Back to my web page

Assault Weapons: Did The 1994 Federal Ban Do Any Good?

Back in 1994, the news media were hammering away at the NRA for their opposition to a ban on assault weapons. There were a few lonely souls who kept pointing out -- in the few newspapers that would print both sides of the argument -- that there wasn't really that much criminal misuse of assault weapons. For that reason, the proposed law wouldn't make much of a difference in murder rates.

Now we have a new and persuasive report that echoes those lonely voices -- and the Clinton Administration paid for that report. The National Institute of Justice is the research arm of the U.S. Department of Justice. Among its activities is providing money to academics to study crime control -- and I am pleased to report that the taxpayers seem to be getting their money's worth. The 1994 federal assault weapons ban included a requirement that the Attorney General provide a report to Congress within 30 months evaluating the effects of the ban.

The National Institute of Justice gave Jeffery A. Roth and Christopher S. Koper, two researchers with the Urban Institute, a grant to determine what the 1994 federal assault weapons ban actually did. My reading of the report shows that Roth and Koper are honest enough, and careful enough, to find conclusions that cannot have sat well with Attorney General Reno and President Clinton.

For those of you who haven't studied statistics, think about a rare event, such as hail falling on your house in California. (I've only had hail fall twice at my house in five years.) Someone comes to your door, and sells you HailShield™ a device that prevents hail. "Gee, those hailstorms do some damage to the paint on the cars, and I've had two in the last five years. Maybe I could use HailShield™."

Being a gullible sort (much like Congress was in 1994), you spend $20,000 to buy this wondrous gadget. In the five years after you install HailShield™, you have one hailstorm. Did HailShield™ work? Imagine that you are the average blow-dried television reporter. "Well, yeah -- I had two hailstorms in the five years before, and only one in the five years afterwards. Wow! A 50% reduction in hailstorms!" The statistician, however, asks, "Did the gadget cause the reduction? Or was this just coincidence?"

As the events that you are studying become rarer, it gets harder to say with any confidence that a change was "statistically significant." In some parts of the Northeast, where it hails a lot, you could probably figure out in a year or two whether HailShield™ actually works. Where I live in California, it would take a long time to have any confidence that HailShield™ works.

Statisticians have all sorts of methods for figuring out whether a change was coincidence or not. If a change was "statistically significant," it means that it was probably not a coincidence; if it was "statistically insignificant," the change probably was a coincidence.

Murders with assault weapons are much like hail where I live in California -- not unknown, but not very common either. The first surprise of this National Institute of Justice report was an admission on the very first page that they had a hard time "discerning the effects of the ban" at least partly because "the banned weapons and magazines were rarely used to commit murders in this country" before the 1994 ban.[1] Well "Surprise, surprise!" as Jim Nabors used to say on the Gomer Pyle TV show.

Why were these weapons so rarely misused? "As shown in exhibit 1, about half the banned makes and models were rifles, which are hard to conceal for criminal use…. Further, the banned guns are used in only a small fraction of gun crimes; even before the ban, most of them rarely turned up in law enforcement agencies' requests to the Bureau of Alcohol, Tobacco and Firearms (BATF) to trace the sales histories of guns recovered in criminal investigations."[2]

The months of debate about the ban also had some unintended consequences. High demand drove up prices, and in response to the increased demand and prices, "production of assault weapons surged in the months leading up to the ban." Annual production of five categories of assault weapons, "AR-15s, models by Intratec, SWD, AA Arms, and Calico-and legal substitutes" more than doubled.[3] Once again, gun control groups have proven to be America's best gun salesmen.

Roth and Koper conclude that rising prices reduced the use of assault weapons by criminals, based on a drop in the number of assault weapon trace requests received by BATF. (When police departments seize a gun during a crime-any crime-and they ask BATF to find out all the lawful buyers and sellers, this is a "trace request.") But Roth and Koper also warn that trace requests are not typical of gun crimes, and so it's not clear how much of the drop in criminal use was real. But the real surprise is the report's conclusion that criminals reduced their use of assault weapons by 9-10% "due to substitution of other guns for the banned assault weapons in 1995 gun crimes."[4]

If criminals substituted other guns for assault weapons, doesn't this make the streets safer, because assault weapons can kill so many more people than regular guns? So Roth and Koper tried to figure out if the ban reduced the number of victims per mass murder, the number of wounds per gunshot victim, and the number of gun murders of police officers.[5] After all, if the bad thing about assault weapons was those big magazines and the ability to spray bullets everywhere, you would expect to see mass murders decline -- or at least the victims would have fewer holes in them!

So what did they find? They found a 6.7% reduction in murder rates in the 15 states where the federal ban could have made a difference. (I would explain why just they studied just those 15 states, but you would fall asleep part way through.) But Roth and Koper also admitted that this reduction was not statistically significant. Because assault weapons had been used in a tiny percentage of murders before the ban, "it is highly improbable that the assault weapons ban produced an effect this large…."[6]

What about the effects of rapid fire and large capacity magazines? "The ban did not produce declines in the average number of victims per incident of gun murder or gun murder victims with multiple wounds."

What about "protecting police officers," the excuse offered repeatedly for the ban? There was a decline in assault weapons used to murder police officers, but Roth and Koper also admitted that "such incidents are sufficiently rare" that it impossible to determine whether the law reduced total gun murders of police officers.[7]

It's possible that the federal assault weapon ban did some good -- but if it did, the effect was so subtle that even at a national level, statisticians couldn't measure it over a 24 month period and honestly say, "Yes, it seems to be helping."

Clayton E. Cramer is an historian. Encourage your local public library to order a copy of his fifth book, Concealed Weapon Laws of the Early Republic: Dueling, Southern Violence, and Moral Reform (Praeger Press, 1999). His web page is at


1. Jeffrey A. Roth and Christopher S. Koper, "Impacts of the 1994 Assault Weapons Ban: 1994-96," NCJ 173405, (Washington: National Institute of Justice, 1999), 1. You can find this report here.

2. Roth and Koper, 2.

3. Roth and Koper, 4-5.

4. Roth and Koper, 6.

5. Roth and Koper, 7.

6. Roth and Koper, 8-9.

7. Roth and Koper, 9.