Unobserved heterogeneity and labor market discrimination

Sexual minorities have historically been subject to many kinds of discrimination. Prejudicial treatment in the labor market could arguably be one of them. Despite that, economic literature has remained mostly silent on the topic. This paper fills that void by leveraging on a novel longitudinal data set that collects detailed information on sexual orientation. I develop an empirical strategy that exploits the fact that sexuality is not a dichotomous trait but rather a wide assortment of sexual preferences. I use empirical models that rely on the identification of unobserved heterogeneity, in the forms of skills and sexual orientation, to allow schooling, employment, and income to be endogenously determined. I find that, after controlling for differences in skills distributions, there are no income gaps against employed homosexuals. However, consistent with the existence of discrimination, homosexuals are 10-20 percentage points less likely to be employed than heterosexuals. These gaps cannot be explained by differences in observable characteristics or skills and are larger among men and the college educated. The results suggest that selection on the employment margin contributes to the elimination of the income gaps as only the highly skilled homosexuals - and thus, higher paid - are employed.