Since 2008, the United States has spent more than $3 billion to help stabilize Mexico and stop its surge in extreme violence. The U.S. gun industry and interest group lobbyists undercut these efforts by advocating lax regulation and law enforcement. This created chaos south of the border, whose impact has already impacted the United States, thereby increasing the illegal drug trade and organized crime.
Meanwhile, the 2003 TIAHRT Amendment established a federal neighborhood of gun tracking data sharing made it difficult to track illegal U.S.-Mexico gun trade to study these effects.
We are professors and investigative journalists in economic development and we spent a year sifting documents to follow the flow of illegal weapons trafficked from the United States to Mexico and quantifying this traffic effort.
This is what we do.
Summarize the data together
We collected records from many sources and created a gun database that was sold by licensed U.S. gun dealers and then trafficked to Mexico. These businesses have been licensed by the U.S. government to sell or manufacture and sell guns and ammunition in the United States, including independent gun stores, chains and pawns.
We collected two sets of data obtained through a request for information to the Defensa nacional (called Sedena) of the Mexican Secretariat and orchestrated Mexico by blocking American weapons:
We track data records from two other sources:
Hankterist group Guacamaya leaked information from the Mexican government in 2022, including gun tracking data from November 2018 to December 2020, from the U.S. Alcohol Agency, national analysis, program and intelligence centers for tobacco, guns, guns and explosives. This dataset contains information about more than 24,000 guns; 15,000 of them are purchased in the United States
Our data sets generated from 100 court cases in the United States involve small arms trafficking from the United States to Mexico and contain information about more than 4,200 guns. It includes context details not found in many leaked data. For example, about 1,900 instances include purchase dates.
Of the 28,000 records in the combined trace record, nearly 13,000 were associated with specific licensed gun dealer addresses in the United States in most years, the ATF was unable to determine the final buyer among the end buyers of its U.S. purchase guns that it recycled in Mexico. We removed records that seem to be dead ends related to the address of the large manufacturer and kept 9,014 records, which we can confidently say goes back entirely to a particular gun seller.
We also used a request from a nonprofit advocacy group Brady: United, through the Information of Inlover Act, to request its gun store transparency program, initially a dataset of data on gun violations against gun violence. Between 2015 and 2018, 4,000 regulations that violated the federal gun laws violated the ATF inspection report, including warning letters, warning meetings and licence revocation recommendations.
Using these four datasets, previous estimates and two other studies, we modeled the number of guns trafficked across borders each year. This allows us to understand the characteristics of gun trafficking to Mexico from multiple perspectives.
We have two main goals:
Weapon flow
We started using the so-called capture turn method to start analysis using the weapon count. This type of counting is often used to estimate the number of animals in the wild, where a full count cannot be performed: capture, mark and release population samples, and then a second sample is taken to see how many labeled individuals are recaptured.
We applied this method to SEDENA and leaked ATF data. The SEDENA dataset contains weapons captured by Mexican troops, while the leaked ATF dataset contains guns seized by Mexican agencies. For ideal capture harvest comparisons, the leaked data will include all SEDENA records, but only 26.5% overlap. Despite its imperfection, this allows us to develop econometric equations to estimate the annual cycle of illegal guns in Mexico.
We combine this with estimates derived from five other sources:
We used an approach similar to the average of political voting, which also incorporates multiple sources to simulate possible trafficking rates. Our model estimates that between 72,819 and 258,101 guns were sold in the United States and were trafficked to Mexico in 2022.
This led to an intermediate estimate of 135,000 guns trafficked from the United States to Mexico in 2022.
To understand the impact of this weapon flow on people living in Mexico, we compare trafficking estimates with homicides in Mexico, from the United Nations Office on Drugs and Crime through the World Bank, 1990-2022 through the World Bank.
We refine our model by looking at the differences in this relationship during the 1994-2004 U.S. federal assault weapons ban. This suggests that a 1% increase in trafficking in persons will lead to a 0.48% increase in homicide rate. According to ATF numbers, more than two-thirds of guns recycled in Mexico and traced back to buyers originated in the United States. Our model shows that only about 6% of these weapons are legally imported.
We assume that a large number of illegal seizures may cause police to buy more guns, and armed police may push criminals to buy more weapons, creating an arms race. Our test results support this theory: We found that from 2006-2018, police gun orders increased by an average of 10%, resulting in an average of 1.4% increase in illegal weapons seizures the following year. And, instead, in those years, 10% of illegal weapons seizures rose, and the police gun orders increased by 18.5% the following year. Economic climate and non-irrigation crime rates may also be related to the purchase of police guns, but we do not control those factors.
Supervision and law enforcement
We also use court case data to track dates of guns purchased in the United States and dates they often reclaim in or on their way to Mexico. This allows us to verify leaked ATF data. We exclude weapons purchased or sold for trafficking to Mexico only through the influence of undercover law enforcement personnel.
About 2,900 shots in the court case data have a serial number, 19 of which match leaked ATF gun tracking data. Details in court records, such as recovery locations in Mexico or dealers who purchased guns, are the same details in the leaked data. These 19 games came from nine court cases in five states, and court cases were independent of leaked data. This confirms the authenticity of the leaked data.
We also used trace data to test eight gun laws that require reporting at the point of sale to test what impact these laws may have on trafficking in Mexico’s guns.
We calculated the possibility that guns remain in the illegal market. We control for possible spillover effects, such as a state’s gun sales suppression, pushing buyers to neighboring countries. One of our calculations uses distances between countries to measure the impact of gun sales and state laws.
We found:
Most gun laws reduce the time spent on the illegal market.
Clearing the law requires that gun shop owners or states keep records of purchases or background checks, resulting in a 126% to 257% increase in the likelihood of illegal firearms being recycled on any day.
In the surrounding countries, more gun laws increase the likelihood of guns sold in the state by about 1.9%.
Our findings coincide with previous research suggest that stronger gun purchase laws in the United States reduce the number of guns that ultimately end up in illegal markets at the national and state levels.
We also assume that law enforcement has known that some federal gun licensees related to weapons records in the data already know that we may be able to measure the impact of law enforcement actions.
To test this, we used the Brady dataset of 2014-2018 ATF gun dealer code violations to match the tracked guns to the licensee.
We analyzed the data using three different methods, all three showing that more citations in violation of gun laws resulted in fewer illegal guns.
Furthermore, we found from 2014 to 2018:
The ATF citation reduced the number of guns sold to traffickers by 20% to 44%.
Licensees accepting law enforcement lawsuits - warning letters, warning meetings or license rulings to warn - contribute 5% to 21% to guns in U.S.-Mexico traffic.
The ATF issues law enforcement lawsuits to about 12% of gun support every year.
Back to the United States
In 2024, the U.S. ATF regulated more than 75,000 federally licensed gun dealers, including independent stores, chains, local brokers and manufacturers.
Many traffickers and gun licensees in court are linked to U.S. crime, and nearly 300 illegal weapons in these data sets match the licensees in the ATF's Letter 2 program, which recovered from U.S. crime scenes within three years and had 25 or more guns sold within one year.
Of the two-year data we obtained, about one-fifth of the list of demand plans for 2022 and 2023 matched the Mexican Crime Gun. Overlapping shows that our impact on Mexican criminal guns, including more compliance checks, leads to fewer illegal guns also applicable to the illegal U.S. market.
Read the full survey: Mexican drug cartel uses hundreds of thousands of guns purchased from licensed American gun stores - Come on in Mexico, U.S. drugs and border immigrants