TTAG reader Bryan Peters [not shown] writes:
Back in the 90s, when everyone else was out participating in the greatest job market in history, I figured it would be a good idea to go to grad school to learn economics. While I was there, the department invited John Lott to present a working paper he and David Mustard authored, entitled, “Crime, Deterrence, and Right-to-Carry Concealed Handguns.” This was the paper that changed the academic discussion about the effect on crime of firearms held and used by the general public. Lott and Mustard found that . . .
concealed carry decreased violent crime and did not increase accidental deaths, a result that proved to be extremely controversial for, well, ever since. In general, some papers have found results that confirm the hypothesis, while others don’t find evidence of either harm or benefit from the laws.
In one amusing side note (if you’re in to extremely snarky and personal academic disputes), one of the people who attacked Lott’s hypothesis was Steven Levitt, who famously advanced his own crime reduction hypothesis: legalized abortion was pre-emptively killing future criminals.
That Levitt’s study suffered from serious issues of data and methodology made those fights all the more interesting. Levitt defended his results, of course, and counter-attacked, only to be counter-counter-attacked by Lott, etc. And so it goes: academic work is not at all less political and polarized than the rest of our country.
But this is a little story of the origins of this debate and how that debate was settled…partly settled, anyway.
When the Economics Department invited Lott to discuss his paper, they also hosted a roundtable discussion with the Sociology Department. Now, that was fun. The discussion with just the economists was all, “Did you control for the simultaneous heteroskedastic moments of the….zzzzzz,” and detailed discussions of the modeling process. Boring science stuff.
As readers may know, the core of Lott and Mustard’s argument was that concealed carry, by raising the average cost of predation (getting shot being a very high cost of doing business) might reduce the overall level of risk for the public, whether any particular person was carrying or not. And the numbers that Lott and Mustard got supported the idea: higher likelihood of potential victims carrying a gun reduces the number of crimes against all potential victims.
Pretty straightforward economic analysis, really: when the cost something goes up (in this case, crime), you get less of it.
But the discussion with the sociologists wasn’t like that boring science stuff at all. In response to Lott and Mustard’s empirical paper, the sociologist responded with – I am not making this up – a hypothetical story about a mugging that might go wrong because the mugger would decide to shoot you when you reach to give him your wallet, thinking you were instead reaching for your concealed weapon.
And that was the state of the debate in the late 90s. In fact, much of the academic work preceding Lott and Mustard on gun control, and concealed carry in particular, was simply comparisons of average crime rates across different locales.
Those papers wouldn’t even attempt to control for any of the many other factors that might be expected to be related to varying crime rates. Factors like he probability of arrest and conviction, incomes, the length of sentences, or anything else. In short, most of the prior work was really awful science, and when Lott and Mustard began looking at the question with the tools available to economists, it was a genuinely new analysis.
And against this, the sociology department at a major university offered up a scary story about how a hypothetical somebody might hypothetically act. Awesome.
In the nearly 20 years since that happened, the academic work has advanced…not much at all. To be fair, there were a few pieces of analysis more along the lines of the techniques Lott and Mustard used and mostly published in economics journals. As you might expect, a furious responses and counter-responses lasted for a few years.
Since then, it appears that the science is divided into those studies that find increases in gun availability – especially concealed weapons – lead to lower (particularly violent) crime, and those studies finding no effect either way. It’s these latter studies that should clinch the political case for relaxed gun laws.
Here is why: to take a fairly typical “no-result” paper, when a study refutes the Lott hypothesis (which, simplified, is: more guns means less crime) it will tend to find that there is no effect on crime. It is very unusual to find any papers which find a result opposite of that hypothesis. That is, few studies find more guns mean more crime.
Think about what that means. Either:
More guns do not cause more crime, or else
More guns would cause more crime, except that potential victims have armed themselves, thus effectively deterring the crime.
Now, whichever of these is more likely (hint: it’s B), it should serve as the end of the so-called “reality-based” arguments against guns. There simply isn’t any reason to believe removing guns will reduce crime.
In the first case, guns are entirely neutral on crime (a point the anti-gun left will never admit, true or not) which means reducing guns is pointless. On the other hand, if there is an equilibrium between criminals and non-criminals, then there is just as much justification in the data for increasing guns as there is for decreasing them.
In some ways, then, we are right back to the late 90s, despite all the academic work. The great majority of the analysis supports the idea that guns (at a minimum) don’t increase crime, and the anti-gun types are back to arguing with anecdotes and hypothetical stories. Only now, instead of stories about how a guy will shoot you because of concealed carry, we have stories about how some sort of magic law would prevent lunatics from killing their moms and stealing her guns.
Which brings us back to a discussion group the sociology guys ran after Lott’s talk back in the late 90s. One of the professors gathered some of the students afterward to interview them in a group setting about their views of guns after the presentation. For whatever reason, he asked me to attend, too.
In the group, I brought out the point that Lott’s argument was entirely empirical and that, therefore, the only valid responses would also have to be empirical; yet we spent time refuting them with made-up stories. It wasn’t what students at a university should be doing, I said.
“Why do you suppose the students were arguing against Lott’s paper, then?” the sociologist asked me.
“Because it means you’re pure.”