Ability, Sorting and Wage Inequality

Wage inequality has increased dramatically over the last 30 to 40 years in the US and in the UK (see, for example, Autor and Katz, 1999). Economists have devoted a substantial amount of work to the documentation and interpretation of the evolution of wage inequality and several of its dimensions, such as the skill premium, or racial and gender wage gaps. One variable whose role is often stressed is schooling. Schooling is an important determinant of wages and therefore it is likely to substantially affect inequality. In the large empirical literature on the effect of schooling on wages, usually labeled the return to schooling, it is emphasized that this parameter varies across individuals and that individuals sort themselves non-randomly across schooling levels (see, for example, Card, 1999, and Carneiro, Heckman and Vytlacil, 2003). As a result, considerable care is needed both in the estimation of such returns and in the interpretation of standard estimates. In contrast, in standard studies of inequality, such as Katz and Murphy (1992) or Card and Lemieux (2001), the role of heterogeneity in returns and self-selection into schooling is often ignored. On the other hand, the role of heterogeneity and self selection into employment has been seriously considered in the study of inequality. For example, Heckman and Todd (2003) show that accounting for differential labor force participation rates of blacks and whites dramatically reduces black-white wage convergence in the last 40 years, because of selective dropout of blacks from the labor force. ∗We thank participants in seminars at the 2005 AEA meetings in Philadelphia, Norwegian School of Economics, Kent, Maynooth and Ente Einaudi for useful comments.