The Analysis of Inequality in the Bretton Woods Institutions

The IMF and the World Bank, like the United Nations, are creatures of the postwar international political and economic order. They (but not the United Nations) were established at an international conference held at the Mount Washington Hotel in Bretton Woods, New Hampshire, in July 1944. Forty-four countries were represented and were the initial signatories of the Articles of Agreement of the two institutions. Perhaps unsurprisingly, the ideas and intentions that dominated the conference and shaped the nascent multilaterals came from the countries that were winning the Second World War and, in particular, from the United States and the United Kingdom.

Whereas the United Nations, whose Charter was formulated at a conference in San Francisco in 1945, was to have a fundamentally political role in international relations, the Bretton Woods institutions were intended to provide a framework for the economic governance of the international system.³ More specifically, the Fund was to oversee the fixed exchange rate system that operated at the time, in which major international currencies were convertible to gold at fixed rates. Under a fixed exchange rate system, payments imbalances (arising from either the current or capital accounts) are not eliminated by the price mechanism (i.e., changes in exchange rates) and must therefore be addressed in some other way. In the long run, the theory went, maintenance of fixed exchange rates would require adjustment in domestic productivities, wages, and prices. In the short run, someone had to provide liquidity to countries experiencing deficits large enough to exhaust their foreign currency reserves. This was, in a nutshell, the fundamental purpose of the (aptly named) International Monetary Fund.

Whereas the Fund was to serve as guardian of the international monetary and financial stability, the International Bank for Reconstruction and Development (IBRD) was intended to provide resources for postwar reconstruction and, secondarily, for the development of poorer countries. The prevalent view at the time was that economic development equaled growth, and that growth was first and foremost a matter of capital investment. Investment is financed by savings, either domestic or foreign. If informational problems, risk aversion, or legal uncertainties prevented savings from flowing from rich countries to profitable investment opportunities in poor countries, then an international public sector institution would serve as a conduit. Unlike the Fund, the Bank was meant to borrow in the international financial markets and to make loans at self-financing interest rates, preferably without calling on its paid-up capital.

It is interesting to note that both institutions have outlived their original purposes. The fixed exchange rate system the Fund was supposed to guarantee collapsed in 1971. The Bank’s reconstruction function, which provided the rationale for its first loans (to countries such as the Netherlands, France, and Denmark), was “soon recognized to be far beyond the financial capacity of the young Bank” (Kapur, Lewis, and Webb 1997, 1: History:10) and was taken on bilaterally by the United States, through the Marshall Plan. The development function remained, of course, but almost no one today would argue that the main constraint to the economic development of poorer nations is a shortage of foreign savings (see, e.g., Ravallion 2016).

Nonetheless, the two institutions proved adept at reinventing themselves. The end of the Bretton Woods fixed exchange rate system (as it was widely known) was very soon followed by the first oil price shock of 1973, which represented a major global terms-of-trade shock. This was followed by a second shock in 1979 and, partly as a consequence, by the US interest rate rises of 1980–81 and the ensuing developing country debt crisis of the 1980s. Payments imbalances had not only not gone away with the adoption of flexible exchange rates; if anything, they became sharper and more turbulent in the 1970s and 1980s. The Fund adapted to the new circumstances and remained relevant and globally active.

The World Bank’s early pivot from reconstruction to development loans was facilitated by the consolidation of the Cold War mindset of the late 1940s and 1950s. As reconstruction loans were first replaced by the US’s bilateral Marshall Plan and eventually became less needed as Europe recovered, the Bank turned its growing lending toward poorer countries. With Western countries firmly in control of the institution, the Bank provided a useful conduit for Western finance and influence in the context of the geopolitical struggle between the two postwar superpowers for allegiance and alignment in the “Third World.” Under its third (and first long-tenure) president, Eugene Black, the Bank created the International Finance Corporation (IFC), through which it could lend to the private sector, in 1956; attained its much-desired AAA credit rating in 1959; and launched the International Development Association (IDA), which raised additional capital from shareholders to subsidize concessional loans to poorer countries, in 1960.⁴ Having lent an annual average of US$1.19 billion (in 1995 prices)⁵ during its first years (1946–49), the Bank had net commitments of $54.4 billion in fiscal year 2020⁶ (World Bank 2021).

Having briefly introduced the two main actors in our story, it is worthwhile briefly considering the meaning of the principal object of interest—namely, inequality. Although many might say that they know intuitively what inequality is, it turns out that it is a multifaceted concept that can mean very different things to different people. Meanings can vary, for example, along the following dimensions:

1. Inequality of what?⁷ Most economists think of inequality as one feature of a distribution. More specifically, they think of it as relating to the dispersion of that distribution. But a distribution is always a distribution of something, among a population or set of “recipient units.” The “something”—the “what”—is obviously of first-order importance. When one sets out to measure or describe “inequality in Brazil in 2020,” the result depends first and foremost on which distribution one is concerned with: is it the distribution of income; wealth; consumption expenditures; educational attainment (e.g., in years of schooling); educational achievement (e.g., in test scores); some measure of health status; some measure of opportunities; or a measure of agency or voice? The list goes on and on. And it doesn’t end with choosing, say, “income”: is income gross or net of taxes? Is it household income or individual earnings? Is it per capita or equivalized? And so on.

2. Inequality among whom? The “something” is distributed over a set of recipient units, often referred to as a “population.” These units of analysis can also differ. A well-defined income concept—say, disposable household income per capita—may be distributed over households or over individuals, and the summary measures will differ. One can also think of an international or global distribution (say, again, of income) as being defined over countries (weighted or unweighted by population) or over individuals. Those three distributions will be quite different, and their inequalities may differ in both levels and trends.

3. Vertical versus horizontal inequalities. Many people—including perhaps most sociologists—think of inequality as gaps or differences between groups, not individuals. The groups may be determined by sex, race, occupation (or class), place of birth, etc. These inequalities are sometimes referred to as horizontal inequalities, whereas the straight measure of dispersion in a distribution is described as “vertical.” Horizontal inequalities are also sometimes known as “categorical” inequalities.⁸

4. Absolute versus relative inequalities. Even once one has a well-defined distribution—which requires a precise definition of the variable (the “what”) and of the recipient units (the “whom”)—and regardless of whether one is most interested in overall dispersion (“vertical”) or gaps between certain groups (“horizontal”), one can still summarize the gaps or dispersion in many different ways. One important distinction is whether the measure one uses to summarize inequality—the inequality index—is built around ratios or absolute differences. Formally, the indices built upon ratios are “scale invariant.” If everyone’s income rises by, say, 10 percent, then inequality is unchanged. But one can also have measures built upon differences, which are known as “translation invariant”. With these indices if everyone’s income rises by, say, $10, then inequality is unchanged. The reader can immediately see the difference. If Jeff Bezos’s income and the reader’s own both go up by 10 percent, then relative (or scale invariant) inequality may not have changed, but absolute inequality has risen. This matters a great deal to the story one tells about what has been happening to inequality. See, for example, Ravallion (2004).

