Game theory and guns: why universal background checks are a debate — and how to solve it

If you follow gun stuff, you've been hearing a lot about universal background checks. So let's get the lay of the land. This article is especially for people who are roughly aware of the issue but don't know the details of it, or the details of why it's a debate at all.

And yeah, why is it a debate? The question “Would you favor or oppose background checks on all potential gun buyers?” really does poll at 90% support. That's surprising. You can't get 90% of people to agree on a flavor of ice cream, let alone a gun law.

What's going on here? There are three key questions to answer:

1. If universal background checks have 90% support, why haven't they already become law?
2. Why does each side insist on a path that the other side genuinely believes is insane? Why does each side only try to win by force?
3. Is there a law that will give everyone what they want at the same time? What would that look like?

Let's answer them one by one.

If universal background checks have 90% support, why haven't they already become law?

This is the simplest of the three, and it comes down to a difference between support for the general concept and support for how it works in real life. The thing that polls in the 80-90% range is this exact wording (with minor variations in different polls): “Would you favor or oppose background checks on all potential gun buyers?”

Background checks have been mandatory under federal law for all gun store purchases since 1993. That requirement doesn't apply to sales between private individuals, because forcing background checks for all sales (e.g. sales between friends) was politically unpopular and didn't have the votes in Congress. So private sales were offered as a compromise to get the law passed.

"Universal background checks" is the term for now revoking that compromise and expanding the law to cover all firearm transfers, whether the seller/transferor is a gun shop or just a private individual. But an actual law has to go into more detail than:

Run background checks on all potential gun buyers. Thanks.

– Congress

What kinds of detail? I'm going to set a timer for 60 seconds and start thinking of edge cases a universal background check law would have to account for. Ok, here goes:

• Loaning guns on a ranch
• Loaning guns to a hunting buddy
• Inheritances
• Gifting to family members
• Gifting to a friend
• Loaning to friends or family
• Repair
• Safe keeping for trav

That last word should be "travel", but I stopped typing when the 60 seconds were up.

If you don't live in a high-gun-ownership area, I know these edge cases seem silly. "What, you're going to give your teenager a rifle for Christmas?"

The answer, believe it or not, is "Actually, yes." In much of the country, gun ownership is swirled into foundational cultural rituals like it's Neapolitan soft serve. Giving your kid a rifle for Christmas isn't just utterly normal, it's a time-honored rite of passage. Telling a ranch hand to grab the rifle out of your truck before he leaves to guard some livestock for a few days isn't just a convenience, it's how you keep the ranch running. Loaning your new pistol to a friend so he can try it out isn't just a favor, it's a way to deepen that friendship just like loaning him a book or going out to dinner.

(For more on edge cases, see David Kopel's detailed breakdown of the universal background check bill from earlier this year.)

(And while I'm writing parentheticals, suicidal ideation is an edge case I didn't think of in the 60-second sprint above. But it's a big one. Two-thirds of shooting deaths are suicides. We should all want it to be incredibly, trivially easy for someone to say to a friend or family member, "I'm going through a deep depression. Could you hang onto my guns for a while?" Universal background checks make that illegal. People with severe depression aren't going to drag a friend to a gun store, involve the gun shop owner in their private mental health issues, and pay a transfer fee for each gun. BJ Campbell wrote a piece specifically about this non-obvious suicide risk of UBCs.)

Again, this will all seem easy to dismiss if you live in a place where these things aren't normal. But that's the point — there are a lot of places where these things are normal. People aren't going to drive an hour to the gun store with their friend or family member, pay a fee, do a background check, loan the gun, and then do the whole thing over again when the person gives the gun back. They're just not. And not only are they not going to do it, they're going to resent whoever tried to make them. That's all purely descriptive — whether a UBC law is good, bad, or indifferent, as an empirical matter people will break it pretty much universally.

Cam Edwards described this with a different cultural touchstone that could help bring the problem closer to home: "Trying to enforce a law requiring background checks on gun owners selling one of their firearms to another individual would be like trying to enforce a law requiring you to show ID before you can get a beer from your buddy's cooler at the neighborhood BBQ."

