The Expiration of The U.S. Assault Weapons Ban Increased Homicides in Mexico

That is the key conclusion from a recent working paper (pdf, ungated) by Arindrajit Dube (UMass), Oeindrila Dube (NYU) and Omar Garcia Ponce (NYU):

Does access to arms promote violent crime? We exploit a natural experiment induced by the 2004 expiration of the U.S. Federal Assault Weapons Ban to examine how the subsequent exogenous increase in the availability of lethal weaponry affected violence in Mexico. The expiration relaxed the permissiveness of gun sales in border states such as Texas, Arizona and New Mexico, but not California, which retained a pre-existing state-level ban. Using mortality statistics over 2002-2006, we show that homicides, gun-related homicides and crime gun seizures increased differentially in Mexican municipios located closer to entry ports in these other border states, relative to entry ports in California. Our estimates suggest that the U.S. policy change caused at least 239 additional deaths annually in municipios near the border during post-2004 period. The results are robust to controls for drug trafficking, policing, unauthorized immigration, and economic conditions in U.S. border ports, as well as drug eradication, military enforcement, and trends in income and education in Mexican municipios. Our findings suggest that U.S. gun laws have exerted an unanticipated spillover on gun supply in Mexico, and this increase in arms has fueled rising violence south of the border.

Below is just one graph with raw data . There are many more graphs, tables, and maps in the paper.

This paper (pdf, ungated) by Luke Chicoine reaches a very similar conclusion with a different methodology:

In the four years following the expiration of the U.S. Federal Assault Weapons Ban (AWB), the homicide rate in Mexico increased 45 percent. Over the same period, over 60,000 firearms recovered in Mexico have been traced back to the U.S. A difference-in-difference approach is used to estimate the effect of the expiration of the AWB on homicide rates in Mexico; states with a strong pre-2005 drug cartel presence are defined as the treatment group. The baseline estimates suggest the expiration of the AWB is responsible for at least 16.4 percent of the increase in the homicide rate in Mexico between 2004 and 2008.

24 Responses to The Expiration of The U.S. Assault Weapons Ban Increased Homicides in Mexico

  1. Scott Monje December 17, 2012 at 11:39 am #

    It’s hardly surprising that lax gun laws attract gun smugglers. Similarly, our lax banking regulations made us the go-to place for money launderers from al-Qa’ida and Iran. Perhaps more surprising is that the people who insist on making access to these weapons legal and easy are the same ones who raised such a stink about Fast and Furious and the failure of the government to stop what they started.

    • Kyle December 17, 2012 at 1:32 pm #

      Your partially correct. I agree lax gun laws would attract more smugglers. But saying the people who want these weapons to be easily accessible are the ones making a stink, I hardly agree. Let’s face, you can have the most strict gun laws the world has ever seen, and I will never stop gun crimes. Making stricter laws just takes guns out of the hands of law abiding people, and your civilian sheep dogs. The only way to stop an armed threat is with an armed response. Might as well outlaw cars, pencils, knives, planes, and common hand tools. All can be used as deadly weapons.

      • Scott Monje December 17, 2012 at 3:09 pm #

        Kyle, I think you’ve fallen into the “canned response” syndrome, since that doesn’t address anything I said. My point here was that the people who made gun buying legal and easy also complained that the Justice Department didn’t arrest people for buying guns. Maybe they didn’t intend to make it easy for Mexican gun smugglers, but that’s what they did.

        • Charlie December 18, 2012 at 7:05 pm #

          Scott, I think you’ve fallen for the “Straw Man” syndrome, since the people who complain about Fast and Furious aren’t complaining failure to “arrest people for buying guns”, but for setting up a sting designed to further a political agenda, coercing law-abiding citizens into breaking existing gun laws in furtherance of this scheme, then completely losing control of the weapons they delivered into the hands of criminals. (Not to mention the fact that these guns were later used in the murder of Federal Agents.)

          I don’t see how it’s inconsistent to believe that these weapons should be legal, but that the US Govt shouldn’t illegally traffic them to criminals.

          • Dan Nexon December 19, 2012 at 11:59 am #

            “but for setting up a sting designed to further a political agenda”

            That would be the the “paranoid lunatics” complaining about Fast and Furious. The non-straw men complaints centered on a program that facilitated gun smuggling for limited law-enforcement benefits.

      • Terry December 23, 2012 at 9:18 pm #

        Kyle, but they’re not all used as deadly weapons. UK had 35 shooting homicides last year to our 27, 000 so strict gun laws do stop gun crimes. Don’t know where you get your information. NRA I guess.

    • Michael December 18, 2012 at 4:47 pm #

      How is it inconsistent to claim that the Justice Dept should actually enforce the laws it is supposed to enforce?