In what follows, use of the term “inequality” without any qualification can be taken to mean a relative measure of (vertical) inequality in the distribution of net household income (or consumption expenditure) per capita among individuals within a given country. That has been the workhorse set of choices to examine economic inequalities within countries at the World Bank and the IMF, and by many other analysts.⁹ But, as we will see, part of the story below involves the choices taken within the two institutions about what distributions to focus on, and how to measure inequality in them.

Histories of ideas are often marked by ebbs and flows, with changes in intellectual fashions occurring sometimes sharply, more often gradually, and sometimes cyclically. The history of thought about inequality at the World Bank and the International Monetary Fund is no different. As we have seen, the IMF is primarily concerned with macroeconomic stability and the management of international financial flows. It is therefore unsurprising that relatively little thought was given to inequality during most of the Fund’s history. Perhaps more surprising is the fact that more recently—over the last twenty years or so—the Fund has indeed produced research and had a public discourse on inequality, as we shall see below. At the World Bank, whose broader focus on development means that it might—indeed ought to—have been concerned with distributional issues from the outset, the ebb and flow is more pronounced. It is useful to divide that history into five phases, which are briefly recounted below. (The IMF reappears in the plot from phase 4.)

Our text analysis reveals little evidence of concern with inequality at the World Bank during its first two decades. The word count for the words “equality” and “inequality” in World Bank documents up until 1969 is essentially flat at zero.¹⁰ For clarity, figure 1 plots that combined word count only from 1970 to 2021, both in absolute and relative terms: the upper panel plots the series for the absolute number of words, while the lower panel shows the relative frequency of those words. Both series are five-year moving averages. They draw on all documents available in the Bank’s Open Knowledge Repository (OKR), which contains mostly text produced from 2000 onward. So, the almost flat line near zero up until 1999 in the absolute series should be taken with a grain of salt: while it is true that the absolute number of documents and reports produced by the Bank has grown very markedly since the late 1940s, the sharp discontinuity at 1999/2000 reflects primarily the selection into the OKR sample. The relative series adjusts for this to the best extent possible, by plotting the relative word count—that is, divided by the total number of words in that year. But even here, the figure begins in 1970 because “inequality” does not really register until 1969.

There are two basic reasons why inequality was not an important concern at the Bank during this early period. First, Bank staff – and the staff involved in analytical writing in particular—has always been predominantly made up of economists, and the economics of development in the 1940s, 1950s, and early 1960s was primarily concerned with aggregates. The dominant neoclassical models of economic growth were those of Roy Harrod (1939), Evsey Domar (1946), and Robert Solow (1956), all three of which had their origins in Keynesian equilibrium theory and explored its implications for capital accumulation when savings equal investment. Thinking about growth (which was thought to be pretty much the be-all and end-all of “development”) was entirely aggregate. From the theoretical side, the models were written for averages or totals, with little or no thought to distribution.¹¹ From the empirical side, there was extremely little data on income (or any other kind of) distribution for poor countries at the time.¹² The empirical focus at the time was on national accounts data, only recently developed, which again were for economic aggregates. Neither economic modeling nor data availability was thus conducive to a focus on inequality.

The second reason is that these were the early days of the Cold War, and inequality and redistribution may well have been seen as “socialist” rhetoric. Although the Soviet Union was represented in the Bretton Woods Conference, it never signed the Articles of Agreement of either the Bank or the Fund. Staffed predominantly with economists trained in the United States and the United Kingdom¹³ and with a Board dominated by the United States and its allies, it was probably natural – though not necessarily right – for the World Bank to avoid such rhetoric.

In 1968, at the end of the Lyndon Johnson administration in the United States, Robert McNamara, formerly the US secretary of defense in charge of prosecuting the Vietnam War, replaced George Woods as president of the World Bank. McNamara hired Hollis Chenery in 1970 to be first his economic advisor and chairman of the Economic Committee and, from 1972 to 1982, vice president of research. Although there had been economic analysis at the World Bank before, including by some prominent economists—Paul Rosenstein-Rodan between 1947 and 1952 being a prominent example—internal research capacity grew markedly during Chenery’s tenure. With this growth came a greater diversity of thematic areas and views, including a growing focus on poverty and human development, as well as the Bank’s first memorable work on inequality.

Although the world was on the cusp of the first oil shock—the consequences of which would, in due course, reorient attention again toward macroeconomic management issues—at the World Bank the early 1970s, and 1973 in particular, marked a substantial increase in the importance given to poverty and distributional concerns. McNamara’s speech to the board of directors that year, delivered in Nairobi, emphasized the 40 percent of the world’s population then thought to live in absolute poverty; noted their concentration in rural areas; and pushed for greater attention to rural development. In September of the same year, an academic conference was held at the University of Sussex that was summarized in a conference volume entitled Redistribution with Growth—a joint study by the World Bank Development Research Center and the Institute of Development Studies at Sussex. Edited by Hollis Chenery, Montek Ahluwalia, Clive Bell, John Duloy, and Richard Jolly (1974), and containing contributions by Irma Adelman, Alejandro Foxley, Mahbub ul Haq, Hans Singer, and Erik Thorbecke, among others, the book was published in 1974.

Read today, Redistribution with Growth is a striking preview of arguments that would only return to the World Bank some thirty years later. Right at the outset, Montek Ahluwalia suggests that “The fact of poverty is not new… What is new is the suspicion that economic growth by itself may not be able to solve or even alleviate the problem within any ‘reasonable’ time period” (Chenery et al. 1974, 3) Reviewing the book, Thiesenhusen (1976) notes that “…the theme running through nearly all the essays is that both more growth and income equality than presently exist are essential for a great number of social and economic reasons and that government policy must intervene” (629, emphasis added).

Previewing an argument later made in the World Development Report 2006—to which we will return—the authors also questioned whether the trade-off between equity and efficiency must hold for all policies and time periods: “In the longer run, however, it can be argued that the transformation of poverty groups into more productive members of society is likely to raise the incomes of all” (47). Even the idea of shared prosperity (introduced as one of the Bank’s twin goals in 2013) was already spelled out here: Ahluwalia and Chenery argued that GNP—which ignores distributional considerations—is an inadequate measure with which to compare economic performance across countries. They suggested that it should be replaced with a weighted sum of income growth across the five quintiles of the income distribution, with declining weights. As we will see, this formulation is uncannily similar—and perhaps preferable—to the definition of the shared prosperity indicator the Bank adopted forty years later.

