Special Report

More Handguns, Less Crime — or More?

The new, expanded edition of John Lott's famous book promises to revive a heated, unresolved debate.

By 6.21.10

In 1998, John R. Lott Jr. dropped a bombshell on the academic and legal worlds with More Guns, Less Crime. Lott had conducted perhaps the most detailed study of crime in history, using data from every county in America, and concluded that right-to-carry (RTC) laws -- which allow citizens to carry concealed guns, but typically require them to receive training and pass a background check first -- reduce crime.

The University of Chicago Press has just released a third edition of the book, with updated numbers and more than 100 pages of new material. In addition to expanding Lott's argument with data from the last decade, it provides an excellent chance to look at the current state of the gun-policy debate.

Three things are clear. One, despite years in the academic spotlight, this book's central argument about right-to-carry laws has failed to create any kind of consensus. Two, the dire predictions of right-to-carry opponents have not come true. Three, the next great gun-control debate will concern handgun bans -- and, Lott's attempt to demonstrate that they increase murder notwithstanding, their effect on crime rates is no clearer than right-to-carry laws'.

To understand the lasting uproar over More Guns, Less Crime, it helps to start with the fact that to arrive at his conclusion, Lott uses what's called "regression analysis" -- a statistical technique that allows researchers to account for, or "control," confounding variables that might throw off their findings.

His basic results, for example, are presented in graphs that show various crime rates rising before concealed-carry laws are enacted and falling thereafter, suggesting that the laws reduce crime. But these aren't graphs of actual crime rates. Rather, each graph combines many states' crime data into a single line, with year zero representing the year the RTC laws went into effect. (They usually look like this, only less dramatic.)

States enacted these laws at different times, and varying trends (rising or falling crime, changing demographics) were taking place when they did so. Lott has to adjust the numbers to control for these variables -- and there are a lot of them, even though it's impossible to factor in everything. Here's a list he provides before explaining one of his new analyses:

Just as was done in the previous chapters, the estimates are going to account for not only all the law enforcement variables (arrest, execution, and imprisonment rates), income and poverty measures (poverty and unemployment rates, per capita real income, as well as income maintenance, retirement, and unemployment payments), the thirty-six measures of demographic changes, and the national average changes in crime rates from year to year and average differences across states (the fixed year and state effects). In addition, the estimates account for the differences in various concealed-handgun laws and other types of gun-control laws. 

This level of detail is standard practice in "econometrics," the sub-field of economics into which these analyses fall. And initially, it's impressive and convincing. But whenever these techniques are applied to a controversial subject, different researchers -- controlling for different variables and using different techniques to do so -- come to different conclusions. And so it has been with the more-guns-less-crime hypothesis.

Since Lott released his first batch of findings in 1997, many experts have weighed in. Plenty have been supportive; he provides a list on his website here. But others don't buy it, and they're not all anti-gun hacks. In 1997, the University of Chicago's Jens Ludwig published an essay called "Do Carry-Concealed Weapons Laws Deter Crime? No." Yale's Ian Ayres and John J. Donohue III wrote a 2003 article claiming that RTC may increase crime. The same year, the Centers for Disease control released a report on gun control; its task force had "found insufficient evidence to determine the effectiveness of any of the firearms laws or combinations of laws reviewed on violent outcomes." In 2004, a National Academy of Sciences panel -- which included econometrics superstar Steven Levitt -- similarly failed to find convincing evidence that right-to-carry laws or other forms of gun control either reduce or increase crime. (Panelist James Q. Wilson dissented.)

How to solve this dispute? Lott and his critics, being academics, turned the debate into a soap opera. There's too much drama to recount it in full here, but follow the links if you're not up to date: Lott's suspicious survey; his alter ego Mary Rosh; statements by Donohue that Lott has "blood on [his] hands" for encouraging RTC laws; Levitt's e-mail to another economist claiming that Lott had bribed an academic journal to run a special, non-peer-refereed issue containing studies that were friendly to his theory; Lott's lawsuit against Levitt regarding not only that accusation but also a statement in Levitt's book Freakonomics; the startling admission by Levitt (after losing one of the counts of that lawsuit) that Levitt himself had been a peer referee of the journal issue in question.

And where does the debate about the more-guns-less-crime theory stand today? Right where it was five, maybe even ten or more years ago. Half the time, not even the names have changed. Ayres and Donohue released a paper just last year called "More Guns Less Crime Fails Again" (pdf). Ludwig, frequently with co-author Philip Cook, still writes about how guns don't make crime go down. And of course, the new material in Lott's third edition includes another round of responses to his critics. He says that no peer-reviewed study has found an increase from RTC, but even he concedes the following: "Not too surprisingly, depending on the precise methods used and the exact data set, the results have varied. Some claim no effect from these laws."

With the econometricians still going back and forth after more than a decade, there's really no way for everyday citizens to choose sides -- as much as we may want to, and as much as activist organizations might urge us to. And as the National Academies panel suggested, it may be that econometric techniques are not precise enough to settle this debate, no matter how long we let the professionals fight it out.

That doesn't mean we know nothing about the effects of right-to-carry laws, and it certainly doesn't mean that More Guns, Less Crime is obsolete. If we take the emphasis off crime rates, and turn instead to counting individual crimes, we see quite clearly that right-to-carry laws, at the very least, don't increase crime.

