Economics analysis of mandatory liability insurance for gun ownership


Posted by AzBlueMeanie:

I have posted about this previously, Public Policy of Personal Responsibility: Require Mandatory Liability Insurance for Firearms. Brad Plumer at Ezra Klein's Wonkblog has some economics analysis of this idea in The economics of gun control:

First, here’s a recent paper (pdf) by Duke’s Philip Cook and Georgetown’s Jens Ludwig trying to quantify the “external cost” of gun ownership. The two economists wanted to figure out precisely what sorts of costs gun owners impose on the rest of society.

That’s not an easy question to answer. For starters, there aren’t even airtight estimates of how many people actually own guns in the United States. So Cook and Ludwig created a data set that used the number of suicides by firearm in a county as a proxy for gun ownership — and checked it against a variety of existing survey data.

The next step was to figure out the “social cost” of owning a gun. The two economists determined that a greater prevalence of guns in an area was associated with an increase in the murder rate, but not other types of violent crimes (guns, the authors argue, lead to “an intensification of criminal violence”). Why does this happen? One possibility: The two economists found evidence that if there are more legal guns in an area, it’s more likely that those guns will be transferred to “illegal” owners.

When the two economists added up the costs of gun ownership—more injuries and more homicides—and weighed them against various benefits, they concluded that the average household acquiring a gun imposed a net cost on the rest of society of somewhere between $100 to $1,800 per year. (The range depends on the assumptions used—and note that they are not including the increased risk of suicide that comes with owning a gun.)

Now, normally when economists come across a product that has a negative externality—like cigarettes or coal-fired plants—they recommend taxing or regulating it, so that the user of the product internalizes the costs that he or she is imposing on everyone else. In this case, an economist might suggest slapping a steeper tax on guns or bullets.

* * *

[This] brings us to John Wasik’s recent essay at Forbes. Instead of a tax on guns, he recommends that gun owners be required to purchase liability insurance. Different gun owners would pay different rates, depending on the risks involved:

When you buy a car, your insurer underwrites the risk according to your age, driving/arrest/ticket record, type of car, amount of use and other factors. A teenage driver behind the wheel of a Porsche is going to pay a lot more than a 50-year-old house wife. A driver with DUI convictions may not get insurance at all. Like vehicles, you should be required to have a policy before you even applied for a gun permit.  Every seller would have to follow this rule before making a transaction.

This is where social economics goes beyond theory. Those most at risk to commit a gun crime would be known to the actuaries doing the research for insurers. They would be underwritten according to age, mental health, place of residence, credit/bankruptcy record and marital status. Keep in mind that insurance companies have mountains of data and know how to use it to price policies, or in industry parlance, to reduce the risk/loss ratio.

Who pays the least for gun insurance would be least likely to commit a crime with it. An 80-year-old married woman in Fort Lauderdale would get a great rate. A 20-year-old in inner-city Chicago wouldn’t be able to afford it.

Gun insurance for gun owners does exist right now, but it isn’t required — as Wasik notes, only 22 cities even require gun dealers to carry liability insurance. And, yes, under this proposal, people would no doubt still acquire guns illegally and evade the insurance requirements. 

Granted, this proposal isn’t likely to garner much political support — even the Illinois state legislature, which has often looked favorably on gun-control laws, swatted a gun-insurance bill down pretty quickly in 2009.

Megan McArdle at The Daily Beast has a lengthy economic analysis of Wasik's proposal in Should People Be Forced to Buy Liability Insurance for their Guns?:

Novel gun control ideas continue to percolate through the commentariat.  The latest idea is requiring liability insurance for gun owners, which seems to have first been suggested by John Wasik blogging at Forbes.  Reihan Salam, one of my favorite thinkers, says it's an idea seriously worth considering.

I too have been mulling this since it started making the rounds, mostly because I've been spending a fair amount of time thinking about insurance for the book I'm writing.  In the end, I think it might be a fine idea to help a small number of people, but it wouldn't do what proponents are imagining in terms of controlling criminal behavior.  Mostly, it would be a way to compensate some victims of gun accidents.