Redistribution with Growth was not the only important contribution to the study of inequality to come out of the Bank during the 1970s. At the Research Department (then called the Development Research Center), Graham Pyatt and collaborators made a number of important contributions to the decomposition analysis of inequality measures, and of the Gini coefficient in particular (Pyatt 1976; Pyatt, Chen, and Fei 1980). Similarly, in the introduction to his well-known book on Income Inequality and Poverty (Kakwani 1980), Nanak Kakwani writes that “… the major research for this book was done at the Development Research Center of the World Bank […] from July 1974 to February 1976.” Not long after, and also at the Development Research Center, Sudhir Anand conducted his pioneering work on racial inequality in a developing country, using data from Malaysia’s 1970 Post-Enumeration Survey (Anand 1983).¹⁴ On the data front, Shahil Jain’s (1975) compilation of inequality measures for a number of countries was an important precursor to later efforts by the Bank to assemble and curate data on income distribution—a subject to which we will return.

As usual, these analytical developments at the World Bank did not take place in an intellectual vacuum. The early 1970s saw increased interest in issues of distribution in economics, philosophy, and in the social sciences more generally. John Rawls’s (1971) A Theory of Justice was perhaps the most important philosophical challenge to utilitarianism in the twentieth century. Equality of opportunity was one of its key principles of justice, and it also proposed a very low tolerance for inequalities of any kind: they should be tolerated only insofar as they worked to the ultimate benefit of the worst off. Within economics, an early and illuminating attempt at linking long-run growth theory to the size distributions of income and wealth was provided by Stiglitz (1969). Atkinson (1970) placed the measurement of inequality on a sounder and more robust basis, which would provide the foundation for most of the work on measurement that followed, for decades.¹⁵ Amartya Sen’s (1973) book On Economic Inequality was both pathbreaking and influential, and his article in Econometrica (Sen 1976) first proposed that the measurement of poverty ought to explicitly incorporate an aversion to inequality. Some of the aforementioned work on inequality measurement and decomposition carried out at the World Bank during this time was influenced by—and in turn exerted its influence on—the work by Atkinson, Bourguignon, Kolm, and Sen.

Inequality, both between groups of countries and within developing countries, was also being explored at this time by a group of influential Latin American and European economists and sociologists, including Raul Prebisch, Celso Furtado, Andre Gunder Frank, and Fernando Henrique Cardoso, among others. While their so-called “dependency theory” of development is best known for its analysis of power asymmetries and unequal terms of exchange between industrial and poorer countries, it also incorporated an analysis of class dynamics and distributive conflict within countries as central to an understanding of underdevelopment. The theory’s influence in international debates was enhanced by its association with—perhaps even adoption by—the United Nations Economic Commission for Latin America and the Caribbean (ECLAC or CEPAL).

Back in the World Bank, the concern with poverty and human development, and the acknowledgment of the importance of distributional considerations in addressing them, was also reflected in the World Development Report 1980, the third-ever installment in the series of reports that would become the Bank’s most important flagship publication. The report’s theme was “Human Development and Poverty,” and it echoed many of the themes first previewed in Redistribution with Growth, including seeing investment in the human capital of poor people as both an end in itself and as a means to subsequently faster growth. The report also explicitly acknowledged the risk that policies aimed at investing in the poorest people might fall victim to vested interests (Stern and Ferreira 1997, 574).

In the 1980s the intellectual pendulum swung back. The decade marked a departure from these concerns with poverty and distribution, both in mainstream economics and in the World Bank’s political “authorizing environment.” In academia, the oil price shocks of 1973 and 1979; the large interest rate increases of the US Federal Reserve in 1980–81; and the ensuing recession refocused minds on economic fluctuations and business cycles. Although theoretical interest in long-term economic growth would resume later in the decade, economists in the early 1980s were most preoccupied with cycles and fluctuations. To this problem, they brought the elegant mathematical modeling of forward-looking behavior known as the rational expectations hypothesis, developed in the 1970s by Robert Lucas (whom we have met earlier in this paper), Thomas Sargent, and others. Among other characteristics, this approach to modeling the economy tended to rely on “representative agents”—that is, the performance of the economy could be understood while abstracting from differences among people and focusing on averages, whether empirical or hypothetical.

Development economics, within which increasing attention had been paid to issues of agriculture, poverty, and human development in the 1970s, suffered a similar thematic shift, in this instance brought about by the debt crisis of the 1980s.¹⁶ Macroeconomic issues—in particular debt, current account imbalances, and how to adjust to them—took center stage in developing countries too and so, understandably, among many academic researchers and at both the World Bank and the IMF.

Turning to the political authorizing environment: In 1979 Margaret Thatcher became UK prime minister, and in 1981 Ronald Reagan was inaugurated president of the United States. The so-called “Thatcher-Reagan revolution,” with its emphasis on lower taxes, smaller governments, and a greater role for markets, was beginning. In June 1981 McNamara stepped down as president of the World Bank, midway through his third five-year term. His replacement, A. W. Clausen, was a former (and future) CEO of Bank of America. He brought both a commercial banking perspective and Anne Krueger—as the first woman (and second ever) chief economist—to the World Bank.

Under Clausen and Krueger, and in the context of an escalating international crisis of sovereign insolvency, the focus changed to debt management and the economy-wide responses that were judged necessary to respond to the payments imbalances that lay at the root of the debt. These responses became known as “structural adjustment” policies. In keeping with the prevailing intellectual climate and—to be fair—in light of considerable evidence of economic rigidities and inefficiencies in developing countries, the emphasis shifted to restoring adequate market incentives, under the mantra of “getting prices right.”

Some of the policies that were embraced and promoted by the World Bank and the IMF during the 1980s—namely, (i) fiscal discipline; (ii) reordering public expenditure priorities; (iii) tax reform; (iv) liberalizing interest rates; (v) allowing for a competitive exchange rate; (vi) trade liberalization; (vii) liberalization of inward foreign direct investment; (viii) privatization; (ix) deregulation; and (x) enforcing property rights—were summarized and given the name “Washington Consensus” by John Williamson (1989).