      This study is bollocks, it completely ignores the basic facts about the violence in Mexico, that less than 20% of guns used in crimes came from the US, that the most dangerous weapons (such as machine guns and explosives) are basically unavailable in the US, and that most of the guns used are sourced either from the Mexican gov’t (through theft, desertion and corruption) or from war zones south of Mexico.

      Now, go write on the board one hundred times: “correlation is not causation”.

      • Andrew_M_Garland December 19, 2012 at 2:22 am #

        Michael, in support,

        An “assault weapon” is defined by the nuances of law, mostly according to cosmetic or add-on features. Having a bigger magazine (attached supply of bullets) or a noise supressor makes a standard rifle into an “assault weapon”. Sometimes it takes two of these additions to qualify as an assault weapon, where either one would not do it alone.

        An assault version takes the same ammunition of the same power and accuracy, with the same semi-automatic character as a non-assault version (each trigger pull fires the gun without manual insertion of a new bullet).

        So, the Mexicans could smuggle non-assault versions, and ship the add-ons separately. This would only be more convenient under the relaxed rules.

        So, why would the ban on “assault weapons” have any effect on the armament of Mexican drug gangs? They are the same weapons, assault or not, with add-on features readily available under both legal regimes.

        Military-Style Weapons
        7/27/12 – National Review by John R. Lott Jr.

        • Dan Nexon December 19, 2012 at 12:01 pm #

          You don’t get to cite John Lott on anything gun-related. Not allowed. Now, rather than call the study bullocks based on the non-responsive 20% figure, let’s see some discussion of whether it establishes its causal claim or not, eh? That figure is, after all, entirely consonant with the claim that the lapse of the ban *increased* violent crime.

  2. Bauerton December 17, 2012 at 5:53 pm #

    Why cherry-pick an odd study on ‘Mexico’ homicide rates versus ‘United States AWB’ … when their are numerous reputable studies ‘directly’ assessing the AWB effects upon U.S. homicide/crime/violence rates ??

    The answer is that there is not a single published academic study showing that the AWB reduced any type of violent crime in the U.S. The AWB was an exceptionally controversial and well-studied issue.

    Also, that “Mexican Study” somehow overlooked “Fast and Furious” in its analysis:

    “Operation Fast and Furious” was launched in 2009 by top Federal DOJ officials, in collaboration with the FBI, DEA and BATF as part of an illegal strategy to deliberately sell arms to Mexican crime cartels thru “straw purchasers” in U.S. border states. A similar Mexican “gun walking” operation began in 2006 — but ‘Fast and Furious’ was the biggest ever DOJ escapade of that type.

    Between 2009 and 2011, Federal ATF agents prompted more than 2,000 firearms (mostly AW-type) to “walk” across the border. These weapons were intentionally delivered to the most dangerous criminal elements in Mexico. About 1,700 of those weapons were ‘lost’, and more than 100 have been found at bloody crime scenes on both sides of the border, including the murder of a U.S. Border Patrol agent in Arizona. ATF whistle-blowers blew the lid off of Operation Fast and Furious shortly that Agent’s death.

    • Oeindrila December 18, 2012 at 5:00 pm #

      It’s extremely difficult to discern the true effect of the AWB on crime rates in America, since variation in guns laws across U.S. states likely reflects current and anticipated crime levels locally.

      For example, in their 2001 study (“The Impact of the 1994 Federal Assault Weapon Ban on Gun Violence Outcomes” in Journal of Quantitative Criminology, Vol. 17, No. 1), Christopher Koper and Jeffrey Roth assess the impact of the 1994 Federal Assault Weapons Ban by comparing homicide rates across states that already had a state-level ban in place to those that did not, before and after the ban. Their estimates suggest about a 10% reduction in homicides (though the effect is somewhat imprecise).

      But the true effect may be even larger if, for example, states that had passed a ban earlier were experiencing decreasing violence trends, or a fall in crime levels with a lag due the early policy adoption.

      The potential responsiveness of local laws to local crime may compromise the ability of such research designs to detect the true effect within the U.S. context, which is one reason we choose to look outside the U.S., to obtain more credible estimates of the FAWB’s impact.

      Also, even if the effect were small, on average, there may be diversity in the effect size across U.S. jurisdictions, with the most volatile or unstable areas seeing larger homicide reduction benefits from gun control. As I discussed in a post this morning, the latest version of our paper finds that effects were much larger in areas of Mexico with greater underlying instability. See:

  3. Frank Youell December 18, 2012 at 12:58 pm #

    There is a presumption here that guns are not portable and/or tradeable. In other words, the supply of weapons on the Mexican side of the border with California is constrained by what can be purchased (or not purchased) in California and then smuggled into Mexico (American guns are generally illegal in Mexico).