* * *

The first question we have to answer is why we want to require the insurance. There are three reasons I can think of:

1) it will simply raise the cost of owning guns to the point where people aren't willing to do it.

2) It will pay for the harm caused by guns

3) It will make insurers into de-facto regulators.

Number one is unfair to advocates of stronger gun control, most of whom say that they do not want to take all guns away from law abiding citizens. I see no reason to doubt them, and so I'm basically discounting any interpretation of this proposal that seems like it would just raise the cost of guns until they were unaffordable for all but the very wealthy.

Number two seems worthy, but unlikely to have a very large effect. The majority of gun deaths are suicides, and it's hard to see to whom the insurance would pay out–there's a reason that life insurance restricts suicide payouts, and presumably, any gun liability policy would have at least the same restrictions. (You could theorize that we would at least make it more expensive to commit suicide, and thereby hopefully reduce the suicide rate, but it's hard to believe that the added cost of a liability policy would substantially deter someone bent on self-harm. The Brady waiting periods did reduce the rate of firearm suicides in people over the age of 55, but not the overall homicide or suicide rates.)

The second largest subset of gun deaths and injuries are deliberately inflicted on another person; along with suicide, this accounts for all but a handful of annual gun deaths. In the case of homicide/assault with a deadly weapon, however, we're talking about a lot of shooters who cannot legally own their guns now, either because they are too young, because they are felons, or because they live in a city which has historically made it very difficult to own handguns. Those people are not going to acquire liability insurance just because you pass a law saying they have to, and given that they are also probably not going to register their guns, there's no way to force them to do so. Moreover, insurance usually excludes criminal acts committed by the policyholder, because of course, there's a huge adverse selection problem.

Which leaves accidents. Accidental death and injury rate from guns is fairly low, compared to both other gun incidents, and other categories of accident: 14,000 injuries and 600 deaths in 2011.  This sounds like a lot, but in a population of 300 million, a lot of people die each year from almost anything: dozens of kids a year drown in buckets. The rate of accidental firearm death or injury is much lower than something like drowning, much less a really common cause like motor vehicle accidents or medical mistakes.  And the damage is already often covered by other forms of insurance.

There would be some benefit to requiring insurance, but overall, we're talking about helping a pretty small number of people compared to the status quo. There might be some kind of justice argument–shift the cost to gun owners rather than your and my health and homeowner's insurance–but we're talking about pennies a year.  It might be symbolically meaningful, but probably not financially.

That leaves making the insurers into quasi-regulators.  The idea seems to be that the insurers will be able to do all sorts of things that the government won't.  If the law won't keep dangerous people from getting guns, or make you keep them under lock and key, maybe Traverlers can do it for them.

Problem Number One: The Insurance Laws Are Highly Political

* * *

I will… note that hopes of getting the insurance market to do something that the government finds politically unpalatable often end with the government intervening to stop insurers from doing that very thing: see flood insurance for people who build houses on flood plains, HMOs, and so forth. If you can't get the government to do it directly, the indirect method is going to face the same political pressures.

Problem Number Two: Criminals Don't Buy Insurance

Most gun homicides are committed using handguns, and a high percentage of those crimes are committed by young men, often poor men with criminal records and connections to the drug trade.

* * *

It's hard to see how an insurance requirement will stop people who are already willing to break the law, so it's not clear what we achieve by making it really, really expensive to procure the gun insurance they are not going to buy.

And even if you got them to buy it, how do you get them to keep paying their insurance premium? One month of even very expensive insurance is not such a big barrier when the gun itself costs hundreds of dollars. [Her example: One in seven drivers in America is uninsured.]

* * *

One way you might sidestep these problems is by requiring proof of insurance to buy ammunition–but you'd run into the "straw purchaser" problem even more than with guns. Someone who comes in to buy fifteen handguns reasonably gets the fisheye. Someone who comes in to buy fifteen boxes of ammunition could just like to buy in bulk; it doesn't take that long to go through a box of cartridges at the range.

Problem Number Three: Non-Criminals Aren't Usually Held Liable for Criminal Activity by Others

* * *

Insurers just write policies on your estimated liability. [Wasik's] effectively advocating an enormous change in the liability law to make people responsible for what happens with their property after they have sold it, lost it, or had it stolen.

