We’ve heard for years that there are more guns in the U.S. than people, but a precise accounting remains elusive. Federal legislation that would track gun sales or establish a nationwide handgun registry has been proposed — to much resistance from the gun lobby. In lieu of exact figures, we have gun owner surveys, industry disclosures, and federal gun background check figures, none of which are comprehensive.
Pinpointing the number of guns in circulation could help us better understand the relationship between gun sales and gun violence. Researchers have consistently found that more guns means more gun deaths, but gun rights advocates continue to argue the opposite. Meanwhile, annual gun deaths have soared to unprecedented levels.
Here, we’ll try to quantify America’s civilian gun stock, explaining through the process why it’s so complicated to do so — and how the numbers we have correlate with gun deaths.
How many guns have been manufactured for the U.S. market?
Every gun that’s manufactured by a licensed U.S. gunmaker has a paper trail. But what happens with that paper trail — and who has access to it — has been the topic of hot debate.
There is currently no requirement that federally licensed dealers report sales volume to the Bureau of Alcohol, Tobacco, Firearms and Explosives, so we don’t have complete data on individual purchases. But we do know how many guns are shipped to FFLs from licensed manufacturers every year, because gunmakers are required to report those figures to the ATF. The agency releases gun production numbers, along with import and export figures, in its Annual Firearms Manufacture & Export Report. “Production” is defined by the ATF as “firearms … manufactured and disposed of in commerce during the calendar year.”
According to historical ATF data, more than 465 million firearms have been produced for the U.S. market since 1899. This figure includes imports from foreign gunmakers but excludes exports by domestic gunmakers.
This data includes guns that are purchased by law enforcement, but not the military. The Small Arms Survey, a Switzerland-based outfit that publishes periodic reports on the global gun stock, estimated in 2018 that local, state, and federal police forces in the United States have just over 1 million firearms.
Why are exports excluded?
This story is focused on the guns that are in circulation in the U.S., so we’re not including guns that have been shipped to other countries. But the U.S. is the top exporter of small arms, with more than 25 million weapons shipped abroad since 1899, according to the ATF data.
The biggest exporters of guns, by ATF’s count, are also among the top domestic sellers: Sig Sauer; Sturm, Ruger & Company; and Smith & Wesson.
Do manufacturing figures equal the number of guns in circulation?
No. Production figures don’t account for how many firearms fall out of circulation each year. Guns are durable, and if maintained properly they can last a century or more. But they do rust and break. They’re also confiscated by police and destroyed, and illegally exported to other countries, said Philip Cook, a professor emeritus of public policy at Duke University, who first sought to quantify the nation’s guns in the early 1990s by surveying gun owners.
As he compiled his gun census, Cook realized that he had to come up with a number to subtract from the total civilian gunstock each year to account for broken, confiscated, and illegally exported guns. He said he “struggled” with what this attrition rate should be. “We know it’s not zero, and it would be unlikely to be as high as 5 percent,” he said. “I used 1 percent as a conservative ‘what if’ calculation.” Cook found that the resulting figure aligned with his survey respondents’ reports of how many guns they owned.
The ATF provides a cumulative number of firearms manufactured, imported, and exported for 1899 through 1945, but no annual breakdowns. If you apply Cook’s attrition rate annually starting in 1946, you get a loss of more than 112 million guns, leaving about 352 million guns in circulation.
Some researchers think Cook’s attrition rate could use an update. “Whether that 1 percent figure makes sense, and whether it should be reduced in the last decade or so when gun sales have surged, is an open question,” said Deborah Azrael, a public health researcher at Harvard who conducted her own landmark survey of gun owners in 2016.
In addition, not every gun manufacturer reports their commercial gunstock to the ATF every year, even though it is required by law. On average, 30 percent of active licensed gunmakers neglected to file manufacturing reports between 2016 and 2020, according to an ATF report published in May 2022. That sounds like a lot, but the agency says most non-reporters are smaller manufacturers whose output accounts for “a relatively small percentage of overall production.”
