Journal Issue

Point of view: The Rising Inequality of World Income Distribution

International Monetary Fund. External Relations Dept.
Published Date:
January 2001
  • ShareShare
Show Summary Details

DOES it matter what is happening to world income distribution (among all 6.2 billion people, regardless of where they live)? Amartya Sen, the recent Nobel laureate in economics, warns that arguing about the trend deflects attention from the central issue, which is the sheer magnitude of inequality and poverty on a world scale. Regardless of the trend, the magnitude is unacceptable (Sen, 2001). He is right, up to a point. The concentration of world income in the wealthiest quintile (fifth) of the world’s population is indeed shocking and cannot meet any plausible test of legitimacy. The chart shows the distribution of world income by population quintiles. Ironically, it resembles a champagne glass, with a wide, shallow bowl at the top and the slenderest of stems below.

But still, the trend does matter. Many champions of free trade and free capital movements say that world income distribution is becoming more equal as globalization proceeds, and on these grounds they resist the idea that reducing world income inequality should be an objective of international public policy. Moreover, many theories of growth and development generate predictions about changes in world income distribution; testing them requires information about trends. Indeed, the neoliberal paradigm—which has supplied the prescriptions known as the Washington Consensus that have dominated international public policy about development over the past twenty years—generates a strong expectation that as national economies become more densely interconnected through trade and investment, world income distribution tends to become more equal. And it is a fair bet that if presented with the statement, “World income distribution has become more equal over the past twenty years” and asked to agree, agree with qualifications, or disagree, a majority of Western economists would say, “agree” or “agree with qualifications.”

If they are right, this would be powerful evidence in favor of the “law of even development,” which says that all national economies gain from more integration into international markets (relative to less integration), and lower-cost, capital-scarce economies (developing countries) are likely to gain more from fuller integration than higher-cost, capital-abundant economies (developed countries). Developing countries wishing to catch up with standards of living in the West should therefore integrate fully into international markets (by lowering tariffs, removing trade restrictions, granting privileges to foreign direct investment, welcoming foreign banks, enforcing intellectual property rights, and so on) and let the decisions of private economic agents operating in free markets determine the composition and volume of economic activities carried out within the national territory. This “integrationist” strategy will maximize their rate of development; put the other way around, their development strategy should amount to an integrationist strategy—the two things are really one and the same.

Distribution of world GDP, 1989

(percent of total, with quintiles of population ranked by income)

Source: United Nations Development Program, 1992, Human Development Report 1992 (New York: Oxford University Press for the United Nations Development Program).

Fortunately, the self-interest of the wealthy Western democracies coincides with this integrationist strategy for developing countries, because as developing countries grow richer, their demand for Western products expands and their capacity to absorb their population growth at home also expands, reducing the pressure on the West created by surging immigration. The World Bank, the IMF, the World Trade Organization (WTO), and the other global supervisory organizations are therefore well justified in seeking to enforce maximum integration on developing countries for the good of all.

What does the evidence show?

Therefore, a lot is at stake in the question of whether world income distribution has become more, or less, equal over the past twenty years or so. It turns out that there is no single correct answer, because the answer depends on which combination of measures one adopts. It depends on (1) the measure of inequality (a coefficient like the Gini, or quintile or decile (tenth) ratios), (2) the unit of inequality (countries weighted equally, or individuals weighted equally and countries weighted by population), and (3) the method of converting incomes in different countries to a common numeraire (current market exchange rates or purchasing power parity exchange rates). Treating these as either/or choices yields eight possible measures, each with some plausibility for certain purposes. Then there is the further question of what kind of data is used—the national income accounts or household income and expenditure surveys.

My reading of the evidence suggests that none of the eight alternative measures clearly shows that world income distribution has become more equal over the past twenty years. Seven of the eight show varying degrees of increasing inequality. The eighth—the one that uses the Gini coefficient, countries weighted by population, and purchasing power parity—shows no significant change in world income distribution. This is because the Gini coefficient gives excessive weight to changes around the middle of the distribution and insufficient weight to changes at the extremes and therefore, in this case, gives more weight (than a decile ratio) to fast-growing China; the use of countries weighted by population has the same effect; and the use of purchasing power parity tends to raise low incomes more than high incomes, compared with market exchange rates. Hence this combination generates the least rise in inequality. But a recent paper by Dowrick and Akmal (2001) suggests that the Penn World Tables, on which most calculations of purchasing power parity are based (see Heston and Summers, 1991), contain a bias that makes incomes of developing countries appear higher than they are. The tables consequently understate the degree and trend of inequality. When the bias is corrected, even the most favorable combination of measures shows rising inequality of world income distribution over the past twenty years, although the trend is less strong than the trend based on any of the other possible combinations.

