The earnings gap between female-dominated and other occupations for career women: a bivariate probit selectivity approach
Women's labor force participation in the U.S. has increased dramatically in the past few decades. There were about 200 percent more working women in 1986 compared to the end of World War II. This increase has resulted in a variety of research about women and their work. The purpose of this study is to estimate career women's eamings gap between female-dominated occupations and other occupations. Most previous studies consider all women in the labor force in a particular year, and estimate the eamings gap using a conventional OLS method.
This study presents four new definitions of career women based upon women's previous working history and their important characteristics related to the labor force participation. Unlike previous studies, a bivariate selective approach is used to estimate the eamings gap between female-dominated and other occupations for specific groups of female labor market participants. Because of a different statistical technique, two important decisions for women who must decide before entering the labor market are considered, namely career-path and occupational-selection decisions. In other words, two separate wage equations for female-dominated and other occupations reflect those two important decisions of why women choose a career path and of why they select female-dominated or other occupations. Because this approach corrects the possible selection bias from the career- path and the occupational-selection decisions, the estimates of wage equations are expected to be consistent population parameters.
The empirical results in this study show that there exist significant wage differentials between female-dominated and other occupations for career women. The estimated eamings gap ranges from 20 % to 48% without controlling for industrial dummy variables. When included industrial dummy variables in wage equations, women in female-dominated occupations earn about 19% to 45% less than women in other occupations, once adjusted for endowment differences. This study concludes that differences in the explanatory variables included in wage equations do not explain a large portion of the eamings gap, confirming the existence of a wage disparity between female-dominated and other occupations for career women.