I, like many people, have been thinking about implications of the recent shooting (massacre) in arizona. Both Gabby Giffords and her husband Mark Kelly are alumni members of the US-China Young Leader’s Forum (I believe that’s where they met each other), an organization that I was fortunate to join last year. I haven’t met them personally, but we have multiple friends in common, and the outpouring of love and respect via that group has been both moving and impressive. It looks like she is going to survive and communicate at some level, and hopefully she will continue to recover well.

Separately, I’m a member of an ultrarunning listserv, where I learned that Giffords’ slain aide Gabe Zimmerman was a serious trail runner whose father Ross is a major player in the AZ running community. There had already been discussion of guns on the list because a runner was killed by a hunter on trails at Fort Bragg on new year’s day.

A lot of the media attention has gone to whether or not the tucson attack and similar events can be linked to violent political rhetoric. Seems very unlikely in this case, and probably not in most other cases either. Several people have said that it’s not the violent words or images per se, but rather the climate of hate and fear-of-others that modern politics (as opposed to early politics??) has engendered.

I’m not convinced. I don’t think politicians’ words have the power to directly change behavior to the extent of tangibly affecting the US homicide rate. Not to minimize the arizona shootings, but even 10 or 20 or 50 deaths a year is unfortunately a drop in the bucket here, and it just doesn’t seem plausible that political rhetoric has a [direct] effect even that large. I also don’t mean to minimize the power of political rhetoric: it changes votes and therefore policies, and those policies affect (and sometimes kill) all of us in ways large and small. I wish politicians were more polite, and more substantive, and more truthful, but all of that is a different topic.

So what could actually make a meaningful impact in terms of decreasing these violent deaths? Dealing with mental illness and with gun control. Not only was the AZ gunman fairly clearly mentally ill, but the majority of suicides (which outnumber homicides by almost 2 to 1) and many homicides involve people with mental health trouble. The answer is to support and treat (it works! really) such individuals, rather than to ignore, stigmatize, and/or incarcerate them. My confident guess is that society would pay less for preventive care than we currently suffer due to the after-effects.

Which brings us to gun control. Do more guns lead to more crime or to less crime? Note first that crime is of course not the same as morbidity: even if crime does go down with more guns, it could well be that accidental injuries and deaths (such as the hunting incident linked above) increase more than enough to counteract that effect. Suggestive evidence comes by comparing the US to canada and europe, both of which have much stricter gun policies and lower rates of violent accidents. Note that they also have much lower homicide rates but similar suicide rates to the US, loosely implying that gun control won’t help with suicide (which requires improved mental health treatment, as above) but will help with homicide.

So where’s the evidence, if any? This has been a contentious issue, filled with anecdotes and high emotion on both sides. One scholar, John Lott, has made a solid empirical case that more guns lead to less crime (pdf available online here). An early and careful review of that book by Ayres & Donohue was impressed with it but ultimately unconvinced by the empirical estimation (which is obviously difficult, since there are so many confounding factors). They have continued to battle back and forth in the academic and semi-academic literature.

[As a side-note, I wasn't impressed with Lott's response to current events in the NYTimes, which for instance ignored the fascinating side story that the closest armed citizen (in a state with very lax gun laws, which obviously didn't help here) came extremely close to shooting an innocent man. He doesn't come across as remotely objective. On the flipside, he also seems to be in a battle with Donohue & Levitt regarding their controversial claim linking abortion and crime, and on that topic a colleague of mine (Chris Foote) has convincingly weakened their results, as have later papers. So it seems that Donohue may have some issues as well; hence "semi-academic" above.]

Back to guns and crime, the most recent rigorous empirical study that I could find was Aneja et al. (including Donohue as a coauthor). They cautiously conclude that the evidence is mostly unsatisfactory, which (as an experimentalist) I’m not surprised to hear. They are most confident in stating that right-to-carry laws tended to somewhat increase aggravated assault, with little to no effect on other potentially relevant crimes.

One interesting point comes out of the original Ayres & Donohue review, which is that by necessity all empirical studies look only at marginal effects (small changes), since that’s all we’ve ever seen. Nobody knows what would happen if we outlawed all guns (although here the comparisons to europe are most helpful), which is a completely different situation.

My best guess at this point is that we should probably have much stricter gun laws: no extended clips (this seems like a no-brainer); seriously enforced waiting periods and background checks (which, if they’re working, should rule out non-negligible fractions of the population) even at gun shows; no semi-automatic weapons (basic handguns and hunting rifles seem more than sufficient); etc. The second amendment wasn’t written to include nuclear weapons, and there’s no reason to think that it automatically included high-powered automatic weapons with armor-piercing bullets either. Any gun in a household with a child needs to be unloaded and locked away at all times; violation of this law should lead to revocation of the right (privilege, whatever) to own guns, simply because the potential consequences are so much more dire and are not rare.

That being said, I do believe in general autonomy, agency, and the right to forceful self-defense. [Some readers may be surprised to learn that I have fired various weapons, including a sniper rifle (quite a kick!), and am a former member of the North American Hunting Club - not that I've ever hunted for anything other than easter eggs.] So why not any kind of gun I want on my own property? Because the societal costs as an empirical matter seem to be high, and the extra benefits are small. But this doesn’t affect the underlying right to bear arms. I’m open to the idea of concealed-carry permits, although I’ve yet to be convinced, and in any case certain places like schools, public buildings, and bars should probably be off limits.

The bloviating on both sides of this issue is remarkable, and the desire seems to be to push as far as possible in the fear of losing ground. Just yell “slippery slope” as loudly as you can, and don’t listen to the replies. Maybe that’s the right political strategy for lobbying groups (although I’m not at all sure), but it’s not what a thoughtful attitude would entail. I’m not saying that every thoughtful person will agree with my analysis, but I do think every thoughtful person should acknowledge the tradeoffs involved and arrive at some interior optimum. This means delineating a criterion for where the line should be drawn, rather than simply saying that it should be to the ‘left’ or ‘right’ of where it is now. It also means paying attention to the data, since philosophy (ideology?) is an important input but can’t predict all the relevant idiosyncrasies of human behavior.