Leaving aside whether individual elements of this policy package were or were not suitable for the needs of various countries in the 1980s, one fact now widely accepted is that it lacked an explicit concern with distribution. The second item in Williamson’s list (above) did, in fact, argue for a redirection of public expenditures that would have high economic returns and protect the poor, such as education, primary health care, and certain forms of infrastructure. But in a context where the macroeconomic prescriptions—essentially, expenditure reduction (cuts in government spending) and expenditure switching (devaluations)—were certain to cause significant economic contractions and to hurt living standards, at least temporarily, this was clearly too little. This lack of concern with poverty and inequality in the context of the structural adjustment process proved very costly to the well-being of millions of people in the developing world and, consequently, also to the reputation of the Bretton Woods institutions for years to come.

In a way, the failure of the Bretton Woods institutions to take poverty and inequality seriously during the “lost decade” of the 1980s carried the seeds of its own correction. Falling living standards and increasing inequality in many countries undergoing structural adjustment raised the international profile of those issues and galvanized opposition to the Washington Consensus. Other international institutions took up the cry for a focus on people and, in particular, on poor people. In 1987, along with Andrea Cornia and Frances Stewart, Richard Jolly published Adjustment with a Human Face. Jolly—one of the coeditors of Redistribution with Growth back in 1974—was now a senior official at UNICEF. In 1990 the United Nations Development Program published its first Human Development Report, intended as a more people-centric response to the Bank’s World Development Report (which began publication in 1978). The Human Development Report was led by none other than Mahbub ul Haq, whom we have already encountered as another one of the authors of Redistribution with Growth.

Within the World Bank, too, there was growing concern with the consequences of structural adjustment and with poverty more generally. After A. W. Clausen was replaced by Barber Conable as president and Anne Krueger by Stanley Fisher as chief economist in 1987, those concerns grew strong enough to have the World Development Report 1990 dedicated to Poverty, marking a decade since the 1980 report on Human Development and Poverty. Building on the work of Martin Ravallion and his collaborators, the WDR 1990 launched the $1-a-day poverty line and the global poverty monitoring exercise at the World Bank, which continues to this day.¹⁷ The exercise of attempting to compare living standards internationally and to count the number of people living in extreme poverty globally has been enormously influential and led directly to the international community’s first Millennium Development Goal and Sustainable Development Goal. Quite aside from the 1990 World Development Report, Ravallion and his collaborators developed a research program on the economics of poverty in developing countries that became highly influential, both within the Bank and in scholarly circles.¹⁸

Poverty is not inequality, however. Indeed, the Bank generally continued to be rather careful to speak of the former while avoiding mentioning the latter, unless strictly necessary. But under any rigorous analysis of poverty, sustaining this dichotomy is hard to do. Changes in poverty are mathematically determined by changes in average incomes and changes in the Lorenz curve, which anchors all meaningful (relative) measures of inequality. (See, e.g., Ferreira and Ravallion 2009.) Once one starts to be interested in poverty, inequality becomes increasingly hard to avoid. Indeed, as figure 1 illustrates, the relative frequency of the word “inequality” in World Bank documents—which had risen in the early 1970s and then flatlined near zero throughout the 1980s—revived precisely in the lead up to the WDR 1990.

Aside from the internal dynamics leading to a resumption of interest in poverty, three other factors combined to make the 1990s the decade when inequality gradually became an issue that both the Bank and, soon after, the Fund were prepared to discuss and study seriously. These were, first, the growing availability of household survey data, which was then indispensable for any empirical assessment, measurement, and analysis of inequality in incomes or consumption expenditures. Whereas Ahluwalia, Carter, and Chenery (1979) had disaggregated household consumption or income data for twenty-five countries when they first tried to estimate global poverty, by the end of the 1990s, the estimates from Chen and Ravallion (2001) were based on 297 national surveys from eighty-eight countries.

Second, the political authorizing environment changed once again, away from the starkly rightwing Reagan/Bush period. The fall of the Berlin Wall in 1989 and the collapse of the Soviet Union in 1991 marked the end of the Cold War.¹⁹ Bill Clinton replaced George H. W. Bush as US president in January 1993. In 1995 Clinton backed the appointment of Jim Wolfensohn as World Bank president. While Wolfensohn came to the Bank from Wall Street and was hardly a socialist, he was certainly more open than some of his predecessors to confronting difficult issues, such as corruption and inequality. His first appointment as chief economist (replacing Michael Bruno) was Joseph Stiglitz, in 1997.²⁰ Stiglitz had not yet been awarded his Nobel Prize and was better known for his pathbreaking work on information economics than for his important early work on inequality (such as his previously mentioned 1969 article in Econometrica). Nonetheless, he was the same man who would go on to publish The Price of Inequality in 2012 and The Great Divide in 2015—two of the first “blockbuster” books on inequality by leading economists in the 2010s, to which we will return below.

Third, in academia, too, the pendulum was gradually swinging back toward a concern with distribution and inequality. Interestingly, whereas the advances that had drawn attention to these issues in the 1970s had been mostly in normative analysis and in measurement, in the 1990s it was a literature in macroeconomic theory that had an early and profound impact. This was a set of papers investigating whether and how, in the presence of capital market imperfections, wealth inequality might reduce efficiency and economic growth. It included seminal work by Galor and Zeira (1993), Banerjee and Newman (1993), and Aghion and Bolton (1997). This literature was only one of the factors listed by Atkinson’s (1997) paper, aptly titled “Bringing Income Distribution in from the Cold,” in which he described at greater length the swing in the academic pendulum I only briefly allude to here.²¹ But for the World Bank, concerned as it had always been with growth and efficiency, it was a particularly important factor. It provided theoretical grounding for the then radical idea that redistribution might, in some instances and if done well, support rather than hinder growth.²²

The confluence of a World Bank leadership team nominated and/or backed by the Clinton administration with a broader upswing in the profile of inequality as a subject of study within economics presaged a gradual shift from ideological opposition toward growing interest in the concept at the World Bank. It is probably fair to say that this started in the Research Department, some of whose staff were among the vanguard of the profession’s return to work on inequality. Key inequality-related publications in the mid-1990s included Deininger and Squire (1996), Li and Zou (1998), and Milanovic (1998). But there was nascent interest in issues of inequality in operational units as well, particularly where client countries were themselves concerned—for example, World Bank (1997) on Chile and Ahuja et al. (1997) on East Asia.

