Public economics in an age of austerity (the graz schumpeter lectures)

Tables 1.1 Optimal indirect tax structure under different assumptions about 4.1 Distribution of income among world citizens 1992 79 Figures 1.1 Gross disposable real income of households

www.ebook3000.com

Public Economics in an Age of Austerity

Governments all round the world are facing problems with their publicfinances At a time of austerity, how much should spending be cut and howmuch should taxes be raised? Does the national debt represent a burden forfuture generations? Should taxes on the rich be raised? This book examineshow the tools of public economics can be applied to answer such key questionsand to suggest alternatives to the austerity policies currently being pursued.The fiscal problems faced are not simply the result of the post-2008economic crisis but reflect a deep-seated fault line in modern economies.There has to be fiscal consolidation to provide for an ageing population,increased investment in education, and climate change The book describeshow public economics can help us think about alternative ways of meetingthis challenge It casts doubt on conventionally held views, such as thoseconcerned with top tax rates, the undesirability of taxing capital income, thetargeting of child benefits, and the merging of income tax and social securitycontributions Thefinal part goes beyond national boundaries and considersglobal public economics, focusing on the pressing problem of financingdevelopment

The conclusion of the book is that there are significant choices to be made.Not all austerity packages are the same: there are alternatives It would bepossible to raise taxes more and to cut spending less It is important to con-sider the full range of possible policies In considering these alternatives,modern public economics provides a useful framework, but it has major lim-itations Economists are too often prisoners within the theoretical walls theyhave erected and fail to see that important considerations are missing.Economists have paid too little attention to the ethical basis underlying theirpolicy recommendations

A B Atkinson is a Fellow of Nuffield College, of which he was Warden from

1994 to 2005 and is currently Centennial Professor at the London School ofEconomics He was knighted on 2001 for services to economics, and is aChevalier de la Légion d’Honneur

www.ebook3000.com

The Graz Schumpeter Lectures

Previous titles in the series:

1 Evolutionary Economics and Creative Destruction

J Stanley Metcalfe

2 Knowledge, Institutions and Evolution in Economics

Brian J Loasby

3 Schumpeter and the Endogeneity of Technology

Some American Perspectives

Nathan Rosenberg

4 Consumption Takes Time

Implications for Economic Theory

Ian Steedman

5 Exchange Rates and International Finance Markets

An Asset-Theoretic Approach with Schumpeterian PerspectiveErich W Streissler

8 The Dynamics of Industrial Capitalism

Schumpeter, Chandler, and the New Economy

Richard N Langlois

9 Growth, Distribution and Innovations

Understanding Their Interrelations

Amit Bhaduri

www.ebook3000.com

This page intentionally left blank

www.ebook3000.com

Public Economics in an Age of Austerity

A B Atkinson

www.ebook3000.com

First published 2014

by Routledge

2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN

and by Routledge

711 Third Avenue, New York, NY 10017

Routledge is an imprint of the Taylor & Francis Group, an informa business

© 2014 A B Atkinson

The right of A B Atkinson to be identi fied as author of this work has been asserted by him in accordance with the Copyright, Designs and Patent Act 1988.

All rights reserved No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers.

Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identi fication and explanation without intent to infringe.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data

Public economics in an age of austerity / A B Atkinson.

pages cm – (The Graz Schumpeter lectures)

1 Finance, Public 2 Taxation 3 Social policy I Title.

Typeset in Times New Roman

by Taylor & Francis Books

www.ebook3000.com

www.ebook3000.com

Tables

1.1 Optimal indirect tax structure under different assumptions about

4.1 Distribution of income among world citizens 1992 79

Figures

1.1 Gross disposable real income of households from 1999 to 2012,

2.1 Shares of the top 1 per cent in the United Kingdom 182.2 The evolution of the proportion‘rich’ in the United States since 1960 21

2.4 Transmitted wealth in the UK as percentage of total personal

2.5 Ratio of personal wealth to personal income in the UK 35

4.5 Alternative forms for the social marginal valuation of income 81

Boxes

4.2 The optimal provision of official development assistance (ODA) 63

www.ebook3000.com

I am most grateful for the invitation to present the Graz Schumpeter Lectures

in October 2012 There are at least three reasons for being grateful Thefirst

is that it gave me the opportunity to visit Graz, where my wife and I weremost warmly welcomed We should like to thank Richard Sturn, Heinz Kurz,Christian Gehrke, their families and colleagues for making our stay soenjoyable and interesting Our thanks too to Johanna Pfeifer for making suchefficient arrangements The second reason is that the preparation of the lec-tures gave me grounds for collecting my thoughts about both current austeritypolicies and the underlying state of public economics As noted below, I havedrawn on a number of earlier pieces of work, but the four lecture series chal-lenged me to bring them together It also led to my re-reading, or reading forthe first time, the writings of Joseph Schumpeter, many of which are highlyrelevant to my subject, notably his remarkable 1918 essay on ‘Die Krise desSteuerstaates’ (‘The crisis of the tax state’), republished in translation inSchumpeter (1991) The third reason for gratitude is that the lecture attracted

a large student audience, who posed excellent questions, both inside andoutside the lecture room

The Lectures build on a review article I have written on the work of the

UK team chaired by Sir James Mirrlees, published in the Journal of EconomicLiterature (Atkinson, 2012) The Lectures draw also on material presented inthree earlier public lectures In September 2011, I gave thefirst Amartya SenLecture, with the title ‘Public economics after The idea of justice’, at theannual conference of the Human Development and Capability Association inthe Hague The Sen Lecture has been published in the Journal of HumanDevelopment and Capabilities (Atkinson, 2012a) In January 2012, I gave thefifth Sandmo Lecture, with the title ‘Public economics in an age of austerity’,

at the Norwegian School of Economics and Business Administration inBergen Chapter 4 draws substantially on the Fourth Jelle Zijlstra Lecture(Atkinson, 2005) given at the Free University, Amsterdam, on 12 December

2005, when I was a visiting fellow at the Netherlands Institute for AdvancedStudies

Empirical aspects of the research reported in the Lectures draw on workcarried out in conjunction with a number of colleagues, whom I would like to

www.ebook3000.com

thank most warmly I have taken distributional data from the Chartbook ofEconomic Inequality, available from the website of INET at the OxfordMartin School, prepared together with Salvatore Morelli I have cited results

on the joint distribution of earned and capital income from a project, initiatedwith Emmanuel Saez, and now in progress with Rolf Aaberge of StatisticsNorway, Sebastian Königs, and Christoph Lakner Thefindings on long-runtrends in inheritance in the UK are part of a multi-country project withFacundo Alvaredo and Thomas Piketty, with whom I have also had thepleasure of working on the World Top Incomes Database and associatedvolumes In Chapter 4, I have drawn extensively on joint work with AndreaBrandolini, whose influence indeed runs through the book

Going further back in time, my thinking on public economics has beenmuch inspired by a long period of collaboration with François Bourguignon,who will recognize many of the issues with which we have grappled, and byfruitful discussions over the years with Agnar Sandmo The foundations formuch of my analysis were laid down in the twelve-year research programmefrom 1978 to 1989, Taxation, Incentives and the Distribution of Income,directed jointly with Mervyn King and Nick Stern This was an intellectuallyexciting time, and, although our weekly breakfasts are long in the past, I owethem a great debt A decade before that, I could not have had a better intro-duction to modern public economics than that provided by Peter Diamondand Jim Mirrlees, and subsequently writing a graduate textbook with JoeStiglitz was an unforgettable experience

This book is based on four lectures and covers only a selection of issues inpublic economics I am very conscious that I have failed to survey manyimportant and interesting topics I should therefore refer the reader to themuch more exhaustive treatments provided in the Background Papers for theMirrlees Review (Mirrlees, 2010) and in the excellent survey of the field byRobin Boadway (2012)

Preparing these lectures has led me to reflect on my approach to publiceconomics and to economics in general As a student in Cambridge, England,

in the early 1960s, I was taught by a number of people who held strong views,views that they expressed with great fervour and a combative style Perhaps as areaction, perhaps as a matter of personal temperament, I have alwaysregarded my views about economics– in contrast to my moral principles – astentative and open to revision as I learned more and acquired more evidence

In this process, criticism is to be listened to, not simply combated or missed Perspectives different from one’s own can yield new insights or revealaspects that need to be re-addressed I stress the need for open-mindedness,since economics today seems to have become less open-minded The needs ofmodern media communication mean that it is dominated by those that have astrong message to deliver In academic life there is increasing sub-division intospecialist groups with heavy emphasis on an agreed set of positions andapproach to economics This makes me uneasy

dis-x Preface

At the same time, I remain optimistic about the future of the subject ofeconomics There are many who share my concerns, and I have had thepleasure of working with a great set of co-authors In recent years, I have hadwonderful students and post-doctoral fellows To all of them this book isdedicated.

