Cobra Economics: Why Gun Buybacks Fail by Design

There is a name for what happens when a program pays for killing a problem and ends up breeding it. The term is from colonial Delhi, where the British government offered a bounty for dead cobras, and the city responded by breeding cobras to collect it. Economists call it the cobra effect; Horst Siebert popularized the name, and Hanoi’s rat bounty — which produced rat farms — is the canonical second instance.

The gun buyback is the American specimen, and it is almost a textbook one. The literature already establishes that buybacks do not measurably reduce gun crime. The cobra frame explains why: not because the program is implemented badly, but because it is designed badly, and the evidence of the design error is on the receipt.

The mechanism, in four lines

A buyback is a price floor. It offers $50–$200 for an object, asks no questions, and cares nothing about what the object originally cost or whether it works. That is precisely the structure that creates a market for the cheapest thing that satisfies the criterion.

• Houston, 30 July 2022. The city paid $50 for broken guns, $100 for rifles and shotguns, $150 for handguns, $200 for automatic weapons — no questions asked. One attendee, anonymous by his own choice, collected $50 each for 62 3D-printed firearms that cost about $3 apiece to make, per FOX 26’s report. The seller counted 63 handed across the table. The event took in 840+ guns and paid out about $100,000, and the city changed its rules afterwards so ghost guns would no longer be accepted.
• Utica, 27 August 2022. A New York Attorney General’s office buyback priced unserialized firearms at a premium. One participant, pseudonymous, printed items on a $200 printer using roughly $50 of filament and left with $21,000 in $500 gift cards for 110 items, including 60 printed auto sears priced at $250 each plus $100 for lacking serial numbers. The office changed its rules within weeks.

In both cases the announced response was a rule change: exclude the printed guns, set a $25 price for printed parts, “ensure that no one can exploit this program again for personal gain.” That response is the cobra pattern repeating itself. The program did not fail because its participants cheated. It failed because it was working exactly as written — the rules were what the participant used.

The design error, exactly

The classic bounty fails in one of two ways, and the buyback tends to hit both:

1. The price floor exceeds the production cost of the cheapest qualifying item. $50 for an object you can produce for $3 is not a destruction incentive; it is a manufacturing incentive. No one at city hall priced the object’s floor against its supply curve — but the person holding the gift card did.
2. No quality or provenance filter. The program pays identically for a working defensive pistol and a printed frame that has never been fired. When the payout does not discriminate, the arbitrageur supplies whatever clears the bar for the least cost — which is exactly what the photographic record from these events shows.

The second failure mode is the one defenders point to as a fix, and it is the wrong fix. Adding a functioning-test requirement (as New York did) changes the minimum quality, not the incentive. A price floor with a specification still creates a market for the cheapest item meeting the spec.

The evidence behind the frame

The cobra reading would be an economic curiosity if the programs showed public-safety effects — a designed-in inefficiency, maybe worth the noise. They do not. The NBER working paper “Have U.S. Gun Buyback Programs Misfired?” (Ferrazares, Sabia and Anderson) covers 339 buybacks in 277 cities over 1991–2015 and finds no evidence that they reduce gun crime, gun suicide or gun homicide, with 95% confidence ruling out a decline of more than about 1.1 percent in the year following an event. RAND’s Gun Policy in America review of the same question finds no qualifying evidence of effect.

That pairing is the whole case: the programs do not work even before anyone games them. The incentive problem is not the reason the policy fails; it is the reason the policy can’t not fail — it monetizes the wrong people and the wrong objects simultaneously.

What this site will and will not do with the finding

The documented events — the Houston man, the Utica man, the rule changes, the research — are all from the published press record, linked and checkable. Reporting them is reporting.

What this site will not do is instruct anyone to manufacture firearms to sell into a buyback. Selling your own junk into one is a lawful voluntary transaction and it is already documented approvingly in mainstream coverage. But manufacturing firearms for personal use needs no license, while manufacturing them to sell raises the licensing question of 18 U.S.C. § 922(a)(1)(A), widened by the Bipartisan Safer Communities Act, and volume and intent are exactly what a prosecutor looks at. A reader who needs a lawyer to parse that sentence should not need this article to make the decision easy.

The policy conclusion does not need the stunt anyway. “Buybacks pay above-market prices for the cheapest objects that qualify, and the research says they change nothing” is the strongest version of the argument — it is the version that survives contact with a skeptic, because every number in it is somebody else’s published number.