Gun Control

During my respite from mathematics, on vacation in Sun Valley, Idaho, I had the pleasure of visiting the Sun Valley gun club to shoot trap two times. As you might expect, guns are not exactly my "thing": they have a tendency to make me somewhat nervous, I have only fired one on a few occasions, and let's face it, it's a rather difficult to reconcile my personality and other interests with "gun culture" (whatever that means). Nevertheless, even I find shooting trap to be marvelously fun from time to time--I even did surprisingly well, with an accuracy of about 50% the first time and 80% the second.

In spite of this boasting and glee, as I was shattering my clay pigeons under the warm summer sun, I couldn't stop thinking about the profoundly destructive effect that firearms have on Americans. Not only is the US homicide rate higher than any other developed country, but firearms are the fatal weapon in 68% of homicides in this country.

These statistics, of course, lead one to contemplate various forms of gun control. On one extreme, opponents of guns would ban guns entirely. Even if this were constitutionally possible, it would be ineffective and divisive. Such a move would ignore the fact that there are perfectly enjoyable and safe uses of firearms (such as the one in which I partook). Moreover, it would quash a perfecly legitimate "gun culture" that prevails to a great extent in America. But even if we could quash this culture (and shut down all American gun manufacturers), banning guns is likely to create yet another black market, in which bad guys can still get guns but good guys cannot.

On the other extreme, some (mostly anarcho-libertarians) would put very few limitations on gun ownership. The argument is somewhat like mutually assured destruction with nuclear weapons. It goes something like this: if everyone has a firearm, then since guns are lethal weapons, no one will risk using them because their victim might kill them first. Fortunately, this argument doesn't appeal to many people because it is severely flawed. Even though its flaws are obvious, I'll give a few just for fun. If you want more people to have guns to protect themselves, you have to decrease the barriers to gun ownership. At some point, this makes it easier for criminals to get guns. Firstly, this makes law enforcement harder, since it eats away that the government's monopoly on force. But secondly, it may not actually reduced gun crime. Since criminals are the ones who decide whether or not they will be involved in a gun crime, they will have the ability to develop better skills than their victims who do not know when, if ever, they will be the victim of a gun crime. There is thus an information asymmetry which always works in favor of the attacker. It is unreasonable to expect everyone to own and train regularly on their firearm, so many people probably still will not own guns and even fewer would be skilled enough to deter a well-planned attack. Lastly, there will be some people--children, in particular--who we all agree should never own guns, and will be put at a terrible disadvantage. To adopt the MAD analogy once more, Lichtenstein would be in an awfuly bad situation if the rest of the world had nuclear weapons. Since there is no way that Lichtenstein could ever build its own nukes, Switzerland could steal their neighbor's ski resorts and mountain chateaus by just fingering the fatal button.

So where does this leave us? We've ruled out the extremes, so we must look somewhere in the middle. That probably seemed obvious from the very beginning. But that's basically what we're doing today and it does not seem to be working very well. There are currently all sorts of proposals for gun control mechanisms using complicated sensors and tracking devices. This is all probably good, but there is one thing that is not discussed very much:

Assume you are a risk-averse agent. Practically everyone is risk averse, even people who use firearms (Maybe this could be empirically determined, but for now I'm just stating it as a resonable postulate. After all, if they were risk-loving, they would all take a risk and face their victim in hand-to-hand combat!). Let's say that you're a mugger, and you can flash your illegal gun at your victim and it's basically a done deal that he will fork over all of his money. There's some probability of getting caught, in which case you lose all the money that you stole plus a lot more that you had before the crime. Call this scenario A. Alternatively you could just not mug the man, and you would have a 100% chance of keeping your original money. Call this scenario B. If your expected utility of income in scenario A is greater than your utility in scenario B, you will mug the man (non-muggers don't think like this because they feel guilty about breaking the law; they have some utility of law-abiding). Now suppose for a minute that we have two sub-scenarios A-1 and A-2. In A-1, you lose twice as much money as in A-2 if the police catch you. But in A-1, the odds of getting caught are half as great. So the expected loss in money is the same in A-1 and A-2. However, the expected utility for a risk-averse person is actually lower in A-1 than in A-2. So having higher consequences for committing crimes is actually a more effective deterrant than increasing the probability of getting caught. It's conveniently also cheaper for the government. This may seem a bit awkward at first glance. If you don't believe me, take some risk-averse utility function over income, say U(x)=x^(1/2), and compute some examples.

The implication of this is that one way of improving our current enforcement of gun laws is simply to make people who acquire guns illegally pay a ton of money. If you make it miserable enough, then the extra trouble of creating even more ingenious methods of finding law-breakers might not be as necessary. One could, for example, have fines of several hundred thousand dollars for such offenses.

If you are skeptical of the practical applicability of such draconian tactics, then evidence from Scandinavian drunk driving laws appears to support my approach. In some Scandinavian countries, drunk drivers face fines on the order of tens of thousands of dollars, suspension of driving privileges, and occasionally prison sentences for driving with a BAC over their legal limit of 0.02. In the US, the legal limit varies from 0.08 to 0.10 by state. In Sweden, driving with a BAC of 0.10 is considered "aggravated" drunk driving, resulting in a prison sentence of two years. The result? Sweden and Norway have half the number of traffic deaths per mile driven compared with the US, and their fatalities have been decreasing ever since the instatement of the law. Although there are a lot of practical difficulties in comparing drunk driving with guns, I think there is at least a case that the economic principle can be applied as an enforcement mechanism for both problems.