Firearm Homicides Dropping. Assault Weapons Ban Not Correlated With Decrease In Homicides. No Need For New Restrictions.

I rarely ask this, but please link, forward, email, and favorite this post as widely as possible. See below for copying permission.

Murder in the United States1 is illegal, and has been for over two hundred years. Strong penalties, up to and including the penalty of death, are incurred by those who commit this heinous crime.

Yet, strangely, despite murder’s high illegality, there were in 2011 over 12,000 of them committed! The largest number of murders were in 1991, with nearly 25,000 of these frowned-upon unlawful incidents.

It is difficult to imagine a penalty more severe than death, so it remains a curiosity that so many murderers are found when such strong laws are in place. Perhaps this scourge can be eliminated by even tougher laws?, say death by torture? Or maybe by creating Executive Orders bypassing the hindrance of Congress and Constitutional safeguards? We must protect the children!

But never mind. Let’s instead look at the number of murders and what devices were used in their commission.

This (Fig. 1), according to the FBI2 and the United States Census, is the per-capita murder and non-negligent manslaughter (hereafter, in a slight abuse of notation, just called “homicide”) percentage from 1960 to 2011 (data for 2012 were not yet available). Overlaid are the same percentages for just those bodies accumulated from deaths by any type of firearm.

Figure 1

Figure 1

The first notable is that the two rates track closely, so that whatever is driving changes in the one is likely (but not certainly) driving changes in the other. The second remarkable feature is the abrupt increase in the turbulent 1960s, and the subsequent decline as the people responsible for these cultural excesses began to enter their 50s and 60s, and even to die off in the 2000s and 2010s. The percent in 2011 was the lowest on record.

I want to repeat that: The homicide rate in 2011 was the lowest on record since 1960.

And that’s true for all homicides committed with firearms of any type; that is, the lowest number of homicides by firearm was in 2011—although I only could find data from 1975-2011.

Let’s repeat that, too: the lowest number of homicides by firearm was in 2011.

Is your first reaction panic? Deep concern? Do you feel in your gut the need to do something? Then I suggest switching to a decaffeinated brand.

Now let’s look at the shaded region on the plot, which is the time the Orwellian-named “Public Safety and Recreational Firearms Use Protection Act”, i.e. “‘Assault’3 Weapons Ban”, was in force, from 13 September 1994 until the same date in 2004. There does not appear to be much correlation between this ban and the homicide rate: homicides both decrease and increase during the time which it was in force.

If you think there might be a correlation, then look at this plot (Fig. 2), which is the homicide rate for handguns, separated from other firearms4, knifes, blunt objects, and other instruments, which is a catchall including poisonings, strangulations, drownings, and similar forms of mayhem.

Figure 2

Figure 2

Each of these series exhibit the same pattern as the overall homicide rate. You may say the “assault” weapons ban lowered the rate of homicides committed with firearms, but then you have to explain why poisonings, strangulations and the like similarly decreased. It is of course possible that would-be murderers, feeling deprived of their loss of frightening-looking “assault” weapons were so forlorn that they lost the heart to add cyanide in their enemies’ tea, but it’s more likely that whatever was responsible for the general decrease in bloodlust caused both the decrease in firearm and non-firearm homicides.

Another possibility is that the number of shootings and other forms of violence remained constant, or even increased in recent years, but that people once wounded, because of improvements in medical science, are not dying at higher rates. This necessarily would decrease the homicide rate for the simple reason that if a person survives a shooting, he cannot be considered murdered. But this explanation is not plausible given that violent crimes (which includes attempted murder) are also on the wane, as shown here (Fig. 3).

Figure 3

Figure 3

There is some evidence that medical science might be the cause of some of the decreased homicide rate from this next plot (Fig. 4), which shows the percentage of homicides of all violent crimes: from a high of 3% to now around 1%. But since violent crimes as a whole are dropping, it is even more plausible that people are just becoming less bloody minded; i.e., less in the need of government control and regulation.

