Because most philosophies that frown on reproduction don't survive.

Thursday, October 08, 2015

Is US Gun Ownership Causing a Wave of Gun Deaths?

With another senseless shooting in the news, there has been a wave of news analysis stories examining the connection between the huge number of guns owned by civilians in the United States and the number of people here who die by gunfire. Some of these arguments border on the tautological. Some articles, for instance, have been at pains to point out that having access to guns increases the chances that someone will have a gun accident, commit suicide with a gun, or assault someone with a gun. Well, obviously, one would be hard put to have a gun accident, or to kill oneself or someone else using a gun, if one did not have access to a gun.

Since many of these arguments seem to center around the idea that increases in gun availability cause increases in gun deaths, I thought it would be interesting to look at historical trends in US gun deaths of various types and compare them to gun ownership rates. While I'm fairly pro-gun, what I found honestly kind of surprised me.

Let's start by looking at murder. The US, as is well known, has a fairly high murder rate compared to other developed countries. A lot of those murders are committed with guns. Here's a chart showing the murder rate (number of murders per 100,000 in population) since 1981 for gun-murders and non-gun murders.

[These are drawn from the CDC data sets here and here.]

The spike in non-gun homicides in 2001 is the result of several thousand people being murdered by means of box-cutters and airplanes.

As you can see, although there was an increase in gun deaths during the height of the crack-cocaine inspired gang wars of the late eighties and early nineties, the overall trend of both gun murders and non-gun murders is down. Non-gun murders have been falling somewhat faster than gun murders, and the percentage of murders carried out by means of a gun is relatively high at 69% in 2007, the last year the CDC has data for.

As I said, the US murder rate is fairly high compared to other wealthy countries. Here's a comparison of the US gun and non-gun homicide rates with the homicide rates of the UK and Australia. I picked these two because their data is available in English (a matter of convenience for me) and because both countries had gun bans put in place after famous mass shootings, bans which are often pointed to as examples of what the US should do. Here's a comparison of the rates:

[UK data from here. Australian data from figure 2 here.)

Since the rates are so different in scale, it's interesting to index them to the rate of a starting year and see how they've trended.

As you can see, Australia has an indexed trend much like the US (declining less than our non-gun murder rate but more than our gun murder rate) while the UK has actually seen an increase in murder rate since 1989, though it's come off its record high a bit over the last ten years. I also marked a potential inflection point on the graph which is perhaps most interesting because it doesn't mark much of a change in the murder rate. In the 1996-1997 period, both the UK and Australia responded to horrific mass shooting events by passing draconian gun control legislation and removing a lot of guns from civilian hands.

The UK had already passed its first big round of gun control in 1988, in response to the Hungerford Massacre, placing severe restrictions on semi-automatic rifles (and not just the scary looking "assault rifles" which are so often discussed here in the US.) Then in 1997, in response to the Dunblane School Shooting, the UK passed much tougher gun control legislation, virtually banning civilian possession of handguns.

Australia had a similar legal change in the same 1996-1997 period due to the Port Arthur Massacre, in which a mentally disturbed man killed 35 and wounded 23 more.

Both countries not only restricted the sale and ownership of guns, but also instituted massive collection/buy-back programs which resulted in the collection and destruction of over a hundred thousand guns in the UK and over a million in Australia. The result was a widely touted reduction in gun fatalities, as shown by this graph discussion the situation in Australia:
[Attribution: "Gun deaths over time in the US and Australia" by Volutin - Own work. Licensed under CC BY-SA 4.0 via Commons -]

However, what this type of analysis misses is that guns were not actually that large a source of violence in these countries to start with. The Australian government's Institute of Criminology provides some graphs that look at the percentage of murders in Australia committed using guns over the long term, and at the relative trends in gun murders and knife murders in recent decades
Percent of Australian Homicides performed using a gun

Percent of Australian Homicides using gun vs. sharp instrument

Guns have never accounted for more than about 40% of murders in Australia, and in the early '90s, shortly before the ban, they only accounted for around 20%. The UK appears to have had pretty similar trends, though I haven't been able to find data as specific. By comparison, guns account for 60% to 70% of murders in the US in any given year out of the last 35.

Perhaps this explains why the banning of guns seems to have had no real impact at all on the already low murder rates of the UK and Australia. The UK saw a flat murder rate for thee years after its ban, then a significant increase, then a reduction back to pretty much right where it was at the time of the ban. Australia saw four years of flat murder rates, then a gradual decline fairly similar to the decline which had been occurring already before the ban. Indeed, the US murder rate dropped much more in the years after the UK and Australian gun bans than did the murder rate in the countries that actually banned guns. In 1995 the US total murder rate (gun and non-gun) was 6.7 times that of the UK and 5.4 times that of Australia. In the third year of those countries' gun bans, in 2000 their murder rates were virtually unchanged, but the US murder rate had fallen to 4.2 times that of the UK and 3.8 times that of Australia.

