Fewer Guns Mean Fewer Gun Homicides

Summary of working paper 7967
Featured in print Digest

About one-third of the gun-homicide decline since 1993 is explained by the fall in gun ownership.

Increases in gun ownership lead to a higher gun-homicide rate and legislation allowing citizens to carry concealed weapons does not reduce crime, according to a recent NBER Working Paper by Mark Duggan. After peaking in 1993, gun homicides in the United States dropped 36 percent by 1998, while non-gun homicides declined only 18 percent. In that same period, the fraction of households with at least one gun fell from more than 42 percent to less than 35 percent. Duggan finds that about one-third of the gun-homicide decline since 1993 is explained by the fall in gun ownership. The largest declines occur in areas with the largest reductions in firearm ownership.

Previous research on the relationship between gun ownership and crime has been impeded by a lack of reliable data on gun ownership. But in More Guns, More Crime (NBER Working Paper No. 7967), Duggan uses a new proxy for gun ownership -- state and county-level sales rates for the nation's largest handgun magazine -- to show that guns foster rather than deter criminal activity.

In theory, the effect of gun ownership on crime is ambiguous. If criminals are deterred from committing crimes when potential victims are more likely to possess a firearm, then more gun ownership may lead to a reduction in criminal activity. If instead guns increase the payoff to criminal activity, or simply increase the likelihood that any particular confrontation will result in a victim's death, then an increase in gun ownership will tend to increase the crime rate.

Proving one theory over the other has been difficult because of the lack of adequate data on gun ownership measured across geographic areas over time. But as evidence of the accuracy of the gun magazine subscription data, Duggan shows that sales rates are significantly higher in counties whose average demographic characteristics are similar to those of the typical gun owner according to national surveys. Furthermore, he shows that the death rate from gun accidents and the number of gun shows per capita are positively related to the magazine sales. While Duggan admits that relatively few readers may be criminals, he points out that the majority of firearms used in crime are obtained either from burglaries or from the secondhand market. Thus as the rate of gun ownership in the general population increases, the ease with which criminals can obtain a gun will increase.

Duggan finds that state and county-level changes in the rate of gun ownership are positively related to changes in the homicide rate. His findings suggest that gun ownership causes crime, and does not simply reflect individuals purchasing guns in response to increases in criminal activity. In support of this, he finds that increases in gun ownership are positively related to future increases in the gun homicide rate, but bear no corresponding relationship to non-gun homicides. His findings reveal that the relationship with other crime categories is much less marked, suggesting that guns primarily affect crime by increasing the homicide rate.

He then examines whether legislation that allowed individuals to carry concealed weapons had an important impact on the crime rate. He shows that this legislation did not lead to a substantial increase in gun ownership, nor did it reduce crime relatively more in counties with high rates of gun ownership. This latter finding suggests, Duggan writes, "either that gun owners did not increase the frequency with which they carried their guns or that criminals were not deterred by the greater likelihood that their victims would be armed." Taken together, his results suggest that Carrying Concealed Weapons legislation did not have an important effect on the rate of gun ownership or on the crime rate.

-- David R. Francis