Federally licensed firearm dealers are an important target of regulatory and enforcement efforts aimed at reducing the supply of firearms diverted to the illicit market, but the extent of dealers' involvement in the illegal diversion of firearms is hard to measure.
In a new study using machine learning, researchers examined firearm transaction and crime gun recovery records from California from 2010 to 2021 to identify dealers selling the highest number and largest proportion of guns recovered in crimes within a year of sale, a well-established indicator of possible illegal activity by dealers or traffickers. The prediction models have the potential to support targeted enforcement, helping disrupt the flow of firearms to offenders.
The study, by researchers at the University of California (UC), Davis, appears in Criminology & Public Policy, a publication of the American Society of Criminology.
"Although most gun offenders do not get their firearms directly from licensed dealers and most dealers abide by laws, even a few negligent or corrupt dealers can contribute significantly to the supply of firearms used in crimes," explains Hannah Laqueur, associate professor of emergency medicine at UC Davis, who led the study.
Firearms dealers can facilitate the diversion of guns to the criminal market through practices such as selling to straw purchasers or failing to conduct required background checks. Prior studies have shown that enforcement actions and lawsuits targeting law-evading dealers can deter these behaviors and reduce the flow of firearms into criminal markets.
In this study, researchers used machine learning techniques to develop two prediction models. The first classifies dealer-years in the top 5% of one-year crime gun sales volume (the number of sales of guns recovered in crimes within a year of sale); the second identifies dealer years in the top 5% based on the fraction of sales recovered within a year. Both models had strong discriminative performance, with the first model particularly effective at identifying the highest-risk dealers.
The models generally outperformed simpler regression and rule-based approaches, underscoring the value of data-adaptive models for prediction. Key predictors included prior-year crime gun sales, the average age of purchasers, the proportion of "cheap" handgun sales, and the local gun robbery and assault rate.
Many of the dealers with the highest predicted probabilities not only sold a large volume of guns with very short "time-to-crime" but also consistently sold crime guns over multiple years. This suggests that a relatively small group of dealers could be targeted for enforcement, offering the potential for outsized impact. More consistent and targeted inspections of high-risk dealers, along with citations or license revocations, could strengthen deterrence and promote compliance, helping reduce the supply of guns to offenders.
"Our findings show how machine learning techniques, combined with California's comprehensive firearm transaction and crime gun recovery data, could help identify potentially high-risk retailers," says Laqueur. "This type of identification can improve the efficiency and effectiveness of inspections and enforcement efforts aimed at interdicting negligent or corrupt dealers."
Among the study's limitations, the authors note that dealers selling many short time-to-crime guns may not have violated the law and conversely, non-compliant dealers may not be reflected in short time-to-crime statistics. They also point out that because California is a state with stringent gun laws and extensive dealer regulations, the number of negligent or law-evading dealers in the study may be lower than in states with fewer regulations. However, while the study's findings are specific to California, consistency of risk factors across different jurisdictions and regulatory contexts suggests that the models could inform approaches in other states, the authors say.
The study was supported by the National Collaborative on Gun Violence Research.
