“I am not sure that I will be able to finish [a book he was writing on measuring global poverty, building on Atkinson (2017)], but it is quite interesting to read all the different country studies for places that I scarcely knew existed (like the Solomon Islands!). I keep an atlas on my desk! I am very impressed with the overall quality of the work being produced in statistical offices around the world, and feel that there is a lot of scope for mutual learning.”
This is an excerpt from the last email that I received from Tony Atkinson, a few weeks before his premature death on 1 January 2017 from multiple myeloma, an incurable disease diagnosed three years earlier.
These words are telling of Tony’s personality. They well illustrate the insatiable intellectual curiosity that brought him to explore archives to find and understand the data he needed, but also to develop a wide-ranging and deep knowledge of the economics literature. They show his profound respect for craftsmanship – be that of the official statisticians of the Solomon Islands or the carpenter who had fixed his bookshelves. They reveal a supreme commitment to research, which not even his illness could restrain.
The modern theory of inequality measurement
Tony is universally celebrated for his outstanding contributions to the measurement of poverty and inequality, in theory and in practice. There is no doubt that his 1970 paper radically changed the way that economists deal with these problems and laid the foundations of the modern theory of inequality measurement. Developing an idea originated by Hugh Dalton (1920), Tony viewed income inequality as the loss of social welfare associated with an uneven distribution of incomes. This focus on social welfare allowed him to derive three important results.
The first is that it is sometimes possible to rank one income distribution as less, or more, unequal than another by only agreeing on a few regularity properties of the social welfare function and on the “Pigou-Dalton principle of transfers”, which states that total welfare is increased by a mean-preserving transfer of income from a richer person to an (otherwise identical) poorer person. This can be easily checked by simply verifying that the Lorenz curves for the two distributions do not cross.1
But the ensuing ordering is only partial, as the Lorenz curves may cross: other restrictions could be imposed on the social welfare function, but eventually it might be necessary to specify a single function. This leads to the second result: any social welfare function can be converted into an (in)equality index and vice versa. This mapping helps to uncover the collective preferences underlying, implicitly or explicitly, an inequality index but also to derive new indices.
The third result is the development of a class of indices that makes explicit different views concerning distributional justice by introducing a parameter that captures the degree of inequality aversion assumed in the measurement.
Amartya Sen (1997: 114) attributes to this article primary responsibility for “the revival of a basically utilitarian perspective in evaluating inequality” that “proved to be a very productive move” – although there is nothing inherently utilitarian in Tony’s formulation (Atkinson 1983: 5).
Tony’s insights have been enormously influential in stimulating innovative research on the axiomatic derivation of indices, the criteria for ordinal comparisons and the multidimensional measurement of inequality and poverty (for example, Jenkins 2017).2 All these topics remain active research fields (Atkinson and Bourguignon 2000, 2015, Atkinson and Brandolini 2015).
Taking data seriously
Advancing the measurement of inequality and poverty is not only a matter of theory. It is also important to pay proper attention to the data used – their characteristics and fitness-for-purpose, and their limitations whenever they are only a proxy of the ideal theoretical notion. The scrupulous attention to data quality permeates Tony’s research. Throughout his life, Tony scrutinised original sources and perused footnotes and appendices to understand what the data really measure, to assess how they compare to alternative sources, to improve existing statistics or to estimate new ones.
There are plenty of examples: the evaluation of the quality of income data collected in the UK’s Family Expenditure Survey (Atkinson and Micklewright 1983); the extensive and thoroughly documented evidence collected in volumes about wealth distribution in the UK (Atkinson and Harrison 1978); income distribution in Eastern Europe under communism (Atkinson and Micklewright 1992); earnings distribution in OECD countries (Atkinson 2008); the construction of statistics on top incomes undertaken with Thomas Piketty and co-authors; the continuous advice offered to international endeavours such as the Luxembourg Income Study and the EU Statistics on Income and Living Conditions.
Tony’s concern for data issues represents a warning to researchers that the comparability of data is a prerequisite for reliable conclusions (Atkinson and Brandolini 2001). This point is central in the report of the Commission on Global Poverty, where he outlined the criteria for monitoring global poverty that will guide the World Bank in the coming years (Atkinson 2017). The consideration for statistical production gained him the respect of professional statisticians, as testified by the influence of the Atkinson Review on the measurement of government output (Atkinson 2005a) and by his appointment as member of the European Statistical Governance Advisory Board.
Inequality analysis as the basis for policy
Theoretical and practical progress in inequality measurement was never conceived by Tony as an end in itself. It constituted the precondition for analysis of the causes of inequality and policies that may reduce it.
The review of the historical experience of advanced countries led Tony to conclude that income inequality moves erratically, tracing a sequence of episodes rather than well-defined long-term trends (Atkinson 1997). Tony had an inclination to shy away from mono-causal explanations and the search for a great unifying theory, an attitude that probably distinguishes his work from that of Piketty.
