Testing for Statistical Discrimination Based on Gender

Abstract Gender wage gaps are a prevailing feature of the labour market. Statistical discrimination has been highlighted as a potential source. However, the evidence is scarce. The employer learning literature provides a testable framework. This paper develops an employer learning model which incorporates screening discrimination, stereotyping and prejudiced beliefs. Implications of screening discrimination are found not to be consistent with wage dynamics in the Danish labour market. Stereotyping with prejudiced beliefs is argued to be a more plausible candidate. Workplace characteristics, such as the fraction of women in high‐ranking positions, do not affect the level of screening discrimination by gender.