Sorting, Education, and Inequality

New York University, New York NY

Investigators

Abstract

Many social commentators have expressed concern that increased sorting (or assortative matching) in the US may lead to undesirable outcomes. Although the evidence favoring increased sorting is mixed, increased sorting can theoretically take place in many spheres. Communities, for example, will be more stratified along educational and income lines if individuals increasingly choose to live in neighborhoods with others of more similar levels of education or income. Workplaces will be more homogeneous if the organization of production leads to grouping high-skilled workers with others like them rather than with low-skilled workers. Marriages will have a higher degree of assortative matching if individuals are increasingly prone to marry others similar to themselves in income and education. Lastly, increased tracking in schools (and the potentially increased homogeneity of neighborhoods) may lead to less heterogeneity in the classroom. This project examines both theoretically and quantitatively diverse relationships that exist among education, sorting, inequality and economic development. A benchmark model is developed that allows the study of the interplay among these variables. In particular, the model analyzes the relationship between marital sorting and education and wage inequality. The model is extended to make it more useful to take to the data. These extensions include: an increase in the feasible levels of education; families of the same type with different numbers of children; different matching assumptions; and cross-country comparisons. Alternative models are pursued that focus on borrowing constraints; endogenous fertility; endogenous sorting; and physical as well as human capital accumulation. Why may increased sorting matter? Some individuals worry that less interaction among different types of individuals will lead to a decrease in empathy and result in less redistribution. Others have argued that greater stratification or decreased peer effects or in schooling, neighborhoods or workplaces can lead to efficiency losses. Alternatively, increased sorting in marriages is likely to increase inequality in the distribution of income or education to the extent that there is an intergenerational family link in these variables. Is increased marital or residential sorting likely to have quantitatively significant effects on the distribution of income and education? This is a question that has received relatively little attention. A recent and very provocative paper by Michael Kremer (1997), however, claims that even very large increases in sorting-whether marital or residential-are unlikely to have important effects on inequality. This is an important challenge to many people's beliefs and some of the work in this proposal is an attempt to meet this challenge.