[toyfpgzm] Paradoxical effectiveness of racial profiling of criminals

If the police can detect using racial profiling a certain class of criminals with accuracy higher than random, and the police do do racial profiling, then one would expect people of those races to preferentially not enter that class of crime, and people of races not targeted by racial profiling to preferentially enter, or survive, in that class of crime.  Negative feedback.

Is this happening?  It generally seems not to be, meaning there are more powerful forces at work.

The canonical example is African-Americans and drug dealing.

The standard narrative of those more powerful forces is that people of certain races face such tremendous barriers to entry to other professions due to race that, despite racial profiling, that class of crime still remains the choice with maximum expected utility.