The trickle of the mid- to late 1990s gradually turned into a flood—or at least a more respectable flow—in the 2000s. Continuing the “tradition” of decadal reports on poverty, the World Development Report 2000/01 was entitled Attacking Poverty. Whereas the 1990 report’s three-pronged policy recommendations had centered on (i) economic growth; (ii) investment in the human capital of the poor; and (iii) social protection, the 2000 WDR sought to bring empowerment—understood as addressing deep-seated inequalities of voice, power, and agency that favored the rich and powerful to the detriment of the poor—to center stage.²³ Ravi Kanbur, the report’s first director, had intended to lead the analysis with empowerment, and resigned from the World Bank when opposition from the US Treasury Department derailed that plan. Even as late as 2000, and even under a Democratic administration, it was okay for the highest-profile public document of the World Bank to focus on poverty. But to suggest that political inequalities were responsible for the socioeconomic structures that enabled poverty to persist was, evidently, still a step too far.²⁴

The first World Development Report to focus primarily on inequality was the 2006 installment, entitled Equity and Development.²⁵ This report also had a three-part structure: after an introductory chapter, chapters 2 and 3 sought to describe inequalities first within, and then between, countries. Innovatively for the time, income or consumption were only two among various other indicators discussed. There were sections and figures on inequality of years of schooling; life expectancy at birth; and even time use between men and women. Categorical inequalities featured prominently, with decompositions shown by race, ethnicity, gender, and spatial area. Chapter 3, which focused on between-country inequality, drew heavily on work by Branko Milanovic, then a staff member at the Bank’s research department, on global inequalities.

Part 2 consisted of three chapters addressing the question “Why does inequality matter?” The report argued that there were intrinsic reasons to care about inequality—rooted in normative, ethical preferences for equality—as well as instrumental ones. The latter came in two dominant varieties: first, inequality of opportunity and in access to capital (human or physical) was detrimental to an efficient allocation of resources, and thus potentially bad for growth and development. Second, inequality in voice and agency led to poor governance institutions, prone to favor the interests of dominant elites rather than the broader common good. The novel concept of inequality of agency drew on earlier work by Rao and Walton (2004).

Part 3 of the report turned to policy ideas and recommendations. There was an emphasis on leveling the political and economic playing fields, including the now widely accepted idea that promoting competition was likely to be equalizing by preventing the accumulation of market power. Perhaps more innovatively, there was a substantive discussion on how to promote more effective equality in access to justice.²⁶ Returning to international inequalities, the report discussed issues that remain eerily topical today, such as migrants’ rights and the economic incentives for the production of vaccines oriented to illnesses most prevalent in developing countries.

It is probably fair to say that some of these arguments, which are widely accepted now, were rather more novel in 2005–6. Discussing the report and the “Equity Agenda” it was said to launch, Oestreich (2018) writes: “Presciently, the report’s authors focused on the related problems of equity and equality, within states and between states and regions: ‘presciently’ since the issue of equity has received increasing attention in the international community in recent years, exemplified by Thomas Piketty’s Capital in the Twenty-first Century” (557).

After the World Development Report 2006, some of the Bank’s research and analytical work that followed it continued to have an impact. This is true of at least two areas in particular: One was Branko Milanovic’s work on global inequalities, which had first appeared in his important Economic Journal article (Milanovic 2002). It continued to evolve and lead to a number of influential publications, including his books Worlds Apart: Measuring International and Global Inequality, published in 2007, and The Haves and Have-Nots: A Brief and Idiosyncratic History of Global Inequality, published in 2010. Another was empirical work on inequality of opportunity: although the WDR had drawn on an established economic theory of equality of opportunity, due to Roemer (1998) and others, there was very little empirical work on the subject, and a number of Bank authors contributed to the early development of that literature (e.g., Bourguignon, Ferreira, and Menéndez 2007).²⁷

Beyond the research, inequality issues figured again in some high-level discussions at the World Bank between 2012 and 2013, following the arrival of Jim Yong Kim as president. Kim’s management philosophy relied on setting targets to focus minds and provide programmatic coherence to the institution he led. After a period of internal discussion, the Bank adopted twin goals that were to guide all of its work. The first of these goals was the “elimination” of extreme poverty, understood as driving the proportion of people living below the international poverty line under 3 percent by 2030, and it was agreed upon relatively easily. There was much more discussion about the second target. Those within the Bank who resented the focus on poverty as excessively narrow wanted a second goal that was about growth and prosperity more generally. Those who wanted to preserve a focus on the underprivileged insisted that prosperity had to be adequately shared. A concern with inequality featured prominently in these discussions, which ultimately led to the adoption of the growth rate in the average income among the poorest 40 percent of any country as the metric by which “shared prosperity” should be monitored. The choice of that metric owed much to the new Bank chief economist, Kaushik Basu, who had proposed something similar much earlier, in Basu (2000)—though, as we have seen above, very similar ideas had also been floated at the World Bank some forty years earlier.

The period that followed the WDR 2006 also saw a substantial increase in the profile of inequality in research conducted at the IMF.²⁸ In keeping with the Fund’s macroeconomic mandate and expertise, the main focus of that research was on the (long-debated) relationship between inequality and growth. In particular, Berg, Ostry, and Zettelmeyer (2012) found evidence that higher inequality was associated with shorter, less persistent growth spells in a large panel of countries. Ostry, Berg, and Tsangarides (2014) went further and claimed that lower inequality in net incomes was associated with faster and more durable growth, controlling for the level of redistribution, and that there was no evidence that redistribution was bad for growth, unless it was extreme. Although the second paper has been the subject of some convincing criticism,²⁹ the research has been undoubtedly influential, both in terms of academic citations and on the public debate. At least in part, that impact reflects the fact that the IMF had been previously widely perceived as being impervious to concerns about distribution, so that its findings that inequality can harm growth, and that redistribution needn’t, have caused as much surprise for their authorship as for any other reason.

But there is no doubt that the work by Andrew Berg, Jonathan Ostry, and collaborators has marked a real change in the IMF’s perception of inequality as a legitimate field of inquiry, given its pertinence to macroeconomic issues such as the level and sustainability of growth. Figure 2 plots the combined word count for the words “equality” and “inequality” in IMF documents from 1946 to 2021, in both absolute and relative terms. Analogously to figure 1, the upper panel depicts the series for the absolute number of words, while the lower panel plots the series for the relative frequency of those words. Both series are five-year moving averages. They draw on 48,616 documents downloaded from the IMF eLibrary between July and September 2021. The lower panel shows a tenfold increase in the inequality “word-share” for IMF documents, from less than 0.01 percent in 2011 to approximately 0.08 percent of all words in 2021. This compares with a previous peak of 0.02 percent in 2000. It would seem that, at least in terms of the word count metric, attention to inequality at the IMF has been clearly and unambiguously on the rise, in contrast to a more cyclical pattern at the World Bank. Much as in the case of the World Bank, IMF research staff were responding to the rising profile of inequality in the academic literature (and, perhaps, at the World Bank), as well as to a more supportive authorizing environment. This was particularly the case under Christine Lagarde, who took over as the Fund’s managing director in July 2011.

