This post is based on a paper that was presented at the 2026 Firearms Law Works-In-Progress Conference. The Conference is held each year on a home-and-away basis with the University of Wyoming Firearms Research Center. This post also appears on the FRC’s Forum. This guest post does not necessarily reflect the views of the Duke Center for Firearms Law.
Slide Fire Solutions built a company on a firearm accessory that was lawful, according to repeated statements from the Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF). Over roughly a decade, the agency issued a series of classification letters concluding that the bump stock was a firearm part, not a machinegun. Slide Fire patented the device, sold tens of thousands of units, and licensed the design to other manufacturers. Then, in 2018, after a presidential memorandum directed the Department of Justice to propose a ban, ATF reversed course. The Final Rule reclassified bump stocks as machineguns, required existing owners to destroy or surrender them, and provided no compensation for the items. ATF’s own regulatory-impact analysis estimated the cost in the hundreds of millions of dollars, including roughly $199 million in forgone production and sales. That is the agency’s own estimate of the economic value the rule destroyed. Six years later, in Garland v. Cargill (2024), the Supreme Court held the Rule unlawful: a bump stock is not a machinegun under the National Firearms Act, and ATF exceeded its regulatory authority.
But Cargill restored the technology’s legal status only going forward. It said nothing about the five years and two months during which the unlawful rule had destroyed the value of Slide Fire’s patents. The patents survived the whole ordeal; the company built on them did not. That gap is the subject of our paper: what the government owes for the value the rule destroyed while it was in force. The existing scholarship has largely passed it over. It is now squarely before the Court of Federal Claims.
We argue that federal regulation destroying the economic value of a valid firearms patent is a compensable taking under the Fifth Amendment. The Second Amendment is what completes that claim: it forecloses the police-power defense the government would otherwise raise to avoid paying compensation. The argument rests on three components, and failing any one of them defeats the claim.
The patent-takings literature is lively, but it has grown up around contexts in which the regulated activity enjoys no independent constitutional protection. Gregory Dolin and Irena Manta, on one side, and Camilla Hrdy and Ben Picozzi, on the other, have debated whether administrative cancellation of patents through inter partes review is a taking. Laura Dolbow’s recent work develops the public-franchise tradition as a limit on patent takings, with the Inflation Reduction Act’s drug-pricing program as the leading application. None of this work engages the firearms-patent setting. Firearms add a constitutional dimension that the standard police-power analysis cannot accommodate. The patents at issue cover inventions sold in a market the Second Amendment specifically protects. That difference is what makes the takings question here distinctive, and why it has gone largely unexamined.
Component 1: Patents are Takings Clause property
The first hurdle to our theory is Oil States (2018), which described patents as “public franchises” and is sometimes read to foreclose patent-takings claims altogether. We think that reading asks the case to do more than it can bear. Oil States was an Article III decision. The question was whether Congress could route the review of an already-issued patent to an executive-branch body (the Patent Trial and Appeal Board) through inter partes review, rather than to a federal court. The Court said yes, because patents fall within the public-rights doctrine. The “franchise” label was a tool for answering that procedural question, not a ruling on whether patents are property the government must pay for when it destroys them.
Justice Thomas, writing for the majority, explained that the decision “should not be misconstrued as suggesting that patents are not property for purposes of the Due Process Clause or the Takings Clause.” He pointed to James v. Campbell and Florida Prepaid as the authorities keeping the question open. Those cases anchor a long tradition. James v. Campbell (1882), a general patent case, held that the government cannot use a patented invention “without just compensation, any more than it can appropriate or use without compensation land which has been patented to a private purchaser.” Florida Prepaid (1999) repeated that “patents have long been considered a species of property.” And in Horne v. Department of Agriculture (2015), the Court held that the Takings Clause protects personal property on the same terms as real property. When it looked for historical proof that the Clause reaches beyond land, one of the authorities it relied on was James, a patent case. That a 2015 takings decision rests in part on an 1882 patent case suggests the tradition is alive. That tradition is general: it does not depend on the Second Amendment at all. Patents are constitutional property on their own footing.
