Total labour income does not only depend on human capital, but also on physical capital. Yet, many common measures of labour income are directly linked to education. In Discussion Paper No. 1149, Casey Mulligan and Research Fellow Xavier Sala-i-Martin consider alternative measures of human capital that are not so quick to identify it with schooling. They argue that a sensible measure of the aggregate value of human capital is the ratio of total labour income per capita to the wage of a person with zero years of schooling and no experience. To compute the wage of a zero-schooling worker, the individual's average weekly earnings are regressed on their schooling and other characteristics such as experience, sex, race, and marital status. The resulting constant term is taken to as the numéraire, the wage of the worker without schooling or experience.

The advantage of this approach over existing measures is that estimated labour income increases with the quality and relevance of topics taught. Furthermore, it can be tested whether the increase in educational attainment is indicative of an increase in the amount of human capital. The drawbacks are that the computation of this measure does not take account of the fact that wages may change for reasons other than variations in human capital, and that it is assumed that the uneducated are perfect substitutes for the rest of the labour force. The main findings suggest that the proposed measure of human capital is positively correlated with existing measures, but that this correlation is far from perfect.

Measuring Aggregate Human Capital
Casey B Mulligan and Xavier Sala-i-Martin

Discussion Paper No. 1149, March 1995 (IM)