We report results from laboratory experiments designed to examine statistical discrimination. Our design expands upon existing research by generating data both on wage contracts and unemployment rates of directly competing worker groups. We find some evidence for statistical wage discrimination against workers having an identical expected productivity but a higher productivity variance. However, those same subjects are less likely to be unemployed, suggesting that our employer‐subjects view hiring choice and wage contracts as substitutable. A clear implication is that field data discrimination estimates based on wages alone may overestimate the true impact of such discrimination. (JEL C90, J71)