On firearms, I’m open to robust gun ownership, but I’m not sold on anything. Like most of my positions, my default is to favor the ability of individuals to operate without restrictions, but I’m by no means a gun purist, to the dismay of many more intense libertarians I know. If there were more stringent regulations on firearms purchases, it wouldn’t be something I cared strongly about.
Nonetheless, many people do feel strongly about gun ownership in the United States, and I wonder if this is a position where efficient advocacy could help us understand whether those feelings are warranted. Unfortunately, gun ownership and gun control are complex issues with many different parts.
I. You Can’t Prevent Everything
From what I can tell, the impetus and goal of many gun control advocates is to prevent mass shooting events like the one that occurred in Orlando: a person who goes to a location specifically to murder innocent strangers. This first section will point out some flaws in the goal of trying to stop all mass shooting events.
The first problem is that restrictions on guns aren’t politically easy. Generally speaking, people who want to own guns are very politically motivated, even if small majorities want to increase gun restrictions. Gun control advocates obviously want to change this, but just keep the political challenge in mind when considering the rest of the arguments. Next, it’s worth noting that there are already somewhere around 300 million firearms in the United States. Would a determined mass shooter have trouble getting around a gun restriction when guns are already plentiful? And there are legal limits to making gun restrictions really universal: the Supreme Court has ruled that blanket handgun bans are unconstitutional. Even if they weren’t (assuming not just a political victory, but now a constitutional amendment), I’m not sure having a shotgun would significantly hinder a mass murderer. If you factor in new technology that let’s you 3-D print your own gun, or even machine parts of popular guns, the quest to remove all guns gets even more absurd. To stop these, not only would you have to outlaw guns generally (repealing the second amendment), but also likely confront significant first amendment and free speech issues in outlawing the distributions of schematics.
So let’s think about what would need to be done to stop a mass shooter from getting a gun: widespread restrictions of multiple types of firearms, prohibition on the distribution of engineering knowledge and designs of firearms, passing constitutional amendments that don’t have the necessary political support that affect not only the right to bear arms but also free expression, plus years, possibly decades, of restricted gun buying before the amount of available guns realistically drops down. Even then, in an gun-law environment similar to modern day France, clearly terrorists can still obtain fully automatic machine guns (that are already banned in the US).
But it’s even worse than gun control advocates think, because even if these gun restrictions were passed and no firearms of any kind could be purchased in the United States, the mass killing of strangers by a lone murderer would still be unpreventable. Why? Because the only reason the Orlando shooting is the deadliest shooting in American history is because the Branch Davidian Waco Siege didn’t end in a gunfight; 76 people burned to death in a fire. Whether you blame the FBI for besieging them, believe the conspiracy theories about the FBI, or agree with the government’s story that the Branch Davidians started the fire themselves, 76 largely innocent people were killed as a result of deliberate human actions. The real kicker is that the siege started because the ATF tried to serve a search warrant for…weapons violations.
Of course, if the Branch Davidian fire had never occurred, the Orlando shooting still wouldn’t be the deadliest mass murder in American history because 2 years after the Waco Siege, the Oklahoma City Bombing killed 168 people and caused massive property damage. It was conducted with a truck bomb made from ammonium nitrate fertilizer, nitromethane fuel, and Tovex explosive stolen from a mining site. Even if you banned every firearm from being sold in the United States, it would not have stopped this attack.
And, of course, the 9/11 attacks, by far the deadliest mass killing in American history killing almost 3000 people, were carried out with men who brought mace, tear gas, pepper spray, and utility knives / multitools. No guns were used at all.
Let’s be clear: these facts are related, but do not prove that gun control laws are worthless. Rather, all I am saying is that stopping all mass killings through gun control is literally impossible, because not only would it be politically impossible, not only would political restrictions on guns not stop mass shooters from obtaining weapons among the hundreds of millions available in the US, but in the end, not all mass killings use guns.
II. Random Mass Shooters Are Rare
But why am I arguing this idea? It almost sounds like a strawman; of course we can’t stop all mass shootings, but clearly there is something to be done about reducing the number of mass shootings on the margin. Gun control advocates believe this would occur with restrictions on firearm sales. But if we agree that stopping all mass shootings is impossible, the discussion becomes one involving trade-offs, risk, and nuance, and the argument for basing policy off of mass shootings falls apart rapidly.