The driving factors often emphasised in the academic and public debate – technological progress, globalisation, demographic changes – cannot be separated from specific national factors, such as the choices made by governments on taxation and social protection. Tony has long argued that it is not inevitable that globalisation and technological progress raise inequality – governments retain room for manoeuvre to counter these trends (Atkinson 1999).
Throughout his life – from his first book, Poverty in Britain and the Reform of Social Security (1969a), written when he was just 25, to his last, Inequality: What Can Be Done? (2015), written when disease had already hit him – Tony translated economic analysis into policy actions.
The latter book – which, in many ways, is his intellectual testament – describes a package of concrete measures to reduce inequality that embraces all fields of government action: from public investments to policies for innovation; from the guarantee of a minimum return for small investors’ savings to policies for redistributing income and wealth. People may disagree with specific proposals – Tony typically concluded when presenting the book – but if they hold that inequality is a problem, then they have a duty to advance realistic alternative proposals.
This exposes the ultimate goal of Tony’s research. Economics is a tool for understanding the world and taking informed decisions on policies; but economists must strive to communicate their results beyond the narrow circles of decision-makers, making them accessible for public discussion.
Significantly, he gave his book on the pros and cons of basic income the title of Public Economics in Action (1995). And for over 30 years he was involved in constructing tax-and-benefit micro-simulation models, first for the UK (Atkinson and Sutherland 1988), then for the EU (Atkinson 2005b).
From the outset, he saw these models as tools that could be applied by policymakers, journalists and the general public: “It is essential that the methods used in the analysis should be fully explicit, and the availability of micro-computer programs of the kind produced by [us] is intended to encourage better informed debate about these important issues” (Atkinson et al 1983: 74).
In his Presidential Address to the European Economic Association, he lamented that economists were lagging behind in communication and compared unfavourably with researchers from other disciplines: “It is noteworthy that Hawking in A Brief History of Time almost entirely eschews equations but has some 35 diagrams in his 200 pages” (Atkinson 1990: 245).3
Beyond inequality: A classical economist
It would be reductive to confine Tony’s excellence to a single, narrow field, however fundamental. Tony was an economist in the classical sense. He never conceived the study of inequality as a separate branch of economics and he was keen to maintain inequality as an integral part of the economists’ agenda. His ethical commitment to understanding and providing tools for fighting economic injustice was intimately related to his economic vision.
Tony had a deep appreciation for general equilibrium theory but was fully aware of its severe limitations in interpreting the real world. I remember him making this point in his advanced lectures in microeconomics at the LSE: “There has been a widely-held view in micro-economics that generality is the supreme virtue and that highly structured models are to be avoided as only suitable for the construction of counter-examples. Once we leave perfect competition, however, it may not be possible to make progress in this way, and we may have to specify more of the structure” (Atkinson 1987: 70).4
It is more difficult to assume away distributive considerations as irrelevant when one moves from the abstract general equilibrium assumed by mainstream macroeconomic models to a more complex and realistic framework with imperfections, externalities and multiple equilibria.
Citing Tony’s lecture notes again: “The provision of micro-foundations for macro-economics has been a major part of the theoretical research agenda, but ... the aggregate functions may have different properties from those at the individual level. One cannot simply posit that the macro-economy behaves like some ‘representative’ individual. It is indeed optimistic to suppose that we could derive properties of the aggregate demand functions without specifying anything about the distribution of tastes and income across the population” (Atkinson 1987: 27).
Tony was especially concerned with the design and role of taxation, social protection, and the welfare state. His Lectures on Public Economics (1980), written with Joseph Stiglitz, have represented the advanced textbook for generations of postgraduate students. Yet, his two first academic articles were in macroeconomics.
One stressed that research on growth models was overlooking the speed of convergence to the long-run equilibrium, despite this being a significant prediction of the models: “If we throw away information about the time dimension, we are reducing still further our limited understanding of the relationship between these models and the real world” (Atkinson 1969b: 137).
The other suggested that technical progress does not apply across the board, as usually postulated, but specifically to particular techniques of production; hence, technical development is localised and history matters (Atkinson and Stiglitz 1969). Daron Acemoglu recently noted that this idea “was ahead of its time in emphasising localised and biased new technologies and challenging the orthodoxy in the modelling of technological change” (2015: 456).
The two 1969 articles show salient features of Tony’s future research: the effort to work out all the implications of theoretical models; a willingness to explore analytical solutions different from the received one; and awareness of the limits of models. “Economists are too often prisoners within the theoretical walls they have erected”, he recently wrote discussing austerity policies, “and fail to see that important considerations are missing” (Atkinson 2014: 84).
Had the profession listened more to him, today we would perhaps be less worried about rising inequalities and the risk they pose for our societies.
Author’s note: I thank Giorgio Gobbi, Stephen Jenkins, Marco Magnani, John Micklewright, Luisa Minghetti and Alfonso Rosolia for very helpful comments on this overview of Tony Atkinson's work. The views expressed here are mine and do not necessarily reflect those of the Bank of Italy or the Eurosystem.
