Debunking “More Guns, Less Crime” 3.0

Kenneth Kormanis, right, had dinner with his wife and daughter at All Around Pizza in Virginia Beach, Va., on Monday, Feb. 18, 2013,  where anyone who carries a gun or brings a concealed handgun permit is offered at 15 percent discount.  (AP Photo/The Virginian-Pilot, Amanda Lucier) MAGS OUT

This article is part of an ongoing series debunking the “More Guns, Less Crime” hypothesis.  See also Part 1 and Part 2.   We have also critiqued a number of egregious errors in this debate including the fabrication that there is an “academic consensus” on More Guns, Less Crime, and that all mass shootings occur in “gun-free zones.”  In this article we make a new argument that, not only is there no empirical evidence that more guns produce less crime, but the causal link purporting to show this isn’t even theoretically coherent.

In the aftermath of the Oregon college shooting that left 10 people dead and nine more injured, conservative politicians and pundits offered their familiar prescription for halting mass shootings: even more guns: “If you had a couple of the teachers or somebody with guns in that room, you would’ve been a hell of a lot better off,” said GOP presidential frontrunner Donald Trump. Ted Nugent, an NRA board member, went event further, saying that just about every member of our populace should carry a concealed weapon. “Disarmed and helpless is an irresponsible, suicidal choice that will get you killed. Defend yourself.”

These proclamations stem from the increasingly popular belief that people carrying concealed firearms deter criminals and mass killers. Indeed, the foundational tenet of the National Rifle Association’s agenda — Right-to-Carry (RTC) reciprocity, permitless carry, and forcing schools to allow concealed carry — is that more “good guys” carrying guns in public will reduce crime and make society safer. In states with RTC laws, it is exceedingly easy to become a “good guy with a gun”— one only needs to pass a background check and a basic gun safety course. In some states like Arizona, you don’t even need a permit; so long as you’re at least 21 years old and not a criminal, you’re free to carry. This push for no-hassle concealed carry is almost unanimously shared by Republican presidential candidates. Trump, for instance, recently touted his support of national RTC reciprocity.

Gun rights advocates frequently highlight the fact that from the early 1990s to today, violent crime nationwide has fallen precipitously, with gun homicides declining 49 perecent. This dip in all types of violent crime happens to correspond with a dramatic surge in the number of states issuing concealed carry permits. Those same advocates, usually citing studies conducted by pro-gun researcher John Lott, contend it is the rising number of good guys with guns on the streets that is responsible for the lower crime rate. But this line of argument runs counter to the facts.

In 1997, Lott published the book More Guns, Less Crime, which argues that states with RTC laws experienced significantly lower crime rates than those without such policies. The volume — based on highly complex statistical models — set off an academic firestorm, with some studies supporting his findings and others identifying significant statistical problems. Despite the controversy, Lott’s work provided an impetus for more states to adopt RTC laws. In 1997, 30 states had shall-issue RTC laws (the least stringent form of permits), while seven states prohibited concealed carry altogether. Today, 35 states have shall-issue RTC laws, seven states don’t require a permit to carry, and no states prohibit concealed carry. When Missouri was debating establishing its own RTC law, a pro-gun group sent a copy of Lott’s research to every state senator, an effort that helped secure enough votes to override the governor’s veto.

Since the publication of More Guns, Less Crime, at least three major reviews of Lott’s work have debunked his findings. One particularly decisive critique, a 2003 study published in the Stanford Law Review, used a superior statistical models and extended the time frame under analysis. With those adjustments, the paper found that the alleged reductions in crime rates evaporated. Another critical analysis, this time issued from 15 of the 16 panel members of National Research Council  (NRC), concluded that “with the current evidence it is not possible to determine that there is a causal link between the passage of right-to-carry laws and crime rates.” Then, in 2011, a team of researchers analyzed the NRC panel’s findings and conclude that RTC laws, in fact, increase crime. And these three studies represent only the tip of the iceberg — there are many more cataloging the numerous ways in which Lott has erred.

