Crime minimization and racial bias: what can we learn from police search data?

Is variation by motorist race in the success rate of searches informative about racial bias if police are motivated by crime minimisation rather than success-rate maximisation? We show that the basic idea of extracting information from 'hit rates' may still be valid, provided one can verify some simple restrictions on the joint distribution of criminality by race. We also extend these results to the case where the police minimise the rate of unpunished crime