Tony AtkinsonSeptember 2013Preface xi

1 Public economics and austerity

The subject matter of my lectures is public economics I hope that this is anappropriate topic for the Graz Schumpeter Lectures, since Joseph Schumpeterwas both a professor of publicfinance and – rather briefly – an active policy-maker, as Austrian Minister of Finance in 1919 Wikipedia, with studiedambiguity, describes him as serving ‘with some success’ Schumpeter himselfsaid that ‘it was no pleasure’ (Swedberg, 1991, page 60) There can be nodoubt that it was a difficult task, and would have defeated even a person withgreater diplomatic skills

Today too Finance Ministers face a difficult task For this reason, I havechosen to focus on public economics in an age of austerity This is certainlytopical In 2010, Merriam-Webster’s Dictionary named the word ‘austerity’ asits‘Word of the Year’ based on the number of web searches this word generated.All around the world, there are governments adopting or are being required toadopt‘austerity programmes’ But these programmes are highly controversial.They divide politicians, they divide countries, and they divide governmentsfrom many of their citizens, as shown by riots and demonstrations in the street

In focusing on austerity programmes and longer-term fiscal problems, I

am– in the spirit of Schumpeter – concerned both with the substantive issueand with the state of public economics I shall ask – what are the lessons ofpublic economics for the debate about austerity programmes? But also I shallask – what are the strengths and weaknesses of modern public economics?The interplay between academic research and public policy is of centralimportance to the discipline of public economics In preparing this lecture, Ilooked back at thefirst issue of Journal of Public Economics that was foundedsome forty years ago I was struck by the extent to which the authors in thatfirst issue have consistently been concerned with the application of their ana-lysis to the pressing problems of current policy The authors included onefuture chair of the US Council of Economic Advisers and one future Gover-nor of the Bank of England Since then, there has been much progress in ourunderstanding At the same time, there remains a gulf – a gulf that isparticularly apparent in recent debates about austerity programmes

In these debates, one hears a great deal from macro-economists and fromeconomists specializing in financial markets Yet one hears relatively littlefrom experts in public economics This seems to me surprising Is not publiceconomics concerned with deficits and taxes and government spending? Is notthe national debt one of the essential topics in any course on public finance?Should not students therefore be taking courses in public economics? To takejust some examples, cannot public economics help us think about the followingquestions?

The balance between tax increases and spending cuts is typically of theorder of 25%/75% Are these the right proportions?

Tax rises in a number of countries take the form of raising VAT rates.Should we also consider broadening the VAT tax base?

Past fiscal consolidations have fallen disproportionately on cuttinggovernment capital investment Does this simply pass on the burden to thenext generations in a different form?

Should we cut back on personal services for the elderly or on youth services?

Should we raise taxes on capital?

And there are many more

These lectures are concerned with the contribution of public economics tothe kind of issues just listed Thefirst chaptercovers thefield more generally,and provides an introduction to the key points The next three chapters gointo greater depth, taking a number of case studies The first (Chapter 2)treats a much debated topic How should we tax the rich? What should be thetop tax rate? Should we tax inherited wealth? The title ofChapter 3(‘Modelscan become prisons’) may appear a little mysterious When I explain that itconcerns the way in which the recommendations by economists can be driven

as much by their modelling as by economic reality, then it may appear ratherdry However, I explore the role of economic models in the context ofimportant real-world issues: taxing capital income, child benefits, and themerging of income tax and social security contributions These policy issuesare discussed in Chapters 1 to 3 in a largely national perspective, but theyhave an important global dimension, and Chapter 4 sets out an agenda forglobal public economics Why for example do national governments maketransfers to other countries, notably in the form of official developmentassistance? Should we have international co-operation in taxation and globaltaxes? Should there be a World Tax Authority?

As I have stressed, I am concerned with the application of public economics

to policy-making, and for this reason I shall be referring at a number ofjunctures to the recent major review of taxation in the United Kingdom (UK)carried out by a team chaired by Sir James Mirrlees This was not an officialbody but one established by the Institute for Fiscal Studies (It was a succes-sor review to that carried out by the Meade Committee in the 1970s (Meade,1978).) The report of the Mirrlees Review (Mirrlees, 2011), Tax by design,

2 A B Atkinson

took a long-term view of ‘a good tax system for the 21st century’ It wasaccompanied by a set of background papers, Dimensions of tax design, pub-lished in 2010 The 1,880 pages, written by sixty-three authors from manycountries, represent a major contribution and a rich source of new ideas forpolicy-makers around the world It also provides a convenient vehicle toassess the current state of public economics.

I referred just now to macro-economists, and I begin with the relationbetween public economics and macro-economics A theme running throughthis book is my view that economists have become over-specialized In parti-cular, public economics has become too divorced from other fields, notablymacro-economics This is a loss to both sides, as was gracefully recognizedsome years ago by Robert Lucas:‘as a practising macroeconomist, I must saythat I have greatly enjoyed this excursion into public finance … Howrefreshing it is to spend some time in the company of a group of appliedeconomists who simply take for granted the desirability of using (andextending) the powerful methods of dynamic general equilibrium theory togain a deeper understanding of policy issues’ (1990, page 314)

The divorce between public economics and macro-economics is relativelyrecent In the 1950s, Richard Musgrave devoted a third of his celebrated Thetheory of publicfinance to stabilization policy As he says, ‘an effort is made toview the stabilization function in the broader context of efficient budgetpolicy This policy must account for the provision of social services and dis-tributional adjustments as well as stabilization’ (1959, page viii) Even ifMusgrave conducted his analysis of the public household in terms of thecelebrated three branches (Allocation, Distribution, and Stabilization), he sawthat they needed to be carefully co-ordinated The tools considered by Mus-grave included automatic stabilizers, whose role has been under-played inrecent analyses of the economic crisis One reason for the under-emphasis isthat we remain fixated on what is happening to Gross Domestic Product(GDP), rather than consider the macro-economic variable of greater concern

to most people– what is happening to Household Disposable Income (HDI).Every quarter we are bombarded with statements as to whether the economy

is growing or in recession But these statements are almost entirely focused onwhat is happening to GDP But what does GDP mean to our citizens? It is ameasure of economic activity, but not one that is directly related to theirliving standards

For this reason, it seems to me better to headline the impact of economicperformance on household living standards And this shows a different story

If we compare the evolution of GDP and HDI over the period from 1999 to

2012 in the seventeen-country Euro zone, then we see (Figure 1.1) that in thefirst part of the period, up to the onset of the crisis in 2007, household dis-posable income rose less than GDP Both are expressed in real terms: that is,

Public economics and austerity 3

adjusted for the rise that has taken place in prices Then, when GDP fellsharply in 2008 and 2009 – a fall of 5.7 per cent – household disposableincomes were broadly maintained, at least until the end of 2010 The auto-matic stabilizers and the discretionary stimulus measures worked This is aremarkable, and little heralded, success I find it surprising that the politicalleaders at the time have not made more of this Less surprising is the focus bypoliticians on GDP as a measure of recovery, but they have to recognize thatthis will not be perceived by their electorates until this feeds intoimprovements in household disposable incomes.

