nazi ideology in Germany strongly reinforced his adherence to democracy. 1

2

The reader is referred to his selected papers written between the 1950s and mid-80s which are reprinted in the three-volume Public Finance in a Democratic Society (Brighton: Wheatsheaf Books, 1986). H.J. Aaron, ‘Tribute to Richard Abel Musgrave’ in S. Cnossen and H-W Sinn (eds), Public Finance and Public Policy in the New Century (Cambridge: MIT Press, 2003), xxiii; J. Head, ‘Richard Musgrave Remembered’, (2007) 22(2) Australian Tax Forum, 9; A.B. Atkinson, S. Cnossen, H.F. Ladd, P. Mieszkowski, P. Pestieau and P.A. Samuelson, ‘Commemorating Richard Musgrave’, (2008) 64 Finanzarchiv, 145.

PEGGY MUSGRAVE: REMEMBERING RICHARD MUSGRAVE, 1910–2007

5

Richard also drew from his social philosophy a basic optimism and faith in the ultimate ascendancy of the best in humankind. This was an optimism which was put to many severe tests during his long life. But from day to day he was never discouraged when events seemed to take a wrong turn, for he considered most to be a small stumble in the march of history towards a better world. This optimism was reflected in his emphasis on the normative in economics and his choice of the public sector as his field of study. He often defended the normative to his students by saying that, unless one has a vision of the best, how can one take steps in the right direction? That ‘best’ includes not only the most efficient use of available resources but also a fair distribution of the fruits of those resources, a fairness not necessarily best determined by market forces. The private sector based on a market system with its reliance on individual self-interest, while important for efficiency and economic growth, he found to be less attractive as a focus of his work than the public sector which in a democratic society is founded on cooperation and concern for the welfare of others. His debate with James Buchanan, published in book form in 1999,3 served to highlight his view that, in spite of various deficiencies from time to time in the conduct of the public sector, the ‘public choice’ assumptions of bureaucratic selfserving and entrenched shortcomings of the democratic state are totally unjustified. In fact he disliked what he considered the pejorative term ‘bureaucrat’ applied to those who work in the public sector, preferring that of ‘public servant’, which he believed was a noble calling not to be disdained. He was proud of his students who went on to serve in the public sector and once, much to his pleasure, when attending a meeting at the Social Security Administration, it turned out that all those present had been former students of his! The following paragraphs attempt to encapsulate the principal areas of his contributions to public sector economics, summarizing the extent to which his ideas have endured in today’s thinking. In general it is probably true to say that he laid the foundation for the modern theory and practice of public finance in his 1958 magnum opus.4 The decades passed and, as he expected, new generations of public finance economists further built upon, extended, and sometimes modified or took issue with his work. Even though in recent times the spirit behind the scholarship applied to public sector economics has been more sceptical of government’s role, perhaps reflecting and indeed influencing the politics of the period, Richard felt this was a passing phase which would ultimately give way to more constructive attitudes towards the public sector.

3 4

J.M. Buchanan and R.A. Musgrave, Public Finance and Public Choice: Two Contrasting Visions of the State (Cambridge, MA: MIT Press, 1999). R.A. Musgrave, The Theory of Public Finance (New York: McGraw-Hill, 1958).

6

TAX REFORM IN THE 21ST CENTURY

The Theory of Social Goods Richard’s earliest interest was focused on the central problem of the provision of social goods and services, i.e. those goods and services which have the characteristic of being ‘non-rival in consumption’ (a term coined by him) and therefore do not provide the incentive to revelation of preferences as do private goods which can be bought and sold in the market. In consequence, provision of social goods has to be made through the public sector with such provision nevertheless reflecting individual preferences, as revealed through a democratic voting system. His interest in the topic of social goods had been stimulated by his study of the work of Knut Wicksell and Erik Lindahl on the voluntary exchange model of the public sector and resulted in his first major paper ‘The Voluntary Exchange Theory of Public Economy’ published in the Quarterly Journal of Economics in 1939. This paper introduced the German-language work of Wicksell and Lindahl to Englishspeaking scholars. Among them was Paul Samuelson and it resulted in his pathbreaking 1954 article ‘The Pure Theory of Public Expenditure’. I well remember Richard and Paul getting together on a Sunday morning to discuss the solution to the provision of social goods. In his characteristically gracious manner, Samuelson has recently expressed his debt to Richard in the following words: It is no great stretch to say that it was Musgrave’s Harvard Ph.D. thesis that brought to America news and details about the Wicksell-Lindahl theory of public goods. … Imagining counterfactual history, surely it would have been better if there had been a 1954 item by Musgrave and Samuelson that led to a post-1970 Nobel Prize in economics. Indeed, many discerning academics believe that Stockholm erred in not honoring the author of Musgrave’s Theory of Public Finance – the acknowledged successor to A.C. Pigou’s The Economics of Welfare and A Study in Public Finance.5 Richard greatly admired Paul Samuelson’s clever mathematical exposition of the optimal solution to the problem of social goods provision. However, he was quick to point out that it presupposed the existence of an all-knowing referee to implement the result, while in practice a democratic voting system is needed to make the decisions. In subsequent decades much has been written on the topic not only of social goods but also of private goods which produce social externalities. Those sceptical of the essential role of government in these areas maintain that bargaining among small numbers of individuals might solve the problem without public sector intervention. Travelling through a Spanish town many years ago with the late Milton Friedman (on another debating event), we remarked on the external costs imposed on the inhabitants of that otherwise tranquil town by the incessant

5

P.A. Samuelson, ‘Affectionate Reminiscences of Richard Musgrave’ (2008) 64 Finanzarchiv, 167.

PEGGY MUSGRAVE: REMEMBERING RICHARD MUSGRAVE, 1910–2007

7

roar of motor bikes. Milton responded by saying that, since the residents could freely and individually negotiate with the motor cyclists, it would seem that the noise did not impose a cost on them. Richard, while very doubtful of that reasoning, acknowledged that some small-number situations exist and allow for bargaining, but most involve large numbers and call for government intervention. Indeed in recent times externalities on a worldwide scale (which may be characterized as international social goods or bads) depend on multinational cooperation for their solution. Richard even went so far as to suggest (perhaps half seriously) that externalities had the useful characteristic of calling for cooperation whether among individuals or jurisdictions, a situation to be welcomed in itself.

A Comprehensive View of the Public Sector Until Richard’s treatise The Theory of Public Finance was published, studies in the realm of public finance were largely confined to taxation. Richard broadened the field to include the role of expenditures, the basis of his work being a systematic and comprehensive view of the public sector to match that already established for the private sector from Adam Smith on, and how it might perform efficiently and equitably in partnership with the private sector. He defined and expressed the fiscal functions of government in his model of the 3-branch budget including allocation, distribution and stabilization.6 This he proposed not only as a normative theoretical structure for public sector budgets allowing separation of the three basic functions of government, but predominantly as a useful pedagogical and analytical device for sorting out the myriad items included in government expenditures and revenues in terms of the basic purpose which they served. In reality the three functions of government are inter-connected and overlapping, but the concept of the budget as fulfilling the three fiscal functions of government serves to rationalize and categorize government economic operations. Many students of introductory public finance have found this tripartite division a useful way of thinking about the economics of the public sector. Richard’s own extensive writings ranged over all three fiscal functions – social goods, distributive policies (especially including taxation) and fiscal stabilization policy. Since his three-branch budget was proposed some 50 years ago, national economies of the world have become ever more inter-connected, and each of the three fiscal functions of government should now be extended to allow for policies to deal with these interrelationships, whether arising from trade in goods and services, capital movements, labor migration or internationally-shared externalities.

6

R.A. Musgrave, ‘A Multiple Theory of Budget Determination’ (1957) 17(3) Finanzarchiv, 333; also included in Public Finance in a Democratic Society, n. 1 above.

8

TAX REFORM IN THE 21ST CENTURY

The Role of Taxation Richard early on stressed the role of taxation not merely in raising revenue to pay for public goods and services, whether or not applied on the benefit principle, but even more importantly in providing the means for adjusting the distribution of income. This was based on his firm belief that a free private market system, while contributing to the achievement of efficiency and growth, does not necessarily result in an acceptable distribution of income that is so important for the quality of life and social stability. Thus the equity or distributional effects of public policy were paramount in his thinking, both positive and normative, though the latter may call for broadening the scope of traditional economics into neighboring fields of philosophy and political theory. His vision of the good society led him to place emphasis on equity in taxation, a concept which has tended to become obscured in later general equilibrium models of optimal taxation in which efficiency plays a dominant role. While he was bemused by and admired the intellectual content of such latter-day optimal tax models and the insights they may provide, his major interest was in a system of taxation which would be understandable and acceptable to the electorate. The policy relevance of many optimal tax models he considered marred by the questionable assumptions made and the fact that the equity implications were less than transparent. Moreover, the assumptions such models employ frequently differ and are sometimes in conflict, thus rendering questionable their policy applicability. There are two dimensions to the concept of tax equity – the choice of an appropriate tax base and the rate structure which is applied. Richard was a firm adherent of a broad-based tax founded on the concept of ability-to-pay, most appropriately defined as income and measured, at least conceptually, in ‘accretion’ terms. In recent years there has been much discussion in academia regarding the tax base, and many public finance economists have turned away from income to consumption, especially in connection with the various optimal tax models which seek to incorporate efficiency considerations as well as equity into the design of broad-based taxation. While he was interested in these models and acknowledged the efficiency advantages of the consumption base, he stuck to his belief that income should be the preferred base. In pragmatic terms he doubted the possibility of achieving a degree of progressivity with a consumption base that would be comparable with that of an income base. As to the argument that taxation of income discriminates against saving, his answer was that saving and wealth holding yields utility just as does consumption and therefore should be given equal tax treatment. This, of course, allows for the place of consumption taxes in implementing the benefit principle where appropriate. While he took the view that the corporation income tax should be applied to act as a form of withholding tax against shareholders’ individual income tax liability and ideally credited against it, Richard also studied aspects of the tax as it is applied in

PEGGY MUSGRAVE: REMEMBERING RICHARD MUSGRAVE, 1910–2007

9

practice. Together with his colleague Marian Krzyzaniak, he undertook a pioneering empirical examination of its incidence.7 My impression was that this study, showing a greater than 100 per cent shifting of the tax, left him with a somewhat sceptical attitude to the findings of econometric studies which necessarily fail to capture the complexities of the modern economy. Indeed, since then, further incidence studies have yielded varying results, and Richard’s own work and that of most others in the policy area has followed Harberger8 and assumed that the full burden of the corporate tax rests on capital. An influential article co-authored with Evsey Domar in 19449 addressed the question of whether proportional taxation of business income deters risk-taking.10 To this day it is often asserted that the disincentive effect on willingness to take on business risk constitutes a reason to exempt profits from tax. However, Musgrave and Domar in their article showed that there is no deterrence to risk-taking provided that business losses are fully offset. On the whole I think it is true to say that this view has generally survived the last 60 years – with one notable exception represented by Louis Kaplow in an article11 which came to the conclusion that the problem is irrelevant, based on the somewhat bizarre assumption that governments themselves avoid taxing risky ventures. Richard was bemused by these kinds of models and, while welcoming the work of the new generation of technically proficient analysts, he had serious doubts about the relevance of some of this work for responsible tax policy making.

The Concept of Merit Wants His philosophical bent led Richard to the concept of merit wants, a concept which has aroused a good deal of controversy within the profession since it first appeared in his Theory treatise.12 Perhaps because it has not fitted comfortably into traditional micro-theory founded on the paramountcy of individual self-interest (since Adam Smith the sine qua non of economic analysis), there seems an inclination to expunge it from the economics lexicon; or to force it into the traditional framework where it becomes redundant. Recent public finance textbooks rarely mention it. But there 7 8 9 10

11 12

M. Krzyzaniak and R. A. Musgrave, The Shifting of the Corporation Income Tax (Baltimore, MD: The Johns Hopkins Press, 1963). A.C. Harberger, ‘The Incidence of the Corporation Income Tax’ (1962) 70 Journal of Political Economy, 215. E. Domar and R.A. Musgrave, ‘Proportional Income Taxation and Risk-Taking’ (1944) 58 Quarterly Journal of Economics, 388. It was written while Richard and Evsey were colleagues at the Federal Reserve Board in Washington during World War II. Amusingly, Richard told me that they developed the idea while they were canoeing on the Potomac River. L. Kaplow, ‘Taxation and Risk Taking: A General Equilibrium Perspective’ (1994) 47 National Tax Journal, 789. R.A. Musgrave, ‘On Merit Goods’, reprinted in its revised version in Public Finance in a Democratic Society, n. 1 above, at p. 34.

10

TAX REFORM IN THE 21ST CENTURY

remain some of us who still feel that merit wants are relevant even in a democratic society and that perhaps the social welfare function should be revised to allow for them.13 Next, a few words on the nature of this particular type of ‘want’ or preference which carries Richard’s label ‘merit wants’. He saw such wants as arising from community traditions, values or customary practices, without reference to individual preferences in the community, though accepted by its members. He most emphatically did not view such wants as emanating from the state or administrative unit defined as an ‘organic’ entity with preferences of its own, such as would be found in an authoritarian state. Rather, they reflect group preferences embedded in historical, religious and ethical traditions and are given expression by common consent through democratic government, acting as the guardian of such values. In a democracy such community wants may be laid down explicitly or implicitly in its founding constitution and amendments thereto, as well as in the body of laws that place restrictions on individual behavior and preferences in order to preserve such community values. One does not have to look outside democratic societies to be aware of the existence of merit wants and the goods and services to which they give rise or to attribute them to the fact that all democratic processes, even town meetings, engage in untidy decision-making.. They often occur in the context of distributional issues, where redistribution goes beyond expressing interdependent utility functions or where redistribution takes the form of categorical or in-kind grants which disregard the preferences of recipients. Merit wants also include conservation and preservation practices which are entrusted to government in the interest of future generations. Another policy area which exhibits departure from the standard of individual consumer sovereignty is the choice of discount rate, which has a powerful effect on the level, choice and longevity of public projects as well as their financing. It might be said that any discount rate chosen by government which departs from the private rate of time preference enters the realm of merit or demerit wants. I was reminded of this when reading the Stern Review recommendations on the choice of discount rate for estimating the costs of global warming. Surely social rates of discount contain an element of merit wants! The principal issues regarding merit wants that have engaged economists’ attention have been (1) their normative status (C.E. McLure, Jr)14 and (2) their policy relevance. With regard to their normative status, some have challenged their unique status by folding them into individual preferences via interdependent utility functions, irrational behavior, or by resort to the democratic process itself 13 14

K.W. Roskamp, ‘Public Goods, Merit Goods, Pareto Optimum and Social Optimum’ (1975) 30(1) Public Finance, 61. C.E. McLure, Jr., ‘Merit Wants: A Normative Empty Box’ (1968) 27(2) Finanzarchiv, 474.

PEGGY MUSGRAVE: REMEMBERING RICHARD MUSGRAVE, 1910–2007

11

(G. Brennan).15 In his excellent survey John Head16 proposed a compromise solution by allowing for a hierarchy of consumer preferences with each successive level arising from more information, experience, ethical awareness and so on. There is thus no need to postulate community preferences which override those of the individual. In the face of a barrage of criticism, I think Richard did waver a little in narrowing somewhat the scope of merit wants. However he stuck to his proposition regarding the central role of community values and the institutions that safeguard and give expression to them. Perhaps one of the gains from introducing the discussion of ‘merit wants’ is that it calls for a broadening of the public economics discourse to embrace political science, social philosophy and ethics, subjects that can more easily be bypassed in an economic analysis of the private sector based on consumer sovereignty. Initially economists simply transferred their presuppositions from the private sector, in which the market system allows fulfillment of individual preferences, to the processes of the public sector. Then the private sector paradigm had to be modified by the introduction of social goods and the non-excludability concept. However admission of merit wants requires a broadening of the boundaries of economics to allow for community wants. Some broadening has been done by public choice theorists,17 but it is unfortunate that the field of public choice has been constrained by its negative a priori assumptions regarding governments’ behavior. Although Richard’s concept of merit goods is now an orphan, I think it cannot be swept under the rug just because it does not fit in with the traditional assumptions of economic theory. Human beings are social animals and in many respects have concerns for their communities which transcend their personal preferences. Perhaps it will take an inter-disciplinary study to establish the legitimacy of the merit wants concept.18 Or perhaps the contemporary field of behavioral economics can come to the rescue and provide a link between merit wants and a particular form of individual preferences.

Fiscal Federalism19 Richard Musgrave’s Theory of Public Finance also introduced consideration of how the fiscal functions of government in a federal system might be divided between the central and lower-level jurisdictions. This was developed further in 15

16 17 18 19

G. Brennan, ‘Irrational Action, Individual Sovereignty and Political Process’ in G. Brennan and C. Walsh (eds), Rationality, Individualism and Public Policy (Canberra: ANU Press, 1990), 97. J.G. Head, ‘On Merit Wants’ (1988) 46 Finanzarchiv, 1. J.M. Buchanan, ‘The Theory of Public Finance’ (1960) 26 Southern Economic Journal, 234. W. Ver Eecke, ‘The Concept of a “Merit Good”: The Ethical Dimension in Economic Theory and the History of Economic Thought’, (1998) 27(1) Journal of Socio-Economics, 133. In this section I have drawn on the paper by Peter Mieszkowski referred to in n. 2.

12

TAX REFORM IN THE 21ST CENTURY

several key articles which spelled out his basic principles of fiscal federalism.20 The issues involved are numerous, but I will only discuss the most basic one, namely the assignment of fiscal functions in a federally organized state. Fundamentally he proposed that the allocation function, i.e. provision of public goods and services, should be shared between the central and lower levels of government, depending on the spatial range of their benefit incidence and on the particular pattern of preferences within each of the various jurisdictions. The latter may differ depending on the history, geography, ethnicity, income and other characteristics of members of the federation. This he termed the principle of reciprocity. He held that the distributional role should belong to the central government, primarily because a single standard of distributional justice should apply to all citizens of the state. This is consistent with the view of social justice as a merit want. This assignment was also necessary on efficiency grounds. If each lower jurisdiction applied its own distributional policy, this would result in population migrations from the less to more generous states with an ensuing misallocation of resources. Finally, there is little disagreement that the responsibility for economic stabilization must lie very largely with the central government. This basic assignment of functions has in large measure been accepted and built upon by other writers in the field. The notion that allocative spending responsibilities should be assigned according to the spatial extent of their benefit spillovers, however, requires either intervention by the federal government via matching grants or, alternatively, a pattern of regional authorities with jurisdiction over certain expenditures and boundaries which may overlap and be superimposed on the historically-determined state boundaries. Some writers have challenged Richard’s proposal of a strong central government, in particular suggesting that the distribution functions does not necessarily belong at the central level for two reasons. First, utility inter-dependence is claimed to be confined within the boundaries of individual jurisdictions. Secondly, it smacks of an organic state imposing its policies on lower-level jurisdictions. Some empirical work has given support to this idea by showing that tax-induced migrations are not of significant magnitude. However, there is also evidence that states within the United States do try to match each other in the level of distributional policies which they apply, suggesting that the central government might better be responsible for applying uniform distributional policies nationwide. The case for and design of intergovernmental grants has also been the subject of much literature on fiscal federalism, in which Richard’s 1971 paper played an early part. He suggested that the case for such grants can be made on three grounds: first, 20

R.A. Musgrave ‘Approaches to a Fiscal Theory of Political Federalism’ (1961), in Public Finance in a Democratic Society, n. 1 above, p. 9; ‘Economics of Fiscal Federalism’, (1971) 10 Nebraska Journal of Economics and Business, reprinted in Public Finance in a Democratic Society, n. 1 above, at p. 33.

PEGGY MUSGRAVE: REMEMBERING RICHARD MUSGRAVE, 1910–2007

13

as adjustments to achieve greater efficiency and equity by compensating for those state expenditure-benefit spillovers which have a nationwide incidence; second, to support those state activities which are consistent with federal merit wants; and thirdly, to support fiscal equalization among jurisdictions to help equalize the fiscal resources of rich and poor states through ‘fiscal capacity equalization’ (FCE) grants which would allow all states to provide basic levels of certain services. This principle of fiscal equalization has been subjected to a good deal of criticism by those who claim that it is anti-individualistic, since it treats the individual grantees as groups. Some critics have expressed a preference for ‘horizontal equity equalization’ (HEE) grants. Richard, while allowing that some case can be made for HEE grants, pointed out that unlike FCE grants they cannot address vertical equity among states. He maintained that rejection of FCE did not properly recognize the place of an individual in a federal system, namely as a member of a national group with merit wants embracing certain basic services. This fiscal equalization debate has not in practice stood in the way of a substantial increase in those policies of the US federal government which have major redistributive effects on resources among states. In all, Richard’s views on a well-functioning fiscal federalism have largely stood the test of time and borne fruit in terms of policy applications.

Public Finance in Action Richard was equally interested in the day-to-day performance of the public sector, and was often engaged in policy matters and the positive aspects of public finance. As a tax advisor at the state and national levels, and with a long history of leading missions on tax reform from post-war Germany to developing countries such as Burma, Bolivia, Colombia, and South Korea resulting in several volumes of reports, he relished meeting reality head-on in the search for feasible progress in tax matters. Nevertheless he had few illusions as to the political obstacles everywhere confronting progress towards his ‘good society’, i.e. a society democratic in governance, efficient in the use of its resources and without the divisions caused by severe distributional inequalities.

Family Background As with most of us, Richard’s early background and education exercised an important influence on his intellectual development. He came from a family which epitomized the high degree of integration of 19th century Jewish intellectuals in German society, lending an added tragic element to the Nazi horrors. Both his father and mother were half Jewish, his father being the son of a Jewish professor and an aristocratic lady, and his mother the daughter of a Jewish lady from a business family and a Protestant father who was the Mayor of Stuttgart. An outcome of this racial intermingling was that Richard deplored racial and religious divisions and

14

TAX REFORM IN THE 21ST CENTURY

animosities. Perhaps this led him to take the faculty leadership at Harvard (I believe, courageously) in setting up a Department of African-American Studies. The earliest family influence came from his father who had a doctoral degree in chemistry but spent most of his life as an amateur playwright, pamphleteer and social activist. His target was German militarism prior to World War I, and for this reason he was actually imprisoned for a while because of his campaign against dueling among German officers. He was often declared persona non grata in Germany and spent considerable time in England as a result. There he taught in a school for delinquent boys and acquired the hyphenated name Abel-Musgrave, the latter taken from one of the books of his friend Arthur Conan Doyle. My impression is that Richard acquired from his father an abiding commitment to social justice. At the same time, perhaps in contrast to his father’s somewhat scattered endeavors, he learned the importance of focus on one’s area of study. His mother, an accomplished mountain climber (unusual for a woman in those days), died when he was 14 years old.21 His life spanned both World Wars. He remembered as a small child the bedraggled German soldiers returning home defeated and humiliated at the end of World War I. As a young man he had felt the deep disappointment at the failure of the Weimar Republic, and in his years as a student at the universities of Heidelberg and Munich he witnessed the emergence of the Nazis and their destructive practices. He tells how he returned from the ski slopes to learn of and engage in a protest at the Reichstag burning, and how he witnessed the book burning at Heidelberg. Most of his family fled to England (where one of his uncles-in-law became a doctor to J.M. Keynes). His father went to the US where he arranged for Richard to join him by applying for a student exchange at the University of Rochester. Of course, all this had a deep effect on Richard as it did on so many other refugees, learning how society can be steered in the wrong direction as a result of an evil ideology taking advantage of economic deprivation and the humiliation of defeat. His German experience also made him wary of grandiose, romantic notions of national uniqueness, which had captured many German intellectuals and opened them up to the blandishments of Nazism.

Academic Career Richard frequently referred to his great good fortune in beginning his graduate work at Harvard at a time of momentous developments in economics associated with both the Keynesian revolution and the pioneering work on imperfect competition associated with E.H. Chamberlin and Joan Robinson. He spoke with such admiration and warmth of his many talented fellow graduate students at that time, all of whom made remarkable contributions to the theory and application of economic science. 21

One of his early playmates before he was sent off to boarding school subsequently became Mrs. Peter Drucker.

PEGGY MUSGRAVE: REMEMBERING RICHARD MUSGRAVE, 1910–2007

15

There were also great teachers such as Taussig, Haberler, Schumpeter and Alvin Hansen, a teacher for whom Richard had the greatest respect and affection. The great stimulation presented by this intellectual setting provided an inspiring foundation for his subsequent career. The time spent during the war years at the Federal Reserve, first as a research economist and then as special assistant to the chairman, Marriner Eccles, supplemented the macro-stabilization background so abundantly present during his graduate years at Harvard. However, monetary theory would not prove to be the focus of his chosen area of study.22 After a short period teaching at Swarthmore College, his next appointment was as professor in the economics department at the University of Michigan where again he was blessed with a stimulating and congenial group of colleagues.23 It was at Michigan that his thinking and writing on the public sector following his thesis, came to fruition in his magnum opus, The Theory of Public Finance published in 1958, which he said benefited greatly from his frequent discussions with colleagues and graduate students in seminars and other venues. While reluctant to leave his beloved Michigan, he could not resist the chance, when offered an appointment at The Johns Hopkins University, to move near to Washington where government policies of great interest to him were being decided. During his several years there, he was a frequent participant in various advisory activities. Following Johns Hopkins, he chose to take up a professorship at Princeton where he taught primarily in the Woodrow Wilson School of Public Policy.24 This was convenient for us both, since I had just accepted an appointment at the University of Pennsylvania. Then when Harvard University offered him a joint appointment in the Economics Department and the Law School, he came full circle spending most of his remaining active years where he had begun his academic career in the United States and which carried such intellectually exciting memories for him.25 After retiring from Harvard, he followed me to California where I accepted an appointment as professor of economics at the University of California, Santa Cruz and where Richard spent several years as an adjunct professor of economics. Richard had interests which extended well beyond his scholarly endeavors. From his years in Germany he acquired a strong interest in literature and particularly the theatre. He was also an amateur painter with a strong style reminiscent of the German Expressionists. His large, colorful painting of the Shwedagon Pagoda in 22

23 24 25

He told me, half jokingly, that he knew it was time to move on to another field when Eccles assigned him the task of reporting to him each day on the latest developments in the bond market! He always said that he enjoyed his time at Michigan most of the several venues at which he had taught. His collected papers are to be deposited in the archives of the Mudd Library of Public Policy at Princeton. Richard’s ashes are buried in the Mount Auburn Cemetery, a few steps from Harvard.

16

TAX REFORM IN THE 21ST CENTURY

Rangoon, painted in 1956 when on an economic mission to Burma, reminds me each day of his spirited approach. He also enjoyed various outdoor activities such as skiing, canoeing, and in later years sailing and fishing on Monterey Bay. An accomplished horseback rider, he passed that interest on to my daughter. He loved having visitors with whom he would discuss economics and current political events, and he especially enjoyed colleagues from abroad and home, as well as past and present students, visiting us in our summer home in Vermont. Richard was a very large presence in my life and I greatly admired his tolerance and indeed his admiration for new ideas as well as his defence of his own firmly held beliefs. I know he would have enjoyed taking part in the give and take of this conference. At this commemoration, I cannot do better than to conclude with the following passage out of the preface to his 1958 Theory of Public Finance (p. v): Unlike some economic purists of today. I admit to more than merely a scientific motivation; intelligent and civilized conduct of government and the delineation of its responsibilities are at the heart of democracy. Indeed, the conduct of government is the testing ground of social ethics and civilized living. Intelligent conduct of government requires an understanding of the economic relations involved; and economists, by aiding in this understanding, may hope to contribute to a better society. This is why the field of public finance has seemed of particular interest to me; and this is why my interest in the field has been motivated by a search for the good society, no less than by scientific curiosity.

Pursuing Tax Reform

Chapter 1

The Political Economy of Tax Reform: A Neo-Musgravian Perspective WITH ILLUSTRATIONS FROM CANADIAN, US, AUSTRALIAN AND NEW ZEALAND EXPERIENCE

John G. Head * 1

Introduction and Overview

For Richard Musgrave, the area of tax policy and tax reform was always a matter of central concern throughout his long and distinguished career in public finance. Although he made significant contributions to public expenditure analysis and worked in such specialist areas as fiscal federalism and development, he had a lifelong commitment to tax reform and especially to tax equity. He would no doubt have heartily endorsed the well known political dictum that a fair tax system is the essential hallmark of a well functioning democracy. It seems appropriate therefore that this memorial volume should begin with a review of the tax reform experience of a sample of four Anglophone countries over recent decades. Before we turn to these experiences, however, it is necessary to observe that our analysis will not be limited to the economic or public finance aspects in the traditional narrow sense. Rather we will follow Musgrave’s example and have an eye also for the political economy aspects, since this is the dimension where so many of the most difficult and intractable problems have arisen. In what *

Taxation Law and Policy Research Institute, Monash University. I am indebted to Patricia Apps, Kimberley Brooks, Michael Brooks, Neil Brooks, Jane Gravelle, Rick Krever, Michael McIntyre, Peggy Musgrave and David White for helpful comments and suggestions. The views expressed in the paper remain my own.

20

TAX REFORM IN THE 21ST CENTURY

follows we make no attempt to offer a complete or systematic treatment of these political economy aspects or to develop a rigorous analytical framework for this purpose. Following Musgrave’s own approach to tax policy and tax reform issues, our orientation will be dominated by practical rather than theoretical concerns, whether in relation to the political economy aspects or to more narrowly economic or public finance issues. The path-breaking, if also rather disheartening, experience of Canada from 1967–1971, beginning with the much admired blueprint for tax reform of the Carter Royal Commission (1966), seems an appropriate place to start. This was the first in a series of such inspirational blueprints to appear in the Anglophone countries over the following 20 years. The tax reform outcome of this impressive initiative was, however, very disappointing. The Carter proposals were the subject of intense and hostile public and political debate and were followed by a government White Paper (1969) containing a much less radical package of reforms. These proposals in turn were subjected to intense scrutiny and strong opposition before the major parliamentary committees and the provinces in a process of what was described by the government as ‘participatory democracy’. As a result, little remained at the end of a long ten-year tax reform odyssey to be implemented in the final legislation in 1971. The Haig-Simons heartland of the US provides our second example, in which the ground for reform was long and carefully prepared over decades by the vigorous advocacy of a generation of committed supporters of the comprehensive income principle as the basis of a fair and efficient tax system. Inspirational tax reform blueprints were provided by the US Treasury (1977) and subsequently and more influentially in Treasury I (1984). The US experience of major income tax reform was much more successful, and the considerable achievements of TRA86 in terms of base-broadening and rate-flattening – and in terms of basic tax policy objectives – were widely acknowledged. However, much of this impressive reform achievement has since steadily unravelled. And doubts remain about the likely economic gains because of inadequate attention to tax unit and related aspects (section 4.2 below). Australia provides our third example. Whereas, in Canada and the US the focus was on income tax reform, in Australia and New Zealand the mix of direct income and indirect consumption taxes was a central issue. The Australian tax reform initiative of 1985 was quite ambitious, as reflected in the Draft White Paper proposals of June 1985 and especially the government’s preferred Approach C. This approach would have combined the introduction of a new broad-based retail sales tax with a major tax-mix switch from personal income tax to indirect consumption tax, and a modest first instalment of much needed income tax reform. The sales tax reform and tax-mix proposals failed, however, at the National Taxation Summit in July. This left a small and unsatisfactory package of piecemeal income tax reforms to be legislated in the Summit aftermath.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

21

The fourth and most successful of our sample countries was New Zealand. The proposals of the McCaw Report (1982) served here as a useful springboard, and were vigorously followed up, further refined and successfully introduced by the politically dominant Lange/Douglas government of the mid-1980s. As part of a very large package, a remarkably broad-based VAT (GST), along with a major tax-mix switch and a far-reaching and most impressive base-broadening and rate-lowering reform of the personal and company income tax system were all effectively implemented in a relatively short period. Here again, however, these efforts may have been seriously compromised by lack of attention to tax unit and related issues. Looking further ahead, we will argue that the priorities for tax reform will likely be very different from those prevailing in the past. In order to meet the challenges such as global warming that lie ahead, rapid and sustained public expenditure growth will be required. For decades Musgrave’s differential incidence concept has supplied the dominant framework for tax reform and tax policy analysis. Whether or not this was most appropriate historically, it seems clear that his alternative concept of balanced budget incidence provides a more appropriate neo-Musgravian perspective for the period ahead. This alternative framework will be applied for purposes of illustration to Australia where a review of the tax system is currently under way. With new and charismatic political leadership, the prospects for fundamental and much needed reform look reasonably bright. The outlook for reform may even have improved somewhat as a result of the recent turmoil in financial markets which could help to generate a Rawlsian ‘veil of ignorance’ effect. This in turn could cloud self-interest and assist in promoting principled deliberation on tax reform issues. There seems moreover no reason to suppose that the objective of revenue adequacy should figure any less prominently among the standard tax policy objectives of equity, efficiency and simplicity.

2

Musgrave on Tax Reform

Richard Musgrave was always and pre-eminently a practical tax reformer. His natural concern with theoretical rigour was tempered with a keen awareness of the practicalities of tax policy. This goes far towards explaining his misgivings about the increasing application of hi-tech mathematical modelling to policy problems. His particular concern was that important issues that are technically less tractable, or simply less interesting, could well be pushed aside and neglected (Musgrave, 1994, p. 180). Tax equity seemed to be an obvious case in point. The Haig-Simons income concept, to which he remained passionately committed throughout his remarkable career, was never viewed by Musgrave as a theoretical ideal but rather as a practical second best which could realistically be implemented through the democratic political process. At the political level, he strongly supported Joe Pechman in his efforts to organise the tax policy fraternity as a cohesive lobby

22

TAX REFORM IN THE 21ST CENTURY

group unified in support of a carefully articulated and persuasively argued program of comprehensive income tax reform. As with the economics, so with the political economy of tax reform, Musgrave was less attracted by purely theoretical models of the public choice variety. Indeed he was repelled by what he understandably perceived as their anti-public sector bias. For Musgrave the modern mixed economy exercised a life-long fascination, combining as it does the need for market analysis of the standard variety with the analysis of collective decision-making and democratic participation in the public sector. The Brennan-Buchanan (1977) models of a revenue-maximising Leviathan government and a bloated public sector he found particularly distasteful. And he could hardly wait for Brennan’s close and highly creative collaboration with James Buchanan during his VPI period to end. It would nevertheless have been interesting to have Brennan revisit the Leviathan model some 30 years on, because I believe that we now face yet another period of rapid and sustained public expenditure growth in which modest Leviathan propensities must surely be welcomed as offering salvation to a planet sorely pressed by large scale market failure problems of global warming, demographic ageing, infectious diseases and poverty. There have been similar periods of rapid expenditure growth in the fiscal history of the past century, notably during World War II and, following the war, in the emergence of the modern welfare state in Western Europe and some of the Anglophone countries. Like the problems of global warming and demographic change, these periods of market failure and public expenditure growth call for an analysis of taxation policy and tax reform which is significantly different in emphasis from the familiar equal-revenue comparisons of differential incidence analysis. Musgrave’s long neglected concept of balanced budget incidence provides the more relevant analytical framework. In this context, considerations of revenue adequacy assume a heightened significance alongside the standard tax policy objectives of equity, efficiency and simplicity. It is, however, necessary to consider whether these priorities should be revised in the context of the financial meltdown. In the short term, a tax cut may obviously be required to cushion the impact of the recent crash. In the long term, however, the revenue cost of the financial bailout and associated budgetary measures seems certain to be very large indeed, which further reinforces the case for assigning high priority to the revenue adequacy criterion.

3 The Simons-Carter Tradition For most of the earlier generation of Haig-Simons income tax supporters, their original conversion to the comprehensive income principle dated back no doubt to the unique experience of reading Henry Simons’ path-breaking masterpiece Personal Income Taxation published in 1938. Although Irving Fisher’s pioneering

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

23

work on the personal consumption tax alternative was published around the same time (Fisher, 1937a, 1937b, 1942), Simons’ book was so persuasively written – and the Fisher concept was so unfamiliar – that the comprehensive income approach quickly achieved academic pre-eminence. The general feeling was that income taxation should first be reformed and properly tried before turning to such a radically different alternative like Fisher’s expenditure tax. The comprehensive income concept dates back, of course, to earlier work by Georg von Schanz (1896) and Robert Murray Haig (1921), the direct influence of which is already to be seen in the important studies of Shoup et al. (1937a, 1937b) which ante-date Simons. There can be no question, however, that the decisive intellectual contribution must be traced to Simons’ rigorous analysis and swashbuckling prose. It remained, however, for the 6 volume Report of the Carter Royal Commission (1966) to show how the Simons income concept could be employed to provide the basis for a detailed tax reform program for a major democratic country like Canada (Head, 1970). The Carter Report was the first of several such inspirational blueprints arguing in detail the practicability of near-ideal income tax reform. Early advocates, like Musgrave (1959, ch. 8), of the comprehensive income tax approach were greatly encouraged by this important Canadian development, and Musgrave (1968a) himself wrote an enthusiastic but also typically probing and very thoughtful review. My own fascination with the concept likewise dates back to Simons and Carter, as I arrived in Canada as an academic migrant at the height of the Carter debate in 1968. Although the Carter package represents a pioneering application of the HaigSimons income concept, there are significant differences of detail between the Carter proposals and the original Simons scheme which was designed with the U.S. in mind. While, for example, the Carter Report follows Simons in proposing a full realisations tax on capital gains, the Carter scheme would retain a fully integrated corporate profits tax as a withholding tax on dividends and retention gains, as well as a conventional source tax on foreign direct investment (Carter Report, vol. 4 ch. 19). Simons, by contrast, would simply have abolished the corporate tax (1938, ch. 9). In the difficult area of owner-occupied housing, Simons would insist on taxing imputed rent, whereas Carter would accept the need for exemption on practical grounds. Remarkably, however, both Simons (ch. 6) and Carter (vol. 3, ch. 17) would tax bequests and gifts as income to the beneficiary – though the double tax on the donor would be relieved under Carter in the case of family members. A further important divergence is thus to be observed in the tax unit area, where the Carter reforms feature an elaborate family unit system, in contrast to the singleindividual approach to be found in Simons. The serious inequities and disincentive effects of such a family unit approach were not appreciated by the Commission (section 4.2 below). The Carter Report also deploys the subjective concept of non-discretionary expenses in a purported derivation of the progressive rate scale,

24

TAX REFORM IN THE 21ST CENTURY

which would clearly have been anathema methodologically to Simons (Head, 1970, pp. 208–210). In combination with the considerable base-broadening achieved, the top personal tax rate would come down from 82.4 per cent to 50 per cent, which was regarded (without much evidence) as a critical level if serious disincentive effects were to be avoided. Such rate-flattening was not, however, aimed to achieve vertical neutrality. Rather, a significant increase in vertical redistribution was intended. However, the similarities between the Carter proposals and the original Simons scheme are far more important than the differences of detail, significant as they undoubtedly are. In particular, the common framework of horizontal and vertical equity objectives is of overriding importance both for Simons and for Carter. From a political economy perspective, however, this overarching framework of distributional objectives – and the Haig-Simons income concept which is central to its implementation – is clearly problematic in terms of political acceptance. There was, to begin with, no guarantee that the Simons-Carter framework of equity objectives would find ready acceptance among the general public – or even among opinion leaders – especially in a country like Canada in which these concepts were almost totally unfamiliar. The academic consensus, achieved over a generation of familiarisation with the persuasive arguments of Henry Simons, had no counterpart in the wider public domain. Indeed the process had not even begun. There is, moreover, the obvious problem that those who stand to lose from the resulting tax reform proposals are highly likely to oppose them. Self-interested behaviour by potential losers in democratic political processes must clearly threaten the achievement of Simons-Carter tax reforms. Typically it may well be the case that a substantial majority will stand to gain from such redistributive reforms. These gains at lower-income levels are likely, however, to be small and widely diffused, whereas the losses from the proposed reforms in such areas as capital gains, bequests and gifts and tax incentives for the mining and petroleum industry are likely to be large and concentrated on groups that are already well organised and capable of a strong political response. In a well-functioning democracy, such narrowly self-interested behaviour might in principle ideally be ruled out. The tax system has a quasi-constitutional character, and Rawlsian tax-fairness principles – clearly related to Simon-Carter objectives – should therefore prevail (Head, 1997). Recognising the quasi-constitutional character of tax policy issues, a principled approach would be adopted under which taxpayer-voters would ignore self-interest and deliberate impartially on tax reform issues. This may well be a reasonable expectation for a Tax Review Committee or Royal Commission. For the real world democracies we face in practice, however, such idealised behaviour could not realistically be expected. Modern public choice analysis, which generally assumes self-interested behaviour in political processes, would strongly suggest therefore that Simons-Carter reforms must fail at the implementation stage. If this problem is to be overcome, a strong

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

25

foundation of popular support for Simons-Carter principles of equity and neutrality would seem to be required which could take years of familiarisation and vigorous advocacy. It is only when these principles are already well established as community values that the pressures of self-interest could be successfully resisted. In expressive voting models of the Brennan/Lomasky (1993) variety, this possibility is, however, considerably strengthened by the further observation that individuals have an obvious incentive to express their values rather than their interests, since their votes cannot generally be decisive, or even influential, in large-number electorates. The fact nevertheless remains that the Carter principles and proposals descended upon a totally unprepared community almost out of the blue. Not surprisingly, therefore, the Carter package received a hostile reception, especially from the business community and professional groups (Head, 1973). Except among academics, the Haig-Simons blueprint was unfamiliar to Canadians and was likely to be very difficult to explain and justify. The basic horizontal equity principle, enshrined in the Haig-Simons income concept, was however very neatly characterised from the outset in the Carter Report as ‘a dollar is a dollar’. Although deceptively simple, this characterisation in itself could hardly guarantee public acceptance, or even understanding, of highly complex measures in such areas as capital gains, company tax integration, family taxation, fringe benefits, bequests and gifts, and international tax. Without a strong foundation of public understanding and support, basic principles of horizontal and vertical equity may carry little weight, and specific proposals to tax capital gains or reduce business tax incentives were bound to be opposed by those stood to lose. As we have already suggested, the redistributive gains and the efficiency benefits of a more neutral tax system would generally be small and widely diffused, whereas the substantial losses would be concentrated on organised groups who would vigorously defend their tax privileges. Despite the hostile reception, the government’s initial response to the challenge of Carter in the Benson White Paper (1969) was quite impressive. Whilst the comprehensive income principle was breached in significant respects, for example in the decision to retreat from full to half inclusion of realised capital gains on listed equity shares, compensating measures in the corporate tax area ensured approximately equal treatment of closely-held and widely-held companies. A further such modification was the decision to tax capital gains on listed shares on a quinquennial accrual basis. Tax incentive provisions were more generous under the White Paper proposals than under Carter, but represented nevertheless a considerable rationalisation of the prevailing system. In general it is fair to say that equity and neutrality remained the foundation of a far-reaching program of reform that promised to reduce or eliminate a host of loopholes and discriminatory provisions in the income tax system. Further major hurdles remained, however, to be negotiated, as the White Paper proposals were opened to further public debate and exposed to detailed scrutiny by

26

TAX REFORM IN THE 21ST CENTURY

the major economic committees of the House of Commons and the Senate as well as the provincial governments. The Canadian government was firmly committed from the outset to consider seriously possible objections to its White Paper proposals in what was described as a process of ‘participatory democracy’ (Head, 1972). Substantial further erosion of the original Carter blueprint was therefore inevitable in the face of continuing strong opposition from the business groups and also from the provinces and the parliamentary committees. In the final legislation little remained, therefore, either of the Carter blueprint or the White Paper proposals. The Carter reforms had effectively been subjected to a ‘prior unravelling process’ under which they were mostly rolled back before ever reaching the implementation stage. Half inclusion of realised capital gains was the only significant exception. This latter change could, however, surely have been made without the ten-year tax reform process involving the Carter Royal Commission, the Benson White Paper and the public and parliamentary manifestations of participatory democracy. The direct impact of Carter – after ten long years – was therefore inevitably disappointing. The longer-term influences and effects were, however, of immense importance. The Carter Report was closely studied in subsequent reviews of taxation systems around the world, and the influence of the Carter proposals is clearly visible in tax reform initiatives of the mid-1980s in Australia and New Zealand, to be discussed below. The Report no doubt helped to strengthen the resolve of HaigSimons supporters as they pressed on towards the reform of the U.S. income tax in TRA86. Domestically the Royal Commission exercise served as a training ground for young Canadian tax reform specialists and transformed the level of tax policy debate with important consequences for the reform of the Canadian income tax over subsequent decades. The Carter legacy remains strong and resonates in tax policy discussion to the present day (Head, 1988).

4 The US Income Tax Reform of 1986 Musgrave (1968b) always rightly insisted that tax reform could never be successful without a clearly articulated vision of where we are going. For most of the leading tax economists of his own generation, including Pechman, Shoup, Goode and Vickrey, the Haig-Simons income concept was the touchstone. And great efforts were made over decades to propagate this vision at Congressional committee hearings, in public debate and, of course, in graduate and undergraduate classes. If tax economists could present a united front, both on the big picture and on the details of a comprehensive income tax reform package, then the prospects for success should be optimised. Tax lawyers, led by Stanley Surrey (1973), made their own unique and highly influential contribution through the closely related concept of tax expenditures. Following the Report of the Carter Royal Commission (1966) in Canada, inspirational blueprints followed for the United States, beginning with US Treasury

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

27

(1977) under David Bradford and followed by Treasury I (1984) under Charles McLure and Gene Steuerle. As in the Canadian case, full implementation was hardly likely, and the Treasury I (and subsequent Treasury II) proposals were vigorously opposed over the course of a lively two-year public debate. The Tax Reform Act of 1986 was nevertheless an impressive achievement, and was greeted with some enthusiasm by Musgrave (1987) and most academic commentators. There was nevertheless a significant division of academic opinion, which turned on the different approaches to reform adopted in Treasury I and TRA86 (McLure and Zodrow, 1987, pp. 38–39). The Treasury I proposals would have applied the concept of real economic income directly, both to the individual and corporate tax, by removing economically unjustified exemptions, deductions, exclusions and credits from the tax base, while applying comprehensive inflation adjustment. Timing problems were also comprehensively addressed, and the alternative minimum tax could therefore be repealed. In the corporate tax area, accelerated depreciation and the investment tax credit would be replaced by economic depreciation. The ‘double taxation’ of dividends under the classical system of corporate tax would at the same time be relieved, though not completely removed, through a deduction at the corporate level for one-half of dividends paid to shareholders. The corporate tax share of total income tax revenue would, however, be increased. The considerable basebroadening achieved would allow a significant scaling-down of personal tax rates (to 15, 25 and 35 per cent) with substantial benefits in terms of equity, efficiency and simplicity. By contrast, the focus in TRA86 was to close down tax shelters indirectly through elaborate interest deductibility and passive loss limitations, in association with a more limited range of politically more ‘acceptable’ base-broadening measures. Capital gains, however, would be fully taxed on realisation, though fringe benefits would remain largely untaxed. Inflation adjustment was not attempted and some measure of accelerated depreciation would therefore remain. The complex provisions of the alternative minimum tax were likewise retained. As in Treasury I, large reductions in individual income tax rates, with a top marginal rate of 28 per cent, would apply – funded in part by a substantial shift of the tax burden to corporations. Horizontal equity and investment neutrality would be much improved, though at the cost of considerable complexity. Musgrave (1987), for his part, felt that the rate-flattening in TRA86 had gone much too far, and he would have preferred to see increased progressivity. He also deplored the failure to reform the classical system of corporate tax. Musgrave’s vertical equity misgivings on rate-flattening have since been strongly supported on efficiency grounds in further research by Apps on tax unit issues and female labour supply (section 4.2 below). Although he was inclined to regard the comprehensive indexation scheme of Treasury I as overambitious, Musgrave was nevertheless

28

TAX REFORM IN THE 21ST CENTURY

concerned from a political economy viewpoint about the lack of CPI adjustment in the proposal for a full realisations tax on capital gains under TRA86. Both Treasury I and TRA86 would leave the highly distorting but politically sensitive deduction for mortgage interest untouched. Full investment neutrality would not therefore be achieved. Substantial efficiency gains could nevertheless be expected, assisted by roughly compensating distortions from the mortgage interest deduction for owner-occupied housing and accelerated depreciation for manufacturing. Although the direct base-broadening measures unravelled over time (Steuerle, 2004), it is interesting to observe that the much criticised interest quarantining and passive loss provisions survived and remained effective in closing down most tax shelters. Significant efficiency gains in investment allocation have likewise been sustained. Despite these differing perspectives on the merits of TRA86, the lesson on implementation from the US experience is therefore reasonably clear if somewhat daunting. With a strong professional consensus in active public support of a particular direction of reform, much can be accomplished. Much less attention seems, however, to have been devoted to strategies which might have helped to prevent the subsequent unravelling of the 1986 reforms. Little is to be gained from even the most comprehensive reform unless it can be sustained, preferably over decades. Efficiency gains can only be realised if the specific measures remain in place over a considerable period. The quasi-constitutional character of the tax system – strongly emphasised by Buchanan and others – is frequently ignored in the hurly-burly of electoral politics. The battle is not over with the passage of tax reform legislation like TRA86. On the contrary it is only beginning, as those who lose under the reform redouble their efforts to reverse the more offensive measures in the package. As we have seen, Musgrave (1987) was himself somewhat apprehensive regarding the lack of inflation adjustment under the new provisions in TRA86 for full inclusion of realised capital gains. And it is certainly true that there was very heavy pressure on this and other measures where considerable sensitivity could have been predicted. Given the very short electoral cycle in many countries – 3 years in Australia – the prospects for political survival of key features of the fiscal constitution must be clouded in uncertainty. Since these matters are not properly the subject of annual voting or budgeting, a comprehensive tax reform package like TRA86 should arguably be entrenched in special legislation which cannot be changed over an extended period of, say, 7–10 years. Even if such entrenchment was politically feasible, it is difficult to see how it could really be desirable. Minor adjustments and amendments typically follow in droves as the details of reform measures are sorted out, often over several years, in the light of experience and further analysis. All this assumes moreover that the structure of the reform is basically sound. By the time TRA86 was passed into

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

29

law, however, the original consensus among loyal supporters of the Haig-Simons income concept was already breaking down.

4.1 THE EMERGING CONSENSUS ON PERSONAL CONSUMPTION TAXATION Following the landmark contribution of William Andrews (1974), a new generation of tax economists were gathering in support of the personal consumption tax which had long been persuasively advocated by Fisher (1937b), Kaldor (1955) and the Meade Committee (1978). Bradford’s Treasury Blueprints volume of 1977 had indeed already presented the personal consumption tax as an attractive alternative to the comprehensive income tax. The more sophisticated modelling of intertemporal consumption choice by a younger generation of economists seemed to suggest that the expenditure tax option could offer a substantial boost to savings and intertemporal efficiency as well as horizontal equity. Even Charles McLure (1988), the principal architect of the impressive income tax reform blueprint presented in Treasury I, questioned whether TRA86 should be regarded as ‘tax reform’s finest hour’ or rather as the ‘death throes of the income tax’. Over the 20-odd years since TRA86 was first passed into law, the unravelling process has proceeded steadily, indeed relentlessly, to the point where the present US income tax is almost unrecognisably different from the comprehensive income tax model on which it had originally been based (Steuerle, 2004). In the meantime a new consensus on progressive consumption taxation has become established among leading tax economists which is fully comparable with the earlier consensus on the Haig-Simons income concept. And a similar consensus has since emerged among tax lawyers. Whether we can now expect a second wave of fundamental US tax reform based on the alternative consumption tax blueprint seems, however, highly doubtful. Early claims of substantial gains in terms of savings and efficiency by Summers (1981) and others are now considerably more muted (Auerbach and Kotlikoff, 1987). The expenditure tax model offers the greatest efficiency gains as compared with the labour income tax alternative, but it would be considerably more costly and complicated to administer and comply with – and much the same is true of the associated R+ F business tax which would serve as a source tax on foreign direct investment. Most of any benefit in terms of saving and intertemporal efficiency stems moreover from the lump sum tax on previously accumulated wealth which will extend over a lengthy transitional period as older taxpayers spend in retirement. From an equity viewpoint these benefits are obviously a mixed blessing, as a second layer of consumption tax is applied to saving much of which has already been taxed (Zodrow, 1997). The labour income tax, such as the X-Tax of David Bradford (2005), is much simpler, and would not involve the specific transitional inequities of the cash flow

30

TAX REFORM IN THE 21ST CENTURY

personal consumption tax. Progressive rates would be applied and, as compared with the present US income tax, vertical neutrality could be preserved. The personal tax on labour income would be supplemented by an R-base tax on business income, which would offer similar advantages in terms of neutrality and simplicity. Sophisticated modelling by Altig et al (2001) shows that welfare gains could be realised in every income class for the specific X-tax package modelled. The top personal tax rate required would be a modest 30 per cent. Transitional problems of various kinds remain, however. And there are serious doubts as to the creditability of the associated R-base business tax in the international setting. The apparent advantages in terms of simplification are moreover much reduced if, as Auerbach (2006) has recently argued, a separate tax on financial services is required. These doubts and problems are heavily compounded if, as many staunch supporters of the expenditure tax have strongly argued, some appropriate supplementary tax on wealth or wealth transfers is required. In this case the complications and political difficulties of capital income taxation under the HaigSimons income tax model are merely replaced by analogous difficulties under the supplementary wealth tax.

4.2 THE TAX UNIT ISSUE Over decades of disputation between supporters of the Haig-Simons income tax and Fisher-Kaldor-Meade consumption tax, the tax unit issue has generally been set aside. In the early dispute between Bittker (1967) and Musgrave (1968b) over the usefulness of the income concept as a guide to tax reform, it was agreed that the tax base concept provides in itself no guidance on the tax unit issue; and this would generally be accepted also by supporters of the consumption base, whether postpaid or prepaid. There seems to have been general agreement that the tax unit system raises separate issues and dilemmas which must somehow be resolved independently and on its own particular principles. As Apps (2006a) has shown, however, the equity and efficiency implications of alternative tax unit systems are first-order considerations in the analysis of tax reform and cannot be considered independently. As she has demonstrated in a remarkable series of papers (e.g. Apps, 2006b), the tax and associated welfare systems of many developed countries discriminate heavily against two-earner families. In a number of cases this is simply the result of tax unit systems based on joint or family taxation. In other cases a similar effect is achieved either through the means-testing of family welfare benefits on household income, through the application of ceilings on social security taxes, through flat-rate taxes and/or in other ways. Even now, crucial effects of the tax unit system are either ignored or modelled incorrectly in the specialised literature to rule out, for example, important real world phenomena such as unexplained heterogeneity of female work choices between market work and household production.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

31

It is still insufficiently recognised in the public finance literature that equal taxation of single-earner and two-earner couples, with the same family responsibilities and the same household income, grossly violates horizontal equity. Household income in this context is a seriously flawed index of economic position, as it completely ignores the benefits in terms of child care and other services generated through household production by the non-working spouse. As compared, say, with a flat-rate tax on household income, equity would be much better served by a progressive-rate tax applied under a tax unit system based on the individual. In the case of the two-earner couple, the lower marginal and average rate which would apply under separate filing correctly recognises their lesser capacity to pay. Swingeing disincentive effects on female labour supply found by Apps (2006b) for a number of West European and Anglophone countries are the result of tax and associated welfare systems which approximate joint or family taxation. As a result of such discriminatory treatment, female working hours are held down at a level well below male hours. Female labour supply and income tax revenue could be greatly boosted if these impediments were removed and the system returned to individual taxation or its equivalent. In the context of these important and urgently needed reforms of the tax unit system, the issue of rate-scale reform under the personal income tax can be seen in a completely different light. In particular, the case for rate-structure progressivity – for so long a major concern of Musgrave and the Haig-Simons school – comes back into its own as a method of keeping down average and effective marginal tax rates on low- to middle-income secondary earners with high labour supply elasticity. A similar argument applies to rate-structure progressivity in the context of personal consumption taxation, whether of the pre- or post-paid variety Musgrave would surely have welcomed the finding by Apps that basic tax policy objectives of horizontal equity, vertical equity and efficiency are, after all, fully compatible rather than conflicting over a broad policy range (Apps, 1996). He would also have welcomed the strong presumption against rate scale reform involving flat or linear rate schedules which he strongly opposed. Despite his own misgivings about the excessive rate-flattening under TRA86, Musgrave might, however, have been surprised by her doubts about the standard base-broadening and rate-flattening strategy which, by this time, was very much the conventional wisdom among HaigSimons income tax reformers.

5 Differential Incidence or Balanced Budget Incidence? The choice between the Haig-Simons progressive income tax ideal and the personal consumption tax alternative has been analysed and debated within the standard tax policy framework of differential incidence. Of the two major concepts proposed by Musgrave (1959, ch. 10) for the analysis of the incidence and effects of taxation,

32

TAX REFORM IN THE 21ST CENTURY

the differential incidence framework, involving equal-revenue comparisons of alternative taxes, has dominated theoretical and empirical studies. It is, however, balanced budget incidence, involving matching changes in tax and expenditure, which dominates the fiscal history of the past century – though it has been rather neglected in the literature. For Musgrave the evolution of the modern mixed economy, bringing together the tax and expenditure sides of the budget within the framework of balanced budget incidence, was a matter of central scholarly and practical concern. Provision for social wants such as national defence; provision for merit wants such as health care and education; and provision for social security: these were all important paradigms encapsulating different phases and dimensions of public expenditure growth. In his review of climate change, Nick Stern (2007, xviii, 27) has stated that global warming arguably constitutes the greatest market failure problem that the world has ever seen. In order to meet the challenge of climate change, large inframarginal policy adjustments will clearly be required. Vast additional outlays will be needed on alternative energy sources and technologies and on other farreaching adjustments and adaptations to limit the damage from global warming. In order to finance the huge balanced budget expansion in prospect, large revenue increases will obviously be required. In addition to carbon taxes and the sale of emissions permits, these seem certain to include revenues from the personal income tax, capital income taxes and corporate tax, as well as taxes on consumption and labour income. Tax reform clearly has a major role to play in this regard. In the area of income tax the massive revenue losses which result from tax loopholes exploited by the wealthy can no longer be afforded. Equity arguments are thus powerfully reinforced by considerations of revenue adequacy. The same is even more obviously true of the notorious disincentive effects of joint taxation on female labour supply. As Apps (2006a) has shown, female working hours in Australia have been held down to a mere 50 per cent of male hours by the deterrent effects of high average and effective marginal tax rules in combination with an expensive and inefficient child care system. In this balanced budget perspective, some of the niceties of differential incidence can arguably be dispensed with. The likelihood of a small but useful efficiency gain from the substitution of some form of consumption-based tax for income tax is arguably neither here nor there, since both forms of tax will probably be required if expenditure needs are to be met. Nor are the refinements of vertical neutrality from differential incidence analysis of such importance. The issues at stake are Galbraithian, and attention should be focused on the comparative distribution of expenditure benefits and tax costs (Galbraith, 1958, ch. 22). At this stage the distribution of expenditure benefits may be impossibly difficult to predict. Relevant policy analysis is thus arguably more rough and ready in the present context than the sophisticated modelling of differential incidence that we are accustomed to. This analysis must nevertheless be attempted.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

33

Musgrave (1994, p. 180) argued strongly that the modern predilection for sophisticated mathematical modelling should not be allowed to distract us from addressing the challenge of the real world problems we currently confront. These are clearly market failure problems on an unprecedented scale which call for a balanced budget rather than a differential incidence analysis.

6 The Australian Income Tax Reform Initiative The US and Australian tax reform initiatives of the mid-1980s exhibit interesting similarities and differences. Following the US model, the Australian tax reform process was conducted under Treasury leadership. The Draft White Paper (1985) was our equivalent of US Treasury I, though it remained only a very pale reflection of that extraordinary document, at least in the area of income tax. In the US, comprehensive base-broadening and rate-reduction strategies had long been accepted as the royal road to fundamental tax reform. As we have already noted, this approach had been vigorously advocated over decades by leading tax economists. The ambitious set of income tax reform proposals contained in Treasury I inevitably fell short to some degree at the implementation stage. The ultimate reform package legislated in TRA86 was nevertheless heavily influenced by Treasury I (as diluted in Treasury II), and represents in the context of real world tax reform a quite remarkable achievement. Correspondingly of course, there was much to unravel! In the Australian context, by contrast, the judgement was made that no such grand strategy of income tax reform could possibly be successful. The Asprey Report (1975) had sparked little public discussion, and it was hardly to be expected that the educational influence of our early conference volumes, useful as they were, could match the effects of decades of political involvement by leading US tax economists and tax lawyers in support of comprehensive income tax reform. For Australia the only serious prospect of achieving politically appealing and economically beneficial reductions in income tax rates appeared to lie rather in the introduction of a broadly based indirect consumption tax (BBCT), such as RST or VAT. Under the government’s preferred Approach C in the Draft White Paper (1985), modest income tax reforms would have been combined with a substantial tax mix change, involving a 30 per cent across-the-board cut in income tax rates to be financed by a 12.5 per cent broad-based RST, which would replace the existing narrow-based and highly distorting wholesale sales tax (WST). There was, however, very little support, even from the business groups, for Approach C at the National Taxation Summit in July. Nor indeed was there much support for the income tax reforms contained in alternative Approach A. At the conclusion of the Summit, this latter approach was nevertheless declared by Prime Minister Hawke to have been sufficiently well supported, and legislation was accordingly to be prepared by Treasury to give effect to Approach A. The

34

TAX REFORM IN THE 21ST CENTURY

resulting tax reform package was duly announced by the Treasurer Paul Keating on September 19, 1985. The Keating income tax reform package featured a handful of strategically chosen measures which were essentially of a piecemeal character falling well short of comprehensive income tax reform. While addressing directly some of the more obvious deficiencies of the existing system, these measures could realistically serve only as a first instalment of a more comprehensive set of reforms which might hopefully be implemented in stages as economic and political conditions allowed. It was, however, a serious concern under this piecemeal reform strategy that the original measures in the Keating package would not be followed up with the complementary reforms required to create a comprehensive and sustainable income tax system. It is indeed a fundamental proposition in second-best analysis that particular measures, which may be highly effective as part of a comprehensive income tax reform package, may well be ineffectual and even counterproductive if implemented in isolation. It was obvious, therefore, that the Keating reforms could prove difficult to justify and sustain over the medium to long term. We have already noted Musgrave’s dictum that effective tax reform requires a clear and detailed vision of where we are going. In a country like Australia where, in spite of Asprey (1975), the Haig-Simons income concept was not widely known or understood, there was an obvious need in the Draft White Paper to lay out in some detail an ‘Approach D’ to signpost the way ahead from Approach A to a consistent and sustainable income tax with a broad-base and considerably lower rates. Similarly, in announcing his tax reform package, Treasurer Keating should have been able to refer the community to an appropriate road map laying out future stages of reform. Turning now to the details of the Keating package, as we have already indicated, we are looking at only a handful of significant measures. In stark contrast to the US reforms of TRA86 where the base-broadening measures were estimated to boost income tax revenues by almost 25 per cent, the corresponding Australian reforms were expected to yield only a very modest $1.7 billion, a mere 6 per cent of personal income tax revenue. As a result, only limited reductions could be achieved in personal income tax rates. Although the top personal tax rate was reduced fairly substantially from 60 per cent to 49 per cent, the cut-in points remained essentially unchanged and unrealistically low. Indeed it would only be a few years before workers on average earnings would be facing the new top rate of 49 per cent. The reform of the personal tax rate structure fell far short therefore of what, after decades of neglect, was now urgently necessary. And this was the case even though the rate changes actually made cost some $4.5 billion – well in excess of the $1.7 billion raised through the limited base-broadening measures. Under the US reforms, by contrast, rates were reduced by an average of 25 per cent across the board with a top rate of only 28 per cent, and the new rate structure was fully indexed. The highly compressed and steeply progressive Australian rate structure would, however, remain unindexed.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

35

The limited scope of the base-broadening measures under Approach A ensured therefore that reform of the rate scale would inevitably remain merely cosmetic in basic design terms. Although the Keating Tax Reform Package was very small as compared with TRA86, it is interesting to observe that key measures for reform of the tax base went beyond those of the US reforms in important respects. As in TRA86, realised capital gains were made fully taxable as income, but with a CPI adjustment. The concern that had been expressed by Musgrave (1987) over the lack of such an adjustment under the realisations tax on capital gains in TRA86 was accordingly met. The benefits of such an adjustment applied on a piecemeal basis and confined to capital gains are, however, far from clear. Unless inflation adjustment is applied also in the case of interest income and expense, the problem of tax avoidance based on debtfinanced investment in capital-gains-yielding assets remains undiminished. There is indeed a strong argument that a purely nominal unindexed tax on capital gains would harmonise much better in basic design terms with the prevailing nominal income tax system. However this may be, the full realisations tax on capital gains, both in the US and in Australia, defied the prophesies of the doomsayers for a considerable period before, in the case of Australia, the unravelling set in following the Business Tax Review in 2000. The most important base-broadening measure in the Keating tax reform package was unquestionably the sweeping proposal to a tax non-cash fringe benefits to the employer at the new company tax rate of 49 per cent (equal to the top rate of the personal income tax, if we ignore the small Medicare levy). Together with the related and equally sweeping proposal to deny the entertainment expense deduction, the new FBT would account for well over half the additional revenue to be gained from the package, at least in the early years. In-kind fringe benefits had, at least in principle, long been taxable to the employee at their ‘value to the taxpayer’, but this provision was proving politically almost impossible to enforce. Following New Zealand precedent, it was accordingly decided to substitute an indirect tax on the employer for the direct but unenforceable tax on the employee. It is a well-established principle of tax analysis, going back to Musgrave’s early discussion in The Theory of Public Finance, that under competition, taxes imposed at the same rate on the same base – but on opposite sides of the same transaction – must have the same incidence and economic effects. For top bracket taxpayers the two taxes would therefore be equivalent, whilst for lowerbracket taxpayers the bias in favour of more conventional cash forms of employee compensation was arguably as it should be. Cashing out of in-kind benefits was accordingly very likely. In practical political terms it seems clear that the rapidly growing problems of tax avoidance in this area would never have been adequately addressed without this shift of onus from the employee to the employer. The new FBT has survived some minor structural adjustments, e.g. with the end of rate alignment in 1988, without significant unravelling. A large concession

36

TAX REFORM IN THE 21ST CENTURY

nevertheless remains for the company car, and this concession has recently been further liberalised. Comparable problems of fringe benefits taxation in the US income tax were simply not addressed under TRA86, as noted by McLure (1989) at our Retrospect and Prospect conference in December, 1986. Another obvious deficiency of TRA86, noted with some regret by Musgrave (1987), was the failure to address the distorting effects and inequities of the classical system of corporate tax. Going well beyond the somewhat inconclusive review of these issues in the Draft White Paper, the Keating package offered what appeared to be a blueprint solution to these problems in the form of a full imputation and rate-alignment system modelled on the full integration proposal by the Carter Commission (1966, ch. 19). There is, however, a possible conflict between domestic and international tax policy objectives under the Carter-Keating rate-alignment model, which serves to explain why the new Australian company tax system was abandoned after only one full year (1987–1988). With company tax rates declining in a number of major developed countries overseas, Australia’s 49 per cent rate was fast becoming uncompetitive and was accordingly reduced to 39 per cent. No comparable reduction could at that time be justified in the top personal tax rate which had only just been reduced substantially from 60 per cent to 49 per cent. The top personal and company tax rates were therefore forced apart under pressure from international tax competition. Indeed they have since diverged further as the corporate rate has been reduced to 36 per cent and 33 per cent, and it now stands at 30 per cent as compared with a top personal tax rate of 45 per cent (or 46.5 per cent including the Medicare levy). With a rate gap of 11.25 points in 1988–1989, widening to 16.5 points today, problems of tax avoidance through corporate retentions and income shifting immediately re-emerged. What had seemed the most remarkable innovation of the Keating package was therefore fundamentally flawed and, along with the interest-quarantining measures of 1985, was the almost instant victim of an unravelling process which has dogged the original Keating initiative down the years.

7 The Lessons for Tax Reform in Australia As we have already noted, the Musgrave Memorial Conference marks the 25th anniversary of our inaugural tax policy conference attended by the Musgraves in August 1982. That conference was the first in a series of such conferences and associated volumes extending over a 15-year period and feeding into academic discussion and public debate. The early conferences in the series contributed usefully to the major tax reform initiative conducted under Treasury leadership by Ted Evans and David Morgan in 1985–1986. It is an interesting coincidence of timing that the Rudd Government has just announced a major tax review to be conducted by a committee under Treasury

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

37

leadership by Ken Henry and Greg Smith (a former Treasury officer) with the participation of John Piggott, Heather Ridout and Jeff Harmer. This committee will review taxation and associated welfare measures at both the federal and state/local levels. An interim report on the federal age pension is scheduled for February 2009, and the final report is expected towards the end of the year. After 12 years of unravelling under the Howard government, and 22 years since the Keating legislation, such a review is clearly long overdue. The opportunity arose quite by chance out of the recent 2020 Summit of Australia’s ‘best and brightest’, and may well have rather shocked the present government. Given the chequered history of tax reform in Australia over recent decades – and the overwhelming influence of the business lobbies in tax policy and public debate – it has long been the conventional wisdom among Labor leaders that tax reform is political rat poison, to be avoided at all costs. Large scale tax avoidance by the wealthy has long been tolerated because of the widely-held belief that reform efforts in this area would amount to political suicide. Such a self-serving attitude might be comprehensible under a right-wing Coalition government of the John Howard variety. With a new Labor government in office a different attitude might be expected. History, however, suggests that Labor could well prove equally disinclined to remedy the situation. Even though the negative impact of such a reform would be confined to much less than 5 per cent of taxpayers – with the major impact on the top one or two per cent – it has long been the view that such measures would be politically perilous. Our previous review of the major tax reform efforts in the US and Australia during the mid-1980s certainly suggests the existence of pitfalls and difficulties. On the US model, a comprehensive reform package could require intensive preparation and educational efforts extending over decades. Even then the reform package might subsequently unravel, leaving little to show in the way of tax policy benefits at the end of the exercise. In the case of Australia, with the failure of Approach C at the Tax Summit in 1985, income tax reform was limited to a few strategically-chosen measures which in themselves could contribute only in a very limited way to the achievement of tax policy objectives. Complementary reforms required to achieve a consistent income tax were often not forthcoming or followed only with a considerable lag. And what little was eventually accomplished soon unravelled. At best the Australian tax reform story was one of two steps forward and three steps back. Of the tax reform efforts ‘closest to home’, it is the New Zealand reform package of the mid-1980s which offers the only really optimistic prognosis. These NZ basebroadening reforms were directly sparked by a major economic crisis of the period. Like the Australian reforms, however, overseas developments were also influential. On the income tax side, the Carter Report is known to have greatly influenced Finance Minister Roger Douglas. Other influences would include the US reform of TRA86 and the supporting blueprints such as Treasury I. As in Australia, the GST

38

TAX REFORM IN THE 21ST CENTURY

proposal and that for tax mix change were much influenced by the spread of VAT in Europe and by the generally lower average share of income tax in other OECD countries. The McCaw Report of 1982 was no doubt influential in articulating these views, and clearly had a greater impact than the Asprey Report in Australia – though the views expressed were fairly similar. There was, however, in New Zealand much the same lack of preparedness on income tax reform that we have observed in the case of Australia. The major difference in the New Zealand case was the overwhelmingly strong political leadership provided by Prime Minister David Lange and his Finance Minister Roger Douglas. Having once been convinced by the arguments for base-broadening, rate-flattening, sales tax reform and tax-mix change, Lange and Douglas were able, by virtue of their politically dominant position, to see these far-reaching blueprint-style reforms through to successful implementation. The US influence is clearly discernible in the income tax area where former US Treasury stalwart Eric Toder headed an influential advisory group within Treasury, and other North American as well as an Australian and other local tax specialists were commissioned to report and recommend solutions on particular topics. As David White (2009) has emphasised in his review of the New Zealand experience (below), much of the success of the NZ tax reform initiative must be attributed to innovative consultation and policy review procedures featuring close and effective interdisciplinary collaboration between lawyers, accountants and economists from the public and private sectors. Remarkable features of the NZ reforms included a comprehensive system of accrual taxation for financial derivatives, an end to most timing and deferral problems, a blueprint GST with remarkably few exemptions – and full inclusion of food – and full income taxation of retirement savings (White, 2009). From a political economy viewpoint, the key to political acceptance of the GST was the direct link with a major reform of the NZ welfare system. A near-blueprint system of company tax integration, based on full imputation and rate alignment at the low and internationally competitive rate of 33 per cent, was also achieved and maintained until the rise in the top personal tax rate to 39 per cent in 2000. Most capital gains, however, remained untaxed, reflecting political acceptability concerns and misgivings over possible lock-in effects of a realisations tax. This impressive package of base-broadening measures and tax-mix change permitted large across-the-board reductions and a considerable flattening of personal tax rates, fully matching in this respect the US achievement in TRA86. There has moreover been remarkably little unravelling – until the recent poorly designed concessions for saving. Apart from the somewhat incongruous exclusion of capital gains, the only significant failure was the attempt to introduce a blueprint system of accrual taxation under the ambitious NZ FTC arrangements.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

39

With the new saving concessions, however, cracks are now beginning to appear in the capital income tax system. The 20-year survival of this vast reform package in the face of lobbying pressures and ‘incentive’ proposals must, however, be accounted a remarkable and most encouraging tax policy achievement.

