A fair question to put to this statute at the start is whether it is even worth studying. The Patent and Trademark Law Amendments Act of 1980, Public Law 96-517, is a short law by the standards of modern legislation, codified as 35 U.S.C. sections 200 through 212, and its core command can be stated in a single sentence: inventions made with federal money belong first to the people who made them, not to the government that paid for them. Universities and small businesses were permitted to keep title to federally funded inventions and to patent and license them, subject to conditions. That sounds like a technical fix to a procurement rule, the sort of thing that should have vanished into the archives after a few quiet years of implementation. It did not. It became one of the most consequential and most argued-over economic statutes of the late twentieth century, and the argument has not let up.

The bill that flipped the patent ledger and the fight over a power never fired

The reason to care is best captured as a test. After reading this article, a reader should be able to explain four things without hesitation. First, that before 1980 the federal government generally took title to inventions produced under its research funding and then licensed almost none of them, so taxpayer-financed discoveries sat on a shelf. Second, that the 1980 Act reversed the default and let universities, nonprofits, and small businesses elect to keep title, patent their inventions, and license them for commercial development. Third, that the government did not walk away empty handed: it retained a permanent, royalty-free license to use any such invention for government purposes, along with intervention powers that let a federal agency step in under specified conditions if the invention was being warehoused or withheld from the public. Fourth, and most strangely, that the most contested of those intervention powers, the so-called march-in rights, has never once been exercised by any federal agency in the law’s entire history. By early 2017, thirty-six years after the statute’s passage, every petition asking an agency to march in had been denied.

Hold those four facts together and the puzzle of the statute comes into focus. The law solved an incentives problem with remarkable speed: within a few years of its passage, universities that had treated patenting as a curiosity were building technology transfer offices, filing patents, and negotiating licenses, because the statute gave them both the right and the motive to do so. At the same time, the same law preserved a government backstop that was meant to reassure everyone that public money would not be abused, and that backstop became the center of a decades-long argument about prices, access, and the reach of government power. The argument is about a tool that has never been fired. Whether that fact proves the tool is working as a deterrent, or proves it was never meant to be used, is the dispute that has kept the statute alive in policy debates long after the patent ledgers were rewritten.

That dispute runs through everything else in this story. The defenders of the law say its genius was to pair decentralization with a safety valve: inventors and institutions got the exclusive rights they needed to attract development capital, and the government kept a narrowly defined power to intervene if exclusivity produced monopoly or neglect. The law’s critics say the safety valve was built to fail, drawn so narrowly and administered so reluctantly that it functions as decoration, and they have pressed for decades to use march-in rights as a lever against high drug prices. Neither side has prevailed outright, and that stalemate is part of the law’s character. It is a statute that works well enough that no Congress has been willing to tear it up, and is contested enough that no faction has ever declared victory. The opening third of this article explains how such a law came to exist: the broken system it replaced, the odd coalition that passed it, and the election-eve scramble that nearly killed it.

One more piece of context belongs up front. The Bayh-Dole Act was not the only technology transfer law of 1980. In October of that year, President Carter signed the Stevenson-Wydler Technology Innovation Act of 1980, Public Law 96-480, which pushed federal laboratories to move their own technology into the hands of outside users and set up offices of research and technology applications inside the labs. That law addressed the government’s in-house research enterprise, while the 1980 Act addressed inventions made outside the government under federal grants and contracts. A companion article in this series covers the university research landscape in detail; this article concentrates on the patent ownership change that made the rest of the machinery matter.

How many federal patent policies existed before the act?

There was no single policy to replace. By the late 1970s, more than twenty different federal patent policies were in force across the research agencies, each with its own clauses, exceptions, and habits of administration. What most shared was a presumption: the default answer to ownership was government title.

What figures did the committee report cite for the old system?

The committee report accompanying the bill cited roughly 28,000 federally owned patents accumulated under the government-title regime, and fewer than 5 percent of them had ever been licensed. Supporters repeated the accounting from the Comptroller General’s findings: tens of thousands of inventions, almost none reaching commercial use.

Which prominent admiral testified against the bill?

Admiral Hyman Rickover, the father of the nuclear navy, testified against the bill, arguing from nuclear industry experience that patents were not critical to development and that inventions paid for by the government should be freely available to all. Senator Russell Long of Louisiana, who chaired the Finance Committee, shared that skepticism.

What claim did Senator Bayh make about drugs under the old system?

Senator Bayh’s retrospective statement on the statute’s thirtieth anniversary made the sharpest version of the claim: among the accumulated federal inventions, not a single new drug had ever been commercialized while the government owned the patent. The point became the moral center of the reform case.

Why did exclusivity matter to the reformers’ argument?

An invention is not a product. Between a laboratory discovery and a medicine on a shelf lie years of expensive, risky development that no company will finance without assurance it can capture the returns. A government-held patent licensed to all comers gives no firm a reason to spend millions turning the invention into something real.

The pre-1980 patent problem

To understand what the statute changed, begin with the world it replaced, which was not a single policy at all. By the late 1970s, more than twenty different federal patent policies were in force across the various research agencies, each with its own clauses, exceptions, and habits of administration. There was no uniform rule for who owned an invention produced with federal money. What the policies shared, for the most part, was a presumption: the default answer was government title. A university researcher who made a discovery under a federal grant, or a small company that invented something under a federal contract, could generally expect the rights to flow to Washington. In some cases a contractor could petition an agency to let it keep rights, but the decision rested with the agency, the process was slow and uncertain, and the default posture of the system was that the public had paid for the research and therefore the public’s government should own the results.

