Wages, Employment, and Statistical Discrimination: Evidence from the Laboratory

When membership in a particular group conveys valuable information about an individual’s skills, productivity, or other human capital characteristics, a non-prejudiced agent may still find it rational to statistically discriminate. We frame statistical discrimination in a labor market setting for a series of laboratory experiments. A main objective of our experiments is to examine how varying productivity risk along several dimensions impacts outcomes across worker groups. Our design expands upon existing research by generating laboratory data both on wage contracts and unemployment rates of directly competing worker groups. We find some evidence for statistical wage discrimination against workers with identical expected productivity but higher productivity variance. However, those same subjects are less likely to be unemployed, suggesting that our employers view hiring choice and wage contracts as substitutable. These laboratory results have interesting implications for labor markets where employers select from workers belonging to distinct statistical groups, and suggest that statistical discrimination based on wages alone may overestimate the true effect of such discrimination.