8 The Prospects for Tax Reform in Australia There is, I believe, much to be learned by the Rudd government from the New Zealand experience. There is also much to be learned from a careful consideration of the balanced budget framework of rapid long-term public expenditure growth in which we now find ourselves. It is, however, necessary to consider how the economic priorities and political economy of tax reform may have changed as a result of the global meltdown in financial markets over the recent period. As in the case of climate change, it will no doubt be strongly argued by business groups and other major beneficiaries of current concessions and gaps in the income tax base that now is not the time to bite the bullet of a genuinely principled and democratic tax reform. Indeed there will probably be intense lobbying for large personal and company income tax cuts and for additional ‘incentive’ provisions further narrowing the tax base. And the case for an across-the-board cut in the personal income tax to help cushion the impact of the meltdown on ordinary taxpayers is certainly compelling. The arguments for base-broadening in the income tax area, for closer coordination of the personal and associated company tax system, and for maintaining a substantial rate of corporate tax are, however, just as strong as they were in the pre-meltdown period. The long term revenue cost of the financial bailout and of associated fiscal and budgetary measures seems likely to be very large indeed, which further reinforces the case we have already made for strengthening the foundations of the present tax system and enhancing its revenue potential. The sense of injustice so keenly felt and strongly expressed by voter-taxpayers, scandalised by the abuses of the financial meltdown, has reinforced the drive for regulatory reform. With appropriate leadership, this enhanced awareness of financial abuse could easily be refocused on comparable abuses in the tax area which have continued over decades. In a situation of heightened uncertainty, a practical ‘veil of ignorance’ (Rawls, 1971) could well apply, considerably enhancing the prospects for principled tax reform and reducing the influence of vested interests. In this setting, an appropriate package of recommendations and a well-argued report from the Review Panel could provide the basis for fundamental, far-reaching and politically acceptable reform. Taxing the Wealthy: For large numbers of high-income taxpayers, income taxation has for too long been a voluntary ‘impost’. Major gaps in the tax base in such areas as capital gains, financial derivatives, interest deductibility and superannuation allow the wealthy to easily avoid tax. As a result, there is no longer any progressivity at the top of the personal tax scale, and the tax system as a whole

40

TAX REFORM IN THE 21ST CENTURY

remains approximately flat across the entire range of incomes – as indeed has been the case for decades (Warren, 1997, p. 590). Artificial tax avoidance in the narrow legal sense is still a problem, and tax evasion involving a variety of offshore manipulations occurs on a very substantial scale. Like David Lange and Roger Douglas in their heyday, PM Kevin Rudd currently occupies a dominant political position in Australia. If the Review Panel can come up with a comprehensive and well-argued package of income tax reforms, it is likely in the current political climate that Rudd and Treasurer Swan could implement it. Exploitation of tax loopholes by the wealthy costs multi-billions annually and grossly offends any reasonable sense of tax equity. Only a small minority of taxpayers would be much affected by the relevant measures – less than five per cent, and probably less than one per cent. Laid out clearly and comprehensively in the final report of the Review Panel – and hammered home by the government – the present abuses could well excite public outrage and fuel demands for reform. The measures required could well become an election-winner. Apart from the obvious measures to rein in artificial tax avoidance and prevent evasion, the central requirement in any such reform is to apply the comprehensive income principle much more consistently across the board. The 50 per cent discount for the inclusion of realised capital gains – so costly to investors and damaging to the economy in the current meltdown – should clearly be withdrawn. Accrual taxation of financial arrangements on the highly successful New Zealand model should immediately be implemented. Negative gearing abuses should be ended with effective quarantining of interest deductions on passive investments. And the present costly, inequitable and unsustainable superannuation tax concessions, introduced in the last years of the Howard government, must clearly be pruned back. Tax Incentives for Working Families: It may well be the case, after twelve years of John Howard, that a considerable constituency has been built up in support of the traditional single-earner family. As a result of welfare measures favouring such single-earner families – notably the Family Tax Benefits, Parts A and B – Australia’s two-earner ‘working families’ have been severely disadvantaged. As Apps (2006a) has clearly demonstrated, female working hours have been artificially depressed by a combination of high average and effective marginal tax rates and a costly and inadequate system of child care. Given our looming expenditure needs resulting from global warming and demographic change – and the huge backlog, after decades of neglect, in education, health care and infrastructure – the energies of our reserve army of well educated, prime age and totally frustrated women are now urgently needed to help meet the challenge of the future. In order to achieve these reforms, it is not necessary to withdraw the support we currently confer on single-earner families. All that is required is that similar benefits should be extended to working families. This could clearly be achieved in stages by liberalising the means test on Part A benefits. Part B benefits should arguably be abolished.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

41

It can hardly be argued that removing the present discrimination against working families would be to risk political suicide. On the contrary, such measures must provide the lynch-pin of any meaningful reform of our tax and welfare system aimed at benefiting Australia’s working families. With quality leadership from Kevin Rudd, this is surely another election-winner, and one to which the new Labor government must inevitably be committed. If, as many shrewd observers of the Australian political scene are inclined to believe, Rudd is capable of providing the leadership we need to confront the challenge of climate change, it is surely true that he is also the leader we need in tax reform. Indeed we cannot have the one without the other, since massive additional revenues will be required to cover these and other expenditure needs. The huge revenues we currently sacrifice through the recent superannuation tax concessions and tax avoidance by the wealthy simply cannot be afforded. And the same is clearly true of the even more massive revenues which are sacrificed as a result of disincentive effects on female labour supply. With a comprehensive and consistent income tax reform package, politically unacceptable rate increases can easily be avoided. Company Income Taxation: Integration or Rate Cut?: As we have already seen in Section 6 above, the present system of personal and company income taxation has long been very poorly coordinated with a large gap between the company tax rate and the top personal tax rate, a gap which has increased over the past 20 years. Tax avoidance problems, in the form of income shifting to corporate entities, have not been adequately controlled, and other inequities and distorting effects remain. In view of the present 16.5 per cent gap between the company tax rate and the top personal tax rate, there is likely to be strong pressure for realignment through a large cut in the top personal tax rates which are often said to be too high. To realign the rates in this way would, however, be massively costly in terms of revenue. It would also be vertically highly regressive and inequitable. The current top personal tax rate of 45 per cent is not significantly out of line with rates in comparable countries, nor is there any reason to suppose that disincentive effects on work effort or saving could be a problem. It has occasionally been suggested that radical base-broadening in the capital income area under the personal income tax could be employed to make up the lost revenue and help preserve vertical equity. Removing the 50 per cent discount on realised capital gains, accrual taxation for financial derivatives and limitations on nominal interest deductibility could provide examples. It is doubtful, however, whether such measures, desirable as they are, could easily be implemented in the face of long-standing opposition. The signs are indeed that personal and company tax rates are likely to be forced further apart, as strong pressures continue to build for a company tax rate cut. The familiar argument is that rates must be kept internationally competitive in order to

42

TAX REFORM IN THE 21ST CENTURY

maintain the level of foreign direct investment (FDI) in Australia and avoid capital outflow. It is necessary, however, to remember that the company income tax is our major policy instrument for securing an appropriate share of the benefits from FDI. If Australia’s natural resources are simply dug up and carted away by foreign-owned companies, there could be relatively little benefit remaining to the nation. Moreover, lowering the company tax rate in the case of resident companies only serves to increase the attractions of the corporate form as a tax shelter for the wealthy. There is a well known proposition due to Roger Gordon (1986) that a zero corporate tax rate would be optimal in the case of a small country like Australia or New Zealand. In the case of Australia, however, there are good and sufficient reasons to resist current pressures for rate reduction. As we know from recent experience, large amounts of much needed revenue can be raised without significant distortion by the taxation of location-specific rents from rich natural resource deposits. Indeed a recent study by the OECD (2007, ch. 5) clearly demonstrates that the international supply of capital is not in general highly elastic, either for large or small countries, as the zero-rate argument of Gordon and others would require. In a somewhat different context, Roger Gordon (1985) has in fact demonstrated that the corporate profits tax is not necessarily unfavourable in its effects on risky investments, whether domestic or foreign. Provided loss-offset provisions are adequate, the imposition of a company income tax may reduce the expected rate of return on investment, but it will also reduce the rate of return or risk premium required by investors. Distorting flows of investment capital may therefore present a much less serious problem under corporate profits taxation than has traditionally been argued. For a variety of economic reasons, therefore, Australia should think long and hard before following international trends towards lower rates of company tax which are much more common in countries lacking a strong natural resource base or other sources of location-specific rents. Other important economic priorities, notably revenue adequacy, clearly override simplistic arguments for rate cutting. Payroll Tax Reform: In the relevant balanced budget perspective, other proposed reforms also appear in a somewhat different light. For decades business groups have argued disingenuously for payroll tax abolition – or as in Fightback, for its replacement by the GST. In its most naive form the argument has been that the payroll tax (PRT) encourages capital for labour substitution and therefore destroys jobs. The fallacy in this argument was already definitively exposed by Samuelson (1961). The close similarity between the economic effects of a tax on labour income (such as PRT) and a consumption tax (like GST) has long been recognised in the public finance literature (e.g. Meade Report, 1978, ch. 8). For equal revenue taxes imposed at a single rate on a comprehensive base, and making standard incidence assumptions, the effects of PRT and GST must be virtually identical, as Freebairn (1993) has concluded.

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

43

It is certainly true that state payroll taxes do not conform to the design ideal of single rate and comprehensive base. This is, however, an argument for closer co-ordination and base-broadening under state PRTs rather than abolition and/or replacement by GST. This argument is moreover strongly reinforced in the balanced budget setting by the further consideration that both PRT and GST will be needed if our looming expenditure needs are to be met. Scarce GST revenue cannot be wasted on the replacement of state payroll taxes. GST Rate Increase: A similar argument would seem to suggest that it would be unwise to rule out a GST rate increase, as has already been announced under the terms of the Tax Review. The Australian GST rate of 10 per cent is at the low end of developed-country VAT rates, and an eventual rise to 15 per cent might need to be considered in the context of our looming expenditure needs. Any such rate increase in Australia requires the unanimous agreement of the six state governments and the federal government, and is therefore politically extremely difficult to achieve. The fact that all but one of these governments are now Labor could well offer a unique opportunity for, say, a small two per cent increase – an opportunity that may not recur. There may, however, be grounds for hesitation. A study by Apps (1997) shows clearly that a tax-mix switch involving substitution of GST for income tax raises EMTRs on second earners with high labour supply elasticities while lowering rates on high-income primary earners with low elasticities. The net effect of such a tax substitution is therefore to reduce work effort and increase tax distortion. The Apps analysis also shows that savings could well be reduced, since the bulk of saving comes from two-earner families. It might seem therefore, that a rise in the GST rate could be unambiguously bad news in terms of basic tax policy objectives. This argument too may need to be modified in the balanced budget setting where tax-mix change is not the issue, and additional GST revenue may be required to meet expenditure needs. The issue is therefore Galbraithian, and there is no presumption that the balanced budget package would necessarily be regressive. The regressive impact of the GST may well be balanced by a progressive distribution of expenditure benefits in areas such as health care and education (Galbraith, 1958, ch. 22).

9

Conclusions and Further Reflections

The political economy of tax policy remains to a significant degree uncharted territory on the road map to tax reform. Of the 4 Anglophone countries in our sample, Canada set out on the road to reform armed only with the impressive Carter blueprint and with minimal preparation of the electorate. Then followed the Carter debate, the Benson White Paper, and careful scrutiny by the major parliamentary committees and the provinces. After repeated testing in the crucible of public and political opinion, little remained at the end of a long ten-year odyssey that could

44

TAX REFORM IN THE 21ST CENTURY

be implemented, though the international spillovers have proved to be of great importance. The US reform effort was much more successful and was marked by an energetic and drawn-out public education program conducted by the public finance fraternity and directed at opinion leaders, which extended over several decades. Counting conservatively from the US Treasury Blueprints volume in 1977, the US reform effort rose to impressive heights with Treasury I in 1984 before subsiding somewhat under further review in Treasury II. The ultimate achievement in terms of basebroadening and rate-flattening – and the gains in terms of equity and efficiency – were highly praised by leading public finance specialists. Over a 15-year period, however, most of this achievement soon unravelled; and the likely benefits may anyway have been exaggerated. After a false start with the Asprey and Mathews Reports in 1975, the Australian tax reform initiative began in earnest with the publication by Treasury of the Draft White Paper in June 1985. The proposals from the DWP were considered at a National Taxation Summit the following month. The proposed new RST and associated tax-mix switch failed, however, at the Summit leaving a small but significant piecemeal income tax reform for implementation in the Keating Tax Reform Package in 1985–1986. Essential supplementary measures came through uncertainly and in stages over a 15-year period. And the original measures mostly unravelled, likewise in stages but some almost instantly, over the same period. Much the most successful of our sample countries was New Zealand. The proposals of the McCaw Report (1982) were vigorously followed up, further developed and effectively implemented under the dominant Lange/Douglas government of the mid-1980s. A classic broad-based VAT (GST) was introduced, and major income tax base-broadening and rate-flattening were successfully accomplished and combined with close integration Carter-style of the personal and corporate tax at a low and internationally competitive rate of 33 per cent. Introduction of the GST was greatly facilitated by an extensive reform of the welfare system. Until the most recent period, unravelling was not a problem, but the benefits of the reform may nevertheless have been exaggerated. Looking further ahead, the priorities for tax reform seem likely to be very different, with a heavy emphasis on rapid and sustained public expenditure growth required to meet the challenge of climate change, demographic ageing, infectious diseases and poverty which face so many countries. This challenge calls for a change of analytical framework from differential incidence to balanced budget incidence. As a result, considerations of revenue adequacy assume a heightened significance alongside the standard policy objectives of equity, efficiency and simplicity (or transparency). For purposes of illustration, this new framework is applied to Australia where a review of the tax system is just beginning. The outlook for reform appears in

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

45

political economy terms to be reasonably promising with a strong and charismatic Labor leader in control. The prospects for fundamental and much needed reform may even have improved if the recent turmoil in financial markets serves to generate great uncertainty resulting in a Rawlsian ‘veil of ignorance’ effect. This might help to mute self-interest and promote principled deliberation on tax reform proposals in which the standard tax policy objectives of equity, efficiency and simplicity would figure more prominently. There seems, moreover, little reason to suppose that the issue of revenue adequacy would rank any less highly. Though there may be grounds for optimism in tax reform when we look forward into the period ahead, it would be foolish indeed not to acknowledge the sorry record from the recent past as reviewed in this paper. Some of the difficulties which have been encountered reflect no doubt a variety of continuing economic disputes of a theoretical and/or empirical character, such as the choice between income and consumption base. Most, however, arguably reflect an array of what, for want of a better term, we have called political economy problems, which we have noted from time to time as they arose. Both types of problems – of economic analysis and of political economy – were of deep concern to Musgrave, both at a theoretical and perhaps especially at the practical level. I shall conclude therefore with some reflections on the political economy of tax reform suggested by the foregoing review. Our review of reform initiatives in a sample of Anglophone countries shows strong evidence of backsliding and unravelling. With the notable exception of New Zealand, it is a story of disappointment and ultimate failure rather than of sustained success. There is still considerable uncertainty and dispute over the best approach to the formulation and implementation of a tax reform package. Should we start by appointing an independent review committee or Royal Commission, or should we rely upon a Treasury-led initiative? Might an independent body such as the UK Institute for Fiscal Studies initiate a wide-ranging and in-depth inquiry? As we have seen in the foregoing discussion, even the proximate objectives of tax reform, including the choice between personal income and consumption tax, the issue of corporate tax integration and the most appropriate structure of specific tax types, are all still vigorously debated. In terms of basic public finance analysis, much has been resolved but a range of important issues remain in dispute. Even when an adequate reform package like TRA86 has been successfully implemented, our sample review of experience in four countries also highlights the danger of unravelling. The public finance literature has yet to deal fully and systematically with this issue. Progress in this area to date has consisted mainly of useful but scattered analytical insights and a few institutional initiatives, such as qualified majorities or a unanimity requirement for particular types of tax change, such as the Australian GST. Richard Musgrave’s celebrated multi-branch budget serves as a valuable heuristic in this regard, along with his insightful analysis of benefit and ability-to-

46

TAX REFORM IN THE 21ST CENTURY

pay taxation. Public choice analysis of the median voter variety has also contributed valuable insights. Buchanan’s insistence on the need for a properly constitutional attitude in tax policy making is of fundamental importance and extends beyond implementation issues to the unravelling question. And his interesting book with Congleton sets out an innovative approach to tax design and tax reform with a strong emphasis on the sustainability of the tax system over the medium to long term (Buchanan and Congleton, 1998). If these broader issues of implementation and sustainability cannot be satisfactorily addressed, there seems little justification for the continuation of the endless efforts which had been put into the analysis and design of tax reform. A continuation of past performance by tax economists into the 21st century would largely be a waste of resources. It is encouraging therefore to observe the attention devoted to these political economy issues by Alt et al. (2008) in the context of the Mirrlees Review. See also the interesting study by David White (Ch. 4, below). References Alt, J., I. Preston and L. Sibieta (2008), ‘The Political Economy of Tax Policy’, study for the Mirrlees Review, Institute for Fiscal Studies. Altig, D., A.J. Auerbach, L.J. Kotlikoff, K.A. Smetters and J. Walliser (2001), ‘Simulating Fundamental Tax Reform in the United States’, American Economic Review 91: 574. Andrews, W.D. (1974), ‘A Consumption-Type or Cash Flow Personal Income Tax’, Harvard Law Review 87: 1113. Apps, P. (1996), ‘Taxation of Families: Individual Taxation Versus Income Splitting’, in J.G. Head and R. Krever (eds), Tax Units and the Tax Rate Scale (Sydney: ATRF) 81. Apps, P. (1997), ‘A Tax-Mix Change: Effects on Income Distribution, Labour Supply and Saving Behaviour’, in J.G. Head and R. Krever (eds), Taxation Towards 2000 (Sydney: ATRF) 103. Apps, P. (2006a), ‘Family Taxation: An Unfair and Inefficient System’, Australian Review of Public Affairs, 7, 77. Apps, P. (2006b), ‘Female Labour Supply, Taxation and the New Discrimination’, Presidential Address presented at the XX Annual Conference of the European Society for Population Economics, Verona, 22–24 June. Auerbach, A.J. (2006), ‘The Future of Capital Income Taxation’, Fiscal Studies 27: 399. Auerbach, A.J. and L.J. Kotlikoff (1987), Dynamic Fiscal Policy (Cambridge, UK: Cambridge UP).

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

47

Australia (1985), Reform of the Australian Tax System, Draft White Paper (Canberra: AGPS). Bittker, B.I. (1967), ‘A “Comprehensive Tax Base” as a Goal of Tax Reform’, Harvard Law Review 80: 925. Bradford, D.F. (2005), ‘A Tax System for the Twenty-First Century’, in A.J. Auerbach and K.A. Hassett (eds), Toward Fundamental Tax Reform (Washington, DC: AEI Press) 11. Brennan, G. and J.M. Buchanan (1977), ‘Towards a Tax Policy for Leviathan’, Journal of Public Economics 8: 255. Brennan, G. and L. Lomasky (1993), Democracy and Decision (Cambridge, UK: Cambridge University Press). Buchanan, J.M. and R.D. Congleton (1998), Politics by Principle Not Interest: Toward Non-Discriminatory Democracy (Cambridge, UK: Cambridge UP.) Canada, Benson, E.J., Minister of Finance (1969), Proposals for Tax Reform, White Paper (Ottawa: Government Printer). Canadian Royal Commission on Taxation (1966), Report (Carter Report) (Ottawa: Government Printer). Fisher, I. (1937a), ‘A Practical Schedule for an Income Tax’, The Tax Magazine, July, 1. Fisher, I. (1937b), ‘Income in Theory and Income Taxation in Practice’, Econometrica 5: 1. Fisher, I. and H.W. Fisher (1942), Constructive Income Taxation. (New York, NY: Harper Bros). Freebairn, J. (1993), ‘The GST and the Payroll Tax Abolition’, in J.G. Head (ed.), Fightback! An Economic Assessment (Sydney: ATRF) 97. Galbraith, J.K. (1958), The Affluent Society (London: Houghton Mifflin). Gordon, R.H. (1985), ‘Taxation of Corporate Capital Income: Tax Revenues versus Tax Distortions’, Quarterly Journal of Economics 100: 1. Gordon, R.H. (1986), ‘Taxation of Investment and Savings in a World Economy’, American Economic Review 76: 1086. Gordon, R.H. (1992), ‘Can Capital Income Taxes Survive in Open Economies?’ Journal of Finance 47: 1159. Haig, R.M. (1921), ‘The Concept of Income – Economic and Legal Aspects’, in R.M. Haig (ed.) The Federal Income Tax. (New York, NY: Columbia University Press). Head, J.G. (1970), ‘Henry Simons Regained: Report of the Canadian Royal Commission on Taxation’, Finanzarchiv 29: 197.

48

TAX REFORM IN THE 21ST CENTURY

Head, J.G. (1972), ‘Canadian Tax Reform and Participatory Democracy’, Finanzarchiv 31: 48. Head, J.G. (1973), ‘Evolution of the Canadian Tax Reform’, Dalhousie Law Journal 1: 51. Head, J.G. (ed.) (1983), Taxation Issues of the 1980s (Sydney: ATRF). Head, J.G. (ed.) (1986), Changing the Tax Mix (Sydney: ATRF). Head, J.G. (1988), ‘The Carter Legacy: An International Perspective’, in W.N. Brooks, (ed.) The Quest for Tax Reform: The Royal Commission on Taxation Twenty Years Later (Toronto: Carswell) 367. Head, J.G. (ed.) (1989) Australian Tax Reform in Retrospect and Prospect (Sydney: ATRF). Head, J.G. (ed.) (1993), Fightback! An Economic Assessment (Sydney: ATRF). Head, J.G. (1997), ‘Tax Reform: A Quasi-Constitutional Perspective’, in G.S. Cooper (ed.) Tax Avoidance and the Rule of Law (Amsterdam: IBFD) 155. Head, J.G. and R. Krever (eds) (1996), Tax Units and the Tax Rate Scale (Sydney: ATRF). Head, J.G. and R. Krever (eds) (1997), Taxation Towards 2000 (Sydney: ATRF). Institute for Fiscal Studies (1978), The Structure and Reform of Direct Taxation (Meade Report) (London: Allen and Unwin). Kaldor, N. (1955), An Expenditure Tax (London: Allen and Unwin). McLure, Jr. C.E. (1988), ‘The 1986 Tax Act: Tax Reform’s Finest Hour or Death Throes of the Income Tax?’ National Tax Journal 41: 303. McLure, Jr. C.E. and G.R. Zodrow (1987), ‘Treasury I and the Tax Reform Act of 1986: The Economics and Politics of Tax Reform’, Journal of Economic Perspectives 1: 37. Musgrave, R.A. (1959), The Theory of Public Finance (New York, NY: McGrawHill). Musgrave, R.A. (1968a), ‘The Carter Commission Report’, Canadian Journal of Economics 1(1) Supplement: 159. Musgrave, R.A. (1968b), ‘In Defense of an Income Concept’, Harvard Law Review 81: 44. Musgrave, R.A. (1987), ‘Short of Euphoria’, Journal of Economic Perspectives 1: 59. Musgrave, R.A. (1994), ‘The Longer View’, International Tax and Public Finance 1: 175. New Zealand, Task Force on Tax Reform (1982), Report (McCaw Report) (Wellington: Government Printer).

JOHN HEAD: THE POLITICAL ECONOMY OF TAX REFORM

49

OECD (2007), Tax Effects on Foreign Direct Investment: Recent Evidence and Policy Analysis, No. 17. Rawls, J. (1971), A Theory of Justice (Cambridge, MA: Harvard University Press). Report of the Committee of Inquiry into Inflation and Taxation (1975), Inflation and Taxation. (Mathews Report) (Canberra: AGPS). Samuelson, P.A. (1961), ‘A New Theorem on Non-Substitution’, in J.E. Stiglitz (ed.) The Collected Scientific Papers of Paul A. Samuelson (Cambridge, MA: MIT Press). Schanz, G.v. (1896), ‘Der Einkommensbegriff und die Einkommensteuergesetze’, Finanzarchiv 13: 1. Shoup, C.S., R.M. Haig et al. (1937a), Facing the Tax Problem (New York, NY: Twentieth Century Fund). Shoup, C.S. and R. Blough (1937b), Report on the Federal Revenue System, submitted to Undersecretary of the Treasury Roswell McGill, Sept. 20. Simons, H.C. (1938), Personal Income Taxation (Chicago, Ill.: Chicago UP). Stern, N. (2007), The Economics of Climate Change: The Stern Review (Cambridge, UK: Cambridge UP). Steuerle, C.E. (2004), Contemporary US Tax Policy (Washington: Urban Institute). Summers, L.H. (1981), ‘Taxation and Accumulation in a Life Cycle Growth Model’, American Economic Review 71: 533. Surrey, S. (1973), Pathways to Tax Reform (Cambridge, MA: Harvard University Press). Taxation Review Committee (1975), Full Report (Asprey Report) (Canberra: AGPS). US Treasury Department (1977), Blueprints for Basic Tax Reform (Washington, DC: Government Printing Office). US Treasury (1984), Tax Reform for Fairness, Simplicity and Economic Growth (Washington: Government Printing Office). Warren, N. (1997), ‘Recent Trends in Australian Taxation and their Impact on Tax Incidence’, in J.G. Head and R. Krever (eds), Taxation Towards 2000 (Sydney: ATRF). White, D. (2009), ‘Income and Consumption Taxes in New Zealand: The Political Economy of Broad-base, Low-rate Reform in a Small, Open Economy’, ch. 4, this volume.

50

TAX REFORM IN THE 21ST CENTURY

Zodrow, G.R. (1997), ‘On the Transition to Indirect or Direct Consumption-based Taxation’, in R. Krever (ed.) Tax Conversations (London: Kluwer Law International) 187.

Chapter 2

International Prospects for Consumption-Based Direct Taxes: A Guided Tour Charles E. McLure, Jr and George R. Zodrow* 1

Introduction

The modern discussion of consumption-based direct taxation – with its ups and downs and ins and outs – has sometimes resembled a never-ending series of rides in an amusement park. Our purpose is to review key aspects of that discussion, noting that many of the issues that have bedeviled tax policy experts over the past 30 years are still unresolved. Since we have ridden some of the rides, often together, we will draw on our own personal experiences to motivate some of the discussion. We begin with a bird’s-eye view of the historical development of the amusement park and then zoom in on particular issues, before beginning the guided tour.

2 The Bird’s Eye View The modern discussion of consumption-based taxes began with Irving Fisher (1939), who emphasized that proportional consumption taxation achieves neutrality toward the choice of when to consume.1 By comparison, the taxation of capital income *

1

Senior Fellow, Hoover Institution, Stanford University and Cline Professor of Economics and Rice Scholar, Tax and Expenditure Policy Program, Baker Institute for Public Policy, Rice University, respectively. This paper is based on an article that appeared in (2007) FinanzArchiv 63: 285. Earlier advocacy of consumption-based taxation by Hobbes was based on the philosophical argument that individuals should in principle be taxed on what they took from society, as

52

TAX REFORM IN THE 21ST CENTURY

under an income tax increases the relative price of future consumption and thus favors present over future consumption. Subsequently, consumption tax advocates have noted that consumption-based taxation achieves horizontal equity between taxpayers with a given lifetime endowment (defined in present value terms), regardless of whether they earn early or late in life or choose to consume early or late. In addition, they have stressed that concerns that an expenditure tax would reduce the progressivity of the tax system are, at least in principle, misplaced, since under a direct consumption tax vertical equity could be maintained by appropriately adjusting the rate structure.2 In 1955, in An Expenditure Tax, Nicholas Kaldor advocated basing direct taxation on consumption. At that time most tax experts thought that implementing a consumption-based direct tax would be infeasible, because taxpayers would need to maintain either detailed records of expenditures or balance sheets, allowing saving to be deducted from income to calculate consumption.3 Kaldor showed that consumption could be calculated much more easily indirectly, by using accounting similar to that under an income tax except that individuals receive a deduction for net saving and are taxed on net withdrawals from saving. In 1974, in a seminal paper published in the Harvard Law Review, William Andrews (1974) questioned the conventional wisdom of that time that consumptionbased taxes are more difficult to administer than income taxes. Indeed, he argued convincingly that the exact opposite is true, as taxation based on consumption

2

3

measured by their consumption, rather than what they contributed, as measured by their income (Musgrave, 1959, pp. 161–164). For example, the President’s Advisory Panel on Federal Tax Reform discussed at length a ‘progressive consumption tax’ option that it estimated would have replicated the distributional burden of the current income tax system across income classes. Moreover, given widespread concerns about recent increases in income inequality, it seems unlikely (although certainly not impossible) that a true flat rate tax would be politically feasible. The President’s tax panel rejected a national retail sales tax, partly on the grounds that it would not be appropriately progressive, as required by President Bush’s charge to the panel. Robert Hall, one of the developers of the original flat tax proposal, has recently concluded that increasing income inequality implies that under a tax prepaid plan two or more individual marginal rates, in addition to the zero-rate applied to income covered by the personal exemption, would be desirable (Hall, 2005). Slitor (1972, pp. 229–30) notes that John Stuart Mill, A. C. Pigou, and John Maynard Keynes all doubted the feasibility of an expenditure tax, but concludes (p. 257) that an expenditure tax as a permanent adjunct to the income tax ‘is probably practicable within a relatively restricted scope of high-income, substantial wealth taxpayers’. Kelley (1970, pp.237, 253) suggests, ‘A fresh consideration of the problem may suggest, however, that an expenditure tax would not be substantially more difficult to impose than other forms of direct taxation currently enforced by developed and developing countries. … In the context of highly developed societies, there seems little reason to doubt the possibility of effectively imposing an expenditure tax’. Kelley notes that analysts at the US Treasury Department had earlier reached a similar conclusion regarding a wartime tax on spending. By comparison, Shoup (1969, p. 349) writes, ‘[T]he formidable administrative problems facing a mass expenditure tax make detailed discussion of it unfruitful at this time’.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

53

eliminates the timing problems (e.g. of deductions for depreciation, depletion and amortization) that greatly complicate the income tax. Shortly after Andrews refuted the charge that a consumption-based tax was not administratively feasible, the US Department of the Treasury (1977) and the Meade Committee in the UK (Institute for Fiscal Studies, 1978) produced the first detailed studies of alternative methods of implementing consumption-based direct taxes. One of their many important contributions was to show the conditions under which taxes based on consumption and taxes on labor income are equivalent. Nevertheless, it is important to remember that there are important differences between the two approaches, especially during the lengthy period of transition from an income tax, during which a consumption tax would apply to income from old capital while a wage tax would exempt deferred capital income from tax. In 1983, Hall and Rabushka proposed what has become the most famous form of consumption-based direct tax – and perhaps the most widely misunderstood – the ‘flat tax’, which combines a single rate income tax on wages and salaries above a sizable personal exemption at the individual level with a business level tax levied at the same rate on real cash flow, with immediate write-off (expensing) of all expenditures, exemption of interest income, and the elimination of interest deductions (Hall and Rabushka, 1983, 1985). These various proposals for consumption-based direct taxation were part of the background during the development of the US Department of the Treasury’s 1984 report to President Ronald Reagan, Tax Reform for Fairness, Simplicity and Economic Growth. Given the widespread interest in consumption taxes, that report’s endorsement of a traditional income tax – explained in Section 4.2 below – may have surprised many. David Bradford (1986) subsequently proposed that graduated rates be applied to the individual level base of the flat tax, producing his celebrated X-tax. Similarly, we proposed progressive consumption-based taxes for Colombia, Bolivia and other developing countries (McLure, Mutti, Thuronyi and Zodrow, 1990; Zodrow and McLure, 1991, McLure and Zodrow, 1996a, b), and McLure made similar proposals for Russia and other eastern bloc countries (McLure, 1992 a, b, c). We called the proposed tax the Simplified Alternative Tax (a term coined by McLure), to stress that it would be a simpler alternative to the income tax. Twenty years after the Treasury Department proposals, the Tax Reform Panel appointed by President Bush proposed consideration of a hybrid system in which an individual level flat rate tax on financial income would be combined with what is essentially the Bradford X-tax. The academic debate on introducing a consumption-based direct tax has focused on three issues: (a) the effects on investment, labor supply, economic output, and welfare, which depend in turn on (b) transition and wealth effects, and (c) international issues. Our discussion will concentrate on the issues of simplicity in

54

TAX REFORM IN THE 21ST CENTURY

administration and compliance and the international aspects of consumption taxes, issues that have been especially prominent in our work in developing countries and countries in transition. For discussions of the equally important issues of economic efficiency, equity, and transition, see Zodrow (2007) and Diamond and Zodrow (2007), and the literature cited there. In addition, we neglect any potential effects of tax reform on revenues; that is, we make the standard assumption that all proposals would be revenue neutral.

3

Some Preliminaries

3.1 TWO APPROACHES TO TAXING CONSUMPTION A tax on income that is not saved is a tax on consumption. Since consumption can be financed by borrowing or by drawing down previous saving, the consumption tax base must include the proceeds of loans and dissaving and must allow deductions for saving and loan repayment. Traditional Individual Retirement Accounts (IRAs), with a deduction for contributions, tax-free accumulation of interest, and taxation of both principal and interest (or other earnings on investment) when withdrawn, implement this approach. Because tax is postponed until saving is withdrawn for consumption, this can be characterized as the ‘tax postpaid’ approach to taxing consumption. The US Treasury (1977) described such a plan, which we shall refer to as Blueprints-1 (BP1), as did Aaron and Galper (1985). A key theorem underlying the economic case for consumption-based taxation is that such a tax does not reduce the return to marginal saving and investment and is thus neutral with respect to the choice between present and future consumption. Thus, assuming constant tax rates and ignoring bequests, inheritances, and abovenormal returns, a tax that excludes the return on investment such as a wage tax is economically equivalent to a tax on consumption.4 The Roth IRA operates in this way; no deduction is allowed for saving, but interest (and other returns to capital) are never subject to tax. Thus the tax that would be paid under the traditional IRA when principal and interest are withdrawn is ‘prepaid’ at the time saving occurs. The US Treasury (1977) also described such a plan, which we shall refer to as Blueprints-2 (BP2), and the Hall-Rabushka flat tax, the Bradford X-tax, the Simplified Alternative Tax, and (for individuals only) the McLure-Zodrow (1996 a, c) proposal for a hybrid consumption tax all follow this approach.

4

This equivalence can be extended to the case of uncertain returns (Kaplow, 1994; Zodrow, 1995) and above-normal returns (Hubbard, 2002). Nevertheless, these assumptions are certainly stringent, and the equivalence would break down if they were relaxed. For example, bequests and inheritances are an important component of wealth at the top of the income distribution, so their tax treatment must be taken into account, and the assumption of constant tax rates is implausible, given both recent history and the prospect that there will be large future federal budget deficits if current rates are not increased.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

55

Under most consumption-tax proposals, the financial transactions of business would be treated in the same way as those of individuals. In addition, all business expenditures would be deducted immediately in the year in which they are incurred, rather than being capitalized and deducted over time via depreciation allowances, amortization, depletion, etc. Another theorem says that, for equity-financed business investment, expensing implies that the marginal effective tax rate (METR) at the business level – the tax rate on income from marginal investments – is zero (Brown, 1948). Combining this treatment of expenditures on real assets with tax prepaid treatment of financial transactions yields what the Meade Committee called the real business cash flow tax base, or ‘R-base’, under which loans are ignored for tax purposes. By comparison, combining expensing with the tax postpaid approach yields the ‘real plus financial’ business cash flow tax base, or ‘R+F base’, under which interest income and the proceeds of loans are taxable and the repayment of principal and interest expense are deductible.5 The hybrid we proposed would employ the tax prepaid method for individuals with the tax postpaid method for businesses.6 To avoid confusion, we summarize the basic features (ignoring many complications) of all of these alternative consumption-based direct tax options in Table 1.7

5

6

7

See Institute for Fiscal Studies (1978). Two additional approaches deserve mention, although they have not received much attention in the US debate and will be discussed only in passing in this paper. The tax base under an S-base consumption tax, a version of which was adopted in Estonia in 1994, consists solely of all business distributions to shareholders. From 1994–2000 Croatia employed a tax that allowed an Allowance for Corporate Equity (ACE) – a tax that provides for deductions for economic depreciation but also allows an additional deduction for the cost of equity capital equal to the product of the book value of equity capital and a riskfree interest rate, and Belgium introduced such a tax at the beginning of 2006. See Boadway and Bruce (1984), IFS Capital Taxes Group (1991), and Devereux and Freeman (1991) for the intellectual origins and development of the ACE tax, Rose and Wiswesser (1998) and Keen and King (2002) for descriptions and evaluations of the Croatian experience, Genser and Reutter (2007) for recent reforms in Austria, Belgium and Italy that move in the direction of an ACE tax, and Gerard (2005) for a description of the new Belgian tax. In yet another alternative, the USA (Unlimited Savings Allowance) approach recommended by Senators Nunn and Domenici would combine the tax postpaid approach at the individual level with a value added tax (implicitly a prepaid tax) at the business level; see Weidenbaum (1996). In the interest of simplicity, we do not consider the more familiar indirect tax options, such as the value added tax or a national retail sales tax; for recent discussions of these options, see Gillis, Mieszkowski and Zodrow (1996).

56

TAX REFORM IN THE 21ST CENTURY

Table 1: Key Characteristics of Alternative Forms of Consumption-Based Direct Taxes BP1, AaronGalper

Flat Tax

BP2, XTax, SAT

McLureZodrow

USA Tax

Characterization

Tax postpaid

Tax prepaid Tax prepaid Hybrida

Hybridb

Business tax rates

Flat rate

Flat rate

Flat rate

Flat rate

Flat rate

Depreciable assets

Expensed

Expensed

Expensed

Expensed

Expensed

Business loans

CFc

Ignoredd

Ignoredd

CFc

Ignoredd

Individual tax rates

Graduated rates

Flat rate

Graduated rates

Graduated rates

Graduated rates

Individual loans

CFc

Ignoredd

Ignoredd

Ignoredd

CFc

Notes: (a) In the McLure-Zodrow hybrid, the tax on individuals is prepaid, but that on businesses is postpaid. (b) The USA tax is comprised of a postpaid tax on individuals and a (prepaid) value-added tax (VAT) on businesses. (c) CF or cash flow treatment of loans implies that interest income and the proceeds of loans are taxable, and repayment of interest and principal are deductible. This treatment characterizes a ‘R+F’ tax base that includes both real and financial transactions in the form of loans. (d) The case in which loans are ignored corresponds to an ‘R’ base, in which only real transactions are taxed; in particular, interest income is not taxed and interest expense is not deductible.

3.2 WHAT’S FLAT ABOUT THE FLAT TAX? WHAT’S UNIQUE? The Hall-Rabushka flat tax proposal consists of five separable components that, when combined, create its simplicity: business real cash flow and the labor income of individuals are taxed separately, business expenditures are expensed, financial transactions are ignored, there are no itemized deductions, and a single ‘flat’ rate is applied to the tax bases of both individuals and businesses.8 Unfortunately, the term ‘flat tax’ is often applied to tax systems and proposals for reform that exhibit 8

In the case of the individual tax, there is a large personal exemption that has the effect of creating a tax bracket that is subject to a zero tax rate.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

57

only the last of these features, the flat rate. Thus when it is said that a country has a flat tax or that someone has proposed one, it is unclear what this means. An income tax of the traditional type that is based on depreciation allowances and deductions for, and taxation of, interest (such as those recently enacted in Russia and several European countries in transition from socialism9) is a far cry from the prototypical flat tax, even if it has a flat rate and especially if it applies different tax rates to businesses and individuals and/or is shot through with exemptions, deductions, and credits. Of the five features of the flat tax, only one, the expensing of business expenditures, is unique to consumption-based taxation. By comparison, all the other features of the flat tax are consistent with an income-based tax, and all have been proposed, if not implemented, in that context (Slemrod, 1997; McLure, 1997).10 In a fundamental sense, the X Tax and the SAT are more like the flat tax than are ‘income’ taxes with flat rates.

3.3 THE CASE FOR CONSUMPTION-BASED TAXATION: SIMPLICITY ISSUES Consumption-based taxation is arguably simpler than income-based taxation, especially in an inflationary environment, and much of our work stresses the administrative and compliance aspects of the consumption vs. income tax debate (McLure and Zodrow, 1990), as do Bradford (1986), Slemrod (1996), Gale and Holtzblatt (2002) and Bankman and Schler (2007). 3.3.1 Basic Simplicity The Holy Grail of income tax reform has long been to implement the comprehensive Haig-Simons income tax base, defined as annual consumption plus the change in net worth over the year. At the very least, achieving this goal would require keeping track of changes in the value of assets and liabilities. In the United States, to a greater degree than in many countries, the values of these changes for tax purposes do not need to be the same as what is recorded on the balance sheets in a company’s financial accounts, although they must be reconciled in a separate tax schedule. Much worse, knowing how much net worth changes in a particular year entails, among other things, knowing how much an asset has depreciated, to what extent a mineral deposit has been depleted, how the current value of intellectual property compares with last year’s value, the change in the value of real estate and other 9 10

See Ivanova, Keen and Klemm (2005). This statement is most suspect with regard to ignoring financial transactions. The proposal for a comprehensive business income tax (CBIT) in US Department of the Treasury (1992) would ‘flip’ the treatment of interest income and expense, making interest non-deductible at the business level and exempt at the individual level. The CBIT would continue capitalization and depreciation, rather than allowing expensing.

58

TAX REFORM IN THE 21ST CENTURY

unique assets, etc. Inaccurate answers to these questions compromise the accuracy of the measurement of income, and imply that the essential economic neutrality goal of the income tax cannot be achieved.11 Moreover, because it is exceedingly difficult to establish empirically the true pattern of decline in asset values, tax provisions are susceptible to alteration for political reasons. Consumption-based taxation avoids most problems of income measurement, as its base does not include change in net worth.12 This is most easily seen in the case of the tax prepaid version. Neither principal nor interest transactions affect the tax base. There is no need to distinguish between debt and equity, since both are treated the same way. Capital gains are exempt, so there is no need to account for basis. Since all business expenditures are written off immediately, there is no need for depreciation accounting, amortization, depletion, etc. Being based on currentperiod cash flow, the tax postpaid method also avoids almost all timing issues. 3.3.2

Immunity to Inflation

Although commonly stated as above, the Haig-Simons definition of income should be consumption plus real change in net worth. That is, income measurement should be independent of the rate of inflation. Inflation erodes the value of depreciable assets, inventories (if valued at historical cost, especially under FIFO), the basis of capital gains, and the principal of debt. Unless the measurement of income accounts for inflation, income will be mismeasured.13 Unlike the base of an income tax, the base of a consumption tax is independent of the rate of inflation. Since consumption taxes are based on cash flows, there is no opportunity for inflation to erode the real values of the quantities that enter calculation of the tax base.

3.4

INTERNATIONAL ISSUES14

Bilateral tax treaties are employed to prevent double taxation of income. Double taxation – by the country where the income is earned (the source country) and the 11

12

13

14

By attempting (at least with regard to many issues) to create a conceptually attractive definition of taxable income, the Tax Reform Act of 1986 demonstrated just how complicated an income tax can be (McLure, 1988). The clear exception occurs when expensing results in an excess of deductions over gross income, what would be called net operating loss under an income tax. To achieve the economic neutrality and equity benefits associated with consumption-based taxation, it is necessary to carry such excess deductions forward with interest. This is problematic, because of the need to know the right rate of interest to use. Chile has long had a comprehensive balance-sheet-based system of inflation adjustment (Thuronyi, 1996). While the system is relatively simple, it cannot be as simple as consumptionbased taxation, which requires no inflation adjustments to be accurate. Our discussion will focus on the issue of creditability; for more general discussions of international issues, see Ballard (2002) and Bradford (2004).

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

59

country of residence of the recipient of income – can be avoided in either of two ways: a) the country of residence can exempt such income, or b) it can provide foreign tax credits for income tax paid to the source country, up to the amount of the residence-country tax on the income in question.15 Virtually all existing treaties are modeled on the three most important model tax treaties – the OECD model treaty, which is by far the most important since virtually all treaties involving developed countries are based on it, the US model treaty, and the UN model treaty. The model treaties, as well as all extant treaties, are, if only implicitly, based on the assumption that the contracting parties tax income and that it is avoidance of double taxation (or non-taxation) of income that is being sought. Consumption-based direct taxes of either the R or R+F types differ significantly from a common definition of income, the former because it does not tax interest income or allow a deduction for interest expense and the latter because it includes the proceeds of borrowing in the tax base and allows a deduction for repayment of principal. Any country that adopted a consumption-based direct tax of either of these types might thus need to renegotiate its tax treaties – a daunting task, not only because negotiating a treaty ordinarily takes many years, during which time international transactions would be conducted under a cloud of uncertainty, but especially because the negotiators would be operating in unknown territory, thus exacerbating the uncertainty.16 It has thus long been recognized that a consumptionbased direct tax might not qualify for relief from double taxation.17 Since the United States is the ‘800-pound gorilla’ in the world economy – and also has the most detailed and restrictive rules regarding crediting – most attention has focused on whether the US would allow a foreign tax credit for a consumption-based tax. No country seeking to attract investment from the US could afford to levy a tax

15

16

17

To the extent income is earned by foreign branches (permanent establishments) operating in the source country, the statement in the text applies without qualification. In the case of income earned by a foreign subsidiary, credit is allowed upon repatriation for income taxes paid by the subsidiary (and perhaps by one or more levels of sub-subsidiaries), as well as for withholding taxes on dividends paid to the parent. The US Department of the Treasury (1992, p. 48) acknowledges that elimination of the interest deduction under the CBIT, one aspect of the R-base, would require ‘extensive international discussions with tax authorities and market participants’. Somewhat anomalously, both an S-based tax and an ACE tax would be much less likely to be denied creditability. US rules on creditability can be traced to the need to avoid providing tax credits for production taxes levied on petroleum. Thus the emphasis has been on whether the tax in question is a tax on net income. A tax that allows no deduction for interest (the R-base tax) or that includes the proceeds of borrowing in the tax base (the R+F–base tax) is suspect, whereas an economically equivalent tax such as an S-based tax (which is applied only to corporate distributions) or an ACE tax (which allows a deduction for imputed interest on equity) is not problematical. Keen and King (2002) note that the creditability of the Croatian tax of the latter type that was levied from 1994 to 2000 was never challenged. Note, however, that Hall and Rabushka (1983) do not address the foreign tax credit issue, and devote barely a page to international issues.

60

TAX REFORM IN THE 21ST CENTURY

that would not be creditable, thus subjecting domestic-source income in the form of above-normal returns earned by US multinationals on FDI to double taxation. (Since normal returns are untaxed under a consumption tax, the US tax would be the only tax on such returns.)

3.5

INVESTMENT INCENTIVES

Apparently focusing on the fact that expensing of equity-financed investment is equivalent to a zero effective tax rate on the normal return to capital – a result that presumably is favorable to investment and growth and that does not depend on the rate of inflation – some seem to believe that expensing can – and perhaps should – be allowed in the context of an income tax. This is, however, unwise, especially in an inflationary environment. First, the combination of expensing and debt finance creates negative marginal effective tax rates at the entity level, and probably for the economy as a whole, implying that it might be profitable from the private after-tax point of view – but not from a social point of view – to borrow to make unproductive investments. Second, expensing allows growing firms – and debt-financed firms that are not even growing – to pay no tax. Aside from the undesirable effects on the perception of fairness, this undermines economic neutrality by encouraging mergers and acquisitions that are motivated primarily by the benefits of using tax losses immediately rather than deferring them (until the growing firm has positive profits) and perhaps losing them altogether. These problems were illustrated by the US experience of the early 1980s when a combination of investment tax credits and accelerated depreciation created a tax system that was more generous than expensing at prevailing inflation rates. A provision for ‘safe-harbor leasing’ was enacted which, in effect, allowed firms to sell excessive deductions and credits to firms that could use them. The specter of large and profitable multinational corporations paying no tax helped fuel the US Tax Reform Act of 1986 (Birnbaum and Murray, 1987).

4

Riding the Rides

Our experiences over the past two decades illustrate the importance of the issues discussed in the previous section.

4.1

JAMAICA

During the early 1980s, McLure participated briefly in the tax reform project in Jamaica headed by Roy Bahl, before his participation was cut short by his decision to accept a Treasury Department position with the Reagan administration. Had he continued in the Jamaica project, McLure almost certainly would have examined – and perhaps championed – the case for a consumption-based direct tax, but only after examining its creditability. As it was, he focused on explaining why it is unwise

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

61

to grant expensing of investment in the context of an income tax, especially in an inflationary environment.18

4.2 TREASURY I In the fall of 1983, McLure became the Deputy Assistant Secretary of the Treasury for Tax Analysis. During the State of the Union address in January 1984, President Ronald Reagan instructed Treasury Secretary Donald Regan to develop a tax reform plan ‘for fairness, simplicity, and economic growth’, and within a few days Regan had assigned McLure primary responsibility for developing that plan. As a visiting economist in the Office of Tax Analysis (OTA), Zodrow participated in preparation of the Treasury report, which Regan submitted to the White House in late November 1984. The tax reform plan became known as ‘Treasury I’, in recognition of the fact that it was, in Regan’s words, ‘written on a word processor’ and thus could be modified by the White House before being submitted to the Congress. Two things about the process and product of Treasury I surprised many observers: that the proposals could be kept under wraps until near the end and that they did not include a consumption-based direct tax alternative to the income tax.19 We will concentrate on the second. As noted previously, by 1984 a consensus was developing among public finance economists that a consumption-based tax would be preferable to the existing income tax – and even to an ideal ‘reformed’ income tax. Nevertheless, what Treasury I proposed was, despite proposed continuation of some features of then-current law, including consumption-tax treatment of many pension plans, essentially a HaigSimons income tax, with a primary objective of ‘taxing all income uniformly and consistently, without regard to source or use’. That objective does not allow exempting income from capital or allowing a deduction for saving. Even now many economists question the decision to opt for reforming the income tax, rather than replacing it with a consumption-based tax. (Indeed, it appears that some would have favored enactment of expensing in the context of an income tax, despite the pitfalls described earlier.) To understand that choice, it is necessary to appreciate the political context in which the Treasury Department was operating. The President of the United States had asked for a plan to reform the tax code. There was at least some chance that a presidential proposal that resembled whatever was proposed could be enacted. If Treasury had proposed a consumption-based tax, perhaps President Reagan could 18 19

See Section 3.4 above and McLure (1991). Treasury I did, however, include an entire volume devoted to examination of the value added tax, an indirect consumption-based tax, because McLure did not want to miss the opportunity to have the IRS take a close look at the tax. The VAT was not given prominence because of the Reagan Administration’s aversion to admitting that deficits were a problem or that higher taxes might be needed.

62

TAX REFORM IN THE 21ST CENTURY

have sold it. That would have been truly historic. We will never know, but it was not a sure bet. Moreover, the Treasury staff was concerned that if it put too many eggs in the consumption-tax basket, it might encounter ‘show-stoppers’ – problems for which it had no solution – late in the game. Three problems seemed particularly troublesome. First, given the political climate, it seemed unlikely that bequests would be treated as consumption and thus made subject to the consumption tax – as recommended, for example, in the R+F-based plan proposed by Aaron and Galper (1985). Many on the Treasury staff felt that such an outcome would be highly inequitable. Second, the international aspects of moving to a consumption tax raised serious concerns. In particular, adopting a consumption tax while our trading partners continued to tax on the basis of income would create a variety of opportunities for tax avoidance, tax evasion, and long or indefinite deferrals of tax. In addition, as discussed above, enacting a consumption tax would imply that all foreign tax treaties would have to be renegotiated, and existing provisions for relieving the double taxation of foreign source income via foreign tax credits would have to be revised. Finally, many on the Treasury staff were concerned that the transition to a consumption tax would be especially difficult. For all these reasons, the consumption tax route was abandoned early in the process of fundamental tax reform (McLure and Zodrow, 1987).

4.3

COLOMBIA

In 1986, shortly after passage of the historic Tax Reform Act in the US, McLure was invited to undertake a statutorily mandated study of the need for inflation adjustment of the income tax in Colombia. Based on the manifest advantages of a consumption-based direct tax, especially in an inflationary environment, McLure proposed successfully that the terms of reference be interpreted to include a detailed examination of the case for such a tax. This examination focused on the administrative aspects of implementing taxes based on both the R and R+F models (McLure, Mutti, Thuronyi and Zodrow, 1990). Notably, we did not fully examine the issue of creditability. In the final analysis, the government of Colombia settled on a system of inflation adjustment that resembled that pioneered by Chile some years earlier. This choice was fully justified, given the short time between submission of our report and the statutorily imposed deadline for modifying the system to introduce inflation adjustment, the novelty of consumptionbased direct taxation, and the unresolved issue of creditability.

4.4

BOLIVIA

In the spring of 1994, Juan Cariaga, the former Finance Minister of Bolivia, invited us to La Paz to discuss with President Gonzalo Sánchez de Lozada the latter’s ideas for a consumption-based direct tax. Despite warnings that such a tax might not be creditable in the US, both the President and Cariaga wanted to proceed.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

63

Bolivia was an ideal place to introduce a consumption-based direct tax. First, there was no corporate income tax; the consumption-based tax would be an additional source of revenue. Thus from an administrative-compliance viewpoint the transition would have been simpler than in virtually any other western country, and capital levy issues would have been much simpler. Second, the Bolivian value added tax (VAT) was one of the ‘cleanest’ of any developing country; it had few exemptions and only one rate. This would have greatly simplified linking compliance and administration of the two consumption-based taxes. Third, this was the President’s idea; it did not come from foreign advisers or the Ministry of Finance, and was thus much more likely to be enacted or at least considered seriously. 4.4.1

The Hybrid

We were convinced that the ‘tax prepaid’ method should be used for the taxation of individuals in Bolivia, in order to avoid the complexity of the ‘tax postpaid’ method. We were, however, reluctant to propose the same approach for business, as that would imply the exemption of the margin of financial intermediaries – not a politically attractive alternative, regardless of any economic rationale for such treatment.20 Moreover, despite its considerable theoretical appeal, the tax prepaid approach suffered from several additional serious practical problems. For example, expensing not combined with cash flow treatment of loans implied that the government would initially suffer relatively large revenue losses. For the same reason more firms would initially incur ‘losses’ for tax purposes, putting greater stress on the choice of the interest rate used to adjust carryforwards, an issue that would no doubt be controversial in any case.21 In addition, although the fact that financial transactions are ignored under the tax prepaid approach is the source of many of its simplicity advantages, it also creates problems, especially in transactions with foreign companies that are subject to an income tax. In particular, we felt that various avoidance schemes involving the manipulation of taxable receipts or deductible expenses and non-taxable interest income or non-deductible interest expense posed serious problems for the pre-paid approach that were not amenable to easy solutions. (For more details, see McLure and Zodrow (1996 a, b, c)). All of these problems are much more manageable if the postpaid or R+F approach is utilized at the business level. We ultimately decided that Bolivia could sensibly do what no one had ever proposed – apply the tax prepaid method to individuals and the tax postpaid method 20

21

Since loans are ignored under the R-base, financial intermediaries would neither pay tax on the interest income on their loans nor take deductions for interest expense paid to their depositors. As a result, their interest rate spread (the difference between their lending and borrowing rates) would be untaxed. For example, the carryforward rate should at least equal the risk-free rate of return, but since there is typically some probability that a firm will never realize the carryforward, an upward adjustment of an unknown magnitude would be appropriate.

64

TAX REFORM IN THE 21ST CENTURY

to businesses, without creating an asymmetrical system that would be an open invitation for tax arbitrage. We described and appraised obvious opportunities for arbitrage under this ‘hybrid’ system and concluded that the risks of arbitrage were less than the clear costs of either ‘pure’ system. 4.4.2

The US Obstacle

As noted above, a consumption-based direct tax was likely to be a non-starter if it were not eligible for foreign tax credits in capital-exporting countries, especially the United States. In Bolivia, despite not being lawyers, we undertook a serious attempt to determine whether such a tax would be – or should be – creditable in the US. In a potentially felicitous coincidence, a representative of the IRS came to La Paz while we were there and accompanied us to see President Sánchez de Lozada. She should have seen that the proposed hybrid tax was a well-conceived plan to improve the nation’s economic performance, and not a scheme to raid the US Treasury. (Indeed, the amount of Bolivian taxes credited in the US would almost assuredly have been less than under a conventional income tax.) But she took a very legalistic stance against creditability. We argued that the inclusion of the proceeds of borrowing in the tax base was exactly offset in present value by the subsequent deduction of debt repayment, but were told the IRS looks at form, not substance! We even undertook, with some assistance from the IRS representative, a demonstration that the business tax base under the hybrid tax would depart by only a few percentage points from that of an income tax. Again, she could not be persuaded.22 Finally, back in the US, we met with the technical staff of the Office of Tax Analysis of the US Treasury Department, who, much to the consternation of the IRS representative, seemed to accept our economic arguments favoring creditability. (The OTA staff consists of economists.) We then visited Joseph Guttentag, the International Tax Counsel, and received a far different reaction. First, we made the mistake of being accompanied by an attaché from the Bolivian embassy, which elevated the meeting to a ‘state-to-state’ plane and eliminated all possibility of open and frank discussion. Second, we suspect that the US Treasury would not have cared very much if only a Bolivian tax were at stake. But agreeing that a consumptionbased Bolivian direct tax would be creditable might pave the way for a request for creditability for a similar tax from other much larger countries. Bureaucratic risk aversion could easily explain the cool reception our proposal received. 4.4.3 Presidential Reluctance President Sánchez de Lozada had initially told us that he wanted to go ahead with the proposal for a consumption-based direct tax, despite our concerns about creditability. Ultimately, however, reason prevailed and he realized that his country 22

See McLure and Zodrow (1996a, b, c) for further description, discussion, and documentation.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

65

could not afford to impose a tax that might not be creditable in the US, as doing so could bring direct investment from the US to a halt. Somewhat surprisingly, we were asked to return the following year to examine the possibility of imposing an R-based tax on the resource sector in Bolivia – an exercise that ultimately also came to naught due to creditability concerns. In many respects, it was unfortunate that we were representing Bolivia in our efforts to have the IRS deem a consumption tax to be creditable, rather than a country to which the US could not so easily say ‘No’, such as Russia or China – and that the proposal was not floated before an administration that would presumably have been more receptive to the idea of consumption-based direct taxation and thus to creditability, such as that of George H.W. Bush. Indeed, the IRS in 1999 granted partial creditability to an Italian income/origin-based VAT – a tax that does not allow deductions for interest expense or wages. However, the relevance of this treatment for the general issue of the creditability of consumption taxes is limited as, consistent with earlier IRS positions on this issue, the amount of credit granted in the US reflected an adjustment that reduced the credit by implicitly introducing deductions for interest expense and wages in the calculation of the tax base (Rossi, 2002).

4.5 THE FORMER SOVIET UNION A consumption-based direct tax would have been highly appropriate for countries in transition from socialism. First, the stimulus to investment provided by expensing and the associated zero METR seems appropriate for countries needing to replace the outmoded industrial infrastructure inherited from the Soviet period. A consumptionbased direct tax would be far superior in that regard to the tax holidays that surely would be – and in fact were – proposed and adopted to achieve the same purpose under an income tax. Second, a consumption-based direct tax would be immune to the inflation that would almost certainly occur. Third, the simplicity of a consumptionbased direct tax would be a great advantage for a relatively poor region that did not have the resources to devote to administering and complying with an extremely complicated tax system; moreover, a simple new tax system might be easier to introduce successfully in countries that lacked a long history of tax compliance. Finally, transition problems did not seem to be much of a problem; since massive transition to something was inevitable, it might as well be to a consumption-based direct tax. Accordingly, McLure floated the idea of a consumption-based alternative to the income tax in a number of countries in transition from socialism (McLure, 1992a, b, c). 4.5.1

Russia

Shortly before the demise of the Soviet Union, a group of Hoover scholars began visiting Russia and working with successive Chairmen of the Supreme Economic

66

TAX REFORM IN THE 21ST CENTURY

Council of the Russian Federation. In 1992 several, including McLure, were even appointed foreign advisers to that body. In that capacity McLure (1991) proposed that Russia adopt the SAT. As it turned out, the Economic Council did not have – or could not retain – enough influence to matter.23 4.5.2

Kazakhstan’

In Kazakhstan a representative of Chevron, then by far the largest foreign investor in the country, told McLure, in effect, of the SAT, ‘Forget it. It would not be creditable’. Since that view echoed the concerns that McLure had expressed earlier in Jamaica, Colombia, and Bolivia, it was not hard to accept.

4.6

COLOMBIA REVISITED

The overwhelming reelection of President Alvaro Uribe in Colombia in May 2006 prompted interest in tax reform, especially of the business tax which is currently characterized by the relatively high tax statutory tax rate, including a temporary 10 per cent surcharge, of 38.5 per cent. Toward that end, Zodrow was invited to discuss options for fundamental tax reform. Many of the issues discussed in the paper thus far have played a prominent role in the deliberations. In particular, there is great interest in Colombia in both the Hall-Rabushka Flat Tax and the flat rate income taxes enacted in Russia and other countries in transition from socialism. The appeal of a consumption-based tax regime that is relatively favorable to foreign (and domestic) direct investment is tempered by concerns related to immediate revenue needs, creditability in the US, and design issues related to limiting avoidance opportunities. Similarly, an interest in the simplicity benefits of a flat rate is tempered by concerns about the distributional implications of flattening the individual rate structure. Finally, there is increasing recognition that although a consumption-based business tax system is desirable because it eliminates distortions of marginal investment decisions while taxing location-specific economic rents at the statutory rate, that rate simply cannot be too high, especially given increased evidence of international tax competition and the prevalence of ‘parasitic’ tax havens.24 In particular, a high statutory rate exacerbates problems with transfer pricing and other financial accounting manipulations, as well as discouraging investment by multinationals with relatively mobile firm-specific rents – highly

23

24

Another obstacle to the enactment of consumption-based tax reforms has been the opposition of the IMF. Indeed, at some point Milka Casanegra of the IMF, in effect, said to McLure, ‘I fought you in Colombia and I fought you in Venezuela and I will fight you here’, but never explained her – or the IMF’s – antipathy to consumption-based direct taxation. See Slemrod and Wilson (2006) and references cited there. See Zodrow (2006) for a recent discussion of the tax policy trade-offs faced by governments attempting to attract highly mobile international capital in the face of increased tax competition, including that from tax havens.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

67

desirable investments that generate positive externalities, especially in the area of technology transfer. It remains to be seen what lies at the end of this latest ‘ride’.

4.7 THE RECOMMENDATIONS OF THE PRESIDENT’S TAX REFORM PANEL IN THE US The two alternatives for tax reform recently presented by the President’s Panel on Federal Tax Reform reflect the current status of debate regarding the relative desirability of income-based and consumption-based direct taxes.25 Specifically, there is widespread agreement that an ‘ideal’ or comprehensive accrual-based tax on real economic income is not administrable, and significantly less but still considerable agreement that the taxation of the normal returns to capital that is inherent under an income tax is relatively undesirable. On the other hand, many observers are unconvinced that a movement to a true consumption tax is desirable or could be implemented in practice, citing uncertainty about the magnitudes of efficiency gains and improvements in administrative and compliance simplicity, as well as concerns about the distributional implications of such a reform and transitional problems.26 Reflecting this lack of consensus, the panel was unwilling to recommend either a true consumption-based tax or significant movement toward a more comprehensive income tax. Instead, the panel recommended replacing the current hybrid incomeconsumption tax system with one of two alternative hybrid systems. The first, the Simplified Income Tax, is an integrated income tax system that follows the traditional base-broadening, rate-lowering approach, especially for the individual income tax, but nevertheless includes a wide variety of consumption tax features. The second, the Growth and Investment Tax, is best described as a consumption tax system based on the X-Tax, supplemented with a layer of flat rate capital income taxation at the individual level. Adoption of either of the panel’s recommendations would move the US tax system closer, but not all the way, to a system of direct taxation based on consumption. By comparison, the alternative of a pure consumption tax based on the X-Tax, discussed at length in the report as its ‘Progressive Consumption Tax’ option, was unable to achieve the unanimous approval required for the panel to recommend it.

25

26

For further evaluation of the panel’s report, see Zodrow and McLure (2006). Zodrow, along with John Diamond, assisted the US Treasury Department in calculating the estimates presented in the panel’s report of the economic effects of the various reforms proposed, using a dynamic, overlapping generations, computable general equilibrium model similar to that utilized in Diamond and Zodrow (2006, 2007). For recent collections of articles that reflect the current status of the debate on these issues, see Zodrow and Mieszkowski (2002b) and Aaron, Burman and Steuerle (2007).

68

5

TAX REFORM IN THE 21ST CENTURY

Concluding Remarks

As we stagger from the amusement park, perhaps a bit dizzy from the rides, we note that enthusiasm for consumption-based direct taxation has ebbed and flowed over time. Indeed, our own enthusiasm has depended on the context in which we considered the case for a consumption-based direct tax. For example, we reached different conclusions regarding policy for Bolivia, for countries in transition from socialism, and for the United States in the period leading up to the Tax Reform Act of 1986. Exuberant advocacy of a switch to such a tax initially followed Andrews’ article debunking the view that a consumption-based direct tax would be administratively infeasible. However, this exuberance gave way to more measured support as tax experts examined more carefully the economic effects of such a switch, which depend crucially on how the transition is handled, the progressivity of the new system,27 and the administrative problems raised by various versions of consumption-based taxation. On economic grounds the case for the switch may be strong, but it is not the proverbial ‘slam dunk’ (Zodrow and Mieszkowski, 2002a). Moreover, hovering over the park is the black cloud of international issues. On the one hand, if the US (or any other country) were to enact a tax of either the R or R+F varieties, it would need to undertake the daunting task of renegotiating its double taxation treaties. On the other hand, no other country can afford to enact such a tax without assurance that the US would allow foreign tax credits for it. At the same time, however, two prominent US reform proposals – the Progressive Consumption Tax option discussed by the President’s tax reform panel and the Comprehensive Business Income Tax recommended by the US Treasury (1992) – would have denied deductions for interest expense and thus would fall into the category of ‘non-creditable’ taxes by current IRS standards. Only time will tell whether these recent high-level expressions of interest in tax structures that differ from a standard income tax will ultimately lead to more flexibility on the part of the IRS in deeming consumption-based direct taxes to be creditable against the domestic tax liability of US. References Aaron, H.J., L. Burman and C.E. Steuerle (eds) (2007), Taxing Capital Income (Washington, DC: Urban Institute Press). Aaron, H.J. and H. Galper (1985), Assessing Tax Reform (Washington, DC: Brookings Institution). Andrews, W.D. (1974), ‘A Consumption-Type or Cash Flow Personal Income Tax’, Harvard Law Review 87: 1113. 27

For example, see Altig, Auerbach, Kotlikoff, Smetters and Walliser (2001) and Diamond and Zodrow (2007).

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

69

Ballard, C.L. (2002), ‘International Aspects of Fundamental Tax Reform’, in G.R. Zodrow and P. Mieszkowski (eds), United States Tax Reform in the 21st Century (Cambridge: Cambridge University Press), 109. Bankman, J. and M. Schler (2007), ‘Tax Planning Under the Flat Tax/X-Tax’, in H.J. Aaron, L. Burman and C.E. Steuerle (eds), Taxing Capital Income (Washington, DC: Urban Institute Press) 245. Birnbaum, J.H. and A.S. Murray (1987), Showdown at Gucci Gulch: Lawmakers, Lobbyists, and the Unlikely Triumph of Tax Reform (New York: Random House). Boadway, R. and N. Bruce (1984), ‘A General Proposition on the Design of a Neutral Business Tax’, Journal of Public Economics 24: 231. Bradford, D.F. (1986), Untangling the Income Tax (Cambridge, MA: Harvard University Press). Bradford, D.F. (2004), The X Tax in the World Economy (Washington, DC: American Enterprise Institute). Brown, E.C. (1948), ‘Business-Income Taxation and Investment Incentives’, in L.A. Metzler et al. (eds), Income, Employment, and Public Policy: Essays in Honor of Alvin Hansen (New York: Norton), 300. Devereux, M.P. and H. Freeman (1991), ‘A General Neutral Profits Tax’, Fiscal Studies 12: 1. Diamond, J.W. and G.R. Zodrow (2006), ‘Consumption Tax Reform: Changes in Business Equity and Housing Prices and Intergenerational Redistributions’, paper presented at a conference on ‘Is It Time for Fundamental Tax Reform?’, sponsored by the Tax and Expenditure Policy Program, Baker Institute for Public Policy, Rice University. Diamond, J.W. and G.R. Zodrow (2007), ‘Economic Effects of a Personal Capital Income Tax Add-On to a Consumption Tax’, FinanzArchiv 63: 374. Fisher, I. (1939), ‘Double Taxation of Savings’, American Economic Review 29: 16. Gale, W.G. and J. Holtzblatt (2002), ‘The Role of Administrative Issues in Tax Reform: Simplicity, Compliance and Administration’, in G.R. Zodrow and P. Mieszkowski (eds), United States Tax Reform in the 21st Century (Cambridge: Cambridge University Press), 179. Genser, B. and A. Reutter (2007), ‘Moving Towards Dual Income Taxation in Europe’, FinanzArchiv 63: 436. Gerard, M. (2006a), ‘A Closer Look at Belgium’s Notional Interest Deduction’, Tax Notes International 41: 449. Gerard, M. (2006b), ‘Belgium Moves to Dual Allowance for Corporate Equity’, European Taxation 46: 156.

70

TAX REFORM IN THE 21ST CENTURY

Gillis, M., P. Mieszkowski and G.R. Zodrow (1996), ‘Indirect Consumption Taxes: Common Issues and Differences among the Alternative Approaches’, Tax Law Review 51: 725. Hall, R.E. (2005), ‘Guidelines for Tax Reform: The Simple, Progressive ValueAdded Consumption Tax’, in A.J. Auerbach and K.A. Hassett (eds), Toward Fundamental Tax Reform (Washington, DC: AEI Press). Hall, R.E. and A. Rabushka (1983), Low Tax, Simple Tax, Flat Tax (New York: McGraw-Hill). Hall, R.E. and A. Rabushka (1995), The Flat Tax (Stanford: Hoover Institution Press). Hellerstein, W. (2007), ‘The U.S. Supreme Court’s State Tax Jurisprudence: A Template for Comparison’, in R.S. Avi-Yonah, J.R. Hines and M. Lang (eds), Comparative Fiscal Federalism: Comparing the European Court of Justice and the U.S. Supreme Court’s Tax Jurisprudence (New York: Kluwer) 67. Hubbard, R.G. (2002), ‘Capital Income Taxation in Tax Reform: Implications for Analysis of Distribution and Efficiency’, in G.R. Zodrow and P. Mieszkowski (eds), United States Tax Reform in the 21st Century (Cambridge: Cambridge University Press), 89. IFS Capital Taxes Group (1991), Equity for Companies: A Corporate Tax for the 1990s, Commentary No. 26 (London: Institute for Fiscal Studies). Institute for Fiscal Studies (1978), The Structure and Reform of Direct Taxation: Report of the Committee Chaired by Professor J.E. Meade (London: Allen and Unwin). Ivanova, A., M. Keen and A. Klemm (2005), ‘Russia’s “Flat Tax”’, Economic Policy 20: 397. Kaldor, N. (1955), An Expenditure Tax (London: Allen and Unwin). Keen, M. and J. King (2002), ‘The Croatian Profit Tax: An ACE in Practice’, Fiscal Studies 23: 401. Kelley, P.L. (1970), ‘Is an Expenditure Tax Feasible?’, National Tax Journal 23: 237. McLure, C.E., Jr (1988), ‘The Tax Reform Act of 1986: Tax Reform’s Finest Hour or Death Throes of the Income Tax?’, National Tax Journal 41: 305. McLure, C.E., Jr (1991), ‘Expensing’, in R. Bahl (ed.), The Jamaican Tax Reform (Cambridge, MA: Lincoln Institute of Land Policy), 324. McLure, C.E., Jr (1992a), ‘Income Tax Policy for the Russian Republic’, Communist Economies and Economic Transformation 4: 425. McLure, C.E., Jr (1992b), ‘A Consumption-Based Direct Tax for Countries in Transition from Socialism’, The Wayne Law Review 38: 1697.

C.E. MCLURE, JR & G.R. ZODROW: DIRECT CONSUMPTION TAXES

71

McLure, C.E., Jr (1992c), ‘A Simpler Consumption-Based Alternative to the Income Tax for Socialist Economies in Transition’, World Bank Research Observer 7: 221. McLure, C.E., Jr (1997), ‘The Simplicity of the Flat Tax: Is It Unique?’, American Journal of Tax Policy 14: 283. McLure, C.E., Jr and G.R. Zodrow (1987), ‘Treasury I and the Tax Reform Act of 1986: The Economics and Politics of Tax Reform’, Journal of Economic Perspectives l: 37. McLure, C.E., Jr and G.R. Zodrow (1990), ‘Administrative Advantages of the Individual Tax Prepayment Approach to the Direct Taxation of Consumption’, in M. Rose (ed.), Heidelberg Congress on Taxing Consumption (New York: SpringerVerlag), 335. McLure, C.E., Jr and G.R. Zodrow (1996a), ‘A Hybrid Consumption-Based Tax Proposed for Bolivia’, International Tax and Public Finance 3: 97. McLure, C.E., Jr and G.R. Zodrow (1996b), ‘Creditability Concerns Doom Bolivian Flat Tax’, Tax Notes International 12: 825. McLure, C.E., Jr and G.R. Zodrow (1996c), ‘A Hybrid Approach to the Direct Taxation of Consumption’, in M. Boskin (ed.), Frontiers of Tax Reform (Stanford: Hoover Institution Press), 70. McLure, C.E., Jr, J. Mutti, V. Thuronyi and G.R. Zodrow (1990), The Taxation of Income from Business and Capital in Colombia (Durham: Duke University Press). Mill, J.S. (1921), Principles of Political Economy (London: Ashley, Longmans, Green and Company). Musgrave, R.A. (1959), The Theory of Public Finance (New York: McGraw-Hill). President’s Advisory Panel on Federal Tax Reform (2005), Simple, Fair, and ProGrowth: Proposals to Fix America’s Tax System (Washington Rose, M. and R. Wiswesser (1998), ‘Tax Reform in Transition Economies: Experiences from Participating in the Croatian Tax Reform Process of the 1990s’, in P.B. Sorensen (ed.), Public Finance in a Changing World (Houndmills: Macmillan Press), 257. Rossi, A.A. (2002), ‘Relief from Double Taxation under the New U.S.-Italy Tax Treaty’, Tax Notes International 28: 487. Shoup, C.S. (1969), Public Finance (Chicago: Aldine Publishing). Slemrod, J. (1996), ‘Which Is the Simplest Tax System of Them All?’, in H.J. Aaron and W.G. Gale (eds), Economic Effects of Fundamental Tax Reform (Washington, DC: Brookings Institution Press), 355. Slemrod, J. (1997), ‘Deconstructing the Income Tax’, American Economic Review 87: 151.

72

TAX REFORM IN THE 21ST CENTURY

Slemrod, J. and J. Wilson (2006), ‘Tax Competition with Parasitic Tax Havens’, National Bureau of Economic Research Working Paper W12225. Slitor, R.E. (1973), ‘Administrative Aspects of Expenditures Taxation’, in R.A. Musgrave (ed.), Broad-Based Taxes: New Options and Sources (Baltimore: Johns Hopkins University Press), 227. Thuronyi, V. (1996), ‘Adjusting Taxes for Inflation’, in V. Thuronyi (ed.), Tax Law Design and Drafting (Washington, DC: International Monetary Fund), 434. US Department of the Treasury (1977), Blueprints for Basic Tax Reform (Washington, DC: US Government Printing Office). US Department of the Treasury (1984), Tax Reform for Fairness, Simplicity, and Economic Growth (Washington, DC: US Government Printing Office). US Department of the Treasury (1992), Integration of the Individual and Corporate Tax Systems (Washington, DC: US Government Printing Office). Weidenbaum, M. (1996), ‘The Nunn-Domenici USA Tax: Analysis and Comment’, in M. Boskin (ed.), Frontiers of Tax Reform (Stanford: Hoover Institution Press), 54. Zodrow, G.R. (1995), ‘Taxation, Uncertainty, and the Choice of a Consumption Tax Base’, Journal of Public Economics 58: 257. Zodrow, G.R. (2002), ‘Transitional Issues in the Implementation of a Flat Tax or a National Retail Sales Tax’, in G.R. Zodrow and P. Mieszkowski (eds), United States Tax Reform in the 21st Century (Cambridge: Cambridge University Press), 245. Zodrow, G.R. (2006), ‘Capital Mobility and Source-Based Taxation of Capital Income in Small Open Economies’, International Tax and Public Finance 13: 269. Zodrow, G.R. (2007), ‘Should Capital Income Be Subject to Consumption-Based Taxation?’, in H.J. Aaron, L. Burman and C.E. Steuerle (eds), Taxing Capital Income (Washington, DC: Urban Institute Press). Zodrow, G.R. and C.E. McLure, Jr (1991), ‘Implementing Direct Consumption Taxes in Developing Countries’, Tax Law Review 46: 405. Zodrow, G.R. and C.E. McLure, Jr (2006), ‘Time for US Tax Reform? The Tax Reform Panel’s Recommendations’, Bulletin for International Taxation 60: 134. Zodrow, G.R. and P. Mieszkowski (2002a), ‘Introduction: The Fundamental Question in Fundamental Tax Reform’, in G.R. Zodrow and P. Mieszkowski (eds), United States Tax Reform in the 21st Century (Cambridge: Cambridge University Press), 1. Zodrow, G.R. and P. Mieszkowski (2002b), United States Tax Reform in the 21st Century (Cambridge: Cambridge University Press).

Chapter 3

Taxing Corporations in the European Union: Towards a Common Base? Sijbren Cnossen* 1

Introduction

The future of capital income taxation in the European Union (EU) hinges importantly on the future of the corporation tax (CT). Under the EU treaty, the Member States do not have to harmonize their CT rates or bases. Harmonization is to be ‘approximated’ only if required for the functioning of the internal market. But greater approximation of capital income tax systems could promote investment, improve the tax burden distribution and, last but not least, reduce compliance and administrative costs. While the normal return on mobile capital cannot be taxed at the same high rates as labour income, tax coordination should enable the Member States to capture some of that return. After all, capital is less mobile in the EU as a whole than between individual states. Tax coordination should also make it possible to tax firm-specific rents more effectively (although not at the same high rates as location-specific rents, if separately identifiable). Furthermore, there is no reason *

Professor Emeritus Erasmus University Rotterdam and University of Maastricht. Currently, Advisor, CPB Netherlands Bureau for Economic Policy Analysis and Senior Fellow, Taxation Law and Policy Research Institute, Monash University, Melbourne. This article is a revision and update of the author’s ‘Coordinating Corporation Taxes in the European Union: Subsidiarity in Action’ (Cnossen, 2008). The author paid tribute to Richard Musgrave at the 63rd Congress of the International Institute of Public Finance held at Warwick, 27-30 August 2007. The address was published as ‘Richard Musgrave’s Quest for an Index of Fiscal Equality Underlying the Horizontal Equity Concept’, FinanzArchiv 64(2): 147 (June 2008). Furthermore, he co-edited and contributed to Public Policy and Public Finance in the New Century, CESifo Seminar Series (Cambridge, MA: MIT Press, 2003) published in celebration of Richard Musgrave’s 90th birthday.

