Statistical Discrimination, Employer Learning, and Employment Di¤erentials by Race, Gender, and Education

Previous papers on testing for statistical discrimination and employer learning require variables that employers do not observe directly, but are observed by researchers or data on employer-provided performance measures. This paper develops a test that does not rely on these speci…c variables. The proposed test can be performed with individual-level cross-section data on employment status, experience, and some variables on which discrimination is based, such as race, gender, and education. This paper shows that if employers statistically discriminate among unexperienced workers, but learn about their productivity over time, then the unemployment rates for discriminated groups will be higher than those for non-discriminated groups at the time of labor market entry and that the unemployment rates for discriminated groups will decline faster than those for non-discriminated groups with experience. Evidence from analysis using the March Current Population Survey for 1977-2010 supports statistical discrimination and employer learning. Keywords: Employer Learning, Statistical Discrimination, Unemployment Rate JEL Classi…cation Number: J71 0I am grateful for support from the WCPC Emerging Poverty Scholar Small Grant.