The fifth and final phase takes us from the publication of Thomas Piketty’s (2014) Capital in the Twenty-First Century to the present day. Its dominant characteristic is an explosion in both scholarly and public interest in inequality in the world’s rich nations and, first and foremost, in the United States. In that country, the sustained and pronounced increase in income inequality since the 1980s—however measured—caused the topic, previously seen by many as a fringe left-wing concern, to become entirely mainstream. Within economics as a discipline, interest in inequality had been growing steadily since the early 1990s, but the widespread use of administrative tax data to shed light on the incomes of the very rich has led to a new, burgeoning literature, as well as to a resumption of mainstream work on the design of tax systems and an upswing in interest in public economics as a field. Though there are now very many people working in this subfield, it is probably fair to say that the two key pioneering papers were by Piketty (2003) and Piketty and Saez (2003).

After a gradual, careful, and politically constrained effort to push inequality back to the center of the debate on development, Bank (and now also some Fund) staff suddenly found themselves overtaken by the predominant intellectual wave. Of course, the Bank and Fund have continued to contribute in meaningful ways to what has now become a mainstream topic in the social sciences. These contributions came in at least three varieties. First, some innovative and widely cited research on inequality continued to be produced on both sides of 19^th Street, including Lakner and Milanovic’s (2016) paper on changes in global inequality “from the fall of the Berlin Wall to the Great Recession,” which gave birth to the now iconic “Elephant Curve” and was published in the World Bank Economic Review. A volume entitled Fair Progress? Economic Mobility across Generations around the World (Narayan et al. 2018) also deserves mention: it presented results from a large effort to collect and process for comparability data on intergenerational educational mobility around the world, on a scale never previously attempted.

IMF research on inequality continued to grow during this period, as figure 2 suggests. A revised version of the original paper on “Redistribution, Inequality and Growth,” first circulated as an IMF document (by Ostry, Berg, and Tsangarides) in 2014, was published as Berg et al. (2018). In the following year, Ostry and Berg, along with Prakash Loungani, published Confronting Inequality: How Societies Can Choose Inclusive Growth (Ostry, Loungani, and Berg 2019), a book aimed at disseminating the Fund’s message that inequality is bad for growth to a wider audience. Senior management support for work on inequality, as well as gender issues and other topics relatively new to the IMF, continued under Kristalina Georgieva, who left the number two job at the World Bank to replace Christine Lagarde as managing director in 2019.

Second, both institutions have continued to make a contribution to the monitoring of inequality changes around the world. At the World Bank, this took place in the context of a biannual flagship publication known as the Poverty and Shared Prosperity Report (PSPR), which was intended to monitor the world’s progress toward the institution’s self-assigned “twin goals” of eliminating poverty and promoting shared prosperity. There have now been three such reports (World Bank 2016, 2018, 2020), all jointly produced by some combination of the Research and Data Departments, in partnership with the Bank’s central operational hub for poverty work, which during this time has been known as the Poverty and Equity Global Practice. Although these reports focus explicitly on poverty and “shared prosperity,” they inevitably discuss inequality as well since, as noted above, it is very hard to discuss changes in either without a broader understanding of changes in the distribution. Indeed, the first PSPR (2016) was entitled Taking on Inequality and made the point that growth alone—even if it continued at the relatively high rates of the previous two decades—would be insufficient for the world to reach the 3 percent target for extreme poverty by 2030. Doing so would require a reduction in inequality as well.³⁰

Third, a lot of the work summarized immediately above builds on one of the Bank’s key comparative advantages—namely, its unique ability to assemble, curate, and analyze microdata from a very large number of countries in the world. These data underpin the work on global inequality; on international comparisons of economic mobility; and the monitoring exercises reported in the Poverty and Shared Prosperity reports. Summary statistics from the data sets and, increasingly, the ability to access the microdata directly to produce simple calculations on the fly are made available through a website known as the Poverty and Inequality Platform, which has succeeded the online platform PovcalNet. These websites have been important public goods provided to the academic and policy communities around the world by the Bank’s Research (and now Data) Departments. PovcalNet owes its origins to the early work on global poverty measurement started by Martin Ravallion and colleagues such as Shaohua Chen and Prem Sangraula. Today, it is a large-scale operation that requires seamless collaboration between the research and data groups and the Bank’s operational poverty economists working on each individual country.³¹

Nonetheless, despite all this good recent work on inequality at the Bank and the Fund, the growth in research and public debate on inequality elsewhere has been so large that the Bretton Woods institutions have become relatively smaller intellectual players in the field. In and of itself, this is clearly no bad thing. Inequality truly is one of the fundamental challenges facing today’s human societies, and the greater visibility and popularity of the topic is hugely welcome. Research institutes such as the World Inequality Lab at the Paris School of Economics, which hosts the World Inequality Database and publishes an annual World Inequality Report, are producing important, pathbreaking work extending our knowledge of inequality around the world. Other examples include the team at UNU-WIDER, who host the World Income Inequality Database (WIID) and have had a long tradition of work on both income and wealth inequality, as well as the Stone Center on Socio-Economic Inequality at the City University of New York. Of course, most academic research on inequality is produced at universities around the world, whether they host one of these institutes or not.

From the perspective of the two institutions, however, two features of this particular surge in external research pose a challenge. First, the bulk of the inequality research currently conducted at the world’s leading research universities and institutes focuses on rich countries. This is not exceptional; the same is true of most other topics. Authors, journal editors, and referees residing in the United States and Europe tend to be very interested in what is happening in their own countries. Second, a considerable share of the new work on inequality appears to focus on rich people. This, too, is perhaps unsurprising, given the rapid rise in top incomes—and top income shares—in most countries for which reliable data is available. It is also part of the key innovation that spurred much of this research, which, as noted earlier, was the use of alternative data sources to supplement the inadequate information contained in most household surveys on the richest households.

There is plenty of evidence that the very rich are becoming richer, both in the Unites States and in many (but not all) other countries, both rich and poor. There is also evidence to suggest that as relatively small, powerful groups come to control a disproportionate share of a country’s economic resources, fairness and competition suffer both in markets and, perhaps most insidiously, in the body politic. These mechanisms of growing concentration of power—market and political—may constitute real threats to the well-being of poor and middle-class people in many places. So, the focus on rich people and rich countries is understandable.

