During the twentieth century, economists began analyzing the growth and development of economies to determine their welfare-improving effects on citizens. One of the ways these effects have been measured is by calculating the impact of economic growth on a country’s income distribution. Simon Kuznets (1901-1985) was the first economist to attempt to do this. Kuznets (1955) conducted a study into the evolution of the distribution of personal income from 1870 to the 1950s and found that the relative distribution of pretax income in the United States, Germany, and England had been gradually moving toward equality. Per capita income in these countries had been increasing over the entire period, but at some point the share of the lower-income group began increasing more rapidly than that of the higher-income groups. The developed countries saw increasing income inequality with growth, then experienced decreasing income inequality. In developing countries, however, income inequality had been increasing along with increases in per capita income.

Economic theory suggests that a higher income share among a country’s wealthy will lead to increased savings, thus increased wealth. Economic growth also usually leads to an increase in industrialization and urbanization, with an increasing disparity in income between rural and urban sectors. These trends suggest that there should be increased inequality with economic growth. Therefore, the trend toward income equality for these developed economies was a puzzle to Kuznets.

W. A. Lewis’s (1954) dual economy model provides some explanation. According to Lewis, an economy starts with unlimited supplies of labor in its agricultural sector. As the economy develops, it creates another more industrialized sector—a manufacturing sector. This change causes a movement of labor from the agricultural sector to the new sector, and leads to worsening income inequality as the new sector enjoys better returns. However, increased movement of labor across the two sectors will, in the long run, reach a turning point, causing incomes in the agricultural sector to improve and leading to greater income equality.

Another plausible explanation for the effect of development on income inequality is its effect on education and the production of human capital. At the initial stages of development, only the rich may be able to afford education, augmenting their skills and thus their income. This will increase income inequality. However, as per capita income increases and more people are able to afford education, the income of the poor will converge with that of the rich.


The inverted-U hypothesis paradigm originated from Kuznets’s initial observation of the growth and distribution of income inequality in the United States, Germany, and England. Some economists argue that this theorem can be applied to all economies, and they have been using empirical evidence to prove this theorem and to determine the causal roots of this process. To be able to successfully test this theorem, however, long-term income distribution data is required. The lack of such data has led to many creative means of testing and to a lack of consensus among economists on how the hypothesis works and what happens when there is increasing per capita growth.

Felix Paukert (1973) used cross-sectional data of countries (although he acknowledged the need for long-term data) and found a tendency toward equality among developed countries, as evidenced by an increase in the share of income of the bottom 60 percent. Among developing countries, however, there is a tendency toward inequality, although Paukert was unable to offer definitive conclusions due to the lack of data in the sample. He therefore analyzed a cross section of forty-three countries and found a Gini ratio of 0.467 among developing countries and 0.392 among developed countries. Since the Gini ratio measures inequality of a distribution ranging from 0 to 1, where 0 corresponds to perfect equality (everyone has the same income) and 1 corresponds to perfect inequality (all but one person has zero income), Paukert’s findings suggest that developed countries have moved toward greater equality relative to developing countries.

Figure 1

Figure 1

Montek Ahluwalia (1976) followed up on the work of Paukert and determined the relationship between income inequality and per capita growth rate using crosscountry data. His sample included sixty countries—forty developing countries, fourteen developed countries, and six socialist countries. Ahluwalia’s results showed that an inverted U-shape could be fitted onto the countries at various stages of development and income inequality. However, the slope of the fitted curve changed when he looked at the sample of developing countries. These results were, however, obtained by controlling for socialist countries, because their policies tend to move countries toward development. Ahluwalia suggested that these results may be "stylized facts" for which an explanation may not be possible.

The use of cross-country regressions with countries at different stages of development is not an effective method for testing the Kuznets hypothesis. This method assumes that the turn from increasing to decreasing income inequality happens around the same level of per capita income for all countries. It also assumes homogeneity across countries. Ashwani Saith (1983) questioned whether the empirical work done to confirm the U-hypothesis is in vain. There exists diversity between countries that cannot be accounted for by controlling for whether a country is socialist or developing. A logical way to test the hypothesis may be a time series analysis for individual economies or simultaneous equations on all countries. Nonavailability of the necessary data has kept this project from being undertaken.