While I understand that there is some liability if you carelessly make it easy for criminals to use your property to bad ends, expanding liability the way that he suggests would be pretty novel.

[McArdle loses her way here. This is not about criminal liability, but rather civil liability, and imposing civil damages for the "external costs" imposed on society for negligence and accidents. An insured gun would require the notice of cancelation of insurance by the seller when a gun is sold, and proof of insurance by the purchaser, the same as an automobile. If a gun is lost or stolen, the owner must notify the insurer of this potential claim, the same as an automobile. It is the reporting process that satisfies the public policy goal of responsible gun ownership. There is no liability for this. Wasik's proposal to extend liability for the acts of a third-party would be an enormous change in liability law, and would not have support in the law.]

* * *

The standard way to handle liability is that as long as you're paying for gun insurance, your insurer is liable for any claims against you (or, perhaps, for claims that originate from actions undertaken while the policy is active).  When you stop paying, the insurer stops being liable. From that day forward, you, or your new insurer, is liable for any damages.

This is a quite workable (indeed, it works all the time), but there's a problem: it seems like it doesn't do much about straw purchasers. They can buy whatever insurance they need to get the gun, and then let it lapse. The insurer doesn't care, because hey, they're not liable; the insurer bases their premiums on the damages they're likely to pay, not whether the purchaser is likely to be a bad person who buys guns for known felons.  You've gained a workable insurance rule, but now the whole point of the insurance–to make it cost-prohibitive for likely criminals to get guns–has been disastrously impaired.

* * *

Problem Number Four: Insurance Often Excludes Criminal Behavior by the Policyholder

Even if you could prove it, however, the liability insurance wouldn't necessarily pay.  Criminal behavior is increasingly explicitly excluded from liability policies. That's what makes it possible for those policies to be issued at an affordable price–and also, makes them less-than-useful for controlling crime. Since most of the uses we're concerned with are criminal, this further decreases the usefulness of liability insurance as a way to reduce gun deaths, unless we actually mandate that insurers cover criminal liability.

* * *

If we did this, the policy costs for likely criminals would go up–yay! But there's a little problem: while you've increased the cost of being a potential criminal, you've just decreased the cost of actually committing a crime. Probably not a lot, because as I mentioned, felons and people willing to buy guns for felons are likely to be judgement proof. Nonetheless, I'd want a lot of solid econometric evidence that benefits of the former outweigh the costs of the latter. Moral hazard is never desireable, but in the case of homicide, it's horrible indeed.

Problem Number Five: Holding Original Insurers Liable Might Destroy the Insurance Market

There is a way to force the insurance companies to care, of course:  have this new, effectively unlimited liability accrue to the insurer at the moment of purchase. This is what I think Wasik has in mind.

That would create quite a good incentive for insurers to prevent the wrong people from buying guns–perhaps too good.  This is the way that courts decided to handle asbestos claims in the 1980s.  And what they did was not limit asbestos insurance to "responsible asbestos owners," but destroy the market for insurance.

[I think McArdle's example of insurance company responsibility for asbestos manufacturers' product liability claims is misplaced. This is different from individual liability insurance. And Congress has exempted gun manufacturers from product liability claims.]

But as I wrote last week, you cannot use semantic games to do an end run around an enumerated constitutional right.  The Supreme Court ruled in Heller that you cannot ban a large class of weapons such as handguns. . . If your insurance scheme makes it functionally impossible for many people who are legally entitled to own guns to actually buy one, you risk running into the "emanations and penumbras" of Heller.  The better the law works to keep guns out of the hands of lots of citizens, the more likely it is to be struck down.

Problem Number Seven: Liability and Insurance Law is Mostly Handled By the States

The Economist might be forgiven for not remembering this, but American writers have also ignored this aspect. While I don't know that it's impossible to pass a federal liability law on this stuff–the commerce clause confers very broad powers–it does seem fairly unlikely; most of these suits are going to involve shooters and victims from the same state.