But attention spans are limited, and there is a real risk that that focus could detract from interest in poor people and poor countries and, most of all, in poor people in poor countries. These are the people who should matter the most to progressive multilateral institutions, and therein lies the challenge to the World Bank and the IMF. Those staff in the two institutions who would like to promote greater equity must embrace the great recent expansion in global attention to inequality, while at the same time working to keep the bright light of evidence shining on those most deprived and least able to protect themselves from challenges like climate change: the poor in poor countries.

Although intellectual fashions are hard to predict, the intense public and scholarly concern with inequality that marked phase 5 above is unlikely to be a passing fad, at least in the short to medium run. Automation and digitalization have led to a process of occupational polarization in most advanced industrial economies. This has been characterized by falling demand for occupations that previously occupied “the middle” of the wage distribution (see, e.g., Autor, Levy, and Murnane 2003; Acemoglu and Autor 2011), and thus by rising inequality. In addition, growing concentration in a number of industries has led to the rise of superstar firms and, more generally, to a greater concentration of market power in these same countries (see, e.g., Eeckhout 2021). The ensuing increase in economic (or supernormal) profits has contributed to an increase in the income share of capital, broadly understood (Barkai 2020).

These two trends seem likely to ensure a continued interest in inequality both among mainstream academic economists and among policymakers in the countries with greatest influence in the governance of the IMF and the World Bank—the two conditions we identified above as influential in determining the degree of attention to inequality within the two institutions. Additionally, both climate change and the fact that China will soon become a contributor to rising, rather than falling, inequality between countries (as it grows past the world’s average income) are likely to keep upward pressure on global inequality as well.

Altogether, these powerful global trends will probably ensure that large inequalities—between and within nations—will continue to constitute one of the most pressing challenges facing both developed and developing countries for the foreseeable future. But they do not, by themselves, ensure that sufficient attention will be paid to “inequality at the bottom.” The persistence of abject deprivation in the world’s poor countries—more than 700 million (almost 1.8 billion) people living in households that subsist on less than US$1.90 (US$3.20) per person per day at PPP exchange rates³²—is not the focus of most of the work on economic inequality being published today. Nor are its causes. Therein lies an excellent reason why research on inequality by World Bank and IMF staff remains deeply relevant and much needed.

What that research should focus on and how it should be conducted is a matter for the excellent researchers working in the two institutions. My own view, for what it is worth, is that it should cover at least four areas, all of which already build on their comparative strengths—either at the World Bank or at the IMF.

First, a more systematic and in-depth exploration is needed of the quality and comparability of the data on incomes and consumption expenditures assembled by the Bank from national statistical offices.³³ Household surveys were difficult to conduct even before the COVID-19 pandemic, and missing incomes at the top are only one of their many shortcomings. Experimental evidence has suggested that (apparently) small changes in questionnaires—for example, in recall periods for the consumption of certain items—can make a large difference in estimates of well-being (Beegle et al. 2012). Household durable goods, like refrigerators and televisions, which are purchased infrequently but provide service for long periods, are valued differently in different countries. Similar differences afflict the imputation of rental service values of owner-occupied housing. The coverage and granularity of consumption items affects the extent to which expenditures by middle- and upper-income households are captured. And all of these issues are prior to the perennial matter of comparing (and sometimes combining) inequality and poverty estimates for income and consumption distributions.

Staff at the World Bank are cognizant of these issues, and efforts have been made to address them. A recent report on poverty in sub-Saharan Africa (Beegle et al. 2019) used a color-coded system (red to green) to indicate the authors’ assessment of how reliable different surveys on the continent were. Acknowledging such differences, and the existence of serious limitations, is an important first step. But the quality issues are substantial and seeking to address them will require much greater investment. Progress here is not constrained by awareness or incompetence. Indeed, the cross-departmental team currently in charge of the Bank’s ongoing global poverty monitoring exercise now produces a Global Poverty Monitoring Technical Note series that carefully and transparently documents the main methodological details, adjustments, and limitations involved in the analysis. These notes are available online and are a great source of information on the “nuts and bolts” of poverty and inequality measurement at the Bank. Instead, the binding constraint on further progress is largely a matter of resources: at present the Bank simply does not have enough trained staff to attempt this task, given the amounts of data that are received and curated every year.

A second area where the Bank has had some leadership—and to which the Fund has recently contributed as well—is the area of measuring inequality of opportunity. I have argued elsewhere that inequality of opportunity is the “bad cholesterol” of inequality: it is the part of observed inequality in outcomes that is most ethically objectionable, and it is also likely to be the most detrimental to growth and efficiency (because those deprived of opportunities fail to be as productive as they might otherwise have been). Because inequality of opportunity largely reflects the extent to which predetermined circumstances shape life outcomes, it is very closely related to the concept of intergenerational mobility (or lack thereof). Family background characteristics (captured by parental income, but possibly also by parental education, occupation, locational choices, etc.) are among the most important circumstances shaping people’s life chances and outcomes—including in education and income.

Unfortunately, reliable data on income, wealth, or other living standard indicators for two or more linked generations—that is, where it is well measured for both parents and their adult children in representative samples (or administrative databases)—are still rare in most developing countries. Given their frequent interactions with national statistical offices, the Bank and the Fund can play an important role in pushing for more and better data collection on family background in poor countries. Beyond documenting the extent of intergenerational persistence or inequality of opportunity more broadly, there is much work to be done in disentangling the causal pathways for that persistence, unpacking genetics from the nurture effect of families, those effects from that of communities and neighborhoods, and those in turn from those of schools, including teachers and peers.

Third is the intersection between persistent inequality and climate change. Much interesting work has already been done—at the World Bank and elsewhere—using model-based simulations to examine what might happen to food prices, occupational and migration choices, incomes, etc., given changes in temperature, rainfall patterns, and sea levels. There is also work on the actual effects of the climate change that has already occurred including, for example, on the effects of high temperatures on learning (Goodman et al. 2019). But I imagine that more of this kind of work should be possible as, unfortunately, the effects of a warming planet become more evident and large changes in the climate and in the frequency of natural disasters take place. Specific causal attribution of any particular event to the broad process of climate change is both notoriously difficult and not particularly necessary. As long as we understand that, on average, certain events are becoming more likely and how, on average, their costs and consequences are distributed, we will have made much progress.

Still under the interaction between climate change and inequality, another crucial set of questions relates to possible redress mechanisms to compensate poorer countries—and especially the poorer people in those countries—for the consequences of climate change on their lives and livelihoods. It is commonly accepted that many of those who will fall victim to some of the large costs of climate change—be it through cyclones, desertification, flooding, higher food prices, etc.—are people whose lifestyles benefited very little from the comforts that carbon emissions have made possible. Yet some of the obvious mechanisms to restore efficiency to this colossal coordination failure, which involve taxing carbon emissions, could generate massive revenues that, if centrally collected, could be used to make major investments in the lives of those people.