Figure 2

Figure 2


Income inequality began increasing in the late 1960s and early 1970s in the United States, England, and Germany, creating yet another puzzle for economists. Rati Ram’s (1991) time series study of U.S. data on income inequality from 1947 to 1987 showed that income inequality followed more of a U-shaped pattern than an inverted-U. His data do not include the earlier period of increasing income inequality. This finding has led to the S-curve hypothesis, which implies that with growth in per capita income, an economy will begin to experience increasing and then decreasing income inequality, and after a period of adjustments, income inequality will begin to rise again. Other economists have argued that the inverted-U is just repeating itself, and thus we should expect to see another period of declining inequality. However, the why and when remain unanswered.

John List and Craig Gallet (1999) analyzed data from seventy-one countries for the 1961—1992 period to test this hypothesis, and found that countries could be placed along all the turns of an S-curve. For example, third-world and developed countries are located on the upward-sloping portion of the curve, while emerging economies are located on the downward-sloping portion. Unfortunately, in addition to covering only a short time period, their data did not include all years for all countries. What they ultimately prove is the relationship between per capita income and income inequality in the world. They fail to address historical change in inequality with changes in per capita income.

Romie Tribble (1999) associated the turn from decreasing income inequality to increasing income inequality with another change in the important sector that affects economic growth. The first turning point in the S-curve can be attributed to economies moving from an agricultural to a manufacturing sector. The second turning point occurs when the service sector becomes dominant, a development characterized by increasing returns to education, thus leading to economic growth. Using U.S. data from 1947 to 1990, Tribble proved that the S-curve fit the data. The shifts in the sectors matched the years considered, but he was unable to test this finding. If his analysis is correct, it implies that some developing countries that never build up their manufacturing sector and seemingly move directly from an agricultural sector to a dominant service sector may never experience a period of decreasing income inequality.


An example of whether Kuznets’s hypothesis can be translated to other developing countries can be found in the economic growth of Asian countries, especially the "tigers." The four big East Asian tigers—Singapore, Taiwan, Hong Kong, and South Korea—achieved high growth rates between the 1960s and the 1990s through export-driven economies. They were able to maintain this growth rate by an increasing shift to the manufacturing sector, industrialization, and improvements in education. However, armed with the experiences of developed nations, these economies offered protection for their agricultural sector in the form of better property rights and subsidies. A graph of income inequalities as measured by the Gini-coefficient over this period shows Hong Kong with a U-shaped curve, and its minimum inequality occurring in 1980 at a Gini of 0.39. Taiwan experienced declining income inequality up to the 1980s, and its Gini of about 0.29 remained fairly constant thereafter. Singapore and South Korea’s income inequality also appears to have remained fairly constant over the time period (data obtained from the UNU/WIDER World Inequality Database, Version 2.0, developed from the Deininger and Squire [1996] dataset).

Income inequality has been increasing in most countries, with the worst increases occurring in developing countries. Countries that should be experiencing decreasing inequality according to the inverted-U hypothesis do not experience it. Increases in cross-country trade, factor mobility, and the ease of capital movement across nations have led to increased globalization. Growth in per capita income could thus be occurring in a nation, but its benefits may not be realized by its population.


Empirical work done on the Kuznets hypothesis generally begins with an assumption about what the relationship between growth and inequality should be, leading to an estimation technique that confirms the assumption. Growth in per capita income cannot fully explain changes in income inequality. Economic growth does not in itself change the income of the poor; rather, it presents opportunities for improving the population’s welfare through the policies adopted in a country. Analyzing this relationship requires a better understanding of these policies and their effects.

Another complication in this analysis is the reverse effects that income inequality could have on economic growth. Economists have theorized that more money in the hands of the rich would lead to increased investments and thus growth in the nation. However, isolating the effects of economic growth could prove daunting.

Next post:

Previous post: