Global Inequalities (2009)
How large is global inequality? How has it been changing over the last century and, most importantly, over the last two decades? The answers to these two factual questions are important starting points for thinking about the “inclusive globalization” the World Bank wishes to promote.
Inequality is not a uniquely defined concept. It can be measured in many different ways. First, there is the question: “Inequality of what?” The levels and trends can be quite different for inequalities in income, consumption, education or health status, for instance. Second, even within each one of these dimensions, there are differences in concepts and measurement methods.
Perhaps the most interesting concept of global income inequality seeks to capture full interpersonal income differences among the world’s people, irrespective of where they live. (This is in contrast to alternative approaches that ignore inequality within countries and, in some cases, even differences in population size.) The latest available Gini coefficient for global inequality is for the year 2002. The estimate is 0.66 (rising to 0.70 if one uses the new purchasing power parity exchange rates calculated by the International Comparison Program). The Gini coefficient is a widely used measure of income dispersion. It ranges from zero, if everyone has the same income, to one, if all the income is held by a single household. A few examples help grasp the magnitude of global inequality: the Gini for the UK is 0.37 (in 2002). For the US, it is 0.40 (in 2000), for China, 0.42 (2004), and for Brazil and South Africa, it is respectively 0.56 (in 2004) and 0.57 (in 2000).
The best available estimates of global inequality between people suggest that it rose steadily during the 20th Century until the 1980s or early 1990s, when it stabilized and may have started to decline (see Figure 1). Both the long period of rising global inequality prior to the 1980s and the decline in the last decade or two were driven primarily by changes in inequality between countries. By the year 2000 (not in Figure 1), inequality between countries contributed roughly twice as much to global inequality as within-country inequality.
Inequality within countries has not changed as much as between-country inequality over the last century. The period has not seen a universal trend in inequality within countries. Between (roughly) 1994 and 2004, out of 51 countries for which data is available, inequality rose in 16, fell in 29, and remained stable in 6. However, since some of the increases have been large in magnitude, and have taken place in large countries (such as China), the overall “within-country” component of global inequality has risen over the last two decades.
The last two decades have, therefore, seen a reversal of the previous secular trends: inequality between countries has begun to fall, (aggregated) inequality within countries rose, and the combined effect was a small decline in total global inequality (which remains at very high levels). It is important to note that the decline in inequality between countries is driven, to a considerable extent, by growth in China and India, which jointly account for one-third of the world’s population.
There are huge inequalities in other important dimensions too, like health status and educational attainment, both within and across countries. Whereas life expectancy at birth (LEB) is 79 for men and 86 for women in Japan, it is 40 for both genders in the Central African Republic. Between 1960 and 1990, international inequality in LEB fell consistently, reflecting both economic growth and progress in public health and sanitation. Between 1990 and 2000, however, inequality in life expectancies increased again, largely due to the effect of the HIV-AIDS epidemic on life expectancy in Africa. (There were also declines in LEB in Russia, and parts of Central Asia.)
Figure 2 illustrates similarly large disparities in infant mortality rates; not only across countries, but across social groups within a country. While a child born in El Salvador has a 2.5% chance of dying before her first birthday if she is born to a well-educated mother, this probability climbs to 10% if the mother has had no formal schooling. Family differences within countries matter a great deal: even though Brazil is considerably richer than Madagascar, an infant born to an uneducated Brazilian mother is almost twice as likely to die in her first year as one born to an educated mother in Madagascar.
There has also been a secular decline in inequality in the distribution of completed years of schooling, with the mean log deviation falling from 0.392 in 1960 to 0.144 in 2000 (as mean years of schooling rose from 3.4 to 6.3). However, educational attainment is not the same as educational achievement, and it is not clear that the latter are also equalizing. One study reported the reading competence of the average Indonesian student in 2001 as equivalent to a student in the 7th percentile of the distribution in France (World Development Report 2006).
The policy concern with these inequalities is increasingly driven by what they imply for the distribution of opportunities. Children born to a poor and uneducated mother in Mozambique are much likelier to die young than if they are born to educated parents in Colombia (see Figure 2), not to mention Sweden or the United States. Even if they survive, they are unlikely to have access to schools of similar quality. Access to various kinds of infrastructure, as well as to both credit and goods markets, are vastly different in poor countries (or poor areas of a large country) than in richer ones.
The resulting inequality of opportunity is seen as unfair by many, and some argue that it may also be inefficient (in that it leads to much wasted human talent and potential).
The ongoing research focuses on three areas. First, development of more solid data bases, as in the other areas of economics, are crucially important for the understanding of economic processes and for influencing policy makers. In this area, development of new data bases is particularly time- and resource-intensive because of the need to have access to information across entire distributions of individuals and households, and this typically requires large surveys. Therefore, improving the quality of income, education and health surveys, and launching and using new types of surveys (e.g., on wealth, prices or time-use) is of key importance.
The second area concerns studies of joint distributions of income and education, income and health, or education and health. This involves complex interactions, not only because, for example, the gradient of (better) health with respect to income may be more or less steep depending on the country but because the causality may go both ways: greater income “buys” better health, but better health may enable people to earn higher income.
The third priority area is work on better definitions and understanding of what “equal opportunity” in a national and global context means. To find out what conceptually “equal opportunity” is and how it can be measured is indispensable to better anchor the research on inequality of opportunities which is increasingly regarded as a way in which political and economic inequalities are perpetuated across generations.
Francisco H. G. Ferreira, email@example.com, 202-473-4382
Branko Milanovic, firstname.lastname@example.org, 202-473-6968
Most World Bank research documents cited in this summary are available through the World Bank’s research archives at http://econ.worldbank.org/docsearch or the Bankwide archives at http://www-wds.worldbank.org/.
1. B. Milanovic. 2008. “Where in the world are you? Assessing the importance of circumstance and effort in a world of different country incomes and (almost) no migration.” Policy Research Working Paper 4493, World Bank, Washington, DC.
B. Milanovic. 2008. “An even higher global inequality than previously thought: A note on global inequality calculations using the 2005 ICP results.” International Journal of Health Services 38(3): 421–29.
2. Figures 1 and 2 come from the World Development Report 2006: Equity and Development. Washington, DC: World Bank.
3. In terms of their contributions to the total mean log deviation, a commonly used measure of inequality which, unlike the Gini, is additively decomposable. See World Development Report 2006: Equity and Development, p. 64.
4. F. Bourguignon and C. Morrisson. 2002. “Inequality among World Citizens: 1820-1992.” American Economic Review 92(4): 727–44.
F. H. G. Ferreira, and M. Ravallion. 2009. “Poverty and Inequality: the Global Context.” In Oxford Handbook on Economic Inequality, Chapter 26, ed. W. Salverda, B. Nolan, and T. M. Smeeding. New York: Oxford University Press.
5. See Chapter 5 of the World Development Report 2006: Equity and Development for further discussion.