Income distribution varies considerably across countries; it tends to become more equal with development in some countries, but just the opposite occurs in other countries. This paper provides a theoretical investigation of the persistent differences in income distribution across countries over time. Motivated by the relationship between income distribution and public spending at different school levels for a broad range of countries over the past 30 years, the analysis centers on the role of public education where specific investments interact with political involvement by different socio-economic groups. Socio-economic groups may form lobbies to influence education policy making. The formation of lobbies is endogenous. Persistent inequality is caused by persistent lobbying efforts of the wealthy that lead to an allocation of public education spending more biased toward them.
Persistent differences in income distribution across countries are a well recognized phenomenon. An existing strand of research focuses on the role of redistributive policies, in particular, public education policy. Theoretical analysis of the relationship between income distribution and total public education spending, however, is in general not consistent with empirical findings. This paper seeks to develop a theory of income distribution dynamics through the mechanism of public education policy that is consistent with empirical observations. Its point of departure is a more realistic view of public education: public education is not a single public good; on the contrary, different school levels appear to be different public goods. Public spending at different school levels exhibits distinct relationships with income distribution, and they tend to benefit different socio-economic groups. As a result, socio-economic groups have an incentive to influence the allocation of public education spending in their favor, and the resulting policies tend to preserve initial income distributions.
I formalize these ideas in a simple general equilibrium political economy model with overlapping generations, where an incumbent government determines public education spending on different socio-economic groups, and current education determines future income. The government might however be influenced by socio-economic groups. I model this political influence as lobbying by socio-economic groups in a common agency framework à la Bernheim and Whinston (1986). The uniqueness of the present model is that socio-economic groups' political influence is not assumed; rather, whether a socio-economic group will engage in lobbying is endogenously determined by its own utility maximization calculation, and is crucially dependent on current income distribution, the government’s valuation of social welfare relative to political contribution, and the cost of entering the political competition. In equilibrium, different sets of lobbies may form, but the rich are invariably more likely to lobby. When different lobbies form, public education policies will be different, and over time, income distributions will evolve on different paths. In the long run, all economies may reach one unique steady state income distribution regardless of the initial distributions, or economies with different initial income distributions may reach different steady state distributions. In the latter case, we observe persistent equality or persistent inequality in different economies.
The income distribution dynamics generated by this model are in sharp contrast to that implied by the traditional median-voter model (see, for example, Meltzer and Richard, 1981, Alesina and Rodric, 1994 and Persson and Tabellini, 1994). In the median-voter model, greater inequality translates into more redistributive policies; it implies that income distribution converges to one steady state in the long run. In particular, Glomm and Ravikumar (1992) and Saint-Paul and Verdier (1993) show that more unequal economy tends to support public education, and continued support of public education lowers income inequality over time. Empirical studies such as Perotti (1996) and studies reviewed in Bénabou (1996), however, find no relationship between inequality and aggregate public education spending.2
This paper provides a political economy explanation of the persistent differences in income distribution across countries over time. In contrast to the previous works that focus on total public education spending, the mechanism here is the allocation of public education spending on different schooling levels. Empirical regularities suggest that public spending on different schooling levels has distinct relationships with income distribution and is in general targeted at different socio-economic groups. Socio-economic groups may form lobbies to influence education policy making. The formation of lobbies is endogenous. Different lobbies may form in economies of different income distributions, but the rich are invariably more likely to lobby. When different lobbies form, public education policies will be different; over time, income distributions will evolve along different paths. In the long run, all economies may reach one unique steady state income distribution, or economies with different initial income distributions may reach different steady state distributions. In the latter case, we observe persistent equality or persistent inequality for different economies.