Firstly, there are the same mitigating factors from the last section that reduce the effectiveness of any gun control measures, and would also mitigate the reduction in mass shootings. These are the many guns that already exist in the US, the ease of substituting shotguns or hand guns for ambiguously termed “assault rifles”, the substitution of explosives for firearms, and constitutional limits to gun restrictions. Given the high cost, how many mass shootings would actually be stopped by legislation? Well, for that matter, how many mass shootings are there now?
Mother Jones, who I otherwise have little good to say about, has an easy to access database of mass shooting events. To an order of magnitude, in the last 10 years between 10 and 100 victims are killed in mass shooting events each year, with an average of 4.1 shooting sprees a year and 37.4 fatalities per year. These events are tragic, but quite rare compared to other events we worry about (and many we don’t). For example, despite Republicans’ (and Democrats’) obsession with the War on Terror, you should not worry about dying in a terrorist attack, as they are incredibly unlikely (over the last 20 years, an average of 175 Americans have died in terrorist attacks per year, with a median of 16 per year). Yet you’re almost 4.7x as likely to be killed in a terrorist attack than a mass shooting. Planes are one of the safest modes of transportation, yet you’re still 11x more likely to die in a plane crash than a mass shooting. Police kill over 1000 Americans each year, so you’re at least 27x more likely to be killed by the police than a mass shooter. Having national legislation to address something that is 27 times less likely than being killed by an actual agent of the state just seems like terrible prioritizing.
In fact, if you’re worried about non-medical deaths, the CDC has a whole page dedicated to accidental deaths. That’s 30,000 deaths (802x more likely than a mass shooter) due to falls, 34,000 deaths (909x more likely than a mass shooter) from motor vehicle accidents, and then 39,000 (1043x more likely than a mass shooter) from accidental poisonings. Are the deaths of innocent people in car crashes less important than the deaths of those in mass shootings? I personally knew two people who died in car accidents, and, statistically, whoever is reading this knows more car crash victims than mass shooting victims. Given what we’ve said about the mitigating effects of constitutional restrictions, and the many decades needed before reducing the number of firearms available in the US, there’s an important question to ask about self-driving vehicle technology. If “universal access to self-piloting cars” is 20 years away and “it is difficult for mass shooters to obtain any firearms” is also 20 years away, which should you focus on making happen faster? Well, if you achieve the first situation, you can save ~1000 times as many lives as the second. There are also no constitutional restrictions on putting federal research funding towards achieving self-driving cars and fewer political opponents.
I would argue that given other problems identified here (not to mention areas of serious need identified by GiveWell) and the incredible rarity of mass shooting events, spending any resources, be those political or financial, on preventing mass killings in the United States would be irresponsible and inconsiderate to the thousands of people who die every year from preventable events.
III. Innocent Victims: Mass Shootings and War
Ok, so we can agree that not only can mass shootings not be prevented entirely, but even trying to mitigate them misallocates resources and ignores huge problems that need attention. But I’m sure most people will have realized by now that by focusing on mass shootings, I’ve ignored most homicides, one of the top 20 causes of death in the United States. And since most homicides occur with firearms, that seems pretty relevant to the conversation. This is absolutely true, and we are going to talk about homicides (and gun homicides) in the next section.
However, before we get there, I think it’s worth asking why people have such strong reactions to mass shooting homicides even though most gun violence is nothing like mass shootings. One answer is certainly the scale of violence in a single event. Another is that gun control advocates believe that similar policies would fix the problems of mass shootings and other gun violence, and we’ll get to that in the next section. But I suspect the most important point is that the victims of mass shooters are essentially random and could have done nothing to prevent their fate; they were just in the wrong place at the wrong time. Everyone can empathize with just going about their normal lives and suddenly being in danger, and most people will have a strong sense of unfairness of that type of scenario. I suspect more “average” homicides invoke less reaction because they involve people who may have connections to criminal activity, or occur among people who already knew each other, and who likely know what kind of people they were acquainted with. That’s not to say that gun control advocates don’t care about gun violence outside of mass shootings, but just that there are certain situations that really trigger our desire to do something.
This particularly scary concept that people uninvolved in criminal activity or unsavory environments could be attacked so ruthlessly is reminiscent of another awful scenario: war. But just like we have a weaker reaction to every day homicides, we also have weaker reactions to wars that occur overseas in places with corrupt governments and bad institutions. And just like we shouldn’t discard the importance of non-mass shooting homicides, we shouldn’t discard the importance of war casualties, since they often share many of the qualities of mass shooting victims: innocence, powerlessness, and being in the wrong place at the wrong time. Since 2011, the Syrian Observatory for Human Rights estimates that Syrian Civil War has taken the lives of anywhere between 280,000 to 400,000 people, with around 90,000 being civilians. That civilian death total over five years is roughly similar to the amount of homicides in the United States over the last five years, which means it’s roughly five hundred times more likely a random human died as a civilian in the Syrian Civil War than as a victim in an American mass shooting in the last 5 years. Yet I suspect there has been less money raised to stop the Syrian Civil War than to stop mass shootings, at least in the US.
