Substitution Effects and Racial Profiling

Regardless of whether you think racial profiling is a form of evil discrimination or a necessary part of security, be sure to read Posner's latest entry on this topic. His main point is that racial profiling will cause substitution effects, which he expects to be quite large: if one race is targeted, then criminals and terrorists of other races will emerge as successful groups. If this empirical assumption is correct, then racial profiling will have little effect on crime rates or national security. My own priors are that the substitution effects are smaller than he assumes.