The most important question is this: When people get concealed-carry permits, do they go on to misuse their guns? People often express concern, for example, that in the adrenaline rush of a car accident, an angry armed motorist might shoot the guy who ran into him.

Lott has been reporting on the behavior of permit holders for years, but the third edition of More Guns, Less Crime includes his most comprehensive list to date. Very few concealed-carry permits are revoked: In the 14 states that keep detailed records, revocation totals range from .01 to .25 percent, and overwhelmingly, revocations occur not because the permit holder misused a gun, but because he violated some other law. (In Kentucky, for example, the most common reason for revocation is a lack of vehicle insurance. In Utah, it's "alcohol violations.") Lott was able to find only 23 examples of permit holders committing murder with guns from 1990 to July 2008 -- assuming his list is comprehensive, that's a murder rate of 1/182nd that of the general population.

Meanwhile, in a much shorter period of time -- Dec. 14, 2008, to Jan. 11, 2009 -- ten permit holders stopped violent crimes. So, without resorting to regression analysis, we can prove reasonably convincingly that RTC does virtually no harm and some good.

This methodology, of course, fails to count the crimes that RTC deters -- the times when criminals don't even attack, out of fear that their victims might be armed. But once one has shown that permit holders don't commit crimes -- and thus that granting them additional freedom doesn't harm society -- any deterrence they provide is simply icing on the cake.

The Supreme Court is in the process of sorting out whether Chicago's gun ban is constitutional. As far as the Court is concerned, the empirical question of whether gun bans increase or decrease crime is tangential. But if the Supremes do strike down the ban, Chicago and other cities that ban guns will have to decide whether to try to preserve their bans in lesser form, inviting further court challenges, or just give up and let their residents own guns.

If it can be proven that gun bans increase crime, the decision becomes easy: Libertarian gun laws are superior in terms of both the Constitution and the desire to control crime. Using four case studies -- D.C., Chicago, Jamaica, and Ireland -- Lott attempts to prove that this is the case. The evidence he marshals is less than convincing, however, and some countries are conspicuous for their absence.

Take Washington, D.C., which banned guns in 1976. Compared with the nationwide murder rate, D.C.'s rate skyrocketed in 1987 -- but that was when the crack epidemic hit the city. Prior to that, D.C.'s rate didn't move much before or after the ban. It was about four times the national average from 1968 to 1974, fell to about three times the national rate by 1976, stayed there until about 1980, and ticked upward slightly before the crack wars came. Similar patterns appear when Lott compares D.C.'s rate with Virginia and Maryland's, and with the average rate of the 48 other largest cities (the top 50, excluding Chicago, which also had banned guns, as well as D.C. itself). This is hardly strong evidence in either direction.

The results Lott presents for Chicago are no more conclusive. Chicago banned guns in 1983, yet its murder rate held remarkably steady for six years before and after that point -- about three times the national average from 1977 through 1989. There was a spike in 1990, after which point the rate settled at about four times the national average -- murder fell in Chicago in the 1990s, but not as much as it did in the rest of the country (most notably Giuliani's New York). Again, there's no evidence that the ban had any dramatic effect in either direction.

Regarding Ireland and Jamaica, Lott presents only raw murder-rate data, with no comparisons to similar countries. Both banned guns in the early '70s, and both experienced significant increases in murder thereafter, but because crime rose in many countries in the '70s, it's hard to prove a cause-effect relationship.

It's also worth noting that the international gun-ban comparisons Lott has used most often in the past -- Great Britain and Australia -- are downplayed here. Great Britain's 1997 handgun ban gets a mention; Lott notes that gun-crime-related deaths and injuries skyrocketed from 1998 to 2005 and asserts that "rates of serious violent crime, armed robberies, rapes, and homicides have soared." He does not note that last year, the murder rate for England and Wales hit a 20-year low.

Meanwhile, Australia banned most firearms in 1996, and Lott in the past has noted (pdf) that violent crime -- not so much murder, but especially armed robbery -- rose thereafter. But the country's name appears nowhere in the new material or the index here, and for good reason: Homicide is at near-record lows there, and armed robbery has come down significantly since its peak at the turn of the century, suggesting that the previous increase may have been a natural fluctuation, not an enduring result of the gun ban.

If gun bans do anything to increase or decrease crime, the effect isn't dramatic enough to show up in these overall trends.

The more-guns-less-crime theory is more than plausible, and it retains the support of many academics. In the end, however, it has become a distraction. In addition to being virtually impossible to prove in a meaningful way, it has placed the burden of proof where it does not belong.

Gun-rights supporters shouldn't have to prove anything. They are on the side of freedom. Gun controllers, by contrast, want to restrict freedom, and thus must prove that their policies provide benefits that are worth that freedom. Whether the topic is RTC, handgun bans, buyback programs, assault-weapon restrictions, or registries, there is simply no evidence whatsoever indicating that to be the case. That's one thing that Lott and the debate he inspired have proven -- whatever the merits of the claim that gun control actually increases crime.


Like this Article

Print this Article

Print Article
About the Author
Robert VerBruggen is an associate editor at National Review. You can follow his writing here.