What happens to burglary rates when local media publish the names and zip codes of concealed carry holders? According to one new study, a significant (if short-lived) decline in burglaries where there are high rates of carry licenses, and on today’s Bearing Arms Cam & Co we’re taking a look at the study, its results, and whether the findings are worth the invasion of privacy of gun owners.

The study itself comes from researchers at Carnegie Mellon University and MIT who decided to see what happened after the Memphis Commercial-Appeal newspaper published an online, searchable database of concealed carry permit holders in Memphis in late 2008. The NRA and other Second Amendment activists decried the publication at the time; warning that it was giving criminals a shopping list of locations where they were likely to find guns to steal, but also that the release of permit holders names and zip codes could put non-gun owners at risk as well by giving criminals a better idea of who the unarmed residents of a particular neighborhood might be.

According to the researchers, it was the second scenario that actually played out, with burglaries in areas with a lot of concealed carry holders declining while zip codes with fewer permit holders seeing an increase in the crimes.

Open government holds promise of both a more efficient but more accountable and transparent government. It is not clear, however, how transparent information about citizens and their interaction with government, however, affects the welfare of those citizens, and if so in what direction. We investigate this by using as a natural experiment the effect of the online publication of the names and addresses of holders of handgun carry permits on criminals’ propensity to commit burglaries. In December 2008, a Memphis, TN newspaper published a searchable online database of names, zip codes, and ages of Tennessee handgun carry permit holders. We use detailed crime and handgun carry permit data for the city of Memphis to estimate the impact of publicity about the database on burglaries. We find that burglaries increased in zip codes with fewer gun permits, and decreased in those with more gun permits, after the database was publicized.

Specifically, parts of the city with a large number of concealed carry holders saw a roughly 20% decline in burglaries in the weeks following the publication of the online database, while burglaries increased in areas with fewer permit holders.

Does that mean that there’s actually some sort of public benefit to being outed as a concealed carry holder? Possibly, though personally the possibility that a subset of burglars might be scared away from breaking into my home is outweighed by the loss of privacy. And as the study’s authors admit, there are some limitations to the study that are worth keeping in mind.

It is appropriate to point at a number of limitations to qualify our current results. First of all, it is worth noting that the changes in crime we detected came as a result of the publicization of the database. This implies two things. First, our results do not distinguish between the effect of information about gun permit holders per se, and the combined effect of that information and its publicization. Second, they suggest that information alone may not be sufficient to influence criminal behavior: The decrease in crime was much more significant after its publicization. As we wrote in the Introduction, Varian (1996) once wrote that public information becomes “too” public under new information technologies that lower the cost of access. Our results suggest that, even in presence of such lowered costs, catalytic events are needed to raise the public’s awareness of those data. Furthermore, seasonality (as well as countless other factors) may have affected the general crime trends in the Memphis area.

Second, and more pertinently to our analysis, we should stress that our econometric approach was aimed at teasing out differential patterns in the number of crimes — and, specifically, burglaries — across zip codes with different numbers of guns. Therefore, it should not be concluded, based only on these data, that an overall increase in gun permits will correspond to overall net decreases in crime, since – in an arms-race dynamic – even zip codes with an absolutely high number of permits may still be perceived as less dangerous by criminals, relative to zip codes with even higher numbers of permits.

Third, it is open to debate whether a privacy cost had to be actually paid for the decrease in crimes following the publicization of the database. On the one hand, one may conjecture that an anonymous publication of gun permits statistics – such as the number of permits in a given zip code – may have achieved similar results, without invading individual holders’ privacy. On the other hand, part of the appeal of the database, and the reason why it drew significant traffic, was arguably the fact that it included individual permit holders’ names. Fourth, the kind of crime data that we have access to does not contain information about the specifics of the crime. Therefore, we cannot distinguish between, say, burglaries where many things were stolen and burglaries where little was stolen.

Finally, the location where the gun permits publication occurred has several state-specific features that may make the findings non-generalizable. Unlike in most states, in Tennessee there is no requirement to conceal a weapon. This, in theory, should mute the effects that we study, and may explain why we observe a reduction in burglaries but not in person-to-person crimes. Tennessee, and Memphis in particular, also have many gun permits (as high as one gun permit for every three Memphis dwellings). It is not clear whether criminals could be deterred by publicly available information on gun permits if there were fewer gun permit holders. The Commercial Appeal redacted street address information, making their information available at the zip code level. It is not clear whether the pattern of a zip-code-wide reduction in crimes would hold if gun permit street addresses were also available.

It’s an interesting study but it’s hardly the last word on the issue. I know that there’ve been several other papers over the years who’ve done much the same thing, including the Cleveland Plain-Dealer, and I’d be curious to know if any similar impact could be detected in northern Ohio.

This is also becoming more of a moot point in many states with the adoption of Constitutional Carry, which remove the mandate that legal gun owners get a permission slip from the state before they can lawfully bear arms. Gun owner privacy isn’t generally a major talking point for pro-Constitutional Carry activists, but maybe it should be given the proclivity of anti-gun media outlets to shine a bright spotlight on those hoping to discreetly bear arms in self-defense.