Figure 4

Figure 4

What should be particularly clear from the Figure 2 are two things: (1) homicides by any type are decreasing, and (2) handguns, and not “assault” rifles or indeed any other type of firearm, are always the most-used weapon. This plot (Fig. 5) emphasizes the significant role of handguns

Figure 5

Figure 5

This is a conditional plot, showing the percent handguns and other firearms are used in homicides. Notice that the percent of homicides committed with non-handguns actually increased during the life of the “assault” weapons ban. Handguns hover around 50%, a little more than 3 times as prevalent as non-handguns. If there are calls to ban anything, one would therefore guess it would be handguns which are demonized, not rifles and shotguns. Alas, the mind of the politician is a difficult thing to grasp.

Now let’s look at the same plot again (Fig. 6), this time including all types of killing methods:

Figure 6

Figure 6

The two rivals, equaling or exceeding in lethal importance to “assault” weapons (and other non-handguns), are knives and other types of weapons, such as poisons, strangulations, and fire. Yet we never hear even rumors of politicians wishing to ban fire. Though we do hear, all too often, of impaired officials banning children for pointing their fingers.

Summary: to use a phrase coined by Father Z, the “biological solution” appears to be lowering homicide rates quite well, with no government intervention required (or desired).

Now for the punchline. None of the statistics presented here are new or unknown. They are available to every politician, and indeed every lawmaker with the word “ban” on his lips knows them well (otherwise they are incompetent). Each of these people, like you now, knows that limiting “assault” weapons will do little to change the homicide rate. Yet still they want to ban. Why?

Could it be—this is reasonable to ask—that they have a different agenda in mind? Did we not hear many elected officials (from both major parties) tell us that guns are “only for hunting”, and did not some call for the confiscation of all guns? I suspect that this is the sole reason for the current flurry, the drive to “never let a crisis go to waste”: to strip citizens of their guns. Not all at once, for that would lead to rebellion, but slowly, incrementally, a death through many small paper(work)cuts.

Addendum The number of mass public shooting incidents per decade have also been decreasing and are uncorrelated with gun ownership. Full details about Fig. 7 are here.

Figure 7

Figure 7

Update Comparisons of the enormous, socially and racially heterogeneous United States with small, relatively more homogeneous European countries are not persuasive. “Denmark has lower gun violence!” somebody will proudly say, forgetting that Denmark on crowded day has only three-fourths the population of New York City.

Update As predicted. “Our children’s safety.

Update On one of the pages that linked to this story (thank you), one person called me an “economist.” Grr. His contention was that I did not focus on “assault” weapons, but only “other firearms.” This individual forgot that “assault” weapons are a subset of “other firearms”, and so are handled in that manner. Murders with them are limited to the number of murders of “other firearms.” And then—and this it the point—the FBI does not have readily available individual statistics separating “assault” weapons. Which makes me wonder how this individual (or anybody) could be so confident that “assault” weapons were such a scourge.



1Murder has not always been illegal everywhere. For example, citizens who were deemed “counter-revolutionary” or “anti-government” were routinely slaughtered, quite legally, in countries with international and national socialist governments. These countries, perhaps entirely coincidentally, banned their citizens from owning most or all firearms. I define murder as unlawful under the law given to us by a Higher Authority.

2Here is the data, which was compiled from several government sources, such as here, here, here, here, and here. Multiply Population by 1000. The data is not 100% certain. I found, on the FBI’s own site, different numbers for homicides for the same years. The discrepancies were never more than a couple of hundred coffins, but this still indicates imperfection. And that means we should lessen the certainty we have in any conclusions we draw from this data.5 If anybody out there can find mistakes or additions to this spreadsheet, please do so by amending it and emailing it back to me at I will then (at some point) redo and create a new post.