The conclusion I draw is that we should not lose confidence in the

effectiveness of short-term stabilization policy We need to retain automaticstabilizers and to use discretionary policy As argued by Peter Diamond,

in today’s economic environment of high unemployment in many tries and high vulnerability to the risk of significant further shocks,shocks that have sizable probabilities,fiscal policy has a far more impor-tant role to fill than in the earlier post-war US recessions Indeed, thefailure to take further fiscal steps to help the recovery is viewed as amajor political shortcoming in the eyes of many analysts, including me

coun-(Diamond, 2012, page 2)

Figure 1.1 Gross disposable real income of households from 1999 to 2012, compared

with GDP, Euro area (17)

Source: Eurostat website, series namq_gdp and sector accounts

I fully share his view, but I will not say more about this Rather, I want tolook ahead to the medium- and longer-termfiscal problem While the title of

my lectures refers to ‘austerity’, I believe that the fiscal problems that weare facing are not simply the consequence of the crisis, but reflect moredeep-seated fault lines in modern economies Indeed, Figure 1.1 contains awarning in this regard Over the longer-term, household disposable incomecannot be expected to grow in line with GDP The annual growth ratefrom 1999 to 2007 was 2.5 per cent for GDP, but only 1.9 per cent forhousehold disposable income In the future, the growth rate of household dis-posable income may well be lower We may indeed beflat-lining, particularlywhen account is taken of population growth, not allowed for in the chart

In emphasizing the longer-term fiscal problem, rather than the immediatecrisis, I am following in the footsteps of Joseph Schumpeter in his essay‘DieKrise des Steuerstaates’ (‘The crisis of the tax state’) Writing at the end of theFirst World War, he argued that it was not the case that ‘an otherwise per-fectly healthy tax state had suddenly become impossible owing to the worldwar and its aftermath’ (1991, page 101) He goes on to refer to ‘a much morebasic inadequacy of the particular society whosefiscal expression the tax state is’(page 101) What is the source of this more basic inadequacy? As I haveargued, household disposable incomes cannot be expected to continue to rise

in line with GDP This is the case for well-known reasons In part this arisesfor life-cycle reasons – at both ends of the age spectrum Resources arerequired to meet the ageing of the population, where, even with longerworking lives, there is going to be a larger dependent population Resourcesare essential for investment in education for young people At the same time,there is the need for expenditure in facing the challenge of climate change andenvironmental degradation – either now or soon And, perhaps less widelyrecognized, we need to build up social net worth In recent decades, the state,

at least in the UK, hasfinanced its tax cuts by running down the net worth ofthe state This can no longer be continued, and indeed I believe that the stateneeds once more to establish a positive net worth position

All of these claims on our resources need to be taken into account indesigning austerity programmes – and medium-term fiscal policy How doespublic economics help us think about these issues? What are the lessons thatmay be drawn? Can one indeed learn anything? Some people– like the Spanishfootball manager, Vicente del Bosque, do not believe that economists haveanything useful to say According to him,‘no one knows anything about eco-nomics It’s the great lie of the economists By contrast in football people mighthave contrasting opinions, each of which has some validity But the economistsalways speak in conditionals– what a mess’ (interview in the Guardian, 8 Sep-tember 2012) I hesitate to take issue with a celebrated football manager, but Ibelieve that he is wrong with regard to economics As I have already explained

in the Preface, I believe in a more modest role for economists Conditionals areprecisely what economists should be offering; they should not be makingdecisions but spelling out the alternatives And there are alternatives

Public economics and austerity 5

1.3 There are alternatives

The first lesson is that there are choices to be made Austerity programmesare often presented as though there is no substitute Mrs Thatcher in Britainfamously used the phrase ‘there is no alternative’, and this has been muchechoed in recent years by politicians in many countries But the function ofeconomists, in my view, is to make clear that there are alternatives

Of course, our choices are limited The government is constrained by itsbudget constraint Revenues have to cover the cost of government spendingand the excess of debt interest over the increase in the debt On account of theeconomic crisis, the choices have become more limited The automatic stabi-lization has come at a price in terms of the government budget, leading tolower revenues and to increased transfer payments The rescue of banks wasfinanced by an increase in public debt Slower growth rate meant that lessnew debt could be issued without raising the debt to GDP ratio The moststraightforward argument for austerity programmes is therefore that the gov-ernment’s budgetary position has worsened This is an age-old argument Inwhat used to be the standard UK public finance textbook, Pigou said that,after a war,‘certain government expenditures, which it used to be worthwhile

to undertake, a country may no longer be able to“afford”’ (1947, page 32).This is a genuine‘austerity’ argument for cuts in public spending

What though does austerity imply? The most obvious answer is that wehave to reverse some of the increases that took place as the economy grew inthe past The revenue constraint becoming tighter implies a combination oftax rises (reducing private spending) and government spending cuts ‘Afford-ability’ would imply returning along the previous expansion path, as illu-strated inFigure 1.2by the dashed arrow This would involve something likethe 45/55 ratio shown in thefigure, where 45 per cent of the cut would fall ongovernment spending and 55 per cent on private consumption (via increasedtaxes on incomes and spending) However, the austerity packages introduced

in many countries have usually focused more on reducing government spendingthan on reducing private spending The typical response in OECD countries hasbeen more like 75 per cent public spending cuts and 25 per cent tax rises This isillustrated inFigure 1.2by the heavy arrow The path chosen by the UK is par-ticularly steep, with thefiscal adjustment loaded on cuts in public spending TheOffice for Budget Responsibility (2013) forecasts in 2013 showed the ratio oftaxes to GDP as broadly constant for the period 2011 to 2017 but the ratio ofmanaged public spending to GDP as falling by some 5 percentage points.Why has an adjustment path been chosen that loads the austerity measuresheavily on the side of cuts in government spending? One can understand thatthere may be short-term reasons why it is easier to make urgent adjustments

to spending and that some stability in taxation is desirable There has been

a– much debated – macro-economic proposition that spending cuts are more

effective in securing fiscal re-adjustment But we are talking here aboutmedium-term policies – over some five years or more Why not simply move

6 A B Atkinson

back down the expansion path? The point I wish to stress there is that this is

a choice A variety of arrows can be followed inFigure 1.2, as illustrated bythe lighter lines The choice of austerity programmes concentrated on spend-ing reductions must reflect other reasons Has there, for example, been a shift

in preferences against public spending? Are austerity programmes being used

as a means for scaling back the size of the state? Or is government spendingseen as redistributive and there has been a shift against redistribution? Whateverthe underlying agenda, it needs to be made explicit We need a reasoned case.What is more, there are evident choices within the categories of spendingand taxation We can cut current spending or we can cut government capitalinvestment We can cut spending on goods and services or can cut transfers If

it is spending on goods and services, then there are many acute choices.Should we scale back services for the elderly, like home help, or services forthe young, like youth training? On the tax side too there are many options,and it is to these that I now turn

1.4 Choices for taxes

Just as there is a choice regarding the balance between spending cuts and taxrises, so too the form of taxation offers a range of possibilities PublicFigure 1.2 Alternative routes for austerity

Public economics and austerity 7

economic analysis can help us in considering the options, but it is essential torecognize that the answers obtained depend crucially on (1) the range ofinstruments considered and (2) the underlying model of the economy and the(often implicit) economic assumptions These two points are the principalsubject of this section.

Range of instruments

In order to illustrate the way in which the choice of taxes depends on therange of instruments, I take the example of the age-old question of the bal-ance between direct and indirect taxation This is a much discussed issue inpublic finance, although you may wonder why After all, it makes little dif-ference whether the government reduces your take-home pay through anincome tax or reduces what you get for it through a value added tax Thereare, however, two differences The first is that the income tax can be madeprogressive, so that better-paid people pay more on an extra€1,000 than lesswell-paid people, whereas a shopkeeper does not ask you about your incomebefore deciding what rate of VAT to charge The second is that the incometax is the same however you spend your income, whereas VAT can be levied

at different rates, so it can be set to zero on food or children’s clothing Doesthis mean that income taxes should be directed at equity goals and indirecttaxes at raising revenue efficiently? The answer is ‘no’, as we can see by con-sidering the optimal role of indirect taxes under different assumptions aboutthe scope for direct taxes– see Table 1.1

To begin with, let us consider the situation where no income tax is possibleand indirect taxes are expected to do all the work Then the design of the taxstructure has to balance both equity and efficiency objectives In loose terms,one wants on equity grounds to tax more heavily goods and services that areluxuries, and on efficiency grounds to tax more heavily goods and servicesthat do not respond much to taxation (their demand is inelastic) These maypoint in opposite directions and a balance has to be struck

But if one allows the simplest direct tax, which is a uniform poll tax, thenthe role of indirect taxes changes Apart from migration in and out of thecountry, the poll tax does not affect behaviour – it is purely lump sum FromTable 1.1 Optimal indirect tax structure under dierent assumptions about theffpossibilities for direct taxation

tax = most efficient way to raise revenue(leaving aside externalities)

3 Fully variable non-linear direct tax Under certain conditions of separability,

and absence of externalities, uniformindirect tax, since equity ensured by direct tax