* * *

Problem Number Eight: Mass Shooters Won't Be Deterred  

Of course, this might be a fine law even if it just deterred regular homicide and left the mass shooters alone, so its effect on rampage shootings isn't exactly an objection. . . James Alan Fox, a criminologist who studies mass shootings, has debunked a lot of the myths of mass shooters.  The two things he says that are relevant here are, first, that it's not as easy as people think to identify potential shooters. Most mass shooters do not have criminal records or histories as a mental patient, and the "warning signs", like social isolation and a tendency to blame others, are both pretty common and pretty hard for the government or your insurer to detect.  His second relevant point is that mass shooters are very, very determined to get the guns they use, planning their attacks for months or years and carefully amassing their arsenal. Short of a ban and confiscation of most of the guns in the US, they will steal, borrow, or buy them however they need to.  An insurance requirement is a trivial obstacle compared to the others they generally overcome in order to execute their deadly plans.

* * *

Problem Number Nine: Liability Insurance is Usually Required for Public, Not Private, Uses  

Contrary to popular belief, you don't have to have a license and insurance to have a car; just to drive it on public roads. Nor do you have to have homeowner's insurance to own a home, just to obtain a mortgage.  This would be an expansion of the way we usually handle insurance requirements.

That doesn't mean it wouldn't pass legal muster–but that the way people are thinking of this (just like a car!) doesn't actually work. Moreover, if there are legal challenges, these requirements will face heavier, not lighter scrutiny, because driving is not a constitutional right.

Liability insurance might provide some level of damages to some people who are accidentally injured, while protecting the assets of the occasional affluent gun owner who inadvertently shoots someone.

* * *

But the fact that the [insurance] you can buy from the NRA costs so little seems to indicate that insurance would not be paying a lot of money to a lot of people. Before I offer my full throated support, I'd want to know how many people we're actually going to help, and how much it will cost us to set up the bureaucracy to enforce the insurance requirements.

But whether or not it's a good idea to help the victims of gun accidents, I just don't see how it's going to do much of anything to tackle the problem that proponents actually want to solve.  They're hoping to fix, not the problem of uninsured accidents, but our criminal gun problem–the folks who already aren't allowed to have guns, yet do a lot of the shooting at other people.

The legal changes needed to use insurance to make a dent in criminal possession and use–which so often go together–would be enormous, far reaching, and subject to gimlet-eyed review by appellate court.  And the very political forces that you are trying to end run would rise up and obstruct at every turn. To put it another way: if you could get support for widespread gun registration, you probably wouldn't bother thinking up ways to use insurers as substitute regulators.

Of course, proponents might say that it doesn't have to be a full solution, only a partial solution. And fair enough.  The problem is, it's not clear to me that any law which could actually be enacted would even a partial solution to gun crime.

Further reading:

Cook and Ludwig’s paper is hardly the last word on the social costs of gun ownership. There are some good discussions (and criticisms) of their study by Tyler Cowen and Don Taylor. See the comments especially. 

Previous articleWhen Women Vote, We All Win Rally
Next articleComments on Sean Arce’s arrest
AZ BlueMeanie
The Blue Meanie is an Arizona citizen who wishes, for professional reasons, to remain anonymous when blogging about politics. Armed with a deep knowledge of the law, politics and public policy, as well as pen filled with all the colors stolen from Pepperland, the Blue Meanie’s mission is to pursue and prosecute the hypocrites, liars, and fools of politics and the media – which, in practical terms, is nearly all of them. Don’t even try to unmask him or he’ll seal you in a music-proof bubble and rendition you to Pepperland for a good face-stomping. Read blog posts by the infamous and prolific AZ Blue Meanie here.


  1. Interesting report, especially these two lines: “The majority of gun deaths are suicides” and “The second largest subset of gun deaths and injuries are deliberately inflicted on another person; along with suicide, this accounts for all but a handful of annual gun deaths. In the case of homicide/assault with a deadly weapon, however, we’re talking about a lot of shooters who cannot legally own their guns now, either because they are too young, because they are felons, or because they live in a city which has historically made it very difficult to own handguns. ”

    In sum, the main use of guns is by those committing suicide (which in most states against the law) or those outside the law. Thus how much difference is going to be if either laws or mandatory insurance is put into place?