74

TAX REFORM IN THE 21ST CENTURY

why foreign share- and bondholders should be completely exempt from tax. Beyond that, the CT is needed as a backstop to the individual income tax (PT). Without a CT, the labour income of the self-employed would be retained in corporate form and largely escape the PT. In short, effective if moderate taxation of capital income seems desirable.1 There are two major routes towards greater coordination of corporation taxes in the EU: (a) a complex top-down all-or-nothing but neat ‘harmonization’ approach, favoured by the European Commission; or (b) a gradual bottom-up, largely reversible but messy ‘coordination’ approach, proposed in this paper. Both approaches have their pros and cons, but with the demise of the harmonization attempt, perhaps the time has come to explore ways to chart a course towards greater coordination by starting explicitly with existing corporate tax systems. In any case, as will be seen, the coordination approach ends up where the harmonization approach starts. Until recently, the Commission’s staff has been working hard to reach agreement on a common consolidated corporate tax base (CCCTB, for short) as the blueprint for harmonization. From the start, CCCTB had three essential elements: a common definition of taxable profits, a common definition of the groups of companies doing business in the EU which could opt to be taxed on their EU-wide consolidated profits, and an agreed-upon formula for apportioning those profits among the Member States where the groups carry on business. Similar to the formulas used in the United States, apportionment would be based on a combination of each group’s assets, wages, number of employees, and sales. Following apportionment, Member States would be able to apply their own (non-harmonized) tax rate and permit certain tax credits, if desired.2 These proposals constitute a fundamental break with current practice under which profits are determined separately for each Member State under the arm’s length accounting principle. CCCTB was launched at the beginning of the century (European Commission, 2001). The Commission soon found out, however, that several Member States were not enamoured of its harmonization approach; accordingly unanimous support – a constitutional requirement – was unlikely to be forthcoming. Subsequently, the Commission proposed, therefore, that the CCCTB should be considered viable as long as at least eight Member States agreed to proceed to harmonize their corporation tax systems. In late 2004, a CCCTB Working Group was convened to design the new system in time for an enabling directive by the Commission to be placed before the European Parliament and the European Council by the end of 2008.3 The Working Group met 13 times, but in April 2008, when it became clear that no agreement was in sight, it discontinued its deliberations. 1 2 3

For the rationale of retaining the CT, see Bird (2002) and for arguments for the positive role of the CT in a globalized capital market, see Zodrow (2006). For useful assessments of the CCCTB, see Fuest (2008) and McLure (2008). For the Commission’s communications and reports on the Working Group’s meetings, see the website of the European Commission. In addition to the CCCTB, the Commission initiated

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

75

Although the CCCTB seems to be a bridge too far, nevertheless the arguments for coordinating the capital income taxes, not just the corporation taxes, of the Member States appear persuasive. Tax competition might bring this about, but would be unlikely to be welfare enhancing. In the spirit of the subsidiarity principle,4 this paper proposes a gradual, bottom-up approach which would leave Member States both time and room for adjusting their capital income taxes to a model that would be defined more precisely as the tax reform proceeds. Also, a broadly based approach encompassing the taxation of all forms of capital income seems preferable to confining the coordination efforts to corporate profits. In search of a common coordinated approach, Section 2 reviews the existing CT regimes and attempts to find common features that could form the building blocks for further coordination. Following this, Section 3 outlines the steps that could be taken to promote CT coordination between the Member States. Section 4 concludes.

2

Survey of Corporation Taxes

Table 1 shows the CT systems that are found in the various EU Member States. The statutory CT rates range from 10 per cent in Bulgaria and Cyprus to 35 per cent in Malta. Statutory CT rates have been reduced by almost 11 percentage points since 1995 (McLure, 2008), following the liberalization of capital markets in the EU. In 2006, the CT ratio (CT revenue expressed as a percentage of GDP) ranged from 1.5 in Estonia to 5.5 in Cyprus (Eurostat, 2008). Since 1995, the weighted average ratio for all 27 Member States rose from 2.3 to 3.4 in 2006. Accordingly, the statutory rate reductions were more than offset by base broadening measures (and, possibly, other factors). Interestingly, in 2009, CT rates in the 12 new Member States are on average some ten percentage points lower than in the 15 old Member States. A partial explanation may be that agglomeration matters: core states appear to have higher CT rates than peripheral states (Garretsen and Peeters, 2007).

4 5

a pilot project on home state taxation for small and medium-sized enterprises. Home state taxation, i.e. taxation of EU-wide profits in the home state of a business, was another proposal in the original report on company taxation in the internal market issued by the European Commission (2001). For more on the subsidiarity principle, see Ederveen, Gelauff and Pelkmans (2006). More than full relief is possible under the CTs in Member States that permit the payment of dividends out of exempt profits without imposing a compensatory tax at the corporate level.

76

TAX REFORM IN THE 21ST CENTURY

Table 1. European Union: Corporation Taxes (CTs) and Individual Income Taxes (PTs) in 2009 (rates in %)

CT–PT System

CT Ratea,b

Double taxation Ireland

12.5e

Imputation system Malta UK

Tax Treatment of Dividends at ResidentShareholder Level

PT on Capital Gainse,d Ordinary Shares

Substantial Holdings

PT rate 20/41

22

22

35 28

Tax credit 35 /65 of dividend 1 /9 of dividend

— 18

15 18

Reduced PT rate Austria Bulgaria Cyprus Czech Republic Denmark France Greece Hungary Italyg Lithuania Poland Portugal Romania Slovenia Spain Sweden

25 10 10 20 25 34.43 25 20 27.5 20 19 26.5 16 21 30.01 28

PT ratef 25* 5* 15* 12.5* 28/43/45 18 10* 10 12.5* 20 19* 20* 16* 20* 18 30

— — — — 28/43/45 — — 20 12.5 15 19 — — 20 18 30

½ of gain 10 — — 28/43/45 29 5% gross 20 2 /5 of gain 15 19 10 16 20 18 30

Dividend exemption Finland Germany Latvia Luxembourg Netherlands Slovak Republic

26 29.83 15 29.63 25.5 19

Size of exemption 3 /10 of dividend ½ of dividend Full ½ of dividend Full Full

28 — — — — 19

28 ½ of gain — ½ of gain 25 19

ACE system Belgium

29.68h

PT rate 25*

—

—

No CT Estonia

—

PT(CT) rate 21

21

21

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

a

b

c

d

e f g h

77

CT rates include (i) a surtax in Hungary (4 per cent), (ii) surcharges in Belgium (3 per cent), France (3.3 per cent), Germany (5.5 per cent), Luxembourg (4 per cent), Portugal (1.5 per cent) and Spain (0.75 per cent–0.01 per cent; deductible from profits), and (iii) local taxes in Germany (effectively 14 per cent on average) and Luxembourg (6.75 per cent). A flat minimum tax, creditable against the final CT, is levied in Austria. Lower or graduated CT rates apply to lower amounts of profits or to small businesses in Belgium, France, Hungary, Luxembourg, the Netherlands, Portugal, and the UK. PT rates shown are for long-term capital gains (in Slovenia, the capital gains tax rate declines by 5 percentage-points for every 5 years of the holding period). Austria, Belgium, the Czech Republic, Germany, Luxembourg, Portugal, Rumania and Slovenia tax speculative capital gains on shares held less than a specified period. Short-term gains (generally, arising in a period of less than 12 months) are taxed at higher (effective) PT rates in Denmark, Germany, Portugal, Rumania, Spain and the UK. Various Member States exempt small amounts of capital gains or tax them at a lower rate. Generally, capital gains are not adjusted for inflation, except in Ireland. In addition, individual net wealth taxes are levied in Spain (0.2-2.5 per cent), France (0.55– 1.8 per cent) and the Netherlands (1.2 per cent; called income tax). In Luxembourg, companies are subject to a 0.5 per cent net wealth tax. Ireland applies a 10 per cent rate to the profits of manufacturing companies. An asterisk (*) indicates that the PT rate is a final withholding tax, which is optional in Portugal (and Belgium). Italy also levies a 3.9 per cent regional tax on productive activities (IRAP) in the form of an income-type value-added tax. In Belgium, the CT rate of 33.99 per cent is reduced by an allowance for corporate equity (ACE), which was 4.307 per cent in 2008.

Source: Author’s compilation from International Bureau for Fiscal Documentation, Microsoft Internet Explorer, Europe – Corporate Taxation and Europe – Individual Taxation, accessed March 5, 2009. Most information in the data base was updated in the fall of 2008, but the data for some Member States relate to earlier years. Reference has also been made to Eurostat (2008) and to OECD (2006).

2.1

CORPORATION TAX REGIMES

As indicated in Table 1, CT regimes can be distinguished depending on whether and to what extent they reduce the double tax that arises when corporate profits are subjected to the CT and again to the PT when paid out as dividends. Relief for this double tax can be provided at the shareholder level or at the corporate level. No relief is provided at either level when, as in Ireland, dividends are taxed at the full PT rate, although the overall effect is mitigated by the low Irish CT and PT rates. 2.1.1 Shareholder Level Relief Under the imputation systems, found in two Member States, shareholders are given a full (Malta) or partial (UK) tax credit against their PT for the CT that can be imputed to the dividends (grossed up by the tax credit) received by them. Accordingly, imputation reduces the excess CT+PT burden on profit distributions

78

TAX REFORM IN THE 21ST CENTURY

in proportion to the marginal PT rates of shareholders.5 Under full imputation, as in Malta, distributed profits are taxed at the marginal PT rate of shareholders. In the UK, the relief is so small that the effective CT+PT burden differs little from the burden under double taxation. Imputation systems, long supported by the European Commission, are the most structured form of dividend relief at the shareholder level. They had dominated the CT picture in the EU, but the cross-border implications were found to be discriminatory6 and overly complicated. The double tax can also be mitigated at shareholder level by subjecting dividend income to a separate or schedular PT rate lower than the top PT rate – the approach adopted by 16 Member States. Consequently, the relief is proportionately greater for high-income-bracket PT payers than for low-income-bracket PT payers. This regressive result can be mitigated but not eliminated, by permitting low-incomebracket PT payers whose marginal ordinary PT rate is lower than the reduced PT rate to opt for full double taxation of their dividend income (with a credit for any PT withholding tax imposed at the corporate level), as is possible in Portugal and Belgium. In ten Member States the dividend withholding tax constitutes the final PT liability. Furthermore, exempting dividend income from the PT, fully or partially, can provide dividend relief. Six Member States follow this approach. A full exemption is equivalent in effect to a schedular PT rate of 0 per cent. More generally, a partial exemption expressed as a fraction, α, of the total dividend, is equivalent to α multiplied by the ordinary PT rate under the schedular approach. The exemption approach, however, does not permit the imposition of a final withholding tax at the level of the corporation, which would save on compliance control costs. 2.1.2 Corporate Level Relief Relief from double tax, in principle equivalent to the relief under imputation (Cnossen, 1997), can be provided at the corporate level under a split-rate system (i.e. a lower CT rate on distributions than on retentions) or a dividend-deduction system. None of the Member States applies these forms of relief. Instead, two Member States exempt part (Belgium) or all (Estonia) profits from CT. Belgium does this in the form of an allowance for corporate equity (ACE), called ‘notional interest on corporate capital’. Interest is set at the rate payable on ten-year government bonds issued in the previous year; presumably, this rate approximates the normal rate of return on capital. The rate – 4.307 per cent in 2008 – is applied to the company’s

6

7

Presumably, for this reason, Malta imposes a 15 per cent tax on dividends paid out of untaxed profits. In Manninen (Case C-319/02), issued 7 September 2004, the European Court of Justice held that the Finnish imputation system was in violation of the free movement of capital principle in the EU Treaty. By following a low rate/large ‘tax base’ philosophy, Ireland has snatched sizable revenues from other Member States. Ireland’s CT/GDP ratio was 3.8 in 2006 compared with an EU-15

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

79

‘risk capital’, i.e. its equity shown on the balance sheet. Belgium introduced the ACE system to stimulate the self-financing capability of corporations. The tax rate reduction is inversely proportional to a corporation’s profitability. In contrast to Belgium, Estonia exempts the whole of corporate profits from CT. Instead, a 21 per cent tax is levied upon the distribution of corporate profits to individuals or other corporate entities. The profits distribution tax has not eliminated CT revenues, which still are 1.5 per cent of GDP, somewhat less than half the weighted average for the EU. 2.1.3

Capital Gains

Double taxation also occurs when retained profits are subjected to the CT and again to a capital gains tax at the shareholder level on increases in share values – increases that, among others, reflect the corporation’s greater net worth as a result of profit retention. Table 1 indicates that 13 out of 27 Member States tax capital gains on (widely-held) ordinary shares (e.g. quoted on national stock exchanges). Twentythree Member States tax gains realized on the sale of other (non-traded) shares, which represent a controlling interest (called substantial holding, variously defined) in (closely-held) corporations. Capital gains on these holdings are more widely taxed than gains on traded shares because they often represent labour income sheltered in the corporate form at a CT rate that is lower than the marginal PT rate on other labour income. The capital gains tax rates shown in Table 1 are the nominal rates. Deferral and various tax base preferences result in low effective rates. Furthermore, it should be noted that no Member State makes a systematic attempt to alleviate the double tax on retained profits (as Norway did until 2006) by allowing shareholders to increase the acquisition price of shares by the corporation’s retained profits net of CT.

2.2

COMPARATIVE ANALYSIS OF CT REGIMES

A thorough review of the corporation tax regimes in the EU, including tax bases, tax incentives and anti-tax avoidance measures (see Cnossen, 2005) indicates that the CTs in the EU Member States are levied at widely differing rates applied to widely differing tax bases. No state heeds the normative implication of the accretion concept of income that the taxation of corporate earnings (profits and interest) should be integrated with the PT of shareholders. Generally, dividend income is taxed at reduced PT rates or (partially) exempted, while capital gains on shares are taxed at very low effective rates (or fully exempted). Furthermore, interest is taxed at lower rates than apply to retained profits or dividend income. Overall, capital income is taxed separately from and at much lower rates than labour income which is subject to the PT and various hefty, regressive social security contributions. Corporate profits are determined on the basis of International Financial Reporting Standards (IFRS), the European-wide rule from 1 January 2005 for companies listed

80

TAX REFORM IN THE 21ST CENTURY

on EU stock exchanges. For tax purposes, however, assets are generally valued at historic cost and capital gains are taxed on a realization basis. The state-specific rules for ascertaining taxable profits are broadly in line with what can be expected, but extremely generous tax incentives, e.g. the tax holidays in the new Member States, reduce the tax base to on average three-quarters of what it otherwise would be. It is difficult to gauge the effectiveness with which the CTs and PTs on capital income are enforced. In all but two Member States, pension and investment funds are not taxed and can hence be used as conduits for not paying tax on the normal return on capital. To some extent, this may be prevented by the use of final source withholding taxes and restrictions on thin capitalization. Little inbound debt capital appears to be taxed. All Member States are reluctant to impose effective withholding taxes on the interest payable on this capital for fear of scaring away foreign investment. Most states attempt to apply appropriate transfer pricing rules based on OECD rules, but half of all Member States do not have legislation for controlled foreign corporations (CFCs). More generally, tax competition forces appear to be at work. Particularly, the 12 new EU Member States thus far seemed intent on emulating the Irish economic miracle (now evaporating) of promoting economic growth and revenue collection by adopting low nominal CT rates and generous tax incentives to stimulate domestic and foreign investment. Initially, corporate tax revenues may rise notwithstanding the low rate, because multinational companies channel their income to the low-tax states (without necessarily changing their production locations) through transfer pricing manipulation, thin capitalization, and particularly royalty payments (Hohohan and Walsh, 2002).7 However, as more Member States join the low-tax club, something like a no-win situation is likely to emerge.

2.3

BROAD FEATURES OF CT REGIMES

The actual taxation of corporate earnings and other capital income surveyed above yields a number of insights that have a bearing on the future of the CT in the EU. These insights can be summarized as follows. (a) All Member States tax capital income and labour income separately. Often capital income appears to be taxed jointly with labour income, but in practice no Member State does so. This situation could be recognized more formally by adopting a dual income tax (Cnossen, 2000) that would eliminate various ambiguities and tax capital income more effectively (Zee, 2002).

8

weighted ratio of 3.3 (Eurostat, 2008), although Ireland’s CT rate is less than one-third of the EU-15’s average rate. Cyprus, the Czech Republic, Luxembourg, Malta and the Netherlands appear to be following a similar strategy. Under a proper cash flow tax, of course, corporations are denied a deduction for interest as well as dividends paid (if not already denied), but they are allowed an immediate write-off

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

81

(b) Capital income is taxed at much lower rates than labour income, by a margin of perhaps as much as one to three. This must be attributed to the greater mobility of capital. If capital were to be taxed as highly as labour (or, more precisely, at a higher rate than the rate in other countries), the incidence of the excess would almost certainly fall on labour. Also, flat rates seem indicated to limit the countless opportunities for tax arbitrage. (c) With few exceptions, distributed profits are taxed at higher CT+PT rates than retained profits, which may distort dividend payout and investment policies. Equal treatment seems worth pursuing. This would be possible if dividend income were exempted under the dual income tax, and the PT rate on capital income set at the same level as the CT rate. (d) Domestic interest income is not taxed if it accrues to tax exempt institutions, such as pension funds. If debt can easily be substituted for equity, this implies that the normal return on capital is not taxed. In that event, the tax on capital income resembles a business cash flow tax, whose tax base is confined to inframarginal profits. Final source withholding taxes (without the possibility of a refund for tax-exempt institutions) or no deduction for interest at the level of the corporation seems the answer if the normal return is to be taxed. This would represent a move toward a comprehensive business income tax (CBIT, for short) under which profits are determined on a normal accrual basis of accounting, but which does not allow a deduction for interest at the corporate level and does not tax interest at the level of the recipient (US Department of the Treasury, 1992). Accordingly, tax-exempt institutions would be taxed implicitly. (e) The tax incentives, particularly in the new Member States, are so generous that investment costs can often be written off immediately. This would convert the CT into a cash-flow tax, because the normal return on capital is not taxed (assuming that interest is actually taxed through, say, (final) source withholding). The abolition of the tax incentives but the retention of the de facto exemption of interest also would make the CT equivalent to a cash flow tax if equity can be fully substituted by debt.8 To the extent that full substitution is not possible,

9

of the cost of business assets. As a result, the return on marginal investments, just making a viable economic return, is exempted. For arguments why cash flow taxation has economic and administrative advantages over a conventional income tax, see McLure and Zodrow (1996). The allowance-for-equity system was conceived by Boadway and Bruce (1986) and given hands and feet by the Institute for Fiscal Studies (1991). Until recently, a modified system

82

TAX REFORM IN THE 21ST CENTURY

an argument can be made in favour of an allowance for corporate equity (ACE).9 (f) Interest on inbound capital generally is not taxed for fear that debtfinanced investment costs will increase and foreign investment decline. Tax coordination is required if this interest is to be taxed. The third country issue remains, but capital is less mobile in the EU as a whole than with respect to (small) individual Member States. Further coordination could be pursued with the US and Japan. It is difficult to choose between these often conflicting directions for change, but – after allowing for the partiality that may be in the eye of the beholder – the common denominator seems to be that the body politic in most Member States appears to want to tax capital income at positive rates, if some way can be found to temper real or perceived tax competition. The choice in favour of income instead of consumption as the primary tax base is supported by economic arguments. As Salanié (2003) and others point out, it may be optimal to tax capital income if bequests are not properly taxed (see also Conesa, Kitao, and Krueger (2009)). Furthermore, another common strand in the tax literature and in actual practice seems to be that capital income should be taxed separately from labour income and at moderate, flat rates.

3

Bottom-up Approach

3.1 AGENDA FOR REFORM In view of these findings, this paper proposes an agenda for capital income tax coordination (and perhaps eventually tax harmonization), which comprises five sequential steps. (a) the introduction of dual income taxes by all Member States under which capital income would be taxed once at a single rate (different for each Member State) to mitigate the distorting effects of the current differential-rate CT+PT systems on corporate financial and investment policies; (b) the introduction of interest withholding taxes by the Member States at the CT rate (or, alternatively, the treatment of interest on par with

10

was in use in Croatia, where it was called the interest-adjusted income tax (IAIT). For a favourable discussion of the system and of the criticisms levelled against it, see Keen and King (2002). At one time, a kind of ACE system was also applied in Austria and Italy (OECD, 2007). Whatever the merits of cash-flow taxation or allowance-for-equity systems, it should be noted that taxes on economic rents would still require tax policy coordination in the EU if location decisions are not to be distorted. For a review and evaluation of the economic and technical aspects of the dual income tax on which this section draws, see Cnossen (2000). For an update on developments in Norway, see

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

83

dividends by disallowing their deduction from profits) to effectively tax the normal return on capital and mitigate incentives for thin capitalization; and (c) the close approximation of the CT rates throughout the EU to eliminate incentives for transfer pricing manipulation and thin capitalization. Following these steps, a fresh review should be made of: (d) the introduction of EU-wide common base taxation with formula apportionment and, subsequently, (e) the adoption of a European CT if and when the EU is given the power to tax. These steps are elaborated below.

3.2

DUAL INCOME TAX10

The dual income tax is a pragmatic approach to the uniform taxation of capital income, which, in the early 1990s, was successfully introduced in the Nordic countries, especially Norway, Finland and Sweden. In adopting the dual income tax, these countries argued that, in (small) open economies, any source-based tax on capital income in excess of the real world rate of interest raises the pre-tax return by the full amount of the tax, so that the after-tax return continues to equate to the exogenously given real world rate of interest. Accordingly, caution in setting the CT rate was advisable. Furthermore, capital market innovation in conjunction with tax arbitrage implied that it would not be possible to tax capital income effectively at progressive rates. Since, for revenue and distributional reasons, these countries were not prepared to lower the top PT rate to the level of the lower CT rate, they decided to tax capital income on a schedular basis. In Norway, Finland and Sweden, all income is separated into either capital income or labour income. Capital income includes business profits (representing the return on equity), dividends, capital gains, interest, rents and rental values. Labour income consists of wages and salaries (including the value of labour services performed by the owner in his or her business), fringe benefits, pension income and social security benefits. Royalties are taxed as labour income or as capital income (if know-how is acquired or capitalized).

11

Christiansen (2004) and for arguments favouring a dual income tax in Germany, see Spengel and Wiegard (2004). For a detailed description and evaluation of the (old) profit-splitting scheme, see Cnossen (2000). This scheme avoids most of the deferral and lock-in effects of the tax that various

84

TAX REFORM IN THE 21ST CENTURY

Table 2. Dual Income Taxes in Norway, Finland and Sweden, 2009 Particulars

Norway

Finland

Sweden

Year of introduction

1992

1993

1991

Corporate profits

28

26

28

Other capital income

28

28

30

Labour incomeb

a. Tax rates (%)a

28–47.8

29.5–52.5

28.89–59.09

b. Costs of earning income

Deductible at basic rate

Deductible at basic rate

Deductible at basic rate

c. Basic allowance for capital income

Yes

No

No

d. Offset of capital income against labour income

In first bracket

Through tax credit at basic rate

Through tax credit at basic rate

e. CT-PT integration method

Shielding method

Basic exemption

No integration

28% but exemption for risk-free return calculated as cost base of shares times fixed interest rate (after-tax return on 3-month government paper)

Quoted shares

30%

28% on 70% of dividend income

Part of dividend income from closely-held companies is taxed as labour income

28% on capital gains minus unused risk-free amounts

28% on capital gains

30% on capital gains

15

15

15

f. PT on corporate profits (%) Distributed profits

Retained profits

Non-quoted shares Basic exemption of up to €90,000 + 28% on 70% of excess

g. Withholding taxes (%)c Dividends Portfolio

0; 10; 15

0; 10; 15

0; 10; 15

Interest

Direct investment

–

–

–

Royalties

–

0

–

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

Particulars

Norway

Finland

85

Sweden

h. Unincorporated businesses Partnerships

Proprietorships

a b c

28% CT + 28% PT on 72% of distributed profits minus risk-free amount

Capital income fixed at 20% of previous year’s net equity; remaining profits considered labour income

30% on profits deemed to be interest on equity

28%–51% on profits in excess of risk-free amount

Same as partnerships

Same as partnerships

28% on remaining retained profits + labour income tax upon distribution with credit for 28%

Including local taxes, if levied. Including non-deductible social security contributions which increase marginal tax rates. Treaty countries.

Source: See Table 1.

The main features of the three dual income taxes, summarized in Table 2, are the following. (a) Tax rates. Basically, all capital income is taxed at the proportional CT rate, while labour income is subject to additional, progressive PT rates. To minimize tax arbitrage, the tax rate on labour income applicable to the first income bracket is set at (approximately) the same level as the proportional CT rate. (b) Costs of earning income. Generally, costs of earning income are deductible only against income subject to the basic rate. This implies that mixed expenses, containing an element of personal consumption, have the same tax value for high- and low-income groups. This does not apply to partnerships and proprietorships, however, in which case the mixed expenses are netted out against profits before these profits are split into a capital and a labour income component. Accordingly, this feature discriminates against costs incurred in earning only labour income. (c) Basic allowance for capital income. Finland and Sweden tax capital and labour income entirely separately. Alternatively, in Norway, the two forms of income are taxed jointly at the CT rate, while net labour

86

TAX REFORM IN THE 21ST CENTURY

income is subsequently taxed at additional, progressive PT rates. The separate taxation of capital income (which accrues mainly to higherincome groups) without permitting a basic allowance for capital income makes it possible to impose flat final source withholding taxes, if desired, as is done in Finland. On the other hand, joint taxation of capital and labour income in the first income bracket, as in Norway, enables the application of basic allowances to either kind of income. (d) Offset of capital income against labour income. Joint taxation, as in Norway, permits the offset of negative capital income against positive labour income. But the same effect is achieved in Finland and Sweden by permitting a tax credit for capital income losses (calculated at the basic rate) against the tax on labour income. (e) CT-PT integration method. In Norway, the double taxation of distributed profits at the corporate level and the shareholder level had been avoided through a full imputation system. Alternatively but equivalently, double taxation could be avoided by exempting dividend income at the shareholder level, as Finland did. Under either approach, compensatory taxes guaranteed that no dividends were paid out of exempt profits without having borne the CT. In recent years, these integration systems were changed. In Norway, it was argued that there was no economic reason to mitigate the double tax on above-normal returns on capital. Accordingly, it introduced the ‘shielding method’, under which only the normal return on capital, called the risk-free return, is exempt from the additional PT. On similar grounds, Finland permits a basic exemption for dividend income at shareholder level. Sweden, on the other hand, fully taxes dividend income at the capital income tax rate. (f) PT on corporate profits. In Norway, the risk-free return for distributed profits is calculated by applying a fixed interest rate (equivalent to the after-tax return on three-month government paper) to the cost base of shares. In Finland, the basic exemption for dividend income – presumably representing the normal return on capital which is only subject to the CT – is 30 per cent of dividend income (in addition, a basic exemption is provided for dividend income from non-quoted shares). By contrast, Sweden taxes profits distributions at 30 per cent, but taxes part of the dividend income of closely-held companies as labour income. The double taxation of retained profits at the corporate level in conjunction with the taxation of realized capital gains at the shareholder level had been avoided in Norway by permitting shareholders to write up the basis of their shares by the retained profits net of the CT. Similarly,

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

87

the basis was written down if the corporation incurred losses or profits were distributed out of previously accumulated earnings. Appropriate adjustments were also made if capital was paid in or paid out. The first in/first out principle applied if part of the same shareholding was sold. The method dealt both with the danger of excessive distributions of retained profits and with the unwarranted exemption of realized gains at the shareholder level due to unrealized gains at the corporate level. In Finland, the double tax on retained profits was mitigated by exempting 30 per cent of capital gains from PT. Currently, Norway and Finland have moved closer to the Swedish treatment of capital gains by taxing them in full at the capital income tax rate, although Norway permits the deduction of any unused riskfree amounts. To some extent, full taxation is justified because the realization of capital gains can be deferred which reduces the effective rate at which the gains are taxed. (g) Withholding taxes. The single taxation of capital income can be ensured through withholding or source taxes at the corporate level or at the level of other entities paying interest, royalties or other capital income. In principle, withholding or source rates should be set at the level of the CT rate. Consequently, these rates could represent the final tax liability if capital income is taxed separately from labour income and no basic allowance applies. This is the case in Finland and Sweden with respect to interest income. No country, however, imposes a withholding tax on interest or royalties paid to non-residents in treaty countries. Withholding taxes are imposed only on dividends paid to non-resident (portfolio) shareholders. (h) Unincorporated businesses.11 In Finland and Norway, the taxable profits of partnerships and proprietorships (as well as closely-held corporations), conventionally computed, had been split into a capital income component and a labour income component, and these were taxed on a current basis. The capital income component was calculated by applying a presumptive return (the sum of the nominal interest rate plus an entrepreneurial risk premium) to the value of the gross assets of the business (Norway) or to equity (Finland). Residual profits were considered as labour income. Norway has changed this arrangement

12

EU Member States impose on capital gains on substantial shareholdings. Also, the profitsplitting rules of the dual income tax seem easier to administer than some of the tortuous and arbitrary provisions for preventing the undertaxation of the self-employed currently on the statute books in countries without a dual income tax. Slemrod (1995) states that ‘although it is not desirable to tax capital income on a source basis [because source-based taxes are distortionary], it is not administratively feasible to tax

88

TAX REFORM IN THE 21ST CENTURY

to align it with the risk-free return concept introduced for profit distributions.

3.3

INTEREST WITHHOLDING TAXES

The goal of ensuring single taxation under the current dual income taxes is mostly honoured in the breach with respect to interest (and royalty) payments to exempt entities, such as pension funds and foreign debt holders (or suppliers of knowhow). This hole in the capital income tax bucket can only be plugged by imposing a withholding tax at the CT rate on all interest – in effect, treating interest on a par with dividend income, which is taxed only at the corporate level. Arrangements could then be made under which the tax withheld at the business level would be creditable in the residence Member States (hence, capital income could be taxed at different rates by these Member States). Alternatively, the tax withheld would not be creditable but would constitute the final liability in the source state (which would require approximation of tax rates if investment location decisions are not to be distorted).12 Final, sourcebased, withholding taxes on interest would make the dual income tax equivalent to a comprehensive business income tax. This tax, proposed by the US Department of the Treasury (1992), proceeds from the fundamental equivalence between a CT levied at source and an equal-rate PT on corporate earnings with a full credit for the underlying CT. Accordingly, no deductions are allowed at the corporate level for dividends and interest paid to shareholders and debt holders, but these income items are not taxed at the level of the recipients, be they individuals, corporations, exempt entities or non-residents. This makes the debt-equity distinction irrelevant and greatly reduces the distinction between retained and distributed earnings (depending on the treatment of capital gains).13 The comprehensive business income tax can be introduced while largely maintaining the present rules for determining taxable profits, including those applicable to depreciation and inventory accounting. Exempt entities and nonresidents would be treated the same as resident individuals or corporations. They would not be eligible for a refund of the tax, nor would they have to pay any additional tax in the form of a withholding tax or otherwise. Corporations receiving income that had been subjected to the comprehensive business income tax as dividends

13

14

capital on a residence basis’. Furthermore, paraphrasing Slemrod (1995a), an EU featuring (equal-rate) source-based capital income taxes would be more efficient than an EU featuring fully enforced residence-based taxes (if feasible of implementation) only because the cost of enforcement is lower for the system of source-based taxes. For a probing comparative analysis and evaluation of the dual income tax, the comprehensive business income tax, as well an allowance for corporate equity, see Sørensen (2007) and Radulescu and Stimmelmayr (2007). McLure (2004) is even more apprehensive about another proposal of the European Commission, i.e. home state taxation (HST) under which participating Member States would

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

89

or interest would also not be taxed on such income. To ensure that dividends and interest are not paid out of exempt earnings, a compensatory tax should be levied on exempt income (made available for distribution as dividends or interest). Capital gains on shares would be taxed only to the extent that they exceed the acquisition cost stepped up by the corporation’s retained profits net of the CT. The main problem of the (final) withholding tax on interest under the dual income tax and the comprehensive business income tax is that they would raise capital costs and dampen (debt-financed) investment, because the normal return on capital (i.e. interest), even if received by exempt entities and non-residents, would be implicitly taxed. Although the introduction of interest withholding taxes would seem a goal worth pursuing, gradual and concerted action is called for.

3.4 APPROXIMATION OF CT RATES The exemption of dividend income at the personal level and the taxation of interest income at source should reduce the need for concerted tax harmonization at the central EU level. The problem of thin capitalization would be solved and the schemes for CT-PT integration would become redundant. Manipulation of transfer prices, however, could still affect the allocation of the corporate tax base across the Member States. To limit this form of tax arbitrage, a minimum rate, as proposed by the Ruding Committee (1992), would have to be agreed to. Presumably, rate approximation would be easier to achieve following the introduction of dual income taxes and interest withholding taxes.

3.5

COMMON BASE TAXATION?

The dual income tax and the comprehensive business income tax would still proceed from the separate-accounting approach in determining the taxable profits of affiliated corporations in different Member States. Provisions for the removal of cross-border obstacles to economic activity and business restructuring, therefore, would still be needed. As pointed out by the European Commission (2001), a comprehensive solution to these problems, if desired, can only be achieved through common base taxation, i.e. the joint determination of the profits of firms with crossborder operations on the basis of consolidated accounts and, subsequently, the assignment of those profits to each of the Member States in which the firms carry on business on the basis of each firm’s weighted share in various economic activities, represented by such factors as its sales, payroll and property (in other words, formulary apportionment – widely practiced in the United States and Canada). The advantages of common base taxation with formula apportionment are fewer distortions, less tax arbitrage and lower compliance costs. Cross-border loss offset would occur automatically. But the path to common base taxation would not be easy, as pointed out by McLure (2004) in a cogent assessment of the European

90

TAX REFORM IN THE 21ST CENTURY

Commission proposals. According to McLure, under common base taxation, first, there would be the problem of the diversity of existing definitions of profits (see Section 2) and the lack of an objective standard against which to judge those definitions. Secondly, there is no clearly best way to define groups of firms for purposes of consolidation. Thirdly, no apportionment formula is conceptually and theoretically superior to others. And finally the administration of the common base tax would require unprecedented cooperation among participating Member States.14 Agreement would probably be easier to reach, however, following the introduction of dual income taxes, the taxation of interest accruing to foreign bondholders, and the approximation of CT/PT rates on capital income.

3.6 A EUROPEAN CT? EU-wide unitary taxation would fully reduce distortions and compliance costs only if applied by a joint administration under a common code uniformly interpreted by the European Court of Justice. Indeed, common base taxation would probably not be possible without these conditions. Accordingly, the logical conclusion of the tax coordination and tax harmonization steps outlined above would be a European CT the revenue from which would either be shared by the Member States on the basis of some formula or flow into the EU’s budget. A truly European CT, however, would require fundamental changes in the EU’s constitution moving it in the direction of a federal (tax) system.

4

Concluding Comment

This paper has argued for tax coordination in a form that relinquishes tax subsidiarity gradually but also reversibly. It has not come out in favour of unbridled tax competition, although it should be acknowledged, particularly in the EU, that tax competition can serve as a discipline on the ‘profligacy of Princes’ (Adam Smith) and present-day governments in the EU (Edwards and Keen, 1996). Neither has this paper advocated the exemption of the normal return on capital by confining the corporate tax base to business cash flow or by introducing a personal consumption tax for which strong arguments can be brought to the fore. It has not taken either of these routes in the belief that economic arguments favour the taxation of the return maintain their own rules for determining taxable profits, but firms with cross-border operations would be taxed by the Member State in which their headquarters are located. Subsequently, the consolidated profits would be assigned to each of the participating Member States on the basis of formulary apportionment. According to McLure, HST has no counterpart in the real world and might impede further evolution toward a harmonized CT system. Also, substantial cooperation would be required in the choice of an apportionment formula and perhaps in the rules for consolidation and cross-border loss offsets. Under HST, moreover, competition for headquarters locations would increase.

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

91

on capital, although at a lower rate perhaps than the tax on labour income. In sum, tax coordination reconciles the requirement of fiscal efficiency with the desire to tax capital income more effectively. References Bird, R.M. (2002), ‘Why Tax Corporations?’, Bulletin for International Fiscal Documentation 56(5): 194. Boadway, R. and N. Bruce (1984), ‘General Proposition on the Design of a Neutral Business Tax’, Journal of Public Economics 24: 231. Christiansen, V. (2004), ‘Norwegian Income Tax Reforms’, CESifo Dice Report, Journal for Institutional Comparisons 2: 9. Cnossen, S. (1997), ‘The Role of the Corporation Tax in OECD Member Countries’, in J. Head and R. Krever (eds), Company Tax Systems (Sydney: Australian Tax Research Foundation), 49. 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), 180. Cnossen, S. (2005), ‘The Future of Corporate Income Taxation in the European Union’, in Capital Taxation after EU Enlargement, Proceedings of OeNB Workshops, Oesterreichische Nationalbank, January 21. Cnossen, S. (2008), ‘Coordinating Corporation Taxes in the European Union: Subsidiarity in Action’, in G. Gelauff, I. Grilo and A. Lejour (eds), Subsidiarity and Economic Reform in Europe (Berlin: Springer) 243. Cnossen, S. (2008a), ‘Richard Musgrave’s Quest for an Index of Fiscal Equality Underlying the Horizontal Equity Concept’, FinanzArchiv 64: 147. Cnossen, S. and H.-W. Sinn (eds) (2003), Public Policy and Public Finance in the New Century (Cambridge, MA: MIT Press). Conesa, J.C., S. Kitao, and D. Krueger (2009), ‘Taxing Capital? Not a Bad Idea After All!’, American Economic Review 99: 1, 25. Ederveen, S., G. Gelauff and J. Pelkmans (2006), Assessing Subsidiarity, CPB Document 133. Edwards, J.S.S. and M.J. Keen (1996), ‘Tax Competition and Leviathan’, European Economic Review 40: 600. European Commission (2001), Company Taxation in the Internal Market, Commission Staff Working Paper, SEC(2001) 1681, Brussels. Eurostat (2008), Taxation Trends in the European Union: Data for the EU Member States and Norway (Luxembourg: Office for Official Publications of the European

92

TAX REFORM IN THE 21ST CENTURY

Communities). Available at: http://ec.europa.eu/taxation_customs/taxation/gen_ info/economic_analysis/tax_structures/index_en.htm. Fuest, C. (2008), ‘The European Commission’s Proposal for a Common Consolidated Corporate Tax Base’, Oxford Review of Economic Policy 24: 4, 698. Garretsen, H. and J. Peeters (2007), ‘Capital Mobility, Agglomeration and Corporate Tax Rates: Is the Race to the Bottom for Real?’, CESifo Economic Studies 53: 263. Hohohan, P. and B. Walsh (2002), ‘Catching Up with the Leaders: the Irish Hare’, Brookings Papers on Economic Activity 1: 1. Institute for Fiscal Studies (1991), Equity for Companies: A Corporation Tax for the 1990s, IFS Commentary No. 26 by the IFS Capital Taxes Group. Keen, M. and J. King (2002), ‘The Croatian Profit Tax: An ACE in Practice’, Fiscal Studies 23: 401. McLure, C.E., Jr. (2004), ‘Corporate Tax Harmonization in the European Union: The Commission’s Proposals’, Tax Notes International, November 29, 775. McLure, C.E., Jr. (2008), ‘Harmonizing Corporate Income Taxes in the European Community: Rationale and Implications’, in J.M. Poterba (ed.), Tax Policy and the Economy 22 (Chicago: University of Chicago Press), 151. McLure, C.E., Jr. and G.R. Zodrow (1996), ‘A Hybrid Consumption-Based Direct Tax for Bolivia’, International Tax and Public Finance 3: 97. OECD (2006), Taxation of Capital Gains of Individuals: Policy Considerations and Approaches, OECD Tax Policy Studies (Paris: OECD). OECD (2007), Fundamental Reform of Corporate Income Tax (Paris: OECD). Radulescu, D.M. and M. Stimmelmayr (2007), ‘ACE versus CBIT: Which is Better for Investment and Welfare?’, CESifo Economic Studies Advanced Access, published online on May 22, available at: http://cesifo.oxfordjournals.org/cgi/ content/abstract/ifm011v1. Ruding Committee (1992), Company Taxation in the Internal Market, COM(2001)582 final, Brussels. Salanié, B. (2003), The Economics of Taxation (Cambridge, MA: MIT Press). Slemrod, J.B. (1995), ‘Comment on V. Tanzi, Taxation in an Integrating World’, in V. Tanzi, Taxation in an Integrating World (Washington, DC: Brookings Institution), 141. Slemrod, J.B. (1995a), ‘Free Trade Taxation and Protectionist Taxation’, International Tax and Public Finance 2: 471. Sørensen, P.B. (2007), ‘Can Capital Income Taxes Survive? And Should They?’, CESifo Economic Studies 53: 172.

SIJBREN CNOSSEN: TAXING CORPORATIONS IN THE EUROPEAN UNION

93

Spengel, C. and W. Wiegard (2004), ‘Dual Income Tax: A Pragmatic Tax Reform Proposal for Germany’, CESifo Dice Report, Journal for Institutional Comparisons 2: 9. US Department of the Treasury (1992), Integration of the Individual and Corporate Tax Systems: Taxing Business Income Once (Washington, DC: US Government Printing Office). Zee, H.H. (2002), ‘World Trends in Tax Policy: An Economic Perspective’, Intertax 32: 352. Zodrow, G.R. (2006), ‘Capital Mobility and Source-Based Taxation of Capital Income in Small Open Economies’, International Tax and Public Finance 13: 269.

Chapter 4

Income and Consumption Taxes in New Zealand THE POLITICAL ECONOMY OF BROAD-BASE, LOW-RATE REFORM IN A SMALL, OPEN ECONOMY

David White * The New Zealand tax system is unsatisfactory in many respects. Its principal faults are: a. it is not delivering sufficient revenue; b. many of its features are contrary to any reasonable efficiency and equity criteria. New Zealand Treasury (1984, p. 210) New Zealand’s current tax system already conforms more closely to the standard objectives of taxation than do the tax systems of many other developed countries. Thus, New Zealand’s tax system is not obviously in need of major overhaul. Still, any tax system, including New Zealand’s, has its flaws and inconsistencies, and seeking improvement is a worthwhile objective. Alan Auerbach (2001, p. 1) *

Centre for Accounting, Governance and Taxation Research, Victoria University of Wellington; Senior Fellow, Taxation Law and Policy Research Institute, Monash University; International Research Fellow, Institute for Fiscal Studies, London. I am grateful to Iris Claus, Malcolm Gammie, Keith Taylor and Eric Toder for providing me with material and to Norman Gemmell, John Head, Eric Toder and conference participants for helpful comments on parts of the draft chapter.

96

TAX REFORM IN THE 21ST CENTURY

The [New Zealand] tax system has long been regarded as one of the most efficient within the OECD. Looking forward, however, the system will face challenges, including risks to the tax base arising from increasingly mobile capital and labour. A clearer strategic direction for the tax system is needed to help maximise living standards in the long term. There are at least two broad options: adapting the system within a comprehensive income approach or moving to a dual income tax system, in which capital income is taxed at a lower rate than earned income. … In any case, weak points within the current tax bases should be re-examined, recognising the merits of a ‘broadbase, low-rate’ approach. Any actions taken in the near-term should … be consistent with the long-term direction eventually adopted. Reforms should also not put long-term fiscal sustainability at risk: a higher GST rate could help achieve this objective. OECD (2007, pp. 8–9)

1

Introduction

For small, open, developed economies, the issue may not be whether they should replace the income tax with a consumption tax, as is often discussed in the United States public finance literature.1 If they have a large tax revenue requirement of more than 30 per cent of GDP and rapidly deteriorating public finances in the global financial crisis, the required average and marginal effective tax rates on a single tax base would need to be very high (see Chart 1 for the size of total New Zealand tax collections in relation to GDP from 1965 to 2007, with some OECD comparisons). The post-World War II experience of many developed countries should not be forgotten. High marginal effective rates of tax increased the benefits from seeking tax preferences, leading to a downward spiral as tax bases became narrower and rates increased. The incentives to avoid and evade tax also increased. This put added pressure on the practical design and maintenance of tax bases. It increased economic costs through distorting the allocation of resources, as well as increasing administrative and compliance costs. Many small, open, developed economies already have two taxes named income and consumption taxes, although the ‘income’ tax may provide consumption tax treatment for significant segments of household assets. The current issue for these economies is more likely to be how to raise more revenue from these and other tax bases at net lowest cost to society and in a way that supports a fair distribution of income. Although many tax bases were broadened in the 1980s, some of them 1

In chapter 5 of this book, Alan Auerbach, who is agnostic about the real benefits of consumption tax (Auerbach, 2007, p. 88), reviews the literature of the last 50 years on the relative merits of income and consumption taxes, focusing on questions of ‘economic content’ and ones that are ‘informed by economic analysis’.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

97

have subsequently narrowed. Among other things, policy-makers must now decide whether to raise revenue by broadening their consumption or income tax bases and what the consumption and income tax revenue mix should be. In thinking about the options, one issue is whether a broad-base, low-rate consumption tax is more likely to be sustainable than a broad-base, low-rate income tax and, if so, why?2 Chart 1: Total Tax Revenues as Percentage of Gross Domestic Product (1965–2007)3 40 35 30 25 20

New Zealand

Australia

OECD America

OECD Europe

2007

2004

2001

1998

1995

1992

1989

1986

1983

1980

1977

1974

1971

1968

1965

15

OECD Total

Source: OECD

Auerbach (2008, p. 55) argues that ‘under standard economic analysis, the hybrid system [with elements of income and consumption tax, like the current United States income tax] would seem to be considerably inferior to a true combination of income and consumption taxes, since it may substantially reduce tax collections

2

3

Zodrow (2007, pp. 50–52) asks a slightly different question in the context of the United States – whether a ‘pure’ consumption tax is more likely than a ‘pure’ income tax. Zodrow is cautiously optimistic that consumption tax reform, ‘enacted in the context of fundamental tax reform, might attain a greater degree of “purity” than would be possible under an incremental reform of the current income tax’. Auerbach (2007, p. 88) is less sanguine, other than perhaps for an indirect consumption tax, which would sacrifice progressivity. He (Auerbach, 2007, pp. 83–84) argues that ‘[t]he evidence suggests that some consumption-tax reforms would be beneficial and others would be damaging, and that knowing in which category a reform falls may be impossible until well after the dust settles and the reform is implemented’. Shaviro (2008) provides a detailed but discouraging account of the extent to which the United States political process might affect the potential simplification benefits of various consumption tax reform proposals; see also commentators, Alm (2008) and Yin (2008). New Zealand Inland Revenue Department (2008, p. 10).

98

TAX REFORM IN THE 21ST CENTURY

without reducing marginal tax rates on saving commensurately’.4 By the term ‘true combination of income and consumption taxes’, Auerbach (2008, p. 54) is referring to a tax system that, for example, combines ‘a broad-based, low-rate income tax along with, say, an add-on VAT’.5 Before the retirement and other savings-tax changes of the last two years (see Table 1), the New Zealand tax system came fairly close to being a ‘true combination of income and consumption taxes’. The New Zealand income tax was relatively broad-based and low-rate. In the year to June 2008, the income tax contributed 71.6 per cent of total tax revenue. The ‘add-on’ broad-base, low-rate GST (goods and services tax, the New Zealand term for its VAT) contributed 19.7 per cent of total tax revenue (Chart 2). New Zealand’s reliance on taxes on income and profits is quite marked in comparison to other OECD countries, largely because New Zealand has no social security tax (Chart 3). The New Zealand comprehensive indirect consumption and comprehensive income tax base reforms were designed and implemented by the Fourth Labour Government from 1984 to 1988. This Government developed a broad-base, lowrate strategy for these two taxes as the most feasible, if second-best, way to rectify the serious fiscal problems it faced.

4

5

Auerbach (2008, p. 55) notes that ‘there is considerable thought that some deviations from standard economic analysis are needed to explain various aspects of observed saving behaviour. These deviations could potentially influence several aspects of the tax system design’. Twenty years ago, John Head (1989, pp. 7–26) proposed that Australia should adopt this combination of taxes for political economy reasons. He proposed ‘a moderately progressive personal income tax and a uniform flat rate sales tax’ to reconcile ‘the conflicting views and priorities of sectional interest groups on fundamental tax policy issues such as the choice between income and consumption tax base and progressive versus flat rate’. At a conference called to celebrate the twentieth anniversary of the New Zealand GST, Neil Brooks (2007) called for a mix of broad-based income and consumption taxes. Even before the current financial crisis, Henry Aaron (2008, p. 66) urged his fellow economists in the United States to turn from the less important income or consumption tax debate and to focus on analysing the ‘huge’ economic effects of the United States federal government needing to raise probably 50 per cent more federal revenue within the next 20 years ‘to pay for the politically irreducible core of publicly financed services and benefits’. He suggested that this was likely to require higher personal income tax rates, improvements to the income tax and a broad-based valueadded tax (see also Burman’s (2009) proposal that the US implement a ‘health VAT’ (2009)). Auerbach and Gale (2009) point out that ‘the future may be upon us much sooner than previously expected’ with a federal deficit ‘larger as a share of the economy than at any time since World War II’, projections based on optimistic assumptions of the deficit averaging at least $1 trillion a year for ten years after 2009, and a fiscal gap in the long run through 2082 of about 7–9 per cent of GDP.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

99

Chart 2: Composition of New Zealand Tax Revenue (Year Ended June 2008)6 Excise duties 2.8 %

Other 5.9 %

Goods and services tax 19.7 % Individual income tax 48.7 % Resident withholding tax 4.9 % Non-resident withholding tax 2.7 % Company income tax 15.3 % Source: The Treasury

In June 1984, New Zealand was a small, closed economy, with foreign-exchange controls, pervasive regulation of the economy and extensive government ownership of economic and social enterprise (Evans et al., 1996). It was also a ‘shattered’ economy (Scott, 2005, p. 60).7 Wages, prices, and interest rates, among other things, had been frozen by government for two years. The Government’s fiscal deficit was 8.4 per cent of GDP in 1984/1985 (IMF Government Financial Statistics basis). Gross official debt was 64.7 per cent of GDP, with nearly 43 per cent of it owed in foreign currencies, and net official debt was 35.3 per cent of GDP (31 March 1984). Forecast net government expenditure was 41.7 per cent of forecast GDP (1984– 1985) (Wilkinson, 1989). The tax system was unable to raise enough revenue to fund fast-growing government expenditure and official debt. Both the consumption and income taxes exemplified poor tax design, with narrow bases and high top tax rates (up to 66 per cent for the income tax and up to 60 per cent for the wholesale sales tax).

6 7

New Zealand Inland Revenue Department (2008, p. 22). New Zealand Prime Minister Muldoon had used the word ‘shattered’ to describe the economy that he had inherited in 1975. Scott expresses his disdain for Muldoon’s stewardship of the New Zealand economy by using the same word to describe the economy that the Fourth Labour Government inherited in 1984, following nine years of Muldoon leadership. Graham Scott was in charge of the New Zealand Treasury economic policy divisions in 1984 and the head of the New Zealand Treasury from 1986 to 1993.

8

Finland

Iceland

Switzerland

United States

Norway

Income and profits

Belgium

Sweden

United Kingdom Ireland

Social security

Germany

Korea

Spain *

Italy

Luxembourg

Goods and services

Mexico ^^

Netherlands ^

Property

Hungary ^^^

Portugal

Payroll

France *

Other

Slovak Republic Poland ** Greece **

Turkey

Czech Republic

Austria

Canada

Japan ***

Australia **

New New Zealand Zealand

Denmark *

New Zealand Inland Revenue Department (2008, p. 24).

Note: Countries are ranked from highest income and profits tax revenue as % of total tax revenue to lowest. * The total tax revenues have been reduced by the amount of capital transfer. The capital transfer has been allocated between tax headings in proportion to the reported tax revenue. ** Data is for 2006. *** Central government taxes only. ^ OECD estimates have been made for other taxes. ^^ Central government and social security funds only. ^^^ Cash basis. Source: OECD

80 70 60 50 40 30 20 10 0

100 90

Chart 3: Tax Revenue as Percentage of Total Tax Revenue (2007)8 100 TAX REFORM IN THE 21ST CENTURY

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

101

From July 1984, a newly-elected Fourth Labour Government quickly transformed New Zealand into an open economy. Within days of taking office, it devalued the exchange rate by 20 per cent. It immediately started deregulating the economy, with foreign-exchange controls being removed a little later, in December 1984. By 1989, extensive tax reform had radically changed the design of the New Zealand consumption and income taxes, with much broader tax bases and lower top tax rates (to 33 per cent for the income tax and a standard rate of 12.5 per cent for the GST). The broader-base, lower-rate New Zealand tax system was judged by the OECD (1989) to be one of the least distorting among member country tax systems. Among the ‘flood of policy initiatives’ in most areas of New Zealand economic policy from 1984–87, the GST and other tax-base-broadening changes were later judged by Treasury Secretary Scott (2005, p. 62) to be the first and second in a list of major economic policy initiatives ‘that worked in terms of the goals set for them at the time’. It was not until 1993–1994, however, that a fiscal surplus was recorded for the first time in 15 years, following expenditure cuts and a cyclical recovery in economic activity (Wells, 1996, and Evans et al., 1996). The New Zealand shift to broader-base, lower-rate consumption and income taxes between 1984 and 1988 excited widespread interest overseas, even though it occurred during an extraordinary decade of world tax reform. John Kay (1990, p. 18) considered that the ‘New Zealand government has implemented one of the most radical reassessments of its system of taxation ever undertaken by a Western country’. Cedric Sandford (1993, p. 226) concluded that ‘by the criteria of the reformers’ own objectives and of sustainability, New Zealand clearly heads the field [of Australia, Canada, Ireland, New Zealand, United States, and the United Kingdom]’. Tax economist Eric Toder and tax lawyer Susan Himes (1992, pp. 357 and 358), both of whom had worked in US and New Zealand governmental tax bodies in the 1980s, argued: New Zealand’s experience demonstrates that bold, fundamental tax reform can be enacted and sustained. During the time that the United States was purporting to invent tax reform, New Zealand was enacting reforms that went even beyond the initial, far-reaching proposals of the US Treasury.9 … Although it is too early to assess the long-run economic benefits to New Zealand from tax reform, its experience in the 1980s is one that should be studied by anyone advocating a truly neutral tax system.

9

The authors refer to the three-volume study widely known as Treasury I, which contained comprehensive proposals to reform the US income tax and which analysed the feasibility of a US value-added tax: US Department of the Treasury (1984).

102

TAX REFORM IN THE 21ST CENTURY

After 1988, there were proposals to deal with the main outstanding issues of certain untaxed capital gains, owner-occupied housing and the tax treatment of rental income, but in each case the proposals were not proceeded with.10 From 1990 to 1999, the Fourth National Government largely retained the broad-base approach to both tax bases. From 1999 to 2008, the Fifth Labour Government increased the top marginal personal tax rate from 33 per cent to 39 per cent and narrowed the company and savings income tax bases in some respects. The Fifth National Government, which took office in November 2008, immediately broadened the income tax base by repealing, from the 2009–2010 income year, the previous Government’s R&D tax credit, in part to reduce personal income tax rates. The indirect consumption (GST) base has remained broad from 1986 to date. A major issue that this chapter examines is why New Zealand’s broad-base, low-rate indirect consumption tax reforms of 1984–1988 have been more stable and have been sustained longer than the broad-base, low-rate income tax reforms of the same period. The particular value of this country case study is that it is possible to hold constant for the implementation period of consumption and income tax reform many of the political economy factors that writers claim can assist major tax reform to achieve its objectives. As this is a one-country study, both sets of broad-base, lowrate tax reforms were then tested year-by-year in the same political, economic and social environment from 1988 to 2008. All of this enables us to focus more heavily on a limited number of variables that might help explain the different outcomes. To keep the chapter within bounds, it focuses on five specific New Zealand base reforms that were all designed and enacted from July 1984 to December 1988: the indirect consumption tax base reform (GST); the income taxation of private savings; the broad financial arrangements income tax base; company tax base-broadening and the imputation tax treatment of dividends; and, the international income tax basebroadening for residents. The main focus will be on the indirect (GST) consumption tax reforms to illustrate how many political economy factors must work together to successfully implement and sustain broad-base, low-rate reform. The chapter is organised as follows. Section 2 describes the approach to the political economy of tax policy taken in this chapter and identifies the political economy factors that were, and were not, constant for New Zealand’s comprehensive consumption and income tax reforms of 1984–1988. Section 3 outlines New Zealand’s broad-base, low-rate tax policy. To analyse the selected reforms, the chapter uses a device that Richard Musgrave (1991) himself used in paying tribute to another great public finance scholar and tax reformer, Carl Shoup: a ‘tableau 10

In 1989, the Fourth Labour Government (Caygill, 1989) issued a 383-page consultative document that proposed to tax all income from capital in a fair, efficient and possibly indexed manner, including some of the real gain on the sale of personal residences. In 2001, an expert committee appointed by the Fifth Labour Government (McLeod, 2001a) to undertake a general review of the New Zealand tax system released an issues paper that, among many other things, raised the possibility of taxing the net equity component of owner-occupied and rental houses using a risk-free return method.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

103

fiscale’ and commentary. A table summarises New Zealand’s 20-year experience of the five sets of tax reforms. The next three sections provide commentary on the table. Conclusions end the chapter.

2 The Political Economy of Tax Policy Although the political economy of tax policy is largely ‘uncharted territory on the roadmap to tax reform’ (Head, 2009), work using historical or political science frameworks has been growing recently, some of it including New Zealand tax reform (Kato, 2003; Ganghof, 2006; and Marriott, 2008). Economists, most notably Persson and Tabellini, have begun developing a more robust analytical framework for considering constitutional effects on economic policy and performance (see their theoretical work (2000), empirical analysis (2003), and later articles by each author). More attention is also being given to political economy in tax reform studies and initiatives. The United Kingdom Institute for Fiscal Studies’ Meade Report (1978) may have set a benchmark for fiscal economic analysis, but it included little political economy. By contrast, the Institute for Fiscal Studies’ Mirrlees Review (2008a) of what makes a good tax system for an open developed economy in the 21st century includes ‘the political economy of tax policy’ as one of its 13 commissioned studies. In part, this has been done to make it more likely that the proposed reforms would be politically sustainable. The debate between the authors of the Mirrlees Review study on the political economy of tax policy (Alt, Preston, and Sibieta, 2008) and the three commentators (Riddell, 2008; Tabellini, 2008; and Wales, 2008) on the issues, and on the advantages of taking a political science, historical or economic approach, is vigorous. Indeed, at the Mirrlees Review conference held at Cambridge University in April 2007, at which each of the 13 working groups presented their studies for commentary and debate, one suggestion (not adopted) was that each commissioned study should have included a political economy section. In this New Zealand country case study, it is possible to hold constant for all of the income and consumption tax base reforms of 1984–1988 many political economy factors that writers claim would assist major tax reform to achieve its objectives: • ‘Adverse macro-economic background’ (Sandford, 1993; Alm, 2008). For both sets of tax reforms, key players in both the public and private sectors worked together constructively in a framework allowing for greatly increased participation and consultation, while New Zealand faced two major economic crises: the 1984 national foreign exchange and economic crisis, when a large fiscal deficit and high net public debt gave the country with narrow-base, high-rate taxes, little choice but to expand

104

TAX REFORM IN THE 21ST CENTURY

•

•

•

•

•

the tax system’s base; and, then, a major international economic crisis, with the 1987 stock market crash (Wilkinson and Dickson, 1989); ‘Discontent with the current tax system’ (Sandford, 1993; Alm, 2008). In 1984, there was widespread public discontent among politicians, professionals, business people, officials, and economists with both the narrow and arbitrary sales and income tax bases and the high and arbitrary tax rates (Douglas, 1980; McCaw, 1982; New Zealand Treasury, 1984; Dickson, 1989; Lojkine, 1989; Stephens, 1989); The leadership of the ‘political entrepreneur’, which according to Sandford (1993, p. 229) is the ‘essence of a theory or model of successful tax reform’ (see also Conlan et al., 1990; Peters, 1991; and Shaviro, 1990). Roger Douglas was the New Zealand Minister of Finance and the ‘political entrepreneur’ for both sets of tax reforms (Douglas, 1993; Dickson, 1989); ‘The form of government and the rules for elections’ (Shaviro, 1990; Persson and Tabellini, 2003). From 1984–1988, New Zealand was a unitary state with power concentrated in central government; and, within central government, most power resting with the executive, made up of a Cabinet elected from a relatively small, unicameral Parliament (which was elected by a ‘first-past-the-post’ system of single-member constituencies), and the public service (Palmer, 1987); ‘Party government’ (Peters, 1991). From 1984–1988, a single (Labour) party held a similar absolute majority in Parliament, which produced ‘the fastest law in the West’ according to a former New Zealand Prime Minister (Palmer, 1987), with strong tax policy-making power resting with the Minister of Finance, Roger Douglas, and his associate ministers, the Prime Minister, David Lange, and Cabinet, all elected by the Labour Party caucus in Parliament (James, 1992); and, ‘Public bureaucracies’ (Kay, 1986; Peters, 1991). From 1984–1988, the same bureaucratic institutional structure existed for tax policy-making, with Treasury, mainly staffed by economists, dominating economic and tax policy-making, and the Inland Revenue Department responsible for tax policy implementation (Boston, 1989; Arnold, 1990).

As we will discuss in subsequent sections, there were also political economy factors that played differently in the implementation of the New Zealand comprehensive consumption and income tax reforms of 1984–1988. These further factors seem to be part of the explanation for the variable outcomes, including the relative difficulty in designing, implementing and sustaining comprehensive income tax base reform: • ideas (Conlan et al., 1990; Kay, 1990; Shaviro, 1990);

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

105

• underlying economic forces (Kay, 1990, 1994); • external influences, like regional countries and international agencies (Sandford, 1993); • the sequence of the reforms (Stephens, 1990 and 1993); • the packaging of the reforms (Stephens, 1990; Sandford, 1993; Douglas, 2005; Alm, 2008); • the support of the political leader (Sandford, 1993 and 2000), which, in the case of a parliamentary system of government like New Zealand, is the Prime Minister; • interest groups (Shaviro, 1990; Peters, 1991; Sandford, 2000); and, • the media (Shaviro, 1990). In this chapter, the author, who worked in the New Zealand Treasury tax policy group from 1987–1992 and 1994–2000, adopts an eclectic, multidisciplinary approach to reflect on the intricate interplay of ideas, individuals, institutions and the economic, political and legal environment for the sustainability of indirect consumption and income tax base reforms. Adopting the evocative metaphor with which Isaiah Berlin (1953) started a famous essay on the power of ideas, individuals and underlying economic forces in history, we choose to be a ‘fox’ rather than a ‘hedgehog’.11 The academic tradition requires each of us to know ‘one big thing’ (the hedgehog), and great advances in our understanding of the ideal tax structure have occurred as a result of rigorous theoretical and empirical analysis from experts within their disciplines. Just occasionally, however, it may be helpful to relax those constraints and to bring together insights from different disciplines and research approaches to aid understanding of a problem. In honouring Richard Musgrave, we have licence to range more widely and freely. He (Musgrave, 1986) never limited himself to one discipline in his research or teaching.

3

New Zealand’s Broad-Base, Low-Rate Tax Policy

In contrast to Musgrave’s (1968, 2001) emphasis on the equity objective, the key drivers of New Zealand tax reform in the period from 1984 to 1988 were revenue adequacy and efficiency (New Zealand Treasury, 1984, p. 210).

11

In the first line of his essay on Tolstoy’s view of history, Berlin (1957, pp. 7–9) quotes a fragment from the Greek poet Archilochus: ‘The fox knows many things, but the hedgehog knows one big thing’. The words may be interpreted many ways. In his essay, Berlin construed them figuratively to divide writers and thinkers into two categories of those who knew one big thing (hedgehogs) and those who knew many little things (foxes), while warning of the dangers of pressing the classification too far. A single writer or thinker can, of course, operate as, say, a fox at one level of analysis and as a hedgehog at a deeper level of analysis – which is how Berlin himself has been categorised by one writer (Crowder, 2004, pp. 1–12).

106

TAX REFORM IN THE 21ST CENTURY

The aim of the broad-base, low-rate approach to indirect consumption and income tax policy-making that developed in New Zealand from 1985 was to reduce both the opportunities and advantages for tax-favoured investment, and to reduce taxpayers’ compliance and government’s administrative costs. It was strongly influenced by the earlier downward spiral resulting from narrowing bases and high top tax rates (up to 66 per cent for the income tax and up to 60 per cent for the wholesale sales tax). New Zealand’s differential rates prior to 1984 had not been optimal tax rates. Lower tax rates on an activity had usually resulted from the strong preferences of a Cabinet Minister, at times following political lobbying. These concessions had led to reduced revenue, which led in turn to higher tax rates on other activities, thus imposing larger tax wedges between prices paid and received by individuals and firms and higher welfare losses for New Zealand. Concessions in one tax base encouraged further rent-seeking in that tax base or elsewhere in the tax, expenditure or regulatory systems. The broad-base, low-rate approach, however, has not always been applied to New Zealand tax policy-making. For economic, legal and political reasons, governments have: • used the income tax to support the vertical equity objective, primarily through the progressive personal income tax scale; • used the tax system to change the behaviour of individuals and firms, for example by imposing excises and duties on alcohol, tobacco, and petrol; • considered that the administrative and compliance costs involved in taxing the good, service or activity were too high; • considered that the economic costs involved in taxing the good, service or activity were too high, for example, because it may lead highly mobile capital or skilled labour to go elsewhere (an important consideration for a small, open economy with a significant need to import capital, technology and skilled labour); • accepted that income tax double tax treaties, whether in their general structure that is not negotiable, or in the provisions of particular treaties that might be negotiable, constrained policy options; or • considered that it was politically unacceptable to tax the good, service or activity, most notably not taxing all capital gains, or owner-occupied housing. Additional revenue from broadening the income tax base and introducing the new GST was used, among other things, to simplify and reduce income tax rates. To some extent, it could be argued that the retention of a progressive personal income

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

107

tax scale and the accompanying reduction of the rewards and opportunities for tax planning by the well-off resulting from the broad-base, low-rate approach assisted tax equity. Most of the burden for achieving the government’s equity objectives, however, now fell on its social welfare, government spending and income tax credit policies. Table 1 summarises the 20-year New Zealand experience of five broad-base, low-rate tax reforms. In addition to the GST, these include the income taxation of private savings, company tax base-broadening and the dividend imputation system, the taxation of financial arrangements, and the international income tax base-broadening for residents. The table also refers to several occasions on which the New Zealand Treasury seriously considered a direct consumption tax. Table 1: Implementation and Sustainability of Several New Zealand Broad-Base, Low-Rate Reforms: 1984–2008 Consumption tax Direct consumption tax?

The Meade Report had ‘an immense influence on the development of tax policy thinking in New Zealand’, making the formulation of tax policy advice ‘systematic and principles-based’ (Dickson and White, 2008) but did not persuade policy-makers to propose a direct consumption tax. Comprehensive direct consumption tax reform was considered by New Zealand Treasury officials from 1982–1985 and from 1996–1998 but was rejected, largely because of the uncertain gains and high risks involved for a small country pioneering this consumption tax option. Broadened or sustained base; or, reduced rate

Indirect consumption tax: GST base

1986–current: broad GST base. GST introduced with immediate compensation for low-income families with children and lower personal income tax rates: e.g., top rate of 66% reduced to 48%

GST rate

1986–1989: 10% GST standard rate, ‘with no reduced rates or super reduced rates, exemptions or zerorates – other than those necessary to define the appropriate base’ (Dickson and White, 2008)

Narrowed base or increased rate

1989–current: GST standard rate increased to 12.5%

108

TAX REFORM IN THE 21ST CENTURY

Income tax Broadened or sustained base; or, reduced rate

Narrowed base or increased rate

i. Debt: Financial instruments

1986–current: broad base for broad range of debt instruments, derivatives and other instruments. In most cases, accrue expected gains and losses, realise and accrete unexpected amounts

ii. Equity: Company tax base

1984–1989: base broadening with most notable exception being exclusion of many domestic capital gains. 2009–2010: R&D tax credits repealed by Fifth National Government

1992: narrowed for forestry expenditure. 2008–2009: narrowed with R&D tax credits for 1 year only

Company tax rate

Mostly 33% (range 28–48%). 2008: rate reduced to 30%

Mostly 33% (range 28%–48%)

Dividends

1988–current: full dividend imputation for NZ company tax. 2003: imputation extended to Australian-resident companies operating in NZ and paying NZ tax

Company and top personal tax rates (and trust tax rates)

1986–1988 & 1989–1999: company and top marginal personal tax rates aligned

1988–1989 & 2000–current: all three tax rates not aligned. 2000–2008: Personal tax top rate raised to 39% ending alignment with 33% company tax rate, which applied from 1989–1999. Reducing difference between top marginal personal and company rate: 8% in 2009 & 7% in 2010. Note 38% (2009) and 37% (2010) top marginal personal tax rate is also higher than trust tax rate of 33% and portfolio investment entity rate of 30%

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

109

iii. Taxation of residents’ international income:

Personal residence rules

1989–2006: broader rules, so easier to become a resident and harder to cease to be a resident

Company residence rules

1988–current: broader rules, so easier to be resident

International trust regime

1987–current: broad base with resident-settlor and foreign trust regimes

Foreign portfolio investment (foreign investment fund (FIF) regime)

1988–1993: first FIF regime for tax haven (1988) and nontax haven investment (1989) deferred each year to 1993. 1993–2007: second FIF regime, which mostly applied accruals capital gains tax (cf. local shares mostly free of capital gains tax), was largely ineffectual because of generous country exemptions and possible low levels of compliance 2007 onwards: third FIF regime broadened base by substantially reducing country exemptions. Regime is now a presumptive tax for most taxpayers (cf. local shares still largely free of capital gains tax). Likely compliance levels open to question

Foreign direct Investment (controlled foreign companies (CFC) regime)

2006: foreign-sourced income exemption for transitional residents

All NZ CFC regimes have had a broad base, taxing active and passive income. 1988: taxhaven CFC transitional regime. 1988–1993: non-tax haven CFC regime deferred several times. From 1993, full CFC regime applies but around 70% of NZ’s FDI is exempt with generous country exemptions

110

TAX REFORM IN THE 21ST CENTURY

iii. Taxation of residents’ international income (continued):

2009: Bill before Parliament proposes to exempt active income, as in most other countries’ CFC regimes, but to reduce country exemptions to Australia only. Government expects less revenue than from current regime, which has applied since 1993 iv. Retirement and other private savings: Contributions

1987: contributions fully taxed

Member (2007) and employer (2008) contribution tax credits in superannuation (pension) schemes widely used by New Zealanders (KiwiSaver schemes)

April 2009: Fifth National Government reduced minimum employee contributions, capped both the compulsory employer contributions and employer contribution tax exemption, and discontinued tax credit paid to employers whose staff are enrolled in KiwiSaver schemes Superannuation (pension) scheme investment earnings

1988: superannuation scheme investment income fully taxed at company tax rate, which is now 30% (2008). Strong incentives to over-invest in non-capital gains taxed assets, especially owneroccupied housing and rental property. Housing is more than 70% of total NZ household savings (cf. OECD average of less than 50% (McLeod, 2001b, p. 29))

Withdrawals

1990: pension and lump sum benefits tax-free

2007: new tax rules for NZ resident individual investing in ‘portfolio investment entities’ (e.g., superannuation funds, unit trusts and certain listed entities) to ensure both final top marginal rate from 2008 on of 30% (cf. 38% top marginal personal rate for other income in 2009), and tax exemption for gains on the sale of shares in NZ companies and certain Australian companies

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

4

111

Indirect VAT/GST Consumption Tax Base Reform

It was taxing personal consumption that Treasury tax policy officials considered offered greatest scope for raising tax revenue at lowest efficiency cost in mid-1984. Only about one-third of personal consumption was subject to wholesale sales tax at the time, without taking into account services or the retail value added (New Zealand Treasury, 1984, pp. 223–224). In the public briefing document to the incoming Fourth Labour Government, Treasury officials discussed and dismissed the personal direct expenditure tax option in three paragraphs (1984, p. 218) and a cash flow business tax in one paragraph (1984, p. 221). From 1982–1985, however, New Zealand Treasury tax policy officials had been fully analysing both direct and indirect consumption tax options. Ultimately, officials decided not to recommend to government the comprehensive direct consumption tax reform option of taxing net cash flows. Their reasoning was largely pragmatic. The New Zealand tax system’s failings were pressing. A small country pioneering this new tax base would be likely to face high implementation costs, large tax design and political marketing problems, and uncertain revenue and tax administration risks, particularly in the transition and at the international border. From 1996–1998, another New Zealand Treasury project team considered converting the New Zealand income tax into a direct consumption tax by making limited modifications to it (for example, allowing tax deductibility for business expenditure and making dividend and interest payments non-taxable for the recipient). Treasury managers decided to stop this project for similar reasons to those that caused the termination of the 1982–1985 project. ‘The necessary transitional rules have proven to be more complicated than originally envisaged. Further, there are no immediate opportunities for large tax cuts … the main possible motivators, being a desire for greater efficiency and, possibly, greater national savings, seem too theoretical to be very [politically] persuasive’, given ‘the mental adjustment’ that a direct consumption tax required of the public (Katz, 1999, p. 16).

4.1 THE GST The 1984 Treasury briefing document took the indirect tax options much more seriously (New Zealand Treasury, 1984, pp. 222–227) and it was an indirect consumption tax that was implemented in 1986. The twin policy pillars of the GST are its comprehensive coverage and its single domestic rate. ‘There are no reduced or super-reduced rates, exemptions or zero-rates – other than those necessary to define the appropriate base of the tax’ (Dickson and White, 2008, p. 2). Over more than 20 years, there has also been relatively little pressure to exclude food and essentials from GST. Where it has surfaced, as it did with recent high food prices, it was limited to several small political parties and social groups.

112

TAX REFORM IN THE 21ST CENTURY

The distributional effects of the GST are discussed by Creedy (1997, 2006) and Stephens (2007), who reports the different results from specific incidence, differential incidence and balanced-budget incidence studies. Adjustments to the state-funded universal superannuation scheme have provided protection to pensioners, at least with respect to these pension payments. Social assistance to those low-paid workers and beneficiaries who were identified to have above-average impacts was provided through income supplementation, but there were periods in the 1990s and early 2000s when the assistance did not keep up with costs. A recent study of the theoretical, empirical and practical issues involved in VAT/ GST tax design for open, developed economies in the 21st century has come to the same conclusion as the designers of the New Zealand GST model. The authors of the Mirrlees Review study on VAT (Crawford, Keen and Smith, 2008) have concluded that there are too many empirical and practical obstacles to imposing different VAT rates on different commodities to assist the less well-off, or to help reduce the disincentive to work created by the income tax, by imposing higher VAT rates on leisure-related goods and services, for example. As one illustration of what is possible, the authors of the Mirrlees Review VAT study calculate that ending the current United Kingdom zero and reduced rates, except for housing and exports, while increasing ‘means-tested benefit and tax credit rates by 15% … would leave the poorest three-tenths of the population better off on average while raising £11 billion to cut other taxes or to spend in other ways’ (Institute for Fiscal Studies, 2008b). What has been the result in New Zealand? It is easy to see why GST is a New Zealand tax policy-makers’ dream, both in revenue and cost terms. It has been ‘reliable and unexciting’ so far (White, 2007).12 The New Zealand GST’s comprehensive coverage allows a large amount of revenue to be collected at a relatively low rate. Its ‘C-efficiency’ (consumption efficiency) – the ratio of GST revenue to the product of aggregate consumption and the standard rate – is nearly 100 per cent, 43.5 per cent above the OECD average (see Chart 4). With a relatively low standard rate of tax compared to other OECD countries, it collects a relatively high effective rate of tax (see Chart 5).

12

Difficulties with the tax, however, should not be discounted. In June 2008, New Zealand tax officials (2008, p. 14) issued a 56-page discussion document on major areas of GST government revenue and business cost risk, exploring legislative solutions to current problems. With the increasing number of persons registered for the GST (over 660,000 in April 2008 from 500,000 in 2000) and the large number of GST returns processed each year (more than 3 million in the 2008 financial year), the officials’ issues paper conceded that ‘the effectiveness of current audit strategies can be tested’.

13

9 3 .5 %

Denmark

Finland

Iceland

Canada

Ireland

Korea

Japan

Switzerland

Luxembourg

Australia

New Zealand Inland Revenue Department (2008, p. 25).

* N o V A T /G S T . S o u rc e : O E C D a n d In la n d R e v e n u e c a lc u la tio n s

0 .0

1 0 .0

2 0 .0

3 0 .0

4 0 .0

5 0 .0

6 0 .0

7 0 .0

8 0 .0

9 0 .0

NewZealand Zealand New

1 0 0 .0

Chart 4: Consumption-efficiency Ratios of VAT/GSTs in OECD (2006)13

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND 113

United States * Mexico

Turkey

Italy

United Kingdom

Greece

Poland France

Portugal

Belgium Sweden

Hungary

Czech Republic

Germany

Slovak Republic

Netherlands

Austria

Spain

Norway

14

Hungary

Austria

Ireland

Norway

VAT/GST rate

Mexico

Korea Turkey Luxembourg

Italy

Germany

Spain

Czech Republic

United Kingdom

Greece

France

Netherlands Belgium

Slovak Republic

VAT/GST revenue as % of gross domestic product

United States * Japan

Canada Switzerland Australia

Poland

Finland

Portugal

New NewZealand Zealand

Denmark

Sweden

Iceland

New Zealand Inland Revenue Department (2008, p. 25).

Note: Countries are ranked from highest VAT/GST revenue as % of gross domestic product to lowest. The comparisons include all levels of government. * No VAT/GST. Source: OECD

0

5

10

15

20

25

30

Chart 5: VAT/GST Tax Rates and Revenues in OECD (2006)14

114 TAX REFORM IN THE 21ST CENTURY

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

115

New Zealand GST tax returns are so simple that many businesses complete their own returns. Large New Zealand professional firms have a very small number of specialist GST partners. Neither Inland Revenue nor the Treasury need to deploy many policy resources to monitor it. New Zealand academics give little priority to it in their teaching or research. It receives little attention in the media (for detail and empirical research, see White (2007)). The more complicated issue of the compliance cost benefits of a comprehensive, single-rate GST have been considered in Dickson and White (2008, pp. 7–10). Reflecting on the 20-year GST timeline, it is clear that the broad-base and single, low-rate strategy could easily have been derailed in the design and implementation stage, and perhaps also in its early life. This raises two questions. The first is how did political economy factors enable the Fourth Labour Government to design and implement a broad-base, low-rate indirect consumption tax in 1984–1986? The second is to what extent can the sustainability of the broad-base and single, low-rate strategy over this period be attributed to its successful political implementation?

