Social scientists have been trying for many years to blame homicides on the presence of guns. A favorite tool in this quest is the use of studies that show a correlation between gun ownership and crime. These studies are then reported as “evidence” that the presence of guns causes crime.

But there’s always been a problem with this attempt at showing causality between guns and homicides: causality can just as plausibly go the other way. That is, in times and places where the local population feels they are in danger of being crime victims, people are more likely to purchase guns for protection. So, rather than saying “guns cause crime,” we should be saying “crime causes guns.”

New Gun Purchases Soar as Uncertainty and Violence Increase

We’re likely seeing this phenomenon at work now.

In recent months, according to the firearm industry’s trade group National Shooting Sports Foundation (NSSF), Americans have purchased millions of guns:

The early part of 2020 has been unlike any other year for firearm purchases—particularly by first-time buyers—as new NSSF® research reveals millions of people chose to purchase their first gun during the COVID-19 pandemic.

Fox News reports:

Gun sales have skyrocketed during the past three months, and a record-breaking 80.2 percent increase in sales was reported in May compared to last year, according to the shooting foundation. April’s data showed a 71.3 percent increase from 2019, and there was an 85.3 percent increase in March, according to information previously released by Small Arms Analytics and Forecasting.

Many new gun owners during this period feared general unrest as a result of the government-mandated lockdowns. Potential first-time buyers still on the fence about buying a firearm in May were perhaps confirmed in their fears by the riots that erupted after the killing of George Floyd by Minneapolis police officers. Then, in the wake of the riots, serious violent crime appeared to spike. It was widely reported, for instance, that homicides in New York City spiked “21 percent in first six months of 2020.” Crime in other cities increased as well, ranging from a jump of over 200 percent in Nashville to 23 percent in Kansas City, Missouri.

Naturally, seeing these news stories, many potential gun owners are more likely to conclude that they need a gun for personal protection. This is especially true when combined with a perception that police organizations cannot be relied upon to engage in crime prevention and enforcement. And this has indeed been the perception in many places where police have appeared unwilling to intervene in June’s riots.

Many normal people would see these events as an illustration of how gun purchases result from fears over crime and uncertainty.

But now, perhaps predictably, left-wing media organizations like The Washington Post are trying to turn this narrative around: people aren’t buying guns as a reaction to violence and social disarray, the Post insists. All those new gun purchases are what’s causing the violence in the first place.

Says the Post:

Americans purchased millions more guns than usual this spring, spurred in large part by racial animosity stoked by widespread protests over the killing of George Floyd in Minneapolis, as well as anxiety over the effects of the covid-19 pandemic.

That gun-buying binge is associated with a significant increase in gun violence across the United States.

The Post cites two new reports, one from the Brookings Institution and another from the University of California, both of which conclude that the rise in gun purchases has likely caused more “gun violence.”

Note the careful use of language here, though: the gun purchases are “associated” with an increase in gun violence, since causality cannot be established. Indeed, near the bottom of the Post article, the author admits:

The authors [of the Brookings and UC reports] caution that a study of this nature cannot prove causality, particularly at a time of massive social upheaval in a country dealing with an unprecedented public health crisis as well as a nationwide protest movement.

Of course, if one is already committed to the idea that guns cause crime, it makes perfect sense that millions of Americans in early 2020 – after passing a criminal background check – will buy guns, and then almost immediately use those guns to commit crimes.

Moreover, it’s unclear that the two studies referenced by the Post article even imply that homicides result from more gun purchases.

The Brookings study, for instance, is more of an op-ed than a study. It’s simply a review of some past events which were followed by surges in gun purchases, including the Sandy Hook and Parkland shootings. This appears to be true indeed, and is a helpful reminder that people do often purchase firearms in light of concerns over personal safety, or at least in light of concerns about future access to firearms.

The UC study is a bit more specific, but even this is far too general to be of any use in concluding that gun purchases lead to violence. Because of data limitations, the UC report, of course, doesn’t establish that the people who bought firearms this year are responsible for any increase in crime that may be occurring. But it’s not even established that surges in gun purchases correlate with surges in crime at the city or neighborhood levels. This is critical, since trends in homicides are not really on a statewide or even metro-wide level. Homicide trends in the U.S. tend to be dominated by homicides in a relatively small number of cities and neighborhoods. For example, the homicide rate in Baltimore is ten times that of the U.S. overall. But this doesn’t mean homicides in Maryland are remarkably high.

So, have firearms purchases surged near the neighborhoods in Chicago, New York, and Kansas City where surges in crime have also occurred? It’s possible, since people bordering the most violent neighborhoods may feel the most at risk. On the other hand, it’s also entirely possible that firearms sales are occurring in places relatively distant from the places with surging homicides. The UC study only appears to give a state-level reading on this. In other words, the study really tells us very little.

This article has been republished with permission from the Mises Institute.