So the answer to this section's question — "If universal background checks have 90% support, why haven't they already become law?" — is coming into focus. The vaguely-worded question of whether buying a gun should require a background check is what gets 90% support. And yes, you'd expect 90% of people to "sure, why not" that general idea in a quick phone survey. But once people see what the details of a UBC law look like in real life, things change. Rephrase the question to capture those details — "Should it be a felony to lend your hunting buddy a gun if his gun breaks?" or "How much jail time should you get for giving your cousin a gun for his birthday?" — and people sour on it quickly.

That's descriptive, not speculative. Empirically, these laws are a lot more contentious than the 90% talking point suggests. David Kopel summed up how UBCs fare in state-level referendums: "Actual election polls tell a different story. Defeated in Maine, passed by less than 1% in Nevada. Passed with 59% in Washington. Once voters learn how miswritten these bills are, including HR8 in Congress, support plunges."

UBCs only got 59% of the popular vote in Washington, a state that hasn't elected a Republican senator since 1994 or a Republican governor since 1980. UBCs were rejected by Congress in the 1993 Brady Bill, specifically as a compromise to get the bill to pass — and the early '90s were a high-water mark for support for gun control. Shortly thereafter, the public began a 25+ year trend of steadily declining support for gun control across nearly every demographic, and that trend continues through the present. In terms of long-term trends, statistically gun control has never been less popular than it is today.

Taking all this together, it's just not plausible that UBCs actually have popular support. Once people see the details, they read it as an attack on their lifestyle. These laws have gotten traction in places where that lifestyle has been extinguished (or never existed) as a significant cultural bloc. Everywhere else, UBCs simply do not have support outside of phone surveys.

Why does each side insist on a path that the other side genuinely believes is insane? Why does each side only try to win by force?

So we have this empirical evidence that UBCs are popular where guns are culturally irrelevant, unpopular where guns are culturally relevant, and never the twain shall meet. Let's break that down a bit.

On the gun rights side, people see their own personal experience with freely transferring guns between friends and family without issue, and the cultural touchstones built around that. They see studies like the recent one in the Annals of Epidemiology finding no effect from California's implementation of UBCs, or the one in JAMA that found no effect on homicides from the background checks that Congress already passed in 1993. They believe that background checks are irrelevant to crime, because criminals already generally acquire their guns illegally, and mass shooters almost all pass a background check (because they tend to have no prior record). And they see the idea as primarily a way to harass gun owners, the outgroup.

On the gun control side, people are often unfamiliar with gun culture and earnestly see those concerns as insane. They have plenty of their own statistics, and entire university research centers dedicated to running studies. They believe that there's a correlation between gun ownership and homicide, and it's a short jump from there to believing that access to guns is literally killing thousands of people every year. So of course every gun transfer should require a background check. How often does anyone need to transfer a gun anyway, and how hard could it be to run a background check? And in any case, this would save thousands of lives. Opposing that would seem not just wrong, but baffling and insane.

Side note: just about everybody will find something in the previous two paragraphs that seems factually wrong. And some of it probably is. (I certainly believe that some of it is.) That's the point. Each side presents facts, but the facts don't really matter. People have made up their minds, and that's it. If they hear a fact that bolsters their position, they'll spread it. If they hear a counterargument, they double down against it. So the real question isn't what the "truth" is, because empirically, the truth doesn't matter. The real question: why doesn't the truth matter?

The problem is that the two sides are in Nash equilibrium: "a strategy profile is a Nash equilibrium if no player can do better by unilaterally changing his or her strategy". If you want to spread gun rights, would making a unilateral concession to the gun control crowd help you spread gun rights? If you want more gun control, would supporting an advancement for gun rights help you get more gun control? Obviously not, in both cases. In Nash equilibrium, a unilateral concession would by definition only hurt you. So in that sense, of course both sides are trying to cram victory down the other side's throat. Mathemetically, a unilateral compromise can only open the door to more concessions.