The clearest expression of that presumption had been written into law more than two decades earlier for the space program. Section 305 of the National Aeronautics and Space Act of 1958 vested title to inventions made in the performance of NASA work in the United States, making the government-title rule an explicit statutory command rather than just an agency habit. (The details of that provision, and the controversies it produced, are covered in this series’ guide to the Space Act.) Other agencies reached similar outcomes through regulation and contract clauses rather than statutes, but the effect was broadly the same: the federal government accumulated one of the largest patent portfolios in the country, and it had no coherent plan for doing anything with it.

The committee report accompanying the bill cited the figures that became the signature evidence for reform. According to that report, roughly 28,000 federally owned patents had accumulated under the government-title regime, and fewer than 5 percent of them had ever been licensed. The same accounting was repeated by the statute’s supporters from the Comptroller General’s findings: tens of thousands of inventions caught in a web of inconsistent policies, almost none of them making it into commercial use. The numbers were contested at the margins in later years, but the direction was never in doubt. The government owned an enormous stock of inventions and licensed only a sliver of them. Something in the system was producing patents and preventing products.

The explanation offered by the reformers was a piece of commercialization logic that has since become familiar but was not yet conventional wisdom in Washington. An invention is not a product. Between the laboratory discovery and the medicine on a pharmacy shelf, or between the prototype and the manufacturing line, lies years of expensive, risky development work that no company will finance without some assurance that it can capture the returns. A patent that anyone may use for free, which is what a government-held patent licensed nonexclusively to all comers amounts to, gives no company a reason to invest millions in turning the invention into something real. Why would a firm spend the money to develop a technology if a competitor could wait for the work to be done and then use the same invention at no cost? The government-title regime, in this telling, had extinguished the fuel of the patent system itself. It had taken the one instrument that makes private development investment rational, exclusivity, and replaced it with a commons that nobody had an incentive to improve.

That logic carried a political corollary that proved decisive: the taxpayer argument. The federal government was funding roughly half of the nation’s research and development at the time, and the vast majority of basic research, the kind from which breakthrough products eventually emerge. Billions of taxpayer dollars went into that research each year. Under the existing policies, the public paid for the research, paid again through the agencies that administered the patents, and then received almost no return in the form of commercialized products, jobs, or new industries. Supporters of reform framed the old system not as protective of the public interest but as a waste of it. The public was not being protected by government ownership; it was being deprived of the fruits of its own investment by a system that made commercialization economically irrational. Senator Bayh’s own retrospective statement on the statute’s thirtieth anniversary made the point in the sharpest available form: the sponsors had found, among the accumulated federal inventions, that not a single new drug had ever been commercialized while the government owned the patent. That claim, drawn from the same body of Comptroller General and committee findings, became the moral center of the reform case.

There were, of course, dissenters from the commercialization logic, and they were formidable. Admiral Hyman Rickover, the father of the nuclear navy, testified against the bill and argued from experience in the nuclear industry that patents were not critical to development and that inventions paid for by the government should be freely available to all. His view had a distinguished lineage in American public science: the idea that knowledge bought with public money should be a public good, not a private monopoly. Senator Russell Long of Louisiana, who chaired the Senate Finance Committee and admired Rickover, shared that skepticism and was not a man whose opposition could be brushed aside. The reformers answered, through their witnesses and their committee record, that free availability had been tried for decades and had produced a warehouse of unlicensed patents rather than a flourishing of public goods. Free, in practice, had meant unused. The disagreement was not really about whether the public should benefit from research it funded. It was about what mechanism would deliver that benefit: open access to the invention as a piece of paper, or exclusive rights that gave someone a reason to turn the paper into a product.

The reformers also made a structural argument about where decisions should be made. A centralized bureaucracy in Washington, they said, could not know which of 28,000 inventions was worth developing, which university laboratory had the industry contacts to move a discovery forward, or which small company was best positioned to take a risk. The people who made the invention and the institutions that employed them had better information and stronger incentives than any federal patent administrator. Decentralizing ownership to universities and small businesses was therefore not just a reward to those performers; it was a claim about competence. The parties closest to the research would manage the resulting rights more effectively than the government ever could. That claim would be tested, and largely vindicated, in the decades that followed, but in 1980 it was still a hypothesis wrapped in a bill.

A lame duck passage

The partnership that carried the bill had a quality that contemporaries found remarkable and that history has treated as the statute’s signature feature. Birch Bayh was a liberal Democrat from Indiana, a two-term senator best known for his work on constitutional amendments and education policy. Bob Dole was a conservative Republican from Kansas, a figure of the party’s right wing who had been his party’s vice presidential nominee in 1976. On most issues they were opponents. On patent policy they were allies, drawn together by staff-level work that began when Purdue University visited Bayh’s Judiciary Committee staff to describe how federal patent policies were stripping inventions from their creators and letting them waste away in Washington. Dole’s office had developed similar concerns, and the two staffs agreed to work together. The result was a bill that drew co-sponsors from across the ideological spectrum, a coalition that supporters later invoked as proof that the problem was technical rather than partisan. When arch-liberals and arch-conservatives agree on the diagnosis, the thinking went, the diagnosis is probably right.

