Gender differences in competitive positions: Experimental evidence on job promotion

This paper analyzes gender differences in access to competitive positions. We implement an experiment where workers can apply for a job promotion by sending a signal to their employer. We control for gender differences in anticipation of discrimination in a treatment where a computer randomly recruits. Discriminatory behavior by the employer is isolated in a treatment where workers cannot send signals. We find that gender disparity among promoted workers is highest when workers can apply for promotion and employers recruit. Strikingly, the gender composition in competitive position is balanced in the absence of a signaling institution. When signaling is possible, we observe that female workers who do not request a promotion are discriminated against.