8 A B Atkinson

an efficiency point of view, this is how revenue should be raised A ment concerned solely with efficiency should not employ indirect taxes butuse the simplest of direct taxes As was recognized by Adam Smith, ‘incountries where the ease, comfort and security of the inferior ranks of peopleare little attended to, capitation taxes are very common’ (Smith, 1904 (1776),page 482) It is only if the government is concerned about equity that itshould use indirect taxes, which have a higher efficiency cost, since they dis-tort consumer choices By raising revenue from indirect taxes on luxuries andreducing the poll tax, the government is benefitting the poorer households, as

govern-it will if govern-it sets a zero rate for goods that they consume more than nately It is distributional concerns that are central In many areas ofeconomics it is assumed that all consumers are identical – much of macro-economic theory is based on the assumption of identical representativeagents– but such an assumption would make no sense in public finance The keyissues in publicfinance arise only when we allow for the fact that people have

proportio-different capacities to pay, as is assumed here

If we go further, and allow for a general non-linear income tax, with ratesthat vary over different tranches of income, then this allows us to achievemore redistribution Indeed, under certain conditions (set out in Atkinson andStiglitz, 1976– seeBox 1.1), it allows us to achieve all the desired redistribu-tion, and there is no need to use indirect taxes for equity objectives The ear-lier result is over-turned: a uniform rate of indirect tax is optimal One caveatwhich is entered is that efficiency may require taxes to be levied, or subsidiespaid, on goods and services that are socially costly or beneficial – or so-calledexternalities This matters and the notion of externalities has a widerpertinence to which I return

Box 1.1 The Atkinson–Stiglitz theorem

Individuals are assumed to differ only in their wage rate per hour Theymaximize utility derived from the consumption of n goods (x1, … xn) andleisure, denoted by x0 There are no external effects The governmentobserves individual incomes and can apply a non-linear income tax, inaddition to indirect taxes The government maximizes an additive function ofindividual utilities subject to raising a specified revenue The Atkinson-Stiglitz theorem (Atkinson and Stiglitz, 1976) states that if the utility function

is weakly separable in leisure and goods (so that it may be written as U[f(x1,

… xn), x0]), then optimal commodity taxes should be uniform regardless ofthe form of the function f()

As has been shown by Konishi (1995), Laroque (2005), and Kaplow (2006),the result can be extended in an important direction They have shown that,even if the non-linear income tax is not optimally chosen, a move from adifferentiated to a uniform indirect tax can be Pareto-improving if accompanied

by a suitable adjustment of the income tax schedule

Public economics and austerity 9

The result shown in the bottom right hand corner ofTable 1.1is one of theconsiderations that underpin the proposal to move to a broader base for theVAT and, in particular, to do away for zero rates of tax, such as those forfood In the UK, VAT is not paid at present on food, children’s clothing, andother items such as books (In Austria, there is a reduced rate at 10 per centfor food and books.) This has led the recent review of UK taxation – theMirrlees Review – to propose bringing these within the tax net Since theimmediate response is that this will hit those on low incomes, they plan touse more than half the revenue to raise the income tax threshold and toincrease income-tested transfers (and adjust tax rates) The Mirrlees Reviewconcludes that the overall package would be both redistributive and raise anadditional £10 billion (0.7 per cent of GDP), or to double the contribution ofincreased taxation to deficit reduction The package would shift the balanceback towards taxes The report presents evidence showing that the bottomfour decile groups would gain in terms of real income and that the top sixgroups would lose (Mirrlees, 2011, Figure 9.5).

Underlying economic assumptions

There are, however, serious questions to be asked about this application ofeconomic analysis and its empirical implementation Most importantly, itassumes that the effect of the imposition of VAT at 20 per cent is to raiseprices by 20 per cent The Mirrlees Review states explicitly (although only in

a footnote) that‘it is assumed that the incidence of the VAT reform is fully onretail prices’ (2010, page 301n) Such an assumption is valid where there isperfect competition and constant costs of production, but ceases to be sowhen these conditions do not hold Whether the tax is fully passed on tothe consumer – no more, no less – depends on the decisions of retailers,wholesalers, and manufacturers Here one must question the underlying view

of the economy The analysis assumes a competitive exchange (Arrow–Debreu economy), where everyone is a price-taker In reality, many marketsare imperfectly competitive, and much of the economics of industrialorganization is concerned with price-setting rather than price-taking behaviour.Imperfect competition may be associated with either under- or over-shifting

of the tax, and in the latter case it is possible that profits may actuallyincrease with the tax rate, as has been shown by Seade (1985) (For furtherreferences, see Myles (1995, pages 361–63).) Indeed, there are some reasons tosuppose that oligopolistic firms will raise their price by more than the tax.Certainly, the extension of VAT to food may serve as a signal for a co-ordinatedround of margin increases, in which case the loss to poorer consumers may belarger than assumed in the calculations cited above It is also the case that theexistence of imperfect competition affects the optimum tax analysis This hasbeen demonstrated by, among others, one of the authors of the MirrleesReview, Gareth Myles He showed (Myles, 1989) that the conditions foroptimal indirect taxation now include terms which depend on the degree of

10 A B Atkinson

shifting, and how it varies across industries, and on the extent to which profitsare taxed As had been pointed out by Austin and Joan Robinson at the time

of the monopolistic competition revolution (Robinson, 1933), the tendencyfor imperfectly competitivefirms to charge more than marginal cost creates adistortion that leads to households consuming less than they otherwise would.This distortion can be corrected by a subsidy – that is, taxing the good less.The situation is just like that of an externality, which is why I highlighted thisaspect earlier As explained by Auerbach and Hines, the condition for anoptimal tax on an imperfectly competitive industry ‘carries precisely theinterpretation offered by Sandmo for the [optimal] tax conditions in the pre-sence of externalities Intuitively, the “externality” in the case of imperfectcompetition is the outcome of the oligopolistic output selection, resulting inthe extra mark-up’ (2003, page 15) (The reference is to Sandmo, 1975.) Thenecessary corrective terms have been derived in a model of monopolisticcompetition by Reinhorn, who notes that‘major policy errors could occur ifone uses the familiar tax rule for the perfectly competitive case when theeconomy is actually imperfectly competitive’ (2012, page 225)

Seen this way, externalities are much more widespread than the case of theexcises on alcohol, tobacco, and petrol In the real world, they arise, to dif-fering degrees, across many industries In the Mirrlees Review, the issue ismentioned only in a footnote on page 156 of the Report, to be dismissed In

my view, this is too hasty Imperfect competition seems particularly relevant

to the case of one of the industries most affected by the expansion of the VATbase, which is the food industry As is well known, food retailing is dominated

by a small number of giant supermarket chains In the UK in July 2013, thepercentage market shares were 30.1 (Tesco), 17.0 (Asda), 16.5 (Sainsbury’s),and 11.7 (Morrisons) (website of Grocery News, September 2013) Three-quarters of the market is supplied by four firms This means that we need amodel of tax incidence with imperfect competition It is also possible that theincrease in tax may be shifted backward Supermarkets are well known fordriving a hard bargain with their suppliers, and it is possible the pressure will

be increased, reducing incomes What is more, there may be “waterbed

effects” Where the supermarkets put more pressure on their suppliers, thesuppliers in turn may seek to recoup some of the loss by raising the price thatthey charge to independent retailers This in turn may drive out some of thefew remaining independent shops, hence increasing still further the degree ofmarket concentration

For these reasons, the extension of VAT to food may not be so evidently agood way forward How then could we raise revenue? One possibility that hasbeen canvassed is a luxury rate of VAT In addition to the current UK stan-dard rate of 20 per cent, we could levy a luxury rate at 30 per cent on certainitems This appears to fly in the face of the theoretical results cited earlier(Table 1.1), since it would increase the degree of differentiation of the VAT Itmoves in the opposite direction from the view that equity can be achieved bydirect taxes However, we have just seen that we need to take account of

Public economics and austerity 11

imperfect competition and of the additional terms that this introduces intothe optimal tax equations Moreover, there is a further limitation of theunderlying analysis, which is that it focuses on just one dimension of dis-tributional difference – earning power There are other sources of inequalitysuch as inherited wealth These will be discussed further in Chapter 2, butthey mean that we have to consider whether luxury rates of VAT may benecessary on distributional grounds to catch the inequality resulting fromconsumptionfinanced from inherited wealth This was indeed one of the keyarguments for a progressive expenditure tax advanced by Nicholas Kaldor inhis well-known polemic: ‘the real inequities of the system arise not so muchfrom the failure to exempt savings out of“income”, but the failure to tax as

“income” the spending power that is exercised through “dissavings” (orspending out of capital)’ (Kaldor 1955, page 14)

Taxes are one side of the government budget, and I have argued that weneed to re-assess the possibilities of seeking additional revenue to meet themedium-termfiscal challenge But what about debt and spending?