That bipartisanship was real, but it should not be mistaken for smooth sailing. The bill faced serious opposition in the Senate Judiciary Committee, where Rickover’s testimony and Long’s skepticism carried weight, and it nearly died more than once in the regular sessions of the 96th Congress. The reformers deliberately limited the bill’s scope to universities, nonprofits, and small businesses, partly because those performers were the crux of the commercialization problem, and partly to avoid provoking a full-scale fight over large defense contractors and their patent rights. Even with that limitation, the legislation ran out of time. Congress adjourned for the 1980 elections without passing it, and the elections changed everything. Ronald Reagan defeated Jimmy Carter in a landslide. Republicans won control of the Senate for the first time since the Truman administration. And Birch Bayh, one of the bill’s two namesakes, lost his own re-election bid to a young Republican congressman named Dan Quayle on November 4, 1980.

What happened next is one of the stranger passages in modern legislative history, and participants’ accounts agree on its outlines even where they differ on emphasis. Congress had adjourned without completing the federal budget, so it had to return for a lame duck session after the election, a session populated by defeated members serving out their final weeks and by incoming members not yet sworn in. That created one last window for the patent bill before Bayh left the Senate forever. Staffers on both sides scrambled to find a vehicle. Early efforts to attach the measure to must-pass House bills failed for lack of a suitable carrier. Then a deal emerged through the House Judiciary Committee, chaired by Representative Robert Kastenmeier of Wisconsin, who had introduced his own patent legislation as H.R. 6933. Kastenmeier’s committee passed its bill with the Bayh-Dole provisions inserted, and the House passed the combined measure on November 17, 1980. To become law, the Senate then had to pass the identical legislation, and because the rules of a lame duck session leave no time for extended debate, the bill needed unanimous consent. That consent materialized, in an act of senatorial courtesy toward a departing colleague that several accounts single out as decisive, and the bill cleared the Senate.

The final step belonged to a president who had just lost his own re-election and whose administration had not supported the bill. The Carter administration had its own ideas about federal patent policy and had not testified in favor of the Bayh-Dole approach during the hearings. Yet the enrolled bill went to the White House, and Carter signed it on December 12, 1980, in the closing weeks of his presidency, making it Public Law 96-517 and adding sections 200 through 212 to Title 35 of the United States Code. The signing came just over a month after Bayh’s defeat and barely five weeks before Reagan’s inauguration. The statute that would define American technology transfer for the next four decades entered the books as the act of a lame duck Congress and a lame duck president.

Why the timing mattered is straightforward, and it is worth stating plainly because it is easy to misread the episode as a quirk. Without the lame duck session, the bill was dead. It had failed to pass in the regular sessions, its principal Senate champion had been voted out, and the incoming Senate majority had no particular attachment to a Democratic senator’s departing project. The budget impasse that forced Congress back into session after the election supplied the only procedural opening the bill ever got. The unanimous consent that carried it through the Senate was available only because the bill’s scope was narrow, its coalition was bipartisan, and enough senators were willing to grant a defeated colleague a final legislative victory. Had the session not been called, or had a single senator objected, the Bayh-Dole Act as we know it would not exist, and the history of university patenting in the United States would have had to be written some other way, by some other Congress, at some other time.

Several myths have grown up around that passage, and they deserve correction. The first is that the law was a product of the Reagan era, a piece of the 1980s turn toward markets and deregulation. It was signed by Jimmy Carter, passed by a Democratic-controlled Congress, and shepherded by a liberal Democrat; Reagan’s contribution came later, when his administration extended the policy to large contractors by executive action in 1983 and when a 1984 presidential memorandum broadened its reach. The second myth is that the bill’s bipartisan sponsorship meant it was uncontroversial. It was controversial enough to stall for two years, to draw the opposition of one of the most formidable witnesses in Washington in Admiral Rickover, and to require a lame duck maneuver to survive at all. Bipartisanship got it a hearing; it did not get it a free pass. The third myth is that Carter signed it enthusiastically as a capstone of his innovation agenda. The record is more complicated. His administration had competing patent policy proposals and did not champion the bill during its hearings; the signature came at the end of a presidency, on legislation that had been reshaped by congressional staff and rescued by parliamentary circumstance. None of this diminishes the statute. If anything, it clarifies what kind of achievement it was: not a president’s program or a party’s platform, but a staff-driven, coalition-built repair of a broken incentive system, passed in the last weeks it could possibly have passed, by people who understood that the window was closing.

Before Bayh-Dole: memoranda, waivers, and the IPA experiment

For more than three decades after World War II, the United States had no single statute allocating patent rights in federally funded research. Each agency improvised under its own enabling law and its own habits. The result was a patchwork: a contractor might keep title under one agency and lose it under another, sometimes for the same kind of work. Inventions accumulated in government portfolios by the tens of thousands while companies, unsure whether they could ever own or exclusively license the resulting patents, declined to invest in developing them. Before 1963 there was not even a government-wide framework to argue about. Since World War II the federal government had steadily expanded its support for research and development, and each agency’s patent policy had been fashioned without central guidance or overall coordination. Some agencies’ enabling statutes assigned inventions to the government outright, while others left the question to contract negotiation. A contractor’s rights therefore depended less on the nature of the invention than on the identity of the funding agency and the prevailing agency practice of the moment. The first attempt to bring order to this system came not from Congress but from the White House.