With the academic evidence mounting against him, Lott recently published a paper in Econ Journal Watch seeking to move the goalposts of the More Guns, Less Crime argument. In the paper, he argues that dissenting studies examining RTC laws overlooked that those laws often differ in how easy it is to obtain a permit, a difference that greatly influences the numbers of concealed carry permits issued by each state. Rather than focus on the passage of RTC laws, Lott contends that researchers should instead examine the change in the number of permits.

Increased concealed carry permit rates have no impact on crime rates.

Lott’s move was transparently tactical. But his focus on the number of permits instead of the passage of RTC laws does make sense. While it is possible for criminal behavior to be influenced by the news of a law passing, logically it is far more likely that the number of “good guys with guns” on the streets is the relevant factor. However, far from salvaging More Guns, Less Crime, Lott’s pivot actually opens up his theory to even more devastating criticism, an opening that was seized by Dr. Charles Phillips of Texas A&M and his colleagues in a new study.

The Texas A&M paper directly challenges the hypothesis that increased numbers of concealed carry permits reduce crime. The study analyzes a decade of data from every county in Florida, Michigan, Pennsylvania, and Texas, the only states with at least a decade of reported data on permit holders and arrest rates after the implementation of their RTC laws (an explanation of their methodology that, unlike what Lott misleadingly suggests in a rebuttal, is very clearly delineated). Using several statistical models, Phillips found no significant relationship between changes in concealed carry rates and changes in any crime rate. In other words, the study found no evidence that increasing the number of permit holders decreases (or increases) crime.

It doesn’t take complicated economic analysis to disprove “more guns equals less crime.” The argument fails a basic logic test.

While the above studies provide clear evidence that, at the very least, permit holders and concealed carry laws do not reduce crime, they are mired in complex statistical models, which leads to arcane arguments over “model selection” and statistical analysis. One way to sidestep those thickets is to examine the two ways permit holders — and by extension RTC laws — could be reducing crime: through direct or indirect deterrence. Recent empirical evidence and studies show that neither of these pathways can be responsible for the reduction in crime.

Defensive gun use is too rare for “good guys with guns” to significantly lower crime rates.

The first mechanism through which permit holders and concealed carry laws could be reducing crime is through direct deterrence, which occurs when an armed civilian uses a gun in self-defense, thereby stopping a crime. The NRA and gun advocates frequently tout surveys conducted by criminologist Gary Kleck indicating that there are around 2.5 million defensive gun uses every year, which would mean millions of criminals being directly deterred from crime.

However, widespread defensive gun use is a myth. The survey results used to extrapolate millions of DGUs suffer from a severe false positive problem and present crime prevention numbers that are mathematically impossible. In fact, as we have detailed in previous articles, not only is defensive gun use no more effective at preventing injury than taking no action at all during a crime, but the best empirical evidence to date from the Gun Violence Archive could also only find 1,600 verified DGUs in 2014. This means that 99.936 percent of Kleck’s claimed DGUs are nowhere to be found, despite those very surveys stating that more than 50 percent of DGUs are reported to the police (meaning there should be a record of them). With so few DGUs, it is not possible for permit holders and concealed carry laws to be reducing crime through direct deterrence.

People’s estimates of concealed carry rates are wildly off — so concealed carry can’t be indirectly deterring criminals, either.

The second way permit holders and concealed carry laws could be reducing crime is through indirect deterrence — when criminals are deterred by the mere threat of confronting an armed civilian. As Lott states, “By the very nature of these guns being concealed, criminals are unable to tell whether the victim is armed before they strike, thus raising criminals’ expected costs for committing many types of crimes.” If criminals perceive that the risk of encountering an armed civilian has increased, they will likely avoid that jurisdiction and choose a different venue for their illegal activities, or not commit the act at all. This requires that criminals are actually sensitive to the prevalence of guns in their environment, and are particularly aware of changes made to state legislation that could potentially influence the quantity of concealed carriers. Until recently, no studies had challenged these assumptions.

Research published last year by Dr. Fortunato of the University of California examined the feasibility of indirect deterrence by conducting a survey asking 1,000 citizens to estimate how many people (out of 1,000) carry a gun in their state. These responses were then cross-referenced with their state’s concealed carry policy (“may issue”, “shall issue”, or “no policy”) and the number of active permits in the state. The paper also controlled for other factors that may influence a population’s belief about the concentration of perceived concealed carriers, such as the depth of legal and illegal firearm market, measurements of state ideology, and rates of firearm violence.