The FBI says murder and non-negligent manslaughter are:

the willful (nonnegligent) killing of one human being by another. Deaths caused by negligence, attempts to kill, assaults to kill, suicides, and accidental deaths are excluded. The program classifies justifiable homicides separately…Deaths of persons due to their own negligence, accidental deaths not resulting from gross negligence, and traffic fatalities are not included…

The murders on “September 11th” are not included in these charts, but the Oklahoma City bombings are.

3It has been said that an “assault” weapon is that which looks frightening to a lawmaker. Given the increase in non-manly politicians—by which I mean folks unable to appreciate a John Wayne movie—more and more weapons will be so categorized.

4A 2009 Attorney General report from California showed that fully automatic weapons, like “machine” guns, are used in 1% of crimes in that state. How could this be? Aren’t these guns illegal everywhere? If it’s one thing you can count on, it’s that a criminal has no respect for the law.

5Footnote to the footnote! Firearm laws vary across state and, within states, across time and across regions. The federal government, jealous of the power of the states, enacts its own laws, which also change through time and by region. The laws everywhere vary in strength, too. Further, citizens move from one state to another, or they move within a state to areas which have different laws. The compositions of the folks living in these great United States has also changed radically since 1960. Therefore, any statistical analysis—usually some high falutin regression model—which purports to have figured out the true influence of firearms regulations will be full of—what our esteemed Veep Joe “Wakka Wakka” Biden called—malarkey.

One thing we can say with certainty: threats to increase restrictions on firearms increase the sales of firearms: heres one example of many.

William M. Briggs, the “Statistician to the Stars!” is, and has been for many years, a proud NRA member. He received no consideration of any kind from anybody for this post.

Permission is granted to copy and reprint this post, under the proviso that my name and a link to my site accompanies it.


Firearm Homicides Dropping. Assault Weapons Ban Not Correlated With Decrease In Homicides. No Need For New Restrictions. — 124 Comments

  1. Joshua Postema,

    I did misread you comment. I apologize. There is no evidence in this blog to support my guess (a guess… hence maybe) of latent factors of any sort. None at all!

    Regarding my comments about feeling compelled to make changes with the anticipation of saving life – you see, you have infer much more than I have expected. Take it at face value. I am thinking of several great emperors in Chinese history.

    Whether/how people act on such feeling or whether they can come up with efficient, reasonable, and agreeable changes is another story.

    I have no illusion that it’d be possible to eliminate guns completely in the US during my lifetime. My main issue is that I haven’t seen any logical arguments or evidence to accept claims such as “it causes fewer deaths” and the one made by MattS.

  2. JH,


    Violence / murder have existed since the dawn of the human race.

    Therefore violence / murder are inherent to the human condition and predate the invention of the gun.

    Therefore eliminating guns won’t reduce violence / murder

    Where is the flaw in this argument?

  3. JH,

    Here’s an empirical argument against gun control.

    Stats from

    The murders most likely to be stopped by gun control are non-premeditated murders.

    If it were true that a significant percentage of non-premeditated murders occur only because a gun is present then you would expect guns to be more prevalent in those categories where premeditation is least likely.

    With one exception this is not the case. Except for brawls due to the influence of narcotics the percentage committed with guns is significantly lower for categories where non-premeditated murder is most likely.

    Percentage of murders committed with guns:
    Note: Felony murder here refers to the felony murder rule (any homicide committed in the course of a separate felony is automatically first degree murder). Felony murder is by law always considered premeditated.
    All murders: 67%
    Felony murder: 72%
    Suspected felony murder 64%
    All non felony murder 63%
    Romantic Triangle: 62%
    Child killed by babysitter 7%
    Brawl due to influence of alcohol 40%
    Brawl due to influence of narcotics 72%
    Argument over money or property: 58%
    Other Arguments: 60%
    Not Specified: 56%

    Again, if the presence of a gun were a significant factor in the occurrence of non-premeditated murder I would expect that the percentage of murders committed in the categories where premeditation is least likely to be HIGHER than the percentage of all murders committed with guns. This is NOT the case.