Unfortunately, things are also worse when it comes to foreign wars because of the policy impacts of American voters. While it’s hard to say whether the US can prevent all wars from happening, just like it’s hard to say how many mass shootings can be prevented through policy, it’s not hard to say that, unlike mass shootings, sometimes the US government creates wars though its policies. The 2003 Iraq War was started unilaterally by the United States and resulted in the deaths of over 100,000 civilians. This unfortunately dwarfs the effect of any mass shootings over that time period by around 300x. If we include the spillover effects from Iraq in helping to start the Syrian Civil War, we find American voters have a lot to answer for.
So the question in this election year is: did people who want to prevent mass shootings vote for a specifically anti-war candidate? Or did they vote for people like Hillary Clinton who supported the Iraq War? Given that Republicans picked Donald Trump, one of the most unpopular candidates in history, and most people aren’t familiar with Gary Johnson, markets are picking Hillary Clinton to win in November. That means that the Democratic primaries had huge impacts on who will become president, and with Hillary, the president’s foreign policy risks being quite aggressive. Given that civilian warfare victims share many of the qualities of mass shooting victims, and that Hillary supported a unilateral war that created many of those victims, how much blame do Hillary supporters get?
It’s certainly conceivable (and likely) that a person could have a nuanced position where they want to avoid mass shootings and also forced themselves to vote for Hillary out of necessity on other issues. One could also argue that Hillary voters are only responsible for her future actions, and she isn’t even president yet. But if people feel so strongly about mass shootings (Senate Democrats tried to pass four gun control amendments and House Democrats have staged a sit in on the House floor), then I can only imagine the horror they feel at those that supported actual unilateral wars, events which are literally hundreds of times worse.
IV. Homicides and Methods
So I think we’ve shown several different ways how the focus on mass shootings is a bad idea. However, gun violence encompasses many more things than mass shootings. In 2013, the CDC reports that 16,121 homicides occurred in the US, with 11,208 resulting from firearms. According to that same report, a small percentage of firearms deaths were from accidental discharges (~600), with the majority of firearms deaths occurring by suicide (roughly 21,000). We are going to set aside suicide for now, and focus on homicides.
I’m going to be referring a lot to Scott Alexander’s post on the state breakdown of gun ownership and homicide data, as well as Brian Doherty’s in depth analysis of gun violence at Reason. Interestingly, if you do a simple regression of state firearm ownership to state gun homicide rate there is actually a loose negative correlation. If you do the same with gun suicide rate, the result is positively correlated. Doherty points out many related observations in gun data that seem to link gun ownership with fewer homicides. However, if you follow Scott’s work in his first section, he finds that there is a link between homicides and guns. but it can be thrown off by the urban/rural divide (gun ownership is high in rural areas, but homicides usually occur in cities) and Southernness. Additionally, if you control for the robbery rate, you can probably avoid the reverse causality issue where high crime rate causes people to buy guns for protection. Controlling for all this, you can find a correlation between gun rates and homicide rates.
However, Scott can’t account for the entire difference in American gun homicides and other countries’ rates just with the gun ownership rate. In fact, Scott finds that America’s non-gun homicide rate (about 5000 or so homicides) is itself higher than many other countries total homicide rate. Scott infers that there is a certain element he calls a “culture of violence” in the US that also increases homicides regardless of gun ownership. By his work, a bit under half of American homicides can be attributed to cultural elements, with the remainder explained by the firearm ownership rate.
Does this make sense?
In 2014, the FBI reported on 8,124 firearm murders of which they only knew the circumstances of 4,857. But of those roughly 5,000 homicides, 1,262 were under felony circumstances, i.e. robbery, narcotics, etc. Another 667 were gang related. Another large chunk were unknown circumstances, but from what we do know, it appears a sizable percentage of homicides occur due to criminal activity which results in murders. The implication could be that making certain guns illegal wouldn’t matter to criminals already breaking the law. I tend to agree with this sentiment, especially in America where guns are pretty easy to obtain, and would likely remain so for decades even after any gun control laws. However, it could be that to reduce the chances of being caught, criminals would avoid having firearms if there were more legal restrictions; since there aren’t restrictions today, criminals keep them on their person, and when things go poorly they end up being used when they otherwise wouldn’t. Moreover, on the margin, if guns were slightly harder to obtain, some amount of criminals would simply forego obtaining firearms because it would be too much effort.