4.2

KEYS TO SUCCESSFUL IMPLEMENTATION

Douglas (2007, p. 3) himself traced the successful implementation of the GST to five key elements involving people and process: ‘political will, the right people, the way in which the proposal was packaged, an effective consultation process, and an effective communication process’. In summarising his findings about the keys to successful tax reform from a comparative study of six countries, Cedric Sandford (1993, pp. 227–229) pointed to the three essential elements as being people, process and the environment. The most important of these, he argued, was ‘political entrepreneurship’,15 which, in the Westminster parliamentary system, needs to be shown by the Finance Minister. ‘His or her conviction, energy, astuteness and toughness, more than anything else’ was necessary. Using these criteria, Sandford (1993, p. 203) judged Roger Douglas, the New Zealand Minister of Finance who introduced the New Zealand GST, to be ‘the most successful tax reformer of the 1980s’. There is evidence to support Sandford’s assessment: • Douglas had been thinking about tax reform long before becoming Minister of Finance, had read the Canadian Carter Commission Report, and had written a short book that included his ideas on tax reform, including shifting the tax base towards consumption by introducing a 15

Guy Peters (1991, pp. 287–288 and 299), a United States political scientist, also placed considerable weight on ‘political entrepreneurship’ in explaining the centrality of tax reform in the politics of industrialised democracies during the late 1980s and early 1990s. He gave considerable prominence to the example of Roger Douglas’ role in New Zealand tax reform, along with examples from North America and Europe, in his account.

116

TAX REFORM IN THE 21ST CENTURY

•

•

•

•

•

•

•

16

10 per cent retail sales tax on all goods, but exempting housing and food (Douglas, 1980, pp. 66–69); Douglas preferred ‘big-league change’ and, according to the Labour Party MP who succeeded him as Minister of Finance, Douglas was ‘easily the most imaginative politician I was ever privileged to work with, though we disagreed on a number of occasions about what was practical as well as necessary’ (Caygill, 1999, p. 53); Facing a fiscal deficit of 8.4 per cent of GDP and net public debt of 35.3 per cent of GDP on first becoming Minister of Finance, Douglas asked officials to develop a ‘quick and dirty’ 12.5 per cent retail sales tax to be implemented immediately; but, on being persuaded of the merits of a VAT (less susceptible to evasion and closer to commercial business practice), he supported the design and implementation of a best-practice model GST (Dickson, 2007, pp. 47–50), even though this involved a two-year wait for the revenue;16 Douglas himself recommended the GST to Cabinet, as senior New Zealand Treasury officials, wanting to maintain the impression of impartiality, would not make a recommendation to him – unlike 1982, when they did recommend a VAT to the previous Minister of Finance (Dickson, 2007, pp. 50–51); In developing and implementing the GST, Douglas initiated, or agreed to, a ‘revolution’ in New Zealand tax policy-making and consultation processes that substantially opened up those processes for the GST and for all future tax policy-making (see the subsequent discussion in this section); Douglas was prepared to appoint the best people for the job, whatever political party they supported (for example, he appointed a former parliamentary candidate for the opposition National Party to chair the GST Advisory Panel (Douglas, 2007, p. 7)); Although one of Douglas’ (2005, p. 58) reform dicta was that ‘speed is an essential part of any reform programme’, he agreed, reluctantly, to a six-month delay in GST implementation to make changes and to allow a second round of submissions, this time before a parliamentary select committee; and In line with another of Douglas’ (2005, p. 58) reform dicta that ‘packaging reforms into large bundles … [is] in fact the key to being able to sell changes politically’, Douglas ensured that there was immediate compensation for low and middle-income families with children on

For some of the risks involved in taking an incremental approach to VAT implementation, see Bird and Gendron (2007), especially pp. 3–5, 96–97.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

117

introduction of the GST. Personal income tax rates were also reduced at the same time that GST was implemented. The result was that the GST approval rating went from a high of 30–35 per cent before implementation to 65 per cent two weeks after implementation (Douglas, 2007, p. 6). But it would be a mistake just to focus on the political entrepreneur, Roger Douglas, important as he was. He (2005, pp. 56–59) does not do so in his reflections on the record of the Fourth Labour Government. He had the good fortune to follow a Minister of Finance, who by the time he was voted out, had clearly demonstrated the folly of relying on narrow-base, high-rate taxes or narrow-tax-base, high-borrowing strategies for funding government. Douglas also had the full support of Prime Minister Lange and Revenue Minister de Cleene, who both greatly helped market the new GST tax, as well as very supportive associate finance ministers and key backbench MPs (Douglas, 2005, p. 57). So to the political entrepreneur it is necessary to add widespread dissatisfaction with the existing tax system, and the support of the Prime Minister, key Cabinet Ministers and members of his or her political party. And Douglas had even more good fortune. New Zealand bureaucrats had had a dry run at the indirect tax reform issues and arguments several years beforehand (Dickson and White, 2008). Under the influence of the Meade Report (1978), the formulation of New Zealand Treasury tax policy advice had become systematic and principles-based. ‘Meade could give New Zealand policymakers no direct guidance on how to structure [government policy to rely more on indirect taxation], nor on the pressing question of the day, which is better, VAT or retail sales tax? Meade, however, inspired an investigative approach to these problems that eventually resolved these questions’ (Dickson and White, 2008, pp. 4–5). In 1981–82, New Zealand officials had concluded that the VAT model was superior and had analysed the most difficult issues, including food, clothing and necessities. They had developed a strategy that carried risks but also created the opportunity to enact a broad-base tax – win on the most difficult group of issues, and the public would be less likely to support others seeking an exclusion from the tax (Dickson and White, 2008). The head of the New Zealand Treasury indirect tax team from 1984–1986 has acknowledged that ‘there is no denying the traditional boundaries observed by public servants in the policy arena were crossed from timeto-time in the interest of gaining a particular desired state of affairs’ (Dickson, 2007, p. 51). The bureaucracy clearly played a key role. Ideas also played a part. In the last 50 years, the VAT idea has swept the world. The acceptance of the value-added tax model in around 130 countries has been described as ‘probably the most important tax development of the latter twentiethcentury, and certainly the most breathtaking’ (Ebrill et al., 2001, p. xi). National, regional and international bureaucracies also influenced GST design in New Zealand: the political and bureaucratic experience with VAT in European

118

TAX REFORM IN THE 21ST CENTURY

countries, especially the United Kingdom; at the regional level, the European Union experience; and, at the international level, the analysis, country experience and practical advice given New Zealand by the OECD and the IMF (particularly, from Alan Tait). To political entrepreneurship and the other factors discussed above, it is necessary to add the power of ideas, and the range of responsive and activist roles played by national, regional and international economic bureaucracies in VAT acceptance and reform. Finally, it is necessary to add the participation of the private sector, including tax professionals, business people, academics, and others, and the open consultation process. The GST marked a turning point in New Zealand tax policy-making and implementation. In the past, tax policy had been made by politicians and bureaucrats in secret and had been announced on Budget Night, often with immediate legislation to follow. A small group of bureaucrats still had a vital role (Dickson, 2007), but the whole tax policy-making and implementation process was broken open in four major ways. The GST was the first time that a New Zealand government had appointed a consultative committee independent of government and the bureaucracy to review submissions on government tax proposals. The committee’s report was published along with the minister’s decision on each recommendation (Green, 2007). The GST was also the first time that a New Zealand government had appointed a group, independent of government and the public service, to coordinate a public information programme on the whole package of reforms and to develop an education programme for taxpayers (Todd, 2007). Thirdly, the GST was one of the first times that tax legislation had been submitted by a New Zealand government to a parliamentary select committee to receive public submissions on the legislation.17 Finally, the GST was the first time that a research institution outside government (the Victoria University of Wellington Institute of Policy Studies) had brought together academics, officials, tax professionals, business people and others to fund, provide researchers and oversee a programme of published research on the policy design, legislative detail, and the incidence of the tax before it was implemented (Stephens, 2007).

17

The first government tax Bill to be referred to a parliamentary select committee was the Income Tax Amendment Bill (No 2) 1985 and that was done by the Fourth Labour Government in March 1985 (Jackson, 1987, p. 127). The GST Bill was introduced into Parliament and referred to a parliamentary committee for a second round of submissions in August 1985. The Minister of Finance, Roger Douglas (Douglas and Callen, 1987, p. 218), has acknowledged that this occurred, ‘despite [his] own reservations’.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

4.3

119

KEYS TO SUCCESSFUL SUSTAINABILITY

The second question that the 20-year New Zealand GST experience raises is more difficult to answer. To what extent can the sustainability of the broadbase and single, low-rate strategy over this period be attributed to its successful political implementation? A good case can be made for it being a major factor in the sustainability of the broad-base, low-rate GST strategy, but as we shall see, more fundamental reasons lie at the heart of the GST sustainability story. We turn first to the contribution from the tax’s political implementation. The successful implementation and acceptance of the tax in 1986 meant that the main opposition (National) party persisted only for a short period with its proposal to replace the GST with a narrow-base retail sales tax. After the Fourth National Government was elected in 1990, its Minister of Finance described the GST as a ‘model of tax reform’ (Todd, 2007). At the time of the first and only general review of the GST to date, two other National Government finance ministers described the GST as an important part of New Zealand’s tax system, which was acknowledged internationally ‘as a welldesigned indirect tax, influencing the design of indirect taxes put in place in other countries’ (Birch and English, 1999). The New Zealand GST had become more than just a tax with bipartisan, domestic political support. It had become a policy and tax that was considered ‘international best practice’ and that was marketed this way by international agencies, like the IMF and private sector groups and individuals around the world. New Zealand tax policy-makers came to consider that they had an international duty to protect this ‘light upon the hill’. Ministers seeking to achieve an economic, social or political policy objective were encouraged to use expenditures, regulation, and, as a last resort, the income tax system, but never the GST. Unbelievable as it might sound in talking about a tax, the GST had become a matter of pride for many New Zealanders. This was true for politicians, bureaucrats, business people and many members of the public, especially those who travelled to countries with narrow-base VAT/GSTs, and who experienced or read about the VAT ‘exception politicking’, or the more than ten years of legal argument about whether the Marks & Spencer ‘teacake’ is a ‘biscuit’ or a ‘cake’, for example. Here was something that New Zealand had been able to avoid that others had not. Among what is a relatively small group of decision-makers in this small country of some 4 million people, this has been a very strong, sustaining force for the tax. A good case can be made for electoral rules systematically shaping economic policy (Persson and Tabellini, 2003), but not even the most radical change that New Zealand has made in its electoral rules in over 150 years has rocked the GST ‘boat’. Up until 1993, Members of Parliament (MPs) were elected under a single-member, simple-plurality system, as in many Westminster systems of government. From the 1996 general election, a mixed-member proportional electoral system, largely based on the German system, has applied, with marked changes to how government

120

TAX REFORM IN THE 21ST CENTURY

operates, and how policy is made and implemented (James, 2000). Since 1998, there have been coalition governments, which have relied on the support of parties outside of government for supply and confidence. The current Minister of Revenue, Hon Peter Dunne, is the leader of a small political party and its only MP in Parliament. He is a Minister outside Cabinet in the National-led Government. There are, however, three general factors that are not specific to the New Zealand GST and that strongly underpinned the GST’s sustainability. First, from his time on the Meade Committee, and possibly before, John Kay18 has developed a compelling case for the sustainability of the consumption tax base, dealing as it does with ‘real’ rather than ‘hypothetical’ transactions. His (Kay, 1994, p. 18) principal argument is that the evolution of tax structures is not so much driven by ‘political fashions but by underlying forces in the world economy’. These underlying economic forces have been reducing government and geographical barriers to trade and investment. Advances in computer technology have made measuring and monitoring daily transactional taxes feasible and understanding the tax effects of sophisticated financial instruments more difficult. Tax bases that rely on ‘assessing’ income from capital over a whole year have become increasingly problematic compared to taxes that have a real, day-to-day transactions base. In short, Kay concludes the trend in tax bases is towards taxes that rely more on real transactions, like the VAT/GST, than those that rely more on hypothetical transactions, as is more often the case for taxes on capital, corporate income and wealth. There is much truth in this argument but it too does not capture the whole picture for our purposes. A second key factor supporting the sustainability of the VAT/GST model in general is its more realistic jurisdictional scope. The GST, for example, does not seek to tax New Zealand residents on their worldwide consumption, including their daily consumption when they are abroad, outside the jurisdiction of the New Zealand revenue authorities and the New Zealand courts. The New Zealand income tax, by contrast, attempts to assess the annual income of New Zealand residents from the 190-plus countries in the world, with limited assistance from the 35 countries with which New Zealand has double tax treaties and from two other countries with which New Zealand has tax information exchange agreements. As our honorand (Musgrave, 1992, p. 364) sadly noted, ‘A destinationbased value-added tax is neutral, and even a personal expenditure tax is less open to avoidance by foreign consumption than is an income tax by foreign investment. Once more, income taxation seems to be at a disadvantage’. Thirdly, New Zealand is a small, unitary state separated by 2000 kilometres of sea from its nearest neighbour, Australia. It has no land borders with any other country. It is not a federation, where a federal VAT would need to be coordinated with many sales tax jurisdictions at state and local levels of government, as in the 18

Kay (1986, pp. 7–10); Kay (1990, pp. 22–23, 29–30, 67); Kay (1994).

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

121

United States. Opportunities for VAT/GST fraud exist (White, 2007; New Zealand Officials, 2008) but they are not nearly as great as those available in federations with sales taxes levied by different levels of government, or those available across land borders or narrow water boundaries, especially in economic unions (Crawford, Keen and Smith, 2008).

4.4

CONCLUSIONS

To sum up, the comprehensive, single-domestic-rate GST does not support the adage that an old tax is a good tax. This ‘new’ tax raises about 19–25 per cent of revenue efficiently and there is widespread public acceptance of its twin policy pillars (Dickson and White, 2008). Even if a minister of finance were to introduce a tax exception for food, clothing or other necessities, there is no guarantee that the full value of the forgone GST would be passed on to the consumer. It is little wonder that politicians in the major political parties do not want to change its comprehensive base. Successful implementation of broad-base, low-rate reform required a high level of contribution from all major players and the support of all the political economy factors identified in section 2 of this chapter. A good deal of the sustainability of the comprehensive, single-domestic-rate tax policy over more than 20 years can be attributed to the open, patient, political implementation of this feasible, ‘secondbest’ tax at the outset. The sustainability of this GST strategy for more than 20 years, however, has also had essential and on-going underpinning from the identified economic, legal, geographic and constitutional factors.

5

Selected Income Tax Base Reforms

Table 1 has sketched just some of the Fourth Labour Government’s major reform that took the New Zealand income tax system closer to the Haig-Simons comprehensive accretion base than any other OECD country tax system: an accruals regime for debt obligations, debt equivalents, and certain derivatives; an accruals regime for foreign portfolio investment but with generous country exemptions; accrual taxation for all (active and passive) income earned from foreign direct investment, whether the income was distributed or not, again with generous country exemptions; income tax treatment of private savings; a dividend imputation system that, as closely as possible, integrated the company and personal income tax systems, backed up by alignment of the top personal and company tax rates for many years. This commentary to the table briefly describes and analyses the political economy of these four income tax reforms to enable comparisons to be drawn in section 6 with the implementation and sustainability of the broad-base, single-low-rate GST reform. John Kay (1994, p. 15) has suggested that in the 1980s New Zealand and Australia were ‘perhaps unique among (OECD) reforming countries in taking

122

TAX REFORM IN THE 21ST CENTURY

the comprehensive income tax ideal relatively seriously’. As we shall see, this commitment was widely shared by the Finance Minister, officials and key private sector experts, with the notable exception for some of them of not wanting capital gains tax to lock investors into holding domestic assets at a time of huge restructuring of the New Zealand economy. The extent to which key New Zealand private sector experts were prepared to support the comprehensive income tax base concept, and the conditions upon which that support rested, is reflected in a report sent to the Minister of Finance in June 1987 by a consultative committee set up to consider accrual tax treatment of income and expenditure (the Brash Committee). This committee was made up of two economists, two tax lawyers, one banker and one tax accountant. Concerned that other deficiencies in the income tax system might negate its accruals system recommendations, the Brash Committee sought, and was given, a mandate by Finance Minister Douglas to provide a general report to him. Comprehensive Tax Reform and Possible Interim Solutions (Brash, 1987) was the title the Brash Committee chose for its report. In just 23 pages, it made general recommendations on income tax rates, comprehensive international income tax base reform, company tax, capital gains, interest deductibility, tax law drafting, and tax administration. With one dissent in relation to capital gains on land and shares, among other things, the Brash Committee (1987, pp. 2, 3, 16–17) said: Our own strong preference would be a comprehensive tax on all gains (whether income or capital) properly attributable to New Zealand residents, at an internationally competitive rate (perhaps 25–30%). We believe that most taxpayers would readily accept the trade-off between a much more comprehensive definition of ‘income’ for tax purposes, and a lower tax rate. … The Committee in fact sees a clear connection between reforms which broaden the tax base and ensure that some minimum amount of tax is paid, especially by corporate taxpayers, and a reduction in rates. …So long as the Government is able to demonstrate its ability to reduce tax rates significantly, continue with its programme of Government expenditure reductions, and reduce the deficit over time, we consider that there will be widespread support for a more comprehensive income tax regime. The ‘base-rate-expenditure’ trade-off that this group of private sector experts, and many others in business, supported is clearly expressed in this report to Finance Minister Douglas.

5.1

INCOME TAX TREATMENT FOR PRIVATE SAVINGS

By international standards, perhaps the most remarkable reform of the New Zealand income tax system over this period was the decision to extend the comprehensive

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

123

income tax principle to all forms of private saving, including occupational superannuation.19 The broad array of tax treatments in mid-1984, most of which were highly concessional to savings, were all converted into a neutral income tax treatment of TTE (taxed contributions, taxed fund earnings and exempt emerging benefits) by 1 April 1990. This decision was motivated by the desire to remove distortions and inefficiency in the financial system caused by the tax-favoured treatment of retirement saving. The removal of these superannuation tax expenditures was estimated to allow a reduction of overall income tax rates by about 2.5 cents per dollar for all taxpayers (Douglas, 1987, p. 21). In accordance with the competitive market philosophy of Finance Minister Roger Douglas, the aim of this reform, in conjunction with the dividend imputation system, was to pursue an objective of tax neutrality towards the various forms of saving and investment. People’s retirement saving needs would instead be met primarily through the universal, state-funded and (since 1998) non-means-tested20 New Zealand Superannuation scheme. Supplementary savings would have to be made privately out of after-tax income without the support of government subsidies or tax incentives. Housing, however, remained heavily tax-favoured. In most other countries, large amounts of revenue have typically been devoted to providing tax concessions for retirement savings. This has long been the prevailing practice, even though there is no clear evidence that such incentive provisions are effective in increasing net personal saving (Toder and Khitatrakun, 2006, pp. 17–25). It is more likely that the main effect of tax incentives is to change the composition rather than the volume of savings, as the various forms of saving (debt, equity shares, housing, superannuation, etc.) are very close substitutes. Tax incentives for private saving nevertheless enjoy overwhelming political support in many jurisdictions. Following the election of the Fifth Labour Government in 1999, and the change in the personal income tax rate scale, pressure began to build for the reintroduction of tax concessions for retirement savings. These pressures eventually led to the introduction of the work-based KiwiSaver scheme in 2007, related tax concessions for employer and employee contributions and other saving vehicle concessions (see Table 1). New Zealand’s mixed-member proportional electoral system contributed to a ‘slippery slide’ (St. John, 2007) of concessions begetting more concessions, presently arrested by the current National-led Government (see Table 1). On the basis of a nationwide survey (before family savings data becomes available), it appears that each dollar of KiwiSaver account balances represents 19 20

For historical context, see Preston (2008), and for analysis, see St. John (2007) and Marriott (2008). For 13 years from 1985 to April 1998, there was effectively an income test on people receiving National Superannuation through a taxation surcharge on the other income of National Superannuitants (Preston, 2008).

124

TAX REFORM IN THE 21ST CENTURY

only 9–19 cents of new saving created by reduced consumption. Further, KiwiSaver seems to lower national savings, taking into account its revenue, administrative, and compliance costs (Gibson and Le, 2008). It has been argued, with some merit, that ‘KiwiSaver and its associated tax incentives will increase future inequality in lifetime incomes and lead to diverging living standards for the elderly’ compared to New Zealand’s retirement saving policy from 1990 to 2007 of a universal, noncontributory, flat pension and no tax incentives (Gibson, Hector and Le, 2008, p. 16). There is little hard evidence that private saving or national saving in New Zealand has been unfavourably affected by the income taxation of private savings, even though national savings are low compared to most OECD countries. What is clear, however, is that the additional revenue raised in this and other ways enabled the implementation, inter alia, of the internationally-competitive company tax imputation and rate-alignment system that was unique to New Zealand and greatly reduced tax distortions and tax avoidance problems over the 12-year period of its successful operation from 1986–1988 and 1989–1999. This rate-alignment system of company and personal income tax is further discussed below.

5.2

COMPANY INCOME TAX BASE AND FULL IMPUTATION

In July 1984, the narrowness of the company income tax base was a pressing problem. The total fiscal cost of special allowances and credits was 42 per cent of the total income tax collectable in the absence of concessions. The corporate income tax made up just 6.7 per cent of total tax revenue. The business tax system was not efficient, discriminating between different types of business organisation, financing, investments and sources of income.21 It did not adjust for inflation. The classical company tax system was also inequitable as it imposed a disproportionately heavy burden on low income shareholders (New Zealand Treasury, 1984, pp. 219–222). Implementing a comprehensive company tax base was difficult, and the base never was made fully comprehensive, most notably in relation to all capital gains (Burman and White, 2003). Indeed, ‘failure to close the holes in the corporate tax system faster than the vandals could cut new ones’ was one of the five failed policies of the Fourth Labour Government, according to a former head of the Treasury (Scott, 2005, p. 63). Nevertheless, research using King-Fullerton effective tax rate methodology concluded that the New Zealand tax system in the late 1990s was significantly more neutral with respect to investment in different types of capital assets and different methods of financing than the system in the early 1980s (Moes, 1999).

21

In 1983, average effective tax rates for companies ranged from –50 per cent for the forestry and wood sector to 39 per cent for the retail sector (Bevin, 1985, pp. 26, 39).

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

125

The company tax had also more than doubled, both its share of total tax revenue from 1984 to 200822 and as a percentage of GDP from 1985 to 2006,23 even though the resident company tax rate was 45 per cent in 1984 and 33 per cent in 2006 and 2008.24 The expanded company tax base may be part of the explanation, but higher company profitability, an improved economy, the base-protecting effect of the New Zealand dividend imputation scheme and a larger corporate sector may also be factors. The latter is highly likely given the incentives to incorporate with a lower company tax rate (at the time, 33 per cent) than the top personal marginal tax rate (at the time, 39 per cent) (New Zealand Inland Revenue Department, 2008, pp. 33–36). The high relative importance of company tax revenue to New Zealand (as a percentage of GDP) and the relatively high company tax rate, compared to most other OECD countries, is illustrated in Chart 6. A dividend imputation credit regime, proposed by Benge and Robinson (1986), was implemented in 1988 and has been sustained. In 2003, the imputation scheme was extended to Australian resident companies operating in New Zealand and paying New Zealand tax. Alignment of the top personal and company tax rates, however, is an essential complement to the imputation system if most distortions arising from the choice of business organisation are to be eliminated and the progressive personal income tax is to be protected. Alignment was achieved at 48 per cent from 1986–1988 and, more importantly, at the internationally-competitive rate of 33 per cent from 1989– 1999. This near-ideal imputation/integration system finally broke down when the top personal tax rate was raised to 39 per cent by the incoming Fifth Labour Government to take effect in 2000. Reflecting the tension between domestic and international tax policy objectives under rate alignment, this rise in the top personal tax rate signalled the higher priority assigned to vertical equity under the new Government. By itself, however, this change would have had minimal distributional effects as lower entity rates, and, then, the narrowing income tax base for private savings, enabled high income earners to shelter income from higher effective marginal rates (for data and a case study, see New Zealand Inland Revenue Department, 2008, pp. 36–42). Currently, the top marginal personal tax rate is 38 per cent and there are various entity rates of 30 and 33 per cent. Benge and Holland (2009) analyse the approaches that could be taken to make the rates more consistent: the realignment of the top personal and company tax rates; integrity measures to ‘mind the gap’; and the Nordic split rate system. 22

23 24

The New Zealand company tax share of total tax revenue was 15.3 per cent in 2008 and just 6.7 per cent in 1984 (New Zealand Government, 2008, p. 46 and New Zealand Treasury, 1984, p. 213). New Zealand company tax collections increased from 2.6 to 5.8 per cent of GDP between 1985 and 2006 (New Zealand Inland Revenue Department, 2008, p. 35). From 1 April 2008, the company tax rate is 30 per cent.

25

Korea

United Kingdom

Spain *

Denmark *

Japan

Czech Republic

Luxembourg

C om pany in co m e tax rate

Iceland

Poland Greece

Slovak Republic France *

Portugal

Switzerland

United States

Netherlands

Finland

Italy

Sweden Canada

Belgium

Ireland

C o m pan y incom e tax reven ue as % o f gro ss dom estic pro du ct

Mexico ** Turkey

Germany

Austria Hungary

New Zealand Zealand New

Australia

Norway

New Zealand Inland Revenue Department (2008, p. 26).

N o te: C ou ntries are ranked fro m high est com pany in co m e tax revenue as % of gross do m estic product to low est. T he com parison s inclu de all levels of go vern m ent. * T h e total tax reven ues have been reduced by th e am oun t of capital transfer. T he capital tran sfer has been allocated betw een tax h eadings in proportio n to th e reported tax revenu e. ** C om pany in co m e tax revenu e not available. S ource: O E C D

0

5

10

15

20

25

30

35

40

45

Chart 6: Company Income Tax Rates and Revenues in OECD (2006)25

126 TAX REFORM IN THE 21ST CENTURY

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

127

The current Government has a medium-term ‘aspirational’ goal of an aligned 30-30-30 tax rate schedule for the top personal, trust and company tax rates. Two factors that will make it difficult for the Government to reach this goal are that the 30 per cent New Zealand company tax rate is already above the average company tax rate of 26 per cent in other small OECD countries (New Zealand Treasury, 2009), and rising deficits make rate reductions of this magnitude unlikely in the near future.

5.3

FINANCIAL ARRANGEMENTS INCOME TAX BASE

With the deregulation of the financial sector and the abolition of many government controls in 1984–1985, tax planning schemes increasingly exploited asymmetries and uncertainty in existing law to defer income for tax purposes and advance deductions for expenditure. Major revenue losses occurred. The Government published far-reaching proposals to tax financial arrangements and appointed the Brash Committee to review the proposals and public submissions on them. The Brash Committee prepared draft legislation that further extended the scope of the regime. A Bill was introduced to Parliament and referred to a select committee, where further submissions and changes were made to the Bill. In March 1987, a radical accrual regime was enacted to tax all returns on financial arrangements progressively over the term of the arrangements. The regime applied to most arrangements where there were timing differences in flows of money or value, including debt obligations, debt equivalents, and certain derivatives. In many cases, all of the income and expenditure was required to be spread over the term of the financial arrangement. It taxed expected, and some unexpected, returns. It included remitted debt in income and allowed most but not all capital losses on financial arrangements to be deductible. Like the GST, this regime for the taxation of financial arrangements was more comprehensive than others in the OECD, and it has been sustained over 20 years. One of the major ongoing issues with the regime has been its scope. A governmentappointed consultative committee of private sector experts (Valabh, 1991, p. 5) reviewed the operational aspects of the regime in the early 1990s because, in its view: That legislation has not been trouble free. It has been subject to fairly regular amendment, frequently with the intention of narrowing the scope of the legislation or removing ambiguities. Despite those amendments, it can fairly be said that the accrual rules are the subject of continuous complaint by taxpayers and practitioners on the grounds that aspects of the rules are perceived to be unfair or unworkable. Most of the changes that this committee suggested would have narrowed the scope of the regime to tax expected returns from debt obligations and debt

128

TAX REFORM IN THE 21ST CENTURY

equivalents (see Edgar, 2000, pp. 166–171, for an excellent conceptual discussion of the issues). These proposals were not enacted, but some of the issues the committee identified were dealt with in subsequent amendments. On a number of occasions, the financial arrangements definition has been narrowed to exclude certain arrangements, sometimes retrospectively back to 1986. A major review of the regime by the Government was undertaken in 1997. The resulting amendments sought to clarify the rules and to reduce compliance costs.

5.4 THE INTERNATIONAL INCOME TAX BASE FOR RESIDENTS As New Zealand had had foreign exchange controls since 1938, it was understandable that the Treasury’s tax policy briefing for the incoming Government in July 1984 largely had a closed economy perspective. It had very few references to cross-border tax issues other than one paragraph on ‘international compatibility’. That paragraph noted, with prescience, that a tax system may ‘provide unintentional incentives to transfer capital from one country to another’, but reform options may require renegotiation of double tax treaties, which may make unique New Zealand forms of tax ‘particularly difficult’ (New Zealand Treasury, 1984, p. 212). Once the Fourth Labour Government removed exchange controls in December 1984, New Zealand residents could freely invest offshore through foreign companies and defer any liability for New Zealand tax until the income was repatriated and distributed to New Zealand resident individual shareholders (at the time, intercorporate dividends were exempt from New Zealand company tax). A resident could set up a ‘tax-haven run-around’ structure so that a single loan from a New Zealand bank could be turned into ‘two’ loans, producing double interest deductions for the resident, for example. Residents could also shelter income from New Zealand tax in trusts with non-resident trustees, as well as in foreign life insurance policies and superannuation schemes. The Government was slow to react. It timidly proposed increasing the Inland Revenue’s information gathering powers (July 1986) and, then, limited ad hoc measures, like tax haven disclosure requirements (December 1986). Two and a half years after the lifting of foreign-exchange controls, the Government finally responded to the serious challenge to the domestic New Zealand tax base. In June 1987, it announced more systematic anti-tax haven measures that would have taxed New Zealand residents on ‘passive’ income of controlled tax haven entities. These proposals were not unique to New Zealand. They were modelled on other countries’ controlled foreign company (CFC) legislation (Prebble, 1987). Following the October 1987 stock market crash, the New Zealand Government responded with a huge tax and regulatory reform package on 17 December 1987. In a complete shock to investors, already reeling from the comparatively large New Zealand stock market fall, the Government announced proposals to tax, on an accruals basis, all foreign source income earned by New Zealand residents through

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

129

companies and trusts. With insufficient information to prepare New Zealand tax accounts for the foreign company, most taxpayers would have had to pay accrual capital gains tax on their foreign-company shares year-by-year, as the shares recovered in value in the wake of the crash. The Government appointed a private-sector consultative committee to review Government proposals and submissions. The consultative committee recommended more limited but still comparatively comprehensive international tax measures: strengthened individual and company residence rules; a controlled foreign company (CFC) regime for both active and passive income earned through a company resident outside seven grey-list countries (about 70 per cent of New Zealand foreign direct investment went into these seven countries); foreign investment fund (FIF) rules for portfolio investment in foreign companies, and foreign life and superannuation policies; and, a settlor regime for taxing trustee income of trusts with a resident settlor. In December 1988, most of the consultative committee’s recommendations were enacted, with staggered implementation dates from 1987 to 1990. These foreign entity regimes were the most comprehensive in the world. The broad-base trust regime was implemented and remains in place today. The FIF regime and the CFC regime, beyond its application to tax havens, were much more difficult to implement. In part, this was caused by the heated debate, within and without the public sector, over whether the government’s international tax objective should be narrow (tax base protection) or broader (in addition, seeking to improve the quality of investment decision-making by New Zealand residents, to reduce the cost of capital for all New Zealand firms, and to attract foreign investment and technology). To the extent that the broader objectives were accepted, there was little consensus about how best to implement them. In part, the implemention difficulties resulted from the chequered history of government international tax announcements and the fact that the regimes were not bedded down before the Labour Government lost the 1990 election. The new National Minister of Finance, Ruth Richardson, did not enjoy two of Roger Douglas’ advantages: she had not prepared herself well on tax policy when she was in opposition (Richardson, 1995, p. 124); and, she and the National Minister of Revenue were frequently at loggerheads, including over international tax policy (Richardson, 1995, p. 123–130). In the end, it took five years of deferral and transition before new CFC and FIF regimes were enacted and implemented in 1993. There are some questions about the extent of compliance with the second FIF regime for taxing foreign portfolio equity, which applied from 1993 to 2007, but there is no hard data to test this concern. From 2007, a third, more comprehensive, FIF regime applies. This FIF regime broadened the base by substantially reducing the country exemptions from seven countries to certain investments in one country only (Australia). Very broadly, for most taxpayers, this regime applies a presumptive tax on a maximum of 5 per cent of the opening market value of offshore shares each

130

TAX REFORM IN THE 21ST CENTURY

year and no tax is payable if the shares make a loss. The early days of the regime see many shareholders facing real, large losses in the wake of a global fall in stock prices but paying tax year-by-year using a presumptive taxation method. The CFC measures to tax foreign direct investment income on an accrual basis went far beyond anything attempted by other capital-exporting countries and would approximate the ideal residence tax principle of capital export neutrality. Of course, to achieve capital export neutrality in practice all the major capital exporters would need to implement it. Also, this treatment of active income was regarded as unduly harsh by New Zealand business competing with foreign companies paying little or no home country tax in some low-tax jurisdictions. Strong pressure to reduce or remove the tax on active income was therefore to be expected. Proposed changes to the CFC income tax base in 2009 will make it less comprehensive, introducing an exemption for active income of CFCs, as in other countries. The Government’s (Cullen and Dunne, 2006) concern was that the foreign accrual CFC regime was, on balance, detrimental to the internationalisation of New Zealand business. It did not give enough weight to the national gains arising from access to offshore capital, ideas and technology when New Zealand enterprises operate offshore. For nearly 20 years, a small, open economy has implemented a conceptuallycorrect residence CFC regime by itself, taxing income on around 30 per cent of its outward foreign direct investment on accrual. Where New Zealand has led, however, major capital-exporting countries have not followed. In that time, both the government and business have sought the best advice available on how a small, open economy should tax its exported foreign direct investment. But as Slemrod (1997, p. 456) puts it, ‘providing clear policy advice in this environment is certainly difficult, even if we put equity issues aside and concentrate on efficient tax policies’. The first generation of empirical research established conclusively that taxes matter to multinational corporations. The next question of how taxes matter requires ‘integrating models of the nature of the multinational enterprise and opportunities for tax avoidance into normative models of tax policy’ (Slemrod, 1997, p. 457). This remains to be accomplished.

6 The Political Economy of the Income Tax Base Reforms Eight political economy factors differed for the New Zealand comprehensive indirect consumption and income tax reforms of 1984–1988. They help explain the different outcomes. First, there was the sequence of reforms. The most successful tax reforms (the GST and the financial arrangement income tax rules, for instance) were implemented in the first term of the Fourth Labour Government when its electoral mandate was fresh.26 The Government was prepared to put deficit reduction behind 26

There are many examples of new governments introducing major tax reform in their first term when their electoral mandate was fresh. Particularly striking are the two direct quotations

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

131

gaining acceptance of the reforms (Dickson, 2007; Stephens 2007), and it was more prepared to be patient and consultative. The less successful income tax reforms were enacted in the second term of the Government when raising revenue was given greater priority, and the December 1987 package seemed ‘hastily conceived’ and poorly promoted (Stephens, 1990; Dickson, 1989). Second, the comprehensive packaging of reforms, so that each group could see both losses and gains in a package (Douglas, 2005), did not always work. The trade-offs in the first package (Douglas, 1985) introducing the new broad-base, 10 per cent GST, abolishing the wholesale sales tax, lowering personal income tax rates (the top marginal personal rate from 66 per cent to 48 per cent) and implementing immediate family benefit reforms were widely accepted. The tradeoffs in the second (1987) package were not. This second package sought to change the tax mix in favour of indirect tax and to make both taxes impersonal and flat rate. The trade-offs in this package (Douglas, 1987) included a 2.5 per cent increase in the GST to 12.5 per cent, a ‘single nominal rate of personal [income] tax [about 23 to 24 per cent]’, more benefit assistance, many of the comprehensive income tax reforms and extensive regulatory reform. When the Prime Minister unilaterally deferred the flat-rate income tax and guaranteed minimum family income proposals one month after they had been announced (see below), the whole package unravelled into a series of seemingly unrelated proposals. Some smaller packages introduced in 1987 did not work either: taxpayers accepted the full imputation system of company taxation, but many opposed the ‘stringent’ international tax regime, as the Government described it on announcement. The Government’s attempt to link popular gains with unpopular losses, by sending both income tax proposals to the same consultative committee, failed. Third, the support of the Prime Minister was essential. David Lange provided vital support for the first major tax reform package (Douglas, 2005) but not for the second package. About one month after Cabinet had approved, and he and other ministers had announced, the second major tax reform package that included the proposed flat rate of income tax, Lange unilaterally announced deferral of the flat tax and guaranteed minimum family income aspects of the package (Stephens, 1990; James, 1992; Chapman, 1992; Douglas, 1993; Lange, 2005). The fourth factor is ideas. At the time, support for the VAT/GST tax base was growing worldwide, while support for progressive income tax rates was eroding. from United Kingdom Chancellor of the Exchequer Howe and Prime Minister Thatcher cited in Riddell (2008, p. 4). Sir Geoffrey Howe (1994, p. 130) argued ‘that our first Budget provided “our only opportunity to make a radical switch from direct to indirect taxation and thus honour the commitment on which our credibility depends”’. In the view of Thatcher (1993, p. 43), ‘Income tax cuts were vital, even if they had to be paid for by raising Value Added Tax (VAT) in this large leap. The decisive argument was that such a controversial increase in indirect taxes could only be made at the beginning of a parliament, when our mandate was fresh?’.

132

TAX REFORM IN THE 21ST CENTURY

It is doubtful, however, whether the New Zealand public would have accepted disengaging or divorcing all taxation instruments from vertical equity considerations and progressivity by adopting flat-rate, impersonal consumption and income taxes. That would have left income distribution issues to be dealt with solely through the welfare system. Further, there is much more room for principled disagreement about what ‘good taxation’ policy requires for a tax that relies upon assessment over a year and attempts to take into account offshore factors (Bird, 1987), as the income tax does, as opposed to a tax that is based on day-to-day transactions within a jurisdiction, as the VAT/GST does. The debates within, and between, the Mirrlees Review groups dealing with value-added tax and excises, the base of direct taxation, taxing corporate income, and international capital taxation, for example, illustrate this point. Fifth, tax structure is greatly affected by underlying forces in the world economy, with economic globalisation leading to more competition for goods, services, capital and labour and each tax having different impacts on investment decisions and productivity at national, industry and firm levels (New Zealand Treasury, 2009, pp. 6–7; Gemmell et al., 2009).27 In section 4.3, we discussed John Kay’s thesis that reduction in government barriers and advances in technology mean tax bases that rely on ‘assessing’ income from capital over a whole year are becoming increasingly problematic compared to taxes that have a real, day-to-day transactions base. Kay (1990), it should be quickly added, was not arguing for an impersonal, flat-rate tax system. He (2008) has gone as far as arguing that for the Meade Committee, of which he was a member, and ‘most ordinary people, questions of fairness and taxable capacity would seem to be of critical importance – even exclusive importance – in determining the household tax base’. As VAT and ‘the existing means of taxing labour income’ had a very limited capacity to secure distributional objectives, Kay (1990, p. 67) saw the challenge for the 1990s and beyond to be broadening ‘the range of instruments consistent with the transactions base. ... The transformation of the income tax into a progressive personal expenditure tax, and of the corporation tax into a tax based on corporate cash flows, provide the solutions to these problems’. Sixth, external influences played a key role in the sustainability of the reforms. In the case of the GST, two international organisations championed the reforms (the IMF and the OECD) and one provided technical assistance on implementation (the IMF). In the case of the income tax, by comparison, one international organisation provided advice (the OECD) while international coordination mechanisms, in the form of New Zealand’s double tax treaty network, imposed constraints on comprehensive income tax reform. 27

With New Zealand’s high levels of international economic integration for capital and highly skilled labour and its population ageing, a switch towards less mobile consumption and property tax bases has been suggested in the longer term (New Zealand Treasury, 2009).

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

133

The seventh factor is interest groups. With the clean break from the former secretive tax policy-making process, senior tax professionals played crucial and constructive roles on consultative committees for both the indirect consumption and income tax reforms, receiving public submissions, commenting on government proposals and making implementation suggestions. There was, however, greater uncertainty over what was good income tax policy for a small, open economy. In addition, assessing the economic and political costs involved with an annually assessed tax that attempted comprehensively to tax worldwide income produced more differences of opinion and some minority reports in the income tax consultative committees. Finally, as the Government proposed taxing more forms of capital income in the wake of the 1987 international stock market crash, while government expenditure remained high and the flat-rate income tax proposal was abandoned, opposition to those comprehensive income tax base-broadening measures greatly increased. The final factor is the media. The ‘new’ GST had the benefit of an unprecedented information, education and coordination effort from an office staffed by people from the private and public sector (Todd, 2007). The publicity for the ‘old’ income tax reforms was handled by departments and Ministers’ offices. In addition, many of the ‘old’ income tax reforms were implemented after the Prime Minister and Minister of Finance had publicly fallen out over the flat-rate income tax and guaranteed minimum family income issues in January 1988, which inevitably spilled over to affect media commentary on all Government policy and actions.

7

Conclusions

This chapter has reflected on the political economy of New Zealand tax policy design, implementation (in 1984–1988), and sustainability. The resulting broaderbase, lower-rate New Zealand tax system was judged by the OECD (1989) to be one of the least distorting among member country tax systems. Reflecting on the reforms 20 years on, a major issue that this chapter has examined is whether a broad-base, low-rate indirect consumption tax is more likely to be sustained than a broad-base, low-rate income tax. For this study of selected income and indirect consumption tax base reforms, it was possible to hold constant six political economy factors that writers claim can help assist major tax reform to achieve its objectives: macroeconomic adversity; widespread discontent with the tax system; a political entrepreneur for the tax policy reforms; the form of government and electoral rules; party government; and the bureaucratic institutional structure for economic and tax policy-making and implementation. This enabled our study to focus more heavily on eight political economy factors that differed for the New Zealand comprehensive indirect consumption and income tax reforms of 1984–1988 and that help to explain the different outcomes:

134

TAX REFORM IN THE 21ST CENTURY

ideas; underlying economic forces; external influences like regional countries and international agencies; the sequence of the reforms; the packaging of the reforms; the support of the political leader (Prime Minister); interest groups; and the media. Our discussion has sought to show the intricate interplay of all of these factors and to warn against simple, single-factor explanations. The most stable and enduring of the two New Zealand broad tax bases over more than 20 years has been the indirect VAT/GST consumption tax base. The twin GST policy pillars of comprehensive coverage and a single domestic rate have been sustained with the full support of the major political parties and players in the tax system. As explained in the chapter, the sustainability of this tax over more than 20 years can be attributed to the open, patient, political implementation with larger economic, legal, geographic and constitutional factors providing essential underpinning. As Musgrave would be quick to remind us, however, two of the GST’s strengths are also its weaknesses. First, its limited jurisdictional scope means that its design and administration is not so influenced by open economy considerations but also that it is a less fair measure of ability-to-pay. It does not include worldwide consumption. It does not tax a person’s lifetime income, unless there is a tax on gifts and bequests, and it makes no allowance for ‘the utility derived from holding wealth prior to its consumption’ (Musgrave, 2001, p. 77). Secondly, impersonal, flat-rate taxation might be administratively convenient, but it cannot match the contribution that personal, progressive taxation can make to achieving a fair distribution of income. For Musgrave (2001, p. 79), ‘visible taxation and personal participation in the taxation process is important as a matter of fiscal discipline and, recalling my Wicksellian links, is needed also to assure fiscal discipline and to offer guidance in the conduct of fiscal affairs’. Of course, with source collection of much income tax today, as Musgrave (2001, p. 79) himself advocated for simplification purposes once major loopholes had been closed, and income tax returns no longer required for many taxpayers in countries like New Zealand, the income tax itself is becoming less visible. As Musgrave (2001, p. 75) was prepared to concede, there are ‘no ultimate solutions’. A final point to make on the GST is that it is not the revenue work horse of the New Zealand tax system. It is an ‘add-on’ revenue raiser. It aims to raise only about 20–25 per cent of total tax revenue. If it were called upon to fund 70 per cent of total tax revenue, as the New Zealand income tax currently does, there is no reason to believe that the twin GST policy pillars of comprehensive coverage and a single domestic rate would survive intact. In addition, higher rates of VAT/GST increase the incentives for avoidance and fraud. On the other hand, the New Zealand income tax problem in 1984 was more complex than the indirect tax problem of how to tax the personal consumption that was not taxed at the time under the wholesale sales tax. The income tax was required

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

135

to raise about 70 per cent of total revenue. It was an old tax that had developed over nearly a century in a haphazard way. It had an ambitious, worldwide jurisdictional reach. It was designed for a closed economy but, from the abolition of exchange controls in December 1984, it had to operate in an increasingly open economy. Up until mid-1984, it had been supported by pervasive government regulation of the economy and widespread government ownership of economic and social enterprise. From July 1984, it had to operate in an economy that was rapidly deregulated, with many government trading enterprises first corporatised and, from 1987, privatised. The thoroughly inadequate and incoherent design of the old tax was completely exposed. From 1984 to 1988, New Zealand made a serious attempt to implement a HaigSimons comprehensive accretion income tax base. It had some remarkable successes that have eluded other countries. Many remain intact today. Of the income tax bases selected for this study, the most enduring have been the financial arrangements income tax base, the full dividend imputation credit system and the international trust income tax regime (a relatively insignificant tax base, even in the context of the New Zealand international income tax base). Alignment of the top personal and company tax rates, however, is an essential complement of the imputation system if most distortions arising from the choice of business organisation are to be eliminated and the progressive personal income tax is to be protected. This was achieved from 1986–1988 and from 1989–1999, but the system broke down with the rise in the top personal tax rate to 39 per cent in 2000. Reforming the company income tax base was difficult, and the base never was made fully comprehensive, most notably in relation to capital gains. Nevertheless, New Zealand’s reliance on company tax today is significant with a ratio of company tax to GDP well above the OECD average. The comprehensive retirement savings income tax reforms were sustained for nearly 20 years. Unfortunately, tax incentives for saving introduced in 2007, together with the large rate gaps already discussed, have created opportunities for tax planning and real concerns about the integrity of the income tax system. The broad-base approach to the foreign portfolio and foreign direct equity investment income of residents initially proved difficult to implement but, after five years of deferral and transition, stable regimes were implemented in 1993. Recent changes have made the foreign portfolio equity investment income base more comprehensive. Proposed changes to the foreign direct equity investment income tax base will make it less comprehensive. The crunch for the New Zealand comprehensive income tax experiment arose with the Fourth Labour Government’s December 1987 announcement of a ‘a single nominal rate of personal tax on all taxable income from 1 October 1988’ and a guaranteed minimum family income – and the Prime Minister’s unilateral deferral of the two proposals one month later. The proposal would have made both the income

136

TAX REFORM IN THE 21ST CENTURY

and consumption taxes impersonal and flat rate. It was the logical extreme of the ‘broad-base, low-rate’ approach to tax policy-making, which had almost become a mantra in New Zealand. But the ‘broad-base, low-rate’ approach is not first-best economics and nor is it first-best public finance, in the best traditions that Richard Musgrave upheld. It is simply a practical tax policy-making guide for efficient taxation in some, but by no means all, circumstances. At one level, the dispute between the Finance Minister and the Prime Minister was about the ‘fiscal arithmetic’ of the December 1987 package but, at a deeper level, it was about different views on social policy and equity, as well as the distributional effects of the tax and transfer system (Douglas, 1993; Lange, 2005). It underlined the importance of analysing and understanding our value judgments and the distributional effects of taxes, the transfer system and public spending,28 as well as the distribution of wealth and earnings in a country.29 Good taxation, as Musgrave (2001, p. 75) argued, is ‘a key test of democracy and a major piece of social capital’. An enduring legacy of the New Zealand tax reforms of 1984–1988, of which Musgrave would approve, has been the opening up of tax policy-making and implementation to the public. The extent and ways that governments have consulted the public from 1984 have differed, but we all have the opportunity to be much more involved in ‘taxing ourselves’30 than we did before 1983. References Aaron, H.J. (2008), ‘Discussion’, in J.W. Diamond and G.R. Zodrow (eds), Fundamental Tax Reform: Issues, Choices, and Implications (Cambridge, MA: MIT Press), 61. Alm, J. (2008), ‘Discussion’, in J.W. Diamond and G.R. Zodrow (eds), Fundamental Tax Reform: Issues, Choices, and Implications (Cambridge, MA: MIT Press), 125. 28

29

30

Over more than ten years now, John Creedy (for example, 1997, 2006), sometimes with a coauthor, has been writing on the distributional effects of the income tax, GST and excises in New Zealand. For more recent data and how New Zealand Treasury officials analyse current equity and distributional issues, see New Zealand Treasury (2009, pp. 12–15). For New Zealand, Atkinson (2008, p. 301) notes the estimates that from 1984 to 1997 the income shares of the bottom decile fell and that of the top decile rose significantly. On the other hand, Crawford and Johnston’s (2004) comparison of final household incomes in 1997/98 with 1987/88, taking into account payments to, and benefits from, central government (in cash or kind), argues that the final incomes of less well-off households at these two reference points were maintained. In his Rodolfo Debenedetti lectures on the changing distribution of earnings in OECD countries, Atkinson (2007, p. 83) points out that ‘it is important to remember that the distribution of earnings is not outside our control. It is true that governments are constrained by the global economy and by the pace of technological change, but policy has a role. … Progressive taxation is still possible. All of these are matters over which citizens have some say; they are not purely spectators of an inevitable process’. That enviable title of a citizen’s guide to the great debate over tax reform in the United States by Slemrod and Bakija (2008).

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

137

Alt, J., I. Preston, and L. Sibieta (2008), ‘The Political Economy of Tax Policy’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, Chaired by Sir James Mirrlees, available at www.ifs.org.uk/mirrleesreview/ index.php (as at April 2009). Arnold, B.J. (1990), ‘The Process of Tax Policy Formulation in Australia, Canada and New Zealand’, Australian Tax Forum 7: 379. Atkinson, A.B. (2008), The Changing Distribution of Earnings in OECD Countries (Oxford: Oxford University Press). Auerbach, A.J. (2001), ‘Comments on Tax Review 2001, Issues Paper, June 2001’, mimeo. Auerbach, A.J. (2007), ‘Comment’, in H.J. Aaron, L.E. Burman and C.E. Steuerle (eds), Taxing Capital Income (Washington, DC: Urban Institute Press), 83. Auerbach, A.J. (2008), ‘Tax Reform in the Twenty-first Century’, in J.W. Diamond and G.R. Zodrow (eds), Fundamental Tax Reform: Issues, Choices, and Implications (Cambridge, MA: MIT Press), 27. Auerbach, A.J., and W.G. Gale (2009), ‘The Economic Crisis and the Fiscal Crisis: 2009 and Beyond’, mimeo. Benge, M. and T. Robinson (1986), How to Integrate Company and Shareholder Taxation: Why Full Imputation Is the Best Answer (Wellington: Victoria University Press for Institute of Policy Studies). Benge, M. and D. Holland (2009), ‘Company Taxation in New Zealand’, mimeo, February. Berlin, I. (1953), The Hedgehog and the Fox: An Essay on Tolstoy’s View of History (New York: Simon & Schuster). Bevin, P. (1985), How Should Business Be Taxed? An Examination of Defects in Business Taxation and Suggestions for Reform (Wellington: Victoria University Press for Institute of Policy Studies). Birch, B. and B. English (1999), GST: A Review (Wellington: Inland Revenue Department). Bird, R.M. (1987), The Taxation of International Investment Flows: Issues and Approaches (Wellington: Victoria University Press for Institute of Policy Studies). Bird, R.M., and P.-P. Gendron (2007), The VAT in Developing and Transitional Countries (Cambridge: Cambridge University Press). Boston, J. (1989), ‘The Treasury and the Organisation of Economic Advice: Some International Comparisons’, in B. Easton (ed.), The Making of Rogernomics (Auckland; Auckland University Press), 68. Brash, D. (1987), Comprehensive Tax Reform and Possible Interim Solutions (Wellington: New Zealand Government Printer).

138

TAX REFORM IN THE 21ST CENTURY

Brooks, N. (2007), ‘An Overview of the Role of the VAT, Fundamental Tax Reform, and a Defence of the Income Tax’, in R. Krever and D.I. White (eds), GST in Retrospect and Prospect (Wellington: Brookers), 597. Burman, L. and D.I. White (2003), ‘Taxing Capital Gains in New Zealand’, New Zealand Journal of Taxation Law and Policy 9(3): 355. Burman, L. (2009), ‘A Blueprint for Tax Reform and Health Reform’, Virginia Tax Review 28: 287. Canadian Royal Commission on Taxation (1966), Report (Carter Report) (Ottawa: Government Printer). Caygill, D. (1989), Consultative Document on the Taxation of Income from Capital (Wellington: Government Printer). Caygill, D. (1999), ‘Managing Risk: A Ministerial Perspective’, in A. Sundakov and J. Yeabsley, (eds), Risk and the Institutions of Government (Wellington: Institute of Policy Studies), 46. Chapman, R. (1992), ‘A Political Culture Under Pressure: The Struggle to Preserve a Progressive Tax Base for Welfare and the Positive State’, Political Science 44: 1. Conlan, T., M. Wrightson and D. Beam (1990), Taxing Choices: The Politics of Tax Reform (Washington, DC: CQ Press). Crawford, I., M. Keen and S. Smith (2008), ‘Value-Added Tax and Excises’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, Chaired by Sir James Mirrlees, available at www.ifs.org.uk/mirrleesreview/index. php (as at April 2009). Crawford, R. and G. Johnston (2004), ‘Household Incomes in New Zealand: The Impact of the Market, Taxes and Government Spending, 1987/88–1997/98’, New Zealand Treasury Working Paper 04/20, available at www.treasury.govt.nz (as at April 2009). Creedy, J. (1997), The Statics and Dynamics of Income Distribution in New Zealand (Wellington: Institute of Policy Studies). Creedy, J. and C. Sleeman (2006), The Distributional Effects of Indirect Taxes: Models and Applications from New Zealand (Cheltenham: Edward Elgar). Crowder, G. (2004), Isaiah Berlin: Liberty and Pluralism (Cambridge: Polity). Cullen, M. and P. Dunne (2006), New Zealand’s International Tax Review: A Direction for Change, A Government Discussion Document (Wellington: Policy Advice Division of the Inland Revenue Department). Dickson, I. (1989), ‘Taxation’, in S. Walker (ed.), Rogernomics: Reshaping New Zealand’s Economy (Auckland: New Zealand Centre for Independent Studies), 137.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

139

Dickson, I. (2007), ‘The New Zealand GST Policy Choice: An Historical and Policy Perspective’, in R. Krever and D.I. White (eds), GST in Retrospect and Prospect (Wellington: Brookers), 45. Dickson, I. and D.I. White (2008), ‘Tax Design Insights from the New Zealand Goods and Services Tax (GST) Model’, Centre for Accounting, Governance and Taxation Research Working Paper Series Working Paper No. 60 April 2008, available at http://www.victoria.ac.nz/sacl/CAGTR/CAGTRhomepage.aspx (as at April 2009). Douglas, R. (1980), There’s Got to Be a Better Way! (Wellington: Fourth Estate Books). Douglas, R. (1985), Statement on Taxation and Benefit Reform 1985 (Wellington: Government Printer). Douglas, R. and L. Callen (1987), Toward Prosperity (Auckland: David Bateman). Douglas, R. (1987), Government Economic Statement, 17 December (Wellington: Government Printer). Douglas, R. (1993), Unfinished Business (Auckland: Random House). Douglas, R. (2005), ‘How We Did It’, in M. Clark (ed.), For the Record: Lange and the Fourth Labour Government (Wellington: Dunmore Publishing), 56. Douglas, R. (2007), ‘The New Zealand GST Policy Choice and its Political Implications’, in R. Krever and D.I. White (eds), GST in Retrospect and Prospect (Wellington: Brookers), 3. Ebrill, L., M. Keen, J.-P. Bodin and V. Summers (2001), The Modern VAT (Washington, DC: IMF). Edgar, T. (2000), The Income Tax Treatment of Financial Instruments: Theory and Practice (Toronto: Canadian Tax Foundation). Evans, L., A. Grimes, B. Wilkinson and D. Teece (1996), ‘Economic Reform in New Zealand 1984–95: The Pursuit of Efficiency’, Journal of Economic Literature 34(4): 1856. Ganghof, S. (2006), The Politics of Income Taxation (Colchester: ECPR Press). Gemmell, N., R. Kneller and I. Sanz (2009), ‘The Growth Effects Of Corporate and Personal Tax Rates in the OECD’, mimeo. Green, R. (2007), ‘Consulting the Public in Developing a GST’, in R. Krever and D.I. White (eds), GST in Retrospect and Prospect (Wellington: Brookers), 13. Head, J.G. (1989), ‘Australian Tax Reform: An Overview’, in J.G. Head (ed.), Australian Tax Reform in Retrospect and Prospect (Sydney: Australian Tax Research Foundation), 1.

140

TAX REFORM IN THE 21ST CENTURY

Head, J.G. (2009), ‘The Political Economy of Tax Reform: A Neo-Musgravian Perspective with Illustrations from Canadian, US, Australian and New Zealand Experience’, ch. 1, this volume. Howe, G. (1994), Conflict of Loyalty (London: Macmillan). Jackson, K. (1987), The Dilemma of Parliament (Wellington: Allen & Unwin New Zealand in association with Port Nicholson Press). James, C. (1992), New Territory: The Transformation of New Zealand 1984–92 (Wellington: Bridget Williams Books). James, C. (ed). (2000), Building the Constitution (Wellington: Institute of Policy Studies). Katz, D. (1999), ‘Towards a Practical Cash Flow Tax’, New Zealand Treasury Working Paper 99/1, mimeo. Kato, J. (2003), Regressive Taxation and the Welfare State: Path Dependence and Policy Diffusion (Cambridge: Cambridge University Press). Kay, J.A. (1986), ‘Tax Reform in Context: A Strategy for the 1990s’, Fiscal Studies 7: 1. Kay, J.A. (1990), ‘Tax Policy: A Survey’, The Economic Journal 100: 18. Kay, J.A. (1994), ‘Tax Reform: A Perspective Longer Than the Life of One Parliament’, in The Irish Dilemma: How to Achieve Fiscal Reform, Proceedings of the Ninth Annual Conference of the Foundation for Fiscal Studies (Dublin: Foundation for Fiscal Studies), 10. Kay, J.A. (2008), ‘The Base for Direct Taxation: Commentary’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, Chaired by Sir James Mirrlees, available at www.ifs.org.uk/mirrleesreview/index.php (11 December 2008). Lange, D. (2005), My Life (Auckland: Viking). Lojkine, S. (1989), ‘Tax Reform: The Consultative Process 1985–1988’, The Accountants’ Journal, October: 22. Marriott, L. (2008), ‘The Politics of Retirement Savings Taxation: A Trans-Tasman Comparison (1972–2007)’, a Ph.D. thesis available at http://researcharchive.vuw. ac.nz/ (as at April 2009). Moes, A. (1999), ‘Effective Tax Rates on Capital in New Zealand – Changes 1972– 1998’, New Zealand Treasury Working Paper 99/12, available at www.treasury. govt.nz (as at April 2009). Musgrave, R.A. (1968), ‘The Carter Commission Report’, The Canadian Journal of Economics 1 (1 Supplement): 159.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

141

Musgrave, R.A. (1986), ‘In Retrospect’, in Public Finance in a Democratic Society, vol. 1, Social Goods, Taxation and Fiscal Policy, (London: Harvester Press), vii. Musgrave, R.A. (1991), ‘Tableau Fiscale’, in L. Eden (ed.), Retrospectives on Public Finance (Fiscal Reform in the Developing World) (Durham: Duke University Press), 351. Musgrave, R.A. (1992), in R.A. Musgrave and T.A. Wilson, ‘Reflections on CanadaU.S. Tax Differences: Two Views’, in J.B. Shoven and J. Whalley (eds), Canada-US Tax Comparisons (Chicago: University Of Chicago Press), 359. Musgrave, R.A. (2001), in J.M. Buchanan and R.A. Musgrave, Public Finance and Public Choice: Two Contrasting Visions of the State (Cambridge, MA: MIT Press), 63. New Zealand, Consultative Committee on the Taxation of Income from Capital (1991), Operational Aspects of the Accruals Regime (Valabh Committee), (Wellington: Government Printer). New Zealand Government (2008), ‘Financial Statements of the Government of New Zealand for the Year Ended 30 June 2008’, available at www.treasury.govt.nz (as at April 2009). New Zealand, Inland Revenue Department (2008), ‘Briefing for the Incoming Minister of Revenue – 2008’, available at www.ird.govt.nz (as at April 2009). New Zealand Officials, (2008), ‘Options for Strengthening GST Neutrality in Business-to-Business Transactions’, an officials’ issues paper, available at (as at April 2009). New Zealand, Task Force on Tax Reform (1982), Report (McCaw Report) (Wellington: Government Printer). New Zealand, Tax Review 2001 (2001a), Issues Paper (McLeod Issues Paper) (Wellington: Government Printer). New Zealand, Tax Review 2001 (2001b), Final Report (McLeod Final Report) (Wellington: Government Printer). New Zealand Treasury (1984), Economic Management (Wellington: Government Printer). New Zealand Treasury (2009), ‘Medium Term Tax Policy Challenges and Opportunities’, mimeo, available at www.treasury.govt.nz (as at April 2009). OECD (1989), New Zealand Economic Survey (Paris: OECD). OECD (2007), New Zealand Economic Survey (Paris: OECD). Palmer, G. (1987), Unbridled Power: An Interpretation of New Zealand’s Constitution and Government, 2nd ed. (Auckland: Oxford University Press).

142

TAX REFORM IN THE 21ST CENTURY

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 (Cambridge, MA.: MIT Press). Peters, B.G. (1991), The Politics of Taxation: A Comparative Perspective (Cambridge, MA: Blackwell). Prebble, J. (1987), The Taxation of Controlled Foreign Companies (Wellington: Institute of Policy Studies/Victoria University Press). Preston, D. (2008), ‘Retirement Income in New Zealand: The Historical Context’, mimeo, December, available from www.retirement.org.nz (as at April 2009). Richardson, R. (1995), Making a Difference (Christchurch: Shoal Bay Press). Riddell, P. (2008), ‘The Political Economy of Tax Policy, Commentary’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, Chaired by Sir James Mirrlees, available at www.ifs.org.uk/mirrleesreview/index. php (as at April 2009). Sandford, C.T. (1993), Successful Tax Reform: Lessons from an Analysis of Tax Reform In Six Countries (Bath: Fiscal Publications). Sandford, C.T. (2000), Why Tax Systems Differ: A Comparative Study of the Political Economy of Taxation (Bath: Fiscal Publications). Scott, G. (2005), ‘One Adviser’s View of the Economics and Politics of Economic Policy’, in M. Clark (ed.), For the Record: Lange and the Fourth Labour Government (Wellington: Dunmore Publishing), 60. Shaviro, D. (1990), ‘Beyond Public Choice and Public Interest: A Study of the Legislative Process as Illustrated by Tax Legislation in the 1980s’, University of Pennsylvania Law Review 139(1): 1. Shaviro, D. (2008), ‘Simplifying Assumptions: How Might the Politics of Consumption Tax Reform Affect (Impair) the End Product?’, in J.W. Diamond and G.R. Zodrow (eds), Fundamental Tax Reform: Issues, Choices, and Implications (Cambridge, MA: MIT Press), 75. Slemrod, J. (1997), ‘Comment’, in A.J. Auerbach (ed.), Fiscal Policy: Lessons from Economic Research (Cambridge, MA: MIT Press), 453. Slemrod, J. and J. Bakija, (2008), Taxing Ourselves. A Citizen’s Guide to the Debate over Taxes, 4th ed., (Cambridge, MA: MIT Press). Stephens, R.J. (1989), ‘New Zealand Tax Reform’, in J. Head (ed.), Australian Tax Reform in Retrospect and Prospect (Sydney: Australian Tax Research Foundation), 65.

DAVID WHITE: INCOME AND CONSUMPTION TAXES IN NEW ZEALAND

143

Stephens, R.J. (1990), ‘Flattening the Tax Rate Scale in New Zealand’, in J.G. Head and R. Krever (eds), Flattening the Tax Rate Scale: Alternative Scenarios and Methodologies (Melbourne: Longman Professional), 103. Stephens, R.J. (1993b), ‘Radical Tax Reform in New Zealand’, Fiscal Studies 14(3): 45. Stephens, R.J. (2007), ‘The Economic and Equity Effects of GST in New Zealand’, in R. Krever and D.I. White (eds), GST in Retrospect and Prospect (Wellington: Brookers), 65. St. John, S. (2007), ‘New Zealand’s Experiment in Tax Neutrality for Retirement Saving’, The Geneva Papers 32(4): 532. Tabellini, G. (2008), ‘The Political Economy of Tax Policy, Commentary’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, Chaired by Sir James Mirrlees, available at www.ifs.org.uk/mirrleesreview/index. php (as at April 2009). Thatcher, M. (1993), The Downing Street Years (London: HarperCollins). Todd, J. (2007), ‘Implementing GST – Information, Education Coordination’, in R. Krever and D.I. White (eds), GST in Retrospect and Prospect (Wellington: Brookers), 27. Toder, E. and S. Himes (1992), ‘Tax Reform in New Zealand’, Tax Notes International 5 (17 August): 347. Toder, E. and S. Khitatrakun (2006), Final Report to Inland Revenue ‘KiwiSaver Evaluation Literature Review’ (IR/2006/022/EVA/T), Tax Policy Center, Washington DC. United Kingdom Institute for Fiscal Studies (1978), The Structure and Reform of Direct Taxation (Meade Report) (London: Allen and Unwin). United Kingdom Institute for Fiscal Studies (2008a), Mirrlees Review, see the conference presentations and pre-publication versions of chapters on the Institute for Fiscal Studies website at www.ifs.org.uk/mirrleesreview/index.php (as at April 2009). United Kingdom Institute for Fiscal Studies (2008b), ‘Simplify VAT to Cut Costs, Raise Revenue and Help the Poor, Says Study Prepared the Mirrlees Review’, mimeo. US Department of the Treasury (1984), Tax Reform for Fairness, Simplicity, and Economic Growth (Washington, DC: Government Printing Office). Wales, C. (2008), ‘The Political Economy of Tax Policy, Commentary’, prepared for the Report of a Commission on Reforming the Tax System for the 21st Century, Chaired by Sir James Mirrlees, available at www.ifs.org.uk/mirrleesreview/index. php (as at April 2009).

144

TAX REFORM IN THE 21ST CENTURY

Wells, G.M. (1996), ‘Fiscal Policy’, in B. Silverstone, A. Bollard and R. Lattimore (eds), A Study of Economic Reform: The Case of New Zealand (Amsterdam: NorthHolland), 215. White, D.I. (2007), ‘Twenty Years of GST: The Best Path Forward’, New Zealand Journal of Taxation Law and Policy 13(3): 357. Wilkinson, B. and I. Dickson (1989), ‘Chronology of Major Fiscal Announcements and Measures’, in S. Walker (ed.), Rogernomics: Reshaping New Zealand’s Economy (Auckland: GP Books), 225. Wilkinson, B. (1989), ‘Fiscal Policy and Government Expenditure Reforms’, in S. Walker (ed.), Rogernomics: Reshaping New Zealand’s Economy (Auckland: GP Books), 93. Yin, G.K. (2008), ‘Discussion’, in J.W. Diamond and G.R. Zodrow (eds), Fundamental Tax Reform: Issues, Choices, and Implications (Cambridge, MA: MIT Press), 133. Zodrow, G. (2007), ‘Should Capital Income be Subject to Consumption-based Taxation?’, in H.J. Aaron, L.E. Burman and C.E. Steuerle (eds), Taxing Capital Income (Washington, DC: Urban Institute Press), 49.

Personal Tax Base: Income or Consumption?

Chapter 5

Income or Consumption Taxes? Alan J. Auerbach* 1

Introduction

This is the third time I have had the opportunity to present a paper at a conference honoring Richard Musgrave.1 His participation will be missed this time, but he will continue to influence the proceedings in his capacity as a public finance superego, forcing us to step back from facile conclusions that fail to account for important institutional and ethical considerations. As in the two previous instances, I will use suitable quotes from The Theory of Public Finance to frame and motivate the presentation. Naturally, Richard’s 1959 text discusses the distinction between consumption taxes and income taxes. Indeed, the book anticipates subsequent analysis motivated by optimal tax considerations (e.g. Atkinson and Stiglitz, 1976, Feldstein 1978) by pointing to the relevance of the relative substitutability of leisure for current versus future consumption (Musgrave, 1959, p. 250). But the passage of time since 1959 is evident from a quote that appears earlier in the book, in the context of choosing an index of equality: ‘Strangely enough, it has been argued by a long succession of distinguished authors that a tax on income is inequitable because it is said to involve the double taxation of saving’ (p. 161). Whether a consumption tax is fairer than an income tax may still not be obvious, but the affirmative position is certainly no longer strange. Finally, Musgrave exposes a fallacy that still haunts the comparison of the two taxes, having to do with the inherent progressivity of these two approaches to taxation:

* 1

University of California, Berkeley; Senior Fellow, Taxation Law and Policy Research Institute, Monash University. The previous efforts were Auerbach (1994) and Auerbach and Hines (2003).

148

TAX REFORM IN THE 21ST CENTURY

As a matter of historical experience, we find that the income tax has been the vehicle of progressive taxation, and that the major sources of sales and excise taxation have been regressive. At the same time, there is no logical necessity for this. Once the transition is made from a tax on commodities to a personal tax on consumer expenditures, the spendings tax may be applied with progressive rates, no less than the income tax (pp. 163–164). Each of these three passages relates to an important element of the ongoing debate over the relative merits of income and consumption taxes. The question of whether consumption taxes are more efficient than income taxes has continued to be explored in a succession of contributions utilizing more and more sophisticated models of economic behavior. The relative fairness of consumption and income taxes relates to whether one should analyze fairness from a lifetime perspective, or perhaps using a shorter (or longer) period of assessment, a question that remains open to debate. And whether a realistic consumption tax really can be as progressive as a realistic income tax is also a live subject of inquiry, with further subtlety contributed to the comparison by the development of individual-based taxes other than the ‘spendings tax’ to which Musgrave refers, such as the flat tax as laid out by Hall and Rabushka (1985) and the multi-rate variant that Bradford (1986) called the X tax. In this paper, I review the state of discussion on many key issues relating to choice between income and consumption taxes. In doing so, I will rely on my two recent surveys (Auerbach, 2008, 2009) as well as on the chapter from the Mirrlees review project on the individual tax base (Banks and Diamond, 2008), all longer papers to which the reader is referred for more detailed discussion.

2

Economic Efficiency

Over time, there have been several different arguments favoring the consumption tax from the perspective of economic efficiency. The first, as cited by Feldstein (1978), relies on the simple Corlett-Hague (1953) analysis of proportional taxation applied to the relative complementarity of leisure to first- and second-period consumption. The logic is that, unless second-period consumption is more complementary to leisure than first-period consumption, then second-period consumption should not be taxed more heavily than first period consumption, as it implicitly would be under an income tax. A second round of optimal tax analysis dates to the frequently cited paper by Atkinson and Stiglitz (1976), who showed, under the assumption that preferences are weakly separable into consumption and leisure, that a progressive labor income tax is optimal by itself, i.e. that no variation in commodity taxes can improve social welfare. If one interprets the different consumption goods in the Atkinson-Stiglitz analysis as consumption in different periods, then the result may be interpreted

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

149

as calling exclusively for a progressive tax on labor income or, in their model equivalently (because there are no initial assets and no transfers to other individuals), a progressive tax on lifetime consumption. Even if the Atkinson-Stiglitz restriction on preferences is not satisfied, this violation does not necessarily imply that one would want to have higher taxes on future consumption than on current consumption, as would effectively be imposed by a capital income tax. A priori, a capital income subsidy is just as likely to be optimal as a capital income tax. Moving to a longer horizon, though, led to a third round of results, represented by the findings of Judd (1985) and Chamley (1986) that, for a system with taxes on labor and capital income, the long-run tax on capital income should be zero. Thus, the result calls for consumption tax treatment in the long run. This conclusion is quite general in some respects, not depending on individual preferences taking a particular form, for example. The intuition is also very simple. For any given set of individual preferences, we might improve efficiency by taxing consumption in different periods at different rates. Taxing future consumption more heavily than current consumption could be achieved by taxing capital income. But it is implausible that we would want to tax consumption more and more heavily as we move into later and later periods, as would be the case if positive capital income taxes continued to apply. Thus, at some point, regardless of the exact form of preferences, the capital income tax will have to converge to zero. Although Judd and Chamley derived their results in simple representative agent models without ability differences or progressive taxation, the logic would apply just as well in the Atkinson-Stiglitz framework: even if one didn’t want uniform consumption taxation, one wouldn’t want to have arbitrarily large distortions facing consumption at distant future dates. Further, the intuition associated with the Judd-Chamley result still resonates for finite horizon models, in that a positive capital income tax in every year would still impose a very high effective tax rate on consumption, say, 50 years in the future. Finally, analysis of tax reforms, rather than static comparisons of different tax systems, has highlighted an important component of consumption tax adoption, the taxation of consumption from existing wealth. As discussed by Auerbach, Kotlikoff and Skinner (1983), this tax on existing wealth accounts for a significant fraction of the efficiency gain in a transition from an income tax to a consumption tax. In summary, there have been a series of developments supporting the view that consumption taxes are more efficient than income taxes, or at least than a uniform income tax that imposes the same tax rate on labor and capital income. But there have also been many challenges to these conclusions.

2.1

CONSTRAINTS ON AGE-BASED TAXATION

As in simpler optimal tax models, the Atkinson-Stiglitz analysis assumes that each household has just one form of labor supply. Leaving aside the issue of how

150

TAX REFORM IN THE 21ST CENTURY

different members of a household should be treated at a given point in time, this assumption also ignores the fact that there is labor income in different periods. Even if consumption were separable from all forms of leisure, there is no presumption that leisure at different dates should be subject to the same tax schedule. One might think that age-specific labor income tax schedules would make sense. This point was emphasized by Kremer (2001) and in more recent analysis by Weinzierl (2008). In the life-cycle context, Erosa and Gervais (2002) find that, absent such age-specific labor income taxation, it is optimal to have a positive capital income tax. The intuition is that if we cannot increase the tax on a certain age cohort’s labor income, then we can tax this labor income indirectly by imposing a tax on capital income that hits this cohort’s future consumption. For this mechanism to work, the capital income tax must fall more heavily on the cohorts whose labor income we wish to tax more heavily, as would be the case if higher-saving cohorts are also those with lower labor supply responses. Erosa and Gervais find this condition to hold, although the capital income tax they find to be optimal is still much lower than the labor income tax.

2.2

OTHER SECOND-BEST TAX CONSIDERATIONS

In addition to constraints on age-based taxation, the literature has highlighted other examples in which positive taxation of capital income might improve the efficiency of the tax system. While, a priori, a violation of the Atkinson-Stiglitz separability condition might make a capital income subsidy just as likely to be optimal as a capital income tax, the capital income tax would be optimal if saving propensities are positively correlated with ability, as has been suggested to be the case (Saez, 2002). Also, taking general equilibrium effects into account in a model with skilled and unskilled labor, one would wish to tax capital income if capital is a relative complement to high-skilled labor, for this would improve the before-tax distribution of income (Krusell et al., 2000). Finally, taking human capital accumulation into account influences the design of capital income taxes. It is customary to think of human capital accumulation as facing a low or zero effective tax rate, since at least one important component of its cost, foregone earnings, is implicitly deductible. But with a progressive labor income tax (that is not age-dependent), a rising age-earnings profile will make future earnings face a higher tax rate than initial deductions, thereby imposing a positive tax rate on human capital accumulation that might call for some taxation of capital income to achieve a more even treatment of the two forms of investment.

2.3

DYNAMIC OPTIMAL TAXATION

Applying a static analysis like that in the Atkinson-Stiglitz paper to behavior over time ignores the fact that we observe the labor supply and consumption decisions of early periods before those of later ones. Thus, we might wish to implement a tax system

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

151

in future periods that is conditioned on current decisions, leaving aside the incentive for governments to deviate from a previously announced policy. Considering such a dynamic model, Golosov et al. (2003) found that the Atkinson-Stiglitz result about uniform commodity taxation still applied across commodities within a given period, but that a positive capital income tax was generally optimal. The intuition is that higher wealth makes it more difficult for the government to impose redistributive labor income taxation, because the cost to high ability individuals of not working hard is reduced by the presence of additional resources. Taxing capital income therefore weakens the self-selection constraints that the government’s tax system must satisfy, allowing it more scope for redistribution.

