Recently I interviewed criminologist Gary Kleck about gun ownership and crime. In the course of that interview, Kleck expressed skepticism regarding economist John Lott‘s claims that expanded concealed carry of handguns reduces crime. I asked Lott if he wanted to reply, and he graciously agreed to do so. He also addressed a number of other questions regarding gun ownership and crime.
I think there’s still much work to be done to square all of the seemingly conflicting claims regarding guns and crime in the United States and across different regions in and out of the country. (Perhaps I’ll look more deeply into such issues in the future.) However, Lott’s remarks provide some excellent leads. Below are his unedited replies to my questions. —Ari Armstrong
December 23 Update: See also “Gary Kleck and John Lott Offer Closing Thoughts in Dispute over Gun Research.”
Ari Armstrong: Americans who want to regulate guns more heavily (or even ban guns) often invoke international comparisons. Rates of violent crime generally, and rates of gun-related violent crime specifically, are higher in the United States than in Canada, Australia, and various western European nations. What do you think is the significance of such international comparisons?
John Lott: We can learn something from international comparisons, but it is important to recognize that crime rates vary across countries for a large number of reasons. A common comparison is to look at the low homicide rate in the UK compared to the US and assume that is just because of the UK’s strict gun control laws. But after the January 1997 ban on handguns, the UK’s homicide rate actually rose by 50 percent over the next eight years. It only declined after that back down to its pre-ban rates after an 18 percent increase in the number of police.
Indeed, the same thing happens elsewhere. In every country in the world that has banned guns, murder rates have gone up.
The UK has lower homicide rates than the US, but their homicide rate actually rose relative to the US after they had a gun ban. There is something else besides gun ownership that explains the difference between the two countries.
Australia’s crime rates also haven’t changed the way that gun control advocates would have predicted. The buyback in 1996 and 1997 resulted in more than 1 million firearms being turned in and destroyed. This reduced the number of guns in the country from 3.2 million to about 2.2 million guns. Since then, however, there has been a steady increase in the number of privately owned guns. In 2010, private gun ownership was back to 1996 levels.
Their firearms homicide rate had been falling for a decade prior to the buyback. It continued falling at the same rate after the buyback. There was no sudden drop, just a fairly constant decline that continued even as gun ownership rose back up to previous levels.
Armstrong: What does your phrase, “more guns, less crime,” summarize with respect to concealed carry? Is your claim that liberalized concealed carry laws led to people owning more guns for defensive purposes, carrying guns more frequently, or both? I know of people in Colorado who purchased handguns because of the liberalized carry laws and who started sometimes carrying a concealed gun because of the laws, but I don’t know how typical such practices are. What’s the best estimate of the effects of the laws in terms of gun ownership and carry?
Lott: Just as law enforcement deters crime with higher arrest or conviction rates or longer prison sentences, the fact that victims can defend themselves also makes committing crime riskier. This point applies not only to guns in the home but also to concealed handgun permits.
The number of concealed handgun permits has exploded in recent years—rising from 4.6 million in 2007 to at least 12.8 million this year. This underestimates the increase because the number of states where permits are required in all or virtually all of a state has increased from 5 to 10.
Armstrong: As you’re aware, criminologist Gary Kleck has criticized both aspects of your claim, “more guns, less crime.” He has said that liberalized concealed carry laws didn’t actually result in more people owning guns or even carrying them concealed more frequently. He has also said that there was no discernible drop in crime resulting from the laws. I know this debate is enormously complex, but can you summarize some of the main evidence here?
Lott: Gary and I have a very different view on how people behave. Economists believe that there is something called the law of demand: As something becomes less costly people do more of it. That applies to people buying more apples as the price falls and to getting more concealed handgun permits. Indeed, there is a lot of evidence that the number of permits increases as the cost of getting them in terms of either fees or training costs goes down.