There are also a fair amount of gun deaths that occur due to simple arguments that escalate when the parties involved have firearms. On the one hand, people who are unlikely to follow gun laws are probably over-represented among people who would shoot someone in an argument, but many of these arguments are among family members or acquaintances, people who would probably regret shooting the victim after the heat of the moment had passed. If firearms were not readily available to escalate many of these situations, there would likely be fewer deaths. Nonetheless, people who would shoot someone else in an argument seems to be the textbook definition of “culture of violence”.
I think the FBI numbers bear out some support for Scott’s position, but there are some caveats. One is justifiable homicide by private citizens with firearms (about 230 or so a year). The same regulations that would make it harder for people to obtain guns to reduce homicides would also make it harder for people to commit justifiable homicides. Since these are justifiable, it seems that people agree these homicides likely saved someone’s life. This complicates the calculations. Brian Doherty goes even further, citing some studies that put defensive gun use as high as over 2 million crime prevention incidents a year! That seems way too high to me, as it would mean something like 1 in 4 people have participated in a defensive gun use by the time they are 50 years old *(see footnote).
Another caveat is the relationship between the culture of violence and guns; without a culture of violence, guns may not be harmful at all (look at high gun ownership Wyoming). With it, guns make it easier for violent people to kill each other. Is it easier to change the culture of violence or to reduce gun ownership? I honestly am not sure, but it’s clear that homicide rates, along with all other crime, have been steadily decreasing for 20 years (did you know that?). Of course, it’s possible that the culture of violence is actually a lot worse than we think if we are to believe Doherty’s cited study, so perhaps guns actually squelch how violent Americans really are. Not sure how much I buy that, but it’s a possibility.
There’s also the question of “protection against tyranny” that the second amendment is supposed to provide. How valuable is this protection? It’s hard to say. On one hand, there haven’t been many revolts against the government in American history. One was highly successful (the American Revolution), one was not (the Civil War), and most others were on a much smaller scale, and weren’t particularly successful. But, on the other hand, the United States has never become a dictatorship, so you could argue the system is working well. Of course, one could also argue that the current American government taxes more strongly and is less democratic than the British Empire ever was.
Scott dismisses any loss of second amendment protections against tyranny by noting that a gun buyback program would reduce gun ownership from 32% to 22%, and any protection provided with 32% gun ownership should be similar to 22% ownership. However, Doherty notes that there is a lot of confusion over what the gun ownership rate actually is, with some estimates as high as 42%. Regardless, I am not so sure that changing the gun rate would have that low of an impact on tyranny protection; if there is any tyranny protection provided by the second amendment, marginal reduction in gun ownership should cause similar marginal reduction in tyranny protection. And with Donald Trump as a major candidate for president, it seems that tyranny protection (or protection against state-sponsored violence) is something that could be pretty useful in the upcoming years.
The last point is that Scott’s calculations are the best I’ve seen given the current data we have, but they are nonetheless very rough. Doherty’s cited studies come to more pro-gun conclusions than Scott’s, but Scott’s analysis controls for more factors, and overall I buy his argument more. Additional studies would be needed to know the real impact of different policies. Whether the CDC should use government funds to study guns is highly controversial, since many Republicans believe that CDC administrators are gun control advocates and would be biased in their studies. I suspect it’s true that CDC administrators would favor gun control, although I’m not sure about their likelihood of bias. Of course, as a libertarian, I have to ask why anyone would favor waiting for the government to fund a study when you could obviously just study it with private funds; money is money, and it seems to me that raising money to study the bad effects of firearms would easy for left-leaning interest groups to do.
Gun ownership and suicide is a pretty straightforward relationship.
Basically, if your method of suicide is effective enough, it will work. Guns are pretty effective at killing someone at point-blank range, so if guns are easy to obtain, suicides will go up. If guns aren’t easy to obtain, substitutes, like overdosing on medication, are slower or require specific knowledge and thus fail more often. Suicides are often ignored in the gun debates, and I admit that I would assign them lower moral weight considering that it is not one person willingly killing another, but a single person willingly killing themselves. From a preference utilitarianism standpoint, these people are technically fulfilling their preferences. Nonetheless, many people who attempt suicide later regret it and change their preferences. And many who do attempt suicide are often diagnosed with mental disorders where medication actually causes a change in their mental state where they do not wish to kill themselves. Given that these preferences can change, but suicide is permanent, it seems that doing things to reduce suicides is a worthwhile endeavor that even suicidal people would endorse. If you remain unconvinced that suicidal people should not be saved, check out the link above to Slate Star Codex for a more in depth argument.
