Dollar and Kraay Reply

Is global inequality rising or falling? We argued in our article that inequality rose dramatically over the two centuries prior to 1980. Since then, inequality has stabilized, and it has even declined somewhat during the recent era of globalization. We are neither the first nor the only ones to make this point. Our observation on the long historical trend in inequality is based on work by economists Francois Bourguignon and Christian Morrison (forthcoming in the American Economic Review). Recent work by Xavier Sala-i-Martin at Columbia University shows a sharp reduction in all measures of global inequality between 1980 and 1998.

The authors of the three letters disagree, with Pitts even going so far as to claim that there is "widely accepted empirical evidence that inequality within and between countries has increased over the last 200 years." How can there be so much disagreement over such a basic and important fact? We offer three explanations.

First, there may be confusion over concepts. We reiterate that by "global inequality" we refer to the distribution of income across individuals around the world. This reflects both differences in average per capita incomes across countries and income inequalities within countries. Many poor countries have grown more slowly than rich countries in the past 20 years, and within some (although certainly not all) countries, inequality has increased. But global inequality has fallen precisely because populous poor countries (notably China and India, but also Vietnam and Bangladesh) have seen sharp accelerations in their growth rates. This acceleration in growth has helped nearly half the inhabitants of the developing world narrow the income gulf between themselves and the rich world. Large countries therefore figure prominently in measures of global inequality, rather than "distort[ing] the picture" as suggested by Pitts.

Second, the data used to measure inequality matter. There is only one comprehensive source of data on income inequality within countries: representative household surveys carried out by countries' statistical agencies. Although such data certainly have flaws, they are the best we have, and so we use them. Moving from comprehensive measures to narrow measures such as intersectoral pay differences in industry (as advocated by Galbraith in his letter) seems inappropriate for our purposes. In most developing countries, only a small fraction of income consists of wages in the formal manufacturing sector, and most poor people work in agriculture and the informal sector. So it seems unlikely that trends in manufacturing-wage inequality will provide a reliable picture of overall inequality trends, which are often dominated by rural-urban and regional income gaps (as is the case in China, for example).

Third, some of the "facts" offered by the authors of the responses are simply wrong. Wells-Dang asserts that "relative inequality has risen ... dramatically" in Vietnam. This statement is not true. A comparison of data from the only two comprehensive household surveys in the 1990s shows that there has been very little change in the Gini index of inequality, from .33 in 1993 to .35 in 1998. Wells-Dang is also incorrect to suggest that a country in which half the population earns $50,000 and the other half earns $500 would show zero inequality according to the mean log deviation measure of inequality we use. The correct figure is 1.6, which is roughly twice as high as the global inequality we measure.

Finally, what is the role of globalization in all of this? Our point is not (as Pitts suggests) that trade, or even growth for reasons other than trade, lowers inequality within countries. In fact, our research shows that countries that grow faster or trade more are as likely to see inequality decrease as increase. Our point is simply that countries that have become more globalized, in the sense of becoming more open to trade and direct foreign investment, have grown faster. This observation is supported by a large body of cross-country and case-study evidence that has tried carefully to separate out the effects of trade from other factors driving growth. Merely observing that growth in China accelerated in the late 1970s and early 1980s before trade took off does not reverse the conclusions of this research, as Galbraith and Pitts suggest. If anything, it does the opposite. A more careful reading of China's reforms clearly shows that most of the initial acceleration in growth after 1978 was driven by the one-time gains from the dismantling of communal agriculture, and that reforms in state enterprises had at best disappointing effects on performance. It is hard to imagine that China would have been able to sustain as high a growth rate over the next two decades without the benefits of further integration with the world economy.

In short, some of the countries that have become more integrated with the world economy in the past two decades have also been among the poorest and most populous in the world. The acceleration of growth in these countries has been an important force in ending a 200-year-long pattern of rising global inequality.