The paper found no statistically significant relationship between a states’ concealed carry policy and people’s perception of the number of firearm carriers in their state. As Fortunato states: “Because beliefs over the distribution of firearm carriers are impervious to permitting policies and do not respond positively to the true distribution of carriers,” increasing the number of concealed carry permits in a state “cannot deter crime.” In other words, people cannot be deterred by the number of permit holders if, on average, people have no clue how prevalent those permits actually are. The paper goes on to concluded that by passing concealed carry laws — which likely increased gun ownership nationwide — “at best, we increase the probability of accidental discharge. At worst, these policies open the door for more violent, potentially deadly, escalations of altercations — altercations that may have ended peacefully if not for the presence of a firearm.”

Some gun rights advocates may argue that while this lack of awareness of the actual number of permit holders is true among the general population, criminals may have superior knowledge of the number of law-abiding citizens carrying firearms, and therefore the notion of indirect deterrence still holds. But there is no substantive evidence to support this claim. Indeed, if criminals do have an internal gun radar of any sort, it is of gun ownership in an area, not the prevalence of concealed carrying. And instead of a being a deterrent, a2002 study by Dr. Phillip Cook indicates that more guns in an area mean more burglaries, as criminals capitalize on the lucrative opportunity that stolen guns present.

Even if criminals were responsive to indirect deterrence from concealed carry permit holders (which the evidence shows they aren’t), the areas with the most per capita permits are already low crime areas. This is clearly seen in the Illinois concealed carry data we previously analyzed, which showed permits are concentrated in white, rural areas with low levels of crime. If there are few criminals in an area to deter, indirect deterrence cannot be significantly reducing overall crime rates.

We don’t know what caused the big two decade drop in crime rates. We do know it wasn’t concealed carry.

So if concealed carry cannot be responsible for the significant, two decade drop in crime rates, what is? A report from the Brennan Center for Justice published earlier this year sought to tackle that question. Analyzing 40 years of data from all 50 states, the report examined a number of potential causal factors, including RTC laws, and found that the most likely causes were various demographic and socioeconomic factors, the end of the crack cocaine epidemic, and superior policing techniques. However, even these factors were insufficient to explain the entire crime decrease, and the authors concluded that we still don’t fully know why crime dropped precipitously.

But what we do know is that rigorous studies on RTC laws and permit holders, combined with empirical data on defensive gun use and studies of people’s perception of gun prevalence, provide powerful evidence that concealed carry does not reduce crime.



Related posts

  • Pingback: There’s A Petition Out There You Really Should Sign. | mikethegunguy()

  • gzuckier

    The frequent assertion of the MG=LV folks, central to the thesis of Lott and the others who publish similar analyses, is that there is a constant barrage of potential crimes all around us which are deterred by firearms without getting to a stage where they are recorded by law enforcement or anyone else (except the folks telling you the story, either directly or via a phone survey by Lott).
    This concept quickly falls apart as soon as you examine it, though. If you make the most generous assumptions in favor of the hypothesis; that half the population are carrying a concealed firearm at any given moment, that every attempt on a person carrying a weapon fails, that every attempt on an unarmed person succeeds, the inescapable conclusion is that either unarmed people are being victimized at an astoundingly high rate many times larger than the official figures, or else armed people are being preferentially targeted at a rate many times higher than unarmed people. And in reality, of course, far fewer than half the population are armed at any given time, criminal attempts on armed people sometimes succeed and on unarmed people sometimes fail, all of which make the math come out even worse for the MG=LV theory.
    Ironically, Lott’s work did pretty well at showing no significant link between gun ownership rates and violent crime rates, which ought to have been enough for the gun rights folks to defend their ownership; but it’s that extra drive to paint themselves as John Waynes and Clint Eastwoods keeping America safe that trips them up yet they can’t stop themselves from making that identification, and the Lotts of the world can’t stop themselves from pandering to it.

  • I saw a freakonomics study that concluded a big part of the drop in the crime rate was attributed to the legalization of abortion. Which kind of points to the observation I have which is unwanted children are more likely to become criminals.