Example on the gun rights side: in 2009, Montana passed a law declaring federal gun laws null and void for almost all firearms made in the state. The deliberate, stated intent was to create a massive loophole that would essentially repeal most federal gun laws within Montana. A federal court quickly struck the idea down, but the point is that it passed (signed by a Democrat governor no less). That’s illustrative. Similar laws quickly followed in Alaska, Arizona, Idaho, Kansas, South Dakota, Tennessee, Utah, and Wyoming.

Example on the gun control side: in 2013, New York state passed a ban on rifle features like pistol grips and adjustable-length stocks. Only 4% of covered gun owners complied with the ban, so there are now ~1 million New Yorkers with a 15-year felony sitting in their safe. This is viewed as a success, and the gun control group Giffords gives New York an A- for its gun laws, second only to California and New Jersey.

Choose one of the two paragraphs above and really embrace its mindset. Believe fully that you're on that side. Then ask yourself: what would happen if you made a unilateral concession to the other side? Would it leave you better off? Would you feel safe? Would you trust them to say, "Ok, thanks, that's all we wanted. We're all set now"?

Naturally, you'd only expect them to come back for more next year. As a rational actor in a culture war Nash equilibrium, you'd be crazy not to be stubborn. The only crazy thing would be compromise.

If you want to change that, first you have to change the incentive structure.

Is there a law that will give everyone what they want at the same time? What would that look like?

There's a key subtlety to Nash equilibrium: you can't benefit from changing your strategy, assuming the other player doesn't change theirs. But what if the other player is willing to change?

This is where people use the word "compromise". But on guns, what they usually mean is, "Fine, let’s compromise: we'll do none of what you want and only half of what I want." Gun owners know the game theory well enough by now that they can never allow another unilateral loss. (You know how well they know it? If I just say "cake meme", they all know what I'm talking about.) Gun control groups are in the same boat incentive-wise, and so can never sign onto a reform of existing gun laws.

What a genuine compromise looks like is that each side gives some things, and each side gets some things. And both feel like they net-benefitted.

Ok, so what would that look like?

(I should note that this give-and-take idea isn't an argument for how things ought to be, it's a description of how things are. Because of the incentives here, this is the only way a major gun law for either side can pass at the federal level. There is an outside chance that given the right emergency, a one-sided law could pass by force. But that could — for these same game theoretic reasons — set off a chain-reaction of backlash that gets arbitrarily bad. Like anywhere from Waco bad to civil war bad.)

For the gun control side: Swiss-style universal background checks

Yup, the big enchilada. Gun rights people often worry that UBCs will turn into the government tracking (and later confiscating) everybody’s guns. Statistically, most of the states that have created a registry have indeed gone on to use their registry to ban certain guns (specifically: California, Connecticut, DC, Hawaii, Maryland, Massachusetts, New Jersey, and New York), so this system prevents that while still mandating that every gun purchase goes through a background check. It’s modeled closely on Switzerland’s system. Here’s how it works:

1. Any gun buyer can log into the NICS background check system and enter their personal information. The system gives them a check number that expires in 1 week. (For reference here is ATF Form 4473, the background check form.)
2. The buyer can then buy firearms from any legal seller. They have to meet face-to-face (or ship the gun to a licensed dealer for the buyer to do the check with), and the buyer shows the check number. The seller verifies the buyer's ID, enters the check number into the NICS system, and the system returns just one word: “approved” or “denied”. If the check is approved, they can proceed with the sale.
3. The system doesn’t collect any information at all on the item(s) being sold/transferred (type, make, model, quantity, etc.) — its only job is to check on whether the buyer is legally allowed to purchase firearms. After one week, when the check number expires, the system doesn’t retain any records. (That information is already archived for 20 years on the Form 4473 for all gun shop sales.) The system collects no information about the seller, as it’s designed to work without even knowing the seller’s identity.
4. Transfers between family members are exempt. Firearm loans of up to 14 days are also exempt — this is to accommodate a situation where, say, two people are on a backcountry hunting trip and one needs to lend the other a gun during the trip. They need some way to do that without committing a felony.

