[Part I; Part II]

Yesterday I provided a glimpse at a study claiming that, while most forms of gun control are ineffective, requiring a license to own a gun does seem to reduce crime. The authors suggested that a key element of these laws might be that they require background checks on prospective gun owners—something that gun-control advocates have been pushing as a policy in itself in recent years, at both the federal and the state level. Today I’ll look at the issue, with a special focus on a new analysis the economist John R. Lott presents in his book The War on Guns.

When talking about background checks, liberals love to bring up Missouri and Connecticut, both of which have experimented with “permit-to-purchase” handgun laws. Under these laws, anyone seeking to buy a handgun must first acquire a permit from a law-enforcement agency, where a background check (and often an interview) are conducted. Unlike the federal background-check requirement, these laws apply to private sales between individuals, not just purchases from licensed dealers.

In the left’s telling, the law was a spectacular success in both states. The reality is more complicated.

The liberal narrative holds best—incredibly well, actually—in Missouri. When the state repealed its permit-to-purchase law in 2007, the homicide rate immediately rose, driven by an increase in gun homicides. At the exact same time, trace data from the Bureau of Alcohol, Tobacco, Firearms and Explosives indicated that a higher percentage of Missouri crime guns were originally purchased within the state.

The data are visually striking. You can see the ATF numbers in this article, and in the chart below I’ve divided Missouri’s rates of gun and non-gun homicide by the corresponding national rates. (This allows us to see what trends unfolded in Missouri that didn’t occur nationwide.)

mochart

But Connecticut is a very different story. It enacted a permit-to-purchase law in 1995, and … not a whole lot happened.

ctchart

To deny this simple fact rooted in readily available data, liberals rely on a poorly conducted 2015 study that credited the law with a whopping 40 percent reduction in gun homicides. The study cut the data off in 2005 and, rather than using national or regional trends as its primary baseline, compared Connecticut’s crime trends with those of a “synthetic state”—a statistical abstraction mostly based on Rhode Island, which experienced a crime wave shortly after Connecticut’s law went into effect.

Enter Lott. In his book, he does spend some time going over the Missouri and Connecticut cases. (His assessment of Connecticut is the same as mine, but he reads the Missouri trends much differently than I do.) But far more importantly, he notes that numerous other states have enacted various background-check laws too, and he conducts analyses on all of the states at once, using data that run from 2000 to the most recent (2013 to 2015, depending on the variable). This approach avoids cherry-picking, and he helpfully provides these laws’ enactment dates in an appendix for anyone who would like to explore the data themselves.

As Lott notes, he actually studied the effects of background-check laws on the murder rate years ago, finding no significant impact. In his new book, he looks at a variety of other outcomes, including suicide rates, killings of women, and mass shootings, again finding no effect. His models control for things like guns coming in from other states (which liberals say undermine state-level background-check laws), demographics, and divorce rates. He also runs separate analyses for background checks on “at least some private transfers” vs. “universal background checks.”

This isn’t the final answer on this question. There’s always a different way to perform these kinds of studies, often leading to different results. (I would have liked to see earlier data included, for example, because most of the states with these laws enacted them before 2000.) But this is a good step toward building a real literature about the effects of state background-check laws, one where researchers refrain from cherry-picking states that fit—or can be made to fit—their preferred narrative. This is especially important because several states have enacted these laws just in the past few years, and a steady flow of new data can enhance our understanding if analyzed properly.

One last note: I’m not sure this research says a whole lot about what a federal background-check law would do. Lott does control for interstate gun trafficking, but if such trafficking renders state-level laws ineffective, there is no effect left for any study to find, save for the rise in interstate trafficking itself. I suspect that state-level laws mainly drive criminals to bring in guns from other states with weaker laws, while a federal law might make it (however marginally) more difficult for them to get guns at all.

As I’ve said repeatedly in the past, I’m not nearly as gung-ho about universal background checks as many liberals are, but such a policy is probably the gun-control idea most likely to work.

Robert VerBruggen is managing editor of The American Conservative.