ATF's 2022 Firearms Trace Data Is Out
Trying to make sense of the most maddeningly incomplete Federal data release.
The ATF has an annual release of data covering all “crime guns” that were recovered and traced in the previous year which they verrrrrry quietly released for 2022 a few weeks ago. The data was released with so little fanfare that ATF did not even realize that the links to download national data all went to 2021 rather than 2022 data until this enterprising blogger told them a week later.
The ATF firearms data has information on each firearm that was recovered and traced to the source where it was initially purchased. You can explore all that data in some historical detail on our firearms data dashboard, for what that’s worth. Here is how ATF defines what it is they’re tracing:
Firearm traces are designed to assist law enforcement authorities in conducting investigations by tracking the sale and possession of specific firearms. Law enforcement agencies may request firearms traces for any investigative reason, and those reasons are not necessarily reported to the federal government. Not all firearms used in crime are traced and not all firearms traced are used in crime.
Firearms selected for tracing are not chosen for purposes of determining which types, makes or models of firearms are used for illicit purposes. The firearms selected do not constitute a random sample and should not be considered representative of the larger universe of all firearms used by criminals, or any subset of that universe. Firearms are normally traced to the first retail seller, and sources reported for firearms traced do not necessarily represent the sources or methods by which firearms in general are acquired for use in crime.
It’s all useful and interesting though the data is maddeningly incomplete in certain areas. The firearms trace data gives a standardized fact sheet of firearm traces from each state which is interesting but isn’t particularly meaningful without greater context.
The real meat of the data release is the national totals of firearms recovered and sourced. There were nearly 400,000 crime guns recovered and traced to a source in 2022, a 10 percent increase from 2021, a 47 percent increase from 2019, and a 172 percent increase from 2010 when the data collection begins.
There are a ton more guns in the US now compared to a decade ago and they're showing up more often in crimes. The increase in the number of recovered and traced crime guns is not uniform throughout the country. California had only a slight increase from 2021 to 2022 and was the source state (where the firearm was initially legally purchased) for roughly the same number of recovered firearms in 2022 as in the years just before COVID. Texas, by contrast was the source state for 9,000 more recovered firearms in 2022 than in 2021, accounting for about a quarter of the national increase and a huge leap from the state’s total before COVID.
The change in firearm recoveries doesn’t inherently fit on a neat red-blue matrix though. Kentucky had only a small increase from 2021 to 2022 while Rhode Island and Colorado had sizable increases.
The other interesting nuggets (to me) from this report revolves around what ATF calls “Time-to-crime”. This is a measure of the amount of time between when a recovered firearm was legally purchased and when it was recovered in a crime (though, again, not all recovered firearms were recovered in crimes and not all firearms used in crimes are recovered and traced).
The average time-to-crime for firearms has plunged in recent years while the share of recovered firearms that were recovered within 2 years of purchase increased by 15 percent between 2019 and 2021. That share actually fell in 2022 after surging in 2020 and 2021.
The average national time-to-crime rose slightly in 2022, from 6.24 years in 2021 to 6.34 years in 2022, but the average recovered firearm in 2022 was still far newer than it was in 2015 when the average was above 10 years.
So we have new guns making up a larger share of all guns recovered and traced by ATF while the average age of a traced gun has fallen 40 percent in 7 years.
It’s a little bit tough to say what it all means. There were a ton more guns recovered and traced in 2022 than a decade prior, but the increase of about 35,000 was smaller than the increases in 2020 (~+38,000) and 2021 (~+56,000). Moreover the declines in the share of recovered guns that were new likely reflects the same trend that the FBI’s background check request data (NICS) is picking up: namely that new gun sales are falling after a post-COVID surge.
I wish I had a more constructive, in-depth bottom line to this data release, but the data lacks a level of granularity that would allow for more concrete conclusions. I’d love to know what share of homicides with a recovered firearm involved a new gun, the types of crimes that involved guns from Texas that are recovered in California, and the types and sources of crime guns that were recovered in New Orleans.
Sadly, the ATF data doesn’t go this deep.
There is no data on individual (anonymized) recovered firearms nor is the information on time-to-crime categorized by type of crime to see if new guns fueled the increase in homicides in 2020. Additionally, little city-specific data is available on an annual basis. The ATF did release data on some major cities in 2019, but the list was not exhaustive and has not been updated since.
The ATF data is interesting and potentially important, but the lack of depth to the data releases means its usefulness in answering questions about firearms and gun violence in the United States is unfortunately limited. As with almost all crime data, this ATF data can add to our understanding of the nation’s crime trend but its flaws require it to be taken with a grain of salt.
Useful summary Jeff
It immediately struck me -- is the _average_ time to crime the interesting bit of information here? if 4 people buy handguns and then immediately rob liquor stores, and 1 guy steals his grandfather's handgun which was registered 30 years ago and has never been involved in crimes, and then robs a liquor store with it, the average will tend to obscure attempts to discover if 'people buying guns so that they can go commit crimes with them' is a real problem or not.