After having lunch with Art Carden today, I decided that the following stuff was sufficiently sensible and interesting to justify a blog post.
The typical economist's argument in favor of punishment for a crime is deterrence. Crimes that cause more damage, other things equal, should be punished more severely. Of course, this is subject to complications due to marginal deterrence. It has been my impression, however, that most economists are uncomfortable with hate crime legislation. Hate crime legislation raises the punishment for offenders who commit crimes for bigoted reasons, or adds additional criminal charges (I am ignoring legislation that makes certain kinds of speech illegal, as I think that is different and much more likely to be a bad idea). Either way, it raises the expected punishment for such crimes. Does this make any sense? Surely what matters is the damage from the crime, rather than reason the criminal committed it, right?
But suppose that, other things equal, a criminal who commits a crime because he or she did not like the victim's race is more difficult to deter. That is, suppose such a criminal is more willing to put up with punishment (either because he or she believes strongly in his or her cause, or because he or she is blinded by rage, or whatever). Then raising the punishment for this sort of crime might make sense--it is harder to deter, so the punishment must be more severe, even if the damage is the same. Legislation might not be necessary to achieve this; judges could simply decide to issue harsher sentences for crimes apparently motivated by bigotry.
I thought this argument was somewhat novel. I see on Econlit that Dhammika Dharmapala has a few papers (coauthored with others) that seem to make similar arguments. I do see a problem with this form of legislation, however, in that it might be simple for criminals to avoid hate crime charges by hiding their racism. Evidence is required to show that a crime was racially motivated; it surely would not be very difficult to conceal this evidence (or never let it come into existence).
Art and I also discussed the difference between punishment for premeditated crimes and "crimes of passion". In general the law penalizes the former more severely than the latter. It seems to us, however, that this is backward. A murderer who carefully considers his crime is much more likely to take into account the possible punishment than someone who acts on the spur of the moment. To get the latter criminal to consider the punishment, the punishment had better be more severe if he acts impulsively. That is, we want to design the law so that it makes people ask "is this a good idea?" before acting. Having lighter punishments for people who act irrationally on the spur of the moment is exactly the opposite of what we want. If, as Bryan Caplan suggests, people have downward-sloping demand curves for irrationality, then if we want them to consume less irrationality, we want to raise the price.
There is some evidence that probability of punishment is more important for deterring young criminals, and severity of punishment is more important deterring adults (Lee and McCrary have a paper on the subject, I think). Perhaps there is a similar distinction for hate crimes; maybe greater probability of punishment would be more effective than more severe punishment. On the other hand, increasing the probability of punishment may be more costly than increasing the severity (depending on whether hiring more police is more costly than housing prisoners for a longer period of time).