This article explores the basic assumption of statistical discrimination th
eory which holds that women and minorities earn lower wages because they on
average, have lower productivity. Employer exploitation of women and minor
ities and social closure by advantaged employees are advanced as alternativ
e explanations for the lower wag es of women and minorities. The authors fi
rst demonstrate that there are substantial gender and racial wage penalties
net of human capital for a sample of employees. The primary analysis focus
es on the sample of private-for-profit establishments in which these indivi
duals are employed Establishment productivity as well as aggregate salaries
and wages and profits are regressed on the sex and race composition of the
establishment with other factors that may influence establishment producti
vity. Findings show that neither the sex nor race compositions of the workp
lace are associated with productivity. The authors Interpret the results to
be most consistent with a social closure account of gender and racial earn
ings inequality.