On Facebook Bryan Caplan bring up an old post of his in which he discusses statistical discrimination.
Judging everyone as an individual is expensive, and relying on statistical generalizations is a cheap and effective alternative. You don’t clutch your purse when you see a bunch of little old ladies approaching on a deserted street. You don’t offer a policeman a joint. You don’t hire a guy with a mohawk as a receptionist at a law firm – even if he promises to get a hair cut. Why not? Because on average, little old ladies don’t commit violent crimes, policemen arrest people for possession of marijuana, and guys with mohawks have trouble with authority.
Of course, the inevitable existence of some statistical discrimination doesn’t make the practice immune to criticism. You can grant that it’s OK to some degree, but – even if the law is silent – still limited by ethics and/or etiquette. But precisely what limitations do you think are justified, and why?
It seems to me that statistical discrimination arises out of a form of social pollution. They are multiple players in the production of this pollutant but the discriminator is one of them.
Suppose that John is bald. Suppose also that many bald men are extremely irresponsible employees, drink on the job and shout sexual obscenities at female coworkers.
Absent either legal or social pressure employers would then be less likely to hire John than a similarly qualified full headed person. John has suffered because of the poor actions of other bald people. He is the victim of an externality – sort of like pollution.
It seems natural to say that the other bald men should be forced to internalize the externality. That they should bear an extra cost when they behave badly. However, the optimal solution should involve both a reduction in bad behavior on the part of the bald guys and a reduction in discrimination on the part of employers.
Well both parties are enjoying some benefits at John’s expense. The other bald guys are presumably getting enjoyment from their bad behavior and the employer is lowering costs by discriminating against bald men. Together both of these actions lead to bad consequences for John but John had nothing to do with either of them.
Then in a Pigouvian world we would place a tax both on the bad behaving bald men and the employer. We could think about it as first putting a tax on the employer because this is the direct action harming John. However, then the existence of the employer tax means that the bad behaving bald guys are now imposing an externality on the employer which they are taxed for.
However, suppose for reasons of monitoring or general liberty, it is not possible to place a tax on the bald guys. It is still the case that some tax should be placed on the employer.
Again, the employer is taking an action which helps himself – statistical discrimination. However, this action hurts John, even though John did nothing to bring this situation about.
For example, the employer could engage in discrimination abatement, a sort of capture and storage program for snap judgments. He could monitor his employees especially closely after they are hired or conduct a more careful interview.
Now, sharp readers will object that the employer is only imposing a pecuniary externality on John and those don’t count. That is, John could simply get around this by offering his services for less.
However, is that how people really feel? If you are told that you probably won’t be hired at such an such business because you didn’t go to college you might feel upset and you might offer to “prove yourself at a lower wage.” However, the fact in theory you could have gone to college quells your ill feelings.
Now imagine that you are told that you probably won’t be hired because you are a conservative, or a Jew, or bald. The offer to prove yourself is not as readily forthcoming and the rage or sadness is more forthcoming.
For most people, being judged by a criteria which they cannot control is deeply upsetting. For that reason its a real externality and not just a pecuniary one.