Statistical Discrimination in a Competitive Labor Market

48 Pages Posted: 8 Apr 1999 Last revised: 8 May 2000

See all articles by Jonathan Berk

Jonathan Berk

Stanford Graduate School of Business; National Bureau of Economic Research (NBER)

Date Written: January 1999


This paper studies the effect of employee job selection in a model of statistical discrimination in a competitive labor market. In an economy in which there are quality differences between groups, a surprisingly strong condition is required to guarantee discrimination against the worse qualified group --- MLRP must hold. In addition, because of the self-selection bias induced by competition, the resulting discrimination is small when compared to the magnitude of the underlying quality differences between groups. In cases in which the discrimination results because employers' ability to measure qualifications differs from one group to another, the conditions under which one group is discriminated against are much weaker. In general, the group employers know least about is always favored. The economic impact of discrimination that is derived from quality differences between groups is shown to be quite different to the economic impact of discrimination that derives from differences in employer familiarity between groups. In the latter case, for a set of equally qualified employees, it is possible for members of the group that is discriminated against to have higher wages. Finally, we show how the results can be used to explain a number of empirical puzzles that are documented in the literature.

Suggested Citation

Berk, Jonathan B., Statistical Discrimination in a Competitive Labor Market (January 1999). NBER Working Paper No. w6871. Available at SSRN:

Jonathan B. Berk (Contact Author)

Stanford Graduate School of Business ( email )

655 Knight Way
Stanford, CA 94305-5015
United States

National Bureau of Economic Research (NBER)

1050 Massachusetts Avenue
Cambridge, MA 02138
United States

Here is the Coronavirus
related research on SSRN

Paper statistics

Abstract Views
PlumX Metrics