The barriers here are not technical or economic. They are predominantly political, and therefore all the more intractable. Whether we like it or not, it is not easy to convince miners in West Virginia, manufacturers in China, or cattle ranchers in Brazil to accept cuts to their living standards (which would ensue from the taxes) while the revenues are spent far away in Bangladesh, Vanuatu, or the Sahel. This is exactly why multilateral institutions, such as the Bretton Woods pair, are best placed to at least make the calculations and inform the global public of what could be achieved, both in terms of the efficiency consequences of the taxes (internalizing externalities and reducing emissions) and in equity terms (from spending the revenues on the poorest people on the planet).

The fourth and final area I will single out as meriting special attention from researchers at the Bank and Fund concerns the design and performance of policies that can assist individuals and households at the bottom of the distribution in escaping poverty in a durable and sustainable manner. Is this best achieved by giving people cash, or by investing in public services they use? If cash can help, should it be transferred conditionally on some behavior, such as school attendance, or unconditionally? Should it be transferred gradually—say, as a monthly payment—or in a large lump sum? Are there combinations of cash, assets, and services—such as in the so-called “graduation programs”—that do better than single-instrument approaches?

This topic is, of course, a burgeoning subfield within development economics, perhaps beginning with many studies of conditional cash transfers but also including studies of large unconditional cash transfers (e.g., Haushofer and Shapiro 2016), “intervention packages” such as BRAC’s Ultra Poor Graduation Programmes (e.g., Banerjee, Duflo, and Sharma 2021), and basic (or minimum) income programs, whether in poor or rich countries (e.g., Verho, Hämäläinen, and Kanninen, forthcoming). World Bank and IMF researchers have already contributed extensively to this area (a very small sample includes Baird, McIntosh, and Ozler 2011; Coady and Le 2020; Filmer et al. 2021; and Macours, Schady, and Vakis 2012), but much more is needed.

Which policies can reduce inequality by promoting the opportunities and achievements of those at the bottom of the global income distribution—poor citizens of poor countries—is a question that lies at the core of (certainly) the World Bank’s mandate, and we are still far from a consensus of what works best under what circumstances, particularly in the long run. Aside from original research on specific policies, which is critical, another role that is well suited to these institutions—given that its policy importance vastly exceeds the rewards it attracts in academia—is that of systematic stock-taking of the evidence, for a broad assessment of what the balance of findings across studies implies. Fiszbein and Schady (2009) is a good example of how the research resources of these institutions can be usefully harnessed to produce excellent syntheses that are subsequently much used by researchers and policymakers elsewhere.

Whether or not the Bretton Woods institutions will continue to play an important, constructive role in the debate about the nature, causes, consequences, and remedies to inequality remains to be seen. It will most likely depend on the same two factors that I have argued have shaped their performance so far: the external profile of these questions in the broader global scholarly community (especially but not exclusively among economists); and support from the authorizing environment that is ultimately shaped by political currents in key shareholder countries and transmitted to staff via senior managers in the two institutions.

Regarding this second factor, perhaps the largest threat looming on the horizon is the rise of political authoritarianism across a wide range of countries, including many—like the United States—that have hitherto been thought of as bulwarks of liberal democracy. Authoritarian regimes are less interested in—or tolerant of—the truth than their democratic counterparts. Support for impartial inquiry in, say, gender inequality or the consequences of climate change has not grown as a result of rising authoritarian tendencies in Russia, Saudi Arabia, Brazil, India, and the Philippines, to mention only a few. If, as happened during the Trump administration, the United States and European nations also succumb to authoritarian populism, multilateralism as a whole will suffer profoundly. The standing of the Bretton Woods institutions will diminish in tandem with the importance of international cooperation to those governments. Naturally, research and analysis in declining institutions will itself wither and decay.

But even in the absence of that extreme (but sadly no longer so unlikely) scenario, the ability of the Bank and Fund to continue to contribute to policy-relevant knowledge on inequality—or indeed on anything else—is uncomfortably fragile, as it depends on the whim of a revolving cast of senior managers with frequently limited understanding of the “intellectual” aspects of the institutions they run. Most ominous, at least in the case of the World Bank, is the frequently asked question of whether the institution really needs its in-house research department. Couldn’t it simply “outsource” the research function to academics in universities?

This view has recently been associated with Paul Romer, a Nobel laureate macroeconomist who held a singularly unsuccessful tenure as World Bank chief economist, but it is unfortunately more widely held. It is an immensely naive view, as can be readily ascertained by considering the quality of the advice and discourse arising from those other multilateral institutions that have not invested in in-house research capacity. Codified knowledge requires certain skills not only for its production but also for its consumption. Monitoring academic research, assessing its quality and relevance, and engaging with its authors and producers requires an in-house team that understand how the sausage is made. It requires exactly the kind of highly trained researchers, capable of publishing in leading peer-reviewed journals, that the two institutions currently have. Indeed, for addressing the four questions I have listed above, more of those researchers and resources are needed, not fewer. With fewer than one hundred researchers on its seventeen-thousand-strong staff, the World Bank, at least, could certainly afford the additional investment. Indeed, it can ill afford not to make it.

Francisco H. G. Ferreira is the Amartya Sen Professor of Inequality Studies at the London School of Economics, where he is also director of the International Inequalities Institute. Ferreira is an economist working on the measurement, causes, and consequences of inequality and poverty in developing countries, with a special focus on Latin America. His work has been published widely and been awarded various prizes, including the Richard Stone Prize in Applied Econometrics and the Kendrick Prize from the International Association for Research in Income and Wealth.

He currently serves as president of the Latin American and Caribbean Economic Association (LACEA). Prior to joining the LSE, Ferreira had a long career at the World Bank, where his positions included chief economist for the Africa Region and senior adviser in the Research Department. He has also taught in the faculties of Economics at the Pontifícia Universidade Católica do Rio de Janeiro and at the Paris School of Economics. He was born and raised in São Paulo, Brazil, and holds a PhD in economics from the London School of Economics.

The author has no competing interests to declare. For full disclosure, I was a member of staff at the World Bank between 1996 and 2020, working mostly in its research department. Most of my own research has always been about inequality. This can therefore not be a fully dispassionate account and will inevitably be colored by my personal perspective. It should be read in that light.

I am grateful to Branko Milanovic, Biju Rao, three anonymous referees, and the two special issue editors, J. P. Singh and Michael Woolcock, for very helpful comments on earlier versions. I am also thankful to Hillary Vipond for superb research assistance.