If suicidal people are worth saving, at least a little bit, then suicides are an additional reason to reduce the prevalence of guns in the United States. Brian Doherty does have some counterpoints here. For example, the US suicide rate is still much lower than many other developed countries despite our high firearms ownership rate. I suspect that this is due to cultural differences in the US that reduce suicide (maybe we’re happier or more religious), but our suicide rate could be even lower with fewer guns. Nonetheless, it is an important point that even if we were to get rid of guns, the secular trend is for more advanced societies to see more suicides; firearms suicides are just a symptom. Solving the suicide problem may require much deeper problem solving.
Does the homicide and suicide data mean we should ban all guns or some guns? There are certainly some good points here that gun ownership ought to be reduced, but it also seems a blanket measure like that might not be socially optimal; most guns, after all, are not used for violence. Should the abuse of something by some mean we should take away the rights of everyone? We don’t seem to think that’s true in the case of alcohol or tobacco. And that thinking is also very dangerous for other things criminals can use, like encryption, or the right to remain silent. There is likely a better policy.
VI. Solution: Taxes Are Better Than Bans
These are the conclusions I’ve come to:
- We cannot prevent all mass shooting events, so we have to figure out on the margin, whether it is worth reducing the number of mass shootings.
- Mass shooting events are so rare as to be not worth basing any policy decisions on.
- Warfare is also pretty bad in many of the same ways as mass shootings, but affects thousands of times as many people and can be changed or prevented by American policy.
- Gun homicides more generally are still a big problem, although the data is rough and hotly debated. It’s likely due to both gun ownership and cultural factors.
- Gun homicides is an area where debate makes sense (also gun suicides), unlike mass shootings.
- Reducing gun ownership would definitely result in fewer gun deaths and likely fewer overall deaths, both homicides and suicides. I am slightly more confident about suicides.
- Taking into account costs of reducing gun ownership (protection against tyranny, defensive gun use, political capital, constitutional change), not all policies would work, and I wouldn’t be in favor of sweeping gun ownership changes.
It seems that the best policy which would reduce gun ownership would be a large sales tax on gun purchases, perhaps as high as 50%. This would still allow people who really want guns to get them, but slow the number of guns sold. Since people getting shot by guns are an externality on the gun-buyer/gun-seller transaction, it makes sense to tax the transaction. Guns are dangerous, and gun purchasers should pay for the risk they are imposing on society. A 50% tax isn’t actually paying for much risk; if a human life value is in the millions of dollars, a few hundred dollars risk expense on a gun means the gun buyer is paying for about a 0.01% chance of a gun death. With about 300 million guns and about 30,000 gun deaths, the chance of any given gun causing a gun death is about 0.01%–the same as the risk paid for in a 50% gun tax.
Now that’s considering a gun with a price tag over $1000 which tends to be fancier rifles. Cheaper guns would reduce the risk paid, even if they don’t reduce the risk of a gun death. We can counter with the point that guns probably have some positive externalities as well between defensive gun use and tyranny protection. But the good news is that a gun tax can be adjusted, so if it seems too low, the tax can go up and gun buyers will pay more for their higher risk hobby. This would also be a great area for gun debates to focus on, unlike mass shootings.
If this seems unsatisfactory to gun control advocates, I would say that this is simply a more economically efficient implementation of the Australian gun-buyback program, which I believe to be relatively admired by American progressives. Moreover, it would not require a constitutional amendment like a gun ban would. Incentives will work; people will buy less of more expensive things. Guns should be more expensive because they are riskier than buyers realize; if after factoring in that additional risk through a tax a gun hobbyist still wants to purchase a firearm, they should have the ability to do so.
The bottom line is that dire fears about mass shooters are not worth worrying about, much less making policy about. Homicides and suicides are a better subject worth debate and discussion, but it’s important to remember that homicide rates are overall on the decline. I think a good solution would be a hefty tax on firearm sales, but reasonable people can disagree on the effectiveness and weight of the various factors involved in this multi-faceted issue.
* 2 million defensive gun incidents a year means 318/320 million people don’t experience a defensive gun use each year. Raise 318/320 that to the 50th power to account for every year a 50-year-old has lived = 73% of 50-year-olds have never had a defensive gun use, or 27% have).
Picture Credit: Gun Wall by Michael Saechang. Licensed under CC-BY-2.0.