On October 10, 1963, President John F. Kennedy issued a memorandum to the heads of the executive departments and agencies carrying a Statement of Government Patent Policy. The memorandum was a compromise. Rather than imposing one uniform rule, it offered agencies a flexible framework with two poles: a title policy, under which the government would take principal or exclusive rights, and a license policy, under which the contractor would keep title while the government received a royalty-free license. Section 1 of the statement laid out guidelines for choosing between them, keyed to factors such as the contractor’s established commercial position and the field and purpose of the work. The government also reserved march-in rights. Under section 1(f) it could require a contractor permitted to own an invention to grant licenses if the invention was not brought to practical application within three years after the patent issued. Under section 1(g) it could act when the invention was required for public use under government regulations or to meet health needs or other public purposes stated in the contract. To coordinate this discretionary system, the memorandum established an interagency Patent Advisory Panel under the Federal Council for Science and Technology.

The waiver was the pressure valve of the Kennedy framework, and it ran on petitions. A contractor, or a prospective contractor negotiating an agreement, could ask the funding agency to waive the government’s claim, in whole or in part, for an identified invention or for a whole class of inventions that might arise under the work. In the case of an identified invention, even the individual inventor could petition. The agency then decided whether granting greater rights than the standard clause allowed would serve the public interest and the purposes of the contract. Nothing in the memoranda told the agency how much weight to give commercialization prospects versus open dissemination, so the standard was applied differently from one agency to the next. In practice this meant case-by-case decisions by individual agencies, each applying its own reading of the guidelines. A Department of Health, Education, and Welfare review later described more than twenty-two different agency policies that organizations had to navigate when administering grants and contracts. Outcomes depended on which agency funded the work and how generously it exercised its waiver authority. Contractors bore the burden of the petition process, and investors bore the uncertainty: a favorable decision from one agency said nothing about the next.

For universities the burden was especially acute. A university laboratory might hold grants from half a dozen agencies, each with its own patent clause and its own waiver habits. Research administrators had to track which inventions belonged to which regime and whether a petition was worth the time and the legal cost. A company considering an exclusive license to a university invention could not be sure the university would even be allowed to grant one, because the underlying rights might still rest with the funding agency.

Congress noticed the disorder. In 1965 an attempt to impose greater uniformity through legislation failed, and the failure mattered. With no statute on the horizon, the executive branch was left to govern patent rights by memorandum, and the agencies were left to implement those memoranda more than twenty-two different ways. The memoranda had one structural weakness that no revision could fix: they bound the executive branch only as long as the sitting president’s policy held, and each agency still implemented them in its own way.

On August 23, 1971, President Richard Nixon issued a revised memorandum and statement of government patent policy, printed at 36 Federal Register 16889, drawing on the experience gained under the Kennedy framework. The Nixon statement preserved the flexible architecture and the waiver system but tightened several provisions. The march-in rights were broadened. The working requirement that an invention be brought to practical application was extended to assignees and licensees, and the government could require the granting of an exclusive license to a responsible applicant on reasonable terms, rather than only nonexclusive licenses. The health and public-use march-in authority was carried forward. Agencies were directed to follow the new uniform provisions, and the petition-and-waiver process continued as the principal route by which contractors obtained title.

Meanwhile the nonprofit sector, and especially research universities, had begun pressing for a different solution: advance agreements that settled patent rights before the work started, eliminating the need for post-invention petitions. The Department of Health, Education, and Welfare answered with the Institutional Patent Agreement. Under an IPA, an institution with an approved patent policy could retain title to inventions arising under its grants and contracts from the moment of contracting, subject to the government’s license and march-in protections. HEW had experimented with such agreements since 1955, executing seventeen of them, but the 1968 reorganization of the Department prompted a fresh start. The old agreements were terminated and a uniform IPA was developed in collaboration with the Department’s patent staff, operating agencies, and senior leadership. The first uniform IPA was executed on December 31, 1968, and the Department went on to execute seventy-two in total.

Qualifying for an IPA was not automatic. An institution had to demonstrate that it maintained a genuine patent management program: procedures for surfacing invention disclosures, forms ensuring that patent rights vested in compliance with university policy, and a delivery system designed to move inventions to the marketplace in a manner most likely to benefit the public. A university that made that showing was entitled to acquire an institutional patent agreement, and under it title vested in the university at the time of contracting, removing the after-the-fact petition entirely. For the universities that depended on HEW grants, the program set the terms of daily academic patent management.

The IPA program was the first crack in the wall of government title. It demonstrated that universities could be trusted to manage patent rights in advance, with title vesting in the institution at the time of contracting rather than after a petition, and it gave other agencies a working model to study. When nonprofit organizations dealing with other agencies asked for something similar, an ad hoc committee under the Federal Council for Science and Technology was formed in 1971 to explore a uniform government-wide policy. After four years of interagency meetings its July 1975 report recommended a modified HEW IPA approach. Still, the IPA was an HEW program, not a statute, and it covered only institutions that qualified.