One of the things that doesn't get discussed much when making comparisons between the US and Australia in terms of gun policy is that the two countries were fairly different in terms of gun ownership and culture before the 1997 Australian gun ban. In the period directly after its gun ban, Australians turned in roughly 1,000,000 guns, which were destroyed by the government. Estimates are that this represented roughly one third of guns in Australia. [source] (Others are either still owned and used at gun clubs under increasingly restrictive gun laws, or have entered the "grey market" of illegal guns.) A little quick math:

In 1997 Australia had 18.5 million residents. If their estimate of 3 million total guns is correct, that represents one gun for every 6.17 people. The congressional research service estimated that in 2009 the US had 310 million guns in civilian hands, and it's possible to use ATF reports on the number of guns manufactured and imported into the US to take the estimate backwards and forwards in time. That gives me an estimated 247 million guns owned by US residents back in 1997. The US population at that time was 273 million. So the US had one gun for every 1.1 people, nearly six times as many guns on a per capita basis as Australia. This means that even pre-ban (when Australia had more guns than now and the US had less) the ratio of guns to homicides in Australia was slightly higher than in the US, and post ban it became much higher.

Now let me be clear: This does not mean that taking guns out of the hands of law abiding Australian citizens caused a murder wave. The number of murders was almost identical pre and post ban: 299 in 1995 and 308 in 1999. (I avoided 2000 when there was a small one year murder spike in Australia and 2001 when 9-11 represented a non-gun-related one year murder spike in the US.) That's what drives the apparent spike in the number of Australian homicides per 100,000 guns: the number of homicides per year remained roughly the same while the number of guns fell by a third. In the US, during the same period, the number of homicides fell (22,895 in 1995 to 17,287 in 1999) while the number of guns in civilian hands increased (236 million to 256 million).

This leads to the data than genuinely surprised me, even as a gun rights supporter. I had known that the US murder rate as a whole had been falling for the last twenty years, despite a large increase in the number of guns. While this seems like an argument against the idea that more guns equals more gun homicides, there is at least the potential counter-argument that since crime and violence is falling overall these larger social trends swamped the effect of increased gun availability. What I hadn't realized until I got into the CDC data (linked above) is that gun suicides and gun accidents have been falling as well.

The gun homicide rate is down 22% since 1986. The gun suicide rate is down 24%. The gun accident rate is down a whopping 67%. By comparison: The non-gun homicide rate is down a steep 45%. However, the non-gun suicide rate is up 8%. Let's look at all of these metrics indexed against 1986 so that we can see their relative changes more clearly.

The number of guns, and thus it seems fair to say gun availability, has increased steadily throughout this period. However, no type of gun related deaths has increased in a way that correlates with the increase in gun availability. The only trend which mirrors gun availability somewhat in recent years is non-gun suicides, but it seems near impossible to argue that the availability of guns is causing people to commit suicide using methods other than guns.

This, I think, provides a useful corrective to some of the rhetoric on gun availability and gun deaths. Clearly, gun related deaths are subject to gun availability. You can't accidentally shoot yourself with a gun if you do not have access to a gun. You can't shoot yourself or another person with a gun if you do not have a gun to shoot with. However, the number of guns in civilian hands has increased by over 50% since 1986, and during that same period the rate of accidental gun deaths has decreased by 67% while the rates of both gun homicides and gun suicides have decreased by over 20%. It's hard to make the case that the increasing number of guns in the hands of US residents is increasing the chances of gun deaths when what we actually see is that the number of guns per person is rising steadily (.78 guns per capita in 1987 to .98 guns per capita in 2007) while the rate of gun deaths is falling.  The CDC data cuts off in 2007, but FBI data on the overall homicide rate shows it falling significantly from 2008 to 2013 while the rate of gun purchases has doubled during the Obama administration versus the period before.

Does this mean that removing all guns, or virtually all guns, from our country would not reduce homicides, suicides and accidents? No. If you really could get all or nearly all guns out of the US civilian hands (a total of well over 360 million guns by now, based on ATF estimates) it would probably decrease violent deaths a bit. But even when the UK and Australia reduced gun ownership rates to levels far below what the US has had at any point in recent history, they saw virtually no impact on their total homicide rates. This probably puts change by this means well out of reach for the US. However, the rapid increase in the number of US guns has not led to an increase in the rates of homicide, suicide or even gun accidents. Indeed, the rate of all of those types of death has been dropping even as the number of guns per capita has increased. The continuing increase in the number of guns in the hands of US civilians has not led to a wave of gun deaths.


Anonymous said...

