The Gender Wage Gap in Bulgaria: A Semiparametric Estimation of Discrimination

Abstract

Bulgaria's transition to a market economy has coincided with a large increase in wage inequality. Given the emphasis on wage leveling in pre-transition Bulgaria, the rise in wage inequality may be due to managers rewarding more productive workers; or it may be the result of rewarding non-economic characteristics such as gender. Using data from the 1995, nationally representative Bulgaria Integrated Household Survey, I examine whether gender discrimination is an important factor determining the gap in wages between men and women and the extent to which gender discrimination affects wage inequality. I model wage determination with a correction for sample selection as a Type III Tobit and estimate this model with the Honore et al. (1997) semiparametric estimator. Unlike the classic Heckman correction for sample selection, this estimator is consistent in the presence of heteroscedasticity. I bootstrap to estimate standard errors. Using separate wage regression estimates for men and women, an Oaxaca decomposition indicates that women's wages are 25 percent lower than men's wages and 85 percent of this differential is due to discrimination, or more precisely, due to differences in how men and women are rewarded for the same characteristics.