Acemoglu, D (2015), “Localized and Biased Technologies: Atkinson and Stiglitz’s New View, Induced Innovations, and Directed Technological Change”, Economic Journal 125: 443-57.
Atkinson, A B (1969a), Poverty in Britain and the Reform of Social Security, Cambridge University Press.
Atkinson, A B (1969b), “The Timescale of Economic Models: How Long is the Long Run?”, Review of Economic Studies 36: 137-52.
Atkinson, A B (1970), “On the Measurement of Inequality”, Journal of Economic Theory 2: 244-63.
Atkinson, A B (1983), Social Justice and Public Policy, Harvester-Wheatsheaf.
Atkinson, A B (1987), “M.Sc. Micro-Economics”, mimeo, London School of Economics.
Atkinson, A B (1990), “Public Economics and the Economic Public”, European Economic Review 34(2-3): 225-48.
Atkinson, A B (1995), Public Economics in Action: The Basic Income/Flat Tax Proposal, Oxford University Press.
Atkinson, A B (1997), “Bringing Income Distribution in from the Cold”, Economic Journal 107: 297-321.
Atkinson, A B (1999), “Is Rising Income Inequality Inevitable? A Critique of the Transatlantic Consensus”, Third WIDER Annual Lecture.
Atkinson, A B (2005a), Measurement of Government Output and Productivity for the National Accounts, Palgrave Macmillan.
Atkinson, A B (2005b), “EUROMOD and the Development of EU Social Policy”, EUROMOD Working Paper No. EM1/05.
Atkinson, A B (2008), The Changing Distribution of Earnings in OECD Countries, Oxford University Press.
Atkinson, A B (2014), Public Economics in an Age of Austerity, Routledge.
Atkinson, A B (2015), Inequality: What Can Be Done?, Harvard University Press.
Atkinson, A B (2017), Monitoring Global Poverty: Report of the Commission on Global Poverty, World Bank.
Atkinson, A B and F Bourguignon (eds) (2000), Handbook of Income Distribution Volume 1, North Holland.
Atkinson, A B and F Bourguignon (eds) (2015), Handbook of Income Distribution Volume 2, North Holland.
Atkinson, A B and A Brandolini (2001), “Promise and Pitfalls in the Use of “Secondary” Data-sets: Income Inequality in OECD Countries”, Journal of Economic Literature 39: 771-99.
Atkinson, A B and A Brandolini (2015), “Unveiling the Ethics behind Inequality Measurement: Dalton’s Contribution to Economics”, Economic Journal, 125: 209-20.
Atkinson, A B and A J Harrison, (1978), Distribution of Personal Wealth in Britain, Cambridge University Press.
Atkinson, A B, M A King and H Sutherland (1983), “The Analysis of Personal Taxation and Social Security”, National Institute Economic Review 106: 63-74.
Atkinson, A B and J Micklewright (1983), “On the Reliability of Income Data in the Family Expenditure Survey 1970-1977”, Journal of the Royal Statistical Society Series A 146 Part 1: 33-61.
Atkinson, A B and J Micklewright (1992), Economic Transformation in Eastern Europe and the Distribution of Income, Cambridge University Press.
Atkinson, A B and J E Stiglitz (1969), “A New View of Technological Change”, Economic Journal 79: 573-78.
Atkinson, A B and J E Stiglitz (1980), Lectures on Public Economics, McGraw-Hill; updated edition: Princeton University Press, 2015.
Atkinson, A B and H Sutherland (eds) (1988), "Tax-Benefit Models", STICERD Occasional Paper, London School of Economics.
Dalton, H (1920), “The Measurement of the Inequality of Incomes”, Economic Journal 30: 348-61.
Jenkins, S P (2017), “Anthony B. Atkinson (1944-)”, in R. Cord (ed.), The Palgrave Companion to Cambridge Economics 1151-74, Palgrave Macmillan.
Kim, E H , A Morse and L Zingales (2006), “What has Mattered to Economics since 1970”, Journal of Economic Perspectives 20: 189-202.
Sen, A K (1997), On Economic Inequality, expanded edition with a substantial annex by J E Foster and A K Sen, Clarendon Press.
 The Lorenz curve shows the proportion of total income received by successive proportionate groups cumulated from the bottom.
 Tony’s 1970 paper is among the most cited articles published in leading economics journals between 1970 and 2002 (Kim et al 2006: Table 2).
 Tony’s interest in graphical representation culminated in the Chartbook of Economic Inequality, which he realised with Salvatore Morelli and Max Roser, where a chart for each of the 25 countries considered shows how different dimensions of economic inequality have changed over time.
 Tony’s lecture notes for the advanced course in microeconomics held at the LSE in the fall of 1987 were handed out to students but bore the warning that they had not been checked and should not be quoted or cited. Here, I am consciously infringing that requirement.