"However, no type of gun related deaths has increased in a way that correlates with the increase in gun availability."

What about the type of gun death that has led you to look into the relationship of guns to deaths at this time—mass murders (defined, I believe, as the murder of 4 or more people at any one time/by one individual)? The statistic that is often given is that Australia has had no mass murders since implementing their 1996 laws, and we are up to more than one per day, on average.

Darwin said...


I didn't try to dig into mass shootings because the data on them isn't very good, and there are so few that I didn't think correlations would be very useful.

If you want to look at some data points on the issue, there was an FBI report put out at the request of the administration which looked at "active shooter incidents" (which is what most people think of when they talk about a mass shooting) between 2000 and 2013. It found an increase during that period from 1 in 2000 to 17 in 2013, though with the highest number in 2010 (26 incidents). That's an increase, but in a pretty noise trend line. And since there were already 250 million guns in the US in 2000, it seems a little odd to plot a trend in which that level produces only one mass shooting but the increase from 250 million to 350 million results in a 17x increase in active shooter incidents.

The Homicide Trends 1980 to 2008 data here:

provides data on the percentage of homicides which involve more than two victims during that time range. That seems to show an increase in the percent of homicides which involved two or more victims from 0.5% of homicides in 1986 to 0.8% of homicides in 2008. If I multiply those rates by the CDC total homicide numbers, I get a rate of murder victims in multi victim homicides per 100,000 in population which is flat to declining from 1986 to 2007 at an average of 0.05.

Finally, there are mass shooting trackers put together from media reports. These get around the issue that FBI data is accurate on the number of people killed but sometimes doesn't include details of the incidents like the number of people in each shooting. Here's an example:

They define a mass shooting as one in which at least four people are killed or wounded. Their counts of total dead in mass shootings are:

2013: 502
2014: 383
2015 ytd: 379 (compared to the same time period, that's on track to be the same as 2013)

However, these media aggregated data sources don't go back very far, and it's harder to do this kind of work in retrospect, so I didn't go after it in order to see if there's a trend.

RL said...

The thing that always gets me with this stuff is while necessary to examine in a way that is empirical to combat alleged emperical claims, the reality is it's still largely playing into a false premise.

The problem is cultural at root. Each society or community will respect life and each other to varying degrees and in different manners. Comparing the US to other countries isn't going to be that great. Heck, looking at the differences between something like Texas and Illinois would be more telling. Or how about Cook County vs. the rest of Illinois.

I know some would immediately balk at that citing the porous nature of such a boundary and the ability of gun laws to be breached, but that's just proving the point that it's a cultural thing. There is Facebook meme that cites just by removing 4 or 5 big cities from the mix the US has a relatively low gun homicide rate. I didn't dig into it but it certainly seems plausible. After all would you feel safer from gun violence in a small town in Nebraska where there are 3 guns for every man, woman, and child or in a major city with strict gun control laws and only .25 guns per person?

Darwin said...


Agreed. I just ran into this piece (predictably from Vox) which contains the following asinine paragraph:

"The relationship between gun ownership rates and gun violence rates is well established. Reviews of the evidence by the Harvard School of Public Health's Injury Control Research Center have consistently found that when controlling for variables such as socioeconomic factors and other crime, places with more guns have more gun deaths."

Once your control for socioeconomic factors and crime? So, basically, once you adjust down the murder rate of Detroit so that its as low as rural Utah, taking out the effect of poverty and unemployment and urban decay and drugs and crime -- then you find that after accounting for all those other factors the number of guns appears to have some affect. Seriously? This is called analysis?

Though note to myself: In that article I did fine a link to a more recent CDC data set that goes through 2013. It shows that gun suicides have increased somewhat in the 2007 to 2013 period (though not as much as non-gun suicides -- apparently suicides are simply becoming more common) -- but gun homicides and gun accidents have continued to go down.

Michael Adams said...

Interesting piece, but I have a question. Was there a change is gun-ownership screening (waiting period, etc.)after the index year 1986 (I was too young in those years to notice one way or the other), that might have a correlation with the decline in gun-violence? I honestly don't know the answer, and I do not have the time to look into it personally, at the moment.
As I understand it, violent crime per capita is actually lower now, overall, than it has been in decades (perhaps ever?). However, the public perception is that we are in an 'ultra-violent' period of history. This perception, I think, makes it hard for both sides of a gun-control debate to dig into the details, which is worth doing, I think. Thanks for your time.

Darwin said...


I'm aware of two major pieces of legislation:

- The Brady Bill instituted a background check and a five day waiting period for gun purchases in 1993, however the background check became an instant background check with the institution of the FBI NICS database in 1998. This remains in effect.