  4. Here are few link that I posted in a previous tread:

    See chart 4:

    per 100k the number of murder by other mean than firearm in the USA is a little higher than those found in Canada, UK, and Australia.

    But the number of homicide by firearm are 6x higher than Canada, 15x higher than Australia, and 30x the UK. The USA is the only country where homicide by firearms is higher than murder by other means.

    Here see the first chart p.37:

    The chart show that some regulation have some effect. Although gun violence decreased between 1980 and 1991, it decreased much faster following the imposition of background cheks and safer gun storage. When the gun compulsory registration law was pass the rate of death didn’t change from what was observed after the previous law was passed.

    Some regulation do have an effect on gun violence/death, and some don’t. Background check works, registry from seller also help the police while the registry for gun owner didn’t help at all.

    Here gun owner rate and gun death rate:

    The countries from the developed world with the highest rate of gun owner also have the highest rate of gun death.

  5. Sylvain Allard,

    No one in the US opposes background checks. In fact the NRA was the earliest proponent of background checks. You are correct background checks work.

    “But the number of homicide by firearm are 6x higher than Canada”

    But everything I have seen indicates that gun ownership levels in Canada are similar to the US. This indicates that something other than the guns themselves are the issue.

    “15x higher than Australia, and 30x the UK.”

    Both the UK and Australia passed total gun bans relatively recently. Both saw increases in homicide post ban.

    There are all kinds of differences between the countries with low rates of homicide and high rates of homicide. There is exactly zero evidence that guns themselves are the controlling issue.

    In the US guns are less prevalent in the types of murders where you would expect the presence or not to make the most difference if the presence of a gun made any difference at all then they are in murders over all.

    This is the opposite of what I would expect if the presence or not of a gun made any difference to whether a murder happens or not.

  6. DAV
    “Running down the street to buy a couple of gallons of gasoline to make a Molotov cocktail can be done in less than an hour. Takes almost zero planning. The same with running over people on a crowded sidewalk and doesn’t even require going shopping first. Just exactly how would you go about eliminating those threats?”

    If a premeditated act of mass murder were more easily enacted through the other means you suggest, then the law of least resistance suggests such means would be more prevalent than premeditated mass murder by firearm. And nowhere have I proposed methodology for eliminating these threats and nowhere have I proposed methodology for eliminating the threat of mass murder by firearm.

  7. MattS:
    “In an individual non-premeditated act a gun is seldom the easiest means.”
    Valid point, by the very definition of premeditated.

  8. Pingback: From John Lott « YouViewed/Editorial

  9. Pingback: Guest Post by William M Briggs: Firearm Homicides Dropping « Do_While(True)

  10. MattS,

    I really haven’t time to read everything and to think carefully. Here is what I can say about your comments. Thanks for not giving up.

    (1) Violence / murder have existed since the dawn of the human race.
    (2) Therefore violence / murder are inherent to the human condition and predate the invention of the gun.
    (3) Therefore eliminating guns won’t reduce violence / murder.

    Assume that (1) indeed implies that violence / murder are inherent to the human condition. Under (2), we can conclude that

    eliminating guns won’t REMOVE violence / murder.

    Nothing in (2) implies that in the absence of guns violence / murder won’t become smaller numerically (reduce) as stated in (3).

    Regarding your empirical arguments:

    If you want to show me that “gun control” (or no guns) doesn’t change anything, you need to first define
    (1) variable(s) that quantifies “gun control”. It can simply be an indicator of the existence of a specific gun control measure; and
    (2) variables that can somehow quantify the “change” of interest, for example, murder rate.

    You then collect observations on your variables. Statistics! Your conjectures and the data from year 2009 demonstrate nothing, and I won’t drawing conclusions based on data from a sample unit (year) of 1.

  11. @jshewey:

    I was discussing public relations strategy, not the second amendment.”

