The CEO of the National Rifle Association, Wayne LaPierre, has broken that organization’s silence following the shooting at Sandy Hook Elementary School in Newtown, Conn. by announcing the NRA’s support for armed guards in “every single school” in the United States and “an active national database of the mentally ill.” (There is no evidence the suspect in the Sandy Hook Elementary School shooting was mentally ill and the American Psychological Association has condemned the comment.)
Putting aside the fact that Mr. LaPierre’s answer to gun violence is, apparently, more gun violence, the NRA’s plan to pay for security officers at every elementary, middle, and high school in the country may be incredibly expensive. It might not even make a difference: other gun supporters had already claimed that the shooting would have been averted had the teachers and staff been armed, but an empirical report has shown that a maximum of 1.6% of mass murderers were stopped by armed civilians.
However, I want to give the NRA’s idea a try. Let’s see how putting a gun in every classroom stands up to economics.
First of all, how many schools are there in the US?
According to the National Center for Education Statistics, there were 98,817 public K-12 schools in America in 2010. Sounds good to me.
Now how much would it cost to hire a police officer or private security officer to patrol every one?
Well, according to the Bureau of Labor Statistics, police officers make $55,010 per year at the median, so hiring 98,817 would cost almost $5.5 billion a year, not including benefits – slightly lower than the price of all of the goods and services produced in the entire country of Malawi.
But this protects lives, so it’s worth it!
Is it really, though?
I’m an economist who just took a class on economics of public policy, so I used a cost-benefit analysis to decide, the same method that governments and private companies all over the world to make policy decisions.
If you don't want to read more numbers than you have to, read this:
Using back-of-the envelope calculations, I figure that, over 10 years, a ban on semiautomatic assault weapons that includes a buyback program (which is also very inefficient, but is better than other options) would cost almost than $34 billion less over 10 years than putting a guard in every school and would save more than 8 times the number of lives in mass shootings alone (my calculations do not include benefits from reductions in other types of violent crime).
[Warning: some technical economics terms follow]
How can you value a life, though? Insurance companies and the US government both use a technique called “value of a statistical life” to average the characteristics of people in different risk groups and make the decision “if it’s worth it.” One of the most popular measures assigns a value per quality year of life; most insurance companies use $50,000 per year, but economic research has shown that number might be about 2 and half times too low, so I use $128,000 in my analysis.
For people who die in mass shootings, we can value the lost years using VSL and multiply it by the average number of people killed every year to find the benefit of measures like putting armed guards in every school.
For example, it turns out that between 1991 and 2010, the average number of people killed in school shootings (K-12, including private schools) was about 1.8 per year, while the average age of those killed, excluding the perpetrators, was 19.2. If the average American lives to be 77.8 (which they were predicted to do in 2010), the average victim of a school shooting during this reference period lost 58.6 years of their life, or $7,500,800 of economic value. Multiply this by 1.8 and we find that every year, the US loses an average of about $13.5 million in economic value due to school shootings, just in life lost (not including spending on law enforcement, opportunity cost, etc.). [Cf: MotherJones, LATimes]
$13.5 million is much less than $5.5 billion – we can conclude that putting an armed guard in every school is one of the least efficient ways of mitigating school shootings (I say mitigating because it is not clear that guards have any effect at all: Columbine High School in Colo. had an armed guard at the time of the shooting there).
Over 10 years, Mr. LaPierre’s program would cost about $34.5 billion, including the VSL benefits from people saved and discounting for the time value of money (i.e. “money later is worth less than money now”).
However, an alternative has been proposed: recalling and buying back semi-automatic weapons (which was tried in Australia after that country’s 1996 mass shooting and proved to be very inefficient). It has been pointed out that Australia has – and had – many fewer guns than the US (the Land of the Free has 88.8 firearms for every 100 people, according to a 2007 survey, 34% higher than Serbia, the next highest country, and 40% higher than Yemen, the most violent country in the world). Thus, opponents conclude, it would be prohibitively expensive to buy back even semi-automatic assault weapons in the US.
The numbers don’t support that conclusion. Using the same methodology that I used to evaluate putting guards in schools, I conclude that a buyback of semiautomatic assault weapons would cost, at the low end, slightly more than $1 billion once. It’s true that the government would no longer receive income from permit fees, but this income is negligible, and the amount spent on enforcing a ban may be up to $0.5 million annually. The rewards, however, would be great: 180 people were killed with semi-automatic weapons during the reference period – 8 times more than in school shootings. The net costs would be about $662 million, giving a savings of almost $34 billion over the NRA’s proposal. (I assume the program would only buy guns of the type used in the Conn. shooting, the AR-15 assault rifle, of which there are an estimated 3,750,000 in the US. There are many other types of assault weapons, but the effective number would probably end up about the same, since buyback schemes have very low participation rates.)
Again, I want to emphasize that these are back-of-the-envelope figures. I have not done a formal cost-benefit study of gun control or armed guards in America, although I wish someone would. These are numbers meant to bring some kind of rationality into this otherwise completely irrational debate. They do not represent a coherent policy of any kind, and should not be used to form such (that’s what the CBO is for). This is just me, my curiosity, and some policy and economics tools I learned to use this past semester. That said, I hope it informs anyone who’s bothered to read this far.
I realize that some people will be offended by my use of the value of a statistical life, but if that offends you, you should leave modern society; all risk analyses that involve humans (even if they don't involve death) value your life in some way. It’s a necessary evil, so we may as well put it to good use.