Inequality Across Countries
Those countries are less equal to this day. Acemoglu, Johnson, and Robinson (2001) argue that settler mortality determined whether countries attracted large European migrations (which then largely eradicated the existing native populations) or smaller-scale migrations that then interacted with native and slave labor. Galor, Moav, and Vollrath (2004) present a more general theory of land distribution and long-run inequality. If populations with heterogenous educational backgrounds are the initial condition determining modern inequality, the actual distribution of income today also depends on a large number of government policies towards schooling and redistribution since that point. Many Asian countries, which once had small European elites and large masses of uneducated natives, are now both well-educated and egalitarian. These transformations reflect the results of government policies towards education in countries like Singapore, Taiwan, and South Korea. By contrast, initial skill differences in countries in ethnically heterogeneous countries were often widened, not eliminated, by subsequent educational policies, which favored the European elites. Figure 34.3 shows the strong negative correlation between inequality and social welfare spending across countries. This correlation reflects both that social welfare reduces inequality and that more initial inequality leads to less redistribution.
While some of this relationship reflects reverse causality, and omitted variables (like income and fractionalization) that drive both variables, there is little doubt that governments have the power through tax policy and through spending to alter the income distribution of their country. Understanding income inequality across countries also requires understanding the reasons that a government decides to invest in its children, or to invest in redistribution more generally. Two factors, apart from inequality itself, appear to be important: ethnic heterogeneity and political institutions. Ethnic and racial fractionalization limits the tendency to redistribute income either because people are less willing to support transfers to those who are ethnically different, or because ethnic differences provide a means of demonizing policies that help the poor (Alesina and Glaeser 2004). Individual data on opinion polls and voting behavior, cross-state outcomes within the USA, cross-national data, and history all support the view that ethnic heterogeneity reduces redistribution. Luttmer (2001) shows that people are more likely to support redistribution when they live around poor people of the same race and less likely to support redistribution when they live around poor people of a different race. Holding income constant, there is also a huge impact of race on the willingness to support redistribution and vote Democratic, suggesting the significance of race in politics. Across states within the USA, the percentage of the population that is black strongly predicts less-generous welfare systems; and across countries, ethnic heterogeneity strongly predicts less redistribution; across countries, redistribution also declines with ethnic heterogeneity (Alesina and Glaeser 2004).
The history of redistribution within the USA points to the important role that race has played at several critical junctures, such as the fight against the Populists in the 1890s (see Woodward 1955). Over the past twenty years, right-wing European politicians including Haidar, LePen, and Pim Fortuyn have all fought the welfare state using rhetoric against ethnically distinct immigrants. Political institutions, like federalism, checks and balances, and proportional representation, also impact the level of redistribution. The literature on majoritarianism and proportional representation (Persson and Tabellini 2003; Miles-Ferretti, Perotti, and Rostagno 2002) suggests that majoritarianism induces governments to cater to the needs of the median voter or voters in swing districts who may or may not be particularly poor. By contrast, proportional representation should at the least focus more attention on marginal classes, who may desire redistribution. In practice, proportional representation is probably more important because it ensures that wellorganized workers’ groups will get representation even in areas where they are a distinct minority. Federalism limits redistribution because out-migration of capital and the wealthy serves as a break on the tendency to redistribute income (Qian and Weingast 1997), except if the government actually wants the rich to flee because the rich represent its electoral opponents (Glaeser and Shleifer 2004). Checks and balances are another force that has tended to restrict the growth of welfare states. While there is little doubt that political institutions do matter for the level of redistribution in society, the greater question is whether these institutions should be taken as first causes or as endogenous factors that reflect deeper social forces (Glaeser et al. 2004).
Political institutions are hardly all that permanent. France acquired proportional representation in 1946, with the Fourth Republic, lost it after the failed right-wing coup that ended that government and put DeGaulle in power, reacquired it under Mitterrand, and then lost it when the socialists realized that proportional representation was helping right-wing extremists. The relative stability of Anglo- American traditions has often led scholars to overstate vastly the degree to which political institutions are durable. I will turn to the causes of institutional change in Section 4 when I compare the USA and Europe..