The drafters of the 1980 Act worked against this background. The Nixon statement’s expansion of march-in to assignees and licensees, and its provision for exclusive licensing of a responsible applicant, showed the executive branch groping toward the same tools Congress would later codify. The IPA program showed that advance allocation of rights to performers could work in practice. What remained missing was the one thing memoranda could not supply: a single, durable, government-wide rule that an investor could rely on regardless of which agency funded the work or which president sat in the White House.

That was the fundamental limit of the whole pre-1980 regime. Even the Nixon revision betrayed it. The 1971 statement was issued precisely because experience under the Kennedy framework had exposed its rough edges, yet the fix was another memorandum carrying the same structural weakness. The Federal Council for Science and Technology spent four years studying the problem, and its July 1975 report recommended a modified HEW IPA approach for government-wide use, but a recommendation is not a law. Agencies could adopt it, adapt it, or ignore it. By the late 1970s the pattern was clear: the executive branch could tinker with the framework indefinitely without ever producing the uniform, reliable allocation of rights that contractors and investors needed. Presidential memoranda could be revised or abandoned by the next administration. Waivers were discretionary, unpredictable, and granted agency by agency. Institutional patent agreements covered only a subset of grantees at a single department. None of these instruments could give every contractor, in every agency, the same reliable rule: that the performer of the research could elect to keep title, subject to defined government safeguards. An investor deciding whether to fund the long development of an early-stage invention could not price the risk that a waiver would be denied or that a new president would change the policy. Congress would eventually supply that uniform, statutory answer. But the questions it answered, and the instruments it replaced, were shaped by seventeen years of memoranda, petitions, and the IPA experiment.

The election mechanics

The core of the statute is a single grant of power, and everything else in the law is scaffolding around it. Section 202(a) of title 35 provides that each nonprofit organization or small business firm may, within a reasonable time after disclosure as required by section 202(c)(1), elect to retain title to any subject invention. That election right reverses the default that governed most federally funded research before the 1980 Act, under which the Government routinely took title to inventions its money had produced and then sat on them. Congress concluded that patents left in Government hands were, as a practical matter, dead patents, and that the path from laboratory bench to marketplace ran through ownership by the institution or company that had done the work.

Eligibility for that right is narrower than many people assume. The statute names only two categories of contractor: nonprofit organizations, a term that in practice covers universities, medical research institutions, and similar tax-exempt research performers, and small business firms, defined under the Small Business Act’s size standards. By its terms the law does not extend the election right to large businesses, and that limitation is a recurring point of confusion. Commentators, licensees, and even some federal employees routinely speak as though the 1980 Act governs every contractor equally, but the text at section 202(a) says what it says: the election to retain title belongs to the nonprofit organization or the small business firm. Any title right a large business contractor holds under a federal funding instrument comes from contract language or later executive-branch patent policy, not from this statute. The distinction matters because the rest of the machinery described here, the disclosure clock, the election window, the Government licence, and the manufacturing condition, attaches to the title the statute gives, and understanding who can receive that title is the first step in applying the law.

The invention that title covers is itself a defined term. Section 201(e) defines a “subject invention” as any invention of the contractor conceived or first actually reduced to practice in the performance of work under a funding agreement. The two halves of that definition each carry weight. Conception is the formation in the inventor’s mind of a definite and permanent idea of the complete and operative invention; actual reduction to practice is the construction and testing of a working embodiment. An invention conceived in private work before the federal money arrived, but first actually reduced to practice while the contractor performed the funded work, is a subject invention. The definition reaches inventions even when conception and reduction are split across the funding line, a breadth that university technology transfer offices learned early to treat with respect.

The “funding agreement” that supplies the other half of the definition is defined at section 201(b) as any contract, grant, or cooperative agreement entered into between any Federal agency, other than the Tennessee Valley Authority, and any contractor for the performance of experimental, developmental, or research work funded in whole or in part by the Federal Government. The term also sweeps in assignments, substitutions of parties, and subcontracts entered into for the performance of the work under the agreement, so a university subcontractor performing research under a prime contractor’s federal contract is within the statute’s reach. Each of these instruments is a different channel through which federal research dollars flow, and the legal differences among the instruments of federal research funding shape how the patent clause gets inserted into the paperwork. But once the clause is in, the election mechanics operate the same way regardless of whether the money moved as a procurement contract, a grant, or a cooperative agreement.

The election right is subject to four circumstances in which a funding agreement may provide otherwise, meaning the agency may take title instead of the contractor. Section 202(a) lists them as clauses (i) through (iv): first, when the contractor is not located in the United States or does not have a place of business located in the United States or is subject to the control of a foreign government; second, in exceptional circumstances when the agency determines that restriction or elimination of the right to retain title to any subject invention will better promote the policy and objectives of the chapter; third, when a Government authority authorized by statute or Executive order to conduct foreign intelligence or counterintelligence activities determines that restriction or elimination of the right to retain title is necessary to protect the security of such activities; and fourth, where the funding agreement includes the operation of a Government-owned research or production facility. The fourth circumstance is stated here in the original 1980 phrasing; the 1984 amendments narrowed it, as the later section on those amendments explains. The statute does not let an agency invoke these exceptions casually. Section 202(b) requires that the agency first determine that at least one of the four conditions exists, and, except for determinations under the intelligence clause, the agency must file a copy of its determination with the Secretary of Commerce within thirty days after the award of the funding agreement. For exceptional-circumstances determinations, the statement must include an analysis justifying the determination, and for funding agreements with small business firms, copies go as well to the Chief Counsel for Advocacy of the Small Business Administration. The Secretary of Commerce, who administers the Government-wide scheme, can then advise the agency head when an individual determination or a pattern of determinations looks contrary to the policies and objectives of the chapter.

