Male-Female Wage Differentials in Taiwan: A Human Capital Approach

Abstract

During the past 3 decades, male-female wage differentials have attracted much attention in all countries for which there are data. The usual approach is to decompose gender wage differentials into two parts: one attributable to an individual's characteristics holding reward structure constant, and the other attributable to a different reward structure holding characteristics constant.' The latter part is usually defined to represent discrimination. Using this approach with data from the 1967 Survey of Economic Opportunity, Ronald Oaxaca found 80% of the observed gender wage differential to be ascribed to labor market discrimination. Mary Corcoran and Greg J. Duncan, using the Panel Study of Income Dynamics, which provides detailed work histories, found a 56% discrimination figure, implying that productivityrelated characteristics accounted for 44% of the observed male-female wage differentials. Because of inherent biases, particularly with regard to measuring human capital, Solomon W. Polachek proposed an alternative way to determine one's characteristics.2 He measured expected human capital, then embedded this measure in a wage regression on pooled male and female data. Over 90% of the male-female earnings gap was explained by gender differences in human capital. Claudia Goldin and Polachek applied Polachek's technique to the 1980 U.S. Census data and accounted for about 80% of the male-female earnings gap.