2.4

MARKET IMPERFECTIONS

All of the arguments for capital income taxation considered thus far relate to the design of an optimal tax system in an efficient, competitive economy. But additional arguments appear once important market imperfections are taken into account. For example, if returns to capital income are not efficiently pooled by capital markets, then capital income taxes may play a role in doing so.2 As discussed further below, this is one facet of the argument over the extent to which income and consumption taxes differ, related to the taxation of returns to risk-taking. A variant of the capital-income-taxes-as-insurance argument arises in cases where individuals self-insure by saving ‘too much’ to protect against the possibility of facing borrowing constraints in the future, in which case it is optimal to impose a positive tax on savings (Aiyagari, 1995). Yet another type of market imperfection involves the weakness of the market for private annuities, in which case some form of capital income taxation may be optimal. In this case, though, the most obvious approach would involve the taxation of estates or inheritances, essentially providing annuities by reducing the level of accidental bequests (Kopczuk, 2003). Finally, not all market imperfections point in the same direction. For example, Judd (1997) argues that capital goods markets are non-competitive, meaning that there is a preexisting distortion discouraging the use of capital in production. In this context, and starting from a model in which, under perfect competition, the optimal capital income tax would be zero, Judd finds that 2

One needs to exercise some care, though, to consider why capital markets are incomplete in the first place. As Weisbach (2005) points out, there are different reasons why individuals might hold undiversified positions. One might be adverse selection. For example, a small business whose owner seeks to lay off some of its ownership risk might be perceived as a ‘lemon’. In this, case, the tax system’s mandatory participation would provide a benefit, as the analysis above suggested. On the other hand, some of the lack of diversification might be due to moral hazard. For example, individuals who insure their human capital risk would have little incentive to work if there is no way to separate the effects on observed income of effort and stochastic shocks. A tax system that provided such insurance would have the same disincentive effects, making the insurance function unattractive.

152

TAX REFORM IN THE 21ST CENTURY

the optimal capital income tax is quite negative in a non-competitive environment calibrated to US evidence.

2.5

RETHINKING TAXES ON EXISTING WEALTH

As stressed above, a major source of efficiency gains in simulated transitions to a consumption tax arise from what effectively is a lump-sum tax on existing wealth. But whether one can or should count on these efficiency gains has been called into question (see, e.g. Kaplow, 2009). First, there is the practical question of whether any realistic tax transition would fail to provide generous transition relief to holders of existing wealth. If such relief is unavoidable, then the potential efficiency gains are substantially reduced (e.g. Auerbach, 1996). Second, the measured efficiency gains are based on the assumption that the capital levy is unanticipated and that it does not increase expectations of future levies. While one might argue that a transition to a consumption tax is sui generis – one cannot make a transition to a consumption tax a second time – and that the impact on future expectations should be small, one might alternatively argue that a government seen to act opportunistically in one instance might be more likely to do so in others. In any event, the extent to which an efficient tax on existing wealth is available is certainly subject to debate.

2.6

HOW LARGE IS THE CAPITAL INCOME TAX WEDGE?

Leaving aside the lump-sum tax on existing assets, the potential economic efficiency gains from a shift to consumption taxation come from eliminating the capital income tax wedge. But how big is this wedge? Numerical simulations have typically been based on a realistic market rate of return, but one strand in the literature suggests that this considerably overstates the capital income tax wedge. Following the analysis of taxation and risk developed in several papers, including Gordon (1985), Kaplow (1994), Warren (1996) and Weisbach (2005), one can show that, under a number of key assumptions, including efficient riskpooling, symmetric taxation, and scalability of asset holdings, the tax on excess returns to capital is neutral; it provides no insurance of the type discussed above, but it also imposes no burden. If the capital income tax wedge is simply the residual tax on the safe rate of return to capital, then the potential efficiency gains are small. Simulations in Auerbach (2009) suggest that this conclusion is true even with the capital levy in place, for the capital levy itself is of less value when the ‘asset’ it provides the government has such a low effective rate of return. If this analysis breaks down, the implications for efficiency will depend on which assumptions fail to hold. On the one hand, if the effective rate of return to capital is higher (as might be true if, as discussed above, there are non-competitive returns), then the capital income tax wedge is higher. On the other hand, if capital income taxation improves risk-pooling, as discussed above, then this reduces the deadweight loss from capital income taxation. Also, deviations from symmetric

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

153

taxation with full loss offset tend to increase the effective tax rate, even as the relevant rate of return to which the tax rate applies may be reduced (Auerbach, 2009). In short, one must trace through the effects of various complications in order to determine their net impacts on the potential efficiency effects of a shift to consumption taxation. Such impacts may well be important but, to date, they have received too little attention in the literature.

2.7

BROAD-BASED CONSUMPTION TAXES AND EFFICIENCY: A SUMMARY

The fact that capital income taxes distort consumption more and more over long horizons establishes a strong case against high capital income taxes. But the case for zero capital income taxes is less obvious. There are a variety of arguments in favor of maintaining some taxation of capital income on efficiency grounds. Still, the case for a Haig-Simons income tax base, grouping capital and labor income together to be taxed at the same rate, has little support on efficiency grounds. Indeed, I am aware of no efficiency-based arguments in the literature in favor of equal tax rates, except for those related to the (quite relevant) administrative problems of distinguishing capital income from labor income, particularly in the case of closelyheld businesses. In short, it is easier to make the case for taxing capital income, on standard efficiency grounds, than for taxing capital income and labor income equally. When specific consumption tax proposals are evaluated against the current income tax system, additional efficiency gains are often claimed, because the current system has many additional distortions not present under an ideal tax system, and because some of the consumption tax proposals would lower marginal tax rates through a reduction in progressivity. But neither type of additional efficiency gains is fairly attributed to a consumption tax reform, unless we want to do away with the current progressivity and added distortions and can do so only by adopting a consumption tax. Indeed, the uncertainty about the net efficiency benefits of a shift to consumption taxation have caused some to focus more on the potential efficiency gains that could be achieved under a shift to any broad-based tax. For example, Hubbard (1997) argues that one should think of the switch from our current tax system to a consumption tax in terms of two steps, the first a move to a clean, broadbased income tax that removes the inter-asset distortions of the current tax system, and the second a move from this reformed income tax to a consumption tax.

3

Rational Choice and Tax Design

Most analyses of the efficiency of a shift to consumption taxes consider the replacement of all or part of the income tax with a broad-based tax on consumption. Existing tax systems already have certain elements of consumption taxation, and this has led some to argue that such a transition is already underway. Under the

154

TAX REFORM IN THE 21ST CENTURY

US tax system, for example, a considerable share of household assets is held in some sort of tax-sheltered asset, whether a retirement account, an education savings account, or an owner-occupied residence.3 Under standard economic analysis, however, such hybrid systems are typically less efficient than a true combination of broad-based income and consumption taxes. First, contributions to tax-sheltered accounts are capped, so that some households may face the full rate of income tax at the margin even if a large share of their assets are in tax-sheltered accounts. Second, contributions to accounts can come from previously accumulated wealth, so that the capital levy associated with the consumption tax is completely absent. Finally, the full deductibility of interest combined with reduced taxation of capital income encourages borrowing to invest in tax-favored assets, rather than saving. But there has been considerable recent thought expressed in the literature that deviations from standard economic analysis are needed to explain various aspects of observed saving behavior, in particular the lack of adequate saving on the part of many individuals approaching retirement. These deviations could potentially influence conclusions about the relative distortions of income and consumption taxes, and also shift the focus more toward finer details of tax system design.

3.1 ARE CAPITAL INCOME TAXES INEFFICIENT? With many individuals possibly saving too little for retirement, one might think that lower taxes on capital income would provide even larger welfare gains than under standard assumptions, by encouraging people to overcome obstacles to saving. But encouraging saving might be ineffective if individuals anticipate that additional saving will simply be squandered by a future self. Depending on the nature of the departure from standard rational choice assumptions, one might be led to tax capital income (Bernheim and Rangel, 2005).

3.2 THE TIMING OF TAX INCENTIVES Under standard assumptions, it does not matter whether tax incentives to save are provided at the beginning or the end of the savings period, i.e. whether the pattern of taxation follows the TEE or ‘tax prepaid’ approach or the EET or ‘tax postpaid’ approach. Both approaches provide a zero effective tax rate on saving when tax rates are constant over time. But given evidence that contributions to deductible Individual Retirement Accounts were greater among households with additional tax due on April 15 (Feenberg and Skinner, 1989), there is some possibility that the immediate deduction is more salient than the eventual tax exemption. 3

According to the Federal Reserve Board’s Survey of Consumer Finances, in 2004 primary residences and retirement accounts represented 74 per cent of net worth of US families in the bottom 90 per cent of the income distribution and 89 per cent of net worth of families in the bottom 90 per cent of the net worth distribution.

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

155

3.3 ACCOUNT RESTRICTIONS The current practice of earmarking tax-sheltered accounts for different purposes (retirement, education, medical expenses, etc.) is inconsistent with the implications of standard models, because it artificially restricts the flexibility of individuals who save. But some models of how individuals deal with self-control problems suggest that maintaining such an artificial separation helps to ‘lock’ funds away from temptation through a form of ‘mental accounting’ (Shefrin and Thaler, 1988). Our knowledge about the extent to which contributions to tax-sheltered retirement accounts constitute new saving is still imperfect. But the possibility that substantial new saving is occurring even when individuals could fund the accounts by transferring assets seems very much at odds with the predictions of standard models of intertemporal choice. This possible existence of new saving may also relate to such mental accounting behavior, with individuals imposing barriers on themselves against transferring assets from one type of account to another. Nor would another type of restriction, on withdrawals from accounts, arise from standard assumptions. However, such restrictions are welfare-improving in models with time-inconsistent preferences, in which individuals wish to bind their ‘future selves’ from behaving in a profligate manner (Laibson, 1997).

3.4

SETTING DEFAULT BEHAVIOR

With full information about their options for saving, rational households would not seem likely to be affected by the default options of employer-sponsored retirement savings plans. But experimental evidence suggests that default options can exert a very strong influence on the level and pattern of contributions beyond the short run (Madrian and Shea, 2001). Our knowledge on this point is clear enough that the US President’s Advisory Pattern on Federal Tax Reform, (2005, pp. 118–19), structured its ‘Save at Work’ plans to have various ‘autosave’ features aimed at increasing the saving of passive employees.

3.5

CONSUMPTION TAXES AND RATIONAL CHOICE: SUMMARY

Saving behavior may not conform to the predictions of standard models, and these deviations have potentially important implications for tax system design. But the nature of the deviations is also important. While there is a tendency to think that an underweighting of the future makes it even more important to encourage saving, this is not necessarily true. And, if saving is to be encouraged, it is not clear whether a broad-based consumption tax is the way to do so. For example, the consumption tax’s broad-based incentive to save may lead to less saving than a series of specific savings schemes if mental accounts play a role in encouraging saving. On the other hand, the confusion that such schemes can induce may reduce their potential benefits if imperfect information is a significant cause of under-saving.

156

4 4.1

TAX REFORM IN THE 21ST CENTURY

Distributional Equity CHOOSING THE HORIZON FOR EVALUATION

There is more than one way in which one might evaluate the relative equity of income and consumption taxes. Perhaps the simplest is in terms of the tax burden on annual income. It is a simple implication of consumption smoothing that consumption will fluctuate less than income, so that the consumption tax base will appear less progressive when evaluated on an annual basis than on a lifetime basis. Even in this case, a more progressive rate consumption tax rate schedule could be used to offset the relative regressivity of the base. But it is also not clear why one would wish to evaluate progressivity on an annual basis. One argument, motivated by some of the evidence from the behavioral economics literature, might be that households are myopic and that the government should use their short horizons in evaluating the tax system. But this argument is weak. First, the behavioral economics literature also stresses the time inconsistency of individual preferences, i.e. that some individuals may ignore the future today but might wish in the future that they had not. Should governments accept the current judgments of individuals when these assessments will change in the future? Second, complete myopia on the part of households is inconsistent with the observed patterns of consumption smoothing. If households vary in the extent to which they are forward looking, perhaps with higher-income households exhibiting less myopia or greater patience, then this amounts to an issue already discussed above, the case for taxing capital income to the extent that saving propensities are related to ability. But it is still not an argument for using annual income as a measure of ability to pay.4 The case for an annual frame of reference seems to require the position that wealth accumulation conveys benefits comparable to current consumption. Moreover, these benefits of wealth accumulation must be distinct from those derived from future consumption, for such benefits (and their anticipated value) will be reduced by the prospect of future consumption taxes. One might also argue that wealth accumulation generates negative externalities through the concentration of wealth and power. There are two problems with this argument, though. First, to the extent that the power conveyed by wealth is based on the power to consume,5 the previous analysis applies. Second, if there are negative externalities to wealth concentration, then this would call not only for capital taxes, but also for capital subsidies for those of more modest wealth. More generally, if household welfare is directly influenced by inequality, perhaps by the consumption one is able to achieve 4

5

Shaviro (2007) discusses various reasons why the lifetime perspective might be inappropriate, including myopia and also incompleteness of capital markets, and concludes that while this might weaken the equity case for pure consumption taxation, it does not advance the case for income taxation. This would be the case even if the wealth were used to buy influence through private transfers, to the extent that the transfers were ultimately used for consumption.

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

157

relative to some reference group, then this has implications not simply for capital income taxation, but also for the design of consumption taxes, particularly if certain types of consumption are associated with wealth.

4.2

LIFETIME INCIDENCE

Even from a lifetime perspective, there is concern that consumption taxes are less progressive than income taxes. Here, one must specify which consumption tax. The evidence, based on numerical simulation results, suggests that consumption tax proposals can achieve progressivity comparable to the current tax system. Simulations in Altig et al. (2001) find that, in the long run, the distributional effects of a switch from the income tax to the X tax would be quite neutral, with very similar percentage increases in lifetime welfare (measured as a fraction of lifetime labor endowments) for all lifetime income classes, ranging from the top 2 per cent of the lifetime income distribution to the bottom 2 per cent. To the extent that there are slight differences in impact, these differences favor those individuals with lower lifetime income. Indeed, roughly the same conclusions hold even when the model simulations incorporate bequests that conform to the actual, highly concentrated empirical bequest distribution. This finding may seem surprising, given that bequests are concentrated among the highest lifetime income groups, in the model as well as in reality. But, at least for the progressive consumption tax analyzed by Altig et al. (2001), bequests are effectively taxed, because the X tax is collected at the business level and hence its future burden on cash flows should be capitalized into the value of bequeathed assets.6 It should also be remembered that these are long-run effects. If a consumption tax is implemented without transition relief, for example, the effects of the capital levy on initial generations could be quite progressive. This would be even more the case to the extent that the assets are inherited, given the concentration of bequests and inheritances. It seems hard, then, to make a general case against consumption taxation (as opposed to certain consumption tax proposals, such as a national retail sales tax) on grounds of distributional equity, unless one is prepared to ignore a taxpayer’s future taxes or the taxes that will be capitalized into the assets that the taxpayer owns. Until taxes are actually paid, of course, a subsequent tax reform might eliminate such future liabilities. Perhaps this is the cause of concern for some. But a tax reform might also increase future tax liabilities, so it is not obvious why the deferral of taxes should be discounted more than by the normal discount rate. Whether the political process would deliver a consumption tax maintaining the current tax system’s degree of progressivity is a distinct question. Also, even in cases in cases where the entire distribution of burdens is not kept more or less constant (as in the X-tax simulations referred to above), there may not be a clear ranking in terms of progressivity. The standard flat tax approach, for example, implemented 6

See Auerbach (2009) for further discussion.

158

TAX REFORM IN THE 21ST CENTURY

with a high threshold for initial taxation, can reduce burdens at both the top and the bottom of the income distribution. The stronger the relative weight given to those toward the bottom of the income distribution, the more progressive such an outcome would appear.

5 Tax Implementation and Administration There are many detailed issues that arise in consideration of particular consumption tax proposals. Many involve how to manage the transition to a new tax system, in particular how to treat existing assets and liabilities. Such transition provisions can become very complicated, particularly if the government is attempting to offer taxpayers some limited form of transition relief.7 Transition considerations are clearly important, but I will focus here on issues relating to ongoing tax system administration. I will also not get into one topic that has already received considerable attention, the relative merits of different types of value-added taxes, in particular the subtraction method from which the flat tax is adapted versus the traditional credit invoice method now in practice. Nor will I discuss which taxes might be creditable under foreign tax treaties, a question that is of considerable importance but essentially devoid of economic content and not informed by economic analysis. Instead, I will focus on a few selected issues that have received attention recently in the literature.

5.1

CROSS-BORDER TRANSACTIONS

Some consumption taxes, like the standard VAT, are implemented on a destination basis, providing border adjustments by relieving tax on exports and imposing it on imports. Others, like the flat tax, would be imposed on an origin basis, with no border adjustments. The distinction between these two approaches has important implications for enforcement. In particular, the origin-based VAT, like source-based income taxes, is subject to the use of transfer pricing by multinational enterprises to shift the tax base among countries. A destination-based tax would not be subject to this problem, but would require that the location of consumption be determined, an issue of at least some concern, particularly given the growth of electronic commerce.8 At the political level, there has been considerable debate about the effects of border adjustments on international trade. At least among countries with flexible exchange rates, however, border adjustments alone should have minimal impact on 7 8

See, for example, Warren’s (1995) discussion of the provisions of the ‘USA Tax’ proposal for a personal expenditure tax developed in the 1990s in the United States. Indeed, this has been an issue of growing importance at the US state level, where states have found it difficult to collect sales tax on internet sales to residents by out-of-state vendors, the problem arising in part from legal decisions limiting the ability of states to require compliance on the part of these vendors.

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

159

trade balances and capital flows, because of offsetting exchange rate adjustments. In particular, the introduction of border adjustments to an origin-based VAT should call forth an exchange-rate appreciation that approximately neutralizes the initial improvement in competitiveness attributable to the border adjustments.9 The primary economic impact should be on asset values, for the exchange rate appreciation would make assets held abroad less valuable at home, and make assets held at home more valuable abroad. These revaluations depend on gross international positions, so even countries with relatively small net external asset balances could experience large revaluations. Auerbach (2009) calculated the potential revaluations for the United States in 2004 (when the US international asset position was small and negative). According to the Bureau of Economic Analysis, at the end of 2004, foreign-owned assets in the United States totaled $11.5 trillion and US-owned assets abroad were $9.1 trillion. The border adjustment of a domestic VAT at a rate, say, of 30 per cent, would require a 30 per cent depreciation of foreign currencies relative to the dollar – a $2.7 trillion loss for US holders of foreign assets, and a gain of $4.9 ($11.5/0.7 – $11.5) trillion for foreign asset holders, relative to a 30 per cent VAT with no border adjustment. As a frame of reference, the net national debt of the United States outstanding at the end of 2004 was $4.3 trillion. The previous analysis applies even in cases where countries manage their currencies, to the extent that they target real exchange rates, i.e. the cost of the commodities they produce relative to the cost of commodities produced abroad. That is, if a country wishes to peg its currency’s real value at a certain level, then it would need to devalue its currency to offset border adjustments adopted abroad. In some cases, however, nominal exchange rates are fixed, the extreme case being when there is a common currency, as in the European Union. With fixed exchange rates, the necessary adjustments would have to take place in domestic or foreign product and labor markets. For example, if a country implemented border adjustments, these could be offset by increased domestic wages and prices. But short-run flexibility is more limited for wages and prices than for exchange rates, so border adjustments could well have short-run effects on trade within fixed exchange rate zones. The story is different, however, in the case of direct taxes. For example, if a personal expenditure tax is imposed, and applies to assets held abroad as well as domestically, then investments abroad would be subject to the same cash flow tax treatment that effectively is imposed by the combination of border adjustments and exchange rate appreciation: a tax deduction for outflows and a tax on inflows and, assuming no transition relief, a levy on existing assets abroad. But this would occur without any exchange rate appreciation.10 9 10

See Auerbach (2009) for further discussion. Note that the offsetting subsidy to existing domestic assets held by foreigners would be absent under this formulation, making the tax base broader.

160

5.2

TAX REFORM IN THE 21ST CENTURY

FINANCIAL TRANSACTIONS

A comprehensive consumption tax would include a tax on net cash flows from both real and financial assets. However, for the economy as a whole, financial assets and liabilities largely offset each other, so it is customary, particularly in the case of indirect taxation, to ignore financial transactions. For example, the typical VAT imposed on a nonfinancial company ignores that company’s borrowing and lending activities. There would, of course, be short-run distributional effects of extending the tax to include financial assets and liabilities, to the extent that the interest rates being paid are fixed. Even in the longer run, though, how to treat financial transactions is a question of increasing relevance, and one could consider including them in the consumption tax base, to adopt what the Meade Committee (Institute for Fiscal Studies, 1978) characterized as an R+F base rather than an R base. An administrative advantage of the R+F base is that use of this base eases the task of taxing the income generated by the financial services sector. An R+F base VAT applied to a financial services company would automatically tax interest-rate spreads that compensate the company for the services it renders. Another advantage of the R+F base is that it would eliminate the incentives of an integrated company offering both real and financial services to shift profits from real to financial activities. Unfortunately, as it eliminates one distinction, the R+F base introduces another. As the Meade Committee’s report (Chapter 12) observed, the transactions of the business sector are related by an identity, R+F+S = 0, where S equals net inflows to the firms from sales of corporate shares and dividend payments, i.e. the net flows to firms from their shareholders. Under an R base tax that ignores financial transactions between the business and household sectors, both F and S are omitted from the tax base. Thus, transactions between a firm’s creditors and shareholders are treated identically. Just as new share issues and dividends are excluded from the tax base, so are borrowing and interest payments. But, once F is included in the tax base, transactions with creditors are subject to a cash-flow tax – with taxes on borrowing and deductions for interest expense and repayment of principal – while transactions with shareholders are still ignored. How serious a problem this distinction is, in how the sources of funds are treated under the R+F base, depends on the extent to which firms can manipulate payments between debt and equity, for example by issuing debt and equity to the same party and overstating deductible interest payments while understating nondeductible dividends. The problem is similar to that under the R base of shifts on the output side between the firm’s real and financial activities, so the decision of whether to group financial flows with the firm’s transactions with its customers, or with the firm’s transactions with its shareholders, depends in part on where the problem of manipulation is greater. In practice, the standard credit-invoice value-added-tax approach follows the R base, rather than the R+F base, with financial services left out or dealt with

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

161

through some alternative mechanism (Ebrill et al., 2001, pp. 94–98). This requires drawing a line between real and financial companies. The same R base approach characterizes the flat tax and the X tax, which are basically variants of the VAT. The natural approach for household-based taxes is to cover both real and financial assets. On the other hand, many variants are conceivable. For example, McLure and Zodrow (1996) advocate implementing a cash-flow tax with an R+F base at the business level and yield exemption (i.e. labor income taxation) at the household level. They argue that it makes sense to impose the tax at the business level, to target potential tax avoidance behavior most directly and to shift the costs of compliance away from less sophisticated household-level taxpayers.

5.3 THE LOCATION OF TAX COLLECTION In addition to the question of taxpayer sophistication and the related question of enforcement, there are other issues that come up in thinking about whether a consumption tax ought to be imposed on businesses or households. For example, in cases where the business-household borderline is unclear, taxes imposed solely at the business level may miss a substantial amount of economic activity. Collecting at least some component of tax at the individual level is also necessary if one is to have a progressive tax, given the extreme inefficiency of using rate variations in indirect consumption taxes to accomplish this. But, as the design of the flat tax, the X tax, and the hybrid tax of McLure and Zodrow just discussed illustrate, the wage portion of the tax base can be shifted to the household level while the remaining cash flows remain in the business tax base. The level at which the tax is collected and, in particular, whether the tax is imposed as an indirect tax or a direct tax, would also affect the measured price level. For example, a value-added tax, or a retail sales tax, added to factor incomes, would have a higher measured price level than if the same tax were collected from individual consumers. This difference simply reflects the standard convention for measuring the price level, but it could have real effects to the extent that government transfer payments and wages are indexed to the price level. Finally, some have argued that indirect consumption taxes levied on purchases, rather than taxes assessed at the household level, are a more effective method of collecting taxes from individuals who do not report income or pay tax under the income tax. The logic is that these individuals currently pay no tax, but would be hit by sales tax or VAT on their purchases. But the underlying logic itself is flawed, basically because it ignores the fact that the tax evader is not simply a consumer, but also a producer, and is no more likely to pay a value added tax than an income tax.11

11

See Hines (2004) for further discussion.

162

TAX REFORM IN THE 21ST CENTURY

5.4 THE (MIS)ALIGNMENT OF INCOME AND TAX LIABILITIES It has long been understood that a consumption tax imposes a zero net liability on saving via offsetting taxes on saving and dissaving. A tax on income, correctly measured, on the other hand, imposes a positive tax on income at the time the income is earned. Thus, changes in income tax rates over time (again, assuming income is correctly measured) affect only the tax rate on income in the periods directly affected. Changes over time in consumption tax rates, however, can impose quite positive or negative tax rates on income, because of the dependence on offsetting tax effects at different dates. This condition led Bradford (1998) among others to suggest replacing the immediate deduction of investment with the procedure of carrying the basis of investment forward with interest and taking depreciation deductions in accordance with the timing of economic depreciation. While these deductions would have the same present value as immediate expensing with tax rates constant, their timing would eliminate the distortion of investment decisions: marginal investments would face a zero effective tax rate regardless of the path of tax rates, as only excess returns would be subject to taxation in any period. This ‘basis with interest’ approach thus achieves the zero marginal tax that would also be delivered by simple yield exemption, but maintains the tax on excess returns. However, this approach also has a major drawback of requiring that investors keep track of asset bases and calculate depreciation allowances over time, and it has not gained favor as yet in the formulation of actual tax policy.12

6

Conclusions

Arguments in favor of replacing the income tax with a consumption tax are not as ‘strange’ as they were 50 years ago. If one were to summarize developments in the literature since then, the arguments for the consumption tax have become more measured as they have become more convincing. Conclusions about efficiency based on simple economic models have waned in influence as the problems of transition have been studied and as various benefits of capital income taxation have been recognized. At the same time, we have come to understand that realistic consumption taxes can be progressive, and to recognize the weakness of the case for equal taxation of labor and capital income. Much more research is needed to determine what the tax structure should look like, but we can be increasingly confident that it will not be a pure income tax or a pure consumption tax. 12

Indeed, rather than attempting to make a cash flow tax look more like an income tax, Auerbach and Bradford (2004) adopted the opposite approach of considering how to make an income tax look more like a cash flow tax. They developed a method of using the cash flow tax with rising tax rates over time to simulate an income tax, suggesting that the advantages of eliminating basis could outweigh the disadvantage of susceptibility to time-varying tax rates, particularly with financial innovation making the identification of specific assets more difficult.

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

163

References Aiyagari, S.R. (1995), ‘Optimal Capital Income Taxation with Incomplete Markets, Borrowing Constraints, and Constant Discounting’, Journal of Political Economy 103: 1158. Altig, D., A.J. Auerbach, L.J. Kotlikoff, K.A. Smetters, and J. Walliser (2001), ‘Simulating Fundamental Tax Reform in the United States’, American Economic Review 91: 574. Atkinson, A.B., and J.E. Stiglitz, (1976), ‘The Design of Tax Structure: Direct versus Indirect Taxation’, Journal of Public Economics 6: 55. Auerbach, A.J. (2009), ‘The Choice between Income and Consumption Taxes: A Primer’, in A. Auerbach and D. Shaviro (eds), Institutional Foundations of Public Finance: Economic and Legal Perspectives (Cambridge, MA: Harvard University Press) forthcoming. Auerbach, A.J. (2008), ‘Tax Reform in the 21st Century’, in J. Diamond and G. Zodrow (eds), Fundamental Tax Reform: Issues, Choices, and Implications (Cambridge, MA: MIT Press). Auerbach, A.J. (1996), ‘Tax Reform, Capital Allocation, Efficiency and Growth’, in H. Aaron and W. Gale (eds), Economic Effects of Fundamental Tax Reform (Washington, DC: Brookings Institution). Auerbach, A.J. (1994), ‘Public Sector Dynamics’, in J. Quigley and E. Smolensky (eds), Modern Public Finance (Cambridge, MA: Harvard University Press). Auerbach, A.J., and D.F. Bradford (2004), ‘Generalized Cash-Flow Taxation’, Journal of Public Economics 88: 957. Auerbach, A.J. and J.R. Hines, Jr. (2003), ‘Perfect Taxation with Imperfect Competition’, in S. Cnossen and H. Sinn (eds), Public Finance and Public Policy in the New Century (Cambridge, MA: MIT Press). Auerbach, A.J., L.J. Kotlikoff and Jonathan Skinner (1983), ‘The Efficiency Gains from Dynamic Tax Reform’, International Economic Review 24: 81. Banks, J.W. and P.A. Diamond (2008), ‘The Base for Direct Taxation’, MIT Department of Economics Working Paper No. 08-11, March 20. Bernheim, B.D., and A. Rangel (2005), ‘Behavioral Public Economics: Welfare and Policy Analysis with Non-Standard Decision Makers’, NBER Working Paper 11518, July. Bradford, D.F. (1998), ‘Transition to and Tax Rate Flexibility in a Cash-Flow Type Tax’, in J. Poterba, ed., Tax Policy and the Economy 12: 151. Bradford, D.F. (1986), Untangling the Income Tax (Cambridge: Harvard University Press).

164

TAX REFORM IN THE 21ST CENTURY

Chamley, C. (1986), ‘Optimal Taxation of Capital Income in General Equilibrium with Infinite Lives’, Econometrica 54: 607. Corlett, W.J. and D.C. Hague (1953), ‘Complementarity and the Excess Burden of Taxation’, Review of Economics Studies 21: 21. Ebrill, L., M. Keen, J-P. Bodin, and V. Summers (2001), The Modern VAT (Washington: International Monetary Fund). Erosa, A., and M. Gervais (2002), ‘Optimal Taxation in Life-Cycle Economies’, Journal of Economic Theory 105: 338. Feenberg, D.R., and J. Skinner (1989), ‘Sources of IRA Saving’, in L. Summers, ed., Tax Policy and the Economy 3: 25. Feldstein, M. (1978), ‘The Welfare Cost of Capital Income Taxation’, Journal of Political Economy 86: S29. Golosov, M., N. Kocherlakota, and A. Tsyvinski (2003), ‘Optimal Indirect and Capital Taxation’, Review of Economic Studies 70: 569. Gordon, R.H. (1985), ‘Taxation of Corporate Capital Income: Tax Revenues versus Tax Distortions’, Quarterly Journal of Economics 100: 1. Hall, R., and A. Rabushka (1985), The Flat Tax (Stanford, CA: Hoover Institution Press). Hines, J.R., Jr. (2004), ‘Might Fundamental Tax Reform Increase Criminal Activity?’ Economica 71: 483. Hubbard, R.G. (1997), ‘How Different are Income and Consumption Taxes?’ American Economic Review 87: 138. Institute for Fiscal Studies (1978), The Structure and Reform of Direct Taxation (London: Allen and Unwin). Judd, K.L. (1997), ‘The Optimal Tax Rate for Capital Income is Negative’, NBER Working Paper 6004, April. Judd, K.L. (1985), ‘Redistributive Taxation in a Simple Perfect Foresight Model’, Journal of Public Economics 28: 59. Kaplow, L. (2009), ‘Capital Levies and Transition to a Consumption Tax’, in A. Auerbach and D. Shaviro (eds), Institutional Foundations of Public Finance: Economic and Legal Perspectives (Cambridge, MA: Harvard University Press) forthcoming. Kaplow, L. (1994), ‘Taxation and Risk Taking: A General Equilibrium Perspective’, National Tax Journal 47: 789. Kopczuk, W. (2003), ‘The Trick Is to Live: Is the Estate Tax Social Security for the Rich?’ Journal of Political Economy 111: 1318.

ALAN AUERBACH: INCOME OR CONSUMPTION TAXES?

165

Kremer, M. (2001), ‘Should Taxes be Independent of Age?’ unpublished manuscript, Harvard University. Krusell, P., L.E. Ohanian, J-V. Ríos-Rull, and G.L. Violate (2000), ‘Capital-Skill Complementarity and Inequality: A Macroeconomic Analysis’, Econometrica 68: 1029. Laibson, D. (1997), ‘Golden Eggs and Hyperbolic Discounting’, Quarterly Journal of Economics 112: 443. Madrian, B., and D. Shea (2001), ‘The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior’, Quarterly Journal of Economics 116: 1149. McLure, C.E., Jr., and G.R. Zodrow (1996), ‘A Hybrid Approach to the Direct Taxation of Consumption’, in M. Boskin, ed., Frontiers of Tax Reform (Stanford, CA: Hoover Institution Press). Musgrave, R.A. (1959), The Theory of Public Finance (New York: McGraw-Hill). President’s Advisory Panel on Federal Tax Reform, 2005, Simple, Fair, and ProGrowth (Washington, DC: US Government Printing Office). Saez, E. (2002), ‘The Desirability of Commodity Taxation under Non-Linear Income Taxation and Heterogeneous Tastes’, Journal of Public Economics 83: 217. Shaviro, D. (2007), ‘Beyond the Pro-Consumption Tax Consensus’, Stanford Law Review 60: 745. Shefrin, H.M., and R.H. Thaler (1988), ‘The Behavioral Life-Cycle Hypothesis’, Economic Inquiry 26: 609. Warren, A.C. (1996), ‘How Much Capital Income Taxed Under an Income Tax Would Be Exempt Under a Cash-Flow Tax?’ Tax Law Review 52: 1. Warren, A.C. (1995), ‘The Proposal for an “Unlimited Savings Allowance”’, Tax Notes 68: 1103-08. Weinzierl, M. (2008), ‘The Surprising Power of Age-Dependent Taxes’, unpublished manuscript, Harvard University. Weisbach, D.A. (2005), ‘The (Non) Taxation of Risk’, Tax Law Review 58: 1.

Chapter 6

Consumption Taxes and Risk Revised Jane G. Gravelle* 1

Introduction

The three broad-based taxes – income, consumption, and wage – are typically compared in the following way: the income tax is a tax on the return to investment and wages, the wage tax eliminates the tax on the return to investment, and the consumption tax adds to the wage tax a lump-sum tax on capital already in place. The lump-sum tax is a large part of what gives the consumption tax its superiority in minimizing distortions, both compared to an income tax and a wage tax. The revenue from this lump sum tax, which continues over time, offsets the cost of providing a deduction for saving or new investment, and permits a tax base that, in the United States, is over 90 per cent of the size of the income tax base. This larger base, as contrasted with a wage base that is probably about 75 per cent as large as the income tax, requires a much smaller increase in tax rates on wages. As a result, the distortions that arise from taxing capital income are eliminated and the additional distortion from taxing wages is limited. As we shall see subsequently, there are some other factors at work, regarding timing, that produce larger effects on the capital stock in some models of the consumption tax. The lump-sum tax that arises from the consumption tax, while a contributor to reducing distortions, presents an enormous barrier to replacing the income tax with a consumption tax, an idea that has been discussed, but not acted on, for many years in the United States. For a consumption tax in the form of a national retail *

Congressional Research Service, US Library of Congress; Senior Fellow, Taxation Law and Policy Research Institute, Monash University. The views in this paper do not reflect the views of the Congressional Research Service or the Library of Congress.

168

TAX REFORM IN THE 21ST CENTURY

sales tax or a value-added tax, the shift in the source of taxation means that taxes must be accommodated in either higher prices or lower nominal wages. The need to accommodate a large price increase (or suffer the consequences of recession, given sticky wages) creates serious macro-economic transition problems. These problems do not arise in a Hall-Rabushka flat-tax approach, where firms deduct wages and the bulk of the tax continues to be collected directly from workers. It does, however, mean the loss of deductions for depreciation, inventory, and basis on the sale of assets, which would also create serious economic and political barriers. Price adjustments are likewise avoided under the personal expenditure tax. And because it does not require a price adjustment, the lump sum tax on old capital falls only on equity assets and not debt. Aside from the lump-sum tax on old capital, the treatment of the return to new investment under a wage tax versus a consumption tax is differentiated by timing. Under wage-tax treatment, no tax is imposed on the investment return. Under the consumption tax treatment, the government provides a deduction up-front for the investment, and, when the investment is converted into consumption, taxes the value (principal and interest). The timing of tax payments for new investors is, therefore, quite different. How is this analysis affected when we introduce risk in the return to assets? While there has been discussion about how to view consumption, wage, and income taxes with this type of risk, most studies of the effects of consumption taxes on the economy are based on models with certainty. Even if the case where there is uncertainty, that uncertainty has been with respect to wage income and life span, leading to a precautionary savings motive, a more muted capital income response, and a smaller welfare gain from switching to a consumption tax. These models, including the model with risk, are reviewed in Bernheim (2002). To address the issue of risk in asset returns, we first review the general issues relating to this risk in the case of income tax, and then turn to the consumption tax. We then consider two existing general equilibrium models of these alternative taxes: the Gordon (1985) model, which examines the economy-wide effects of a tax on the risk component of the investment return, and the Kaplow (1994) model where the government is an active participant in the asset markets. Finally, we consider risk in asset returns in a general equilibrium model where the government addresses its budget constraint by replacing one tax with another, the differential incidence approach most commonly taken in studies of the potential efficiency gains under the consumption tax. We conclude that, while risk is important to take into account in comparing an income tax with either a wage or a consumption tax, it is not the issue of risk that differentiates a consumption tax from a wage tax, but either the issue of timing and income effects. Taking asset risk into account, in a model where agents optimize, appears to greatly weaken the efficiency case for moving to a consumption tax, though at the cost of significant distributional consequences. If agents are less

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

169

sophisticated and save according to rules of thumb, shifting to a consumption tax could potentially be damaging to their welfare, not because of risky aspects but because of timing effects.

2

Asset Risk and the Income Tax

Dating back to Domar and Musgrave (1944), there have been many explorations of the effects of income taxation on risky assets. Domar and Musgrave show that if a tax is imposed only on the risk premium or if the risk-free return is zero, the investor can restore his or her pretax utility by increasing the share of risky assets in the portfolio. In effect, the government is sharing in the investment by taking a share of the return and the variance. If the government imposes a tax of 50 per cent on the risky return, the investor doubles the amount of risky assets in his portfolio, so that his after-tax risk and variance are unchanged. If such an investment shift could be made, the tax is clearly costless to the individual. And, if the government can absorb risk better than private markets, or allocate it more efficiently, this tax can contribute to economic efficiency. There were reservations about these Domar and Musgrave findings, which are well known and are reviewed by Sandmo (1985) and Poterba (2002). The analysis holds only for a proportional tax with full loss offset, which is not characteristic of all income tax regimes. The corporate tax, an important part of the capital income tax, is, however, a proportional tax; and in practice, losses may frequently be offset, either because they can be offset against other income or because of carryback and carryforward rules. In the case of capital gains taxes, often cited as a case where loss offset is highly imperfect, investors may elect when to pay taxes by realizing assets. They therefore have the freedom not only to match capital gains and losses in the current period, but also to delay the receipt of gains relative to losses in order to minimize tax payments. Indeed, the power of this election to pay taxes is the main reason for restrictions on the offset of losses against ordinary income. Moreover, since 1986, the capital gains tax rate has been roughly proportional for most of the higher-income individuals who hold these types of assets. The same flat-rate treatment is also currently applied to dividends. Therefore, the major area where the problems of progressivity and imperfect loss offset may be serious is in small businesses, although even here losses can be offset against other income. In addition, if taxes apply on the risk-free part of the return, the results are affected by opposing substitution effects (encouraging more risky assets) and wealth effects (likely encouraging less). The question of where the risk goes (is it absorbed by the government or redistributed) is also an important issue, as is the general limitation of these initial analyses that did not place the findings in a general equilibrium context. Subsequently, there were attempts to consider general equilibrium effects of the taxation and risk reduction features of the income tax. If the government actually

170

TAX REFORM IN THE 21ST CENTURY

absorbs the risk (because it is better at spreading risk) or can allocate risk more efficiently, a tax may add to rather than detract from economic efficiency. The insights from the analysis of income taxation with risk bear a close resemblance to the insights regarding the difference between consumption taxes and wage taxes – the government is entering as a partner in the investment, sharing in returns and risk. In the case of the consumption tax, as distinct from the wage tax, the government also enters as a partner, sharing investment returns and risk. The basic bottom line of the Domar-Musgrave analysis that is relevant to the issue at hand is that the burden of the tax on the risky return, which tends to account for the bulk of asset returns, is offset by the reduction in variance, making the actual burden of the tax considerably less.

3

Risk and Consumption Taxes

As noted above, no taxes apply to investment returns under a wage tax, but such returns along with principal are taxed when consumed under a consumption tax. In a world of certainty, this tax upon distribution and consumption is exactly equal in present value, given a proportional tax, to the value of deducting the cost of the investment initially. When the US Treasury Blueprints study (1977) proposed a model consumption tax, they allowed taxpayers the option of choosing a cash-flow approach (up-front deduction with tax on payment) or a wage tax approach (exemption of income) for passive investments. Allowing that option was based on the argument that these two treatments were equivalent in present value. Some remnant of that notion remains in the US tax system, where eligible taxpayers can elect a traditional individual retirement account (IRA) or, alternatively, a Roth IRA where there is no up-front deduction and no returns are taxed. The same option was recently extended to employer savings plans (401(k) plans). In an uncertain world, however, it has been claimed that the consumption tax is very different from the wage tax, and closer to the income tax. In commenting on expenditure tax design, and on the Treasury Blueprints option in general, Michael Graetz (1980, pp. 172–3) states that among the factors that must apply for a consumption tax to be equivalent to a wage tax are: (1) no initial wealth, (2) bequests do not exist, or are taxed as consumption in the final tax return, (3) tax rates are not progressive and do not change over time, (4) perfect capital markets with no uncertainty; all can borrow and lend unlimited amounts at a risk-free rate, and (5) all income is either wage income or income accumulated during or after the initial period. Many of Graetz’s points are straightforward: if there is initial wealth or bequests or non-wage income, then the present value of lifetime consumption does not necessarily equal the present value of lifetime wages. The bequest issue is actually a part and parcel of the lump-sum tax on old capital, a tax that is collected as old

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

171

capital is consumed; with bequests, that consumption of old capital is delayed and spread to future generations, although how that burden is borne depends on bequest motivation and behavior. A constant proportional tax rate is necessary to prevent positive or negative taxes on earnings under a consumption tax, if the tax upon deduction differs from the tax on consumption taxes may be positive (if tax rates rise) or negative (if tax rates fall). But it is (4) that is the focus of our concern. Graetz does not expand a great deal on this concept, except to note that the consumption tax approach would narrow the difference between lucky and unlucky investors to the extent that speculative investment opportunities were limited. Bradford (1986) makes a case that the lack of taxation of such outcomes under the wage tax (which exempts the earnings of the lucky as well as the unlucky) is not an important difference for equity purposes in distinguishing the ex ante (wage-type) treatment and ex post (consumption-type) treatment. He discusses as an example the purchase of a lottery ticket, where one could either tax the winnings or tax the ticket, and demonstrates that one could achieve the same outcome by changing the number of tickets purchased. Hubbard (2005), however, makes a case for considering consumption taxes and income taxes as similar, both taxing returns to capital. He states: The truth is that, despite the common perception that consumption taxation eliminates all taxes on capital income, consumption and income taxes actually treat similarly much of what is commonly called capital income. (p. 83) He goes on to indicate that there are four types of capital income: the risk-free return, the expected risk premium, the return to market power or entrepreneurial skill, and luck. The last three, he suggests, are treated the same in an income and a consumption tax. Subsequently, he concludes that criticisms that a consumption tax will be regressive compared to an income tax because it does not tax the returns to capital are not well founded; and he cites data indicating that over a third of the reduction in the share of taxes paid by high-income households is offset by the taxation of risk taking and entrepreneurial returns. Note that Graetz was also concerned about the role of progressive tax rates. This issue would only arise with a graduated direct expenditure tax and not with the variants of consumption taxes frequently under consideration at present: an indirect tax such as value added or national retail sales tax, the flat tax, or the X-tax, a version of the flat tax with graduated rates on the wage portion, but not the business net cash flow portion that comprises the tax on investment outcomes. Similarly, note that there is no issue of loss offset with a consumption tax because the ex post payment is on the principal plus interest; thus, there is symmetric treatment even if the entire investment is lost.

172

TAX REFORM IN THE 21ST CENTURY

4 Gordon’s General Equilibrium Model of Income Taxes The original explorations of the interaction of risk and return with corporate tax rates were effectively limited to partial equilibrium approaches, examining a single investor. Gordon (1985) examined the corporate income tax in a general equilibrium framework, treating the tax as proportional with full loss offset. Gordon’s model shows that if there is no tax on the risk-free return to investment and revenues are returned in a lump-sum fashion, there is no distorting effect of the corporate tax because the burden of the tax is offset by the benefit of risk reduction. The corporate tax is therefore a non-distorting tax. Moreover, if the government can distribute risk more efficiently than the private sector (where trading costs and the limits on distributing across generations may constrain the efficiency of private risk allocation), the corporate tax can contribute to efficiency. He also finds that in a risky environment a corporate tax on risk could cause capital to move from the noncorporate to the corporate sector. Gordon’s fundamental conclusions are in opposition to the traditional analysis of the corporate tax. He finds that the corporate tax may raise significant revenues with very little distortion and may even be beneficial in terms of allocation if the government can allocate risk more efficiently than the corporate sector.

5

Kaplow’s General Equilibrium Model

Kaplow (1994) also has a general equilibrium model in which he examines all three taxes: income, wage, and consumption taxes. In his model, the government has its own portfolio of assets and can affect saving and resource allocation through portfolio adjustments. Kaplow demonstrates a number of tax equivalencies based on this model For the income tax, for example, he shows that the government could, instead of imposing a proportional tax on the portion of the return reflecting the risk premium, simply purchase more risky assets, which would increase the government’s yield and risk in the same way as an income tax. In that case, he concludes that the income tax is the equivalent of a wage tax plus a tax on wealth (reflecting the tax on the riskless portion of the return). This analysis reflects the original Domar-Musgrave notion that the investor, who is subject to tax on the risky return could restore his or her pre-tax risk and return by increasing the share of risky assets in his portfolio. In Kaplow’s model, the government provides the stock of risky and riskless assets with which the investor could, in theory, trade in order to restore his pre-tax outcome. Similarly, he shows that a consumption tax is likewise equivalent to a wage tax plus a tax on existing wealth. It might be helpful, in explaining not only this finding, but also some of the discussion in the following section, by positing a simple twoperiod model with an initial wage tax. At any one time in the economy, there is a set of old individuals, who own all of the capital, and consume out of capital plus

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

173

return, and young individuals who earn all of the labor income and save for old age. Consumption in the economy, assuming the economy is growing at rate g, is: (1) Co = (1+r)K (2) Cy = W(1–tw) – (1+g)K where subscripts o and y represent old and young, C is consumption, K is capital, W is wage income, tw is the wage tax rate, and r is the rate of return. If we add (1) and (2) we obtain the standard economy-wide relationship: Cy +Co = W+rK–gK–tw xW; that is total private consumption in the economy is income (from labor and capital) minus taxes minus net savings (the growth in the capital stock). Note that the young individual will become old in the next period and have consumption of Co = (1+r)(1+g)K. Thus, discounting the second period by the rate of return and adding it to the first period yields: (3)

Co + Cy/(1+r) = W(1–t)

Suppose now a tax-inclusive consumption tax at rate tc is substituted for the wage tax. We now have, at the time of introduction: (4)

Co = (1+r)K(1–tc)

(5)

Cy = W(1–tc) – (1+g)K(1–tc)

The tax on the initial old is the lump-sum tax on old capital. Collecting that tax provides the revenue which allows the young person to deduct his own investment plus some extra revenue if r is greater than g. Thus, the government will have additional revenues. Just as the holder of risky assets could restore his position by shifting the share of risky assets in his portfolio under a tax on the risk premium, the young person can restore his former consumption levels by increasing his investment from K to K/(1–tc), and his budget constraint is unchanged if tc = tw. In Kaplow’s model the government can supply those assets. Hence a consumption tax is equivalent to a wage tax and a tax on existing wealth.

6 Outcomes with Realistic Government Behavior and Optimizing Individual Models Gordon’s 1985 model makes the case that the corporate tax has little distorting effect, but his model, where the government returns the revenue as a lump sum, is not a very realistic guide for predicting outcomes, or even for determining an optimal tax

174

TAX REFORM IN THE 21ST CENTURY

system. Kaplow’s model is even less so, since the government does not behave in the way his model presumes (as he acknowledges in his paper). Governments hold physical assets that are used to produce public goods or that are held as common resources, such as national parks, but these assets are not bought and sold, although the federal government may occasionally sell an asset, such as a mineral right, to permit private production. The government also has a stock of debt but, in the case of central governments this stock of debt is used to manage the economy and finance spending, not to buy and sell as part of an investment portfolio. Governments may impose new taxes to finance spending, or replace an existing tax with a new one. We consider the second approach, which is the current focus of consumption tax proposals in the United States, to consider the merits of switching tax bases in a world of uncertainty. Replacement of the income tax by consumption and wage taxes has been studied using general equilibrium models, including fairly complex computable general equilibrium models, by many researchers. The most complex of these are the multiperiod life-cycle models, with the steady-state outcomes developed by Summers (1981), and a model tracing the transition, developed and eventually presented in detail in a book by Auerbach and Kotlikoff (1987). An alternative model is the infinite-horizon model which has a straightforward steady-state solution (since the after-tax return must return to the same level after a tax change). Most of these models are certainty models, although Engen and Gale (1996) developed a model with wage risk. We begin this discussion by examining the life-cycle model, followed by a discussion of the implications of the infinite-horizon model.

6.1

LIFE-CYCLE MODELS

As indicated above, in the two-period life-cycle model, consumption taxes differ from wage taxes in three ways: when introduced they impose a lump sum tax on old capital spent to finance consumption; they change the timing of taxes over the life-cycle; and they require a lower tax rate on wage income to be revenue neutral, because the wage base is smaller than the consumption base. Similarly, consumption taxes can be differentiated from income taxes, which can be illustrated by comparing consumption taxes to income taxes in our two-period model. In Table 1, the first row is the consumption of the old initially, Co; the second is the consumption of the young initially, Cy; the third is the budget constraint of the young. In the following three rows, in the bottom half of the table, these same variables are expressed in forms of the income tax rate t under conditions of revenue neutrality where (W is wage income, r the rate of return, K the capital stock, and g the growth rate.

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

175

Table 1: Consumption under Three Tax Regimes Wage Tax (tw)

Consumption Tax (tc)

Income Tax (t)

Consumption of the Old

Co = (1+r)K

Co = (1+r)K(1–tc)

Co = (1+r(1–t))K

Consumption of the Young

Cy = W(1–tw) – (1+g)K

Cy = (W– (1+g)K)(1–tc)

Cy= W(1–t) – (1+g)K

Budget Constraint of the Young

Cy +Co/ (1+r) = W(1–tw)*

Cy + Co/ (1+r) = W(1–tc)

Cy +Co/ (1+r) = W(1–t)– trK(1+g)/(1+r)*

Consumption and Budget Expressed in Income Tax Rate (a = capital income share, g/r =b) Consumption of the Old

Co = (1+r)K

Co= (1+r)K(1–t– tab/(1–ab))

Co = (1+r(1–t))K

Consumption of the Young

Cy = W(1–t– t(a/(1–a))–(1+g)K

Cy = W(1–ttab/(1–ab))– (1+g)K(1–ttab/(1–ab))

Cy = W(1–t)–(1– g)K

Budget Constraint of the Young, in t

Cy + Co/ (1+r) = W(1–t)– trK

Cy +Co/(1+r)= W(1–t)–WtgK/ (W+rK–gK)

Cy +Co/ (1+r) = W(1–t)– trK(1+g)/(1+r)

* This budget constraint could also be written Cy+ Co/(1+r(1–t)) = W(1–t), which is its typical form.

Switching from an income tax to a consumption tax imposes a lump-sum tax on the old, although not quite as large as in the case of the shift from a wage tax, because the increased tax applies to the principal at the beginning the period. In this respect the consumption tax differs from the wage tax and the income tax. The old are slightly benefited in a shift from the income tax to the wage tax, but heavily penalized under a shift from either tax base to a consumption tax. Under the consumption tax, the young have the lowest taxes in the first period because they are able to deduct all of their savings, while the consumption tax rate is only slightly higher than the income tax rate due to the small amount of net savings (gK). The taxes they save by investing today are repaid in the next period, but they have positive lifetime income effects from the shift in tax base from consumption to income, as can be shown in the budget constraint written in terms of the income

176

TAX REFORM IN THE 21ST CENTURY

tax rate. If the growth rate is zero, they effectively pay only the wage tax. The consumption tax is only slightly higher and is effectively a wage tax, while under the income tax they pay a second period tax on capital income which is large relative to wage income. Their lifetime income effect is positive under the consumption tax relative to the income and wage tax, as can be seen from the rewriting of the budget constraint in terms of the income tax rate, with revenue neutrality and no behavioral response. The present value of lifetime income is smallest under the wage tax and largest under consumption tax, as long as the rate of return, r, is greater than the growth rate, g. Suppose that the utility function between present and future consumption is Lshaped, so that the same fraction of consumption would be required after income changes. It is clear that under the consumption tax, where there is an overall lifetime income effect when moving from a wage to a consumption tax, savings should be increased from K to K/(1–tc) just to preserve the old level of consumption, and slightly more to reflect the share of the additional income allocated to future consumption. These simple two-period examples are reflected in more complex multi-period models, but there are clearly some important income effects. To shed some light on this issue, we use some findings from Engen, Gravelle, and Smetters (1997), who compare the effects of replacing the income tax with either a wage tax or a consumption tax using different types of models, and Gravelle (1991) who explores the roles of income and substitution effects under similar tax replacements using a compensation scheme. Substituting equal-yield wage and consumption taxes for a 20 per cent proportional income tax in a fixed labor supply model, Engen, Gravelle, and Smetters find that, under the switch to consumption tax the capital stock increases by 34.7 per cent; while, when switching to a wage tax, the capital stock falls by 1 per cent. Clearly, these income and timing effects are powerful. It is easy to see why. In their model, with savings rates very low (around 6 per cent), the consumption tax rate is similar to the income tax rate. If the capital stock were increased from K to K/(1–t), the capital stock would increase by 25 per cent. The bulk of the increase in savings in this model comes from the simple timing of the tax collections in the steady state. In a sense, this is part of the point being made by Kaplow, except that in a model of tax replacement the adjustments have to come, at least in a closed economy, from the interaction in the economy. One obvious difference from the Kaplow type of model is that the expansion of the capital stock will affect the rate of return and other variables in the economy. For example, in a Cobb-Douglas world with the capital income share at 25 per cent, output will rise by (1.25).25 – 1 per cent or 5.74 per cent. This expansion drives down the rate of return back towards the original after-tax return, in fact, moving 75 per cent of the way – the equivalent of a 15 per cent implicit tax on the original return. Although individuals have a lower

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

177

return in the steady state, their lifetime income rises through increases in the wage rate. Thus Hubbard’s (2005) argument that the consumption tax falls on a portion of income appears incorrect. Further savings should accumulate from the overall income effects both from direct changes and output increases. Note also that if this model were applied to a small open economy with perfect capital mobility, the capital stock could expand without affecting the rate of return or wage rates, an outcome similar to Kaplow’s model. In this case, the rest of the world fills the role of the government in providing an expanded stock of assets. Engen, Gravelle, and Smetters’ (1997) analysis, in fact, indicated, in the closedeconomy model, that the final outcome of the consumption tax was not affected by the inter-temporal substitution elasticity at all in the simple fixed labor model with a unitary factor-substitution elasticity and fixed labor (as long as there were no bequests or depreciation) when the model was re-calibrated to the same original economy, as a matter of theory. Nor was it much affected in their endogenous labor model. In that case the labor supply in the steady state fell slightly, but the change in output was negligibly affected. This increase in output from the timing shift would reduce any substitution effects and, in the simple model considered by Engen, Gravelle and Smetters, ultimately causes the capital stock to rise enough to return the rate of return to its original value. Hence, the inter-temporal substitution elasticity becomes irrelevant. Results were, however, affected by the wage tax shift. In their base case with endogenous labor, the wage tax shift reduced output by 0.6 per cent with an intertemporal substitution elasticity of 0.25. If the elasticity were raised to 0.5, output increased by 3.8 per cent, but if it were reduced to 0.05, it fell by 7.6 per cent. Gravelle (1991) examines a model where the capital income tax rate is higher (30 per cent) than the labor income tax rate (15 per cent). The inter-temporal substitution elasticity matters here, but it remains clear that income effects dominate. For the consumption tax, with an inter-temporal substitution elasticity of 0.25 and fixed labor supply, following the model of Summers (1981), the capital stock rises by 31.4 per cent with a consumption tax and falls by 1.1 per cent with a wage tax. However, if agents are compensated with lump-sum payments to reflect the tax changes during each period, the capital stock increases by only 13.3 per cent. Moreover, the welfare gain is much smaller. Lifetime income rises in the steady state by 5.03 per cent for the consumption tax and falls by 4.09 per cent with the wage tax, but the gain under the consumption tax in a compensated model is 0.6 per cent. In the endogenous growth model of Auerbach and Kotlikoff, welfare gains and losses are smaller: the consumption tax increases welfare by 2.02 per cent and the wage tax decreases it by 1.2 per cent, but the compensated gain under the consumption tax is only 0.39 per cent. What is the implication of this modeling for choosing a consumption tax over the income tax in an environment with asset-return risk? It is not that the consumption

178

TAX REFORM IN THE 21ST CENTURY

tax is a tax on risky returns: the same scaling up of the capital stock in response to timing should occur whether the returns are variable or not. The important issue is that the income tax is less distorting in the risky case. Of course, the responses to asset risk are also constrained by the economy. For example, if new, more risky, investments have more variability and lower returns than existing ones, then investors will not be able to fully restore their previous utility, unless they are in a small open-economy environment. But clearly the inter-temporal distortion arising from taxes in the risky asset world should be considerably smaller than the usual assumptions in dynamic models. Already, there is reason to believe that efficiency gains from shifting to a consumption tax are small relative to the dramatic and questionable inter-generational income shifting; considering a shift in a risky world simply reinforces that reservation. Unless there is a good policy reason for imposing a burden on the old in order to increase the capital stock, the case for a consumption tax shift seems weak indeed.

6.2 THE INFINITE-HORIZON MODEL The infinite-horizon model is a much less persuasive model of human behavior as it requires a great many assumptions. Most of those assumptions are flatly contradicted by real world observations. For example, with differential residence and source-based taxes in a world with mobile capital or with different state tax rates, we should observe corner solutions. In the infinite horizon model, the capital stock increases are identical for wage and consumption tax shifts, because there are no income effects. In the Engen, Gravelle, and Smetters (1997) model, they were the same as the consumption tax, which returned the after-tax return to its original value. Consumption falls (and savings rise) not because of timing effects, as the old can preserve their consumption by reducing bequests, but through substitution effects. Although the infinite horizon model is driven by a long-run infinite substitution elasticity, as in the life cycle model, the benefit of shifting to a consumption tax in this model is greatly reduced in an uncertain world. If the tax on capital income is negligible, because the tax is offset by a reduction in variance, the efficiency gains are much smaller.

7

Behavioral Models

Although the life-cycle model has much to recommend itself compared to the infinite horizon model (which has more restrictive assumptions and produces corner solutions in many real world applications, such as those in an open economy), there are also reasons to be skeptical that any of these inter-temporal models describe real world savings behavior given the model’s demands (perfect information, perfect foresight, ability to solve an extremely complicated economic problem that occurs only once in a lifetime, and perfect capital markets). Agents may instead choose

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

179

rules of thumb (see Bernheim, 2002, for a review of both inter-temporal models and alternative theories). These models are often referred to as behavioral or bounded rationality models. In a world of imperfect information and knowledge, and limited capability to solve complex optimization problems, one could also make the case that savers may not focus on the variance-reducing aspects of an income tax. Certainly, the resistance to capital income taxes by most wealthy individuals is suggestive of such a view. If investors do not perceive the risk-reducing benefits, then they may simply view capital income taxes in the standard way as in a certainty world. At the same time, with such a view of behavior, agents also do not respond to these taxes in the same way as the standard certainty models of saving. There is some evidence that agents do not engage in the rational behavior depicted in inter-temporal models or the optimizing behavior with respect to risk suggested by the Domar and Musgrave model. One study that supports this view finds that the default position (opt out or opt in) has a pronounced effect on employee participation in savings plans, and that individuals also choose the default portfolio allocation (Madrian and Shea, 2001). Those findings suggest that individuals are either heavily influenced by inertia, or that they depend on clues from their employers about both the level and allocation of saving. Economists are understandably reluctant to abandon rational optimizing approaches, because it leaves them without any model to guide either description or prescription. Nevertheless, policy should be made to reflect what people do, not what we wish them to do. In “bounded rationality” models, individuals, unable to make the complex calculations required of optimizing behavior, fall back on rules of thumb. In addition to signals from their employers, they may try to follow financial advisors’ advice about investing in the stock market for the long term, and having a mixed portfolio. They may try to save a certain fraction of their income, or they try to achieve a target amount of wealth in old age, or they notice how elderly relatives get by with social security, Medicare and pensions. If they make mistakes, they might adjust the age at which they retire if that is possible. They may tie up their savings in less accessible forms to deal with lack of self-control in consumption. These approaches suggest that the savings rate is relatively unresponsive to rates of return, or may decline if a target wealth is planned. Some empirical work is fairly consistent with this notion. Times series studies mostly suggest small responses with varying signs of savings response rates to changes in rate of return (see Gravelle, 1994, Chapter 1 for a summary). The evidence on IRAs and pension savings plans does not provide much clear evidence that savings increased with consumption tax type provisions (see Gravelle, 1994, Chapter 8 and Bernheim, 2002). There are many problems with these studies, but one cannot ignore the fact that long periods with many dramatic changes in tax rates were nevertheless characterized by little change in savings rates.

180

TAX REFORM IN THE 21ST CENTURY

How does one evaluate policy in such a model? In this case the distortions of the income tax may be relatively unimportant, not because the reduction in variance compensates for the tax, but because individual savers and investors do not react to the tax rate in their savings behavior. That in itself suggests that the efficiency gains from shifting to a consumption tax may be limited, but for different reasons. But can we learn anything further from, at least, a thought experiment, about guides to making policy? Consider the effect of shifting to a consumption tax in a model where the savings rate is a constant share of disposable income and individuals do not recognize either the risk aspect of income taxation or the future taxes associated with consumption tax. And suppose policy-makers believe that too little savings occurs, either because they believe people are naturally myopic, or because the evidence suggests that the standard of living falls a great deal when old. There is some evidence that this is the case, particularly for lower-income households (see Bernheim and Scholz, 1993). In this particular rule-of-thumb world, the outcome could depend on the form of the tax. If the tax were imposed as a direct consumption tax (with a tax on dissaving and a deduction for saving), the old would of course pay the tax. The young in our simple two-period world would have a rise in disposable income, and would consume part and save part. But their additional saving would not be adequate to restore the level of consumption they previously enjoyed because they consume part of their lower tax liability. Such a circumstances already potentially exists in the United States in part, where a large fraction of passive assets are in retirement accounts that experience consumption tax treatment. This effect can be illustrated in our simple two-period model, where, to simplify, there is no net saving so that the consumption tax rate and the income tax rate are the same, absent behavioral changes. Under the income tax, the individual budget constraint when young is Cy = W(1–t)–K, and we assume that K is a fixed share of after-tax income, that is K = sW(1–t). Hence, C = W(1–t)(1–s). Now assume a direct consumption tax is imposed, so that Cy = W(1–t)–K*(1–t). Disposable income is W(1–t) +tK. Part of the additional disposable income from the savings from deducting investment will be consumed, and while savings will rise, it will not rise to K/(1–t), but to K/(1–st). (Note that the new value of savings is derived from solving an equation where disposable income is a function of the new level of savings; if it is perceived to be based on the old level, the new savings will be K(1+st).) Thus, the individual will likely have less consumption in the next period (assuming r and s are small relative to K). Alternatively, suppose the tax were imposed as a VAT or retail sales tax and prices rose to accommodate the tax. The old would experience the wealth confiscation as a loss of purchasing power. The young would see their nominal disposable income rise because there is no longer a tax. The budget constraint would be Cy/(1–t) = W– K, where the price level increases from 1 to 1/(1–t). In this case, K is now sW,

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

181

and the new level of K compared to the old one is K* = K/(1–t). In this world of simple rules, the form makes a difference. Consumption in the first period would not change. Individuals would consume enough of their additional disposable income to cover price increases, and consumption in the second period would rise slightly (because the return is not taxed). In this case, the income effect falls entirely on second-period consumption, and savings after paying tax are slightly larger than before. Note, however, that if the consumption tax rate is higher than the income tax rate because of growth, the income effect would be smaller, and consumption would fall in the first period, so that despite a positive income effect, consumption would fall in the first period and rise in the second. The Hall Rabushka tax, a VAT with wages taxed at the individual level would not require a price rise. The lump-sum tax on old capital falls on equity capital, either through a tax on the sale of physical assets or a fall in the price of corporate stock. For investments in physical capital, it would similar to a direct consumption tax. For investments in debt, the budget constraints would be the same as in the income tax case, and no change in first period consumption would occur by a direct tax reduction for the individual. It would be as if the income tax were still in place. For investments in stock, the price of stock would fall by more than the tax (to account for existing debt) and investors in stock would be able to purchase more shares for a fixed nominal amount, but the price would remain depressed when they sold them for consumption, following the pattern for a regular consumption tax larger than the actual tax. For individuals investing directly in assets it would have effects similar to the direct consumption tax. Thus investors in passive assets would have outcomes similar to a regular VAT with price increases. What is causing the form to matter? Because the saver follows a fixed rule, the perception of income is important. In the VAT case, income rises through tax reductions to fund consumption in both periods at the higher prices and by applying the same savings rates, real consumption is not changed. With the direct consumption tax, disposable income rises by an amount that needs to be saved to pay for future taxes, but if the saver does not recognize this future tax, the additional tax saving is split between consumption and savings, leading to a decline in future consumption. This failure of optimization with a direct consumption tax is actually quite important in the United States where significant amounts of savings are in retirement plans that use this method of up-front deduction and payment on distribution. The outcomes would be different in cases where the target is a fixed amount of capital, with outcomes depending on how individuals incorporate prospective taxes on sale and price changes. Clearly, in a case with price changes, if individuals did not recognize the reduced purchasing power of their assets, they would not change their saving, and would find themselves, when old, in the same position as the initial old. Similarly, if individuals did not recognize that taxes must be paid on spending from principal as well as return, they would save too little.

182

TAX REFORM IN THE 21ST CENTURY

One problem with these models is that it is not clear how rules-of-thumb are formed. With target savings, the targets might adjust in the case of price changes, but perhaps not in the case of direct taxes. But without understanding behavior in these types of models, the outcomes become very uncertain. In none of these cases is there a gain in welfare from eliminating distortions; all of the outcomes reflect how the rule of thumb interacts with the apparent shifts in income to produce particular outcomes. We have done very little research on these issues despite evidence indicating that individuals do not optimize in the manner suggested in our inter-temporal models.

8

Conclusion

We have considered four types of models that might operate in a risky environment: the Gordon (1985) model of a capital income tax imposed in isolation; the Kaplow (1994) model where the government is an active participant in the asset markets, a standard tax-substitution model with inter-temporal optimization; and a behavioral model where households are limited in their sophistication and ability to optimize. Ultimately the Kaplow (1994) model and even the Gordon (1985) model are questionable for examining taxation, consumption taxes, and risky asset returns. The behavior of the government in both these models is suspect, particularly in the Kaplow model. This paper raises the question of how useful such an analysis is in deciding whether adopting a consumption tax as a substitute for an income tax is welfare improving. If we estimate a more realistic model, that allows government to behave as it normally does and allows individuals to optimize, introducing risk suggests that the gain in welfare is negligible, which greatly undermines the already rather weak policy case for a consumption tax. If, however, we introduce the tax change into the relatively uncharted territory, both theoretically and empirically (see Bernheim, 2002, for a review of the evidence), of non-optimizing models, the outcome becomes unclear, but possibly damaging to individual welfare, especially with some forms of consumption tax. If, in this world of failure to optimize, we are concerned about savings behavior, it might be better to approach the issue through policies whose outcomes are more certain and perhaps more consistent with non-optimizing models. They include expanding the coverage of Social Security, and encouraging participation in employer plans by requiring an opt-in rule and a balanced default portfolio, rather than adopting a consumption tax that taxes wealth when spent. Moreover, it might be desirable to move away from retirement savings plans with up-front deductions such as traditional IRAs, and toward Roth-style IRAs that do not allow up-front deductions but do not tax returns, moving tax collections from the future into the present. If the revenue gain from this change were used to reduce the deficit, so as not to undermine national savings in case private savings fell under a traditional model, outcomes might be improved under any modeling approach.

JANE GRAVELLE: CONSUMPTION TAXES AND RISK REVISED

183

Perhaps it is worthwhile to close with a quote from Bernheim (2002), who following a discussion of behavioral models where individuals are unable to optimize or must deal with self-control issues, indicated that one can find references to alternative behavioral hypotheses in conventional analysis, but noted: Yet these references are usually haphazard and mentioned in a rather ad hoc way as possible explanations for otherwise puzzling phenomena. With rare exceptions, alternative behavioral hypothesis have not been used as frameworks for organizing lines of inquiry concerning the effects of taxes on savings. (p. 1203) One might also add that they have not been explicitly incorporated into studies of consumption taxation as a replacement for the income tax either. Perhaps it is time they did. References Auerbach, A.J. and L.J. Kotlikoff (1987), Dynamic Fiscal Policy (New York: Cambridge University Press). Bernheim, B.D. (2002), ‘Taxation and Saving’, in A.J. Auerbach and M. Feldstein (eds), Handbook of Public Economics 3 (New York: Elsevier) 1173. Bernheim, B.D., and J.K. Scholz (1993), ‘Private Savings and Public Policy’, in J.M. Poterba (ed.), Tax Policy and the Economy 7, National Bureau of Economic Research (Cambridge, MA: MIT Press) 73. Bradford, D.F. (1986), Untangling the Income Tax (Cambridge, MA: Harvard University Press). Domar, E. and R.A. Musgrave (1944), ‘Proportional Income Taxation and RiskTaking’, Quarterly Journal of Economics 58: 388. Engen, E. and W.G. Gale (1996), ‘The Effects of Fundamental Tax Reform on Savings’, in H.J. Aaron and W.G. Gale (eds), Economic Effects of Fundamental Tax Reform (Washington, DC: Brookings Institution) 83. Engen E., J. Gravelle, and K. Smetters (1997), ‘Dynamic Tax Models: Why They Do the Things They Do’, National Tax Journal 50: 657. Gordon, R.H. (1985), ‘Taxation of Corporate Capital Income: Tax Revenues Versus Tax Distortions’, Quarterly Journal of Economics 100: 1. Graetz, M. (1980), ‘Expenditure Tax Design’, in Joseph Pechman (ed.), What Should be Taxed: Income or Expenditure? (Washington, DC: Brookings Institution) 161. Gravelle, J.G. (1991), ‘Income, Consumption, and Wage Taxation in a Life-Cycle Model: Separating Efficiency from Distribution’, American Economic Review 81: 985.

184

TAX REFORM IN THE 21ST CENTURY

Gravelle, J.G. (1994), The Economic Effects of Taxing Capital Income (Cambridge, MA: MIT Press). Hubbard, G. (2005), ‘Would a Consumption Tax Favor the Rich?’, in A.J. Auerbach and K. Hassett (eds), Toward Fundamental Tax Reform (Washington, DC: AEI Press) 81. Kaplow, L. (1994), ‘Taxation and Risk Taking: A General Equilibrium Perspective’, National Tax Journal 47: 789. Madrian, B.C. and D.F. Shea (2001), ‘The Power of Suggestion: Inertia in 401(k) Savings Behavior’, Quarterly Journal of Economics 118: 1149. Poterba, J. (2002), ‘Taxation, Risk-Taking and Household Portfolio Behavior’, in A.J. Auerbach and M. Feldstein (eds), Handbook of Public Economics 3 (New York: Elsevier) 1109. Sandmo, A. (1985), ‘The Effects of Taxation on Savings and Risk-Taking’, in A.J. Auerbach and M. Feldstein (eds), Handbook of Public Economics 1 (New York: North Holland) 265. Summers, L. (1981), ‘Capital Taxation and Accumulation in a Life Cycle Model’, American Economic Review 71: 533.

Tax Rate Scale: Equity and Efficiency Aspects

Chapter 7

Taxation, Labour Supply and Saving Patricia Apps and Ray Rees* 1

Introduction

Over the last decade or more, a number of countries, notably the US, UK and Australia, have reduced significantly the progressivity of the rate scale of their formal income tax systems by cutting rates at higher income levels. It is well recognised that these reforms have shifted the overall tax burden towards the middle of the earnings distribution during a period of increasing inequality. The same countries have also introduced new, or expanded existing, tax credit and family support programs with benefits withdrawn on the joint income of a couple. Examples include the earned income tax credit (EITC) program in the US, the child tax credit (CTC) and working tax credit (WTC) in the UK, and the Family Tax Benefit (FTB) system in Australia. It is less well recognised that the withdrawal of benefits on joint income under these programs leads to a system of joint taxation with a marginal rate scale that is no longer progressive but, instead, exhibits an inverted U-shaped profile – the highest marginal rates apply across low to middle family incomes. With joint income as the tax base, the new rate scale raises the effective marginal tax rate on a secondary earner, and therefore tends to shift the overall tax burden towards twoearner couples on low to average earnings. Since the US allows joint filing, the key effect of combining its EITC program with the Federal Income Tax is to replace the progressive marginal rate scale of the latter with a scale that has higher effective rates on low to middle family incomes, and therefore higher effective rates on the income of a secondary earner. In countries *

University of Sydney; Australian National University; University of Technology Sydney; Institute for the Study of Labor; Taxation Law and Policy Research Institute, Monash University; and University of Munich, respectively. This research was supported by an Australian Research Council Discovery-Project grant.

188

TAX REFORM IN THE 21ST CENTURY

such as the UK and Australia, which have apparently individual income tax systems, the programs simultaneously result in a significant shift from individual to joint taxation under the newly created inverted U-shaped rate scale. The aim of this paper is to show, as a detailed case study, how, through a succession of reforms, Australia has moved towards an income tax system of this kind.1 We show first, in Section 2, how increases in the low-income tax offset (LITO) combined with tax cuts at high income levels in recent Australian Government budgets have shifted the overall tax burden towards the ‘middle’. Following this we explain how the FTB system compensates single-earner families on low and average earnings while leaving two-earner families uncompensated. The result is that, though ostensibly based on the progressive taxation of individual incomes, we now have a system that closely approximates one of joint taxation with the highest marginal rates applying across a wide middle band of earnings and to the income of a secondary earner, typically the female partner. The available evidence on wage elasticities indicates that the labour supply of prime-aged males tends to be unresponsive to changes in the net wage, while that of females in the same age category tends to be quite responsive. If we accept this evidence, the ‘new’ tax system can be expected to have significant disincentive effects on female labour supply.2 In Sections 2 and 3 we present data that show female hours of work are not only well below male hours, but they also exhibit a high degree of heterogeneity across seemingly identical households. We argue that both these features of the data can be attributed to high tax rates on second earners, together with a poorly developed and costly childcare sector. We go on to explain why, under these conditions, household income is neither a fair nor efficient tax base. In Section 4 we present a lifecycle analysis that indicates strong negative effects on household saving as well as on labour supply. The central thesis of the paper is that, over the last decade or more, Australia has moved towards an income tax system that fails in terms of efficiency considerations, and cannot be supported on the basis of either vertical or horizontal equity criteria. A concluding comment is contained in Section 5.

2 The Australian Income Tax System Australia is widely viewed as having an income tax based on individual incomes, with a progressive rate scale. However, as already indicated, a country’s income tax is set not solely by the formal rate scale applying to personal income, but in combination with tax credits, offsets, exemptions and cash benefits. The focus of

1 2

For an analysis of the UK and US formal and effective income tax systems, see Apps and Rees (2009). See Eissa and Hoynes (2004, 2006) and Eissa and Liebman (1996) for studies that estimate the disincentive effects of the high tax rates on second earners under the US EITC program.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

189

the analysis in this section is on the structure of marginal and average rates of the Australian income tax system resulting from four key policy instruments: • • • •

Personal Income Tax (PIT); Low income tax offset (LITO); Medicare Levy (ML); and Family Tax Benefit Part A (FTB-A) and Family Tax Benefit Part B (FTB-B).

Section 2.1 examines the rate structure of the Personal Income Tax when combined with the LITO and ML. We then go on to consider in Section 2.2 the change in the structure of rates when FTB-A and FTB-B are included.

2.1

PERSONAL INCOME TAX, LITO AND ML

Table 2.1 lists the formal MTR scale of the PIT and the true scale that applies when the LITO is included.3 Figure 2.1a shows graphically the effect of the LITO on the MTR scale of the PIT, and Figure 2.1b presents the resulting profile of ATRs, with respect to annual taxable income. Table 2.1: PIT Rate Scale and LITO, 2007–08 Taxable Income

MTR

Taxable Income

$0 – $6,000 $6,001 – $30,000 $30,001 – $75,000 $75,001 – $150,000 $150,000 +

0.00 0.15 0.30 0.40 0.45

$0 – $11,000 $11,001 – $30,000 $30,001 – $48,750 $48,751 – $75,000 $75,001 – $150,000 $150,000 +

MTR + LITO 0.00 0.15 0.34 0.30 0.40 0.45

As we can see, according to the specified formal rate scale, we have a strictly progressive, piecewise-linear income tax. However, when the LITO is included, this is no longer the case. A higher rate in the dollar applies to incomes from $30,000 to $48,750 than to the next income band, $48,750–$75,000. The LITO raises the zerorated threshold from $6,000 to $11,000, and it also raises the MTR from 30 cents to 34 cents in the dollar on incomes from $30,001 to $48,750, as shown in Table 2.1. The LITO is in fact an entirely redundant policy instrument that serves only to reduce the transparency of the higher rate of 34 cents in the dollar in the middle of the scale.

3

The 2007–08 LITO is $750, and is withdrawn at a rate of 4 cents in the dollar above a lower income threshold of $30,000.

190

TAX REFORM IN THE 21ST CENTURY

-.1

0

.1

MTR .2

.3

.4

.5

Figure 2.1a: MTRs: PIT Scale and LITO, 2007–08

0

50000 100000 Taxable income, dollars pa MTR (PIT)

150000

MTR (PIT+LITO)

-.1

0

.1

ATR .2

.3

.4

.5

Figure 2.1b: ATRs: PIT and LITO, 2007–08

0

50000 100000 Taxable income, dollars pa ATR (PIT)

150000

ATR (PIT+LITO)

When considered in isolation, the LITO might appear to have the advantage of reducing ATRs at low income levels. The standard argument along these lines is that an offset is a more effective (or less ‘costly’) use of taxpayers money for assisting those on very low incomes than an increase in the zero-rated threshold, which provides a benefit to all taxpayers above the threshold. This argument is fundamentally flawed. It is based on a misunderstanding of the relevant criteria for evaluating a tax system – those of efficiency and fairness. Given that higher income, prime-aged individuals are typically found to have unresponsive labour supplies, a more efficient system is likely to be one that retains rate scale progressivity. In regard to fairness, the argument distracts attention from the true distributional outcome of reforms of this kind.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

191

If we examine successive increases in the LITO and the simultaneous cuts in rates on higher incomes introduced in recent budgets and proposed for future years, we find that the effect is to shift a disproportionate share of the tax burden towards middle income earners. To show this, Figure 2.2 plots the ATR profiles of the PIT scale and LITO for 2004–05 and 2007–08, the proposed rate scale and LITO for 2010–11, and the Government’s ‘aspirational’ rate scale and LITO for 2013–14. The MTRs for 2004–05, 2010–11 and 2010–14 are listed in Table 2.2 below. ATRs are calculated for incomes indexed by average weekly ordinary time earnings, with 2007–08 as the reference year.

-.1

0

.1

ATR .2

.3

.4

.5

Figure 2.2: ATRs: PIT + LITO, 2004–05 to 2013–14

0

50000

100000 150000 Taxable income, dollars pa ATR 2004-05

ATR 2007-08

ATR 2010-11

ATR 2013-14

200000

From around 1996, bracket creep has raised significantly the ATR on average earnings. Figure 2.2 shows that the effect of the LITO, together with tax cuts at higher income levels, has been to deny individuals on average earnings an equiproportional rate of compensation for the failure to index tax bands. This is indicated by the smaller vertical gap between the ATR profiles for 2004–05 and 2007–08, and the close-to-zero gaps between later profiles, across taxable incomes from around $50,000 to $70,000 pa. From the profile of MTRs we can see that, in an economy with rising tax revenues due to bracket creep, the shift to the ‘middle’ is achieved by raising the 30 cents rate to 34 cents across low to middle incomes; for example, across the incomes from $37,000 to $82,500 in 2013–14, while also reducing the top rates. When we introduce the ML there is an even stronger effect of this kind.4 To illustrate, we take 4

The ML is normally calculated at 1.5 per cent of taxable income, with exemption categories or reductions based on family income. There is a surcharge for individuals and families on higher incomes who do not have private patient hospital cover – calculated at an additional 1 per cent of taxable income.

192

TAX REFORM IN THE 21ST CENTURY

the case of a two-parent, single-income family with two dependent children. For this family, the lower income threshold of the ML exemption is $33,841. At this point the exemption is withdrawn at a rate of 10 cents in the dollar. This creates eight income tax bands and results in an MTR of 40 cents on incomes from $33,436 to $39,335 pa. This 40 cent rate is followed by a 35.5 cent rate and then a 31.5 cent rate. We therefore see a further shift towards an inverted U-shaped profile of MTRs. Table 2.2: Tax Scales: 2004–05, 2010–11, 2013–14 Taxable Income

MTR

Taxable Income

2004–05 $0 – $6,000 $6,001 – $21,600 $21,601 – $58,000 $58,001 – $70,000 $70,000 +

(LITO $235) 0.00 0.17 0.30 0.42 0.47

$0 – $7,382 $7,382 – $21,600 $21,601 – $27,475 $27,476 – $58,000 $58,001 – $70,000 $70,000 +

2010–11 $0 – $6,000 $6,001 – $37,000 $37,001 – $80,000 $80,001 – $180,000 $180,000 +

0.00 0.17 0.34 0.30 0.42 0.47 (LITO $1500)

0.00 0.15 0.30 0.37 0.45

$0 – $16,000 $16,001 – $30,000 $30,001 – $37,000 $37,001 – $67,500 $67,501 – $80,000 $80,001 – $180,000 $180,000 +

‘Aspirational’ 2013–14 $0 – $6,000 $6,001 – $37,000 $37,001 – $180,000 $180,000 +

MTR + LITO

0.00 0.15 0.19 0.34 0.30 0.37 0.45 (LITO $2100)

0.00 0.15 0.30 0.40

$0 – $20,000 $21,001 – $30,000 $30,001 – $37,000 $37,001 – $82,500 $82,501 – $180,000 $180,000 +

0.00 0.15 0.19 0.34 0.30 0.40

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

193

Table 2.3: MTRs for 2-Child Family: PIT+LITO+ML, 2007–08 Taxable Income $0 – $11,000 $11,001 – $30,000 $30,001 – $33,435 $33,436 – $39,335 $39,336 – $48,750 $48,751 – $75,000 $75,001 – $150,000 $150,000 +

MTR 0.00 0.15 0.30 0.40 0.355 0.315 0.415 0.465

The family tax benefit system has a more profound effect of the same kind on the MTR profiles of two-parent families, as the following analysis shows.

2.2

FAMILY TAX BENEFITS 2007–08

Table 2.4 illustrates the effects of the FTB system on marginal and average tax rates for a single-earner family with two children under 13, one under 5, as taxable income rises. FTB-A provides a cash transfer of $4,460.30 per child under 13 years. The ‘maximum rate’ is withdrawn up to the ‘base rate’ on family income above $41,138 at a rate of 20 cents in the dollar. The base rate is $1890.70 and is withdrawn at a rate of 30 cents in the dollar on joint income above $95,192. The result is that a MTR of 55.5 cents in the dollar applies across the middle income band of $41,319 – $67,014. At $95,193 the MTR rises to 71.5 cents in the dollar due to the withdrawal of the base rate at 30 cents in the dollar. FTB-B is a cash transfer of $3,584.30 for a child under 5 and is withdrawn on second earner’s income above $4,380.0,5 and so the single-earner family receives the whole amount. The effect of withdrawing the maximum rate of FTB-A from $41,319, together with the higher MTR due to the LITO and the withdrawal of the ML exemption, is to replace the progressive PIT rate scale with one that has very high rates towards the lower end and middle of the income distribution. ATRs are progressive due to family tax benefits. We have a negative income tax system for the single-earner, two-child family up to around $50,700. The impact on the two-earner family is, however, very different. MTRs and ATRs on the second income depend on the primary earner’s income because FTB-A and the ML exemption are withdrawn on joint income. Table 2.5 lists the tax rates on the second income when the primary earner’s income is $40,000 and there are two children under 13, with one under 5. A MTR of 21.5 cents applies to her earnings from $1,319 to $4,380 due to the withdrawal of FTB-A, plus the ML 5

The benefit is fully withdrawn at $22,302.

194

TAX REFORM IN THE 21ST CENTURY

of 1.5 per cent. At $4,380 her MTR goes up an additional 20 cents, to 41.5 cents in the dollar, due to the withdrawal of FTB-B. The withdrawal of FTB-B on the second income, together with the withdrawal of FTB-A on joint income, has the effect of denying the second earner the zero-rated threshold of $11,000. At $11,000 her MTR rises to 56.5 cents in the dollar, because her income hits the 15 per cent rate of the PIT scale. Table 2.4: MTRs and ATRs: Single-Earner Family PIT+LITO+ML+FTBs Taxable Income $0 – $11,000 $11,001 – $30,000 $30,001– $33,435 $33,436 – $39,335 $39,336 – $41,318 $41,319 – $67,014 $67,015 – $75,000 $75,001 – $95,192 $95,193 – $107,797 $107,798 – $150,000 $150,000 +

MTR

ATR

0.00 0.15 0.34 0.44 0.355 0.555 0.315 0.415 0.715 0.415 0.465

–1.21 –0.39 –0.22 –0.15 –0.13 0.12 0.15 0.20 0.26 0.31 –

Table 2.5: MTRs and ATRs on the Second Income Tax Rates on Second Earnings Taxable Income $0 – $1,318 $1,319 – $4,380 $4,381 – $11,000 $11,001 – $22,302 $22,303 – $27,014 $27,015 – $30,000 $30,001 – $40,000

MTR

ATR

0.015 0.215 0.415 0.565 0.365 0.165 0.355

0.015 0.155 0.311 0.440 0.427 0.401 0.389

Primary income = $40,000 pa

Because FTB-A, excluding the base rate, is fully withdrawn at a family income of $67,014, the second earner’s MTR falls to 16.5 cents in the dollar, the rate on personal income plus the ML, at $27,015. At $30,000 her MTR is 36.5 cents in the dollar, the sum of the 30 per cent rate on personal income, the 4 cents withdrawal rate of the LITO, and the ML rate. At $40,000 the second earner has faced such high MTRs on her earnings up to $22,302 that her ATR is close to 40 per cent.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

2.3

195

DIAGRAMMATIC EXPOSITION

This section illustrates diagrammatically the selective taxation of second earners under the ML and Family Tax Benefit system, and the resulting shift towards joint taxation. The figures compare MTRs and ATRs faced by the primary and second earner in a household that switches from being single to two-earner – in other words, a household that switches ‘type’ by changing the labour supply of the female partner as second earner. We compare three types: • SE A single-earner household in which the male works full time in the

market and the female works full time at home • PT A two-earner household in which the male works full time in the

market and the female divides her time equally between market work and household production, and earns half the market income of the male. • FT A two-earner household in which both partners work full time in the market and earn the same incomes. MTRs and ATRs faced by each partner are computed as a function of primary income. We compare the tax rate profiles of the PIT and LITO with those for the full system including FTBs and the ML, for a household with zero non-labour income, no gender wage gap and with two dependent children under 13 (one under 5). We also present profiles of ATRs on the second income for selected levels of primary income. MTRs and ATRs by Primary Income Figure 2.3a plots the MTR profiles for the PIT with the LITO included, and Figure 2.3b, the profiles for the full system, with respect to primary income. The MTR profiles of primary earners are denoted by MTR1 followed by their household type, SE, PT or FT. Similarly, the MTR profiles of the second earner are denoted by MTR2 followed by household type. As indicated in Figure 2.3a, the MTR profiles of primary earners and of the second earner in the FT household coincide because each earns the same income and the tax base is individual income. The second earner in the PT household faces a lower MTR because she has a lower income6 apart from where the withdrawal of the LITO raises her MTR to 34 cents before the 40 cents in the dollar rate becomes effective. When we move to the full system in Figure 2.3b we see that FTBs and the ML introduce much higher MTRs across average earnings for the SE household. For the FT and PT households the higher rates apply at lower income levels. This

6

As shown in Boskin and Sheshinski (1983), a lower tax on a second earner with a lower income is consistent with the Ramsey rule for efficiency. It can also be supported on distributional grounds, as shown in Apps and Rees (2007). See also Apps and Rees (1999).

196

TAX REFORM IN THE 21ST CENTURY

is a feature of joint taxation: at any given level of primary income, the two-earner household faces a higher MTR than the single-earner household.

-.4

-.2

0

MTR .2

.4

.6

.8

Figure 2.3a: MTRs: PIT+LITO by Primary Income

0

50000 100000 Primary income, dollars pa MTR1 SE PT FT

150000

MTR2 PT

MTR2 FT

-.4

-.2

0

MTR .2

.4

.6

.8

Figure 2.3b: MTRs: PIT+LITO+FTBs+ML by Primary Income

0

50000 100000 Primary income, dollars pa MTR1 SE

150000

MTR2 PT

MTR2 FT

Figure 2.4a plots the profiles of ATRs on household income due to the PIT and LITO and Figure 2.4b, shows those for the full system. In each figure the ATR profiles of the three household types are denoted by ATR SE, ATR PT and ATR FT, respectively. The graphs also show ATR profiles of the second earner in the PT and FT households, ATR2 PT and ATR2 FT, calculated as the additional tax the household pays due to the second earnings.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

197

-.6

-.4

-.2

ATR 0

.2

.4

.6

Figure 2.4a: ATRs: PIT+LITO by Primary Income

0

50000 100000 Primary income, dollars pa ATR SE

ATR2 PT

ATR PT

ATR2 FT

150000

ATR FT

In Figure 2.4a the ATR profiles for the SE and FT households also coincide, again because the tax base is individual incomes. Note, however, that while the two household types face the same ATR at any given level of primary income, the FT household pays twice as much tax because both partners work full-time in taxed market production. In effect, they contribute twice as much to financing the FTB system. The PT household pays less than twice as much because the second earner has a lower income.

-.6

-.4

-.2

ATR 0

.2

.4

.6

Figure 2.4b: ATRs: PIT+LITO+FTBs+ML by Primary Income

0

50000 100000 Primary income, dollars pa ATR SE

ATR2 PT

ATR PT

ATR2 FT

ATR FT

150000

198

TAX REFORM IN THE 21ST CENTURY

Figure 2.4b shows the increase in the tax burden on the two-earner household due to the FTB system. The withdrawal of benefits on joint income and on the income of the second earner raises the ATR on the household’s additional income due to the second partner going out to work. The higher ATR on her income raises, in turn, the ATR on household income in the two-earner households. The FT household now pays more than twice as much tax as the SE household. Again, this is a characteristic feature of joint taxation. ATRs on the Second Income Figure 2.5 plots the ATR profile faced by the second earner as her income rises, for four levels of primary income: $30,000, $40,000, $50,000 and $60,000 pa. (Note that Table 2.5 gives the figures for ATRs on a second income for the case of the family with a primary income of $40,000 pa.) The graph shows that in all four cases, the second earner reaches high ATRs at relatively low income levels.

-.2

0

ATR .2

.4

.6

Figure 2.5: ATRs on Second Income at Selected Primary Income Levels

0

10000

20000 30000 40000 Primary income, dollars pa ATR2 30000 ATR2 50000

50000

60000

ATR2 40000 ATR2 60000

Taxes on second earners and their families at the levels indicated, together with a lack of access to affordable, high quality childcare, can be expected to have strong negative effects on female labour supply, not only during the child rearing years but throughout the entire lifecycle, for reasons discussed in detail in Section 4.

2.4

IMPACT ON ‘IN-WORK’ FAMILIES

We now turn to an empirical analysis of the impact of the rate structure we have described on families ‘in-work’, using data for a sample of 1945 couple income units drawn from the Australian Bureau of Statistics (ABS) Income and Housing

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

199

Survey (IHS) (2003–04). The sample is selected on the criteria that the couple income unit has dependent children and at least one parent is employed. Families in which both parents are unemployed or out of the workforce are excluded in order to focus on the income tax and FTB system, rather than on the wider welfare system.7 The sample is also limited to families in which at least one parent earns above $15,000 per annum, earnings are principally from wages and salaries, and neither parent has a negative income from investments or unincorporated enterprises. The parent with the higher private income is treated as the primary earner.8 The male partner is the primary earner in over 87 per cent of records. All incomes are indexed to the 2007–08 financial year. As in the preceding section, the tax burden on the primary earner is calculated as the tax the family would pay if it had only one earner, that is, if the second earner withdrew from work and therefore reported zero earnings. The burden on the second earner is then calculated as the increase in the family’s tax burden when her earnings are included in family income. We present profiles of ATRs for three household types: single-earner (SE), two-earner with the second in part-time work (PT), and two-earner with both partners in full time work (FT).9 Table 2.6 first of all shows the distribution of household types by quintiles of primary income. Note that the three types tend to be fairly evenly distributed across quintiles, apart from a slight tendency for two-earner households to predominate in the middle quintiles. Table 2.6: Household Type by Primary Income Quintiles Primary income SE PT FT

1

2

3

32299 44.7 29.9 25.4

45349 31.0 34.4 34.6

56588 30.4 37.4 32.2

4

5

70157 125060 27.0 36.3 42.2 38.1 30.8 25.6

All 66096 33.9 36.4 29.7

Table 2.7 reports the ATR on the household income of each household type, and on the second income of the PT and FT household (ATR2 PT and ATR2 FT), by quintiles of primary income. The overall data means in the final column show that, on average, primary earners in the SE household pay $7,673 in tax. The PT household’s tax is almost double, at $13,575, because, on average, the second

7

8 9

This excludes very few records. Of male partners in the full sample of families, 83.6 per cent are in full time work, 6.7 per cent are in part-time work and 2.5 per cent are unemployed. 27.9 per cent of married mothers are in full time employment. 37.6 per cent are in part time work and 2.3 per cent report being unemployed. Only a quarter of one per cent of families reports both parents as unemployed. Private income is income from all non-government sources such as wages and salaries, profits, investment income and superannuation. See ABS (2005). ‘Full time’ is defined as employed 35 hours per week or more.

200

TAX REFORM IN THE 21ST CENTURY

earner pays $6,533 on an income of only $21,357. For the FT household the family tax burden rises to $17,437, with the second earner contributing $11,548. A striking feature of the results is that the absolute burden on the second income tends to be relatively constant across the distribution. Because second incomes tend to rise with primary income, the highest ATRs on second incomes appear in the bottom two quintiles. The highest ATR, that of 35.1 per cent, applies to the incomes of PT second earners in quintile 2, where the average second income is only $20,320 pa. What this means is that a married mother in quintile 2 who decides to work part time in the market rather than full time at home will, on average, earn around $20,000 and lose over a third in taxes and reduced FTBs. She will also contribute more to GST revenue, because her additional income will be spent at least partly on GST rated goods and services as substitutes for those she could produce herself by working full time at home. Table 2.7: Tax Burdens by Primary Income Quintiles SE Net tax $pa ATR SE % PT Second earnings $pa Tax on second earnings $a ATR2 PT % Net household tax $a ATR PT % FT Second earnings $pa Tax on second earnings $a ATR2 FT % Net household tax $a ATR FT %

1

2

3

4

5

All

–8081 –22.6

–3051 –6.3

2929 4.7

8578 11.4

39161 26.4

7673 11.3

15753 5017 31.8 –3202 –6.7

20320 7126 35.1 4902 7.3

21411 6653 31.1 8945 11.3

23869 6561 27.5 14397 15.2

23846 7062 29.6 37829 25.3

21357 6533 30.6 13575 14.9

21732 7545 34.7 804 1.4

31964 10248 32.1 8772 11.2

37307 11172 29.9 14976 19.6

43554 13507 31.0 22658 19.6

49422 15293 30.9 41905 25.5

36827 11548 31.4 17437 17.3

An objection that may be raised to the preceding analysis is that it fails to control for demographics; for example, it might be argued that the FT household has fewer and older children, and that this accounts significantly for differences in tax burdens by type. To show that this is not the case, Table 2.8 presents ATRs for the representative two-child family that switches type from SE to PT or FT, as defined section 2.3. Recall that the primary earner works full time in all three cases, there is no gender wage gap and non-labour incomes are zero. Thus the second income in the PT household is half that of the primary earner, and in the FT household, equal to primary earnings. The figures for each quintile are calculated for the data means of primary income reported in Table 2.6.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

201

ATRs on the second income in the lower quintiles are even higher than indicated in Table 2.7. For example, the second earner in quintile 2 on only $22,675 pa faces an ATR of over 44 per cent, almost 10 percentage point-higher than the previous result. Table 2.8: Two-Child Family: Tax Burdens by Primary Income

Quintile SE Net tax $pa ATRH SE % PT Tax on second earnings $a ATR2 PT % Net household tax $a ATR PT % FT Tax on second earnings $a ATR2 FT % Net household tax $a ATR FT %

3

1

2

3

4

5

–8873 –27.5

–2949 –6.4

2975 5.3

9335 13.3

35416 28.3

5280 32.7 –3593 –7.4

10008 44.1 7064 10.4

8688 30.7 11663 13.7

11702 33.4 21037 20.0

17881 28.4 53297 28.6

12842 39.8 3969 6.1

16666 36.8 13717 15.1

21876 38.7 24852 22.0

24066 34.3 33401 23.8

42584 34.1 78000 31.2

Household Income: An Unfair Tax Base

If families with the same wage rates and demographic characteristics were observed to make the same time allocation decisions, then, all else being equal, we could reasonably expect to find a strong correlation between household income and family welfare. Under these conditions, a progressive tax on household income would not necessarily be unfair in terms of its distribution of burdens across household types. It would, of course, discriminate against the second earner, but not against low wage two-earner families. However, with heterogeneity in the labour supply of one parent, typically the mother, this is no longer the case. Furthermore, the problem of errors in a welfare ranking defined on household incomes becomes especially serious when, as the analysis to follow will show, the profile of male earnings, and therefore of primary earnings, for full time work is relatively flat across the middle of the distribution and then rises sharply towards the top. Table 3.1 gives data means for the average market hours of primary and second earners, by quintiles of primary income, for the sample of 1945 ‘in-work’ families from the ABS 2003–04 SIH. The sample is again split into the three household types, SE, PT and FT, based on data for employment status, as in Table 2.6. Since the male is the primary earner in the vast majority of cases, the quintile profiles of hours highlight gender differences in labour supply and the high degree of

202

TAX REFORM IN THE 21ST CENTURY

heterogeneity in the market hours of married mothers as second earners, at every level of primary income. Married mothers employed full time work almost the same hours as married men. Those employed part time work less than half male hours. Table 3.1: Labour Supplies of Household Types by Primary Income Quintile SE Primary market hours pa PT Primary market hours pa Second market hours pa FT Primary market hours pa Second market hours pa

1

2

3

4

5

All

2062

2240

2346

2380

2531

2296

2154 1050

2315 1098

2318 1064

2382 1155

2532 1054

2351 1088

2078 2256

2201 2104

2229 2196

2354 2209

2512 2342

2272 2251

Table 3.2 presents data means for primary and second earnings for each household type, by quintiles of primary income. Figure 3.1 presents the results graphically. A crucial feature of the earnings profiles is the relatively flat segment across the middle quintiles. This means that the position of a family in a ranking defined on household income will be very sensitive to the earnings, and therefore to the labour supply, of the second earner because it will take only a small increase in her earnings to shift a family from a low percentile of family income to a significantly higher point in the distribution. Table 3.2: Primary and Second Earnings by Primary Income Quintile SE Primary earnings $pa PT Primary earnings $pa Second earnings $pa FT Primary earnings $pa Second earnings $pa

1

2

3

4

5

All

32212

44816

54238

70071

131886

66362

31409 15753

44439 20320

55987 21411

67056 23869

117324 23846

65357 21357

32739 21732

44763 31964

55480 37307

68637 43554

104446 49422

60437 36827

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

203

Figure 3.1: Primary and Second Earnings by Primary Income 140000

120000

Earnings $pa

100000

80000

60000

40000

20000

0 1

2

Primary SE

3 Primary Income Quintiles

Primary FT

Second FT

4

5

Primary PT

Second PT

This is illustrated in Table 3.3. The table gives the quintile means for household income, followed by the quintile distributions of the three household types. As we would expect, in contrast to the ranking by primary income (Table 2.6), the vast majority of single-earner families are in the lower quintiles. The upper limit of quintile 1 is $50,824, and the lower limit of quintile 4 is $86.857. A single-earner family with an income of, say, $45,000 will be located in quintile 1. If the family switches ‘type’, with the second partner working full time also for $45,000, it will be re-ranked from quintile 1 to quintile 4. If the household has a preschool child, much of the second net income might be spent on childcare. Clearly, such a household could not be said to have the same standard of living as another in which only one parent needs to work full time to earn $90,000 while the other works full time at home. To argue to the contrary it is necessary to assume that home childcare makes little to no contribution to family welfare. Table 3.3: Household Type by Household Income Quintile Household income $pa SE % PT % FT %

1

2

3

38604 67.2 21.9 10.9

60256 42.8 40.1 17.1

77671 23.0 42.1 34.9

4

5

98271 160411 15.3 21.7 41.8 36.2 42.9 41.8

All 87630 33.9 36.4 29.7

The fundamental deficiency of a household income ranking is that it is defined on an income variable that omits home production, appropriately weighted by price. The ranking is driven by the labour supply of the second earner, and is therefore negatively correlated with time allocated to domestic work and childcare.

204

4

TAX REFORM IN THE 21ST CENTURY

Incentive Effects

This section presents data on lifecycle labour supplies, the allocation of time to childcare and domestic work, and government spending on childcare, that offer an explanation for gender differences in labour supply and wage elasticities. The section also presents profiles of household saving over the lifecycle which suggest that high tax rates on second earners, together with limited public funding for childcare, have strong negative effect on household saving as well as on female labour supply.

4.1

HOUSEHOLD LABOUR SUPPLY

Broad comparisons of participation and employment rates are sometimes mistakenly interpreted to indicate that male and female labour supplies have now largely converged. However, household survey data show that there is still a large gap between male and female hours of work in most OECD countries.10 In Australia around 88 per cent of all males and 72 per cent of all females aged from 25 to 59 years are employed – a difference of only 16 percentage points.11 However, over 80 per cent of the males are employed full time, while only 38.4 per cent of females in the same age category work full time. Similar rates are obtained when the sample is limited to couples: 92 per cent of males and 73 per cent of females are employed. Almost all prime-aged married males – 85 per cent – work full time, but only 34 per cent of prime-aged married females are in full time work. The result is that married women work around half the hours of married men. These figures also reveal the very high degree of heterogeneity in female labour supply. While male labour supply shows relatively little variation, with almost all men working full time, females are distributed more evenly between zero hours and full time work. Heterogeneity in female labour supply is strongly associated with children, as we would expect. However, controlling for demographics, as well as for wage rates and non-labour incomes, leaves much of it unexplained. In fact, significant heterogeneity emerges only with the arrival of the first child. To see this, it is useful to compare the labour supplies of couples across four broad lifecycle phases: a prechild phase;12 a 0–4 child (or pre-school) phase in which the youngest child is under five years, a 5–17 child phase in which the youngest child is 5 years or older, and a post-child phase, in which there are no longer children under 18 years present. Figure 4.1 presents histograms of male and female hours of market work in these phases based on 2005 HILDA data for the ‘usual weekly hours of work’ of partners

10 11 12

For comparisons across Australia, Germany, the UK and US, see Apps and Rees (2009). The figures are based on data from the Household, Income and Labour Dynamics in Australia Survey (HILDA), Wave 5, 2005. This phase includes all records in which there are no children present in the household and the female partner is under 42 years.

50

40

Frequency % 20 30

0-4

5-14

15-24 25-34 35-44 45-54 55-64 65-74

Pre-child phase

0-4

5-14

Females

15-24 25-34 35-44 45-54 55-64 65-74

Child 5-17 phase

Males

75+

75+

10

0

Males

Females

10

0

Frequency % 20 30

40

50

10

10 0

Frequency % 20 30

40

50

0

Frequency % 20 30

40

50

0-4

0-4

5-14

5-14

Females

Males

Females

15-24 25-34 35-44 45-54 55-64 65-74

Post-child phase

Males

15-24 25-34 35-44 45-54 55-64 65-74

Child 0-4 phase

Figure 4.1: Household Labour Supplies – Prime Aged Couples

75+

75+

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING 205

206

TAX REFORM IN THE 21ST CENTURY

aged from 25 to 59.13 The first band of the histograms represents 0 to 4 hours and subsequent bands, increments of 10 hours. In the first phase, the profiles closely match – partners of prime working age tend to work full time and for the same hours, suggesting they have the same preferences.14 In the 0–4 child phase, the proportion of men working full time remains about the same, while that of women falls dramatically. At the same time, a high degree of heterogeneity in female labour supply emerges, with 60 per cent remaining in work but less than 20 per cent in full time work. In the 5–17 child phase, full time female employment rises to 31 per cent. Around 24 per cent continue to work less than 5 hours per week. In the post-child phase, the proportion of females reporting working less than 5 hours per week rises to a third, and around a third work full time. Thus, in the post-child phase female labour supply remains well below its prechild level, indicating a high degree of ‘persistence’ in the labour supply decision made in the pre-school phase.15 While the dramatic change in the profile of female hours from phase 1 to phase 2 indicates a strong association between the decision to have children and the labour supply decisions of married women, the high degree of heterogeneity across seemingly similar households suggests that the relationship cannot be captured by the simple view that children ‘cause’ the reduction in female labour supply. To the contrary, we would argue that, given the decision to have children, the observed changes in female labour supply are driven by the economics of investment in the care and education of children, much of which is directly influenced by government policy, and by the gender wage gap. The argument is straightforward. In phase 1 there is a low demand for homeproduced goods and services because there are few of the kinds of goods and services couples in this phase consume for which there are not good, affordable market substitutes, and so there is a low demand for domestic labour in this phase. Put simply, there’s nothing much to do in the home, and so it would make no sense for either partner to specialise in household production, or for singles who have not yet had children to do so. Moreover, the gender wage gap is likely to be less significant in this early phase. These conditions explain why almost all males and all females who have not yet had children, whether single or married, work full time and have close to the same average weekly hours. 13

14

15

The numbers of male and female records in phases 1 to 4 are: phase 1, 330 male and 284 female; phase 2, 597 male and 582 female; phase 3, 742 male and 749 female; phase3: 415 male and 525 female. There are relatively few records in this phase because there are few young married couples without children. However, when we include singles who have not yet had children, and who are therefore essentially in the same lifecycle phase, we obtain similar results from much larger samples. Almost all men and all women not in higher education work full time prior to having children. This is consistent with the results of panel data studies for the US. See, for example, Shaw (1989, 1994).

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

207

The arrival of children creates a very large demand for their care and for investment in their education. While governments has taken over much of the role of investing in the education of children once they reach school age, it has largely neglected to invest in the care and education of those under school age. The result is that market childcare can be very costly. At the same time, as shown in Section 2, the Australian income tax system reduces significantly the net income of the second earner. These policies undermine the capacity of a second earner to finance childcare, especially when her future wage is uncertain and she faces an imperfect capital market in which the borrowing rate is above the lending rate.16 Childcare can be provided by some combination of parental time and services bought in from the market. The opportunity cost of parental childcare is determined by the present value of the current and future net market income foregone. The higher the effective tax rate on the second earner, and the more costly and difficult it is to access market childcare, the more of it will be provided at home, other things being equal. The demand for childcare then implies a large induced demand for household production and introduces a fundamental change in the work choices of couples, which will reflect the relative costs of each partner’s time. Moreover, withdrawal from the labour market by the female as the lower wage partner in phase 2 can lead to a lower wage due to loss of human capital and career possibilities.17 This effect offers an explanation for the strong persistence of female labour supply decisions made in phase 2 into later phases, including the post-child phase.

4.2

LIFECYCLE LABOUR SUPPLY AND SAVING

In this section we extend the preceding analysis to include seven phases. We present more detailed evidence on the time use of household members outside the market and on labour supply and saving, based on data for all couples drawn from the ABS 2003–04 Household Expenditure Survey (HES) combined with information on time use from the ABS 1997 Time Use Survey (TUS).18 The sample of couples contains 4228 records. To highlight the lack of government support for children below school age, we partition households in the ‘Child 0–4’ phase in Figure 4.1 into two sub-groups: families with pre-school children only and those who have both preschool and older children. We also split the ‘Child 5–17’ phase into two sub-groups: families with school children aged 5 to 15 only and those who also have older dependent children. The latter phase includes families with dependent children aged 18 and over who

16 17

18

See Apps and Rees (2003). An extensive literature on work-related human capital accumulation includes the contributions of Eckstein and Wolpin (1989), Altug and Miller (1998) and, more recently, Imai and Keane (2004) and Olivetti (2006), among others. For further detail, see Apps and Rees (2003).

208

TAX REFORM IN THE 21ST CENTURY

are still in school or in a post-school education program. There are therefore seven phases as follows: (1) (2) (3) (4) (5) (6) (7)

adult members do not yet have children children aged 0–4 only – preschool age children aged 0–4+ older children school children aged 5 and under 15 dependent children aged 15–24 years – high school, tertiary, etc. adults are still of working age but children have left home adults are retired

Table 4.1 presents weighted data means for the allocation of time to market and domestic work, and to home childcare, in each of these phases.19 The table also reports median household saving in each phase.20 Figure 4.2 shows graphically the profiles of male and female labour supplies across the phases and Figure 4.3, the allocation of time to domestic work and child care. Table 4.1: Market, Domestic and Childcare Hours and Household Saving

Childcare

Market

Domestic

1 2 3 4 5 6 7

1820 799 694 1021 1263 1113 70

896 1215 1402 1475 1431 1494 1665

– 2291 2164 1508 234 – –

2140 2125 2076 2046 2036 1953 74

605 668 703 798 637 806 1342

Childcare

Phase

Domestic

Male Hours pa

Market

Female Hours pa

– 872 841 522 144 – –

Median saving*

7754 3315 506 1910 6152 7360 –112

*2003–04 $pa

19 20

The number of records is 503, 385, 383, 644, 592, 762 and 959 in phases 1 to 7, respectively. Saving is computed as total weekly household income less total expenditure excluding the principal component of mortgage repayments, capital housing costs and superannuation and life insurance. Note that total expenditure includes income tax and so saving is, in effect, computed as net income less consumption expenditure.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

209

2000 0

1000

Hours pa

3000

4000

Figure 4.2: All Couples: Labour Supply by Gender

1

2

3

4 5 Lifecycle phase

Male labour supply

6

7

Female labour supply

2000 0

1000

Hours pa

3000

4000

Figure 4.3: Domestic Work and Childcare by Gender

1

2

3

4 5 Lifecycle phase

6

7

Male domestic hours

Female domestic hours

Male dom+ccare hours

Female dom+ccare hours

The time use profiles provide evidence of strong substitution of household production, consisting mostly of childcare, for market labour supply by married women. In the pre-child phase, the allocation of time to household production involves only domestic work and is relatively low. The time allocation decisions of partners in phase 2 are very different. Female market hours drop to well below half their phase 1 level. Time spent on household production rises dramatically, to a total of 3406 hours in phase 2 due to over 2000 hours of childcare. In phase 3, female hours drop even further and childcare hours remain at much the same level. This is

210

TAX REFORM IN THE 21ST CENTURY

not surprising because families in this phase still have an average of 1.16 preschool children, as well as school children. In phase 4 there are no pre-schoolers, and so household production hours fall significantly due to a fall in childcare hours, and market hours rise from their low point of 694 hours pa to 1021 hours pa. In phase 5, female hours rise to 1263 pa, a level that is still well below the phase 1 level and well below male hours, even thought the children are in their late teenage years or in tertiary education. On entering the post-child phase, female market hours tend to fall while domestic work remains significantly higher than in phase 1. Males also increase their total hours of household production from phase 2 onwards, but only to a level that is a small fraction of female hours. The additional hours are due to time spent on childcare. These data indicate that, unlike married females, who substitute away from both market work and leisure from phase 2 onwards, married males tend to substitute away from leisure only. The data means indicate that, on average, females have less leisure than males during the childrearing years. The fall in the leisure of both partners suggests that the household is not using the capital market to smooth consumption.21 From the profile of median saving it might appear that household saving tracks the presence of children. However, as we show later, household saving strongly tracks female labour supply. The household’s consumption expenditure on market goods, as opposed to its implicit expenditure on the allocation of time to domestic work and childcare, also tracks female labour supply because it draws on household income which, as shown in Section 3, tracks female labour supply. The data available on in-kind government benefits suggest that at least part of the observed gender differences in time allocation decisions can be attributed to variation in government support for childcare and education with the age of a child. Table 4.2 presents lifecycle profiles of bought-in childcare and government spending on in-kind childcare and education benefits. The table also lists the number of dependent children in each phase. The lifecycle profile of government in-kind childcare and school benefits shows the dramatic rise in government spending when a child reaches school age. In phase 2, in-kind childcare benefits average only $1413 pa. In phase 3 it is slightly higher, at $1626. However, because families in this phase have an average of 1.4 school children, they receive over $10,000 in additional education benefits, which amount to an in-kind subsidy of over $7000 per school child. There is still an average of 1.2 preschool children present and so it remains very costly for the second earner to go out to work. In phase 4, government spending on in-kind education benefits rises to $13,905 pa. There are no pre-schoolers and so it is not surprising to find that female hours rise significantly, and continue to rise into the next phase.

21

See Apps and Rees (2003).

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

Table 4.2:

211

Government In-Kind Childcare and Education Benefits # Dependent Children 1.41 2.57 1.93 2.85

Phase 2 3 4 5

In-Kind Childcare Benefits* 1413 1626 256 –

In-Kind School Benefits* 424 10795 13905 9219

* 2003–04 $pa

The preceding figures are data means for a sample of households that make very different decisions concerning time use and saving within each phase. To examine further the issue of female labour supply heterogeneity and household saving, we partition households into two ‘types’ across phases 2 to 6, defined on median female hours within each phase. We label households in which the female works less than median hours type H1 and those in which she works above median hours, type H2.

2000 1000 0

Hours pa

3000

Figure 4.4: Labour Supplies by Household Type

1

2

3

4 5 Lifecycle phase

6

7

H1: Male market hours

H1: Female market hours

H2: Male market hours

H2: Female market hours

Figure 4.4 presents the labour supply profiles of each type across phases 2 to 6. The results reflect the especially strong substitution of home childcare for market work by the H1 household, together with substitution away from leisure. In the case of the H2 household, there is much stronger substitution away from leisure, suggesting that married mothers who choose to work longer hours in the market are faced with having less leisure over the entire lifecycle. Note that the average number of children in the H1 household is only marginally higher than in the H2 household in each phase.22 22

The average number of children in the H1 household is 1.51, 2.77, 1.97 and 1.58, in the H2 households, 1.30, 2.36, 1.89 and 1.45, in phases 2 to 6 respectively.

212

TAX REFORM IN THE 21ST CENTURY

Figure 4.5 plots the lifecycle profiles of median saving for each household group, and the profile for the full sample. The figure shows graphically the strong tendency for household saving to track female labour supply. This is not surprising, given that the H1 household has chosen to substitute domestic for market work and therefore has less market income available for the purpose of saving. Median household saving of H1 households is well below that of H2 households from phase 2 to retirement. In phase 1, median saving is high because almost all female partners work full time. In phase 2 there is a sharp fall in the median of the full sample, due to a dramatic fall in the saving of H1. The median saving of the H1 household is actually negative in phases 2 and 3.

5000 -5000

0

Saving

10000

15000

Figure 4.5: Lifecycle Saving by Household Type

1

2

3

4 5 Lifecycle phase

All Saving

6

7

H1 Saving

H2 Saving

The diverse saving decisions of the two household groups cannot be attributed to family size. Controlling for the effects of family size, lifecycle phase and the net income of each partner using regression analysis generates profiles that show a wider gap in household saving. The result is driven by a coefficient on the net income of the female partner that is around twice that on primary net income, indicating a much higher propensity to save from her net earnings.23 An additional child is found to have a large negative effect, but because the two household types have close to the same number of children, family size explains very little of the additional saving of the H2 household, evaluated at data means. A formal model of the joint

23

The result was found to hold across a wide range of model specifications. The models were estimated on a sample that excluded the bottom five per cent of male net incomes, negative female incomes, and the top 1 per cent of male and female net incomes, to remove the effect of outliers.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

213

determination of female labour supply and household saving over the lifecycle that generates similar profiles can be found in Apps and Rees (2003).24 These results suggest that household decisions concerning labour supply and household saving are made simultaneously, and in response to net-of-tax wage rates and the price of market childcare, as well as interest rates.

5

Concluding Comment

We have shown how changes in various tax policy instruments – the PIT scale, the LITO, ML and FTBs – have shifted the overall burden of the Australian income tax system toward the middle of the earnings distribution and towards low- and averagewage two-earner families. This direction of reform has been pursued in a labour market with rising average wages but increasing earnings inequality, and therefore cannot be supported on the basis of conventional equity criteria. The reforms also run counter to efficiency rules, by imposing high tax rates on the second earner, typically the female partner with the more responsive labour supply. High tax rates on the second earner are a characteristic of joint income taxation. In a detailed analysis of the structure of marginal and average rates we show that most Australian families are now taxed on the basis of joint income under an MTR scale that exhibits an inverted U-shaped profile. A system of this kind is widely recognised to have strong negative effects on female labour supply and productivity, and therefore on the future tax base for funding family support. The fundamental limitation of the system lies not in the level of FTBs, but in its effective MTR and ATR structure, created by applying an MTR scale with an inverted U-shaped profile to joint income. Taxing families in this way seriously inhibits the reallocation of female time from the household to the market during a period of declining fertility and therefore of falling demand for domestic labour. With population ageing, the present level of FTBs is therefore likely to become unsustainable due to productivity losses from labour supply disincentive effects. The system is also unfair because it taxes second incomes at very high rates and therefore ignores the fact that of two households with the same total household income, where one has the second earner working entirely in the market, the other entirely in the household, the latter will have a significantly higher standard of living because of its higher level of output of household goods and services.

24

An essential feature of the model is that it takes account of the presence of two adults in the household. Much of the literature on saving behaviour treats the household as a single decision unit. See, for example, Blundell et al. (1994) and the survey by Browning et al. (1996). In models of this kind, high wage H1 households are confused with much lower wage H2 households because, in effect, the models fail to control for wage rates. The studies therefore miss the strong positive association between female labour supply and household saving at a given wage level.

214

TAX REFORM IN THE 21ST CENTURY

The solution to these limitations of Australia’s ‘new’ income tax system lies in a return to the taxation of individual incomes under a progressive rate scale, combined with universal family benefits. There is also an urgent need for public investment in a high quality, education-oriented, childcare sector, to allow the expansion of female labour and to raise the standard of education of the next generation. The combination of these policies can be expected to be more than self-financing, because of the resulting growth in the tax base and productivity gains References Altug, S. and R.A Miller (1998), ‘The Effect of Work Experience on Female Wages and Labor Supply’, Review of Economic Studies 65: 45. Apps, P.F. and R. Rees (1999), ‘Individual vs. Joint Taxation in Models with Household Production’, Journal of Political Economy 107: 393. Apps, P.F. and R. Rees (2003), ‘Life Cycle Time Allocation and Saving in an Imperfect Capital Market’, Paper presented at the NBER Summer Institute 2003 in the session on Aggregate Implications of Microeconomic Consumption Behavior, Boston, July 21–25 (Available as DP 1036 at ). Apps, P.F. and R. Rees, (2007), ‘The Taxation of Couples’, IZA Discussion Paper 2910. Apps, P.F. and R. Rees (2009), Public Economics and the Household, (Cambridge: Cambridge University Press). ABS (2005), ‘Survey of Income and Housing – Confidentialised File’, Technical Paper 2003–04 (revised). Blundell, R., M. Browning and C. Meghir (1994), ‘Consumer Demand and the LifeCycle Allocation of Household Expenditures’, Review of Economic Studies 61: 57. Boskin, M.J. and E. Sheshinski (1983), ‘Optimal Tax Treatment of the Family’, Journal of Public Economics 20: 281. Browning, M. and A. Lusardi (1996), ‘Household Saving: Micro Theories and Micro Facts’, Journal of Economic Literature 34: 1797. Eckstein, Z. and K.I. Wolpin (1989), ‘Dynamic Labor Force Participation of Married Women and Endogenous Work Experience’, Review of Economic Studies 56: 375. Eissa, N. and H. Hoynes (2004), ‘Taxes and the Labor Market Participation of Married Couples: The Earned Income Tax Credit’, Journal of Public Economics 88: 1931. Eissa, N. and H. Hoynes (2006), ‘Behavioral Responses to Taxes: Lessons for the EITC and Labor Supply’, in J.M. Poterba (ed.), Tax Policy and the Economy 2nd ed., (Cambridge, MA: MIT Press) 163. Eissa, N. and J.B. Leibman (1996), ‘Labor Supply Responses to the Earned Income Tax Credit’, Quarterly Journal of Economics 112: 603.

PATRICIA APPS & RAY REES: TAXATION, LABOUR SUPPLY AND SAVING

215

Shaw, K. (1989), ‘Life-Cycle Labor Supply with Human Capital Accumulation’, International Economic Review 30: 431. Shaw, K. (1994), ‘The Persistence of Female Labor Supply: Empirical Evidence and Implications’, Journal of Human Resources 29: 348.

Chapter 8

The Distributional Effect of Consumption Taxes in Tax Systems Neil Warren* 1

Richard Musgrave on Tax Incidence

Richard Musgrave believed strongly in the role of government in society. As he saw it, the role of the economist was to contribute in such a way as to make government function better for the benefit of all in society and to achieve this by bridging the gap between the theory and practice of good government. This focus by Musgrave is no more apparent that in his writings on the issue of tax (and fiscal) incidence as evident in his early contributions on income tax progression (Musgrave and Thin, 1948) and estimating tax incidence (Musgrave et al., 1951; Musgrave, 1953a, 1953b; and Krzyzaniak and Musgrave, 1963). On his visits to Australia to participate in conferences organized by John Head (e.g. Head, 1983), my own empirical studies of tax incidence were the subject of some debate over their obvious limitations. However, at no stage did Musgrave express doubt about the merits of undertaking such studies. This is not surprising given his hallmark pursuit of better government. While empirical studies of the distributional impact of personal income taxes have a long history (including those undertaken by Musgrave in Musgrave and Thin (1948) and Musgrave et al., (1951)) and have since developed a high level of refinement, the same cannot be said for empirical studies of consumption tax *

Australian School of Taxation, University of New South Wales. This paper is based on N.A. Warren, A Review of Studies on the Distributional Impact of Consumption Taxes in OECD Countries’, OECD Social, Employment and Migration Working Paper No 64 OECD, http:// www.oecd.org/els/workingpapers.

218

TAX REFORM IN THE 21ST CENTURY

incidence. This is despite the increasing contribution to government revenue from this tax source. In this paper, we review the progress made over the past four decades into the measurement of consumption tax incidence, while also conjecturing over the future directions for such research. It will be apparent throughout this paper that the influence of Richard Musgrave’s writings on incidence-related issues has been profound – although he would probably be disappointed at how little progress has been made in measuring consumption tax incidence over past decades.

2 Why Distributional Impact Studies Despite early advances in modelling the incidence of all taxes,1 comprehensive tax incidence studies remain relatively few with most studies limiting their focus to the distributional impact of personal income taxes and social security levies (as well as social welfare benefits). The result is that reforms to personal income tax and social welfare policies are almost always now accompanied by detailed analysis of their impact on various household types.2 In contrast, such analysis remains an exception in the case of consumption tax reforms.3 This position cannot be explained by a lack of academic research into consumption tax incidence – as many such studies have been undertaken over past decades. Rather, it reflects two challenges4 confronting tax incidence studies. The first is that, unlike the broad agreement existing on the conceptual approach to adopt when estimating the incidence of personal income tax, no such agreement exists on how to model the incidence of consumption tax on individuals. The second challenge is 1

2

3

4

Studies of the distributional impact of all government taxes on individuals have a long history. Gillespie (1965) and Dodge (1975) undertook early studies in Canada and Pechman and Okner (1974) and Reynolds and Smolensky (1977a, 1977b) in the United States. Bentley, Collins and Drane (1974) and Warren (1979) estimated tax incidence in Australia, while the UK Office of National Statistics produced some of the earliest official estimates of tax incidence in the 1950s (with Australia’s first official estimates by the Australian Bureau of Statistics in 1984 based on the UK methodology). However, such estimates rarely incorporate labour supply responses, including only a static snapshot of the pre- and post-reform situations (i.e. assuming no responses by individuals). Detailed modelling of the distributional impact of consumption tax reforms is, however, conducted in some countries, such as Australia. See Chapter 5 of the Australian Government document outlining the impact of the then-proposed 10% GST, Tax Reform: Not a New Tax, A New Tax System (1998) accessible from . See. Barrett and Wall (2005), Creedy (2001, 2002), Decoster (1995), Decoster, De Swerdt and Verbist (2007), Decoster, Schokkaert and Van Camp (1997), Decoster and Van Camp (2001), Garfinkel, Rainwater and Smeeding (2006), Harding, Lloyd and Warren (2004, 2006, 2007), Kaplanoglou and Newbery (2003), Kaplanoglou (2004), Liberati (2001), Madden (1995), O’Donoghue, Baldini and Mantovani (2004), Newbery (1995), and Tsakloglou and Mitrakos (1998).

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

219

that the data demands of such studies are not readily met. In combination these two facts mean that very different approaches can be taken to modelling the incidence of consumption tax on individuals. This paper takes stock of the progress, past and present, of efforts to estimate the distributional impact of consumption taxes, with the aim of better understanding how such studies are undertaken (Section 4), what have been their findings and what benefits might be derived from including consumption taxes in inter-temporal and cross-country comparisons of the impact of government activities on individuals (Section 5).

3 Consumption Taxes – Important and Different Figure 1 presents data on the importance of the main categories of taxes across all OECD countries. Both the level and mix of tax revenues are shown to vary widely over time, even for similar categories of taxes. In the case of consumptionbased taxes, while their contribution to total tax revenue has decreased, on average, from 37 per cent to 30 per cent over the period 1965 to 2004, this has come about as a result of a significant change in their mix. There has been an increase from 15 per cent to 19 per cent in the share of general consumption taxes (e.g. VAT/ GST) and a fall from 22 per cent to 11 per cent for taxes on specific goods and services (such as excise on petrol, tobacco and alcohol). Further, despite the falling contribution of total consumption taxes to total tax revenue between 1965 and 2004, their size relative to GDP has risen from 9.6 per cent to 10.8 per cent. This is also at a time when the mix of income-based taxes has changed away from personal income taxes towards social security contributions, especially from employers.5 Because of these trends, any study of changes in the distributional impact of government on individuals which omits consumption taxes or social security contributions will lead to biased results. Figure 2 presents data on tax to GDP ratios in all OECD countries in 2004, with countries ranked by the level of their taxes on goods and services (OECD Tax Classification 5000). What is apparent is that while personal income taxes and social security taxes on employees are a significant proportion of GDP, they are nevertheless just part of the total tax burden. For many countries high consumption taxes are accompanied by low personal income taxes and vice versa. In 11 countries taxes on goods and services contributed more to total tax revenue than taxes on personal income and social security levies on employees combined; in 20 countries, taxes on all goods and services exceeded taxes on personal income; 5

In the case of taxes on personal income, there were equally significant changes in the mix. Between 1965 and 2004, the personal income tax as a proportion of GDP fell, especially in recent years, primarily due to cuts in personal income tax rates, while employee social security contributions initially increased but have stabilized over the past decade. This is all at a time when employer social security contributions increased significantly, both as a proportion of GDP and as a share of total taxes.

0

10

20

30

40

50

60

70

80

90

100

1965

14.6

22.5

10.9

8.5

10.3

5.9

27.2

1970

15.7

19.3

10.4

8.2

11.0

6.1

29.3

1975

15.3

16.0

9.0

7.3

14.0

7.0

31.5

1980

16.0

15.0

6.7

7.5

14.8

7.3

32.7

a. Percentage share of total tax revenues

1985

16.6

14.6

7.6

8.0

14.3

7.8

31.1

1990

17.4

12.1

10.0

7.7

14.1

7.9

30.8

1995

17.7

12.4

10.3

7.9

15.1

8.6

27.9

2000

18.6

10.9

10.4

10.1

15 0 15.0

8.2

26.8

Average of 30 OECD Countries

2004

19.2

10.9

11.0

9.6

15 4 15.4

8.5

25.5

Figure 1: Trends in Tax Revenues in OECD Countries

General Consumption

Specific Goods and Services

Other

Corporate income

Employer's social security contributions

Employees' social security contributions

Personal income

220 TAX REFORM IN THE 21ST CENTURY

1965

3.8

5.4

5.8

1970

4.4

2.3

2.9

1.7 3.1

1.5 2.7 2.2 2.8

7.0

8.2

1975

4.6

4.8

2.2 2.7

4.2

2.1

9.4

1980

1985

5.5

4.8

4.7 5.0

2.5

2.6

4.7

2.6

2.4 2.1

4.7

2.3

10.3

10.2

Source: OECD Revenue Statistics (2006) 1965–2005

0

5

10

15

20

25

30

35

40

b. Percentage Share of GDP

1990

5.9

4.1

3.4

2.6

4.8

2.7

10.5

1995

6.2

4.4

3.6

2.8

5.3

3.0

9.8

2000

6.8

4.0

3.8

3.7

5.5

3.0

9.8

2004

6.9

3.9

3.9

3.4

5.5

3.0

9.1

General Consumption

Specific Goods and Services

Other

Corporate income

Employer's social security contributions

Employees' social security contributions

Personal income

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES 221

2 3

4

4

4

5

4

2 3 2

3

10

4

4

4

6

13

5

3

5

3

2

12

4

4

4

4

3

3

6

3

5

3

2

6

6

2 2 4

6

8

6

4

6

3

2

10

4

7

3 1

5

7

3

4

4

4

14

7

3

7

3

4

7

7

3

3

4

1

8

7

4

5

3

7

6

7

4

5

3

3

10

7

4

3

5

4

5

9

2

3

6

15

8

3

6

2

6

10

6

5

4

6

5

7

8

2 2 4

3 3

8

3

3

3

5

5

8

4

2 2

14

4

8

2 4

10

3

10

8

4

3 2

3

5

9

3

4

3

3

16

9

5

4 2

2

13

7

7

2 3

3

5

5

4

1 3

4

2 2

2

7

11 10 11

4

1 4

14

25

Taxes on general consumption (5110)

Taxes on specific goods and services (5120)

Other Taxes (including Payroll and Property Taxes)

Taxes on corporate income (1200)

Employers' social security contributions (2200)

Employees' social security contributions (2100)

Taxes on personal income (1100)

Source: OECD Revenue Statistics (2006) 1965–2005.

Note: Countries are ranked, from left to right, in increasing order of taxes on goods and services (5000) relative to GDP

0

5

2 3 2 2

3

15

10

9

20

25

30

35

40

45

50

Taxes in Percentage of GDP in 2004

Figure 2: Level of Tax Revenues Across OECD Countries

222 TAX REFORM IN THE 21ST CENTURY

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

223

while in 10 countries taxes on general consumption exceeded those on personal income. These differences imply that any inter-country comparison of the impact of government taxes on individuals which omits consumption taxes will yield biased results because of both the different level (Figure 2) and mix (Figures 1a and 1b) of these taxes. Even including only general consumption taxes (such as a VAT/ GST) will not fully address this problem because of cross-country differences in the composition of consumption taxes (Figure 1) and in their size. The case for including consumption taxes along with personal income tax and employee social security contributions taxes in any inter-temporal or cross-country comparisons of the impact of government tax policies is therefore clear. Even if the contribution and composition of consumption taxes remained similar and unchanged over time and between countries, studies focusing only on personal income tax and employee social security contributions will provide only a partial assessment because the incidence of these two groups of taxes is significantly different.

4

Modelling the Distribution of Consumption Taxes: Theory and Practice

Taxes on consumption are typically collected by an intermediary, such as a retailer. However, such an intermediary cannot ultimately bear the burden of a consumption tax as this must rest on individuals – either as consumers, recipients of income or owners of assets. The question is how, in the presence of an intermediary, a consumption tax comes to be passed through to its ultimate bearer. The difficulties posed by this question go a long way to explaining why most tax incidence studies focus on personal income taxes and government cash transfers (which are in effect negative income taxes). In the case of personal income taxes, the convention is to assume that the economic (or final) incidence is on the recipient of the income flow: this enables the economic incidence of personal income taxes to be estimated using survey data on individual incomes. In the case of a consumption tax, even when household surveys collect data on household consumption expenditures, the tax burden on this expenditure is not obvious because the link between the statutory (or legal) and economic (final) incidence of the tax is complex. However, overlooking the distributive implications of consumption taxes simply because of the modelling complexity that this involves is fraught with dangers. In practice there has always been a demand for empirical estimates of the distributional impact of consumption taxes. The challenge is to ensure that such estimates are suitably qualified, fully explained and tested for sensitivity using the best available theoretical knowledge and data sources. It is for this reason that this section sets the scene by addressing three critical issues confronting all empirical studies of the incidence of consumption taxes on individuals within households.

224

TAX REFORM IN THE 21ST CENTURY

(1) How are taxes allocated to individuals? If all consumption taxes are ultimately shifted to individuals, how does this happen; how are these taxes valued; and what approach is used to assess the incidence of the tax? (2) What is the scope of the study? Does it cover all taxes or only selected ones; just one year or a lifetime; and all the population or just some of the population? (3) How is the total burden of consumption taxes distributed between individuals? Important here is the base for inter-unit comparisons (dollars or some welfare unit; annual or lifetime); the unit of analysis (person, family or the household); and the equity measures used. The answers given to each of these questions will reflect the objective of the empirical study and the data sources available. In turn, the decisions taken can fundamentally influence the meaning and interpretation of the results obtained.

4.1

HOW ARE CONSUMPTION TAXES ALLOCATED TO INDIVIDUALS?

4.1.1

Theory

Who bears the tax burden? Answers to this simple question are the subject of considerable debate. What is commonly accepted is the interest in the answer. Not only are communities concerned with equity, efficiency and simplicity of the tax system, so too are policy makers because of the impact of taxes on the economy, politicians and the voting public. The problem is that taxes are imposed in a multitude of ways across a range of different bases. Any study of the distributional impact of taxation must therefore begin with an attempt to estimate the ultimate burden (or economic incidence) of taxation on individuals. This section considers the conceptual issues raised in addressing this question. Details of tax revenue collected by government can be readily obtained from annual budget documents and official statistics. This is, however, only part of the burden of tax. By distorting behaviour, a ‘deadweight loss’ (DWL) or efficiency loss is created through the imposition of the tax burden. In addition, the collection of taxes imposes a compliance cost on taxpayers and a cost on government to administer the tax. In practice, the DWL and the administrative and compliance costs of taxation are rarely comprehensively measured and are almost never considered in studies of the distributional impact of consumption taxation. It could be argued that this approach is not unreasonable in studies focussed on the differential incidence of taxes, since here attention is on revenue-neutral alternatives to the current tax system which implies minimizing (or effectively removing) any first-order (behavioural responses) and second-order (broader macroeconomic) effects. In practice though, most consumption tax incidence studies estimate the absolute incidence of consumption taxes accompanied by the adoption of the simple

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

225

(and strong) assumptions that all consumption taxes are fully passed through to individual final consumers and that DWL and compliance and administrative costs are not important. 4.1.2 Modelling Consumption Tax Incidence in Empirical Tax Incidence Models Four primary data sources are required in any empirical tax incidence study: taxation statistics; aggregate income and expenditure statistics; income and expenditure data for individual households; and information on tax shifting. Data on taxation can be obtained in the form of either aggregate official national data or in disaggregated form. Aggregate tax data is readily collected by all governments and reported in annual budget statements, official statistical publications, and in international compendiums based on some common classification system such as the OECD (2006). This data can be complemented with those reported by the tax revenue collection agencies. However, these statistical sources only report data on what is collected and do not provide guidance as to how such taxes ultimately come to be borne by individuals in households. To this end, two additional statistical sources must be accessed. The first are surveys of household income and expenditure. Most surveys collecting data on individual income also collect information on the personal income tax and employee social security contributions paid. However, only a few surveys also collect data on consumption expenditure, and those that do generally do not provide information on the taxes hidden in this expenditure. What is required is insight into how taxes on the inputs and outputs of producers and distributors come to be passed forward to individuals as consumers. It is here that the National Accounts Input-Output data play a critical role. Input-output tables allow identification of how the statutory (or legal) incidence of consumption taxes ultimately flows through to household final consumers (economic incidence). Figure 3 outlines an input-output framework of how consumption taxes come to be ultimately borne by individuals, assuming their full forward shifting to individual consumers (Warren 1998). This framework provides a way of thinking about how taxes whose statutory incidence is on firms are ultimately passed through to individuals as the final consumers and will form the basis for comparing the different methodologies applied in empirical studies examined in Section 4. Consumption taxes can be divided into two categories: those on inputs into the production and distribution of goods and services (labelled as INTAX in Figure 3) and those on expenditure by final consumers (FDTAX, i.e. FD1 to FD7) of which households are one such consumer. Those consumption taxes which are imposed directly on households when they purchase from a retailer are shown by A. However, this is only part of the total consumption tax collected by government. Also embedded in retail sales to households is the tax on intermediate inputs into the production of those goods shown by B in Figure 3. However, these intermediate

226

TAX REFORM IN THE 21ST CENTURY

Figure 3: Modelling of Consumption Taxes Under the Assumption of Full Forward-shifting

INTAX: Consumption Taxes on Inputs into the Production Process

B

FDTAX: Consumption Taxes directly on Final Demand Expenditure FD2

FD1 FCE Households

GFKE Public Enterprises

FD3 GFKE Private

FD4

FD5

FD6

Increase in Stocks

FCE Government

GFKE Government

FD7 Exports

C1 C2 Household Consumption

A

Impacts on resident households

D2 Impacts on resident households

D1 No impact on any households

E2 Impacts on resident households

E1 Impact on nonresident households

FCE = Final Consumption Expenditure GFKE = Gross Fixed Capital Expenditure

.

Source: Author’s elaboration.

taxes impact not only on households (B) but also on the final consumption of nonhousehold final consumers (FD2-FD7). While the pass through of A+B to household final consumers is relatively straightforward, the important question is how intermediate and final consumption taxes on FD2-FD7 become incident on resident households. Different approaches have been adopted to address this issue, and these are shown schematically in Figure 3. A common approach is to assume that taxes on investment goods by private and public enterprises (or FD2 and FD3) are also inputs into the production of goods for FD1 and FD4-FD7 and should therefore be modelled as passed through (shown by C1) to the other categories of final consumers.6 Even when adopting this approach, the question remains as to how those taxes impacting on FD4 to FD7 impact on resident households. With respect to consumption taxes paid by general government (FD5 and FD6), two approaches are commonly adopted. The first simply ignores this tax 6

It will be shown in Section 5 that this is the approach adopted officially by the UK ONS (2007), Australian ABS (2006) and Statistics Canada (SDSP/M 2007) and in Scutella (1999).

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

227

and consequently reduces the nominal amount of government expenditure by the amount of this tax (D1). The alternative is to allocate this tax burden to individuals (D2) while leaving government expenditure unchanged in nominal terms. Changes in stocks of goods (FD4) are generally either ignored (D1) or allocated to current consumers (D2). For taxes incident on exports (FD7) two basic approaches have been adopted: that they ultimately impact on non-resident consumers when they consume these goods in their countries of residence (E1); or that countries cannot ‘export’ its tax burden if this makes price uncompetitive in international markets. In this second case, the assumption is that a country’s exchange rate will adjust to restore its competitiveness and this impact is distributed across households, as in the case of taxes on resident households (shown by E2). As Figure 3 illustrates, even assuming the full forward shifting of consumption taxes to households does not yield a simple approach to estimating the incidence of consumption taxes on individuals in domestic households. This is true even if we ignore the possibility that some of these consumption taxes might become incident on the owners of factor inputs (labour and capital) into the production of these goods or on the suppliers of produced inputs into their production.

4.2 WHAT IS THE SCOPE OF THE STUDY 4.2.1

Coverage of Taxes

Although this study is only concerned with consumption taxes, what constitutes a consumption tax is not beyond dispute. After all, if interest is in those taxes which are ultimately incident upon consumption, then this may be more than just those taxes which are traditionally seen as levied on consumption. A comprehensive approach might include taxes on capital and labour inputs into the production and distribution process, or those taxes on capital-related transfers, that are ultimately shifted to consumers. A more basic question is whether a study of the distributive impact of consumption taxes also includes subsidies (negative consumption taxes), consumption taxexpenditures, non-tax revenue such as user-pays fees and charges and dedicated (ear-marked) taxes – or the shifting of other taxes onto consumption. Exclusion of such taxes will result in a partial view of the impact of government taxes on consumption. 4.2.2

Population Coverage

As important as the coverage of consumption taxes is the coverage of the population on whom these taxes are incident. At its broadest, this would include both residents and non-residents, the latter incurring a burden as a result of exports of domestically produced goods and services. In practice, consumption tax incidence studies only focus on the domestic population and domestic taxes.

228

TAX REFORM IN THE 21ST CENTURY

One complication with using household survey data is that its coverage is narrower than the overall population. This is because household surveys typically exclude those in non-private dwellings or in remote areas (ABS 2006, p. 34) – resulting, in Australia, in only 98 per cent population coverage. Difference in coverage is common but depends on the approach taken to modelling tax incidence. 4.2.3

Period of Study?

While information on tax collected has an annual focus, this period has no intrinsic economic significance, other than as an accounting concept. In the case of tax incidence studies, more relevant is how the tax impacts over an individual’s lifetime or how that burden varies during their lifecycle.7 Data availability is a key constraint here. While cross-section data provide a snapshot of tax incidence, they can also be applied to the study of lifecycle tax incidence (by for example, examining age ranges for household heads). Crosssection data has also been used to create hypothetical lifetime data. With the increasing prevalence of panel data, lifetime tax incidence is becoming the focus of increased attention. Nonetheless, when interest is in the immediate plight of citizens, the focus of research on the distributional impact of consumption taxes will be on the ‘here and now’. It is therefore not surprising that most studies focus on a point-in-time snapshot of tax incidence with some examination of how it varies across different ages of the household head.

4.3

EVALUATING THE DISTRIBUTION OF CONSUMPTION TAXES

Knowing the total tax collected from some particular commodity is only the first step in assessing the distributional impact of these consumption taxes across different individuals in households. The next crucial step is to assess the distribution of these taxes between different individuals, and this requires a number of decisions which have the potential to critically influence any evaluation of the distributional impact of these taxes. 4.3.1

Unit of Analysis

A key consideration in any distributional impact study is the unit of analysis. In practice, the unit chosen is influenced by the objective of the study and the household survey data available. In the case of the survey data, this is collected data at three levels: persons, families, and households. The problem is that all data is not available for all units. In the case of income data, it is collected in surveys according to who is the recipient (e.g. individuals for wages and salaries and the family level for income flowing from jointly held assets). With expenditure data, it is only available at the household level where the household is defined as a group 7

See further discussion of this issue in 5.3

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

229

of individuals who live together and have common housekeeping arrangements. Moreover, since no two households are alike in terms of their socio-economic and demographic characteristics, focusing on the distributional impact of consumption taxes across households is unlikely to be that informative. The solution is to focus on that unit which is common to both families and households – the individual – while accommodating key differences between households so as to allow interunit comparisons. 4.3.2

Base for Inter-unit Comparisons

Undertaking inter-unit comparisons involves a two-staged process: firstly, determining the means available for satisfying a household’s needs; and secondly, recognizing there are economies of scale within different households. With reference to the first element, most studies rely on an income concept derived from a crosssectional household survey. While annual income is commonly adopted in studies of the incidence of income taxes, there is some debate about the appropriateness of this measure in studies of the incidence of consumption taxes. The issue here is that while consumption taxes appear to be regressive based on annual income, they are likely to be less regressive and even progressive when their effect is assessed over an individual’s lifetime, (Creedy, 1999, 2002; Poterba 1989; Metcalf, 1997; and Fullerton and Rogers, 1993). Critical in this debate is the appropriateness of using annual income as a measure of an individual’s potential well-being when consumption may depend also on an individual’s lifetime income. People know that over their lifetime, their annual income will first be low, peak in middle age and then fall in old age, and they will factor this in when determining annual consumption. The result is a relatively stable level of annual consumption over the lifecycle, despite fluctuations in annual income. This implies that cross-sectional studies of consumption tax incidence based on annual income will present more regressive results for the young and the old (lower-income groups often dissaving) than for the middle aged (higher incomes with positive savings). Clearly, dissaving by low income groups as shown by annual income and expenditure surveys is not sustainable, and care must therefore be taken when interpreting the distributional impact of consumption taxes using such data sources. This argument suggests that a more appropriate measure of well-being for consumption tax incidence studies is an individual’s expected lifetime income. When all people are exactly the same – earning the same income, making the same expenditure and paying the same amount of taxes over their life-course – the lifetime incidence of all taxes would be proportional, regardless of whether lifetime income or consumption are adopted as the measure of well-being. The challenge is how to measure lifetime income in practice. One approach is to use annual consumption as a proxy for lifetime income on the basis that it

230

TAX REFORM IN THE 21ST CENTURY

is less volatile than annual income. Based on this approach, various studies have found that the regressivity of consumption taxes is significantly lower than when assessed on annual income, and could even become progressive when consumption tax credits are considered (Poterba, 1989, 1991; Metcalf, 1994, 1997; Feenberg, Mitrusi, and Poterba, 1996). The attractiveness of annual consumption is that this data is readily available and the approach is simple to apply. The downside is that annual consumption may not be a good proxy for lifetime income because consumption is not stable over the lifetime (Caspersen and Metcalf, 1994). Consumption also follows a lifecycle pattern similar to but less accentuated than income. One solution is to use lifetime rather than annual consumption. Fullerton and Rogers (1991, 1993) estimated tax incidence based on age-income profiles and lifetime income measure. They concluded that while both corporate and individual income taxes appeared to be less progressive in a life-cycle framework than under a snapshot analysis, sales and excise taxes were less regressive – with the result that the overall incidence of the U.S. tax system was similar to that based on annual income.8 In summary, two basic approaches are used in the literature on consumption tax incidence to address concerns about life-course redistribution (and ability to pay taxes, Poterba 1993): (1) Measure annual tax burdens relative to lifetime income (as in Poterba, 1989 ,1991; and Metcalf, 1994, 1997); and (2) Measure lifetime tax burden relative to lifetime income (as in Fullerton and Rogers, 1996). Despite these approaches, most consumption tax incidence studies continue to focus on a snapshot approach measuring the annual tax burden relative to annual income derived from a cross-sectional household survey. As shown in Figure 4, the income measures adopted in practice vary depending on the particular taxes considered and the purpose of the study. Those concerned with the distributional impacts of consumption taxes invariably include these taxes in an income definition that goes beyond the conventional definition of disposable income. In an attempt to reflect lifetime considerations, many studies also attempt to proxy lifetime income by presenting their annual income snapshot by distinguishing individuals based on their age.

8

See also Fullerton and Rogers (1996), Lyon and Schwab (1991) and Metcalf (1997).

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

231

Figure 4. Different Income Concepts and Stages of Redistribution INCOME CONCEPTS AND STAGES OF REDISTRIBUTION

BENEFITS

CASH BENEFITS (age pension)

PRIVATE INCOME Before government intervention (income from employment, investments etc)

A

PLUS

GROSS INCOME

MINUS

DISPOSABLE INCOME

MINUS

POST-TAX INCOME

INDIRECT BENEFITS (education, health etc)

TAXES

B DIRECT TAXES

C

OR

INDIRECT TAXES

PLUS

DISPOSABLE INCOME plus social benefits

D

F

MINUS

PLUS

E

FINAL INCOME

eg http://www.canberra.edu.au/ centres/natsem/home (NATSEM Conference Paper 76) and Jones (2007) p3

EXAMPLES

INDIRECT BENEFITS (education, health etc)

INDIRECT TAXES

FINAL INCOME

eg ABS 2006, Cat 6537, p63

.

Source: Author’s elaboration.

4.3.3

Inter-unit Comparisons

In relation to the measurement of economies of scale in household consumption, most studies adjust income for economies of scale in consumption though ‘equivalence scales’ such that: D W= —

sE

(1)

where W is economic well-being, D is gross (or disposable) income, S is household size (number) and E is the equivalence elasticity (Atkinson, Rainwater and Smeeding, 1995).

232

TAX REFORM IN THE 21ST CENTURY

All studies on the economic well-being of individuals and households agree that some equivalence scale should be adopted but as to what value, there is no clear answer. A commonly used scale is the ‘modified OECD equivalence scale’ – where the first adult in each household has a weight of 1.0, the second and subsequent adults have a weight of 0.5, and dependent children a weight of 0.3; this is closely approximated by E = 0.6 in (1) – as used in ABS (2007b). Other commonly used scales are 1:0.7:0.5 (Harding, Lloyd and Warren, 2005; and Whiteford, 1985), which is proxied by E = 0.75; and E = 0.5, which is the scale used in all OECD reports on the subject. Using SE, each individual can be ranked according to his or her equivalent income or economic well-being. However, ascertaining a household’s well-being does overlook many factors other than income which may impact on a household’s needs, such as conditions of work (including hours worked), life cycle differences in earning and spending patterns, the value of home production, the imputed benefits from owner occupation; receipt of fringe benefits; the impact of unrealized capital gains or the benefit from retained earnings by corporations. 4.3.4

Equity Measures

Having identified the unit of analysis (the individual) and the basis for ranking these units (W in 1), attention turns to evaluating how equitably taxes are distributed between them. This cannot be determined without some notion of what is an equitable distribution. To this end, two basic measures are used: single number measures and other descriptive measures. Single number measures can be divided into those based on the Lorenz curve and those based on the generalized entropy measures, of which Lorenz-curve-based measures are the most commonly used.9 When income is equally distributed, it can be shown as the diagonal line SRX in Figure 5. When income is not equally distributed, the Lorenz curve could be represented as SZX and income inequality be measured by reference to A. The Gini index of income inequality is 2*A, and varies between a value of zero (for perfect equality) and unity (in the case of perfect inequality). The Lorenz curve and Gini index can be applied to study the re-distributional impact of taxes where our interest is either in horizontal equity (how ‘equal individuals in equal circumstances should be treated equally’) or vertical equity (how ‘different individuals in different circumstances should be treated differently’) implications of tax design. If we define A in Figure 5 as the difference between post-tax income (Y) and pre-tax income (X) then the effect of taxes on income distribution can be measured by:

9

For an overview of the Lorenz curve based and other measures, see Creedy (1999), Leigh (2005) and Kesselman and Cheung (2004) Table 2, p724 .

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES A

B

L = CA – C B

233

(2)

where CAA is the concentration (or Gini Index) of after-tax income X with ranking based on after-tax income X and C BB is that for before-tax income Y with ranking by before-tax income Y. This is the so-called Reynolds and Smolensky (1977b) measure of the redistributive impact of the tax system,10 with negative values of L indicating that income inequality is reduced by the tax (which is defined as an ‘income-inequality-improving’ tax) and positive values implying that the tax worsens income inequality. Figure 5: Lorenz Curves and Concentration Indexes X

100%

R

Cumulative Proportion of Income

Z

A

Line of perfect equality

Lorenz Curve or Concentration Curve of Income

B S 0%

Cumulative Proportion of Households

Y 100%

Source: Author’s elaboration.

The redistributive impact of a tax can arise from three factors: the level of the tax, the progressivity of the tax and the re-ranking effect of the tax. To highlight the role of these three factors, Suits (1977) and Kakwani (1977) developed conceptually related measures of tax progressivity. The Suits Index is the ratio of the area under the concentration curve for a tax to the area under a proportional line. Using Figure 5, the Suits Index is the ratio (B/(A + B), so that with a progressive tax the Suits index is positive (and ≤ 1) and with a regressive tax it is negative (and ≥ –1). If 10

This approach is similar to that proposed by Musgrave and Thin (1948) who proposed an index expressed as the ratio of the post-tax Gini and pre-tax Gini or A 1 – CA. B

1 – CB

234

TAX REFORM IN THE 21ST CENTURY

the concentration curve maps the cumulative distribution of tax on the vertical axis and before-tax income on the horizontal, then the Suits Index can be represented as: CB S = 1 – —T K

(3)

where K is the area under the line representing the proportional distribution of tax. Kakwani (1977) built on the approach developed by Suits (1977)11 to decompose L into its two constituent parts. His measure compared the distribution of the tax to the distribution of pre-tax income and defined tax progressivity as: K = CTB – CBB

(4)

where CTB is the concentration index of tax t with ranking by pre-tax income B and C BB is the concentration index of pre-tax income Y with ranking by pre-tax income. If K is positive, then the tax is progressive since a tax which is more unequally distributed than B will improve income inequality. A value of K less than zero has the opposite effect, worsening the distribution of pre-tax income. Of these measures, the Congressional Budget Office (1988) has suggested that the Suits measure is the most commonly used. In more recent times, the full range of these measures is generally applied. The limitations of Lorenz curve based measures are detailed by Creedy (2002). Despite the appeal of simple measures of inequality, by far the most common representation of tax incidence across households is through tabular presentation of tax incidence. In this representation, households are divided into equal groupings such as deciles, quintiles or quartiles. Results are then presented as either the ratio of tax to income (ATR) or in terms of the redistributive impact of taxes on some measure of income inequality. A criticism of this approach is that, while estimates of average incidence are interesting, they need to be complemented with estimates of marginal incidence, which involves examining changes in the burden as income increases (in effect a changing progressivity measure such as a change in liability progression (MTR/ATR) as income increases).