Gary, like sociologists generally, doesn’t believe that prices alter people’s behavior. I don’t think that the evidence supports that view and it doesn’t make a lot of sense to me. I also don’t understand why Gary and other sociologist don’t believe that police deter crime.
That said, Gary claims that while the number of concealed handgun permits has soared from 4.6 to 13 million over the period from 2007 to 2015, no more people are legally carrying guns than they did previously.
Clearly not everyone who gets a concealed handgun permit carries their gun. But unless the percentage of people with permits who do carry has fallen dramatically, the number of people who carry must have increased dramatically over the last 20 years. Just between 2007 and earlier this year, the number of people with concealed handgun permits tripled from 4.6 to 13 million, and that doesn’t take into account the rapidly growing number of states that no longer require permits in carry in all or virtually all of their states.
So if the number of concealed handgun permits goes up when the concealed handgun laws change, there are a lot of studies that show crime rates decline. Indeed, the vast majority of studies find that relationship.
If one doesn’t believe in deterrence, there is one question that I frequently ask people. Would you post a sign announcing that your home is a gun-free zone? Would you feel safer? In my experience, even the most ardent gun control advocate would never put “Gun-Free Zone” signs on their home. That seems like strong evidence that even gun control advocates believe in deterrence.
Armstrong: Kleck claims that “across areas, there is no effect of gun ownership rates on crime rates, including homicide rates.” What’s your response?
Lott: Well, I know that Gary feels very strongly that gun ownership doesn’t make people safer, but I think that the evidence is fairly clear. Take a simple point: Can you name one place in the world where guns were banned where murder rates went down? I can’t. Every place in the world that has banned either handguns or all guns has seen murder rates go up. Americans are familiar with what happened in Chicago and DC, but it is even true for what would be the ideal gun control experiment—island nations that can’t blame a neighbor for their guns. If Gary were correct, you would think that you would see at least a couple clear cases where murder rates remained unchanged.
There is also evidence that the states that have had the biggest increases in gun ownership have had the biggest drops in violent crime rates.
Gary puts much more weight on purely cross-sectional data. I explained earlier why that approach is likely to be very misleading, but in this case, even if I don’t put much weight on this evidence, countries with the lowest gun ownership rates do tend to have higher homicide rates.
Armstrong: Kleck says that U.S. gun laws “don’t have any effect” and are “not even intended to have an effect” on gun ownership rates. What do you think about that?
Lott: Clearly this statement is wrong. For example, if guns are banned, won’t that affect the gun ownership rate? But it is wrong in a simpler sense. If the total costs of getting a concealed handgun permit in Illinois is $500 but in Pennsylvania it is $19, does anyone honestly believe that it won’t cause relatively fewer permits to be issued in Illinois? This gets back to our earlier discussion about the differences between economists and sociologists.
With Gary’s view of the world, I don’t think that you can understand why gun control advocates push for the types of gun control laws that they push. It seems clear, at least to me, that the regulations are aimed at increasing the cost of gun ownership precisely to reduce gun ownership and eventually make it easier to pass gun control. For example, back in 1997, Tom Smith, a gun control advocate and the director of the General Social Survey, told me that a large drop in gun ownership would “make it easier for politicians to do the right thing on guns” and pass more restrictive regulations.
Armstrong: Most people, including most politicians and most academics, have no or almost no training in statistical regression analysis. I took an introductory class in college, so I have a rudimentary understanding of it. Yet much of the debate over the impacts of concealed carry laws hinges on such analysis. How should the layperson approach conflicting claims that rely on regression analysis? Is it fair to say that, if the statistical trends aren’t obvious and that different researchers can interpret them differently, a generalized skepticism is a sensible default position?
Lott: That is an excellent question. It is also a difficult one. There is a truth out there to be discovered. But many people involved in the academic debate figure that people will just look at their conclusions and not how they got there. Still, there are some things to look for in any study:
- Do they use all the data that is available? If someone doesn’t use all the data available, they better have an extremely good reason and I would be very skeptical. Imagine that you flipped a coin 20 times and got 10 heads and 10 tails. If I let someone selectively pick coins, he could get any result he wants (such as picking 5 heads).