Once the contractor holds the election right, the statute imposes a disciplined sequence of obligations, set out in section 202(b) and given concrete timeframes in the implementing regulations at 37 CFR part 401, whose standard patent rights clause at 37 CFR 401.14 is required to appear in funding agreements covered by the law. The first step is disclosure. The contractor must disclose each subject invention to the Federal agency within a reasonable time after it becomes known to contractor personnel responsible for patent matters, and the regulations fix that reasonable time at two months: the contractor discloses to the agency within two months after the inventor discloses the invention in writing to contractor personnel responsible for patent matters. The disclosure takes the form of a written report identifying the funding agreement and the inventors and conveying, to the extent known, the nature, purpose, operation, and the physical, chemical, biological, or electrical characteristics of the invention.

The second step is the election itself. The contractor elects in writing whether or not to retain title to the invention by notifying the Federal agency within two years of disclosure to the agency. The two-year period can be cut short where the clock on patentability is already running: where a publication, public use, sale, or other availability to the public has started the one-year statutory period in which valid patent protection can still be obtained in the United States, the agency may shorten the election period to a date no more than sixty days before the end of that statutory period.

The third step is filing. The contractor that elects to retain title must file its initial patent application on the subject invention within one year after the election, or earlier if the end of the statutory period for obtaining valid patent protection in the United States arrives sooner. If the initial filing is a provisional application, the contractor must follow it with a nonprovisional application within ten months. The fourth element of the sequence is reporting: the agency has the right to require periodic reporting, no more often than annually, on utilization or efforts at obtaining utilization being made by the contractor, its licensees, or its assignees. Agencies also require a report before the closeout of the funding agreement listing all subject inventions or stating that there were none.

The sequence is a use-it-or-lose-it regime. A contractor that discloses late, elects late, or files late gives the Federal agency a path to take title to the subject invention, and the case law treats missed deadlines seriously: in Campbell Plastics Engineering and Manufacturing, Inc. v. Les Brownlee, 389 F.3d 1243 (Fed. Cir. 2004), the Federal Circuit held that an invention lost to the Government when the contractor failed to comply with the disclosure provisions. Technology transfer offices therefore treat the two-month disclosure window and the two-year election window as compliance dates on par with patent filing deadlines. The discipline was deliberate. Congress wanted inventions out of the laboratory and into the patent system on a predictable schedule, and the clocks in section 202(b) and the regulations are the mechanism that keeps that schedule.

The government’s retained licence

No matter how cleanly the contractor navigates the disclosure and election sequence, one encumbrance travels with the title permanently. Section 202(c)(4) provides: “With respect to any invention in which the contractor elects rights, the Federal agency shall have a nonexclusive, nontransferrable, irrevocable, paid-up license to practice or have practiced for or on behalf of the United States any subject invention throughout the world.” Every word in that sentence was chosen to give the Government something it could never lose, and the provision is best understood by reading the grant attribute by attribute.

Nonexclusive means the Government’s licence does not diminish the contractor’s own rights. The contractor keeps title, keeps the power to exclude infringers, and keeps the power to license, exclusively or otherwise. The Government licence sits alongside those rights rather than displacing them; a patent can have a commercial owner and the Government as a parallel licensee without conflict. Nontransferrable means the licence stays with the Government. The agency cannot sell it, assign it to a private party, or spin it off as an asset, which keeps a public right from becoming a private one through the back door. Irrevocable means it cannot be taken back, not by the contractor, not by a later administration, and not by a court revisiting the bargain. Paid-up means no royalties are owed: the Government has already paid, in the form of the research funding that produced the invention, and it will not pay twice.

“To practice or have practiced for or on behalf of the United States” is broader than a simple right of the Government to use the invention in its own laboratories. The phrase “have practiced” lets the Government authorize a contractor to practice the invention on the Government’s behalf, so a defense procurement office can buy a product embodying the invention from a private manufacturer and that manufacturer is covered by the Government’s licence when it makes the product for the Government contract. “For or on behalf of the United States” sets the purpose boundary: the licence authorizes Government use, not private commercial exploitation by private parties for their own account. “Throughout the world” removes any geographic limitation, so the licence is not confined to American soil. The practical result is what the brief states plainly: the Government never loses the ability to use what it funded. A university that patents a federally funded cancer diagnostic, or a small business that patents a federally funded sensor, can license, exclude, and commercialize, but it can never sue the Government for using the invention in Government work, and it can never collect a royalty from the Government for that use.

This standing licence is frequently confused with the march-in authority of section 203, and the confusion is worth dispelling because the two provisions operate on entirely different principles. The licence at section 202(c)(4) is automatic and unconditional: it vests when the contractor elects rights, it requires no agency finding, no proceeding, and no showing of need, and it persists for the life of the patent whether or not the invention is ever commercialized. March-in, by contrast, is an intervention power. Under section 203, a funding agency may require the contractor or its licensee to grant licences to responsible applicants, on reasonable terms, only after the agency makes one of four findings: that the contractor has not taken effective steps to achieve practical application of the invention; that action is necessary to alleviate health or safety needs not reasonably satisfied; that action is necessary to meet requirements for public use specified by federal regulations; or that the domestic manufacturing condition has been breached and no waiver obtained. March-in is contingent, adversarial, and procedural, and through 2017 it had never been exercised. The section 202(c)(4) licence needs no exercise at all. It is a standing property interest of the United States, present in every elected subject invention from the moment of election, doing its work quietly in the background of every licensing negotiation that follows.

