Inequality
Edward L. Glaeser
U.S
NBER Working Paper No. 11511 July 2005
ABSTRACT
This paper reviews five striking facts about inequality across countries. As Kuznets (1955) famously first documented, inequality first rises and then falls with income. More unequal societies are much less likely to have democracies or governments that respect property rights. Unequal societies have less redistribution, and we have little idea whether this relationship is caused by redistribution reducing inequality or inequality reducing redistribution. Inequality and ethnic heterogeneity are highly correlated, either because of differences in educational heritages across ethnicities or because ethnic heterogeneity reduces redistribution. Finally, there is much more inequality and less redistribution in the U.S. than in most other developed nations.
1. Introduction
he insight that economics impacts politics as much as politics impacts economics lies at the heart of political economy. This circle of causation is at the center of research on the political economy of inequality. Democracy, political stability and executive constraints all appear to be more feasible in more equal societies. Public policies towards redistribution and human capital can make societies more equal.
In Section II of this paper, I review the causes of inequality. Recent increases inequality in the developed world are in part the result of skill-biased technological change (Katz and Murphy, 1992), but government policies influence how these technological changes impact different developed countries ( Hanratty and Blank, 1992).Long-standing differences in inequality across countries may reflect colonialism and patterns of agriculture, but initial differences in inequality also influenced political institutions, policies, and structures, which may have exacerbated economic inequality (as in Moore, 1966, Engerman and Sokoloff, 2002).
In Section III of this paper, I review the second half of the causal chain between inequality and politics— the impact of inequality on government. In medial voter models, inequality predicts more redistribution. However, greater inequality can mean that the wealthy have more resources which can be used to reduce influence. Great gaps between rich and poor may also hurt democracy and rule of law if elites prefer dictators who will protect their interests or if the disadvantaged turn to a dictator who promises to ignore property rights.
In the last section of this review essay, I turn to the special question of “American Exceptionalism,” which in this context refers to the question of why there is so much less redistribution and more inequality in the U.S. than in Europe. American exceptionalism appears to be the result of ethnic heterogeneity and European political institutions, which have been influenced by numerous revolutions and two World Wars. A large body of evidence supports the hypothesis that ethnic heterogeneity reduces support for welfare, and the U.S. is far more ethnically heterogeneous than Europe. The U.S. also has political institutions including a majoritarian government, federalism, and checks and balances which have limited the expansion of the American welfare state, while European countries lack comparable institutions. These institutional differences are not exogenous, but rather reflect the fact that many European constitutions were rewritten by left wing politicians when earlier monarchies were defeated militarily.
2. The Political and Economic Causes of Inequality
Most studies of inequality focus of income, but inequality can also be calculated based on wealth, consumption or any other reasonable proxy for well-being. Wealth or consumption have the advantage that they are less subject to short term income shocks, and the inequality of lifetime earnings is probably more important than the inequality of transitory earnings. However, because wealth and consumption data are not available in enough circumstances, most of the empirical work focuses on inequality of annual income. Consequently, I focus my empirical discussion on that variable.
Measuring income inequality also requires transforming the distribution of income or wealth or consumption into a single measure that can be used in standard empirical work. The literature on this issue parallels the industrial organization literature on market concentration. The most popular measure of income inequality is the Gini coefficient which is the difference between the 45 degree line and the Lorenz curve that shows the cumulative distribution of income. A second measure is the share of total national income possessed by various subgroups of the population, i.e. the share of total wealth owned by the richest 5 percent of the population. In some cases, these variables will actually reveal much more than a Gini coefficient, especially if we are interested in knowing whether inequality matters because the rich are particularly rich or because the poor are particularly poor. As different measures are usually highly correlated, different empirical studies that use these different measures often produce quite similar results (e.g. compare Persson and Tabellini, 1994, with Alesina and Rodrik, 1994).
Inequality over Time
These empirical measures are then used by empirical researchers, who have provided a series of facts about the correlates of inequality. Perhaps the most famous relationship is the Kuznets (1955) curve shown in Figure 1. Income inequality first rises and then falls as countries get richer. This curve points to the initial period of industrialization as the point of development where inequality is maximized. Indeed, U.S. history shows a “great compression” during the middle decades of the 20th century (Goldin and Margo,1992) as the relatively equal period between 1950-1975 followed the far greater inequality of the Gilded age and pre-depression America. Figure 1 also illustrates American exceptionalism as the U.S. is much more unequal than other countries of comparable income.
The Kuznets curve is not just an economic phenomenon; it also reflects political factors. The general pattern in industrializing nations is that there are few public efforts to redistribute during early industrialization.As industrialization proceeds, governments almost universally started taking a more active role in redistribution, which is one reason why inequality declines with development. Development increases redistribution for at least three reasons: development is generally associated with greater government size, probably due to increasing governmental competence; development is associated with greater education and political skill on the part of poorer citizens, and development transforms a dispersed agrarian workforce into clustered industrial workers who can readily be organized.
Somewhat surprisingly, given the Kuznets curve, the great compression was followed within the United States, and elsewhere, by a significant increase in inequality since 1975. Katz and Murphy (1992) conclude that the period of rising inequality in the U.S. appears to be have been driven by rising demand for more skilled workers. The rise in demand for the skilled might be the result of a number of different changes including skill-biased technological change, increasing trade and globalization, the decline of manufacturing, and unions. Pride of place appears now to be given to changing technology (Autor, Katz and Krueger, 1998) as the cause of greater inequality, but other factors also matter.
While most authors seem to believe that rising inequality within the U.S. is the result of economic as opposed to political changes, the impact of these economic changes is determined in part by a political filter. Technological changes and increases in world trade should impact most developed countries in similar ways. Yet the U.S. has experienced a much more striking increase in inequality than most other comparable countries (Picketty and Saez, 2003, Blau and Kahn, 1996, Hanratty and Blank, 1992), and economic forces alone do not appear to explain why inequality rose so much more within the U.S.
Political factors surely played some role in the greater increase in inequality with the U.S. At the top of the income distribution, less progressive taxation in the U.S. made it easier for Americans to become rich and increased the incentives for Americans to acquire large fortunes (Picketty and Saez, 2003). Stronger unions and centralized bargaining in Europe increased equality there among industrial and other middle income workers (Blau and Kahn, 1996). Unions have generally fought inequality among their members, perhaps because heterogeneity breaks down union cohesion or because of the quasi-democratic nature of most unions. At the bottom of the income distribution, general unemployment benefits and restrictive labor market regulations ensure that more of the less productive European workers leave the workforce and are therefore excluded from measures of wage inequality.
Inequality across Countries
If changes in inequality within the U.S. and other countries are primarily the result of changing returns to skills and government responses to those changing returns to skill, differences in inequality across countries are more often caused by differences in the distribution of skills. In particularly egalitarian countries, like in Scandinavia, the population is generally well educated and the distribution of skills is quite compact. Conversely, particularly unequal developing countries like Brazil have enormously heterogeneous skill levels between educated urban elites and less educated agricultural workers.
5. Conclusion
The implications of inequality research for public policy are far from clear. Alesina and Rodrik (1994) suggest that inequality deters growth because inequality leads to costly redistributionary policies. According to this model, redistribution is the problem that inequality creates. The literature on the political economy of inequality does not give clear answers on policies, but it does suggest that shifts in inequality may influence polities in far-ranging, substantial, and often unpredictable ways.
上一页 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] ... 下一页 >>