Much of the rhetoric of fiscal consolidation is concerned with the nationaldebt as a burden on future generations Austerity programmes are defended

on the grounds that we need to reduce the debt passed on to the future Manyyears ago, President Eisenhower said that‘I do not feel that any amount can

be properly called a “surplus” as long as the nation is in debt I prefer tothink of such an item as “reduction on our children’s inherited mortgage”’(State of the Union Message, January 1960) President Eisenhower was right

in one respect, even if he was wrong in another

Where he was wrong is in focusing solely on the national debt Just as Ihave argued in the case of taxes, we have to consider the whole range ofinstruments In addition to the national debt, we also pass on to our children:

Pension liabilities

Public financial assets

Public infrastructure and real wealth

Private wealth

The state of environment and stocks of natural resources

This was indeed well illustrated by President Eisenhower himself when, in hisnext, and last, State of the Union Message in 1961, he recorded proudly that

he had been responsible for the ‘largest public construction programme inhistory’ (the Interstate Highway System) and many other major publicinvestments The children and grandchildren are driving along those roadstoday Indeed, one could argue that it is the real values – the roads, schools,and stocks of minerals– that matter, and that the national debt is simply an

12 A B Atkinson

accounting arrangement We owe it to ourselves This is not of course true ofexternally held debt, but let me for the moment focus on internally held debt.President Eisenhower was, however, right in a second respect: national debt

is an inter-generational issue It is a matter of justice between one generationand subsequent generations Generational time should not be confused withcalendar time This matters since generations overlap, as is illustrated sche-matically inFigure 1.3, where the bars denote periods of adult life Two gen-erations are shown, and the debt considered is issued in thefirst period, andrepaid in the second period Generation 1 are adult today and are borrow-ing– say, to bail out the banks, where the alternative is raising taxes today.Does this mean that they are passing on the burden to their children? Notnecessarily It depends on who pays the taxes to repay the debt (and to servicethe interest, where the interest rate exceeds the growth rate)

Where it is Generation 1 that repays the debt, then postponing may be seen

as making no difference: this generation has to save an extra equivalentamount to meet the future tax obligation Their total savings rise and thegovernment debt is an addition to their portfolio: it does not displace privatecapital (so there is no impact on production or wages) Where the tax inthe second period is fully anticipated, then it is the same as reducing currentincome, so that there is no difference between debt and tax finance This is thecase of so-called ‘Ricardian equivalence’ The proposition of equivalence isassociated with the name of the English economist David Ricardo, since heprovided a clear analysis of the issues involved in the choice between debt andtax finance At the same time, as Sandmo has observed, ‘it is something of

Figure 1.3 A simple overlapping generations model

Public economics and austerity 13

BORROW Today

Next period

Generation 1

Who repays debt

Generation 1 pays back debt and interest:

no problem

Children pay (part) of debt: implies inter-generational transfer (not necessarily un-wound via increased bequests)

Generation 2

| | Denotes period of adult life

a paradox that the expression “Ricardian equivalence” in modernmacroeconomics has come to stand for a hypothesis that Ricardo definitelydid not hold’ (Sandmo, 2011, page 85) Ricardo was not persuaded thattaxpayers do indeed correctly anticipate the future taxes necessary to repaythe debt.

What is more, part of the burden of repayment falls on future generationswhose interests are not fully taken into account by the current generation.(Even if members of Generation 1 are concerned about the well-being of theirchildren and grandchildren, their mere existence may be uncertain, and in anycase they have their own claim to enter society’s evaluation.) In terms ofFigure 1.3, where the repayment in period 2 is shared between all of thosealive at that date, part of the burden is then shifted on to the younger Gen-eration 2 (Of course, it is also the case that they may have borne part of theburden in their pre-adult life, but I am leaving this on one side.) The shiftinghappens both directly and indirectly The direct impact is that of the taxesthat members of the younger generation have to pay, reducing their lifetimewealth The indirect effect arises because in this case the total capital avail-able, and hence output and wages, are affected There is less productive capi-tal since thefirst generation does not increase its savings by the full amount ofthe debt, so that there is some displacement of private capital As a result, thewages received by the younger generation are lower

If government borrowing transfers part of the burden to future generations,this is not necessarily an argument against borrowing, but it does mean that

we have – unavoidably – to confront distributional questions We have toconsider explicitly the weight to be attached to the incomes or consumption

of different generations As discussed inChapter 3, this breaks down into twoseparate elements First, we may attach less weight to the consumption offuture generations because we expect them to be better off than today’spopulation Second, we may attach less weight to the well-being derived from

a given level of consumption by future generations simply because they aremore distant in time, sometimes referred to as‘pure’ discounting.1

The appropriate policy on the national debt depends on the judgments justdescribed The same applies to decisions about cuts in government spending.The pattern of current spending can have significant inter-generational effects

A major criticism of the UK austerity programme is that the cuts fall moreheavily on the young than on the elderly There have been major reductions inspending on youth programmes, the abolition of the educational maintenanceallowance, whichfinanced young people to stay on at school, and large rises

in university fees At the same time, in the UK there have been smaller cuts incare for the elderly and, as yet, no scaling back of the programme for free bustravel for those over 60

In an inter-generational context, particular significance attaches to publiccapital formation Public infrastructure investment tends to suffer particularly

in times of austerity In their analysis of thirty-two episodes of lasting andsignificant budget consolidation, Balassone and Franco (2000) found that in

14 A B Atkinson

twenty-five cases the ratio of public investment to GDP had decreased, andthat in twenty-three cases investment fell more than current outlays (notincluding debt interest) This is illustrated by the history of public investment

in the UK In 1964 the UK was investing about 10 per cent gross, and around

6 per cent net (of depreciation), of its GDP in public investment This tinued until the mid-1970s when in the 1976 Sterling Crisis the UK governmentwas forced by the IMF to make major cuts in its public spending At thetime, there was much debate about the extent to which fiscal adjustmentshould be made via higher taxes or by lower spending It was spending thatbore the brunt, and capital spending was particularly affected This, appar-ently short-term, measure was followed by a long-run decline in publicinvestment In part, but only in part, the decline was associated with thedenationalization of a number of industries such as electricity, gas, and tele-communications, and with reduced investment in housing But a significantpart reflected reductions in investment in core public services, such as health,education, and transport (Clark, Elsby, and Love, 2001) The state stillneeded to invest, but by 1997 the level of net investment was close to zero.The Blair–Brown Government reversed this trend and net investment reached3½ per cent in 2009, but under the cuts initiated by Labour and taken further

con-by the Coalition public investment is due to fall back to under 1½ per cent Adifference of 2 per cent, which is over a quarter of the projected total fiscaladjustment, would build each year sixty major hospitals, refurbish 2,500schools, or build a thousand miles of motorway

If we are spending less on the services benefitting younger people and sing on less in the form of capital infrastructure, is this fair? This evidentlydepends on the value judgments regarding inter-generational equity describedabove People will take very different positions Some people hold that thepure discount rate should be close to zero or even zero All generationsshould be treated equally Others apply a substantial pure discount factor Inthe optimal tax analysis of Chamley (1986) discussed inChapter 3, he used 4per cent per year In his calculations of the cost of business cycles, Lucas(1987) employed a discount factor of 5 per cent per year The differencebetween zero and 5 per cent will lead to very different views about policy Butthere are some conclusions that can be drawn The first element in theweighting depends on the expected rate of growth Here it seems clear, as Isaid earlier, that we have to take a less optimistic view Even if growth ofGDP is resumed at past rates, a substantial part of the additional resources isgoing to be required to meet the needs of an ageing population and to offsetenvironmental damage Household real spendable incomes are likely toincrease at a slower rate than GDP This means that future generations aregoing to be less well-off in terms of consumption than we anticipated in thepast If that is correct, then whatever one’s view about the other determinants,

pas-we should be increasing the pas-weight attached to future generations

A shift in weight in favour of younger cohorts should be reflected in thedesign of austerity packages Two immediate, and quite obvious,

Public economics and austerity 15

consequences are that we need to prioritize services for young people, fromearly years interventions to youth programmes, and that public investment forthe future should be safeguarded These are obvious – one might say blind-ingly obvious – but this is the reverse of what is being done in the austerityprogramme in the UK A third implication is that there should be a shift inthe direction of higher taxes on better-off members of the older generations.

1.6 Global public economics

So far I have been talking about the choices open with regard to nationalfiscal policies, with a particular focus on those of a small offshore set ofislands – the United Kingdom But each of the topics has had an interna-tional dimension, and it is clear that this dimension should be playing an evermore significant role in our analysis This is true of stabilization measures,where the G20 meeting of 2009 illustrated the possible contribution from suchco-ordination It is true of tax policy where the threat of tax competitionlimits what national governments can achieve It is true of debt policy when

we allow for external holdings of sovereign debt What is more, it is evidentthat there are global problems that require a global response, such as the globalimbalances, the funding of development and the challenge of climate change.For all these reasons, policy-making has to be seen not just in nationalterms but in global terms as well Here, however, public economics falls short

of what is needed There have been valuable contributions on subjects such astax competition, but the international dimension has not been the centrestage Twenty years ago, Ruben Mendez said that‘What is striking in all theliterature on publicfinance is that practically everything is written at the level

of the nation-state or smaller political entities There are no courses or booksdevoted specifically to the subject in international terms’ (1992, page 35).Here I should enter a mea culpa The textbook by Joe Stiglitz and myself,Lectures on public economics, had nothing to say about the global aspects(Atkinson and Stiglitz, 1980)

Chapter 4 is concerned with global public economics, and particularlywith the challenge it poses to the normative branch of public economics Onwhat basis should we evaluate policy? Presumably we would not simply carryover a purely national evaluation of the costs and benefits? But would wemove to a fully-fledged world view – global cosmopolitanism? This wouldmean giving equal weight to all citizens of the world For many peoplethat would be a step too far, not least because of the limited sphere ofcontrol of national governments The challenge is illustrated by the UKTreasury document setting out the basis on which public investment productsshould be assessed:

All impacts (including costs and benefits, both direct and indirect) onnon-UK residents and firms should be identified and quantified sepa-rately where it is reasonable to do so, and if such impacts might affect the

16 A B Atkinson

conclusions of the appraisal Generally, proposals should not proceed if,despite a net benefit overall, there is a net cost to the UK.

(HM Treasury 2011, page 21n)This appears to be adopting a global perspective, but rejecting a globalevaluation

In reality, national governments attach some weight to the well-being ofthose outside their borders but stop a long way short of giving equal weight

As I argue in Chapter 4, one such intermediate position is that where richcountries take a proportionate responsibility But in the meantime I leavereaders to ponder on their own response to the challenge

In terms of substantive policy, my main conclusion about the lessons fortoday’s austerity programmes is that there are choices to be made Not allausterity packages are the same: there are alternatives To begin with, thebalance between spending cuts and tax rises has to be justified in terms ofmedium and long-term implications We need to be clear how far the presentemphasis on spending cuts reflects a shift in social objectives – a scaling back

of the state– rather than simply fiscal consolidation The issue of the nationaldebt has to be seen in terms of all actions that affect different generations,including the pattern of public spending, levels of infrastructure investment,and the distribution of the tax burden Indeed inter-generational justice sug-gests shifting spending cuts towards older generations, protecting publicinvestment, and raising taxes on older generations

As far as public economics is concerned, I have argued that it provides aframework for analysis that is informative, but that needs to be treated withcaution and enriched Conclusions can depend sensitively on the assumptionsmade about the range of instruments at the disposal of the government and

on the underlying model of the economy The standard economic modelemployed is, I believe, no longer fit for purpose – for example in failing totreat imperfect competition as a central case In this respect, public economicsshould draw more extensively on developments in otherfields, such as indus-trial organization and international economics In the latter case, there is theneed to face the challenge of global public finance and to develop a globalwelfare economics In short, there is much that is highly relevant but alsomuch that needs to be done

Notes

1 The role of the two elements may be seen in terms of the social discount rate used

to evaluate the return to public investment: the mathematical formula isδ + εg,whereδ is the pure discount rate, ε is the elasticity of the marginal value of con-sumption and g is the growth rate of consumption per head

Public economics and austerity 17

2 Taxing the rich

2.1 Unequal shares

In 1972, I published a book entitled Unequal shares, which was about wealth

in Britain and taxing the rich (Atkinson, 1972) I was concerned with theconcentration of wealth and the failure to tax inherited wealth sufficiently Insome senses, the book was not well timed Readers asked why I was writingabout continuing wealth concentration when, as may be seen fromFigure 2.1,the inequality of wealth had been falling for some fifty years The share ofthe top 1 per cent in total wealth had been close to a half, but by 1972 hadfallen to around a quarter The share of the top 1 per cent in total grossincome had been around 20 per cent but had fallen to some 8 per cent by the

Figure 2.1 Shares of the top 1 per cent in the United Kingdom

Source: Atkinson and Morelli (2013)

— ■— Share of top 1 per cent in total earnings (NES/ASHE)

— a— Share of top 1 per cent in total earnings (Income tax statistics)

* Share of top 1 per cent in gross income

• Share of top 1 per cent in total wealth

1970s Top earnings had not kept up with the overall rise in earnings Itappeared that one had simply to wait.

However, we now know the results of waiting The 1970s were a turningpoint for inequality in the UK Top income shares have risen: the share of thetop 1 per cent in gross income has doubled since 1979, and is back close topre-war levels All three curves – wealth, income, and earnings – show a U-shape There has been a complete turnaround, just as there was in the UnitedStates and in quite a number– although not all – countries The research onwhich this is based was initiated by Thomas Piketty in his study of France(2001) and has been brought together in the World Top Incomes Database, awebsite that has attracted a great deal of media attention

Why have top incomes attracted attention? First, the rise in top sharesraises issues of social justice When 15 to 20 per cent of total gross incomegoes to the top 1 per cent, then many people are led to ask about the princi-ples on which this distribution is determined The post-war settlementinvolved a more inclusive society in which resources were more equitablydistributed than in the past and this is threatened by a return to pre-war levels

of inequality Secondly, rising top income shares have been in the newsbecause of the issue on which these lectures focus– designing fiscal policy in anage of austerity If, as I argued inChapter 1, we need to re-assess the possibilitiesfor raising taxation, then we cannot ignore afifth of national income

Raising tax revenue provides a direct link with policies to help combatpoverty and social exclusion But the connection goes deeper Unequal shareswas my second book Thefirst book (Atkinson, 1969) had been on poverty inBritain and the reform of social security Then, and still today, I see poverty andriches as being closely connected I often quote the statement by Tawney(1913) that ‘what thoughtful rich people call the problem of poverty,thoughtful poor people call with equal justice a problem of riches’ But first

of all, we need to clarify what we mean by‘rich’ This is the subject of tion 2.2 I then tackle in Section 2.3the much discussed issue of the top rate

Sec-of income tax This is highly controversial In France the top rate has beenraised; in the UK it has been lowered These decisions are certainly informed

by the literature on optimal income taxation, but, as I discuss inSection 2.4,there is much that is missing from the analysis The literature has also con-centrated on the taxation of labour income, whereas we need to consider therole of capital income In Section 2.5, I examine the sources of top incomesand the new issues that are raised for the design of taxation One element inthis story is inherited wealth Widely thought to be a matter of the past,inheritance is returning as an economic factor, as is discussed in Section 2.6.The principal conclusions are summarized in thefinalSection 2.7

There is a tendency to identify the rich with a particular percentage of thepopulation, like the top 1 per cent The division between 1 and 99 has

Taxing the rich 19

surfaced frequently in the media ‘The one per cent’ was the title of a 2006movie ‘We are the 99 per cent’ has been a slogan adopted by the Occupymovement However, 1 per cent is not only arbitrary– why 1 per cent and not0.9 per cent? – but also misses one of the key issues which is whether or notthe rich are a changing proportion of the population Is one differencebetween the US today and the US in the 1950s that there is a largerpercentage of rich people?