Component 2: Destruction is not cancellation
The heart of the paper is a distinction the post-Oil States case law has not drawn. There are two ways the government can harm a patent. The first is cancellation: the Patent Office determines, through inter partes review, that the claims never should have issued, and rescinds them. Courts have held that cancellation is not a taking (Christy, Golden), and we accept that. A patentee has no protected interest in claims that were invalid from the start.
The second is destruction: the patent remains formally valid (no tribunal has questioned its validity), but a regulator eliminates the field in which it has any economic value. For example, the bump-stock Final Rule did not invalidate a single Slide Fire patent. With few exceptions, it made the manufacture, sale, and possession of the patented device a federal crime. A patent is a right to stop other people from making and selling the invention, and that right is worth something only where someone would otherwise want to make and sell it. Once the device itself was contraband, there was no lawful market to exclude anyone from. The right to exclude survived on paper and had nothing left to operate on. The closest doctrinal analog is Ruckelshaus v. Monsanto (1984): when EPA disclosed a registrant’s trade-secret data, the trade secrets remained formally valid, but the disclosure destroyed the secrecy that gave them value, and the Court treated that as a taking. Our claim rests on destruction of the operative field, not cancellation of the entitlement.
The government’s best counter-authority is Andrus v. Allard (1979), which upheld a ban on the commercial sale of eagle-feather artifacts. But Andrus turned on facts the Court emphasized: the owners were still allowed to possess, transport, donate, or devise eagle-feather artifacts; only the right to sell such products was regulated away. Horne later drew the dividing line precisely there, distinguishing Andrus because the program before it “require[d] physical surrender” and transfer of title. The bump-stock Rule compelled exactly that kind of surrender and destruction, meaning Andrus does not protect it.
Component 3: The Second Amendment forecloses the police-power defense
Even assuming destruction, the government normally has a familiar escape hatch: the police-power doctrine, under which the regulation of dangerous or noxious things falls categorically outside the Takings Clause. The doctrine traces to Mugler v. Kansas (1887), and the Federal Circuit has extended it to federal seizures in cases like Acadia and Amerisource.
This is where the Second Amendment does its specific, load-bearing work. The police-power doctrine rests on a hidden assumption: that the government has broad authority to regulate the activity at issue. That assumption holds for alcohol, gambling devices, narcotics, and contraband, none of which the Constitution protects. It does not hold for arms. The Second Amendment limits the federal regulatory authority over arms at the source, and Heller, Bruen, and Rahimi give that limit real content. The Court returned to the point in Wolford v. Lopez (2026), which struck down Hawaii’s rule barring licensed carry on private property open to the public. Justice Jackson, in dissent, argued that the case was “about property rights, not gun rights.” The majority answered that States “may not adopt property-law rules that violate constitutional rights,” that a State may not “sidestep the Takings Clause” by enacting a law purporting to extinguish a property interest, citing Tyler v. Hennepin County (2023), and that the Second Amendment “is entitled to no less protection than other constitutional rights.” Although Wolford is a state law case, and it does not decide the federal question we raise, the express pairing of the Second Amendment and the Takings Clause still matters: the Court placed the Second Amendment and the Takings Clause on the same footing against a sovereign’s attempt to define a property interest out of existence. Neither a patent nor any takings claim was before the Court in Wolford; the Takings Clause enters only in a footnote answering the dissent, and we rely on it for that pairing and nothing more. So, the Second Amendment does not override the Takings Clause or beat the police-power doctrine head-on. It knocks out the assumption of broad regulatory authority on which the defense is built. Remove that assumption, and the defense has nothing to stand on; the takings claim then proceeds on components one and two alone.
We want to be precise about the boundaries, because the natural worry is that the argument proves too much. It does not. The framework is narrow and conjunctive. It reaches only federal action that (1) destroys, not merely diminishes, the operative field of a valid firearms patent, (2) falls outside the regulatory authority the Second Amendment permits, and (3) is later held to have exceeded the government’s statutory or constitutional power. By “held” we mean set aside by a court with authority to do so, as the Supreme Court did in Cargill. Whether a single district court’s vacatur is enough, and whether relief running only to the parties would be, are questions the paper takes up rather than assumes. We are not arguing that every federal firearms regulation is a taking; the paper says so expressly. A lawful, authorized regulation that merely affects value is analyzed differently, and the police-power defense carries real weight there. The most obvious adjacent case, pharmaceutical patents under the Inflation Reduction Act, falls outside our framework precisely because drugs are not constitutionally protected and the program regulates price rather than destroying the operative field.
