This paper considers racial profiling in traffic stops as a public policy problem. Efficiency and equity considerations are characterized. I argue that while there is a strong argument that racial profiling produces a violation of fairness, specifically in the treatment of innocent black motorists, the efficiency effects of profiling are not known. One cannot assign probabilities to the possible magnitudes of either deterrent effects or the harms of profiling to individuals. This makes the assessment of profiling an example of decisionmaking under ambiguity. I defend a notion of a “Fairness Presumption” that requires a policymaker to be able to make an affirmative case if a policy is to be implemented that induces unfairness. On this basis, I reject racial profiling as a policy. Steven N. Durlauf Department of Economics University of Wisconsin 1180 Observatory Drive Madison, WI 53706-1393 [email protected]