Ethnic profiling seems to be a hot topic again following the terrorist attacks on London. Heated debates between proponents and opponents can be found on numerous cable news programs. The arguments are loud and furious, but generally poorly made.

Ethnic profiling can be viewed as an empirical question. Does it increase or decrease the cost of enforcement? There should be no doubt that profiling in the short-run can decrease the cost of enforcement. Concentrating your resources on a group that has a higher propensity to commit a crime must lower the number of crimes. The counter argument is that if you target a specific ethnic group, the targeted group will take advantage of that profiling by assuming a different disguise. That is probably correct, but to do so is costly. To the extent that we can increase the cost of terrorism, we make it less likely.

What the cost of enforcement does not take into account are the costs imposed on the innocent in the targeted group. To be stopped and searched imposes material and psychological costs on the individual being searched. The real question is whether the reduced cost of enforcement that is shared by the wider society outweighs the cost that is concentrated on the specific group. Welfare is transferred from the targeted group to society as a whole. Unless the targeted group also perceives a net benefit from profiling, (the added safety offsets the added cost) they are likely to be against profiling.

Assuming that profiling results in a net benefit to society as a whole, a side payment from the non-targeted group to the targeted group may gain their compliance. For example, suppose the authorities pay each individual that was searched $20 out of general revenue. This could be viewed as the price of the information gained from the search. If the cost imposed on an individual being searched was less than $20, then they would receive a net gain in welfare. Moreover, if reduction in enforcement costs and the added safety to society was greater than $20 from a profiled search, then the non-targeted group is also better off. Thus, profiling would result in a Pareto optimal move. An unintended consequence is that some individuals in the non-targeted group may take on the demeanor of those in the targeted group with the hope of getting searched and collecting $20. This would not only be a waste of $20, it would also increase the cost of enforcement.

Isn’t it enough for profiling to be Hicks-Kaldor efficient? We estimate that the marginal benefit exceeds the marginal cost, but a side payment is not made. The problem is that without the side payment the targeted group feels discriminated against. This is likely to affect the long-run cost of enforcement. The long-run cost of enforcement will depend on the assistance of the targeted group. The targeted group can increase the cost of a search through non-compliance and may even be swayed to the side of the terrorist.