Incorporating Occupational Attainment in Studies of Male-Female Earnings Differentials

Abstract

Differences in the occupational distributions of men and women represent an important source of male-female differentials. Yet most studies of wage differences concentrate only on either occupational segregation or the problem of unequal pay for equal work. This study refines and merges these two approaches, allowing for variation both in occupational distribution and in wages resulting from differences in job qualifications and productivity indicators. Our approach uses multinomial logit analysis to predict occupational attainment for men from a set of personal characteristics and then simulates an occupational distribution for women. Wages are estimated as a function of productivity measures for both men and women within each occupation. Our model finds that only 14 to 17 percent of the total wage differential is attributable to differences in endowments and that more of the unexplained difference results from within-rather than across-broad occupational categories.