Following Becker's (1993) suggestion that rests for discrimination sho
uld attempt to infer whether profits differ for products sold to minor
ities and nonminorities, this article tests :he hypothesis that racial
discrimination affects market prices of auto insurance in Missouri. C
ompared with tests for discrimination in lending markets, our results
are less susceptible to bias from omitted variables. Controlling for a
vailable demographic and coverage-related factors, we do not find that
loss ratios at the zip-code level are negatively related to percent m
inority population. This finding is inconsistent with the hypothesis t
hat racial discrimination increases premiums relative to expected clai
m costs for minorities.