    Is it really plausible that Mexicans in Baja can’t cross the border elsewhere to obtain guns? Is there no trade (presumably illegal) in weapons in Mexico? Can the authors show that weapons were more expensive in Baja than in other areas of Mexico after the AWB was lifted?

    If we were talking about non-portable goods (housing or hay) this argument might have some merit. Guns are very portable. They are relatively small and valuable. That makes them highly tradeable.

    Why different homicide rates in different part of Mexico after the AWB was lifted? The cartel wars took off in Mexico after 2000. The Mexico – California border is more tightly controlled than the borders with Arizona, New Mexico, and Texas. Presumably the cartels took their business (and funerals) elsewhere.

    A useful note here is that most drug cartel weapons do not appear to be of U.S. origin. From

    “Research has asserted that most weapons and arms trafficked into Mexico are from gun dealers in the United States.[113] In response to a 2009 GAO report that claimed 87% of Mexican crime guns traced to U.S. origins, the DHS pointed out that DHS officials believe that the 87 percent statistic is misleading (i.e.: out of approximately 30,000 weapons seized in drug cases in Mexico for 2004-2008, 7,200 appeared to be U.S. origin, approximately 4,000 were found in ATF manufacturer and importer records, and 87 percent of those—3,480—originated in the United States)”

    • Oeindrila December 18, 2012 at 11:19 pm #

      There’s a cost associated with transporting guns not because of their weight but because it is risky to drive around with guns in the trunk given potential detection by the authorities. This risk leads smugglers to adopt costly methods for shipping weapons across the border, like transporting 1 to 3 guns at a time in what are called “ant runs.”

      Unfortunately, there’s no systematic data that tracks the price of purchasing illegal weapons in different Mexican muncipios, but there is qualitative evidence of fairly large profit margins associated with transporting guns across the border, for example, from this CRS report:

      And, the price increases when the sale is further south in Mexico, as described by this NY Times article:

      It’s also risky for cartels south of CA to cross over into rival cartel territory to snuggle weapons, because certain cartels tend to dominate certain port cities and have ongoing alliances with gangs on the American side. This adds an additional turf based cost to arbitraging arms across different parts of the border.

      In terms of the types of weapons are getting trafficked, a Violence Policy Center study suggests that 43% were assault weapons and another 30% were semiautomatic guns:

      • Frank Youell December 19, 2012 at 2:11 am #


        Mexicans living in Baja, Mexico seeking American weapons, had (have) a long list of potential sources irrespective of gun regulations in California. First, they could drive to Arizona-Mexico border which is only 150 miles away (at most). Second, they could buy weapons from folks east of Baja, Mexico (if travel is precluded by cartel conflicts). Third, they could purchase weapons in American states (north of the border) other than California, and then bring them into Mexico via the California border.

        Both Mexico and the United States have east/west highways relatively near the border. Transporting goods west on either side of the border isn’t that hard. As you say, there is no evidence that weapons were scarcer on Mexico-California border versus other border areas after the AWB expired.

        I don’t doubt that guns cost more on the Mexican side of the border. Crossing a border with any enforcement at all, imposes some costs. However, the border is almost certainly less important than the fact that guns are legal in the U.S. and illegal in Mexico. Illegal goods generally trade a a premium. Since moving illegal goods costs money, it makes sense that guns would cost more away from the border.

        The NYT article actually has some data on this point. An American $125 handgun is (was) apparently worth $375 in Tijuana and $500 further south (quoting). That’s a $125 premium for moving an illegal gun in Mexico. Presumably it is cheaper to move legal guns in the U.S. (from Arizona to California).

        In any case, the Baja gun price premium (if one ever existed) was (is) likely to be small. Too small to account for large deltas in crime rates.

        • Oeindrila December 19, 2012 at 8:17 pm #

          To be clear, I did not say there is no evidence that guns are more scarce in areas near CA. Our paper in fact shows precisely the opposite. Using actual data on gun seizures from the Mexican government, we show that the FAWB expiration led to differential increases in crime gun seizures (as well as homicides) in areas closer to the other non-CA border states. And, this is only visible for the rifle category that includes assault weapons, but not handguns. This can’t reflect any of the things that we control for, which include law enforcement efforts, economic conditions, drug seizures, and illegal immigration.

          We agree that transporting goods illegally makes things more costly. That is one of the reasons it is more costly to drive a truck full of guns through California, where the Roberti-Roos act has made it illegal to transport assault weapons. Crossing over into Arizona means going into Juarez territory, given cartel segmentation. I am not sure what it means to get guns east of Baja, since that is the Pacific Ocean.