    Perhaps, but the fact remains that public relations are not as important as law, especially Constitutional law. If one of those organizations is attempting to pass laws that are good on PR and bad on Constitutional compatibility, they are definitively worse than those bad on PR and good on Constitutional compatibility; the law (and human nature) remain as they are.

    This is not to say that the Constitution is a perfect document, but that it still rules the day. It certainly has some timeless wisdom that we have forgotten today, like the value of an armed citizenry against a tyrannical government.

  12. MattS,

    «Both the UK and Australia passed total gun bans relatively recently. Both saw increases in homicide post ban.»

    Do you have any fficial data to support that, other than right wing blogs.

    This is what I found about Australia see graph p.16:

    The data shows a steady decline in murder in the last 20 years.

    This is for the UK

    There is a positive trend till 2003, until it started to fall.

    In the UK, a major law passed in 1997 while the murder rate was in a positive trend. It seems to have taken a few years before the law showed some efficiency.

  13. Pingback: Since 1968, “more Americans have died from gunfire than died in … all the wars of this country’s history.” « Sunset Daily

  14. Josh–
    Apples and Oranges. My comments you refer to were simply generalizations about public relations strategies, e.g., the effectiveness with which an entity uses the science of communications to dissipate a message, how to measure said effectiveness, how to evaluate the impact of said message, how to effectively use communications feedback loop.

  15. @jshewey:

    “Apples and Oranges. My comments you refer to were simply generalizations about public relations strategies, e.g., the effectiveness with which an entity uses the science of communications to dissipate a message, how to measure said effectiveness, how to evaluate the impact of said message, how to effectively use communications feedback loop.”

    We may be talking past each other a bit and I apologize if that is the case. My intent was to suggest that supporting what is right over what is popular, even at risk, is the best option. It is not inherently noble to move to a “middle ground” unless that middle ground is inherently noble. On this issue, I believe it not to be the case.

  16. so the homicide rate in the US, even after the recent fall, is 4 times that of the UK. Wonder why?

    According to a Canadian friend of mine, the murder rate among Americans of English ancestry is similar to England, among Japanese-Americans similar to Japan, and so on. States like Idaho, Montana, the Dakotas, Maine, etc. that border Canada have rates similar to Canada. He claimed that if you discount black and white descendants of the Southern “honor culture,” the US would have the same rate as Denmark.

    IOW, there may be cultural reasons. Many in the Southern honor culture came from the Scots borders, where murder and mayhem were routine in the period of emigration and “Blue Bonnets over the Border” was more than just a catchy bagpipe tune. The whole concept of duels over insulted honor or (in modern parlance) “dissing” is a cultural thing, regardless whether they draw their rapiers, their flintlock pistols, or a Glock-9.

    10 highest rates in the Union:
    Louisiana……11.2 (cf. New Orleans)
    New Mexico……7.5
    Maryland……..6.8 (cf. Baltimore)
    South Carolina..6.8
    Michigan……..6.2 (cf. Detroit)
    Missouri……..6.1 (cf. St. Louis)
    Tennessee…….5.8 (cf. Memphis)
    All are Southron states except MI (Detroit) and two border states much impacted by the Mexican drug cartels.

    10 lowest rates in the Union
    Rhode Island…1.3
    New Hampshire..1.3
    That is, New England and other highly Anglo and Scandinavian states, plus Hawaii, with its high Japanese population.

    by comparison,
    (UK has three times the murder rate of Japan!)

    Midwest states bordering Canada
    Washington…..2.4; Brit. Columbia…..2.5
    Idaho……….2.3; Alberta…………2.8
    Montana……..2.8; Saskatchewan…….4.1
    N. Dakota……3.5; Manitoba………..3.3
    Minnesota……1.4; Ontario…………1.5

    Oddly enough, violent crime in Canada also peaked ca. 1992, just like the US.

