Discrimination in Hiring: Evidence from Retail Sales.

This article develops and empirically tests a unified model of racial bias in hiring that integrates three major theories: taste-based discrimination (managers lower hiring standards for same-race applicants), screening discrimination (managers have more accurate information about same-race applicants), and complementary production (same-race worker-manager pairs yield higher productivity). Using administrative data from a large U.S. retailer covering 48,755 newly hired commission-based salespeople and 7,892 managers across 997 stores, the study finds that managers are significantly more likely to hire workers of their own race. Productivity analyses reveal that screening discrimination dominates black–Hispanic hiring patterns, a combination of screening discrimination and complementary production characterizes white–Hispanic pairs, and all three forms of discrimination appear to offset each other in white–black pairs. The authors propose statistical tests based on the supermodularity of mean productivity and variance to distinguish these discrimination types, highlighting implications for firm policies aimed at reducing racial disparities in employment.