Gender and ethnic discrimination in hiring : evidence from field experiments in the German labor market

The present thesis investigates the prevalence of and the reasons for hiring discrimination against women and ethnic Turks in the German labor market. Subsequent to a discussion of how to reveal discrimination, the literature on wage and employment differences inside and outside the German labor market is reviewed. Afterwards, different (economic) theories explaining inequalities in labor markets are presented. In the empirical analyses a field experiment - the so called correspondence testing - is conducted where matched pairs of (fictitious) male and female as well as German-named and Turkish-named applicants respond to, respectively, 656 and 608 (real) apprenticeship offers in predominantly male-dominated jobs. Descriptive results and econometric analyses using probit regressions on various model specifications indicate that the female applicant has a 19 percent lower callback probability compared to her male counterpart. However, differential treatment is both job- and firm-type driven. While callback rates are not statistically different from zero in female-dominated and “gender-neutral” occupations, they prevail in jobs where men are overrepresented. Furthermore, discrimination is restricted to late recruiters, i.e., companies that advertise their vacancies right before the apprenticeship is supposed to start. Similar conclusions can be drawn from the study investigating ethnic discrimination. The 32 percent lower callback probability of the Turkish-named applicant decreases if early rather than late recruiters are addressed. Apart from that, comparing response and callback rates to the candidates using different experimental designs, i.e., sending out single versus pairs of applications, yields no statistically significant differences demonstrating the unbiasedness of the correspondence approach.