- The Federal Assault Weapons Ban was passed in 1994 and expired in 2004. It banned "high capacity magazines" for both rifles and pistols, and also banned guns with more than one of a list of "military" features. The list, however, didn't really ban anything normally used in crimes, it was mostly a ban on military "looking" guns. The ban was not renewed when it expired in 2004.

Agnes said...

This is... interesting. I had no idea thatthe common concept "people in the U.S. have more guns" meant quite so much more guns. If I understand correctly, it means rather more than one gun for every adult (1.1 guns per capita back then, and increasing since...and children obviously don't have them, and without meaning to be sexist, I can't imagine all adult women own guns - sorry if I am mistaken). With this estimate, it seems strange to ask whether there are positive effects of gun ownership (does it actually occur that your average law-abiding citizen has to raise firearms to defend themselves from criminals?) Or is it mostly about honoring the traditions of your forefathers? Also, if so many people either owns guns or has someone in the family who does, who is the anti-gun voice?

Cojuanco said...


In response to your last question - the anti-gun voice tends to be people who live in urban or inner-suburban areas.

I think a lot of the gun control debates boil down to a tension between rural culture and urban culture. In the countryside, guns tend to be seen as just another power tool, like a drill or a chainsaw. You use guns to kill pests, or hunt for food, or for sport shooting. Yes, you can use guns to kill, rural people acknowledge - but it's no different than using a power drill to defend yourself.

In contrast, in urban areas, pests of the kind seen in rural areas, or things to hunt are rather less common to say the least. Furthermore, sport shooting requires significant amounts of space you just don't have in the city. So the only time guns become something worthy of note tends to be either when the police are using them to subdue suspects, or when criminals are using them.

So for country people, the positive effects of gun ownership would be, for example, less loss of livestock or crops due to less vermin; more food to put on the table, and a nice bit of recreation out in open areas. In urban areas, less so. The problem is that most states are a mix of rural and urban areas, and country people and city people have proverbially found it hard to understand each others' culture.

Darwin said...


The polling data that I've seen suggests that something along the lines of 30-40% of US households own some kind of gun, so what that per capita data works out to (and yes, you're reading that right, to the best of our knowledge there are more guns than people in the US) is that there are a minority of adults who own a lot of guns.

I'm something of an example, though I don't know how typical, in that I own five. I have an old .22 rifle that I inherited from my father and a .22 pistol that I bought twelve years ago when I had a membership at a shooting club with an indoor shooting range which was near my work. (This was before kids, and MrsDarwin worked as a stage manager at night, so I had a lot of time to kill in the evenings after work.) Then I own three World War II era rifles which I bought out of historical interest but have been restoring, and I take those out to a local range to shoot every so often.

Some gun owners have self defense as a primary or secondary motivation. Others go hunting every season. Hunting is very common in many rural parts of the US. Personally, I don't expect to ever have to defend myself with a gun (we live in a very safe small town) but I simply enjoy target shooting and the connection with history.

And there is something very rewarding about shooting sports. Our second oldest girls took archery lessons for a while, and I went along to the club as well since I had done some archery back in boy scouts, and in many ways its very similar -- you need to have a lot of precision and control of your motions to aim and shoot well, so there's a sort of enforced calmness that you need to be able to summon up, but there's also a certain exhilaration to releasing a shot and the feel, sound and sight of it flying down range and hitting the target.

Agnes said...

This is funny. i have never thought to connect archery to firearms, although my husband and my 3 kids do archery as well, it is connected to Hungarian cultural heritage. Hunting in Hungary is restricted, hunting rights are owned by hunting and forest preserving companies, one has to be very rich and/or influential to be able to hunt as a hobby. Also, during Communism, civilians obviously weren't allowed to keep weapons, so another 40 years to break any traditions there might have been. Target shooting is not a common hobby either. It may support your original post's theory though, that even with so few civilians owning guns, I found 2 incidences of school/university campus shooting in Hungary in news archives. Without doing statistics, it seems that it is not in proportion to gun ownership. Also, although in Hungary licence to owning a gun is supposed to depend on some kind of psychological testing, it clearly did not filter those individuals responsible.

MrsDarwin said...

Also, I should point out that we don't actually have any ammunition in the house. When Darwin takes the guns down to the shooting range (something that doesn't happen very often), he buys bullets there. We do have little round shells for the air gun downstairs, and Darwin and the kids go downstairs sometimes and shoot at targets taped to cereal boxes. Of all the guns in the house, I myself have only ever shot the air gun, and then only once, though I did do a lot of work helping Darwin refinish the vintage guns.

The best way to understand the American mindset of individualism (of which gun ownership is a part) is to read the books of Laura Ingalls Wilder. I know they've been translated into Hungarian. There's an image of the Hungarian version of Little House in the Big Woods (the first book) here.