Bump stocks may be the most prominent example, but they are not unique. ATF classified forced-reset triggers as machineguns by open letter in 2022; the Northern District of Texas vacated that classification in National Association for Gun Rights v. Garland. ATF classified certain firearms equipped with stabilizing braces as subject to the NFA by rule in 2023; the same district vacated that rule in Mock v. Garland, and the government let its appeal lapse. Go back further: the 1986 Hughes Amendment, which closed the machinegun registry to civilian transfers, destroyed the civilian market for newly manufactured machineguns, and with it much of the practical value of machinegun-related patents. Across four decades and multiple administrations, the same structure recurs: a previously patented firearms technology is classified in a way that destroys its economic value, the patents stay formally valid, the commercial market disappears, and no one is compensated.
The test is easier to understand by way of example. Take the 2023 brace rule.
The destruction question does the most work, and it turns on legal availability rather than price. A rule that taxed braced pistols, or barred their sale to minors, would have made the patents worth less; the market would have shrunk but survived. The brace rule did something else. It reclassified the braced firearm as a short-barreled rifle, so that making, selling, or possessing one without registration carried federal criminal exposure. The lawful market for the patented article closed. The question is not what share of value was lost, but whether the invention could still be lawfully sold at all.
The constitutional question is straightforward here. A braced pistol is a configuration of a handgun, the arm Heller placed at the center of the Second Amendment, and the Mock court found braced pistols in common use.
So is the authorization question. The Northern District of Texas vacated the rule, and the government abandoned its appeal.
Each condition holds, and compensation is owed for the period the rule was in force. That framing also marks the test’s limit. Where a patent keeps a lawful market (government procurement is the obvious one), the field is not destroyed, and the patentee’s remedy for the government’s own use lies in § 1498 rather than the Takings Clause. The takings claim reaches the market the regulation closed, and no further.
What is live now
We do not overstate where the law stands. The bump-stock takings claim has mostly lost so far. The Court of Federal Claims dismissed the original action in Modern Sportsman I (2019); the Federal Circuit affirmed dismissal in McCutchen (2021) on a Lucas “inhere in title” theory; the Supreme Court denied review. That theory holds that some limits come built into property: if the law already barred what the owner wants to do, enforcing that limit takes nothing. On the Federal Circuit’s view, the National Firearms Act’s machinegun ban had always limited title to bump stocks. We engage that adverse authority directly.
But the ground has shifted. Oil States expressly reserved the takings question. Cargill eliminated the statutory premise underlying McCutchen’s “inhere in title” reasoning. And in Modern Sportsman III (2025), the same judge who dismissed the claim in 2019 denied the government’s motion to dismiss the physical-takings claim under Horne and let the case proceed. The takings question the pattern raises is no longer hypothetical; it is being litigated.
Individual Justices have begun to make the connection, though not yet in a holding. Dissenting from the denial of certiorari in Snope v. Brown (2025), Justice Thomas reasoned that after Bondi v. VanDerStok it “seems that ATF could at any time declare AR-15s to be machineguns prohibited by federal law,” and that until the Court resolves whether the Second Amendment forecloses that possibility, “law-abiding AR-15 owners must rely on the goodwill of a federal agency to retain their means of self-defense.” That, he wrote, is “no constitutional guarantee at all.” Writing on the denial of certiorari in Wilson v. Hawaii (2024), he had already drawn the takings analogy directly: constitutional rights are “self-executing prohibitions on governmental action,” a violation of the Takings Clause occurs “at the time of the taking,” and “[t]he same principles apply to the Second Amendment.” Neither writing carries precedential weight. Both describe the position this paper is about, in which a patentee’s rights depend on an agency’s continuing forbearance.
The paper’s contribution is the doctrinal architecture that question requires: the cancellation-versus-destruction distinction, the Second Amendment as a constraint on police-power authority, the individual patent as the unit of analysis, and a compensation measure drawn from the established government patent-use cases. The trend shows no sign of stopping, and each time, the same question arises: who bears the cost of a rule the government was never authorized to issue?