In a 2013 working paper for the Small Arms Survey, economist Jurgen Brauer estimated that about 320,000 newly produced weapons were not reported to the ATF between 2001 and 2010. That’s less than 1 percent of the official total.
Finally, the ATF data does not account for 3D-printed guns and most guns assembled from kits, because until recently, DIY firearms were not required to be reported to the agency. A rule that took effect in 2022 requires privately made firearms to be imprinted with serial numbers, which may result in more accurate information about their manufacture going forward.
How does the ATF’s figure compare to other available estimates?
Many news organizations and researchers use FBI background check figures as a proxy for the number of guns that enter into circulation each month. But background checks do not correspond one-to-one with sales, for a variety of reasons. Some are concealed carry permit applications and renewals, people can buy multiple weapons with a single background check, and at least one state — Kentucky — runs background checks on permit holders once a month. We compared FBI background check figures over the last 25 years to the ATF’s gun manufacturing and import figures over the same period, and background checks outnumbered gun production 2 to 1.
The National Shooting Sports Foundation, the gun industry’s trade group, publishes firearm production reports that draw on a host of sources, including the ATF, the U.S. International Trade Commission, and the Congressional Research Service. In 2020, the group reported that “the estimated total number of overall firearms in civilian possession is 433.9 million.” That figure does not appear to account for guns leaving circulation.
How has gun production changed over time?
American gunmakers produced more than 9 million guns a year on average between 2016 and 2020, while U.S. entities imported just under 5 million firearms annually on average, the ATF data shows. In 2020, the totals were 11 million guns produced and 6.8 million imported.
Annual manufacturing and import figures have historically been much lower. Fifteen years ago, around 7 million guns in total were manufactured and imported in a single year. That ramped up considerably after the election of Barack Obama, the nation’s first Black president. Total firearm production and imports jumped 30 percent between 2008 and 2009, and never returned to pre-2008 levels. Since 2009, combined production and imports have risen 152 percent.
Other surges in manufacturing and imports can be seen between 2011 and 2012, a period that saw Obama run for and win a second term (38 percent), and 2015 and 2016, when Donald Trump declared his presidential candidacy and won the election (25 percent). The biggest year-over-year jump on record was between 2019 and 2020, when a pandemic-related gun sales surge saw an unprecedented 40 million guns fly off the shelves (62 percent). (Pandemic demand likely explains the biggest year-over-year jump in imported firearms: 71 percent between 2019 and 2020.) Conversely, demand for guns appears to have waned between 2016, the year of Trump’s election, and 2017, the year he took office (24 percent).
In 2000, according to the ATF, there were 1,397 guns manufactured for every 100,000 people in the U.S.. By 2020, that figure had jumped to 3,410.
Has the type of guns being manufactured changed?
The majority of guns manufactured in America are handguns, but that wasn’t always the case: Until the early 1990s, rifles and shotguns routinely outsold pistols and revolvers. Handguns accounted for 57 percent of domestic firearm manufacturing in 2021, the most recent year such data is available, with nearly 8 million guns produced, while long guns accounted for just 33 percent, with just under 2 million produced.
That tracks with a cultural shift in American gun ownership that saw fewer people buying guns for recreation and hunting and more buying guns for self-defense. Since 1990, more than a third of rifles produced in America have been semiautomatic rifles like AR-15s, according to the NSSF.
The last decade has also seen a rise in the manufacturing of so-called miscellaneous firearms, which are gun parts, like frames and receivers, that are “sold before being assembled with other components” to make a finished gun, according to the ATF.
What’s the relationship between gun production and gun deaths?
When we charted gun manufacturing and imports alongside annual gun deaths going back to 1968, we found that when gunmakers ramped up production, gun deaths rose.
Researchers cautioned us that just because an increase in gun manufacturing is associated with an increase in gun deaths doesn’t mean it is the cause. But in many instances, gun deaths and production peaked in the same year.
The year with the highest gun death rate on record up until that point, 1974, also saw the most guns manufactured until then. Production fell slightly over the next decade, and so did the gun death rate. In 1993, a new all-time record for production and imports coincided with the second-highest gun death rate on record. Over the next five years, production and imports fell 42 percent, and the gun death rate dramatically declined — from 15 per 100,000 in 1993 to 11 per 100,000 in 1998.