    Canadian Maritime provinces compared to upper New England
    Nova Scotia….1.7; Maine………..2.0
    Newf&Lab…….1.4; Rhode Island….1.3
    New Brunswick..0.9; Vermont………1.3
    PEI…………0.7; New Hampshire…1.3

  17. Excellent piece. Figure 4 shows homicides are now ‘only’ one percent of violent crime. As I recall, the former ‘Great’ Britain’s violent crime has increased 77% since their gun ban in 1998.

    I don’t think I’m comfortable living in a country, much less a town, where one percent of violent crime is reduced to near zero (murder) but the other 99% of violent crime (forcible rape, aggravated assault, and robbery) increases 77%! Besides the lifetime of mental anguish over being raped (and who’s to say it’ll happen only once in your lifetime in Obama’s Pleasantville?), I may live out my life in a vegetative state, in a wheel chair, or with severe disfigurement after a heinous aggravated assault (and who’s to say it’ll happen only once in your lifetime in Obama’s Pleasantville?).

    I think I’ll just keep my guns and let Obama’s voters worry about themselves.

    BTW, Figure 5 shows how more prevalent the use of handguns is over ‘assault rifles’. This is a statistic staring us in the face as to the real reason for the assault rifle ban: A tyrannical government would much rather face a citizenry with pistols than rifles.

  18. These figures are impressive, but they are essentially a red herring. Prof. Briggs, as an expert at picking apart wishcasts and illicit extrapolations “based on” data, should be the first person to realize that this information does not even address the core concern of those pushing for an assault weapons ban or clip size restriction right now. The point is not to lower the total number of deaths. No one is so foolish as to believe that handguns are not far and away the go-to weapon for murder. Few believe even that the total number of mass shootings (defined here as 4+ people) will decrease just because of assault weapons and over-ten-round magazines being banned.

    The point is to lower the number of casualties and injuries in a single shooting. That is what the ban on assault weapons and high-capacity magazines would be useful for, and it’s why people are pushing for it. Imagine, for instance, that James Holmes only had a ten-round Glock. Imagine if the guy who shot up Sandy Hook only had a ten-shot hunting rifle. How much damage could have been avoided? Prof. Briggs’ data takes none of this into consideration. It also treats a collection of thousands of single or double homicides (atrocious as they may be), many related to drugs or gangs and so forth, as being on par with the Sandy Hook massacre. So, again, there’s a red herring fallacy going on here–as well as a few begged questions and instances of non sequitur. Not the best I’ve seen from this blog.

  19. rank sophist: I’ll start by being evil, and say that policy should be about reducing the number of people murdered in the aggregate, not the maximum number murdered per incident. If Policy A results in 1000 total murders in 700 incidents and Policy B results in 50 murders in five incidents, gimme some B.

    The VA Tech shooter was using pistols with fifteen round magazines. Ten round mags meant he would have had to change mags 33% more often. So what? Nobody was able to shoot back, and the time window to realize that the guy you’ve been trying to hide from is reloading, and that now is your chance to rush him from the greatest possible distance you were previously trying hard to achieve is very small. If you don’t mind the mags falling on the ground/floor, you can change pistol mags REALLY fast.

  20. Ed,

    True or not as all of that may be, it doesn’t change the fact that this article is a fallacious use of data. Prof. Briggs has built a powerful case against an argument that does not exist. (Also, a policy that was actually designed to curb the total number of gun deaths would most likely have the NRA inciting war in the streets.)

  21. Maudi,

    You have to be careful about violence statistique. What is often used is the crime reported to police. There are reason to believe that citizen in the UK are more thrustful of their police than the US citizen. How many minority people will contact the police in Arizona for example? Why should they have any confidence in the police. Rape is without a doubt the less reported crime of all, but if police and justices show willingness to prosecute the perps it may be enough for more people to say that they have been raped.

    Besides the claim of 77% seems strange when you look at this figures:

  22. Pingback: Did Assault Weapons Ban reduce Homicides by firearm? | Politisite

  23. Pingback: A Few Curmudgeonly Notes « 36 Chambers – The Legendary Journeys: Execution to the max!