It is often said that purchasing power parity measures should always be preferred to market exchange rates and that countries should always be weighted by population rather than treated as equal units of observation. Certainly, purchasing power parity measures are better for measuring relative purchasing power, or relative material welfare, though the available data are not good enough for them to be more than rough-and-ready approximations, especially for China and, before the early 1990s, the countries of the former Soviet Union. But data problems aside, we may also be interested in income for other purposes. Indeed, for most of the issues that concern the world at large—such as migration flows; the capacity of developing countries to repay foreign debts and import capital goods; the extent of marginalization of developing countries in the world polity; and, more broadly, the economic and geopolitical impact of a country (or region) on the rest of the world—then we should use market exchange rates to convert incomes in different countries into a common numeraire. After all, the reason why many poor countries are hardly represented in international negotiations whose outcomes profoundly affect them is that the cost of hotels, offices, and salaries in places like New York, Washington, and Geneva must be paid in U.S. dollars, not in purchasing power parity-adjusted dollars. Using market exchange rates, the conclusion is clear: all four combinations of measures using market exchange rates show that world income distribution has become much more unequal.

Causes of increasing inequality

What are the causes of the rise in world income inequality? The theory is not exactly what one might call watertight; the causality is very difficult to establish. Differential population growth between poorer and richer countries is one cause. The fall in non-oil commodity prices—by more than half in real terms between 1980 and the early 1990s—is another, affecting especially the poorest countries. The debt trap is a third. Fast-growing middle-income developing countries, seeking to invest and consume more than can be covered by domestic incomes, tend to borrow abroad; and they borrow on terms that are more favorable when their capacity to repay is high and less favorable when—as in a financial crisis—their capacity to repay is low. We saw repeatedly during the 1980s and 1990s that countries that liberalized and opened their financial systems and then borrowed heavily—even if to raise investment rather than consumption—ran a significant risk of costly financial crisis. A crisis pulls them back down the world income hierarchy. Hence the debt trap might be thought of as a force in the world economy that is somewhat analogous to gravity.

Another basic cause is technological change. Technological change of the kind we have seen in the past two decades tends to reinforce the tendency for high-value-added activities (including innovation) to cluster in the (high-cost) Western economies rather than disperse to lower-cost developing countries. Silicon Valley is the paradigm: the firms that are pioneering the collapse of distance themselves congregate tightly in one small space. Part of the reason is the continuing economic value of tacit knowledge and “handshake” relationships in high-value-added activities. Technological change might be thought of as distantly analogous to electromagnetic levitation—a force in the world economy that keeps the 20 percent of the world’s population living in the member countries of the Organization for Economic Cooperation and Development (OECD) comfortably floating above the rest of the world in the world income hierarchy. If we have world economy analogues of gravity and electromagnetism, can the world economy analogue of relativity theory be far behind?


Income divergence helps to explain another kind of polarization taking place in the world system, between a zone of peace and a zone of turmoil. On the one hand, the regions of the wealthy pole show a strengthening republican order of economic growth and liberal tolerance (except toward immigrants), with technological innovation able to substitute for depleting natural capital. On the other hand, the regions of the lower- and middle-income poles contain many states whose capacity to govern is stagnant or eroding, mainly in Africa, the Middle East, Central Asia, the former Soviet Union, and parts of East Asia. Here, a rising proportion of the people find their access to basic necessities restricted at the same time as they see others driving Mercedes.

The result is a large mass of unemployed and angry young people, mostly males, to whom the new information technologies have given the means to threaten the stability of the societies they live in and even threaten social stability in countries of the wealthy zone. Economic growth in these countries often depletes natural capital and therefore future growth potential. More and more people see migration to the wealthy zone as their only salvation, and a few are driven to redemptive terrorism directed at the symbolic centers of the powerful.

Reorienting international organizations

The World Bank and the IMF have paid remarkably little attention to global inequality. The Bank’s World Development Report 2000: Attacking Poverty says explicitly that rising income inequality “should not be seen as negative,” provided the incomes at the bottom do not fall and the number of people in poverty falls or does not rise. But incomes in the lower deciles of world income distribution have probably fallen absolutely since the 1980s; and one should not accept the Bank’s claim that the number of people living on less than $1 a day remained constant at 1.2 billion between 1987 and 1998, because the method used to compute the figure for 1998 contains a downward bias relative to that used to compute the figure for 1987. Suppose, though, that the incomes of the lower deciles had risen absolutely and the number of people in absolute poverty had fallen, while inequality increased. The Bank’s view that the rise in inequality should not be seen as a negative ignores the associated political instabilities and flows of migrants that—all notions of justice and fairness and common humanity aside—can harm the lives of the citizens of the rich world and the democratic character of their states.

The global supervisory organizations like the Bank, the IMF, the WTO, and the United Nations system should be giving the issue of global income inequality much more attention. If we can act on global warming—whose effects are similarly diffuse and long term—can we not act on global inequality? We should start by rejecting the neoliberal assumption of the Bretton Woods institutions over the past two decades, now powerfully reinforced by the emergent WTO, that development strategy boils down to a strategy for maximum integration of each economy into the world economy, complemented by domestic reforms to make full integration viable. The evidence on world income distribution throws this assumption into question—as does a lot of evidence of other kinds. International public policy to reduce world income inequality must include a basic change in the policy orientation of the World Bank, the IMF, and the WTO so as to allow them to sanction government efforts to impart directional thrust and nourish homegrown institutional innovations.

Robert Hunter Wade

Suggestions for further reading:

Other Resources Citing This Publication