The licence travels with two companion obligations that keep the Government informed about what it holds. Section 202(c)(5) gives the Federal agency the right to require periodic reporting, but no more frequently than annually, on the utilization of the subject invention or on the efforts the contractor, its licensees, or its assignees are making to obtain utilization. Inventions reported through the iEdison system, the Government-wide invention reporting platform administered for the National Institutes of Health, are subject to annual utilization reporting, including the status of development, the date of first commercial sale or use, and gross royalties received. The statute treats utilization information gathered this way, including information obtained in a march-in proceeding, as privileged and confidential commercial and financial information, exempt from disclosure under the Freedom of Information Act. Section 202(c)(6) imposes the Government-support statement: when a United States patent application is filed by or on behalf of the contractor or its assignee, the specification of the application and of any patent issuing on it must include a statement specifying that the invention was made with Government support and that the Government has certain rights in the invention. The Patent Office’s Manual of Patent Examining Procedure directs that this appear at the beginning of the specification, in the standard form naming the contract and the funding agency. The statement is the public face of the licence, a flag on every federally funded patent telling the world that the Government’s retained rights are inside.

The manufacturing condition

If the election right is the statute’s grant and the retained licence its price, the manufacturing condition at section 204 is its industrial policy. The section is titled “Preference for United States industry,” and its text begins with the strongest possible override: “Notwithstanding any other provision of this chapter.” What follows is a rule aimed squarely at the commercialization channel the statute was built to create, the exclusive licence:

“No small business firm or nonprofit organization which receives title to any subject invention and no assignee of any such small business firm or nonprofit organization shall grant to any person the exclusive right to use or sell any subject invention in the United States unless such person agrees that any products embodying the subject invention or produced through the use of the subject invention will be manufactured substantially in the United States.”

Three features of the rule deserve attention. First, it reaches only exclusive rights to use or sell in the United States. A nonexclusive licence, or an exclusive licence covering only foreign territory, is outside the provision’s terms. Congress aimed at the transaction that most fully commits the invention’s American market to a single partner. Second, it binds assignees as well as the original title holders. A university cannot wash out the condition by assigning the invention to a patent-holding company that then grants the exclusive licence; the prohibition follows the title into the assignee’s hands. Third, the domestic-manufacture promise covers both products embodying the subject invention and products produced through the use of the subject invention, so a process invention is within the condition just as a product invention is. What the exclusive licensee must agree to is substantial manufacture in the United States. The statute and its regulations do not define “substantially,” and the standard has been read as a flexible one, turning on the facts of the product and its production chain rather than on a fixed percentage.

The statute then provides the escape valve, and it is a narrow one: “However, in individual cases, the requirement for such an agreement may be waived by the Federal agency under whose funding agreement the invention was made upon a showing by the small business firm, nonprofit organization, or assignee that reasonable but unsuccessful efforts have been made to grant licenses on similar terms to potential licensees that would be likely to manufacture substantially in the United States or that under the circumstances domestic manufacture is not commercially feasible.” The waiver is granted in individual cases, not by class, and only by the funding agency itself. The showing has two alternative prongs. The applicant may demonstrate that it tried, reasonably but unsuccessfully, to license on similar terms to potential licensees likely to manufacture substantially in the United States, which requires evidence of a genuine domestic licensing effort, not a perfunctory one. Or the applicant may show that under the circumstances domestic manufacture is not commercially feasible, a standard that reaches cases where no domestic production base exists for the product, where scale or cost structures make American manufacture uneconomic, or where the nature of the technology resists domestic production. The agency weighs the showing, and the burden rests on the party seeking the waiver. Through the history of the statute, waivers have been granted sparingly, and the Department of Energy has built its own elaboration of the condition into its financial assistance awards, including manufacturing plans submitted with funding applications describing how the applicant intends to manufacture domestically, consistent with the exceptional-circumstances authority at section 202(a)(ii).

Why Congress wrote the condition is stated in the policy declaration at section 200, which lists among the objectives of the chapter “to promote the commercialization and public availability of inventions made in the United States by United States industry and labor.” The 1980 Act was born in a period of acute anxiety about American industrial competitiveness. The late 1970s brought stagflation, a rising trade deficit, and the sense that American research was being commercialized abroad while domestic industry stagnated. A statute that let universities and small businesses own federally funded inventions, and then let those owners grant exclusive rights to manufacturers with no domestic-production obligation, would have struck the Act’s sponsors as a taxpayer subsidy for foreign manufacturing. The manufacturing condition closes that loop. The public pays for the research; the public’s industry should get the manufacturing. The condition interacts with the election right as a limit on what the election right can be sold for. The contractor owns the invention and may commercialize it freely, but when it chooses the most potent commercialization instrument, an exclusive licence for the American market, the licensee must commit to American production or the funding agency must be persuaded to waive the requirement. The condition does not restrict the contractor’s ownership. It restricts the contractor’s power to hand the American market to a manufacturer that will build the product elsewhere, and in that restriction lies the statute’s answer to the question of whose industry the public’s research money is meant to serve.