Between 2008 and 2013, as production and imports surged 132 percent, the gun death rate remained essentially unchanged, at around 10 deaths per 100,000. But guns and gun deaths began rising in tandem again a few years ago. In 2020, the first year of the pandemic, gun companies produced and imported more guns than ever before, and there were more gun deaths than any previous year on record — more than 45,000. In 2021, gun deaths rose to a new record of 48,830. Firearm import figures for that year haven’t been released yet, but if they’re similar to 2020 figures, manufacturing and imports could reach 20 million in a single year for the first time.
The ATF releases production figures once a year, and as it does, we’ll be updating our count here.
When we looked into the relationship between gun deaths and gun production in more detail, we found that the relationship between gun production — particularly handgun production — and suicides is stronger than gun deaths overall. But correlation is not causation, veteran researchers said.
David Hemenway, a Harvard researcher who has repeatedly found that easily accessible guns means more gun deaths, said household gun ownership levels might be more important than the size of the civilian gunstock. “There have been periods when the gun stock has increased, but household gun ownership levels have been stable or falling,” he said, citing previous surveys of gun owners. He also said the number of guns in a single household might correlate more strongly to gun deaths than gun production. “Guns last a long time,” he said.
That may have changed since the pandemic, however. In 2020, police recovered almost twice as many guns with a short “time-to-crime” — in this case, recovered within a year of their purchase — than in 2019, according to ATF firearm trace data.
Daniel Semenza, a criminologist at Rutgers University–Camden who has also studied the correlation between gun availability and gun death, said he was struck by how gun production and gun deaths appear to rise and fall together at various points in the last half-century. “It’s not 100 percent correlation,” he said, “but it’s pretty close.” The findings confirm what researchers have observed for years, he said: “When there are more guns, it just increases risk,” in the form of accidental shootings, suicides, and arguments that turn deadly due to the presence of a gun.
The corresponding rise in production and gun deaths raises questions about how much responsibility the gun industry has for the country’s gun violence epidemic. That possibility is being debated in court, where gun rights groups are challenging a recent New York law that allows gun manufacturers and sellers to be sued for endangering public safety. Three other states have enacted similar laws, which attempt to circumvent the 2005 federal immunity that shields gun companies from most lawsuits arising from the criminal misuse of their products. The issue may end up in the Supreme Court.
How many guns do people near me own?
ATF production figures aren’t broken down at the state or local level. So while we know how many people are shot in our neighborhoods, we don’t know how many of our neighbors have guns. Once a gun is shipped to an FFL, that’s where the paper trail ends.
“It would be really useful to know how many millions of guns are in stores,” said Azrael, the Harvard researcher. The way the ATF defines production as “manufactured and disposed of in commerce during the calendar year” assumes that every gun shipped to an FFL is sold. While surges in demand and production, particularly during the pandemic, may give the impression that gun stores are selling out, that’s not necessarily the case, Azrael said.
The ATF does have state-by-state breakdowns of the number of machine guns, short-barreled rifles and shotguns, and silencers in civilian hands. That’s because they’re regulated by the National Firearms Act, and transfers must be approved by the federal government. But that excludes the vast majority of guns sold in the U.S.
Not having gun ownership data at the state level impedes gun violence research, Andrew Morral, a senior behavioral scientist at the RAND Corporation, recently told FiveThirtyEight. Without localized data, researchers can’t accurately assess the impact gun ownership has on firearm mortality.
A more precise accounting of the civilian gunstock is “really important,” Semenza said, adding that he would like to see researchers and journalists use an exact figure in lieu of ranges and estimates, even if there are caveats with the data. He said one reason they don’t is because historical production figures aren’t easy to find.
“This is in a bunch of different documents in really cumbersome language,” he said. “And the farther back you go, the harder that data is to wrangle.” Semenza also pointed to “this persistent myth” that the number of guns in America is an unknowable figure. “It’s almost a self-fulfilling prophecy.”
Graphics and additional reporting by Chip Brownlee.