A different approach, suggested by Stark (1972) is to define the rich asthose whose income (or wealth– I come back to this later) is above a cut-off.Such an approach is implicit in the widely used definitions of the ‘middleclass’ in the sense that the ‘upper class’ are those above the middle class.However, the usually applied cut-offs in defining the middle class, such asincomes below 200 per cent of median, seem to define a class of a size thathardly corresponds to what is typically envisaged In the US in 2010, twicethe mean income is around $100,000 If the‘rich’ were to be defined as thoseabove this level, then we should be including more than a tenth of the popu-lation More plausibly, we could define the rich as millionaires in the sense ofthose with gross income in excess of $1 million a year In 2010 this wouldhave meant, in the US, the top 0.15 per cent of the population Of course, aswith poverty, the cut-off could be a relative line As was noted by Watkins acentury ago, ‘the “large fortune” is a more or less relative quantity … Therich of former days would not even be“respectably poor” in New York Citytoday’ (1907, pages 3–4) But even such an approach does not help avoid thecharge of‘arbitrariness’ Why take a $1 million in 2010 rather than in 2000?Average income does indeed seem an appropriate point of departure, since

it captures a key advantage of high income: the ability to employ others Inhis ‘The crisis of the tax state’, Schumpeter assessed the tax capacity of dif-ferent rulers in terms of the numbers of foot soldiers or armoured horses thatthey could hire The Elector of Brandenburg for example could hire 800 footsoldiers (1991, page 106) Moving to today, a person with twice averageincome is unlikely to be able to hire a full-time person; one would need alarger multiple Thinking in these terms does not lead to a precise number,but it certainly suggests that we should be thinking more in terms of fivetimes average income or ten times Ten times would equate to looking, in

2010, at semi-millionaires, people with $500,000 or more annual grossincome

If in the case of the United States we were to take a cut-off of ten timesaverage income, then the proportion of rich has indeed grown in the US– seeFigure 2.2 Fifty years ago the number of rich, defined this way, was around0.25 per cent of the population (in each case I am counting tax units) Theproportion remained much the same until the end of the 1970s, and thenbegan to rise It rose steadily well into this century, so that we are looking atthe top 0.5 per cent What is more, the average income of this group has risen.They used to have about twenty times the mean; they now have more likethirty times the mean

20 A B Atkinson

2.3 Optimal taxation and the top tax rate on incomes

How should we tax the rich? In this section, I am going to focus on the singlequestion – what should be the top rate of income tax? This is both a muchdebated question and an excellent example of the strengths and limitations ofmodern public economics It is much discussed in the United States– see forexample Diamond and Saez (2011)– and in France, as a result of PresidentHollande’s introduction of a 75 per cent tax rate, but here I concentrate onthe experience of the United Kingdom

In March 2012, the UK Chancellor of the Exchequer announced that heplanned to reduce the top rate of income tax, echoing the famous Budget of

1988, in which Nigel Lawson, the then Chancellor in the Cabinet of MrsThatcher, reduced the top rate from 60 per cent to 40 per cent, to loud cheersfrom his Conservative supporters One of the Conservative MPs was quoted

as saying that he did not have enough zeroes on his calculator to work outhow much tax he was saving! The top rate then remained unchanged for morethan twenty years, the New Labour Government under Tony Blair refusing toraise the top rate It was only in 2009, late in the Labour Government underGordon Brown, that the top rate was raised from 40 to 50 per cent (forbroadly the top 1 per cent of individuals), with the increase taking effect in2010–11

This recent UK history is set in historical context in Figure 2.3 Theintroduction of graduated tax rates was highly controversial in the UK.The UK had been thefirst country to introduce a personal income tax – in

Figure 2.2 The evolution of the proportion ‘ rich ’ in the United States since 1960Source: US microfiles, based on joint work with Christoph Lakner

Taxing the rich 21

1799– but this had been at a flat rate on all income, if your income exceeded

a threshold A succession of rates over different bands was only introducedwith the supertax in a famous and much contested ‘People’s Budget’, intro-duced by David Lloyd George with the support of Winston Churchill, then aLiberal At that time the top rate was 7.5 per cent, but the top rates wereincreased substantially in the following years, particularly during the FirstWorld War, and remained high It should be noticed that the top rate ofincome tax reached 50 per cent before the UK ever had a Labour Government.But it was Labour that recently returned the top rate to 50 per cent, and itwas the Conservative and Liberal Coalition that decided to cut the top rateback to 45 per cent, with the intention of returning to 40 per cent This taxcut, at a time offiscal austerity, was much debated Two aspects of this debatestruck me The first is that the Chancellor, George Osborne, framed hisdefence of the move in terms of the impact on revenue To quote from theBudget speech, ‘the direct cost is only £100 million a year Indeed HMRCcalculate the loss of other tax revenues may even cancel that out In otherwords, it raises at most a fraction of what we were told – and may raisenothing at all So from April next year, the top rate of tax will be 45p’ Thisstruck me because the maximization of revenue can be interpreted ascorresponding to a Rawlsian objective Where the least advantaged are notthemselves subject to tax, their welfare is maximized by the maximization ofrevenue It may appear strange to class Mr Osborne as a Rawlsian, but there

is nothing particularly socialist about the Rawlsian objective It exhibits no

Figure 2.3 Top tax rates in the UK from 1909 to 2011

- Top tax rate

- Top tax rate (on earned income)

concern about inequality among the rest of the population We do not, in theRawlsian case, have to discuss such questions as the weight to be placed ontop incomes relative to the upper middle and lower middle The welfare of all

of these groups is valued, at the margin, at zero

The second striking feature has been the emphasis on economic analysis.Here the Mirrlees Review, referred to inChapter 1, has played a key role Theofficial study of the implications of the proposed tax cut claimed the support

of the Mirrlees Review, describing its estimate of the revenue impact assistent with that contained in the Mirrlees review’ (HM Revenue and Cus-toms, 2012, Executive Summary) The Mirrlees Review does indeed say that

‘con-‘it is not clear whether the 50% rate will raise any revenue at all’ (2011, page109) The underlying theoretical analysis goes back to the celebrated article

by Mirrlees (1971) on the optimal design of income taxation This article islong and complex, containing 141 numbered equations and a further count-less un-numbered equations Some aspects may, however, be presented moresimply Diamond and Saez (2011) present a half-page calculation of theoptimal income tax rate that suffices here, and this is summarized inBox 2.1.(See also Keane, 2011, section 2 for an exposition.) The simplicity of thiscalculation rests on two assumptions The first assumption is that the dis-tribution of (taxable) income, y, is such that there is a Pareto upper tail ThePareto distribution cannot describe the whole distribution, since the density iseverywhere decreasing, which is not true of the typical distribution ofincome.1 But where the distribution can reasonably be approximated at thetop by the simple formula proposed by Pareto, we can use the convenientproperty that, wherever you slice the top incomes at y*, the average incomeabove y* is equal to a multiple of y* The multiple is α/(a-1), where a is thePareto exponent So, if a = 1.5, as Diamond and Saezfind for the US, then ifyou stand at any point in the upper tail, the mean income of the people aboveyou is three 3 times your income The second simplifying assumption is thatthe elasticity of taxable income, β, is constant over the top income range.Armed with these two assumptions, it may be shown (see Box 2.1) that therevenue-maximizing tax rate is given by t = 1/(1+α.β)

BOX 2.1 Determination of the optimal top tax rate

Consider a small change Δt in the top tax rate, applied to incomes, y, inexcess of y* The change in revenue consists of a direct increase due to thehigher tax rate and an indirect decrease due to the fact that top taxpayerswill reduce their incomes On the assumption that the elasticity of taxableincome, β, is constant over the top income range, the change in taxableincome is β [y/(1-t)] Δt (The elasticity is defined positively as the propor-tionate change in y with respect to a proportionate change in (1-t), whichaccounts for the term y/(1-t).)

