On Monday, three Democratic Senators – Dianne Feinstein, Charles Schumer, and Sheldon Whitehouse – released a Congressional investigation finding that the majority of weapons recovered and traced from crime scenes in Mexico originated in the United States. According to ATF figures quoted in the report, of the 29,284 firearms recovered in Mexico and traced by authorities in 2009 and 2010, 70% were found to have U.S. origins. These weapons have played a major role in the bloodshed of Mexico’s ongoing drug war, which claims thousands of victims every year. Among other reasons, Congress allowed the Federal Assault Weapons Ban (FAWB) to expire in 2004, making it harder to keep the most deadly guns out of the hands of drug traffickers. And the “gun show loophole” allows individuals to purchase weapons from private, unlicensed sellers at U.S. gun shows without going through a background check. How, precisely, does this supply of U.S. guns affect levels of violence in Mexico?

The Congressional report doesn’t attempt any specific estimates, and the sheer number of variables affecting the situation on both sides of the border makes it difficult to give a simple answer. However, a recent study has isolated and examined one major component of the problem. In a new paper, Arindrajit Dube of UMass-Amherst and Oeindrila Dube and Omar García-Ponce of New York University consider just one factor – the expiration of the FAWB in 2004 – and bluntly ask: “Do more guns cause more violence?” Noting that the FAWB’s expiration created a “natural experiment” to examine the effect of U.S. laws on Mexico’s crime rate, the authors reviewed data from 2002 to 2006 – looking at years before and after the ban’s expiration – and concluded that “the 2004 change in U.S. gun law had differential effects on homicides in Mexican municipios with greater exposure to the policy shock.” That is, homicides, gun-related homicides, and gun crimes increased because of the U.S. law’s expiration, particularly in regions near the Arizona and Texas borders. There was no similar increase in the areas near California (perhaps, the authors suggest, because of California’s stricter gun laws).

The authors admit surprise at their results, saying the “spillover” effect was “unanticipated,” but they report that their findings were “robust to controls for drug trafficking, policing, unauthorized immigration, and economic conditions in U.S. border ports, as well as drug eradication, trends by income and education, and military and legal enforcement efforts in Mexican municipios.” And they offer a grim measure of this spillover effect: “Our estimates suggest that the U.S. policy change caused at least 158 additional deaths each year in municipios near the border during the post-2004 period.”