          But of course some assault weapons get through to CA — but this type of spillover suggests that our empirical approach will underestimate the true effect of the FAWB expiration.

  4. Manda December 18, 2012 at 2:36 pm #

    Liberal slanted politifact even calls the 90% figure a “half truth.”

    “ATF officials challenge the suggestion that Mexico only sends them guns they suspect are from the United States. In fact, the ATF found about a quarter of the 90 percent were made in other countries and then taken illegally from the United States into Mexico.”

    Other reports correctly indicate that a lot of the drug cartels firearms are coming from China and other markets.

    • Crash December 18, 2012 at 3:04 pm #

      ‘Liberal slanted’ politifact? Where did you get that conclusion from, “Unskewedfact”?

  5. DocMerlin December 18, 2012 at 4:32 pm #

    This is likely spurious, as the assault weapons ban didn’t actually ban anything meaningful, it mostly just banned cosmetic features and made standard-cap magazines more expensive.

  6. TJIC December 18, 2012 at 5:01 pm #

    This is an accidental correlation, as the AWB didn’t outlaw any classes of guns – it merely said that guns could not have certain cosmetic traits.

    The day before the AWB passed, one could buy a semi-auto AR-pattern rifle in .223 with a 30 round magazine.

    The day after the AWB was signed, one could buy a semi-auto AR-pattern rifle in .223 with a 30 round magazine…as long as it didn’t have an (utterly useless) bayonet lug.

  7. Gordon Biggs December 18, 2012 at 6:51 pm #

    My political leanings are fairly conservative, but I am also a realist and I do respond to data. If the study was indeed accurate, then it shows a strong enough correlation to be convincing. I agree that correlations are not always causations. But sometimes they are, and in this case, absent any other cause, the authors’ conclusions are pretty convincing on the face of it.

    If someone can bring up a solid counter-argument, such as an additional legislative, economic, or social change that happened to occur at the same time as the expiration of the assault-weapons ban, then I’d question the study. No matter what quibbles earlier writers have had with things like the definitions of “assault weapons,” my guess is that the legislation had a real and noticeable effect on restricting the free availability of the sort of weapons used by drug dealers.

    Regarding facing up to facts: I would have thought that other right-leaning free individuals would have wised up a bit after the November election results. Leading up to the election, the right puffed itself up and looked at selective data and marketing announcements, while the left worked a tough game and stuck to its job: fooling voters into thinking that another 4 years of BO was good for the country. Ugh.

    One of the benefits of being a right-leaning thinker is the freedom to disagree with others. (This is not so on the left, where holding the communist party line is job #1 and if you disagree you are ostracized.) Si I heartily disagree with those of you who, like TJIC, simply take a position and stick with it, using single-item examples in opposition to what seems like a remarkable study with hard-to-avoid findings. There are still those of us on the right who believe in science and common sense. (Yes, including evolution.) Let’s face up to the reality around us and do better.

    Does this mean that I think assault weapons should be banned? NO. It means we should sit down and think seriously about the data and underlying problems.

    One interesting fact is that the murder rates in the US seem to be closely correlated with the criminalization of mind-altering recreational ingestible substances. The per-capita murder rate peaked during Prohibition and plummeted after 1933 when the law was repealed. The murder rate began to peak in the 1960s as the demands for recreational (and harder-core) drugs increased. Unless we get a better handle on the core pathology we will continue to deal with the symptoms only, and end up no better than we started out.

  8. Frank Youell December 19, 2012 at 2:45 am #

    “The per-capita murder rate peaked during Prohibition and plummeted after 1933 when the law was repealed. “

    No. In 1933, the murder rate was 9.7 per 100,000. In 1934 (with alcohol legal), the murder rate was 9.5 per 100,000. The murder rate does continue to fall after 1934. However, there is no ‘murder cliff’ associated with the (abrupt) ending of prohibition.

  9. Ropingdown December 20, 2012 at 3:57 am #

    Considering that they’ve just found a personally-purchased handgun that was sold to the number two agent of the Phoenix ATF office, George Garrett…at the site of a Sinaloa Cartel shootout…I’m not sure people who study US gun trafficking know what they’re dealing with. I think academics have a hard time imagining the world as it is. Messy.

  10. Justin G December 25, 2012 at 4:06 am #

    I think you forgot to mention they also have 4x the assault rate, 2x robbery, increased crime across the board.

  11. Steve September 7, 2013 at 1:02 am #

    If you took the time to read this article then read this one too:

    – Unless you don’t listen to both sides before making a decision.