An empirical contribution to the study of discrimination against women in the labor market.

Summary Women are paid less than men, for all observable characteristics. However, the extent of wage discrimination appears to be lower in France than in the Anglo-Saxon countries. Two factors can potentially explain the relative weakness of wage discrimination in France. First, it may be the result of a corrective effect of the unions. Research on Anglo-Saxon data suggests a moderate effect of unions on wage discrimination. The main reason put forward is the under-representation of women in influential positions in unions. Our estimates on French data suggest that the presence of union representatives in firms does not reduce wage discrimination, since it induces an equivalent increase in male and female wages, given observable characteristics. Secondly, the relative weakness of wage discrimination may be accompanied by significant discrimination in hiring. Indeed, in the presence of restrictive legislation, discrimination may shift to components where it is less easily detected. The results of the controlled experiment we conducted in the French financial sector in 2002 suggest the existence of discrimination in hiring conditional on women's family situation. It would appear that women's relative access to positions of responsibility is more affected by their probability of maternity than by their family responsibilities per se.