Log in

No account? Create an account
A short rant on gun rhetoric - Suit up, son! You're going to Mars! [entries|archive|friends|userinfo]
Suit up, son! You're going to Mars!

[ userinfo | livejournal userinfo ]
[ archive | journal archive ]

A short rant on gun rhetoric [Dec. 22nd, 2012|12:42 am]
Suit up, son! You're going to Mars!
People talking about guns are using poorer arguments than they ought to be.

Quite frequently, gun control advocates cite studies showing that increased access to firearms results in more deaths by firearms. On the other side of the debate, you hear pro-gun folks presenting facts to the effect that access to guns produces a larger number of "gun defenses," i.e., attempted crimes that were thwarted by guns.

My problem is this: "Gun deaths" and "gun defenses" are meaningless statistics.

Consider three towns, A, B, and C. Each town has an axe murderer that kills 3 people a year, and is the town's only source of violent deaths. Then one year, some event happens that increases each town's access to guns.

In town A, the axe murderer buys a gun, and becomes a gun murderer. He commits his usual 3 murders using the gun instead of an axe.

In town B, the axe murderer commits 3 axe murders that year, and 2 more murders also get committed by other people using newly-available guns.

In town C, a vigilante with a newly-acquired gun shoots the town's axe murderer before he can murder anyone else. That is the only death in the town that year.

In all three of these towns, "gun deaths" went up. But "total violent deaths" went down in town C, went up in town B, and stayed the same in town A.

You can make the same argument about the relationship between "gun defenses" and "total crimes."

The question people should be asking is: How does access to guns affect overall safety? "Overall safety" seems like it should be a hard thing to measure, since it requires you to measure different sorts of risks on the same scale. How many home invasions are the equivalent of one homicide? However, this sort of quantification of risk is exactly the sort of thing that insurance actuaries do for a living. If owning a gun had a significant effect on the overall safety of your home, for example, you would expect that some homeowners insurance would be willing to offer a premium reduction to gun owners, unless of course some market distortion is affecting their calculations. I recommend gathering your own data on that subject.


[User Picture]From: rifmeister
2012-12-24 03:05 am (UTC)
I'd be interested in seeing you flesh out this line of thinking a little more. Do you really want us all to gather our own data? Is the data out there?
(Reply) (Thread)
[User Picture]From: algorithmancy
2012-12-24 05:03 am (UTC)
I think there is some data out there, but I certainly won't claim to have sifted through it. I haven't found a rigorous survey of all insurance policies on the internet. I'd welcome some help in looking for one.

Some googling around suggests that some insurance companies consider gun ownership to be an increased risk, which may be reflected in premiums. It also seems like some liability insurance policies have explicit coverage exclusions for gun violence. I haven't found any policies that give you a premium break for owning a gun. So my quick unscientific scan leans preliminarily in favor of guns making your home less safe, not more. Look around and see if you reach the same conclusion.
(Reply) (Parent) (Thread)
[User Picture]From: arcanology
2012-12-24 03:48 pm (UTC)
You made me curious so I had a look and found a bunch of gun boards bemoaning insurance companies and pointing at certain companies that are good and/or NRA certified.

I was actually thinking along that line earlier that at some level this is an insurance question - mostly for schools rather than idividuals. I would imagine that the company insuring your school against everything would not be thrilled to know that there were ten more guns in the building, to the tune of either a bunch of money or an exclusion so that when someone gets shot accidentally the school district is bankrupt. But until someone tries to get such a policy it would be hard to know.
(Reply) (Parent) (Thread)
[User Picture]From: rifmeister
2012-12-24 03:59 pm (UTC)
So I think this general direction of looking at insurance companies is veyr smart, because insurance companies are at least in theory free actors with clear incentives.

But I wonder how to think about this case. Homeowners insurance generally protects belongings, right? So a higher rate for gun owners would directly indicate an insurance-company perceived correlation between gun ownership and home insurance policy claims, but what does this tell us about actual safety? Maybe people with guns leave their doors unlocked more. Maybe people with guns also tend to have more desirable-to-thieves possessions or tend to live in more crime ridden neighborhoods, even after controlling for whatever other factors the insurance companies know about and are legally allowed to use. Under the covers, the insurance companies are running regressions, and all I can really say is that gun ownership is showing up as a positive coefficient of risk of claims.

Put differently, I think your switch to looking at insurance companies doesn't help that much with your original goal of measuring "overall safety."
(Reply) (Parent) (Thread)
[User Picture]From: algorithmancy
2012-12-24 06:03 pm (UTC)
A standard homeowner's policy covers a bunch of different risks, including some "personal liability" coverage for the policy holder. If I get sued for harming somebody—anywhere, not just in my home—my homeowners insurance will pay the damages up to some amount. I would argue that such policies are much closer indicator of "overall safety."

I agree that policies that cover your valuables aren't as interesting an indicator. For one thing, you would have to correct for the fact that guns are valuables.
(Reply) (Parent) (Thread)
[User Picture]From: rifmeister
2012-12-24 06:32 pm (UTC)
Right, so this gives us at least one working hypothesis, which is that gun owners are more likely to get sued for harming somebody, without saying anything about overall safety?
(Reply) (Parent) (Thread)
[User Picture]From: algorithmancy
2012-12-30 01:04 am (UTC)
Ok, I guess I need to qualify my earlier post: A policy that covers your stuff would be interesting if insurers were willing to give gun owners a premium break. That would be pretty convincing evidence that guns are useful for protecting your stuff.

That said, I would argue that premiums for a comprehensive policy that covers your house, your stuff and your personal liability are a noteworthy (albeit not perfect) measure of "home safety," If insurers perceive gun ownership as moving the risk needle in "riskier" direction, then I would say that there's an argument to be made that guns make your home less safe. The main thing that's missing from the "home safety" picture is the risk of physical/emotional damage from home invaders. So the question is would covering that risk move the needle back the other way. Personally I doubt it, but I would welcome some data. Life insurance policies and injury policies might give us some insight.
(Reply) (Parent) (Thread)
[User Picture]From: firstfrost
2012-12-24 08:43 pm (UTC)
Some links on Facebook which you may also have seen, tell me that research into this sort of thing is apparently being suppressed. :-\
(Reply) (Thread)