A Case Study on Racial Profiling
In the actuarial field, the case of racial profiling on the highways has received perhaps the most attention from social scientists, legal scholars, and public policymakers. As a result, it is the subfield with the most abundant empirical data and economic models. It seems only fair to ask, at this point, what the empirical evidence shows: Does racial profiling on the roads reduce the overall incidence of drug possession and drug trafficking on the nation's highways? Does it produce a ratchet effect on the profiled population? And does it distort our shared conceptions of just punishment?
Unfortunately, as noted in chapter 3, the new data on police searches from across the country do not provide reliable observations on the key quantities of interest necessary to answer these questions precisely. Specifically, the data do not contain measures of comparative offending or elasticity, nor of natural offending rates within different racial groups. Nevertheless, it is possible to make reasonable conjectures based on both the best available evidence and conservative assumptions about comparative offending rates and elasticities. Let's proceed, then, cautiously.
The term offending ratecan be defined in several ways. First, it can refer to the rate of actual offending in the different racial groups, given the existing distribution of police searches. This I will call the “real offending rate.” It is calculated by dividing the total number of members of a racial group on the road who are carrying drug contraband by the total number