4.1

MUSGRAVE’S CONTRIBUTION TO EMPIRICAL STUDIES

Many factors need to be specified prior to undertaking a study into the distributional impact of consumption taxes on households. Richard Musgrave’s contribution to our understanding of these factors is undeniable. It was he who stressed the importance of differential incidence approaches to estimating tax incidence – highlighting the 11

The Suits index is calculated as the ratio of the area under the concentration curve for a tax to the area under a proportional line. Using Figure 9, the Suits Index is the ratio (B/(A+B) so that with a progressive tax the Suits index is positive (and ≤ 1) and with a regressive tax, negative (and ≥ –1).

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

235

problems with absolute estimates. With Tun Thin he contributed to our understanding of income inequality measures and with Krzyzaniak to tax incidence in a general equilibrium context. With Musgrave’s insightful contribution, we now better appreciate how differing approaches to each of the issues raised in this Section can lead to different tax incidence findings from the same data sources. Without such an appreciation, an informed analysis of the results from empirical consumption tax incidence studies could not be undertaken.

5 Findings from Country Studies of Consumption Tax Incidence This section reviews selected empirical studies with a view to learning more about the approach taken to estimating consumption tax incidence in practice and comparing their findings. This is difficult, as available estimates are based on many different approaches, with different modelling frameworks, and different data sources and coverage of taxes and households. As a result, this section provides only a broad review of selected studies, referring the reader to the individual studies for further insight. The primary focus is on the approaches undertaken in a limited number of English-speaking countries, how they differ, how their findings differ and what part could be attributed to differences in data and conceptual approaches. Also, for some studies, it is not possible to separate the effect of consumption taxes from that of other taxes and of public benefits. However, none of the studies reviewed here include consideration of deadweight losses, tax compliance and administration costs, tax gap or tax expenditures. In this review, three key aspects of each study will be identified: • The approach to modelling the shifting of consumption taxes to

households. • The measurement of household well being. • The method of evaluating the distributional impact of consumption taxes across households. As will be noted below, key drivers determining the approach taken in addressing these three aspects of tax incidence studies are the purpose of the study, the data available and whether government is directly involved in its preparation. In general, studies aimed at providing multi-country comparisons have a lesser level of precision than single-country studies. Also, government-based studies generally benefit from a higher level of resourcing, in terms of both access to data and financial support. Section 5.1 distinguishes between government and non-government studies for a single country. Section 5.2 focuses on comparative studies based on the Luxembourg Income Study (LIS)12 project, on recent enhancements to the EUROMOD13 model, and to other multi-country studies. 12 13

. and in particular .

236

TAX REFORM IN THE 21ST CENTURY

Table 1. An Overview of Studies on the Incidence of Consumption Taxes Country/ Multi Country Database s

What Consumption Taxes are Included?

Tax Modelling Characteristics Using Notation in Figure 3

How are Intermediate Taxes Modelled (B/C/D/E in Figure 3)

Incidence Approach

What Tax Shifting Assumptions?

SINGLE COUNTRY STUDIES ABS (2006, 2007a)

Australia

All

A+B+C1+D1+E1

I-O model

Absolute

Fully forward

Harding, Lloyd & Warren (2006)

Australia

All

A+B+C2+D2+E1

I-O model

Absolute

Fully forward

ONS (2007)

UK

All

A+B+C1+D1+E1

I-O model

Absolute

Fully forward

IFS Green Budget 2007, Myck (2000)

UK

Fuel Duty

A

Not modelled

Differential

Fully forward

Congressional Budget Office (2006)

US

Federal consumption taxes

A+B+C2+D2+E2 (Taxes on intermediate goods borne in proportion to their overall consumption.)

A is allocated direct to individuals and (B+C+E) on the basis of share in total consumption

Absolute

Fully forward

Chamberlain & Prante (2007) (Tax Foundation)

US

Federal Excise and Customs, State Sales Tax

A+B+C2+D2+E2

Not modelled; allocated in aggregate to household overall consumption

Absolute

Sales Tax and Customs Duties to consumers bu Diesel Excise split between consumers and corporate tax allocation approach

McIntyre, Denk, Francis, Gardner, Gomaa, Hsu, & Sims (2003) (ITEP)

US

State Sales and Excise Taxes

A+B+C2+D2+E2 (Exporting between states modelled but appears all taxes ultimately allocated to residents)

I-O model with taxes on intermediate inputs & capital Investment allocated)

Absolute

Fully forward

Vermaeten, Gillespie & Vermaeten (1995)

Canada

All

A+B+C2+D2+E1

Not modelled (allocated direct to household consumption)

Absolute

Commodity taxes are borne by consumers, except for the share of such taxes on government purchases which is borne by personal income taxpayers and the share on purchases of capital goods and exports, the common portion of which is borne by consumers and the differential portion of which is borne by labour.

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

What Measure of Welfare (as indicated in Figure 4)

A/B/C/D/E

A/B

Equivalence Scale (1Adult/2+ Adults/ Each Child)

Tax Rates or Nominal Tax Approach?

Modified OECD equivalence scale: 1/.5/.3

Tax rates

1/.5/.3

Tax rates

Distributional Impact Presentation

Quintile, Tabular

Main Results

VAT: Quintile 5/1 = 0.54 65% Taxes on Production allocated to individuals

Decile, Gini, Tabular

Quintile 5/1: VAT/Excise: = 0.38 93% Taxes on Production allocated to individuals

A/B/C/D/E

See Table 2

Tax rates

Quintile, Tabular, Gini

Quintile 5/1: VAT 0.58, Excise 0.37, VAT/Excise 0.49 71% Taxes on Production allocated to individuals

C

Modified OECD equivalence scale:

Tax rates

Figures

Differential results are not directly comparable to absolute incidence estimates. Quintile 5/1: Fuel Duty 0.61, Indirect tax reforms 1997-2001 0.39

B

Square root of household size

Nominal tax distributed according to consumption of the taxed good or service

Tabular

Quintile 5/1: Federal Excise: = 0.24 in 2004

“Each quintile contains equal numbers of people, and thus unequal numbers of households” p16

Nominal tax

Tabular, Suits Index

Ranking of families into quintile without apparent adjustment

Tax rates

B

B (but including only non-elderly families (singles and couples, with and without children))

100% Taxes on Production allocated to individuals

Quintile 5/1: Federal Excise: 0.87, Federal Customs 0.59, State Sales 0.76, State Excise 0.59, Total 0.74 100% Taxes on Production allocated to individuals

Tabular

Quintile 5/1: State General Sales – Individuals 0.45; Other State Sales & Excise – Individuals 0.23; State Sales & Excise on Business 0.32; All Sales & Excise Taxes 0.36 100% Taxes on Production allocated to individuals

Three different definitions of B are applied

Square root of household size

Nominal tax

Tabular, Figures

Quintile 5/1: Commodity taxes: 0.60 Taxes on Production allocated to non-residents according to exports and ownership of factor bearing taxes by non-residents

237

238

TAX REFORM IN THE 21ST CENTURY

Table 1 (continued) Country/ Multi Country Database s

What Consumption Taxes are Included?

Tax Modelling Characteristics

Deussing (2003)

Canada

Federal Excise and VAT

A+B+C1+D1+E1

I-O model

Absolute

Fully forward

Barrett & Wall (2005)

Ireland

VAT and Excise

A

Not modelled

Absolute

Fully forward

Garfinkel, Rainwater & Smeeding (2006)

LIS

VAT and Excise

A

Not modelled (allocated direct to HH consumption)

Absolute

Fully forward

Harding, Lloyd & Warren (2007)

See ONS(2006) and Harding, Lloyd and Warren (2006)

O'Donoghue, Baldini & Mantovani (2004)

EUROM OD countries

Not modelled

Absolute

Fully forward

Using Notation in Figure 3

How are Intermediate Taxes Modelled (B/C/D/E in Figure 3)

Incidence Approach

What Tax Shifting Assumptions?

MULTIPLE COUNTRY STUDIES

A

Source: Compilation by the author.

The key methodological differences and basic findings from these studies are summarized in Table 1. The structure of the table reflects the issues raised in Section 4 as critical issues that all studies of the distributional impact of consumption taxes must address. These include the scope of the study, the range of consumption taxes covered, the modelling of intermediate taxes (A, B, C, D and E in Figure 3), the incidence approach (absolute or differential) used, the tax shifting assumptions and measure of welfare adopted. Attention is also drawn to differences in equivalence scales and whether a tax rate is applied to a household’s consumption to determine their consumption tax liability or if the nominal tax burden for society as a whole is distributed across households based on their share of consumption of the taxed good. The distributional impact measures used to present the results are also noted, including the use of tables and figures (presented using deciles or quintiles) and single number measures such as a Gini-based measure or the Suits index. Findings from these various studies are contrasted using simple summary measures.

5.1

SINGLE-COUNTRY STUDIES

It is quite common for governments to resource agencies to undertake research into tax incidence, or to sponsor research into this issue by non-government agencies,

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

What Measure of Welfare (as indicated in Figure 4)

Equivalence Scale (1Adult/2+ Adults/ Each Child)

Tax Rates or Nominal Tax Approach?

Distributional Impact Presentation

Main Results

Family total income

No apparent adjustment

Tax rates

Tabular, Figures

Results are presented by family income ranges

B

1/.66/.33

Nominal tax allocated by related expenditure

Tabular, Figures

Quintile 5/1: VAT 0.63; Excise 0.56; VAT/Excise 0.61

C, E

Square root of household size

Tax rates

Tabular, Figures

See Table 4

Tax rates

Tabular, Figures

See Tables 5 and 6

C

239

facilitated by their collation of data necessary to undertake such analysis. In some cases, this goes as far as to support the development of microsimulation models for the analysis of how current and prospective tax and welfare policies might impact on households. In single-country studies, the important contribution of government studies will be clear. Amongst the English speaking OECD countries, the United Kingdom, Canada and Australia have all sponsored comprehensive studies or developed models which enable non-government researchers to study tax distributional issues (e.g. Canada). While there are some similarities in approach by these and other single-country studies, there are also differences which help to explain the different findings.

240

TAX REFORM IN THE 21ST CENTURY

Figure 6: Consumption Tax Incidence in Australia, Canada, Ireland and the United Kingdom a.

UK: Average Incomes, Taxes and Benefits by Quintile Groups of All Households, 2005/06 35%

Other 30%

Excise 25%

9.8%

VAT

20%

7 0% 7.0% 8.4%

15%

6.2%

5.6% 4.5%

6.0%

5.6%

8.9%

8.4%

8.0%

6.7%

2nd

3rd

4th

Top

10%

4.8%

3.1% 11.6%

5% 0%

Bottom

Source: ONS (2007), Appendix 1, Table 14A b.

Australia 2001–02 25%

24.9%

GST and Excise Duties

% of Equivalent Gross Income I

20%

14.3%

15%

13.7%

12.7%

11.8%

10.9% 10.2%

10%

8.7%

8.3% 6.4%

5%

0% 1

2

3

4 5 6 7 Deciles of Equivalised Gross Income

Source: ONS (2007), Appendix 1, Table 14A

8

9

10

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

c.

Australia 2003–04

3.5%

Tobacco products (Excise+GST) Alcoholic beverages (Excise+GST)

3.0% 2.5% 2.0%

1.9% 1.1%

1.5%

0.9%

0.9% 0.4%

1.0% 0.5%

1.3%

1.2%

1.2%

1.3%

1.1%

Lowest

Second

Third

Fourth

Highest

0.0%

10%

GST in total selected taxes on production

8% 6% 4%

8.8% 6.9%

6.4%

5.8%

2%

4.7%

0% Lowest

Second

Third

Source: ABS 2006, Cat No 6537.0 2003-04, Table 5

Fourth

Highest

241

242

d. 9 8 7 6 5 4 3 2 1 0

TAX REFORM IN THE 21ST CENTURY

Canada 2000: Distributional Impact of Federal Indirect Taxes and GST Tax Credits Indirect Taxes and GST Credits

2.9

1.8

1.0

0.4

0.2

0.2

GST Credit as a Percentage of Post-Tax, PostTransfer Income 2000 Income,

0.2 0.1

4.8

5.7

5.3

6.0

6.0

5.7

5.4

4.6 Indirect Taxes net of GST credits, as a Percentage of Post-Tax, PostTransfer Income, 2000

Source: ONS (2007), Appendix 1, Table 14A

Effective Taax Rates (%)

e. Canada 1988: Commodity Taxes 16 14 12 10

14.0 12.2

11.9

11.9

11.3

8

10.9

10.2

94 9.4

8.5 7.5

6

5.9

4 2 0

1-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-98 99-100 Source: Vermaeten, Gillespie and Vermaeten (1995)

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

f.

Ireland: VAT 2004: Excise and VAT Paid by Equivalised Income Decile 25

Excise as % of Income 2004 VAT as % of Income 2004

20

% Gross Income

243

6.3 6.7

15

6.3

6.0

6.2 5.8

5.6

10.6

10.8

52 5.2

4.6

10 14.5 5

2.8 12 0 12.0

11 8 11.8

11.2

11 8 11.8

10.6

9.9 6.8

0 1st

2nd

3rd

4th

5th

6th

7th

8th

9th

10th

Equivalised gross income deciles

Source: Barrett and Wall (2005), Table 3

5.1.1

United Kingdom

Government Studies The UK ONS has been a pioneer in the estimation of the effect of taxes and benefits on household income. As early as 1957 it began making results available publicly and has done so each year since. In additional to a detailed publication (ONS 2007),14 findings are also circulated more widely in the ONS publication Economic Trends.15 The ONS approach allocates those taxes to households which can be reasonably attributed to households.16 The ONS study covers, in terms of Figure 3, 14 15

16

The ONS publication ‘The effects of taxes and benefits on household income, 2005/06’ is downloadable from . See ‘The effects of taxes and benefits on household income, 2004/05’ Economic Trends article which can be downloaded from . However, from 2007, this publication goes online and appears to no longer include a summary of the full publication. See ONS (2007, p3) where it was stated that: ‘The analysis only allocates those taxes and benefits that can reasonably be attributed to households. Therefore, some government revenue and expenditure is not allocated, such as revenue from corporation tax and expenditure on defence and public order. There are three main reasons for non-allocation. Some taxes and benefits fall on people who do not live in private households. In other cases, there is no clear conceptual basis for allocation to particular households. Finally, there may be a lack of data to enable allocation. In this study, some £321 billion of taxes and compulsory social contributions have been allocated to households. This is equivalent to 60 per cent of general government expenditure, which totaled £536 billion in 2005 (Table 13). Similarly, £288 billion of cash benefits and benefits in kind have been allocated to households, making up 54 per cent of general government expenditure (Table 13)’.

244

TAX REFORM IN THE 21ST CENTURY

Table 2: The Effects of Taxes and Benefits on Household Income in the United Kingdom, 2005–06 Percentage Shares of Household Income and Gini Coefficients Percentage shares of equivalised income for ALL Original Gross Income Disposable Post-tax Income Income Income Quintile group2 Bottom 2nd 3rd 4th Top All households Decile group2 Bottom Top Gini coefficient (%) Notes

3 7 15 24 51 100

7 11 16 23 44 100

8 12 16 22 41 100

7 12 16 22 43 100

1 33

3 28

3 26

2 28

52

37

34

37

1 This is a measure of the dispersion of each definition of income (see Appendix 2, paragraph 53). 2 Households are ranked by equivalised disposable income. Source: ONS 2007 2005-06, Table 2 p6

Equivalence Value Type of household member Equivalence value Married head of household (that is, a married or cohabiting couple) 1.00 1st additional adult 0.42 2nd (or more) additional adult 0.36 (per adult) Single head of household (adult) 0.61 1st additional adult 0.46 2nd additional adult 0.42 3rd (or more) additional adult 0.36 (per adult) Child aged: 16–18 0.36 13–15 0.27 11–12 0.25 8–10 0.23 5–7 0.21 2–4 0.18 Under 2 0.09 Source: ONS 2007 p44 para 49 http://www.statistics.gov.uk/CCI/article.asp?ID=1804&Pos=1&ColRank=1&Rank=224

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

245

A+B+C1+D1+E1, i.e. around 70 per cent of taxes on inputs into the production process. Results on the distributional impact of taxes on goods and services – whether on final demand or on intermediate inputs into the production and distribution process – indicate that consumption taxes represent around 30 per cent of the disposable income of the bottom quintile, as compared to 14 per cent for the top decile (Table 2 and Figure 6, panel a). The regressivity of consumption taxes is apparent, with VAT less regressive than excise duties, partly reflecting the zero VAT rate applied to food. Table 2 details the equivalence scale17 applied in the ONS study and the impact of all taxes on various measures of income and income inequality – but does not report specific results for consumption taxes. The ONS report is static and based on an absolute incidence approach, implying that its model has limited scope to inform policy making. Non-Government Studies One of the most developed microsimulation models used to estimate differential tax incidence is that developed by the Institute for Fiscal Studies (IFS). In addition to the basic model TAXBEN, IFS have also developed extensions to examine VAT and excise incidence. Behavioural response of consumers to relative price changes due to consumption tax reforms18 were considered in earlier studies (Symons and Warren, 1996) but are no longer maintained. Results from modelling of VAT and excise duties in association with TAXBEN are regularly presented in the annual IFS Green Budget analysis. For example, the 2005 volume investigated the distributional effects of all tax and benefit reforms since 1997,19 while that of 2007 examined the distributional impact of reforms of environmental taxes (in the form of a 5 per cent increase in fuel excise duty, finding that its distributional impact is almost proportional across deciles 2 to 9, but higher than average on the bottom and the top deciles).20 Using the same modelling framework used in the Green Budgets, Myck (2000) finds that changes in consumption taxes introduced between 1997 and 2000 increased income inequality.21

17

18 19 20 21

The UK government is using the modified OECD scale to adjust incomes (see Appendix 3 of Households Below Average Income 1994/95–2004/05 or Appendix A of M. Brewer, A. Goodman, J. Shaw and L. Sibieta, Poverty and Inequality in Britain: 2006, IFS Commentary 101, 2006). See . See IFS Green Budget Chapter 7. See Source; IFS Green Budget 2007, Chapter 11, Figure 11.6, p. 206.

Myck (2000) concludes that ‘changes to consumption taxes since July 1997 have had a negative impact on post-tax incomes across the whole income range, with the most pronounced effects being among the poorest households. On average the effect of reduced VAT on domestic fuel has been outweighed by increases in excise duties on tobacco and road fuel for households in all 10 deciles’.

246

TAX REFORM IN THE 21ST CENTURY

The IFS now adopts the OECD modified equivalence scale, which is marginally different from that used by Myck (2000).22 However, since the IFS results relate to tax changes (or differential incidence estimates) they have limited comparability with those from ONS (absolute incidence). Also, the IFS results do not adequately model A and C in Figure 3. As with most other studies, the IFS methodology assumes full shifting forward of consumption taxes – not allowing for behavioural responses and ignoring the effect of such taxes on intermediate inputs. The household data used in the IFS study is, as with the ONS (2007) study, UK Expenditure and Food Survey (EFS).23 5.1.2 Australia Government Studies The Australian Bureau of Statistics began preparing official estimates of the impact of government benefits and taxes on households in 1987 for the year 1984 (ABS 2006) based on the same methodological approach as adopted by ONS in its UK studies. With each release of a new Household Expenditure Survey (HES), ABS has subsequently released new estimates which are available for the fiscal years 1988– 89, 1993–94, 1998–99 and 2003–04. In terms of Figure 3, the ABS approach models A+B+C1+D1+E1, i.e. including only those taxes that are levied either directly on final consumption of households, or on business investment and on intermediate inputs into these final consumption expenditures. As a result, some 60 per cent of total taxes on production (ABS 2006, Cat 6537, Table A4.1, p. 86) are allocated to domestic household final demand. If taxes on foreign households final demand (through exports) and on government (which can be viewed as churning) are regarded as not incident on domestic households (D1 and E1), then the proportion allocated is greater. Figure 6 (panel c) details the basic findings of this study in relation to consumption taxes with households ranked by equivalised household disposable income, equivalised with the ‘modified OECD’ equivalence scale (ABS 2006, p.79). The study presents the results by various social-economic and demographic groupings as well as the ratio of income in various deciles to that in other deciles (e.g. Percentile 90/Percentile 10 on ABS (2006, p.17)) but no single measure of inequality. Individual consumption taxes are not identified, but the GST component of total taxes on production for the fifth quintile was 0.54 of that for the first quintile. 22

23

In terms of equivalence scales adopted by the IFS and UK government Appendix 3 of the IFS report prepared for the Department of Work and Pensions Households Below Average Income (HBAI) 1994/95–2004/05 2006 report at and . See details of the ONS Family Expenditure and Food Survey at .

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

247

Non-Government Studies Non-government studies have been undertaken by Harding, Lloyd and Warren (2006) using the STINMOD-STATAX microsimulation model.24 STINMOD allocates personal income taxes and social welfare payments to households, while STATAX uses input-output data to model the forward shifting of all consumption taxes to domestic and foreign households (i.e. A+B+C in Figure 3). Lambert and Warren (1999) detailed this model and the estimation of the consumption tax rates applied to unit record Australian Household Expenditure Data. At its simplest, the first phase in STATAX is to identify the consumption tax component in the price of various goods. This is achieved by constructing an InputOutput Price Model using data prepared by the Australian Bureau of Statistics.25 In terms of Figure 3, all consumption taxes (A+B+C2+D2+E1) are assumed to be shifted forward to domestic or foreign household consumers. This model for allocating taxes to households has been applied in two variants. Warren (1979, 1983, and 1998) allocated aggregate taxes to households along with aggregate income using household income and expenditure shares observed in household surveys. In contrast, Harding et al. (2005, 2006, 2007) relied on the conversion of national accounts tax aggregates into ‘effective tax-inclusive tax rates’ which are then applied to the household unit record data used by STINMOD. These two approaches can yield different results. For example, if expenditure on alcohol is under-reported in household surveys, the effective tax rate approach applies the rate to the underreported expenditure and therefore under-allocates the tax on alcohol. In contrast, the approach based on tax aggregates will allocate all the tax on alcohol regardless of the total underreporting or how this differs between groups. Each approach will therefore result in quite different distributional impact estimates for taxes on alcohol. Just as underreporting of expenditure on items can lead to very different tax incidence patterns from the two approaches, underreporting of any particular income source would similarly distort results. Furthermore, in contrast to microsimulation models (based on household unit record data), the income assigned to households includes not only cash income but also imputed elements such as imputed rents, imputed interest and insurance income, retained profits, company taxes (on dividends), other taxes (on dividends), and supplements to wages and salaries (Warren, 1997, Table 3, p. 672). Since the Harding, Lloyd and Warren (2006) study is concerned with evaluating the distributional impact of consumption tax reforms introduced in July 2000 across households using the STINMOD-STATAX microsimulation model, STATAX was used to estimate the effective consumption tax rates on a range of commodities in 24 25

See and NATSEM Technical Paper – TP16. Australian National Accounts: Input Output 1993–94, ABS Cat No 5209.0., Australian Bureau of Statistics.

248

TAX REFORM IN THE 21ST CENTURY

HES as a result of the reforms introduced in July 2000.26 This package of reforms was not revenue neutral and reduced households’ overall tax liability. Figure 6 (panel b) shows the findings for GST and Excise duties across deciles of equivalised individuals ranked by gross income. The impact of these taxes on the lowest decile is nearly four times that for the highest decile: when adjusting the results to quintiles, the experience for the lowest quintile is around 2.5 times as high as that for the highest quintile. Using an equivalence scale the same as that in the ABS study of 1/0.5/0.3, Table 3 outlines findings for Gini-index-based inequality measures. The progressivity of the personal income tax is apparent as is the regressivity of consumption taxes. Not including consumption taxes when examining the impact of government on individuals will therefore result in a more progressive distributional impact than is in fact the case. Table 3: The Impact of the Tax System on Income Distribution in Australia

ATR G* G*–G %G P %P Ratio of Decile’s ATR 9/1 5/1 9/5

Personal Income Tax (PIT)

GST and Excise Duties

Other Taxes

All Taxes excl. PIT

All Taxes

19.5 0.3029 –0.0547 –22.5

9.7 0.3763 0.0187 8.6

15.9 0.3702 0.0126 5.4

25.6 0.3945 0.0369 14.0

45.1 0.3274 –0.0302 –8.5

0.2256

–0.1737

–0.0667

–0.1072

0.0368

265.4

–101.4

–63.9

–165.2

100.0

29.95 18.66 1.60

0.33 0.47 0.71

0.44 0.55 0.80

0.78 0.76 1.03

Notes G* G G 2001–02 G*–G %G P %P

Gini index of post (selected) tax Income Gini index of Gross Income (pre-tax) 0.3576 Gini Index of post-tax income less Gini Index of pre-tax income Contribution to % change in post-tax Gini index Progressivity index (Concentration index of taxes) Percentage contribution to tax progressivity

Source: Harding, Lloyd and Warren (2005), Table 13. 26

Other microsimulation models have been applied to the study of the distributional impact of consumption taxes in Australia. The Australian Federal Treasury developed PRISMOD

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

5.1.3

249

Canada

Government Studies Canada’s Social Policy Simulation Database and Model (SPSD/M) is a static microsimulation model that combines individual administrative data from Revenue Canada’s sample of T1 personal income tax returns and employment insurance claimant histories, with data from the Survey of Consumer Finance (SCF), the Survey of Household Spending (SHS) and the Survey of Income and Labour Dynamics (SLID) on family incomes and expenditures. This model is maintained by Statistics Canada27 and made available for the modelling of tax and welfare policies – being just one of a number of microsimulation models developed by Statistics Canada.28 The SPSD/M model is sophisticated in its modelling of consumption taxes,29 based on input-output data and the reallocation of consumption taxes on non-household final demand to households (i.e. A+B+C1+D1+E1 in Figure 3). However, the model omits consumption taxes levied at the provincial level. Surprisingly little use has been made of SPSD/M in the evaluation of the overall incidence of consumption taxes – its use being primarily for the study of personal income tax and social welfare programs and their reform. Deussing (2003) used SPSD/M to analyse the impact on families of Federal taxes and transfers, including Federal custom import duties, excise duties, excise taxes, other energy taxes, and the Federal GST, with results reported for family income groups.30 While this study omits provincial consumption taxes, it considers the GST credit designed to compensate low income groups for the adverse effects of the GST. The ability of this credit to offset the regressive impact of GST shown in Figure 6 (panel d) highlights the importance of including both the tax and expenditures in studies on the incidence of consumption taxes. For people in 2000 with incomes less than C$20,000, their

27

28 29

30

(Henry and Wright, 1992) for the analysis of the 1985 and 1998 consumption tax reforms. Scutella (1999) detailed an input-output consumption tax model which was applied by Creedy (2002) to the study of vertical and horizontal equity of the consumption tax reforms introduced in Australia in July 2000. See and . This model is then made available to various users such as University students for a fee to model tax and social policy eg See . This is modelled in COMTAX which is a module in SPSD/M. A detailed explanation of the operation of this model is available in the documentation downloadable from the Statistics Canada website at . It is also unclear if family income has been adjusted for family size (see Statistics Canada publication, Income in Canada 2005, Catalogue no. 75-202-XIE. . In other Statistics Canada publications, the equivalence scale used gives a weight of 1 to the oldest person in the family, 0.4 for the second-oldest person, 0.4 for all other family members aged 16 and over, and 0.3 for all other family members under age 16.

250

TAX REFORM IN THE 21ST CENTURY

Federal indirect tax burden pre-GST credit was some 64 per cent higher than for people with incomes over C$100,000; when including the GST credit, however, this difference is reduced to just 4 per cent. This observation highlights the importance of including all taxes, tax credits and welfare payments when examining the adverse distributional effects of a GST. This could be particularly relevant if the welfare payments and personal income tax system is designed to address the regressivity of these consumption taxes – hence the exclusion of the latter may yield a progressive distributional outcome never intended. Non-Government Studies Canada has a long history of tax incidence studies undertaken outside government.31 Two basic approaches characterized most studies. One is that adopted by Gillespie et al. which relies on allocating national account aggregates of income and taxes to families (or households) using data on the distributional patterns evident in household income, expenditure and finance surveys. The results from this approach, which is akin to that used by Warren (1979, 1997) for Australia, were applied to the situation in 1988 by Vermaeten, Gillespie and Vermaeten (1995) and are shown in Figure 6 (panel e). The second approach includes those based around the use of SPSD/M model which involves estimated tax rates being applied to household survey data with the potential for less than comprehensive allocation to household of all taxes collected. For Canada, as observed by Kesselman and Cheung (2004, pp779–780), the problem is that ‘almost all studies… are now quite dated, relying on data sets from the 1970s and 1980s. Few capture the major PIT reforms in Canada of 1988, the adoption of the GST in 1991, or the increasing use of payroll taxes since 1990’. 5.1.5

Ireland

Barrett and Wall (2005)32 present results from a study undertaken for the Combat Poverty Agency to better understand the distributional impact of consumption taxes on those with low incomes. With a focus only on taxation, this study used gross (rather than net) income and the household as the unit of study (with an equivalence scale where the first adult is weighted as 1, all other adults are given a weight of 0.66 and each child is weighted as 0.33). Data on spending patterns was taken from the Ireland Central Statistics Office Household Budget Survey 1999/2000 and information on consumption taxes from the Irish Revenue Commissioners and the Department of Finance. The method for determining the incidence of VAT and 31

32

At the vanguard of this Canadian research has been Irwin Gillespie who undertook such a study for the Royal Commission on Taxation in 1966 ‘The Incidence of Taxes and Public Expenditures in the Canadian Economy’. See Kesselman and Cheung (2004) for a comprehensive review of tax incidence studies in Canada. See .

NEIL WARREN: THE DISTRIBUTIONAL EFFECT OF CONSUMPTION TAXES

251

excise taxes on household expenditure relied on information on the VAT and excise rates applied to each of the different goods and services identified in the survey data, and this was applied to household expenditure on a tax-inclusive basis.33 In terms of Figure 3 this involves only allocating C and omitting taxes on intermediate inputs and those directly on non-household final demand. The modelling of GST reported in Figure 6 (panel f) is therefore incomplete in its coverage of the tax – although the majority of this tax is included. The conclusion of this study is that consumption taxes represent 21 per cent of the gross income for those in the bottom decile but only around 10 per cent for those at the top: adjusting the data to quintiles, the ratio for the lowest quintiles is only 60 per cent higher than the top quintiles. 5.1.6

United States

The US literature on consumption tax incidence can be distinguished according to whether these studies are state or federally focused, and whether their approach is based on a microsimulation model or some national accounts aggregate allocation framework. Government Studies: Federal While the Congressional Budget Office (CBO, 1998) has developed microsimulation models to analyse the effect of income tax and welfare reforms on households (much like STINMOD and TAXBEN), its application to consumption tax incidence is more limited. This is because the US Federal government only applies consumption taxes in the form of excise and customs duties. CBO34 in turn only models the incidence of excise duties – not customs duties – assuming that they are passed through to final consumers, while those that affect intermediate goods and are paid by businesses are attributed to households in proportion to their overall consumption. CBO assumes that each household spends the same amount on taxed goods as similar households with comparable income in the Consumer Expenditure Survey (CBO, 2006).35 The income measure used to rank households is their comprehensive pre-tax household income adjusted for household size based on the square root of the household’s size. Households are ranked by their adjusted income and grouped into quintiles which contain equal numbers of persons (CBO, 2006, p. 4).CBO estimates of the distributional impact of Federal Excise Duties indicate that households in the lowest quintile pay for Federal Excise Taxes around 2 per cent of their comprehensive income, four times as much as those in the top quintile; since 1980 this incidence 33 34

35

If t is the VAT rate (or tax-exclusive rate) then the tax-inclusive rate to apply to nominal expenditure reported in a household survey is t/(1+t). Effective tax rates equal the amount of tax liability divided by income. See Congressional Budget Office, Effective Federal Tax Rates, 1979–1997 (October 2001) and Effective Federal Tax Rates, 1997 to 2000 (August 2003), as well as Web-only updates that extend the period of analysis through 2003. Source CBO (2006) p. 3.

252

TAX REFORM IN THE 21ST CENTURY

has risen for those at the top of the income distribution and declined for those at the bottom (Figure 7). Other US Federal Government agencies that have prepared estimates of the distributional impact of consumption taxes include the Office of Tax Analysis in the US Treasury,36 the Joint Committee on Taxation37 and the Joint Economic Committee,38 but it appears that the CBO is the primary government agency currently preparing consumption tax incidence estimates. Government Studies: State Most US States maintain their own tax impact models. As in Canada, it is also common to find state consumption taxes accompanied by some assistance to low income households such as sales tax credits. What is noteworthy about US State taxation is that a large proportion of the revenue collected by state governments is from taxes on consumption. As a consequence, any major review of the distributional impact of State taxes on households must inevitably model the incidence of consumption taxes. The methodologies adopted in some of these studies have begun to model more rigorously the distributional impact of consumption taxes, i.e. as fully set out in Figure 3.39 However, these State-based studies only focus on incidence of their own taxes, while taking only limited account of the important inter-jurisdictional tax shifting arguments and Federal taxes.40

36 37 38 39

40

See . pBA. Begin with investor A. In the absence of taxes, he will invest in asset A, which for a small investment suppose yields a return higher than asset B. He will continue to invest in asset A until the marginal return from asset A falls to the level of asset B,6 that is until the pre-tax rates of return are the same: pAA = pAB. The position of investor B depends exactly on how we define ownership. Desai and Hines seem to have in mind discrete ownership: either asset A is owned by investor A or by investor B, but not both. In this case, in equilibrium investor B would generally simply hold asset B.7 In the presence of taxes, the conditions are similar, except that the investors’ investment allocations will depend on post-tax rates of return instead of the pre-tax rates of return. The introduction of a completely general tax system could generate an equilibrium in which investor B owns asset A instead of investor A. That is, if A were taxed at a very high rate on the return from asset A, while B were taxed at a very low rate on the return from asset A, then B may value the asset more highly than A. This case would violate the Desai/Hines notion of CON. To ensure CON holds we therefore require conditions on the effective tax rates such that investor A continues to own asset A. So far, however, we have effectively treated the investment by either investor as a new investment. What if B owns the asset, but A wishes to purchase it since pAA > pBA? As noted by Becker and Fuest (2007), the price at which B is prepared to sell the asset should reflect the tax rate he faces on the return. If tAA > tBA, then in the absence of the additional pre-tax return generated by A, A would value the asset at a lower value than B. A would only purchase the asset if the additional

6

7

It is also possible that given his wealth, and lack of any further finance, the rate of return on asset A remains higher than the rate of return on asset B; in this case, investor A would invest only in asset A. However, we do not analyse this possibility. However, if B can hold a minority share in asset A (which is controlled managed and controlled by A), then B could also earn pAA. However, we will not examine this case either.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

507

return generated by A owning the asset at least compensates for this difference in valuation. In general – as Desai and Hines argue – CON will hold if A and B face the same overall effective tax rate as each other on the returns to any asset: that is, tAA = tBA and tAB = tBB. (In this framework, this is equivalent to the condition for capital import neutrality.) This would occur if the returns to the asset were taxed only in the location of the asset, and not in the residence location of the investors (which may be different). This in turn is consistent with the investors’ residence countries exempting foreign source income from tax, the case emphasised by Desai and Hines. However, now let us add, in turn, two further elements. First, we have so far discussed only the position of unspecified investors; we have not yet distinguished between investment undertaken by individuals and investments undertaken by corporations. Second, we have not yet considered in detail the role of competition between companies, including international trade.

2.3

DISTINGUISHING PORTFOLIO AND DIRECT INVESTMENT

Devereux (2000) analyses simultaneously the optimality properties of personal taxes on capital income, and taxes on corporate profit. A useful framework for this distinction is to suppose that individuals (possibly indirectly through investment vehicles) undertake portfolio investment in corporations. These corporations need not be resident in the same country as the individual investor: despite a home bias in portfolio investment, there is considerable cross-border portfolio investment. Corporations, in turn, raise finance from individual investors around the world, and undertake direct investment; again, the direct investment can be domestic, or crossborder. Given perfect capital mobility for portfolio investment, and abstracting from risk,8 then the post-corporation tax rate of return required by companies must be equalised across all companies, no matter where they are located. (Note that in principle the return may be subject to a number of taxes: the general principles discussed here relate to all taxes levied though we shall continue to refer for simplicity just to corporation tax). In equilibrium, taxes levied on individual shareholders may affect the post-corporation tax, pre-personal tax, rate of return, required of companies, but that same rate is required of all companies. We treat personal taxes as being levied on a residence basis: any withholding tax levied on payments to all shareholders is for this purpose treated as a tax on the corporation. A change to any individual’s personal tax rate may affect her allocation of investments, and may affect her post8

There is an extensive literature which analyses portfolio investment decisions in the presence of risks and differential taxes on capital income, see, for example, Brennan (1970), Gordon and Bradford (1980) and Bond, Devereux and Klemm (2007). However, we abstract from these issues here.

508

TAX REFORM IN THE 21ST CENTURY

tax rate of return. However, as long as each investor is small relative to the market (or each group of investors is small) such a change in tax rates will not affect the required post-corporation tax rates of return. Devereux (2000) analyses the required relationship between corporate and personal taxes in such a setting: here we abstract from personal taxes altogether in order to focus solely on corporate taxes. How does such a setting affect the requirements for achieving production efficiency? The key new ingredient created by considering portfolio and direct investment is that the post-corporation tax rate of return required from direct investment by corporations is now fixed. In terms of the framework used above, it is straightforward to reinterpret the investors described above as being corporations. In this case the effective tax rates are effective corporate tax rates. The key difference, however, is that the post-corporation tax rate of return is fixed, and is common across all companies. Like Desai and Hines (2003), Devereux (2000) also considers the case in which the productivity of an asset depends on the company which owns it. However, Devereux (2000) considers a more general problem than simply the ownership of a single asset. Suppose, for example, that each company operates in each country through wholly-owned subsidiaries. Then the pre-tax rates of return on the assets in a single country need not be equal to each other. Consider this case first in the absence of tax. If one firm were more productive than another, it would generate a higher rate of return. This would attract more capital from individual investors. As the company expanded, its marginal rate of return would decline (and that of the other company would increase if it had less capital). In equilibrium, the two companies would earn the same marginal rate of return. But in the presence of corporation tax, post-tax rates of return must be equalised. Differences in effective corporate tax rates must therefore affect only pre-tax rates of return. How do the neutrality concepts work in this context? There is no impact on the implications for how CEN can be achieved, except that the requirements for neutrality now specifically apply to effective corporation tax rates: as long as each company faces the same effective corporate tax rate on all its investments then location decisions will not be distorted. However, CEN on its own is no longer sufficient to achieve production efficiency. Suppose CEN holds, but that the effective tax rate faced by company A exceeds that faced by company B. Then the pre-tax rates of return earned by company A must exceed those earned by company B: production efficiency would not hold. It is also clear that if only one country introduced CEN, this would not generate production efficiency. For example, if only A introduced CEN, the pre-tax rates of return would differ between the two companies, and they would also differ across the different investments undertaken by B. This calls into question the notion of some individual countries (see, for example, HM Treasury and HMRC, 2007) that

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

509

they aim to achieve CEN for their own investors. On its own, this is not optimal either at a global level or for an individual country (see below).9 What of CIN? This is more complex, since to some extent this is a neutrality property in search of a rationale. When introducing this concept, we noted that – in a simple framework – CIN would imply that all investors faced the same posttax rate of return (though pre-tax rates of return would differ across countries). However, this is not an issue for the optimality of corporation taxes if all companies are required to earn the same post-corporation tax rates of return. The ultimate posttax rate of return earned by investors will be affected by their personal tax rates on capital income, but not on corporation taxes. However, the ‘level playing field’ mentioned by McLure (1992) suggests another interpretation of CIN, emphasised by Devereux (1990, 2000). This is that companies from A and B compete with each other in each country. Suppose, for example, that each company produces a similar good in each country, and competes with the other company in the market in that country. The pre-tax rate of return required by each company will affect the price at which it can afford to sell its good. In general, the higher is the relevant tax rate faced by the company on its investment, the higher will be the price at which it must sell its own good. But if CIN holds – tAA = tBA and tAB = tBB – then the companies face the same effective tax rates in each country, so that neither company will gain a competitive advantage. Thus, if producers sell their product only in the country in which the good or service is produced, then levying the same effective tax rate on all producers in that country is a sufficient condition for taxes not to distort competition. This is equivalent to an exemption system in all residence countries. Although this is a form of CIN, it might also be considered a form of CON. In this setting, CON continues to require an exemption for foreign source income (or at least a tax system which is equivalent). That is, to avoid distortions to the corporate ownership of assets in any jurisdiction, all companies which may purchase assets there must face the same effective rate of tax. However, the concept of fair competition outlined here is more general than CON, since the change of ownership considered in CON could be achieved not by acquisition, but by firm A simply entering the market and undercutting B. For example, it is possible that there are reasons why A would not purchase the assets of B, even if it was more productive (for example, due to high transactions costs, or B being unwilling to sell). In this case, A and B would compete in the same market. Ultimately, it is possible that B could go out of business, to be replaced by A (which would take us back to the analysis in the previous section, instead of assuming that both countries operate in both markets). 9

Note though that although some countries may aim broadly to achieve CEN, no country has actually implemented a system which would achieve this. It would involve, for example, rebating foreign tax liabilities which were in excess of domestic liabilities.

510

TAX REFORM IN THE 21ST CENTURY

A broader interpretation of CON is therefore that one company has no competitive advantage over another as a result of differential effective tax rates depending on ownership.

2.4

INTERNATIONAL TRADE

But the sense in which the term CON was used by Devereux (1990) refers to a still broader concept, which also requires consideration of another dimension: international trade. In practice, companies do not always sell their output in the location of production: they can export the output to other countries. This has a profound impact on the requirement for production efficiency. Suppose companies A and B compete with each other in a third country. To avoid any distortion to competition in the third market, it would be necessary for both companies to face the same effective tax rate. Yet, typically corporation taxes are not levied on a destination basis – where the output is sold. So there would not necessarily be any tax levied in the third country. Instead, tax would be levied either in the “source” country, where the investment takes place, or the ‘residence’ country, where the company’s management and control takes place. Yet A and B may compete with each other even if their source and residence locations are different from each other. It is perhaps useful to define a new term – market neutrality – which holds if taxes do not distort competition between companies; that is, one company does not derive a tax-induced competitive advantage over another.10 It is clear from this analysis that market neutrality would require full harmonisation of source- and residence-corporation taxes. That is, the effective tax rates faced by all potential competitors in any third country would need to be the same. This is clearly a much more demanding requirement than achieving CON in the Desai and Hines sense, which can be achieved by an exemption system, generating source country taxation.11

2.5

SUMMARY OF PRINCIPLES FOR GLOBAL OPTIMISATION

In a world with direct and portfolio investment, in which all companies worldwide are required to earn a given post-tax rate of return, the principles of CEN, CIN and CON need some revision. With regard to the optimal setting of corporation tax, the principle that different individual savers should face the same post-tax rate of return is no longer relevant, since corporation taxes in any one country typically have no impact on this rate of return. 10 11

This is the original sense of CON, as used by Devereux (1990), but ‘market neutrality’ is probably a more accurate name. It is for these reasons that the empirical papers of Devereux and Pearson (1995) and Devereux and Loretz (2008) consider the proximity of European tax systems to full harmonisation.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

511

That leaves the principle of production efficiency. It is true that production efficiency is Pareto optimal only under certain conditions. However, there is no comparable and clear principle in cases where these conditions do not hold. As Slemrod (1995) has argued, the situation here is comparable to the principle of free trade. The conditions under which free trade are optimal are well known; while it is possible that these conditions do not hold, most economists still consider free trade to be a useful guiding principle. The same could be said for production efficiency. Other ‘principles’ are therefore subsidiary to production efficiency. Based on the discussion above, production efficiency implies two ‘principles’: 1. Direct ‘CEN’: that taxes should not distort the location of corporate activity 2. Market neutrality: that taxes should not distort competition (even in a very broad sense) between any companies operating in the same market.12 The second of these principles is in some ways a more general concept of CON than the definition of Desai and Hines (2003). If it were the case that any less productive firm was immediately acquired by a more productive firm, then all (remaining) firms would be equally productive. In this case there would be no harm to competition, since all firms would be equally productive. But if such acquisitions do not always occur – as seems likely, given costs of acquisition – then we are left with firms of different levels of productivity co-existing. This introduces the more general principle above. In a world where the post-corporation tax rate of return required for all companies is the same, production efficiency cannot be achieved by residence- or source-based taxes unless they are fully harmonised. In the absence of sufficient international agreement to achieve that outcome, a question arises as to whether is it possible to identify which of these two forms of taxation generate the greater welfare costs? Source-based taxation distorts location choice and competition generated by international trade; residence-based taxation distorts competition generated by cross-border investment and international trade. However, given that all these factors are closely related – for example, decisions of location involve the choice between cross-border investment and trade – any argument in favour of one form of taxation on this basis would be precarious.

3 Economic Principles for National Optimisation So far the discussion has focused on the conditions needed to achieve global production efficiency. But this almost inevitably requires international cooperation. Perhaps a more realistic goal for a national government is to maximise the welfare 12

Devereux (2000) and Devereux and Pearson (1995) summarized the principles as Direct CEN and Direct CIN. If both of these hold, then so do the two principles set out here.

512

TAX REFORM IN THE 21ST CENTURY

of its own citizens. In this section we address the issue of the optimal tax treatment of the returns to outbound investment from the perspective of the domestic economy only. Many of the issues raised in the previous section on global optimisation also apply here. This section begins by summarising the traditional approach, and then asks how the policy prescriptions are affected by factors ignored in the traditional approach. The key new factor is the distinction between portfolio and direct investment, discussed above, which fixes the post-corporation tax rate of return required from a company located in a small open economy.

3.1 THE TRADITIONAL APPROACH Maximising national income, rather than global income, may generate a very different ‘optimal’ tax structure for the taxation of cross-border income. As with the terms CEN and CIN, Musgrave (Richman, 1963, 1969) introduced the term ‘national neutrality’ to describe optimal policy for an individual government. Supported by more the formal analysis of Feldstein and Hartman (1979), the key policy prescription is that a government should tax the worldwide income of its residents, treating foreign taxes as a cost (and hence allowing them to be deducted, but not allowing a credit). The rationale behind this is as follows. Suppose there is a fixed supply of saving, to be allocated between domestic and outbound investment. For the country as a whole, the optimal allocation of investment would equate the ‘social’ rates of return from the two. In a simple framework,13 the social rate of return to the home country is the return net of foreign taxes, but before domestic taxes (since domestic taxes are used to benefit domestic residents). Hence the post-foreign tax rate of return on outbound investment should be set equal to the pre-tax rate of return on domestic investment. Private investors, however, will allocate investment to equalise posttax rates of return. These two allocations are only the same under a worldwide tax system where foreign taxes are deductible from the home country tax base. Slemrod (1995) also justifies this approach with an analogy to free trade. In a basic economic model of a small open economy, it is inefficient to levy import tariffs or a source-based tax on capital income. But from the perspective of an exporting country, the existence of an import levy in the importing country does not justify any rebate to exports. Similarly, if the capital-importing country does have a sourcebased tax on capital income, that is no reason for the residence country to adjust its tax system: the residence country should levy tax on the returns net of foreign tax (which is an expense), and not give any credit.

13

Strictly, where the marginal cost of public funds is unity.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

3.2

513

DISTINGUISHING PORTFOLIO AND DIRECT INVESTMENT

A central assumption of the basic argument – explicitly stated in the Feldstein and Hartman (1979) model – is that there is a direct link between domestic and outbound investment: $1 more invested abroad reduces investment at home by $1. Yet this is not generally the case in practice. Indeed, in the framework of portfolio and direct investment set out above, domestic companies can raise unlimited finance on the world market at the world post-corporation tax rate of return. In this case, there may be no link at all between domestic and outbound investment, and hence no link between the optimal tax rates. This position is studied by Devereux (2004), who considers a small open economy that has both inward and outward portfolio and direct investment, and analyses the optimal relationship between the taxes on these different forms of investment. In this model, outbound investment does not crowd out domestic investment, since at the margin both forms of investment are financed by inward portfolio investment. In this case, the existence of a tax on domestic investment has no implications for the optimal effective tax rate for outbound investment. If the country as a whole has no market power (that is, all its firms together are small relative to the world market), then the optimal tax rate on outbound direct investment is zero. A similar result was implicitly found by Mintz and Tulkens (1996). If the country does have some market power, then it could be optimal to tax the outbound investment, to drive up the world price of the product. This is similar to the market power case for a tax on exports. However, it is not feasible to do this in practice with a general tax system applying all outbound investment. The argument that the appropriate tax on outbound investment is zero is in fact similar to the argument that the appropriate tax on inbound investment is zero (see, for example, Gordon, 1986). This is because if the outbound direct investment is financed at the margin by inbound portfolio investment, then it really can be considered a form of inbound investment. The decision of a multinational company as to where to locate its parent company and headquarters is similar to any other location decision. A high tax rate associated with parent company location may induce the company to choose a different, lower taxed, location. In such a setting, there is no rationale for the government hosting the parent company to tax its worldwide income. In an extreme case in which the capital simply flows through the residence country of the parent, then there is unlikely to be any gain to that country from being an intermediary location. In that case, any tax levied by the residence country is likely to deter the company from routing capital in that way. However, there are likely to be some benefits of hosting the company headquarters, as long as some economic activity takes place there. Moreover, although at the margin finance may be provided on the world market, given home bias in portfolio investment allocations, it is likely that at least some shareholders are resident in the same country. If there

514

TAX REFORM IN THE 21ST CENTURY

are some benefits such as these, then imposing a tax on worldwide income will in general raise the required pre-tax rate of return, possibly giving the company a competitive disadvantage in foreign markets, and hence reducing the attractiveness of the residence country as a headquarters location. In this setting, it is possible to make the case against taxing the returns from outbound investment more forcefully by questioning the ‘nationality’ of the company. Public comment still tends to refer to companies as a ‘UK multinational’ or a ‘Canadian multinational’. But those terms are inherently contradictory. If a company is multinational, then how can it also be Canadian? Perhaps such a term might be justified if the company’s shareholders were all Canadian residents, but with international portfolio investment, that is unlikely to be the case. What proportion of the shareholders would need to be Canadian residents for the term to be reasonable? Would the term be justified if the headquarters of the company is in Canada? That seems implausible as well, since the headquarters may reflect only a small part of the company’s activities. Indeed, Desai (2008) argues that even basic headquarter functions of multinational companies are typically now split, and located in different countries. In a typical multinational company, then, some activity of the company may be taking place in the country of the headquarters, but those activities are best seen as just one of the many and varied activities that the company undertakes worldwide. It is not clear why that justifies that country taxing the worldwide income of the company. Certainly it is hard to make a case that the location of the management, or the place at which the company primarily raises finance, or the place in which it is listed, or some other aspect of the headquarters, is the crucial element of the company which justifies the government of that country taxing the worldwide income of the company.

3.3

INTRODUCING OTHER LINKS BETWEEN DOMESTIC AND OUTBOUND INVESTMENT

In a more complex model, the links between domestic and outbound investment may be more subtle. Devereux and Hubbard (2003) consider the case in which a domestic company may export to the foreign market or produce abroad: in either case it competes with a non-resident company making the same decision. In principle, the domestic government could modify its domestic tax system to manipulate the decisions of the company in a way which would give it a competitive advantage, and hence boost domestic welfare. However, the ‘optimal’ tax system in this setting depends on the economic parameters facing the company: exemption, credit or deduction systems could all be optimal in different situations. Again, it is not feasible to adjust the tax system for outbound investment according to the specific characteristics of the tax payer. Another possible link between domestic and outbound investment, considered by Becker and Fuest (2007), is that there may be

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

515

a limit to availability of management in the company: undertaking an investment in one location precludes undertaking another somewhere else. Taken to an extreme, this would reintroduce the one-for-one relationship between domestic and outbound investment assumed by Feldstein and Hartman (1979). There is now a growing empirical literature examining links between outbound and domestic investment (see, for example Simpson, 2008), which would imply an even more complex calculation of the optimal tax system. Again though, it is unlikely that the optimal system would be the same for all companies. If governments cannot design tax systems which vary according to firm or sector characteristics, but are forced to implement general tax structures, then the reasoning above suggests that the underlying presumption should be in favour of a system which does not tax foreign income of domestic corporations.

3.4

SUMMARY OF PRINCIPLES FOR NATIONAL OPTIMISATION

As with the principles for global optimisation, in a world in which all companies worldwide are required to earn a given post-tax rate of return, the presumption in favour of taxing worldwide income while allowing only a deduction for foreign taxes requires revision. In fact, the presumption should swing in entirely the opposite direction towards exempting income earned abroad from domestic tax. In general, any tax levied would raise the required pre-tax rate of return on the company which would lessen its ability to compete with companies headquartered in other countries. This conclusion is only reinforced by questioning the justification of a government to tax the worldwide income of a multinational whose headquarter is located in that country. There is no obvious special feature of the location of the headquarters that justifies the imposition of a tax on worldwide income, as opposed to income arising as a result of domestic activity. But the nature of the economic argument is important for policy. The argument is that the domestic country would actually be better off if it did not impose a tax on foreign income earned by corporations. This argument does not rely on the policy of the host country. But some countries apply an exemption system only in specific cases, for example to income arising in other countries with which they have a treaty. The argument here implies that they would be better off exempting all foreign source income, irrespective of the policy of the host government or the existence of treaties.

4

Difficulties in Identifying the Location of Profit

Unfortunately, identifying that there is no clear rationale for taxing the returns to outbound investment, and hence exempting income earned by foreign affiliates from domestic tax, is relatively straightforward compared to the problems of

516

TAX REFORM IN THE 21ST CENTURY

implementing a pure source-based tax system. A number of difficult problems arise as income and costs need in principle be allocated between jurisdictions. Yet not only is that difficult to achieve in practice, in many cases there is simply no conceptual basis to support any particular approach. The treatment of different forms of income creates another problem: for example, interest payments are typically taxed in the country where they are received, rather than the country from which they are paid; yet this is in contradiction to the general principle that tax should be levied only where the income-generating activity took place. It is not clear why the form of financing of a foreign affiliate by its parent should turn the international tax system from one based on source (for equityfinanced investment) to one based on residence (for debt-financed investment). In this section we address these two issues in more depth; we finish with a brief discussion of the appropriate treatment of capital gains arising from the sale of an affiliate.

4.1

DOES SOURCE-BASED TAXATION HAVE A SOUND CONCEPTUAL FOUNDATION?

The principle of exempting foreign source income from domestic taxation requires it to be possible to identify where corporate income is earned. A source-based international tax system would require a multinational company to allocate its profit between the taxing jurisdictions in which it operates. However, attempting to define where profit is generated is often very difficult, and in some cases impossible. Both income and costs arising in each jurisdiction in principle need to be identified. In a simple case, we can consider for example a British resident company that wholly owns a subsidiary which is registered, and which carries out all its activities – employment, production, sales – in, say, France. Then France would typically be considered to be the source of the corporate profit. Conventionally, we can also drop sales from the list of activities: if the subsidiary exported all its product to Germany, France would still conventionally be regarded as the source of the profit (although in economic terms it is less clear why this would be the case). Things are less clear, however, if the British holding company owns several subsidiaries in different countries, which undertake different aspects of the multinational’s activities: for example, finance, marketing, R&D, production, sales. The existing system of separate accounting requires all transactions between these different parts of the group to be valued, in order to divide total profit between the countries involved. The contribution made by each would be determined using ‘arm’s length pricing’ – in principle, the price that would be charged by each subsidiary for its services as if it were dealing with an unrelated party. Of course, such a procedure is difficult in practice since in many cases no such arm’s length price can be observed; transactions between subsidiaries of the same corporation

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

517

may not be replicated between third parties. But in many cases, not only is this difficult to administer, it has no conceptual foundation. Further, it is in principle necessary to identify costs of financing with specific activities. For example, a pure source-based tax system would give relief only for interest payments which relate to debt finance used to undertake domestic activities: that is, there is in principle a need to allocate interest expenses between domestic and foreign uses. This is not just difficult in practice, however; it is not clear even whether there is a ‘correct’ allocation conceptually. The multinational company may borrow in order to support its global activities; and it may have an overall desired gearing ratio. This would imply that borrowing for one activity would preclude borrowing for another activity. Which activity is actually financed by borrowing could be arbitrary. In this case, effectively there is simply an overall level of acceptable borrowing, and hence no conceptual basis on which to allocate interest payments between domestic and foreign uses. Conceptual problems in splitting profit between jurisdictions arise in arm’s length pricing as well. For example, suppose that the multinational has two R&D laboratories in different countries. Each R&D laboratory has invented, and patented, a crucial element of the production technology. Each patent is worthless without the other. One measure of the arms length price of each patent is therefore clearly zero – a third party would not be prepared to pay anything for a single patent. Another possible measure would be to identify the arm’s length price of one patent if the purchaser already owned the other patent. But if both patents were valued in this way, then their total value could easily be larger than the value of the final output. More generally, it is far from clear that the arm’s length price used by independent parties would be the same as the price which would be transacted by affiliates within the same group: different organisational forms can be expected to arise because of differences between companies, which may be reflected in cost structures. The distortions created by imposing transfer pricing rules in this setting are explored in several papers.14 In a recent contribution, Devereux and Keuschnigg (2008) analyse how companies choose their organisational form: whether to expand abroad through foreign direct investment, and thereby own the foreign activity directly, or through outsourcing by licensing a third party to undertake the same activities. In this model, the choice of organisational form depends on the success probabilities of the firm. Crucially, however, the arm’s length price chosen in the outsourcing case depends on the characteristics of that case. For example in the absence of tax considerations, if there is a financial constraint, the parent company would not require the foreign subsidiary to make an upfront royalty payment; such a payment would reduce the scale of activities abroad, which would reduce aggregate profit. However, a positive royalty payment would be optimally chosen in the outsourcing case. Treating the royalty payment in the outsourcing case as the appropriate royalty in the FDI case 14

See, for example, Halperin and Srinidhi (1987, 1991) and Harris and Sansing (1998).

518

TAX REFORM IN THE 21ST CENTURY

would induce a distortion to the choice of organisational form and in the level of investment undertaken.15 Identifying where profit is generated is a fundamental problem of conventional corporation taxes in an international setting. In some ways it is a problem with which the world has learned to live, even though allocating profit among source countries is in practice a source of great complexity and uncertainty. But this problem is not just one of complexity and uncertainty: it can – and perhaps should – also affect the fundamental design of the tax system.

4.2

DISTINGUISHING DIFFERENT FORMS OF INCOME

Most countries tax the foreign-source interest, rents and royalties payments received by their residents, whether individuals or companies. The return to equity investment – a dividend receipt – is therefore typically taxed in a very different way to the returns from other forms of investment. This issue has not been directly addressed in the economic literature investigating the optimal taxation for foreign corporate income, and therefore requires some discussion. There are at least two different forms of payment here. The first reflects a return to activity which has taken place in the home country. In principle, for example, if a patent is developed by a British company, and a foreign affiliate is licensed to use that patent in its activity, then it would pay a royalty to the British owner of the patent for the right to use it. The receipt of this royalty would be taxed in the UK (but typically not in the other country). This does not contradict the general principle of exempting foreign income discussed above, since the royalty is in effect a return to the research and development activity which took place in the UK and which resulted in the patent. However, a second form of payment does appear to contradict the principle of exempting foreign income. A classic example here is the distinction between debt and equity. Typically, economists argue that the tax system should not differentiate between investment financed by debt and by equity. Indeed, in practice it is becoming increasingly difficult to distinguish the form of financial contracts, where such contracts may have some of the traditional elements of both debt and equity. However, let us leave to one side here the issue of whether debt and equity should be treated equally, and consider only the international aspect of this. If dividends received from foreign affiliates are exempt home country tax, then what is the appropriate treatment of interest received? A natural starting point here would be to simply refer to the analysis above, which made no distinction according to the form of financial transaction between parent and foreign subsidiary. Suppose a British parent company raised finance on the world market in order to undertake an investment in France. The analysis above 15

Vann (2007) discusses a range of more practical problems with arm’s length principle.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

519

suggests that the UK should not seek to tax the income arising in France. The fact that the company may choose to finance the French subsidiary through debt, rather than through equity, should in principle make no difference. However, the structure of the international tax system is problematic here, since almost all countries would grant relief to the interest paid to the British parent (subject to host country thin capitalisation rules). If the UK were simply to exempt interest income received from foreign subsidiaries as implied above, then this income would not be taxed at the corporate level at all. Indeed, given the preferential treatment of interest payments by the subsidiary, there would be a significant advantage to British companies choosing to lend to their foreign subsidiaries. The issue here is that the basic structure of the international tax system for multinational companies is close to a source-based tax for equity-financed investment, but a corporate-residence based tax for debt financed investment. It is hard to think of a sensible economic rationale for this practice, especially when the finance provided is internal to the multinational company. One possible response to this distinction is to call for a worldwide change in this practice. But, more feasibly, the issue for an open economy such as the UK, is what its optimal policy should be, given the treatment of interest payments elsewhere. This issue is related to the issue of relief for interest payments by the British parent. If the basic aim of the tax system is to tax profit arising in the UK, then in principle relief should only given for interest payments made by the parent to the extent that the underlying borrowing was used to finance activities which took place in the UK. Conversely, if interest relief is not granted for payments for borrowing used to finance foreign activity, then there is no clear rationale for taxing interest received from foreign affiliates. Exempting foreign source interest receipts from taxation would be a radical departure from the international norm, but one which is implied by the principle that only economic activity taking place in that jurisdiction should be taxed. One caveat to the recipient country introducing such a policy would be to consider whether the country could do better than leaving the interest entirely untaxed. Given that interest is usually taxed in the hands of the recipient, then home country collecting some tax revenue on interest receipts might have a negligible impact on incentives (relative to the no tax case) and therefore raise welfare in the home country.

4.3

CAPITAL GAINS

Another form in which the return to an investment can be earned is through a capital gain. In an international case, a parent company may sell a foreign affiliate, rather than continuing to operate it. The price at which the affiliate can be sold should in principle reflect the net present value of post-tax earnings.16 The presumption here 16

This may reflect the value of the selling company or the acquiring company, but in either case, the price should reflect the post-corporation tax valuation.

520

TAX REFORM IN THE 21ST CENTURY

is that the affiliate is earning a return which is subject to corporation tax in the host country; if the home country of the parent does not seek to tax the dividend stream from the affiliate to the parent, then there is little rationale for taxing the capital gain earned from the sale of the affiliate. However, this argument applies equally to domestic affiliates. If a British parent has a British affiliate which is subject to British corporation tax, then the sale price of that affiliate should reflect the corporation tax due on the future returns that will be made. Imposing capital gains tax on the sale would be a form of double taxation which would be avoided in the case of dividends paid to the parent. Note that this argument applies to the sale of companies which generate a return subject to corporation tax. It does not necessarily apply to the sale of an asset which does not provide a stream of taxed returns but which simply appreciates in value, such as a work of art. In that case, imposing capital gains tax would not imply any double taxation. The sale of assets owned by an affiliate is less clear. To the extent that the value of the asset is determined by the net present value of the post-tax return which the asset generates, then in principle capital gains tax would be a form of double taxation. But to the extent to which the asset valuation depends on other factors, then there may be a case for retaining a capital gains tax charge.

5

Defining the Tax Base

There are two fundamental questions in the taxation of the profits of multinational business. First, ‘where should the profit be taxed?’. This paper has primarily addressed this question. The second question is ‘what should be taxed?’. Devereux and Sorensen (2006) and Auerbach et al. (2007) describe a matrix of alternative possible international tax regimes which answer these two questions in different ways. In the context of the appropriate taxation of international profit, it is worth raising the second question, if only to highlight the difficulties in interpreting the ‘effective tax rates’ used, but not defined, in Sections 2 and 3. An important distinction here is between effective marginal and average tax rates. The most common measure used in economic theory is an effective marginal tax rate, which measures the percentage difference between the pre-tax and post-tax rates of return on a marginal investment. This is indeed the most appropriate use in the context of Section 2, where the analysis examines required rates of return, and by implication marginal investment projects. But it is well known that it is possible to design a tax system which in effect taxes only economic rent; since it does not tax a marginal project, the effective marginal tax rate is zero. A flow of funds tax (see, for example, Meade, 1978) or a tax with an ACE allowance (see IFS, 1990, Bond and Devereux, 1995, 2003) has these properties. One way of achieving full harmonisation of effective marginal tax rates would therefore be to have cooperation over agreement to introduce such a tax base.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

521

Unfortunately, however, even this approach would not achieve production efficiency in a broader context. This is because in a world with discrete investment choices, it is the effective average tax rate which affects those choices. The idea is that a company which is choosing between two profitable location decisions, for example, would choose the location with the higher post-tax net present value. The role of tax is in determining how much of the pre-tax net present value remains after tax: this is measured by the effective average tax rate (see Devereux and Griffith, 1998, 1999). Neither a cash flow tax nor a tax with an ACE allowance has a zero effective average tax rate. So, unless statutory tax rates were completely harmonised (as well as bases), investment decisions would remain distorted. Of course, the nature of the optimal tax system is in any case more difficult in such a setting, since it is a setting in which companies may earn an economic rent. As we discussed above, the presumption in favour of production efficiency strictly holds only when there is no economic rent or it is taxed at a rate of 100 per cent. But if we aim to achieve production efficiency as an approximately optimal system, using taxes based on the residence or source of the company, then the presence of discrete choices does indicate that both effective marginal and effective average tax rates need to be harmonised.

6

Conclusions

This paper examines economic principles which underlie the optimal taxation of international corporate profit. Sections 2 and 3 review and extend the analysis of the forms of taxation which can generate global and national optimality respectively. Section 4 discusses issues which arise in attempting to implement a source-based tax, and Section 5 briefly raises issues concerning the nature of effective tax rates. Global optimality would generally be achieved by production efficiency. The analysis in Section 2 demonstrates that – in the context of cross-border portfolio and direct investment, as well as international trade in goods and services – taxes on a source or a residence basis would need to be fully harmonised to achieve production efficiency. The conditions for national optimality depend partly on the characteristics of the economy: the government could, in principle and in some cases, induce a welfare gain by stimulating or reducing outbound direct investment. However, Section 3 argues that a central case is where domestic and outbound investment are both financed at the margin by inflows of portfolio investment, and the country has no market power: in this case, there is no convincing argument for taxing the returns from outbound direct investment, even if domestic investment is taxed on a sourcebasis. This leads to a presumption in favour of source-based taxation. However, Section 4 considers a number of conceptual problems with the implementation of a source-based tax system, based on separate accounting. The first is whether there is, in many cases, a reasonable conceptual basis for allocating

522

TAX REFORM IN THE 21ST CENTURY

the global profits of a multinational company to individual jurisdictions. Broadly, global profits may be higher simply because the company is global: those returns are not directly related to specific economic activity taking place in a specific country. Particular conceptual and practical difficulties arise with debt finance and with transfer prices on the sale of goods and services within a multinational company. Another important issue considered is whether the return to debt finance should be treated differently from the return to equity finance. Many countries exempt foreign source dividends from tax but tax foreign source interest. There is no clear rationale for this distinction, which implies a source-basis for equity finance and a residence-basis for debt finance. Consideration of the feasibility of source-based taxation does not paint a very optimistic picture of the future of taxes on corporate profit. And the paper has not taken account of tax competition, which appears to be driving down statutory rates of tax around the world. Can taxes on corporate profit survive in the long run, without doing too much harm in distorting corporate behaviour? Perhaps a more radical reform is called for. One possibility to consider is a destination-based tax, levied where sales are made to a final consumer (see Bond and Devereux, 2002, and Auerbach et al., 2008). If the final consumer is immobile, then such a tax would not distort the location of economic activities undertaken by companies. If it applies to all companies competing in the same market, it would not distort the pattern of competition. And since income would be measured only by reference to sales to a third party, intra-company transfers would not be taxable, and transfer pricing would become irrelevant. Such a tax therefore meets the objectives of a globally optimal tax system outlined above. It would also be optimal from the perspective of a single country, since no source or residence taxes would be applied in that country. If other countries maintained source-based taxes, then that country would attract higher inward direct investment. The feasibility of such a tax needs to be examined in more detail. References Advisory Panel on Canada’s System of International Taxation (2008), ‘Enhancing Canada’s International Tax Advantage’ (Ottawa: Department of Finance, Canada). Auerbach, A.J., M.P. Devereux and H. Simpson (2008), ‘Taxing Corporate Income’, Centre for Business Taxation Working Paper 07/05. Australian Department of the Treasury (2002), Review of International Taxation Arrangements: A Consultation Paper. Becker, J. and C. Fuest (2007), ‘Taxing Foreign Profits with International Mergers and Acquisitions’, Centre for Business Taxation Working Paper 07/19. Brennan, M.J. (1970), ‘Taxes, Market Valuation and Financial Policy’, National Tax Journal 23: 417.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

523

Bond, S.R. and M.P. Devereux (1995), ‘On the Design of a Neutral Business Tax Under Uncertainty’, Journal of Public Economics 58: 57. Bond, S.R. and M.P. Devereux (2003), ‘Generalised R-based and S-based Taxes Under Uncertainty’, Journal of Public Economics 87: 1291. Bond, S.R. and M.P. Devereux (2002), ‘Cash Flow Taxes in an Open Economy’, CEPR Discussion Paper 3401. Bond, S.R., M.P. Devereux and A. Klemm (2007), ‘The Effects of Dividend Taxes on Equity Prices: A Re-examination of the 1997 UK Tax Reform’, Centre for Business Taxation Working Paper 07/01. Crisco, C. and R. Martin (2004), ‘Multinationals and US Productivity Leadership: Evidence from Great Britain’, OECD Science, Technology and Industry Working Paper. Desai, M. and J.R. Hines (2003), ‘Evaluating International Tax Reform’, National Tax Journal 56: 487. Desai, M. (2008), ‘The Decentering of the Global Firm’, Harvard Business School. Devereux. M.P. (1990), ‘Capital Export Neutrality, Capital Import Neutrality, Capital Ownership Neutrality and All That’, Institute for Fiscal Studies, unpublished. Devereux, M.P. (2000), ‘Issues in the Taxation of Income from Foreign Portfolio and Direct Investment’, in S. Cnossen (ed.), Taxing capital income in the European Union (Oxford: Oxford University Press), 110. Devereux, M.P. (2004), ‘Some Optimal Tax Rules for International Portfolio and Direct Investment’, FinanzArchiv 60: 1. Devereux, M.P. and R.G. Hubbard (2003), ‘Taxing Multinationals’, International Tax and Public Finance 10: 469. Devereux, M.P. and R. Griffith (2003), ‘Evaluating tax policy for location decisions’, International Tax and Public Finance 10: 107. Devereux, M.P. and C. Keuschnigg (2008), ‘The Distorting Arm’s Length Principle in International Transfer Pricing’, Oxford University Centre for Business Taxation. Devereux, M.P. and S. Loretz (2008), ‘Increased Efficiency Through Consolidation and Formula Apportionment in the European Union?’, 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: 335. Devereux, M.P. and M. Pearson (1995), ‘European Tax Harmonisation and Production Efficiency’, European Economic Review 39: 1657.

524

TAX REFORM IN THE 21ST CENTURY

Devereux, M.P. and P.B. Sorensen (2006), The Corporate Income Tax: International Trends and Options for Fundamental Reform. Diamond, P. and J. Mirrlees (1971), ‘Optimal taxation and Public Production I: Production Efficiency’, American Economic Review 8. Edwards, J.S.S. (2005), ‘Gains from Trade in Government Revenue and Paretoefficient International Taxation’, Topics in Economic Analysis and Policy 5: 22. Feldstein, M. and D. Hartman (1979), ‘The Optimal Taxation of Foreign Source Investment Income’, Quarterly Journal of Economics 93: 613. Gordon, R.H. (1986), ‘Taxation of Investment and Savings in a World Economy’, American Economic Review 76: 1086. Gordon, R.H. and D.F. Bradford (1980), ‘Taxation and the Stock Market Valuation of Capital Gains and Dividends’, Journal of Public Economics 14: 109. Halperin, R. and B. Srinidhi (1987), ‘The Effects of the US Income Tax Regulations’ Transfer Pricing Rules on Allocative Efficiency’, The Accounting Review 62: 686. Halperin, R. and B. Srinidhi (1991), ‘US Income Tax Transfer-pricing Rules and Resource Allocation: The Case of Decentralized Multinational Firms’, The Accounting Review 66: 141. Harris, D.G. and R.C. Sansing (1998), ‘Distortions Caused by the Use of Arm’s Length Transfer Prices’, The Journal of the American Taxation Association, 20, supplement: 40. HM Treasury and HM Revenue and Customs (2007), Taxation of the foreign profits of companies: a discussion document (London: HM Treasury). Horst, T. (1980), ‘A Note on the Optimal Taxation of International Investment Income’, Quarterly Journal of Economics 94: 793. Huizinga, H. and S.B. Nielsen (1997), ‘Capital Income and Profit Taxation with Foreign Ownership of Firms’, Journal of International Economics 42: 149. Huizinga, H. and S.B. Nielsen (2002), ‘The Coordination of Capital Income and Profit Taxation With Cross-ownership of Firms’, Regional Science and Urban Economics 32:1. Institute for Fiscal Studies (1991), Equity for Companies: A Corporation Tax for the 1990s, IFS Commentary C026 (London: Institute for Fiscal Studies). Keen, M.J. (1993), ‘The Welfare Economics of Tax Co-ordination in the European Community: A Survey’, Fiscal Studies 14.2: 15. Keen, M.J. and H. Piekola (1996), ‘Simple Rules for the Optimal Taxation of International Capital Income’, Scandinavian Journal of Economics 99: 447. Keen, M.J. and D.E. Wildasin (2004), ‘Pareto-efficient International Taxation’, American Economic Review 94: 259.

MICHAEL DEVEREUX: TAXATION OF OUTBOUND DIRECT INVESTMENT

525

McLure, C. (1992), ‘Co-ordinating business taxation in the single European market: the Ruding Committee report’, EC Tax Review 1: 13. Meade Committee (1978), The Structure and Reform of Direct Taxation (London: Allen and Unwin). Mintz, J. and H. Tulkens (1996), ‘Optimality properties of alternative systems of taxation of foreign source income’, Journal of Public Economics 60: 373. Musgrave, P. (1969), United States Taxation of Foreign Investment: Issues and Arguments (Cambridge, MA: International Tax Program, Harvard Law School). New Zealand Inland Revenue Department and New Zealand Treasury (2007), New Zealand’s International Tax Review: The Treatment of Foreign Dividends and Transitional Issues. Richman, P. (1963), Taxation of Foreign Income: An Economic Analysis (Baltimore: Johns Hopkins Press). Simpson, H. (2008), How Do Firms’ Outward FDI Strategies Relate to their Activity at Home? Empirical Evidence for the UK (Bristol, UK: CMPO, University of Bristol). Slemrod, J.B. (1995), ‘Free Trade Taxation and Protectionist Taxation’, International Tax and Public Finance 2: 471. Stiglitz, J.E. and P. Dasgupta (1971), ‘Differential Taxation, Public Goods and Economic Efficiency’, Review of Economic Studies 38: 151. United States Department of the Treasury, Office of Tax Policy (2007), Approaches to Improve the Competitiveness of the US Business Tax System for the 21st Century. Vann, R. (2007), ‘Problems in International Division of the Business Income Tax Base’, University of Sydney, http://www.sbs.ox.ac.uk/NR/rdonlyres/39E0DC63ACD8-4546-84F6-618E037CA20B/0/Vann.pdf. Voget, J. (2008), ‘Headquarter Relocations and International Taxation’, Centre for Business Taxation.

Controlling Tax Avoidance

Chapter 17

Containing Tax Avoidance: Anti-Avoidance Strategies Chris Evans* 1

Introduction

Traditionally the topic of tax avoidance, or how to contain it, has not been a central focus of the public finance literature.1 Indeed, the 5th edition of the classic Musgrave and Musgrave public finance text, Public Finance in Theory and Practice,2 does not mention tax avoidance at all and only makes the briefest mention of tax shelters in the contexts of interest quarantining and the US alternative minimum tax. This omission is slightly surprising given that tax avoidance, like tax evasion, has been around as long as tax itself and can have a major impact on so many aspects of public finance. But the lack of any major attention in the literature in the second half of the 20th century may not be so surprising when one considers that it *

1

2

Professor of Taxation, Atax, University of New South Wales; Senior Fellow, Taxation Law and Policy Institute, Monash University. The work in this paper draws upon, and updates, my work previously published in this area, including C. Evans, ‘Barriers to Avoidance: Recent Legislative and Judicial Developments in Common Law Jurisdictions’ (2007) 37(1) Hong Kong Law Journal, 103; C. Evans, ‘Nuclear deterrents, snipers, shotguns and more’, ‘Editorial’ (2007) 36(3) Australian Tax Review, 133; and C. Evans, ‘The Battle Continues: Recent Australian Experience with Statutory Avoidance and Disclosure Rules’, in J. Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 37. A notable exception is the chapter by M. Brooks and J. Head, ‘Tax Avoidance: In Economics, Law and Public Choice’, in G.S. Cooper (ed.), Tax Avoidance and the Rule of Law (Amsterdam: IBFD, 1997), p. 53. R. Musgrave and P. Musgrave, Public Finance in Theory and Practice (5th edn, Singapore: McGraw-Hill, 1989).