This has a number of things going for it. First, it hands gun control groups the item at the very top of their wishlist. A background check on every single gun sale in the country. It opens up NICS to private individuals, which gun owners have long wanted — it's nice peace of mind to be able to run a background check if you're selling a gun to someone you don't know well. And it does all of that in a way that preserves family transfers, respects temporary loans, and makes a registry technologically impossible.

For the gun rights side: reform Depression-era laws on barrel length and silencers

Because of an 85-year-old law (the National Firearms Act, or NFA), rifles and shotguns in the US must have a minimum barrel length (16" and 18", respectively). So you can go to the store, pass a background check, and buy a rifle with a 16" barrel. But possessing that same exact rifle with a 15" barrel is a felony punishable by up to 10 years in federal prison.

To buy a short-barreled rifle or shotgun (SBR or SBS) legally, you pay a $200 tax to the ATF, and then you wait. Currently the average wait time is 11 months. Many people incorrectly believe that the long wait is for some kind of James Bond background check that takes 11 months to do. It’s not; SBRs/SBSes are subject to the exact same background check as all firearms are. The 11-month wait is literally just for the ATF to get through its paperwork backlog. Also, if you loan your SBR to a family member, that loan is a felony punishable by 10 years in federal prison. To stay out of jail, you’d have to do the $200 tax + 11-month wait again just to loan the gun to your dad. Then when he gives it back to you? $200 tax + 11-month wait again. All NFA items have to go through this same process. Legally, an SBR is treated identically to a grenade launcher.

As an illustration of how arbitrary that is, a well-meaning gun owner who cut his gun in half as a form of renunciation unknowingly broke this law — his video shows him red-handed doing what the NFA describes as “illegally manufacturing a short-barreled rifle”. For the harmless act in that video, he can be jailed for 10 years.

Contrary to popular belief, the barrel length law has nothing to do with dangerousness; the shorter the barrel on a rifle or shotgun, the less powerful and less accurate it is. The law is actually an interesting relic of how the NFA came to be. The first draft of the NFA included all handguns — in 1934 the $200 tax was equivalent to $3700 today, and it was designed to effectively ban all handguns.

The bill’s writers added the barrel length rules to stop people from buying a rifle, cutting down the barrel and the stock to make it almost as short as a handgun, and then saying, “It’s a rifle, not a handgun.” There was only one problem: there was no political support to include handguns in the NFA. The law wouldn’t pass in that form. So the writers removed handguns from the NFA, but they never removed the barrel length rules that were only there to close the handgun loophole. That is why today a 16” barrel is fine while a 15” barrel is 10 years in federal prison.

Also because of the NFA, silencers, aka suppressors, are currently in this same legal category. What do suppressors do? Well, apart from an explosion or a rocket launch, a gunshot is quite simply the loudest sound you will ever hear. It is deafeningly, dangerously loud, around 165 decibels. Decibels are a logarithmic scale, so every 10 dB is 2x as loud. A jackhammer is 115 dB. A jet airplane taking off 25 yards away is 130 dB. The OSHA standard for sound that will instantly damage your hearing is 140 dB. Guns are another 4-6x louder than that OSHA cutoff.

A silencer lowers a gunshot to about 130 dB. That’s still 3x louder than a jackhammer. Extremely loud, but just quiet enough to not instantly damage your hearing. Gunshot detection systems like ShotSpotter still pick up silenced gunshots, because they listen for sound signature in addition to volume. Silencers are the only way to avoid hearing damage in situations where there’s no chance to put on hearing protection — that's why in countries like Norway, silencers are completely unregulated and it’s considered unsafe to hunt without one.

(We have an in-depth article about silencers if you'd like more details on them.)

This is a proposal to remove short barrels and silencers from the NFA. Instead of putting them in the same category as grenade launchers, treat them like any other gun — available to adults who pass the background check, denied to those who don’t.

Game over

So that's the incentive structure at play here. A player in Nash equilibrium has no incentive to change their strategy. So you have to give them one if you want them to change. Folks who deny that and insist on winning by force are just telling you that they don't understand the game. Folks who embrace it, well they're ready to win.