- Do they use what is called “panel” data? Data that looks at many different places over time. Very few academics look at either purely cross-sectional data (looking across places at one point in time) or purely time series (looking at one place over time). Panel data combines the two and allows one to have lots of experiments and thus disentangle different possible explanations.
- Do they only look at one gun control law? Many liberals argue that all sorts of gun control laws are important, but then they only account for one gun control law in their study. It gives an idea of how much they may have cherry-picked their results.
- Do they at least start with an approach used by other academics? For example, with panel data do they account for geographic and time differences (so call “fixed effects”).
- Do they misdescribe others’ research?
Beyond that it takes a lot of work to read studies (particularly opposing studies), but the more studies you read the more you read the more you will be able to draw your own judgments.
Armstrong: Perhaps the proposed reform of gun laws that strikes the most people as common-sensical is to expand and improve the background check system for gun purchases. Yet you’ve been critical of that system on grounds that it doesn’t work very well and that it actually prevents some people in profound danger from obtaining a gun for self-defense. Do you think the background check system could be improved, or should it just be scrapped? If the latter, what do you think should be done instead?
Lott: Background checks seem to make people feel safer, but the truth is the background check system is a mess. Virtually everyone who is stopped from buying a gun is a law-abiding citizen who should have been able to buy it, but they were stopped simply because they have a name similar to someone the government really wanted to stop.
The president keeps claiming that “background checks have kept more than 2 million dangerous people from buying a gun,” but stopping someone because they have a name similar to that of a felon is not the same thing as stopping a felon from buying a gun.
This is the same problem experienced with the “No Fly” list. Remember the five times that the late Sen. Ted Kennedy was “initially denied” flights because his name was on the anti-terror “no fly” list? His name was just too similar to someone that we really did want to keep from flying. By Obama’s method of counting, that means the “no fly” list stopped five flights by terrorists.
For gun purchases, the Bureau of Alcohol, Tobacco, Firearms and Explosives dropped over 94 percent of “initial denials” after just the first preliminary review. The annual National Instant Criminal Background Check System report explains that these cases were dropped either because the additional information showed that the wrong people had been stopped or because the covered offenses were so many decades old that the government decided not to prosecute. At least a fifth of the remaining 6 percent were still false positives.
All these denials mean delays for many law-abiding gun buyers. Although this is merely an inconvenience for most, initial denials cause dangerous delays for people who suddenly, legitimately need a gun for self-defense, such as a woman being stalked by an ex-boyfriend or spouse.
Beyond the crashes in the computers doing the checks and the initial denials, another 6 percent of checks fail to be completed within two hours, with most delays winding up taking three days.
President Obama ignores what happens to those who suddenly feel threatened. A gun really can make a huge difference in being able to defend against assailants. Indeed, my own research suggests these delays from the background check system likely increase violent crime, even if ever so slightly. Perhaps not too surprisingly, rape appears to be the crime most sensitive to these delays.
Furthermore, there is no real scientific evidence among criminologists and economists that background checks actually reduce crime. In fact, a 2004 National Academy of Sciences panel concluded that the Brady background checks didn’t reduce any type of violent crime. Nor have other later studies found a beneficial effect.
The number of criminals stopped by the checks is also quite small. In 2010, there were over 76,000 initial denials, but only 44 of those were deemed worthy for prosecution and only 13 individuals were convicted. Even those 13 cases don’t tend to be the “dangerous” criminals Obama claims are being stopped.
The delays have other consequences. States that have applied background checks to sales by private individuals have seen around a 20 percent drop in the number of gun shows, eliminating for many poorer people a relatively inexpensive source of buying guns. For gun shows, which usually only last two days, even a three-day delay means that no sale will be made.