530

TAX REFORM IN THE 21ST CENTURY

is only relatively recently that avoidance and evasion have become the major social phenomena that they now are. The general conclusion of commentators in the UK, the US and elsewhere3 is that tax avoidance activity has really only grown significantly in recent decades. Revenue authorities readily concur as to the recent heritage. As noted in a recent South African Revenue Service (‘SARS’) Discussion Paper on Tax Avoidance,4 the growth in tax avoidance activity is a worldwide concern and ‘has been a growing problem internationally during the past ten years’. An army of what Tanzi5 and Braithwaite6 have respectively called fiscal and moral termites has been eating away at tax revenue bases throughout the world in an unprecedented fashion over the last 30 or so years, and with increasing vigour in the last decade. It is therefore appropriate that the topic of tax avoidance – and more particularly of how it can be constrained or contained – should be a key aspect when tax reform in the 21st century is under consideration. The chapter initially explores (in Section 2) the nature of avoidance activity, the reasons for its growth, and the reasons why we should be concerned, in order to provide appropriate background and context. It then considers, in turn, administrative (Section 3), legislative (Section 4) and judicial (Section 5) responses to concerns about tax avoidance activity. The separation of the review into these three convenient areas is somewhat arbitrary, as they are far more integrated and interdependent than such separation suggests. Section 6 offers some concluding thoughts. The principal objective of the chapter is to identify, examine and evaluate some of the key trends in the responses that national governments and international organisations are currently making to the problems of tax avoidance. The conclusion is that national tax authorities, international tax organisations, and national and supranational legislatures and judiciaries, have at their disposal a large range of anti-avoidance strategies that are designed to deal, and are generally capable of dealing, with the onslaught of these so-called fiscal and moral termites. Moreover, there is some evidence that the battle being waged against widespread tax avoidance

3

4 5 6

For example: N. Tutt, The Tax Raiders: The Rossminster Affair (London: Financial Training, 1985) and The History of Tax Avoidance (London: Wisedene, 1989); J. McBarnet and C. Whelan, Creative Accounting and the Cross-Eyed Javelin Thrower (Chichester: John Wiley, 1999); J. Freedman, ‘Defining Taxpayer Responsibility: In Support of a General AntiAvoidance Principle’, [2004] 4 British Tax Review, 332; J. Slemrod, ‘An Empirical Test for Tax Evasion’ (1985) 67 Review of Economics and Statistics 232; J. Bankman, ‘The New Corporate Tax Shelters Market’ (1999) 83 Tax Notes, 1775; J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005). SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005) at pp. 3, 16, 19. V. Tanzi, Globalization Technological Developments and the Work of Fiscal Termites (Washington DC: International Monetary Fund WP/00/181, 2000). J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005).

CHRIS EVANS: CONTAINING TAX AVOIDANCE

531

is tilting in favour of the defenders of the fisc.7 But while perhaps there is no need for fiscal or moral panic, there is no room for complacency either.

2 Tax Avoidance: Background 2.1 WHAT IS TAX AVOIDANCE? Tax avoidance comes under many labels. In Australia it is often referred to (particularly by the Australian Tax Office – ‘ATO’) as ‘aggressive tax planning’;8 in South Africa as ‘impermissible or abusive tax avoidance’;9 in New Zealand and the United Kingdom as ‘unacceptable tax avoidance’;10 and in the US terms such as ‘tax abusive shelters’ are often used. Whichever term is used, tax avoidance is often contrasted with tax evasion, and also with tax planning/mitigation. The distinction between tax evasion and tax avoidance is well recognized. It is the difference between working outside the law and working within the law (though against its spirit).11 The former Chancellor of the Exchequer in the UK, Denis Healey, suggested that ‘the difference between tax avoidance and tax evasion is the thickness of a prison wall’.12 The OECD notes that tax evasion involves ‘illegal arrangements through or by means of which liability to tax is hidden or ignored …[such that]…the taxpayer pays less tax than he is legally obligated to pay by hiding income or information from the tax authorities’.13 But the OECD finds it more difficult to offer such a precise definition for tax avoidance, suggesting, ‘somewhat awkwardly’,14 that tax avoidance is ‘an arrangement of a taxpayer’s affairs that is 7

8 9 10

11 12 13 14

For example, in August 2005 US Government prosecutors secured a USD 456 million settlement with KPMG LLP as part of a deferred prosecution agreement in which KPMG admitted to fraudulent conduct in the design and marketing of a series of tax shelters including Flip, Opis, Blips and SOS. See also D. Weisbach, ‘Comments on Recent Developments on Tax Shelters in the US’, in J. Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 57, in which the author suggests that in the US in 2007 ‘there seems to be a consensus that the large scale retail marketing of tax shelters has slowed significantly’ (p. 60). See, for example, ATO, Aggressive Tax Planning End-To-End Process (ATO Practice Statement Law Administration PS LA 2005/25, December 2005). See, for example, SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005). See, for example, Lord Templeman’s judgment in the case of CIR (NZ) v. Challenge Corporation Ltd, [1987] AC 155, and Lord Goff’s judgment in Ensign Tankers (Leasing Ltd) v. Stokes [1992] 1 AC 655. R. Woellner, S. Barkoczy, S. Murphy, and C. Evans, Australian Taxation Law (18th edn, Sydney: CCH, 2008), at pp. 1484−1488. , 23 October 2006. International Tax Terms for the Participants in the OECD Programme of Cooperation with Non-OECD Economies (Paris: OECD, 2007). See, for example, SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005).

532

TAX REFORM IN THE 21ST CENTURY

intended to reduce his liability and that although the arrangement could be strictly legal it is usually in contradiction with the intent of the law it purports to follow’.15 The distinction between (unacceptable) tax avoidance and (acceptable) tax mitigation/planning is generally recognized as not being as clear as the distinction between avoidance and evasion. There is a great deal of legislation in all countries prompted by the distinction, and the proliferation of litigation in the area of specific and general anti-avoidance provisions suggests that many of those legislative provisions have not hit their targets. This distinction between tax avoidance and tax mitigation was referred to in the Willoughby case,16 where Lord Nolan stated: The hallmark of tax avoidance is that the taxpayer reduces his liability to tax without incurring the economic consequences that Parliament intended to be suffered by any taxpayer qualifying for such reduction in his tax liability. The hallmark of tax mitigation, on the other hand, is that the taxpayer takes advantage of a fiscally attractive option afforded to him by the legislation, and genuinely suffers the economic consequences that Parliament intended to be suffered by those taking advantage of the option. Thus, it is probably reasonable to conclude that tax planning or tax mitigation is ‘concerned with the organisation of a taxpayer’s affairs (or the structuring of transactions) so that they give rise to the minimum tax liability within the law without resort to…impermissible tax avoidance…’.17 Note, however, that seeking to allocate tax evasion, avoidance and planning/ mitigation into neat compartments can over-simplify the position. Not only are the boundary lines between these three concepts not always clear, but some very learned commentators have even cast doubt on the helpfulness of such categorisation. For example, in the McNiven v Westmoreland case, Lord Hoffmann has noted that unless the statutory provisions ‘contain words like ‘avoidance’ or ‘mitigation’, I do not think that it helps to introduce them’.18 In a somewhat different context, Lord Walker of Gestingthorpe has noted that a ‘simple tripartite classification [of] tax evasion – illegal and criminal; tax avoidance – legal but unacceptable; tax mitigation – legal and acceptable…is too crude an analysis to promote understanding of what is a fairly complex subject’.19

15 16 17 18 19

International Tax Terms for the Participants in the OECD Programme of Cooperation with Non-OECD Economies (Paris: OECD, 2007). CIR v. Willoughby [1997] 4 All ER 65, at p. 73. SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005). McNiven v. Westmoreland [2001] STC 257. Lord Walker, Ramsay 25 Years On: Some Reflections on Tax Avoidance (2004) LQR 412, at 416.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

2.2

533

HOW IS TAX AVOIDANCE CHARACTERIZED?

It is possible to characterize tax avoidance activity in any number of ways. For example, one can consider avoidance in relation to its goals, to the nature of its activities, or by reference to its key features. 2.2.1

Goals of Tax Avoidance

Getting to the essence of tax avoidance perhaps requires an understanding of its goals – what tax avoidance sets out to achieve. Clearly the end game is a reduction of tax liability, but that can take a number of forms. It is possible to identify four possible goals that underpin tax avoidance activity: elimination; deferral; recharacterisation; and/or shifting.20 The first possible goal – the permanent elimination of a tax liability needs no further explanation, and is presumably the tax avoider’s nirvana.21 Deferral involves the postponement of the payment of a tax liability, and relies on the concept of the time value of money for its effectiveness.22 Re-characterisation is simply the conversion of the character of an item or transaction – for example from a taxed or a highly taxed item like revenue to a tax exempt or less heavily taxed item like capital.23 The fourth possible goal – shifting – can relate to income or profit shifting (as in shifting a liability from a highly taxed entity to a less heavily taxed or even exempt entity24) as well as value shifting25 where value is shifted between assets. Achievement of any or all of these goals is only possible because of the potential for tax leverage or tax arbitrage that arises as a result of the so-called inconsistencies and discontinuities that exist within national tax jurisdictions and across international tax borders. Of course the inconsistencies and discontinuities that are evident within and between tax systems are not likely to disappear – they exist for very good reasons. Distinctions between, for example, debt and equity, or capital and revenue, or between the assessability of different sorts of entities and different levels of income are fundamental parts of national tax systems, and they exist for both policy and 20 21 22

23

24 25

SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005), at p. 16. In the Australian Spotless case (FCT v. Spotless Services Ltd (1996) 186 CLR 404) nirvana was sought by the taxpayer but on this occasion not attained. The Australian Citylink case (Commissioner of Taxation (Cth) v. Citylink Melbourne Ltd (2006) 80 ALJR 1282; 62 ATR 648; [2006] HCA 35) more than adequately illustrates the advantages of deferral. The 1997 McGuckian case in the UK (IRC v. McGuckian [1997] 1 WLR 991) is a straightforward example of re-characterisation. That case involved a transfer of shares to a non-resident trust, together with the subsequent sale of the rights to dividends from the shares for a lump sum which, it was unsuccessfully contended, was capital in nature. As in the UK Barclays Mercantile case (Barclays Mercantile Business Finance Ltd v. Mawson [2005] STC 1; [2004] UKHL 51). As in the Australian Peabody case (FC of T v. Peabody (1994) 181 CLR 359; 94 ATC 4663).

534

TAX REFORM IN THE 21ST CENTURY

political reasons. But we have to be aware of the consequences. As Jeff Waincymer points out: ‘[p]rogressive rates of taxation encourage income-splitting techniques; tax expenditures in favour of activities deemed worthy of encouragement lead to the creation of tax-inspired shelters; …administrative necessities such as limiting the taxing exercise to a particular period encourage manipulations of the timing of deductions and receipts of income streams’.26 2.2.2

The Techniques of Tax Avoidance

Identification of contemporary techniques for tax avoidance – manipulation of the mismatches within and between tax systems – is not a difficult task, although that is not in any way to under-estimate the ingenuity of their design or the complexity of their construct that is so often a feature of such schemes. Lord Walker of Gestingthorpe, in an unpublished paper27 presented shortly after the decision in the Ramsay case28 was handed down by the House of Lords, identified ‘seven types of tax avoidance’, proceeding from the simplest case to the increasingly complex (and to most observers, increasingly objectionable). These were: (1) (2) (3) (4)

using a relief; finding a gap; exploiting (or abusing) a relief; anti-avoidance karate (by which he meant the capacity for taxpayers to turn to their own advantage statutory provisions designed to prevent tax avoidance); (5) unnatural assets or transactions; (6) pre-ordained transactions; and (7) dodgy offshore schemes. Many would argue (as Lord Walker readily concedes) that the first is not tax avoidance at all – that it falls squarely within the realms of tax mitigation. Professor Willoughby, the taxpayer in the oft-cited Willoughby decision, for example, was not engaged in an elaborate or contrived scheme aimed at tax avoidance. As the judgment in that case clearly shows,29 he was involved in the utilisation of a tax regime enacted by Parliament which provided tax deferral for bona fide long term retirement saving. But before it is readily concluded that ‘using a relief’ should not 26

27 28 29

J. Waincymer, ‘The Australian Tax Avoidance Experience and Responses: A Critical Review’, in G.S. Cooper (ed.), Tax Avoidance and the Rule of Law (Amsterdam: IBFD, 1997), p. 247, at p. 248. Lord Walker ‘Ramsay 25 Years On: Some Reflections on Tax Avoidance’ (2004) LQR 412, at p. 416. WT Ramsay Ltd v. Inland Revenue Commissioners [1982] AC 300. CIR v. Willoughby [1997] 4 All ER 65.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

535

figure in any taxonomy of avoidance, it might be worth bearing in mind that the more recent House of Lords case of Barclays Mercantile,30 which did involve an elaborate and contrived scheme, was – at essence – simply about a finance leasing company taking advantage of a relief (in this case, capital allowances) that Parliament had intended to be of benefit to some companies (the question being whether the intention was to extend it to finance companies). The fact that the taxpayers in both Willoughby and Barclays Mercantile were successful does not mean that simply ‘using a relief’ can be readily discarded from Lord Walker’s hierarchy of avoidance techniques. And by the same token (again conceded by Lord Walker) not all offshore schemes are by any stretch of the imagination worthy of the label of ‘dodgy’. 2.2.3

The Key Features of Tax Avoidance

Many techniques share a series of common characteristics, most of which are very competently summarized in the paper on tax avoidance prepared in late 2005 by the South African Revenue Service.31 In the view of that revenue authority the ‘badges’ or ‘hallmarks’ of avoidance typically include any or all of the following features: • the lack of economic substance (usually resulting from pre-arranged

• •

• • • •

30 31 32

circular or self-cancelling arrangements), with the result that an apparently significant investment proves ultimately to be illusory, and, through various devices, the taxpayer remains insulated from virtually all economic risk, while creating a carefully crafted impression to the contrary; the use of tax-indifferent accommodating parties or special purpose entities, often referred to in the jargon as ‘washing machines’; unnecessary steps and complexity, often inserted to prop up a claim of business purpose, or to disguise the true nature of a scheme or ‘as a device to cloak the tax shelter transaction from detection’; 32 inconsistent treatment for tax and financial accounting purposes; high transaction costs; fee variation clauses or contingent fee provisions; the use of new, complex financial instruments such as derivatives, hybrids and synthetic instruments which have made it possible for promoters to mimic almost perfectly the risks and returns attributable to more traditional financial instruments such as equity shares or ‘plain vanilla’

Barclays Mercantile Business Finance Ltd v. Mawson [2005] STC 1; [2004] UKHL 51. SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005), at pp. 19−27. SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005), at p. 16.

536

TAX REFORM IN THE 21ST CENTURY

debt without incurring, at least in theory, the tax consequences typically associated with them; and • the use of tax havens, particularly in the context of captive insurance companies, captive finance subsidiaries and intangible property holding companies. Of course this is not to suggest that the existence of these characteristics, either alone or in combination, must necessarily point to the existence of tax avoidance activity. That conclusion can only be drawn after a careful consideration of all of the facts. But it is to suggest that, prima facie, the existence of these features, alone or in combination, may indicate avoidance activity. The Anti-Avoidance Group (‘AAG’) of Her Majesty’s Revenue and Customs (‘HMRC’) in the UK has developed a similar list of ‘signposts’, identifying the factors that it considers may indicate avoidance.33 These include • transactions or arrangements which have little or no economic substance or which have tax consequences not commensurate with the change in a taxpayer’s (or group of related taxpayers’) economic position. For example, forex matching abuse where the position is flat in economic substance but the benefit of hindsight can be used to choose which of two equal and opposite positions is taxed. • transactions or arrangements bearing little or no pre-tax profit which rely wholly or substantially on anticipated tax reduction for significant post tax profit. For example dividend buying where the value of foreign tax credit more than offsets a pre-tax loss; and contrived gilt strip losses counterbalanced by gains on exempt assets such as options. • transactions or arrangements that result in a mismatch such as: between the legal form or accounting treatment and the economic substance; or between the tax treatment for different parties or entities; or between the tax treatment in different jurisdictions. For example a transaction where a payment is treated as an expense in one jurisdiction but the corresponding receipt is not taxable income in another jurisdiction. • transactions or arrangements exhibiting little or no business, commercial or non-tax driver. For example film schemes which create tax losses in excess of the genuine commercial investment. • transactions or arrangements involving contrived, artificial, transitory, pre-ordained or commercially unnecessary steps or transactions. For example the use of artificial arrangements designed to take assets out of the inheritance tax regime while allowing the former owner to continue to enjoy the benefits of ownership. 33

Accessed at 17 May 2008.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

537

• transactions or arrangements where the income, gains, expenditure or losses falling within the UK tax net are not proportionate to the economic activity taking place or the value added in the UK – especially where the transactions or arrangements are between associates within the same economic entity and would not have occurred between parties acting at arm's length and/or add no value to the economic entity as a whole. For example the transfer of ownership of an income stream from a UK resident company to an associated company resident in a low/no tax jurisdiction in circumstances where the economic activity giving rise to the income does not accompany the transfer of ownership and/or where no economic benefit accrues to the economic entity as a whole.

2.3 WHY HAS TAX AVOIDANCE BECOME MORE PREVALENT? Many reasons for the growth of avoidance activity in the latter part of the twentieth century and early part of this have been advanced in the literature, including administrative and judicial inertia. Braithwaite34 has identified globalisation, increasing deregulation and changes in market forces as principal causes. Indeed, Braithwaite positions his work squarely within a broader market context. He shows how the waves of aggressive tax planning in Australia and elsewhere have initially been supply driven. It is his contention that a relatively small group of promoters, including some acting out of major financial institutions and more latterly the (now) Big Four accounting firms, have been the driving force behind many of the schemes that have been adopted by taxpayers in Australia and elsewhere. Much the same point is made by Richards, who notes ‘the conventional wisdom is [that] most of the planning and mass marketing emanates from the accounting firms’.35 There is also little doubt that the supply of tax avoidance products in the market place has been fuelled by the availability of talented human resources prepared to work in the area and by ‘rapid advances in computer and telecommunications technology [which have] greatly enhanced the ability of investment banks, accounting firms and other tax advisers to create sophisticated tax and financial models…to market…to multiple taxpayers’.36 This abundance of supply has, in turn, created a demand for aggressive tax planning opportunities from a much larger group of taxpayers, who feel that they should not miss out on what the ‘big end of town’ is able to enjoy. As Pederick37 34 35 36 37

J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005). G. Richards, ‘Finance Act Notes: Disclosure of Tax Avoidance – Section 19’, [2004] 5 British Tax Review, 451, at pp. 453−454. SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005), at pp. 7, 8. W. Pederick, ‘Fair and Square Taxation for Australia’ (1984) 19(6) Taxation in Australia 575.

538

TAX REFORM IN THE 21ST CENTURY

noted many years ago: ‘[c]ynicism grows apace and a race not to be left out of the tax minimisation derby, by hook or by crook, infects the body politic’. The US Treasury38 has noted that changing attitudes may play a large part in the growth in corporate tax shelters in recent decades. It is simply a reflection ‘of more accepting attitudes of tax advisers and corporate executives toward aggressive tax planning’. And the demand is not merely now restricted to those at the top end of town. For example, the growth in the mid to late 1990s of mass-marketed tax avoidance schemes pedalled by the ‘white shoe brigade’39 to high income blue collar workers operating in the resources sectors in the remoter parts of Australia has shown that tax avoidance activity is now a much more comprehensive and extensive phenomenon than was the case in earlier years. That observation is not confined to Australia.

2.4

IS THE GROWTH IN TAX AVOIDANCE ACTIVITY A PROBLEM?

The question does, of course, need to be asked as to whether all of this growth in avoidance activity needs to be a matter of concern. The answer to this question is relatively straightforward. In the first place, and perhaps most importantly, it impacts negatively on the capacity of national tax jurisdictions to collect the revenue needed for the proper discharge of governmental functions. Revenue collection is the primary function of any tax system, and systematic and widespread avoidance activity will clearly have an adverse impact on that function. When tax revenues do not flow as anticipated, or when large amounts of expected revenue are diverted by successful avoidance activity, cuts in government expenditure will follow, ‘with the resulting social and political difficulties that such cuts may bring’.40 Equally importantly, to the extent shortfalls are made up through higher rates on income that is taxed, there is a serious risk of exacerbating tax-induced economic distortions and inefficiencies. There are no reliable estimates on the losses to national treasuries as a result of avoidance activity, but the amounts are likely to be very significant. For example, one 2005 estimate has suggested that tax haven activity alone has resulted in annual revenue losses to other countries in excess of USD 50 billion.41 In the UK, the tax loss from avoidance is estimated to run into several billion pounds across both

38 39

40 41

US Treasury, The Problem of Corporate Tax Shelters – Discussion, Analysis and Legislative Proposals (Washington DC: United States Treasury Department, 1999), at p. 19. A derisive term originally coined to describe the property developers who worked with Sir Joh Bjelke-Petersen, Premier of Queensland, in a number of suspicious deals in the 1960s, 1970s and 1980s. M. Boyle, ‘Cross-Border Tax Arbitrage – Policy Choices and Political Motivations’, [2005] 5 British Tax Review, 527, at p. 531. SARS, Discussion Paper on Tax Avoidance (Law Administration, South African Revenue Service, November 2005), at p. 27.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

539

direct and indirect taxes.42 In 2008 the US Senate Report into tax haven banks suggested that each year the US ‘loses an estimated $100 billion in tax revenues due to offshore tax abuses’.43 But the harmful effects of tax avoidance activity go well beyond their impact on revenue collections. They also significantly affect the efficiency and equity of tax systems. These impacts are neatly encapsulated by Bankman44 where he notes that ‘[t]ax shelters siphon off resources from more productive ventures, redistribute the tax burden and threaten to undermine compliance’. As the OECD has also noted, any ‘proliferation of arbitrary and contrived schemes…leads to a perception that the system is unfair [which can] discourage compliance, even by taxpayers that had not previously engaged in…tax avoidance’.45 It should also not be forgotten that much of the complexity of modern tax systems is a direct result of the introduction of specific integrity measures, involving convoluted anti-avoidance legislation designed to counter real and perceived avoidance abuses. It can, therefore, safely be concluded that the growth in tax avoidance activity is a matter of grave concern, as it can reduce revenue collections, introduce economic inefficiencies by distorting economic behaviour, undermine the integrity of national tax systems and introduce a host of additional and unwanted complexities to those systems. The next sections of the paper review the various anti-avoidance responses that have emerged from national governments and international organisations, and identify the key strategies that have been put in place in attempts to contain and counter avoidance activities. The analysis begins with a consideration of what might be termed tax authority responses – looking in particular at the work of national and international revenue agencies. It then considers legislative and judicial responses. While administrative, legislative and judicial responses are considered separately for the purposes of the paper, it must be emphasized that in practice they are closely intertwined and interdependent.

42

43

44

45

Her Majesty’s Revenue and Customs Anti-Avoidance Group ‘Our Vision and Strategy’ accessed at 17 May 2008. Note that the Trade Unions Congress, in a 2008 Touchstone Report, suggested that tax avoidance in the UK was costing GBP 25 billion in lost revenue. As noted by B. Dodwell, ‘Missing Billions? Where’s the Evidence?’, Tax Adviser, March 2008, this figure, representing some 5 per cent of UK total tax yield, is unlikely. US Senate Permanent Subcommittee on Investigations, ‘Staff Report on Tax Haven Banks and US Tax Compliance’, 17 July 2008, at p. 1. Note that this estimate was itself derived from studies conducted by a variety of tax experts. J. Bankman, ‘An Academic’s View of the Tax Shelter Battle’, in H. Aaron and J. Slemrod (eds), Crisis in Tax Administration (Washington DC: Brookings Institution Press, 2004), at p. 31. OECD, Harmful Tax Competition: An Emerging Global Issue (Paris: OECD, 1998), at para. 30.

540

TAX REFORM IN THE 21ST CENTURY

In addition it needs to be acknowledged that other forces are also at work to counter some of the effects of avoidance. For example, lower tax rates may be associated with declining avoidance activity, and so international pressures for the reduction in tax rates since the 1980s may well have played some part in countering avoidance. Revenue authorities have also become increasingly conscious of the importance of providing incentives, or ‘compliance carrots’, to the great majority of taxpayers who do wish to comply with their tax obligations, rather than simply beating taxpayers with the various sticks that are available to them. This theme of cooperative compliance is mentioned but not fully developed in the paper which, by its nature, is not focused on the compliant.

3 Tax Authority Strategies to Counter Avoidance 3.1

NATIONAL INITIATIVES

National revenue authorities have been very active in ensuring that their administrative machinery is as well-positioned as it possibly can be to identify and counter what they regard as unacceptable tax avoidance activity. Even the most cursory of trawls through tax office websites around the world brings up a host of national initiatives. Examples drawn from just two jurisdictions – the UK and Australia – will illustrate the sorts of initiatives to counter avoidance that currently typify national responses.46 In the UK, the 2006 Review of Links with Large Business (‘The Varney Review’) shapes much of the current thinking about compliance, and therefore about antiavoidance strategies. The focus of the Varney Review is upon outcomes designed to improve the attractiveness of the UK business tax administrative environment, and within this framework it has identified four key themes: certainty; risk management; speedy resolution of issues; and clarity through effective consultation. These themes have prompted a call for new kinds of relationships to be forged between HMRC and the business community where the watchwords are ‘disclosure, transparency, co-operation and proportionality.’47 Within this context, the development, maintenance and delivery of HMRC antiavoidance policies and strategies is the direct responsibility of the Anti-Avoidance Group (‘AAG’). The AAG has an important role in delivering the aims of HMRC and its broader compliance strategy. The AAG’s published strategies include48 making 46

47

48

For an overview of recent US initiatives in countering avoidance activity, see D. Weisbach, ‘Comments on Recent Developments on Tax Shelters in the US’, in J Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 57. D. Hartnett, ‘Boundaries, Behaviour and Relationships: The Future’, in J. Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 179. Her Majesty’s Revenue and Customs Anti-Avoidance Group ‘Our Vision and Strategy’, , 17 May 2008.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

541

tax law robust against avoidance; engaging with its customers about its approach to avoidance; optimising its operational response to avoidance; and changing the economics of avoidance to make it less attractive. More specifically, the AAG identifies a series of actions designed to achieve these outcomes. These are: • to quickly and expertly prevent and close down avoidance by effective

legislation; • to engage with its customers about how to address avoidance; • to know what avoidance schemes or bespoke arrangements are being • • • •

marketed and used; to know which organisations and individuals are more likely to carry out avoidance and organize resources accordingly; to treat those who avoid their tax obligations as higher risk than organisations and individuals who do not; to be proactive at challenging avoidance by investigation and effective litigation; and to take a strategic approach in litigating avoidance cases.

This robust and proactive approach to avoidance activity is based upon real-time intelligence (often made available through the scheme disclosure rules discussed later in the paper), together with the sensible application of the principles of risk management and proportionality. It fits comfortably within the broader cooperative compliance framework espoused in the Varney Review and HMRC publications. Similar trends are visible in Australia. The ATO publishes a comprehensive annual compliance program49 which describes the tax compliance risks it is most concerned about and what it is doing to address them. ‘Aggressive tax planning’, which the ATO defines as ‘the use of transactions or arrangements that have little or no economic substance and are created predominantly to obtain a tax benefit that is not intended by the law’, features as a prominent part of that annual program. The program identifies the ATO’s general approach to aggressive tax planning, certain headline issues, and its current focus and priorities, as well as actions taken and successes achieved in the preceding year. The approach of the ATO to large business was spelled out in a 2007 speech by the Commissioner.50 In that speech the Commissioner notes that ‘the challenge for large business and the Tax Office is to create certainty through transparency and cooperation’, and emphasizes the current ATO mantra of ‘consultation, collaboration 49 50

The latest is the Compliance Program 2008−09, . M. D’Ascenzo, Creating the Right Environment: Transparency, Cooperation and Certainty in Tax, Financial Executives International of Australia Conference, Sydney, 19 June 2007 .

542

TAX REFORM IN THE 21ST CENTURY

and co-design’. Mention is also made of two ‘Forward Compliance Arrangements’ made with the ANZ bank and BP in the areas of goods and services tax and excise tax, designed to promote governance arrangements that reduce the company’s risk of audits, tax litigation, penalties and interest, and also streamline access to support and advice. Further developments in the direction of such compliance arrangements took place in May 2008, when the Commissioner of Taxation announced that the top 50 companies by turnover in Australia will be given the option of entering into an Annual Compliance Arrangement (‘ACA’) with the Tax Office on a voluntary basis.51 The intention of ACAs is to provide these companies with certainty in relation to their tax position. Under an ACA the ATO would issue the taxpayer with a sign-off letter confirming the outcomes of a joint risk assessment (including the development of a risk management plan). To the extent of the disclosure, the ATO would agree not to audit low risk matters, and for higher risk issues the company would know what those issues are. The ACAs are built around the company that enters into the arrangement with the ATO having sound tax risk management processes, and making full and true disclosure. There is an expectation of sound corporate governance, early dialogue and mitigation of tax risks. In summary, the UK and Australian responses to aggressive tax planning have moved beyond the ‘command and control’ frameworks that typified the 1970s and 1980s to one of ‘responsive regulation’ and ‘meta risk management’ from the 1990s onwards.52 Braithwaite explains that under the command and control approach ‘[t]axpayers lodged their returns, the [revenue authority] assessed them and decided how much tax was due. Audits were conducted to detect the provision of false information on returns, which, when detected, typically resulted in the imposition of modest penalties’.53 That command and control mentality involved the seesaw of ‘carrot and stick’ approaches, oscillating between a customer service focus and one which relied on punitive legal action, depending on whichever particular philosophy happened to be in the ascendant at a particular time within the organisation. Responsive regulation involves an enforcement pyramid (now encapsulated in Australia in the ATO’s Compliance Model) in which the bulk of taxpayers engaged in cooperative compliance are situated at the base of the pyramid, while a small hard-core of recalcitrant offenders are at the apex. Little enforcement activity is required for those at the base of the pyramid; essentially they require only the positive encouragement to comply. On the other hand the revenue authority possesses a credible capacity to escalate up the pyramid to progressively more severe sanctions 51 52 53

M. D’Ascenzo, A New Dimension, Corporate Tax Association Convention, Sydney, 12 May 2008. J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005). J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005), at p. 68.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

543

in the face of persistently aggressive non-compliance. In Braithwaite’s terms, responsive regulation involves ‘sending clear signals through concrete enforcement actions that the agency is willing to escalate in order to create a culture where systemic preventive solutions and good relationships with taxpayers will do most of the compliance work’.54 As part of this responsive regulation, meta risk management simply refers to the ‘risk management of risk management’.55 It entails, as in the UK and Australia, revenue authority monitoring of the tax community’s self monitoring and self regulation.

3.2

INTERNATIONAL INITIATIVES

Initiatives undertaken by national revenue authorities such as those outlined above for the UK and Australia have been supplemented by many international activities involving cooperation between tax authorities and work by multinational organisations. Work done on the elimination of tax havens and harmful tax competition by the OECD56 would be just one example of the sorts of initiatives that have been taken in this area. More recently the Forum for Tax Administration – established by the OECD in 2002 and attended by national Taxation Commissioners since 2004 with a mandate to develop effective responses to current administrative issues in a collaborative way by working with member and certain non-member countries – has considered the challenges faced with non-compliance with tax laws in an international context. In September 2006, 35 Commissioners or Deputy Commissioners signed the Seoul Declaration which stated that: Each country differs in the level and structure of its taxes, but all countries – both low and high tax countries, developed and developing – agree that once national tax laws have been enacted, they need to be enforced. Enforcement of our respective tax laws has become more difficult as trade and capital liberalisation and advances in communications technologies have opened the global marketplace to a wider spectrum of taxpayers. While this more open economic environment is good for business and global growth, it can lead to structures which challenge tax rules, and schemes and arrangements by both domestic and foreign taxpayers to facilitate non-compliance with our national tax laws. It is our duty as heads of our respective countries’ revenue bodies to ensure compliance with our national tax laws by all taxpayers,

54 55 56

J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005), at p. 178. J. Braithwaite, Markets in Vice: Markets in Virtue (Leichardt: Federation Press, 2005), at p. 85. See, for example, OECD, Harmful Tax Competition: An Emerging Global Issue (Paris: OECD, 1998).

544

TAX REFORM IN THE 21ST CENTURY

including activities beyond our borders, through effective enforcement and by taking preventive measures that deter non-compliance.57 The Seoul Declaration set up an OECD study to examine the role of tax intermediaries (advisers) in relation to non-compliance and the promotion of unacceptable tax minimization arrangements. The final report58 emanating from the study concluded that some tax advisers do play a critical role in designing and promoting aggressive tax planning, but that it was overly simplistic to focus just on the supply side. Taxpayers themselves represent the demand side for aggressive tax planning, setting their own strategies for tax-risk management and determining their own appetites for tax risk. The report therefore considered the tripartite relationship between revenue bodies, taxpayers and tax advisers, and concluded that appropriate risk management tools were vital for revenue bodies. It further considered that revenue bodies needed to operate using the five following attributes in their dealings with all taxpayers: • understanding based on commercial awareness; • impartiality; • proportionality; • openness (disclosure and transparency); and • responsiveness.

The report noted that if ‘revenue bodies demonstrate these five attributes and have effective risk-management processes in place, this should encourage large corporate taxpayers to engage in a relationship with revenue bodies based on cooperation and trust’, described in the report as an ‘enhanced relationship’59. Such an enhanced relationship would benefit both the revenue authority and the taxpayer (through greater certainty and lower compliance costs). The report also accepted, however, that not all taxpayers would wish to adopt such enhanced relationships and that the demand for aggressive tax planning will not disappear completely. Thus ‘revenue bodies will need to have effective risk-management processes in place to identify these taxpayers and allocate the necessary level of resources to deal with them’.60 A further example of international cooperation between national tax authorities, this time outside the OECD, is the work of the Joint International Tax Shelter Information Centre (‘JITSIC’) in countering international cross-border tax arbitrage activities on a real time basis. 57

58 59 60

D. Butler, ‘Tax Administration in an International Context – The Study into the Role of Tax Intermediaries’, Tax Administration: Safe Harbours and New Horizons, Eighth Atax International Tax Administration Conference, Sydney, March 2008. OECD, Study into the Role of Tax Intermediaries (Paris: OECD, 2008). OECD, Study into the Role of Tax Intermediaries (Paris: OECD, 2008), at p. 5. OECD, Study into the Role of Tax Intermediaries (Paris: OECD, 2008), at p. 6.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

545

JITSIC was established in May 2004 between the tax authorities of Australia, Canada, the UK and the US, with the objective of supplementing ‘the ongoing work of tax administrations in identifying and curbing abusive tax avoidance transactions, arrangements, and schemes’.61 An early Press Release noted that an initial focus would include ‘the ways in which financial products are used in abusive tax transactions by corporations and individuals to reduce their tax liabilities, and the identifications of promoters developing and marketing those products and arrangements’.62 The purpose of this international task force is to: • provide support to the parties through the identification and understanding of abusive tax schemes and those who promote them; • share expertise, best practice and experience in tax administration to combat abusive tax schemes; • exchange information on abusive tax schemes, in general, and on specific schemes, their promoters and investors, consistent with the provisions of bilateral tax conventions; and • enable the parties to address better the abusive tax schemes promoted by firms and individuals who operate without regard to national borders.63 Boyle64 has suggested that there are initial indications that JITSIC’s efforts are working and cites Mark Everson, then the head of the IRS, as saying: ‘We have seen things we either would never have picked up or would have picked up years down the road…. We have seen a series of kinds of transactions, or in some cases particular transactions, that merit follow-up by the individual taxing authorities.’ Boyle also notes, however, that the actual output and practical impact of the organisation is difficult to ascertain as a result of the secrecy that surrounds much of JITSIC’s operations. These specific examples of international collaboration between some of the revenue authorities – and there are plenty of other examples of both broader and narrower cooperation65 – illustrate that governmental responses to tax avoidance 61 62 63 64 65

Joint International Tax Shelter Information Centre Memorandum of Understanding, 27 October 2006. ‘Australia, Canada, UK and US Agree to Establish Joint Task Force’ (IRS Newswire IR2004-61, 3 May 2004). Joint International Tax Shelter Information Centre Memorandum of Understanding, and http://www.ato.gov.au, 27 October 2006. M. Boyle, ‘Cross-Border Tax Arbitrage – Policy Choices and Political Motivations’ [2005] (5) British Tax Review 527, at 533−534. See M. Dirkis, ‘Looking Beyond Australia’s Horizon: The internationalisation of Australia’s Domestic Taxation Information Gathering and Debt Collection Powers’, Tax Administration: Safe Harbours and New Horizons, Eighth Atax International Tax Administration Conference, Sydney, March 2008.

546

TAX REFORM IN THE 21ST CENTURY

activity – just like the avoidance activities themselves – go well beyond national borders. Revenue authorities have shown themselves to be more than willing to adopt a proactive and cooperative administrative stance to complement national activities and initiatives in the legislative and judicial spheres. The paper now considers some of those legislative and judicial initiatives.

4

Legislative Responses to Avoidance Activity

Common law jurisdictions have not been reluctant to adopt direct legislative responses to the threats posed by avoidance activity, and the pace of the introduction of such measures has quickened in direct response to the perceived growth in the threat. Legislative responses have been on a number of broad fronts: • the introduction in all jurisdictions of specific or targeted anti-avoidance

rules (‘SAARs’ and ‘TAARs’) aimed at particular areas where abuse has been identified or where revenue leakage is suspected, together with the continuing refinement and use of general anti-avoidance rules (‘GAARs’) and legal principles in a number of the jurisdictions; • the use of product or scheme disclosure rules to obtain real-time intelligence on avoidance market developments and the use of promoter penalty rules as a further weapon in the arsenals available to revenue authorities; and • the use of principles based drafting to combat avoidance activities. These three broad areas are dealt with in turn, by reference to some major recent developments in some of the common law jurisdictions.

4.1 ANTI-AVOIDANCE LEGISLATION 4.1.1

Specific Legislation Targeting Tax Avoidance

The intellectual dexterity of advisers and the complexity of commercial transactions have, over recent years, caused a mushrooming of schemes and arrangements that have attracted revenue authority attention and later the introduction of specific antiavoidance rules to control their effectiveness. The use of specific integrity measures – as opposed to the use of general anti-avoidance rules which are dealt with later – has the advantage of precision. They are the ‘smart bombs’ in contrast to the ‘carpet bombs’ or ‘weapons of mass destruction’66 that may be represented by general antiavoidance provisions.

66

This analogy is provided by A. Halkyard in ‘Not a Weapon of Mass Destruction: Can the Ramsay Approach Apply to the Inland Revenue Ordinance in Hong Kong?’ (2005) 9(3) AsiaPacific Journal of Taxation 56, at 57.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

547

But they also carry the concern that their aim can be poor or that their impact can be deflected. Sometimes conceived in haste and rushed through national parliaments without due concern for the careful legislative drafting that must accompany any new tax measure, scheme promoters and tax advisers are often able to dodge their impact and even to incorporate them into fresh iterations of avoidance activity – the anti-avoidance karate identified by Lord Walker which was mentioned earlier. The Australian experience with the introduction of specific anti-avoidance measures designed to clamp-down on the alienation of personal services income through the use of interposed entities, introduced to operate with effect from 1 July 2000, is a classic example.67 The initial legislation was well-intentioned but poorly conceived, and it is only after a number of refinements and amendments have passed through Parliament over the ensuing years that the measures have begun to operate as intended. The UK has been among the more prolific of the common law jurisdictions in introducing new specific anti-avoidance measures in recent years. For example, the 2005 Budget saw the introduction of three new and extensive sets of specific antiavoidance rules targeting, respectively, arbitrage, double tax relief avoidance and financial avoidance, as well as even more specific provisions to address abusive film schemes. Two principal factors likely account for the UK’s greater reliance on specific anti-avoidance measures. In the first place, the UK – in contrast to Hong Kong, Australia, New Zealand, Canada and South Africa – does not have a statutory general anti-avoidance provision. While the existence of such a provision will never obviate the need for specific anti-avoidance legislation, and many of those countries who do have such a general rule still manage to generate plenty of specific provisions of their own, those countries with a general anti-avoidance rule do appear to be able to get away with less legislative enactment in this area.68 The second reason for the spate of specific anti-avoidance measures in the UK lies in its recent adoption – which is dealt with later – of a disclosure regime whereby promoters and users of potentially abusive tax avoidance schemes are required to disclose details of those schemes when they are first available for implementation. This boost to real time intelligence has led directly to the introduction of specific legislative measures to counter the identified abuse. 4.1.2

General Anti-Avoidance Provisions

The second legislative response to tax avoidance activity is the continuing refinement and use of GAARs in a number of the jurisdictions. The 1998 Consultative Document on the possible introduction of a general anti-avoidance rule for direct taxes in the 67 68

The anti-avoidance rules are contained in Income Tax Assessment Act 1997, Pt 2-42. See also, D. Weisbach, ‘Formalism in the Tax Law’ (1999) 66 Chicago Law Review 869 in this connection.

548

TAX REFORM IN THE 21ST CENTURY

UK noted then69 that ‘[t]he United Kingdom is unusual among developed countries in having neither a statute nor an established legal principle to counter tax avoidance in general. Many other countries in the developed world have found such a rule or principle to be a very useful remedy for countering tax avoidance, although not a universal cure’. General anti-avoidance rules have operated for many years in most other common law jurisdictions, including Hong Kong (sections 61 and 61A of the Inland Revenue Ordinance), Canada, (section 245 of the Canadian Income Tax Act), New Zealand (sections BG1 and GB1 of the New Zealand Income Tax Act 1994), South Africa (sections 80A to 80L of the Income Tax Act 1962) and Australia. Australian governments use a variety of general anti-avoidance rules to combat what are perceived to be abusive avoidance activities. These include Pt IVA of the Income Tax Assessment Act 1936, section 67 of the Fringe Benefits Tax Assessment Act 1986 and Division 165 of the A New Tax System (Goods and Services Tax) Act 1999. It is difficult to assess the effectiveness of the general anti-avoidance rules in each of the countries where they have operated. Much, obviously, depends upon the perspective adopted. In South Africa, for example, the South African Revenue Service held such grave concerns about the effectiveness of its previous GAAR that a completely fresh approach was adopted with a new provision enacted in late 2006.70 Brian Arnold, writing about recent court decisions in Canada,71 notes that these decisions ‘will inexorably render the rule largely ineffective…’, though Revenue Canada has enjoyed some success in the courts subsequently, particularly in the Mathew case.72 In New Zealand the general anti-avoidance rules have operated with mixed success. Its own High Court recently held in Accent Management Ltd 73 that a forestry investment scheme involving a host of the hallmarks that have been suggested are likely to constitute tax avoidance was in fact tax avoidance. In contrast, the Privy Council somewhat uncertainly held – in the Peterson74 case – that a film scheme which had many of those same hallmarks was not avoidance within the terms of the New Zealand general anti-avoidance rule. The application of the general anti-avoidance provisions has also been anything but certain in Australia. The Commissioner of Taxation has enjoyed considerable success in using the rule to counter aggressive tax planning in the form of mass 69 70

71 72 73 74

Inland Revenue, A General Anti-Avoidance Rule for Direct Taxes: Consultative Document (London: Inland Revenue, 1998), at para. 4.3. E. Liptak, ‘Battling with Boundaries: The South African GAAR Experience’, paper presented at Corporation Tax: Breaking Down The Boundaries, Oxford University Centre for Business Taxation, Oxford, 28 and 29 June 2007. B. Arnold, ‘The Long, Slow Steady Demise of the General Anti-Avoidance Rule’ (2004) 52 Canadian Tax Journal, 488, at p. 488. Mathew v. The Queen [2005] SCC 55. Accent Management Ltd & Others v. CIR (2005) 22 NZTC 19027. Peterson v. CIR [2005] STC 448.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

549

marketed schemes.75 However, the four cases on Part IVA that have been litigated in Australia’s High Court thus far – Peabody, Spotless, Consolidated Press Holdings and Hart76 – have produced mixed results for the Commissioner and have not always provided practitioners with the clear guidance that they often crave as to how the provisions will apply in the particular circumstances in which they are required to advise their clients. An analysis conducted by the UK’s HMRC in 2006 of the general anti-avoidance provisions in a number of countries came to three broad conclusions. Firstly, that the tax landscape of each country was a product of many and diverse factors, including that country’s own history, culture and economic development. As a result, successful strategies in one country could not automatically be transported to another. Secondly, the approach of the judiciary to tax cases in general and avoidance in particular was vital to the effectiveness of anti-avoidance measures. And thirdly, that the impact of a particular anti-avoidance measure was heavily dependent upon the interaction of all the components of a compliance regime. For example, even the strongest GAARs were likely to be ineffective where avoidance schemes were unlikely to be detected in either the marketing phase or at the point at which returns were submitted; and they would also be ineffective if the revenue authority was unwilling to litigate avoidance cases.77 There is, therefore, no obvious conclusion as to whether the arsenal of antiavoidance weaponry is better served with or without a general anti-avoidance measure. The US – where judges have applied a robust, practical and commercial approach to tax avoidance cases – appears not to need such a rule. In the UK the jury is out. In most of those countries that do have a general anti-avoidance rule78 the provisions usually appear to operate successfully, but only – as in all provisions of last resort – where they are used sparingly. Over-use by revenue authorities is an abuse and can only serve to depreciate the value of the provision.

75

76 77

78

See, for example, Howland-Rose and Others v. FCT [2002] FCA 246; 2002 ATC 4200; Vincent v. FCT [2002] ATC 4742; Puzey v. FCT [2003] ATC 4782; FCT v. Sleight [2004] FCAFC 94; [2004] ATC 4477; FCT v. Lenzo [2008] FCAFC 50. In contrast, see FCT v. Cooke & Jamieson [2004] FCAFC 75. FCT v. Peabody (1994) 181 CLR 359; FCT v. Spotless Services Ltd (1996) 186 CLR 404; FCT v. Consolidated Press Holdings Ltd [2001] HCA 32; FCT v. Hart [2004] HCA 26. D. Pickup, ‘Comparative Study of the Legal Frameworks Used by Different Countries to Protect their Tax Revenues’, in J. Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 9. The countries that were examined were the UK, Canada, South Africa, New Zealand, Australia, US, Netherlands and Spain. And none have subsequently dispensed with the rule.

550

4.2

TAX REFORM IN THE 21ST CENTURY

PRODUCT DISCLOSURE RULES AND PROMOTER PENALTY REGIMES

The second area where there have been significant recent developments on the legislative front is the enactment, in many of the common law jurisdictions, of product/scheme disclosure legislation and promoter penalty regimes. Both types of legislative provision are designed to counter, essentially, the mass marketing of what might be loosely called tax exploitation schemes (to borrow the Australian terminology), although both are equally capable of dealing with boutique or one-off avoidance activity. Product disclosure schemes represent a pre-emptive strike, in that they have the capacity to provide revenue authorities with real time intelligence that can allow governments to move rapidly to close down loopholes and block avoidance. Promoter penalty regimes are more reactive and punitive, but can nonetheless act as a significant deterrent to those who might seek to market abusive tax schemes. In 2004, the UK followed the lead of the US79 and Canada by enacting legislation80 designed to provide the tax authority with early information about ‘tax arrangements’ and how they work, together with information about who has used them.81 Tax arrangements were carefully defined to include any arrangement (for example, a scheme, transaction or series of transactions) that will, or might be expected to, provide the user with a tax advantage when compared to adopting a different course of action. When the disclosure regime was introduced in the UK in 2004, disclosure was limited in scope to tax arrangements concerning employment (for example share schemes) or certain financial products. This was significantly widened with effect from August 2006 to the whole of income tax, corporation tax and capital gains tax. There are also disclosure rules relating to Stamp Duty Land Tax, VAT, and, since May 2007, National Insurance Contributions. The introduction of these disclosure requirements in the UK initially caused widespread concern within the tax profession. That concern seems, to some extent, to have been dissipated by the introduction of ‘hallmarking’, which was designed to limit the need to disclose all tax efficient schemes – the intention of the hallmarks (for example, tests such as confidentiality, premium fees, the presence of off-market terms) being to limit disclosure to only those schemes that are new, innovative, or of specific concern. Certainly a House of Lords Select Committee was pleased to note 79

80 81

See D. Weisbach, ‘Comments on Recent Developments on Tax Shelters in the US’, in J. Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 57 for a review of the significant recent changes to US disclosure rules, including changes to reportable transactions and the obligation for tax shelter promoters to provide lists of clients to the Internal Revenue Service. Finance Act 2004, ss 19 and Part 7 (ss 306−319). HM Revenue and Customs, Guidance: Disclosure of Tax Avoidance Schemes (London: HMRC, June 2006), at p. 11.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

551

in 2006 the consensus among private sector witnesses that the rules were working well.82 Those rules have certainly provided HMRC with unparalleled access to realtime intelligence that has enabled it to move swiftly to legislate against avoidance activity deemed to be a threat to the revenue base. Table 1: UK Disclosures Statistics Financial Year 1 Aug 04 – 31 Mar 05 1 Apr 05 – 1 Apr 06 1 Apr 06 – 1 Apr 07 1 Apr 07 – 1 Apr 08

Direct Tax Disclosures

VAT Disclosures

Total Disclosures

503 607 346 274

680 91 65 30

1183 698 411 304

Source: www.hmrc.gov.uk/avoidance/avoidance-disclosure-statistics.htm accessed 18 May 2008.

Table 1 shows scheme disclosures in the first four years of operation of the rules. Note that the statistics do not show the numbers of users of the schemes, as details of clients and users are not required under the rules. Nor do the statistics show the number of generic schemes disclosed, as more than one person may disclose the same type of scheme. Australia (and likewise Hong Kong, New Zealand and South Africa) has not yet implemented a disclosure regime of this nature, though it is a matter of some speculation as to whether it is merely a question of time before the revenue authorities in those countries manage to persuade their political masters of the absolute necessity for such wide-reaching provisions. But, to date and in Australia at least, the recent emphasis has been upon the enactment of legislation designed to impose penalties upon the promoters of what have been labelled ‘tax exploitation schemes’. Legislation83 designed to deter the promotion of tax exploitation schemes in Australia was introduced in early 2006. The legislative provisions seek to deter: the promotion of tax avoidance and evasion schemes (collectively referred to in the legislation as tax exploitation schemes); and the implementation of schemes that have been promoted on the basis of conformity with a product ruling, in a way that is materially different to that described in the product ruling.

82

83

Cited in J. Tiley, ‘The Avoidance Problem: Some UK Reflections’, in J. Freedman (ed.), Beyond Boundaries: Developing Approaches to Tax Avoidance and Tax Risk Management (Oxford: Oxford University Centre for Business Taxation, 2008), p. 67. Division 290 of Tax Administration Act 1953.

552

TAX REFORM IN THE 21ST CENTURY

In summary84 the provisions enable the Commissioner to: • request the Federal Court of Australia to impose a civil penalty on a

scheme promoter or implementer. The maximum penalty the Federal Court can impose is the greater of 5,000 penalty units (currently equal to AUD 550,000) for an individual or 25,000 penalty units (currently equal to AUD 2.75 million) for a body corporate and twice the consideration received or receivable, directly or indirectly, by the entity or its associates in respect of the scheme; • seek an injunction to stop the promotion of a scheme or implementation of a scheme not in conformity to its product ruling; and • enter into voluntary undertakings with promoters or implementers about the way in which schemes are being promoted or implemented. In deciding what penalty is appropriate, the Federal Court can have regard to all matters it considers relevant, including the amount of loss or damage incurred by scheme participants and the honesty and deliberateness of the promoter’s conduct. The Explanatory Memorandum notes85 that the civil penalty regime is not intended to inhibit the provision of independent and objective tax advice, including advice regarding tax planning. Some commentators have, however, expressed serious reservations about the potential impact on tax advisers providing tax planning advice, as well as expressing concerns about many other aspects of the promoter penalty provisions.86

4.3

PRINCIPLES-BASED DRAFTING

Sir Ivor Richardson long ago suggested that traditional legislative drafting techniques – ‘where certainty and precision are sought through the detailed expression of policies in the variety of complex circumstances in which they will operate’ and ‘too often the intent is lost or blurred in a legislative fog’87 – may be inappropriate for tax law. There have been serious attempts in various jurisdictions since then to adopt a more principles-based approach to legislative drafting. In the UK in recent years this has been extended to an attempt to draft principles-based legislation to

84 85 86 87

Chapter 3 Explanatory Memorandum to Tax Laws Amendment (2006 Measures No. 1) Act 2006. Paragraph 3.50 Explanatory Memorandum to Tax Laws Amendment (2006 Measures No. 1) Act 2006. For example, J. King, ‘New Measures Deterring the Promotion of Tax Exploitation Schemes’ (2006) 35 Australian Tax Review, 163. I. Richardson, ‘Reducing Tax Avoidance by Changing Structures, Process and Drafting’ in G.S. Cooper (ed.), Tax Avoidance and the Rule of Law (Amsterdam: IBFD, 1997), p. 327, at p. 338.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

553

tackle avoidance activity. This was part of the simplification package announced by the UK Government in the Pre Budget Report in 2007.88 Late in 2007 HM Treasury and HMRC jointly released a consultation document89 which seeks to use principles-based legislation to tackle avoidance arising from or through disguised interest90 and sales of income streams arrangements.91 Principlesbased legislation strives to embody a principle of UK taxation, and it is proposed that the principle would be accompanied by a statement of how the legislation intends to operate by reference to that principle. Thus, for example, the principle dealing with disguised interest simply states: ‘A return designed to be economically equivalent to interest is to be taxed in the same way as interest’. It is argued that the introduction of such legislation will allow the repeal of significant amounts of existing legislation. Other claimed benefits include the potential for greater certainty, enhanced conceptual simplicity, improved coherence, reduced complexity and compliance costs, and the promotion of fairness and consistency. Against this, the consultation document considers that the disadvantages might be that a principles-based approach might cast a net so broad that unintended transactions might be caught up in it, and that it might be an inadequate or ineffective replacement of existing anti-avoidance provisions. The process of consultation is still on-going, and responses have been mixed. Changes were made to the initial draft legislation relating to disguised interest as a result of a workshop convened in January 2008, but the Government did not proceed, as intended, to introduce principles-based avoidance legislation in the 2008 Finance Bill. It is therefore a little early to assess the likely outcomes of this initiative, but it is clear that this more generic approach to anti-avoidance legislation has strong support in the UK within Government agencies. Other countries are watching with interest.

88 89 90

91

HM Treasury, Pre-Budget Report and Spending Review, London, 2007. HM Treasury and HMRC, Principles-based Approach to Financial Products Avoidance: A Consultation Document, London, December 2007. ‘Disguised interest schemes exploit differences in tax treatment between interest and other receipts such as dividends, and seek to convert taxable interest into an exempt dividend or capital gain. For example, a person subscribes for shares in a company that only has one asset, which increases in value in the same way as an investment with a guaranteed return. When the shares in the company are sold or redeemed, the return on them equates in substance to an amount from, say, a bank deposit.’ HM Treasury and HMRC, Principles-based Approach to Financial Products Avoidance: A Consultation Document, London, December 2007, at para. 2.1. ‘Selling an income stream is a device designed to try to turn economic income into a return that is treated by tax law as capital.’ HM Treasury and HMRC, Principles-based Approach to Financial Products Avoidance: A Consultation Document, London, December 2007, at para. 3.1.

554

TAX REFORM IN THE 21ST CENTURY

5 The Role of the Courts in Countering Avoidance If the legislatures of the various common law jurisdictions have been active in recent years in seeking to counter tax avoidance activity, their productivity has been more than matched by the judiciaries in those regimes. Cases relating to tax avoidance and tax planning have typically constituted one of the major areas of tax litigation in many of these regimes. This is not to suggest that the respective judiciaries have necessarily been prorevenue in their deliberations. Indeed, many of the major cases that have recently been heard have provided disappointing outcomes to revenue authorities. Nor is it to suggest that the deliberations of the courts have necessarily provided any greater clarity on the dividing line between what might be regarded as acceptable tax mitigation or unacceptable tax avoidance. The line remains as unclear as ever. But an examination of recent cases, particularly those heard in the superior courts in common law jurisdictions, does reveal some significant developments in the jurisprudence in this area, and in particular shows that there is now a greater certainty in the approaches that those courts are likely to take in the interpretation of the statutory provisions that exist in their respective jurisdictions. Such an examination also leads to the tentative conclusion that there is some level of convergence in the interpretative approaches in many of the common law jurisdictions, although by no means a consensus in how such cases should be approached. Moreover, that degree of convergence that has occurred has come about from very different starting positions and through very different routes. There has been an emphasis in recent UK superior court judgments in cases such as Arctic Systems,92 West v Trennery,93 Barclays Mercantile,94 Scottish Provident,95 McGuckian,96 and MacNiven97 that the role of the judges is to interpret the words of the statute, and to do so in a purposive fashion. As Tiley notes after considering some of those cases, ‘we are left with the simple fact that tax law is about interpreting statutes and that statutes should be interpreted purposively…and in their context’.98 This represents a clear departure from the view that had developed in the UK and elsewhere in the line of cases starting with Ramsay99 and developing through 92 93 94 95 96 97 98 99

Jones v. Garnett [2007] UKHL 35. [2005] STC 214. Barclays Mercantile Business Finance Ltd v. Mawson [2005] STC 1; [2004] UKHL 51. Scottish Provident Institution v. Inland Revenue Commissioners [2004] UKHL 52; [2004] 1 WLR 3172 (HL). IRC v. McGuckian [1997] 1 WLR 991. MacNiven v. Westmoreland Investments [2001] STC 237 (HL). J. Tiley, ‘Barclays and Scottish Provident: Avoidance and Highest Courts: Less Chaos but More Uncertainty’ [2005] (3) British Tax Review, 273, at p. 273. WT Ramsay Ltd v. Inland Revenue Commissioners [1982] AC 300.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

555

Furniss v Dawson.100 Those earlier cases had suggested that there was an extrastatutory anti-avoidance rule of law, or doctrine – the so-called ‘Ramsay doctrine’ – which was to be applied by the courts when considering avoidance issues. Effectively that doctrine asserted that where there was a preordained transaction or series of transactions which had steps inserted for no commercial purpose, those steps could be ignored and the relevant statutory provisions could be applied to the end result. Andrew Halkyard reminds us of some of the epithets applied to the Ramsay doctrine: ‘the doctrine of disregard’; the principle of fiscal nullity’; ‘a judge-made anti-avoidance weapon’ ‘a weapon of mass destruction’ and ‘a broad spectrum antibiotic which kills off all anti-avoidance schemes’.101 Lord Hoffmann, writing on tax avoidance after being involved in the Barclays Mercantile case, states: The primacy of the construction of the particular taxing provision and the illegitimacy of rules of general application has been reaffirmed by… [Barclays Mercantile]. Indeed, it may be said that this case has killed off the Ramsay doctrine as a special theory of revenue law and subsumed it within the general theory of the interpretation of statutes….102 This does not imply any reversion to a literal approach to statutory interpretation, which was part of the rationale for the emergence of the Ramsay doctrine in the first place. Instead a textual, contextual and purposive approach is likely to underpin judicial reasoning in the UK’s consideration of avoidance cases in the future. As Freedman notes: ‘The House of Lords has now confirmed that the essence of the ‘new approach’ to tax avoidance in the United Kingdom is that the court gives tax provisions a purposive construction in order to determine the nature of the transaction to which it is intended to apply, before going on to decide whether the actual transaction (which might involve considering the overall effect of a number of elements intended to operate together) answers that statutory description’.103 This is not to suggest that the current UK position is stable, or that it provides the certainty of outcome that advisers crave. But, it has removed some of the old chaos104 and at least suggests that it is able to provide certainty of approach, if not of outcome.

100 Furniss v. Dawson [1984] AC 474. 101 A. Halkyard in ‘Not a Weapon of Mass Destruction: Can the Ramsay Approach Apply to the

Inland Revenue Ordinance in Hong Kong?’ (2005) 9(3) Asia-Pacific Journal of Taxation, 56, at 57. 102 L. Hoffmann, ‘Tax Avoidance’ [2005] (2) British Tax Review, 197, at 203. 103 J. Freedman, ‘Converging Tracks? Recent Developments in Canadian and UK Approaches to Tax Avoidance’ (2005) 53(4) Canadian Tax Journal, 1038, at 1040–1041. 104 J. Tiley, ‘Barclays and Scottish Provident: Avoidance and Highest Courts: Less Chaos but More Uncertainty’, [2005] 3 British Tax Review, 273 at 274.

556

TAX REFORM IN THE 21ST CENTURY

The Ramsay doctrine held particular sway in the UK in part because the UK does not have the ‘provision of last resort’ that is represented by a statutory general anti-avoidance rule. In that respect it is not dissimilar to the US where a host of judicial doctrines, including the business purpose test, the step transaction doctrine, the sham doctrine and the economic substance doctrine, have prevailed in cases relating to tax avoidance. But although the UK shares with the US the lack of a general anti-avoidance rule, it does not share – as Barclays Mercantile has made absolutely clear – any tradition in which judicial anti-avoidance doctrines make good statutory deficiencies. There is, however, also evidence of a greater consistency in the interpretative approach to tax avoidance cases taken by the superior courts in countries which do have statutory general anti-avoidance rules. In comparing the outcomes of the two recent anti-avoidance cases to have been heard by the Canadian Supreme Court (Canada Trustco105 and Mathew106) with the UK outcomes in Barclays Mercantile and Scottish Provident, Freedman notes the ‘similarities between the positions reached in the two countries via different routes. In each country the intention of the particular statute concerned, as revealed by its wording construed in context, is paramount; in each jurisdiction the fact that a transaction is motivated by tax saving is not, on its own, fatal to its effectiveness for tax minimization purposes’.107 In cases such as these, both Canada and the UK have also affirmed that there is no place for a business purpose test – along the lines of the doctrine that has prevailed in the US since its seminal case of Helvering v Gregory108 in the 1930s. The emerging jurisprudence on Australia’s general anti-avoidance rule suggests that the courts, in Australia as well as the UK and Canada, must consider the words of the statute interpreted in a purposive fashion, as required by the provisions of the Australian Acts Interpretation Act. The provisions of Part IVA are quite prescriptive, and if interpreted literally could annihilate virtually all tax planning transactions. Part IVA involves a consideration of three basic requirements. Firstly, there must be a ‘scheme’. This is so broadly defined as to encompass virtually any act or transaction, although the judgment by the Federal Court in the 2007 Star City case109 has suggested that the existence of a scheme cannot always be taken for granted. The second is that there must be a ‘tax benefit’, again broadly defined. The final element, which is the area of greatest contention, is that the scheme must have been entered into for the sole or dominant purpose of obtaining a tax benefit. This requires an objective assessment, based upon analysis of eight factors, including 105 The Queen v. Canada Trustco Mortgage Co. [2005] SCC 54. 106 Mathew v. The Queen [2005] SCC 55. 107 J. Freedman, ‘Converging Tracks? Recent Developments in Canadian and UK Approaches to

Tax Avoidance’ (2005) 53(4) Canadian Tax Journal 1038, at 1039. 108 69 F 2nd 809 (1934). 109 Star City Pty Ltd v. FCT [2007] FCA 1701.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

557

the manner in which the scheme was implemented, its form and substance, and its effect. There is some evidence of a tension between the lower and higher courts in Australia over the interpretation of Part IVA. The Federal and Full Federal Courts have shown themselves willing to adopt a more commercial approach to the interpretation of the provisions. This is not to suggest that a ‘business purpose’ test has emerged in those courts, but they are certainly more likely to find in favour of the taxpayer where – on an objective examination of the circumstances surrounding the transaction – they are able to establish a commercial motive for the transaction. In its deliberations, in contrast, the High Court has repeatedly insisted that this is not the appropriate approach, and that the legislation must be construed purposively with a view to establishing the appropriate outcome. Upon occasions this approach has favoured the revenue (Spotless, Hart) and upon others the taxpayers have succeeded (Peabody). Recent judicial decisions in Hong Kong and New Zealand also confirm the purposive approach to the interpretation of taxation statute in common law jurisdictions. For example, the Arrowtown110 case in Hong Kong reasserted ‘the need to apply orthodox methods of purposive interpretation to the facts viewed realistically’.111 The facts in the Peterson112 case in New Zealand, heard by the Privy Council in 2005, bear some resemblance to those in Barclays Mercantile and Canada Trustco, save that the object of investment was films rather than pipelines or trailers. The outcome was also similar, in that the taxpayer was successful in claiming depreciation allowances (the claim was for capital allowances in Barclays Mercantile and capital cost allowance deductions in Canada Trustco). This tends to confirm the view that there is some degree of convergence in the jurisprudence in some of the common law jurisdictions in the approach taken by the courts to avoidance type cases, albeit through the interpretation of very different legislation. The Canadian and New Zealand general anti-avoidance rules have some similarities, but are – in substance – quite different from each other. And both are very different from the specific legislative provisions that were in play in Barclays Mercantile. The Peterson case also illustrates very clearly the absolute difficulty of determining, with any degree of certainty, where the borderline lies between what Lord Millett (who delivered the majority decision in favour of the taxpayer) deemed an ‘acceptable tax advantage’ (for which read tax mitigation) as opposed to an ‘unacceptable tax advantage’ (for which read tax avoidance).113 The Privy 110 Collector of Stamp Revenue v. Arrowtown Assets Ltd [2004] 1 HKLRD 77. 111 A. Halkyard, ‘Not a Weapon of Mass Destruction: Can the Ramsay Approach Apply to the

Inland Revenue Ordinance in Hong Kong?’ (2005) 9(3) Asia-Pacific Journal of Taxation 56. 112 Petersen v. CIR [2005] STC 448. 113 Petersen v. CIR [2005] STC 448, at paras 35−37.

558

TAX REFORM IN THE 21ST CENTURY

Council split three to two in favour of the taxpayer, and it is interesting to note that both the majority judgment (Lord Millett, Baroness Hale and Lord Brown) and the minority judgment (Lord Bingham and Lord Scott) made reference to the fact that this was not a borderline case. The majority regarded it as clear that the general anti-avoidance rule was not applicable; by the same token the minority noted, in their judgment, that ‘a clearer case [for the application of the New Zealand general anti-avoidance rule] can hardly be imagined’.114 Drawing the threads together, it can be concluded that the courts – and particularly the superior courts – in all of the common law jurisdictions examined have reached the position of accepting that taxing statutes, just like other law, are to be consistently interpreted in a purposive fashion, having regard not only to the words of the legislation but also to the intended legislative effect. This convergence exists notwithstanding different starting points and different routes. Above all else though, the inevitable conclusion is reached that despite the large number of cases that have been heard in recent years and despite a greater certainty of interpretive approach by the courts, there is still no certainty of outcome from the courts. The dividing line between acceptable tax mitigation and unacceptable tax avoidance remains as indistinct as ever. As has been noted by Michael Littlewood in writing about decisions of the Privy Council in this area, the line is one of the most difficult in the whole of the law. ‘All in all, to describe the distinction between avoidance and mitigation as ‘vague’ is to understate the problem, for it suggests that there is general agreement as to roughly where the line lies and that the disagreement is only as to marginal cases. But none of their Lordships appear to have regarded any of the cases as marginal. It is difficult, therefore, to extract from them any guidance as to where the line lies’.115 Perhaps it has to be concluded that the courts will never be in a position to provide certainty of outcome, but that consistency of approach is at least a step in the right direction.

6

Conclusions

It may be too cynical to assume ‘the existence of tax avoidance as a constant and perpetual motivation for every taxpayer’,116 but there is no doubt that tax avoidance is widespread and that it presents a major problem for those concerned with public finance issues. There is some evidence that the aggressive retail marketing of tax avoidance products and schemes may have been constrained in recent years, but 114 Petersen v. CIR [2005] STC 448, at para. 96. 115 M. Littlewood, ‘The Privy Council and the Australasian Anti-Avoidance Rules’, [2007]

(2) British Tax Review 175. 116 C.H. Gustafson, ‘The Politics and Practicalities of Checking Tax Avoidance in the United

States’ in G.S. Cooper (ed.), Tax Avoidance and the Rule of Law (Amsterdam: IBFD, 1997), p. 349, at p. 376.

CHRIS EVANS: CONTAINING TAX AVOIDANCE

559

avoidance activity is by its nature opportunistic and ad hoc. Simply raising the price of avoidance (through successful containment, increased regulation and constrained supply) will not choke off demand. Indeed, no single response or approach – whether administrative, legislative or judicial – can adequately or effectively contain avoidance activity. Such containment only begins to occur where strategies drawn from all three spheres complement each other by operating in combination. As Sir Ivor Richardson astutely pointed out some years ago, current requirements for a comprehensive and integrated approach go beyond a more traditional analysis where ‘the legislature…exerts control of tax avoidance through special and general anti-avoidance provisions; the revenue administration contributes in administering those provisions and exercising discretions; and the judiciary is expected to strike the right balance between acceptable and unacceptable tax planning through its interpretation and application of tax legislation.’117 Ultimately, however, corporate and personal taxpayers themselves have to take responsibility for the level of avoidance and the degree of acceptance of such behaviour that exists at any time in any society. The revenue authority, the legislature and the judiciary can play a role in shaping the demand for, and supply of, tax avoidance activity, but such issues belong, in the final analysis, in the realms of moral and ethical behaviour of the taxpayers themselves. Corporate and personal social responsibility – and the reputational damage that excessive and egregious avoidance activity can attract – remains the ultimate deterrent, notwithstanding the impressive arsenal that can be available to those who seek to counter avoidance. Beyond that we should also perhaps be mindful that two of the traditional goals of public finance – simplicity and equity – have critical roles to play in determining social responses to avoidance activity. In recent years these two goals may have been less prominent in tax reform than the efficiency goal that lends itself to easier economic measurement and evaluation. It is paradoxical that the more complex that the tax regime becomes (often in attempts to contain avoidance activity), the more likely it will be that opportunities for avoidance will arise. Avoidance activity thrives in complexity and uncertainty. And where that complexity exacerbates the natural interaction (sometimes mediated by intermediaries) between the taxpayer and the revenue authority such that it becomes frictional rather than cooperative, there will almost inevitably be a higher probability of avoidance activity. Equity also carries with it the important message that tax systems must not only be fair – they must also be perceived to be fair. If vast swathes of the population are not convinced that the tax system does operate fairly (whether it does or not), or 117 I. Richardson, ‘Reducing Tax Avoidance by Changing Structures, Process and Drafting’ in

G.S. Cooper (ed.), Tax Avoidance and the Rule of Law (Amsterdam: IBFD, 1997), p. 327, at p. 327.

560

TAX REFORM IN THE 21ST CENTURY

that sectional interests hold undue sway or receive inappropriate favours (whether they do or not), they are themselves more likely to engage in the sorts of nefarious activities that they condemn in others.

SERIES ON INTERNATIONAL TAXATION 1. Alberto Xavier, The Taxation of Foreign Investment in Brazil, 1980 (ISBN 90-200-0582-0) 2. Hugh J. Ault and Albert J. Ra¨dler, The German Corporation Tax Law with 1980 Amendments, 1981 (ISBN 90-200-0642-8) 3. Paul R. McDaniel and Hugh J. Ault, Introduction to United States International Taxation, 1981 (ISBN 90-6544-004-6) 4. Albert J. Ra¨dler, German Transfer Pricing/Prix de Transfer en Allemagne, 1984 (ISBN 90-6544-143-3) 5. Paul R. McDaniel and Stanley S. Surrey, International Aspects of Tax Expenditures: A Comparative Study, 1985 (ISBN 90-654-4163-8) 6. Kees van Raad, Nondiscrimination in International Tax Law, 1986 (ISBN 906544-266-9) 7. Sijbren Cnossen (ed.), Tax Coordination in the European Community, 1987 (ISBN 90-6544-272-3) 8. Ben Terra, Sales Taxation. The Case of Value Added Tax in the European Community, 1989 (ISBN 90-6544-381-9) 9. Rutsel S.J. Martha, The Jurisdiction to Tax in International Law: Theory and Practice of Legislative Fiscal Jurisdiction, 1989 (ISBN 90-654-4416-5) 10. Paul R. McDaniel and Hugh J. Ault, Introduction to United States International Taxation (3rd revised edition), 1989 (ISBN 90-6544-423-8) 11. Manuel Pires, International Juridicial Double Taxation of Income, 1989 (ISBN 90-6544-426-2) 12. A.H.M. Daniels, Issues in International Partnership Taxation, 1991 (ISBN 90-654-4577-3) 13. Arvid A. Skaar, Permanent Establishment: Erosion of a Tax Treaty Principle, 1992 (ISBN 90-6544-594-3) 14. Cyrille David and Geerten Michielse (eds), Tax Treatment of Financial Instruments, 1996 (ISBN 90-654-4666-4) 15. Herbert H. Alpert and Kees van Raad (eds), Essays on International Taxation, 1993 (ISBN 90-654-4781-4) 16. Wolfgang Gassner, Michael Lang and Eduard Lechner (eds), Tax Treaties and EC Law, 1997 (ISBN 90-411-0680-4) 17. Glo´ria Teixeira, Taxing Corporate Profits in the EU, 1997 (ISBN 90-4110703-7) 18. Michael Lang et al. (eds), Multilateral Tax Treaties, 1998 (ISBN 90-4110704-5) 19. Stef van Weeghel, The Improper Use of Tax Treaties, 1998 (ISBN 90-4110737-1) 20. Klaus Vogel (ed.), Interpretation of Tax Law and Treaties and Transfer Pricing in Japan and Germany, 1998 (ISBN 90-411-9655-2) 21. Bertil Wiman (ed.), International Studies in Taxation: Law and Economics; Liber Amicorum Leif Mute´n, 1999 (ISBN 90-411-9692-7)

22. Alfonso J. Martı´n Jime´nez, Towards Corporate Tax Harmonization in the European Community, 1999 (ISBN 90-411-9690-0) 23. Ramon J. Jeffery, The Impact of State Sovereignty on Global Trade and International Taxation, 1999 (ISBN 90-411-9703-6) 24. A.J. Easson, Taxation of Foreign Direct Investment, 1999 (ISBN 90-4119741-9) 25. Marjaana Helminen, The Dividend Concept in International Tax Law: Dividend Payments Between Corporate Entities, 1999 (ISBN 90-411-9765-6) 26. Paul Kirchhof, Moris Lehner, Kees van Raad, Arndt Raupach and MichaelRodi (eds), International and Comparative Taxation: Essays in Honour of Klaus Vogel, 2002 (ISBN 90-411-9841-5) 27. Krister Andersson, Peter Melz and Christer Silfverberg (eds), Liber Amicorum Sven-Olof Lodin, 2001 (ISBN 90-411-9850-4) 28. Juan Martı´n Jovanovich, Customs Valuation and Transfer Pricing: Is It Possible to Harmonize Customs and Tax Rules?, 2002 (ISBN 90-411-9888-1) 29. Stefano Simontacchi, Taxation of Capital Gains under the OECD Model Convention: With Special Regard to Immovable Property, 2007 (ISBN 97890-411-2549-1) 30. Michael Lang, Josef Schuch, and Claus Staringer (eds), Tax Treaty Law and EC Law, 2007 (ISBN 978-90-411-2629-0) 31. Duncan Bentley, Taxpayers’ Rights: Theory Origin and Implementation, 2007 (ISBN 978-90-411-2650-4) 32. Sergio Andre´ Rocha, Interpretation of Double Taxation Conventions: General Theory and Brazilian Perspective, 2008 (ISBN 978-90-411-2822-5) 33. Robert F. van Brederode, Systems of General Sales Taxation: Theory, Policy and Practice, 2009 (ISBN 978-90-411-2832-4) 34. John G. Head and Richard Krever (eds), Tax Reform in the 21st Century: A Volume in Memory of Richard Musgrave, 2009 (ISBN 978-90-411-2829-4)

Tax Reform in the 21st Century A Volume in Memory of Richard Musgrave John G. Head and Richard Krever (eds.) No government can be sustained without the ability to tax its citizens. The question then arises how can a nation do so in a way that’s fair and equitable to taxpayers without distorting incentives or the allocation of resources, while simultaneously promoting economic growth and providing the state with the funds it requires to adequately address the needs of its citizens? This insightful work, featuring contributions from a stellar array of international tax experts and economists, addresses the crucial issues which developed countries will confront in the early decades of the twenty-first century: - - - - - -

The pursuit of tax reform. Personal tax base: income or consumption? Tax rate scale: equity and efficiency aspects. Business tax reform: structural and design issues. Interjurisdictional issues. Controlling tax avoidance.

ISBN 978-90-411-2829-4





SOIT 34