This means that the total change in revenue is given by

Taxing the rich 23

the sum over incomes y from y* upwards of [(y-y*)– tβy/(1-t)] Δt

In the square bracket, only y varies, so that we can replace this by itsmean, ym The total change in revenue is now

[(ym-y*)– tβym/(1-t)]Δt

and the revenue maximizing tax rate is then such that the square bracket

is zero:

t/(1-t) = (1-y*/ym)/β = 1/ α.β or t = 1/(1+ α.β)

where the second step follows from the fact that, with the Pareto uppertail, we know that the mean above y* is a/(a-1)y*, so that y*/ym= 1-1/α

This calculation underlies the background chapter to the Mirrlees Review

by Brewer, Saez, and Shephard (2010), where they estimate that the elasticity

of taxable income in the UK for the highest 1 per cent is 0.46 To obtain thisestimate, these authors control for other factors by comparing the change

in the income share of the top 1 per cent (affected by the tax cuts) with that ofthe next 4 per cent (assumed not to be affected by tax cuts) (The governmenttakes afigure of 0.45.) Brewer, Saez, and Shephard take a value for a of 1.67,

so that the elasticity of 0.46 implies a revenue-maximizing tax rate of 56.6 percent When account is taken of the social security tax rate and of indirecttaxes, they go on to conclude that the income tax rate should be no higherthan 40 per cent

2.4 What is missing from the analysis?

This appears to be a triumph for modern public economics Theoretical soning leads to a useable tax formula that can be combined with micro-econometric research to yield a definite conclusion that is immediately taken

rea-up by policy-makers In my view, however, we should be more cautious Thereare several reasons for having reservations about this conclusion In this sec-tion, I am going to set out no fewer than six None of these objections arenew Some – but not all – are clearly recognized in the Mirrlees Review.However, the objections have not, in my judgment, been given sufficientattention

Objection 1: Confidence intervals

First of all, there is considerable uncertainty surrounding the estimate of thetaxable elasticity Brewer, Saez, and Shephard stress that ‘as our estimate ofthe elasticity is tentative, so is the estimated optimal top tax rate’ (2010, page

24 A B Atkinson

110) This is echoed in the main report:‘there is no escaping the uncertaintyaround the estimate of a 40% revenue-maximizing income tax rate’ (Mirrlees,

2011, page 109) The reported standard error for the estimated elasticity oftaxable income implies a 95 per cent confidence interval from 0.21 to 0.71,with implied revenue-maximizing overall tax rates from 46 to 74 per cent– awide range around 56.6 per cent These in turn mean that the optimal tax rate

on the top range could be as low as 24 per cent or as high as 62 per cent.The views of most politicians could be encompassed Moreover, while, toquote Keane, ‘the majority of the economics profession … believes laborsupply elasticities are fairly small (i.e well below 0.50)’ (2011, page 966), hegoes on, in a lengthy literature survey, to say that‘there are important reasons

to question whether this majority view is really an accurate reflection of theempirical evidence’ (page 966)

The point I want to stress here is that this uncertainty regarding a keyparameter should enter the optimal tax formula At the moment, we areacting as though the estimate of the elasticity is produced in one part of theresearch, and then applied as though it were a known parameter in the taxoptimization Instead, the calculation of the optimal tax rate needs to recog-nize explicitly that there is uncertainty surrounding the elasticity and notsimply work with a central point estimate (I am referring here to uncertaintyfaced by the government, not to uncertainty faced by individuals.) Takingexplicit account of the distribution of elasticities could well lead us to set alower tax rate than if we were totally confident in the point estimate If thiswere the case, then the UK Chancellor’s argument for a tax cut would bestrengthened

Objection 2: Assumptions about other taxes

Secondly, we should consider the implications of the arithmetic that took theMirrlees Review calculation of the top tax rate from 56.6 per cent down to 40per cent What is being calculated here is the income tax rate which, togetherwith other taxes, leads to a total tax wedge of 56.6 per cent between themarginal £1 paid by the employer and £1 of goods consumed by the worker.The tax wedge, or‘effective marginal tax rate’, takes account of – in addition

to the income tax – employer social security contributions (at that time, 13.8per cent), employee contributions (2 per cent on high earnings), and valueadded tax then at 20 per cent All of these together mean that the marginalincome tax rate that generates a total wedge of 56.6 per cent is some 38 percent In other words, the income tax is only part of the overall tax wedge.Implicit in the focus on the tax wedge is the assumption that people treatthe different elements in the same way: that £1 of income tax is the same as

£1 of social security contributions This assumption– taken for granted in theReview– is open to question, as I discuss in the next chapter But if we acceptthe assumption for the purposes of the present discussion, we can see that itimplies that the optimal income tax rate depends on what other taxes are paid

Taxing the rich 25

by those affected Suppose that the marginal earnings come from employment, or that people are paid via a company, so that the full rate ofsocial security contributions is not payable? In that case the optimal topincome tax rate should be, not 38 per cent, but 47 per cent It is also quitepossible that the top rate taxpayers do not spend all their income on goodstaxed under VAT Suppose that they spend half their extra income on zero-rated goods, like books or food? Or suppose that they save half of their mar-ginal earnings? Or suppose they spend it abroad? Then the optimal tax ratebecomes 51 per cent.

self-Objection 3: Growing inequality

Thirdly, we should not lose sight of the fact that the optimal tax formula citedabove contains two parameters: there is a as well as β Economists tend toassume that it is the elasticity of response that is the core of their subject, butequally central should be the distribution This is particularly the case wherethe distribution of top incomes is becoming more concentrated in the form of

a lower value for α, implying a higher optimal top tax rate The MirrleesReview estimate is based on the UK Survey of Personal Incomes for 2003–4,whereas the 2009–10 data suggest that α could well be lower If α were to be1.5, then the optimal tax rate on a self-employed person spending half theirmarginal income on books would become 53 per cent

Objection 4: Inter-dependency (positive and negative)

The taxable elasticity used in the optimal tax calculation is estimated ing away any inter-dependence between the incomes of different people.2It isbased on the changes in the incomes of those affected (say, the top 1 per cent)relative to the incomes of those not affected by the tax change In otherwords, the non-taxed group’s experience is used to control for what wouldotherwise have happened The hypothesis is that this group– the bottom 99per cent– earn the same amount as in the absence of the tax change How-ever, this rules out some of the most powerful arguments– in both directions

assum-On the side of those who wish to see a lower tax rate is the view that thetop 1 per cent increase their incomes through increased effort and initiative,and that this effort and initiative creates jobs for others So the revenue effectshould include the taxes collected on these new employees (and saving ontransfers paid) This effect is potentially large, and could justify very low taxrates It could be right – from a revenue point of view – that Warren Buffettpays less tax than his secretary if his actions create many more jobs in, say,the rail freight industry

In the opposite direction, it has been argued that the inter-dependence isnegative: that the increase in income of the top 1 per cent comes at theexpense of other taxpayers An obvious example concerns managerial remu-neration In order to consider this, it is useful to go back to the theories of

26 A B Atkinson

managerial firms and the separation of ownership and control developed byWilliamson, Baumol, and others, in the 1960s and 1970s (for references and acomparison of the different approaches, see Solow, 1971) In these models,managers are concerned with their remuneration (both monetary and non-monetary) but also with other dimensions such as the scale or rate of growth

of their firms, and allocate their effort accordingly Here too we may haveseen a shift In the past, faced with high marginal rates of income tax, topcorporate executives may have seen a low return to negotiating higher payand concentrated on securing alternative sources of utility, such as unpro-ductive corporate expenses, but they may also have ploughed back profits intosecuring faster expansion than in the traditional stock market valuationmaximizing firm But cuts in taxation mean that they switch efforts back tosecuring a larger share of the profits This increase in remuneration, orbonuses, comes at the expense of the shareholders So against the increase inmanagerial pay has to be set the smaller amount paid out to shareholders,which– if in the form of lower dividends – means lower tax revenue This is aconcrete example of the bargaining effect that has been identified by Piketty,Saez, and Stantcheva (forthcoming) As they show, this can lead to sig-nificantly higher optimal top tax rates, coming up with a figure of 83 per cent,

in place of the 56.6 per cent with which we began

Objection 5: Departure from the standard competitive model of the labourmarket

The reference to pay bargaining reminds us that the standard optimal taxanalysis is based on a competitive model of the labour market Just as in theprevious chapter, we need a more realistic economic model, drawing onadvances being made elsewhere in economics– in this case labour economics

As is noted by Persson and Sandmo, the real-world labour market ‘deviatesfrom the standard competitive model in a number of important respects …[This] insight has so far had little effects on the theory of taxation, which stillrelies heavily, both for positive and normative studies, on competitiveassumptions’ (2005, page 558) Persson and Sandmo investigate optimalincome taxation in a‘tournament’ model where wages are determined not byproductivity but by one’s productivity relative to other workers As they note,such a model is particularly relevant to the salaries of top executives It maytherefore be a more suitable framework within which to examine the optimaltop tax rate As is illustrated in Box 2.2, the larger the spread of rewardsoffered by the tournament, the higher may be the optimal tax rate The morethat pay is individualized, the greater the optimal tax rate

BOX 2.2 Optimal taxation in a tournament model

Consider the determination of a single income tax, t, for a group of workers

of differing ability n, distributed with cumulative distribution function F(n)

Taxing the rich 27