The nonprofit bargain: licensing conditions and the flow-down

The right of a university or other nonprofit organization to elect title under 35 U.S.C. 202(a) is not a grant of ordinary ownership. Section 202(c)(7) attaches a set of conditions that apply only to nonprofit contractors, and they function as the price of the election right. Together they express a single legislative judgment: a nonprofit that takes title to a taxpayer funded invention holds that title as a trustee of public money, not as a private proprietor.

The first condition, subparagraph (A), is a prohibition on assignment. A nonprofit contractor may not assign its rights in a subject invention in the United States without the approval of the funding agency. The statute provides one exception: assignment is permitted without agency approval to an organization that has as one of its primary functions the management of inventions. Entities such as Research Corporation, the archetypal university patent management organization, fall within this exception. The condition matters because the prohibition on selling the patent outright prevents a university from simply flipping a federally funded discovery to the highest bidder. The exception matters because it channels any transfer toward organizations whose business is the stewardship of inventions, and the statute adds that such an assignee is subject to the same provisions as the contractor, so the conditions travel with the rights.

Subparagraph (B) requires the contractor to share royalties with the inventor. Subparagraph (C) requires that the balance of any royalties or income earned by the contractor with respect to subject inventions, after payment of expenses incidental to the administration of subject inventions including payments to inventors, be used for the support of scientific research or education. The two provisions operate as a unit. The individual whose laboratory work produced the invention receives a share of the proceeds, and everything that remains must flow back into the research and educational mission rather than into general revenue. The implementing regulations recognize one exception to subparagraph (C): funding agreements for the operation of a government owned, contractor operated facility carry a different royalty disposition scheme, in which royalties up to a defined portion of the facility budget support research, development, and education consistent with the facility mission, and amounts above that threshold are divided between the contractor and the Treasury.

Subparagraph (D) imposes a small business preference. Except where it proves infeasible after a reasonable inquiry, a preference in the licensing of subject inventions must be given to small business firms. The implementing clause, Federal Acquisition Regulation 52.227-11(i)(4), gives the preference operational content: the contractor must make efforts that are reasonable under the circumstances to attract small business licensees, and it must give a preference to a small business concern when licensing if the contractor determines that the concern has a plan or proposal for marketing the invention which, if executed, is equally as likely to bring the invention to practical application as any plans or proposals from applicants that are not small business concerns, provided the contractor is satisfied that the small business concern has the capability and resources to carry out its plan. The decision in any specific case rests with the contractor, and the Secretary of Commerce may review the contractor’s licensing program and negotiate changes when the review discloses that the contractor could take reasonable steps to implement the preference more effectively.

Alongside these nonprofit specific conditions, every funding agreement carries the manufacturing condition of section 204, which the nonprofit licensing program must observe: no exclusive right to use or sell a subject invention in the United States may be granted unless the licensee agrees that any products embodying the invention or produced through its use will be manufactured substantially in the United States. The section 204 condition applies to all contractors, and for nonprofits it is one more term of the trust.

Congress imposed these conditions because the Bayh-Dole bargain was built on a specific theory of the university. The university was not receiving a reward for its work; it was receiving custody of a publicly funded asset on the theory that it was the institution best positioned to move that asset toward practical use. The assignment restriction prevents the university from treating the patent as an ordinary alienable asset. The royalty provisions prevent it from treating licensing income as a private windfall. The small business preference prevents it from reserving commercialization opportunities for its most powerful industrial partners. The domestic manufacturing condition prevents the commercialization it fosters from delivering its production value elsewhere. Each condition limits, in a different way, the resemblance between nonprofit title and ordinary private ownership.

The same trust theory runs through the flow-down provisions. Section 401.14(g) of title 37 of the Code of Federal Regulations requires the contractor to include the standard patent rights clause, suitably modified to identify the parties, in all subcontracts, regardless of tier, for experimental, developmental, or research work to be performed by a subcontractor. The subcontractor retains all rights provided for the contractor in the clause, and the contractor may not, as part of the consideration for awarding the subcontract, obtain rights in the subcontractor’s subject inventions. In other words, a prime contractor cannot use its position in the contracting chain to strip patent rights from a university or small business subcontractor, and the small organization performing the actual research keeps the same election rights the prime holds. Paragraph (g)(3) goes further: where the prime award with the federal agency is a contract, rather than a grant or cooperative agreement, the mutual obligations of the parties created by the clause constitute a contract between the subcontractor and the federal agency with respect to the matters covered by the clause. The subcontractor’s relationship with the government is direct, not mediated through the prime.

This flow-down structure explains why the nonprofit conditions are not an appendage to Bayh-Dole but a statement of its character. A subcontractor that is a nonprofit organization inherits the same assignment restriction, the same royalty obligations, and the same small business preference as the prime, and it enforces its rights directly against the agency. The statute treats the university, the small firm, and the subcontractor alike as stewards of the public investment, each answerable to the agency and to the conditions that made their title possible. Nonprofit title under the Act is therefore ownership with duties attached: duties to the inventor, to the research mission, to small business, to domestic manufacture, and to the government that funded the work.

The 1984 amendments and the large-business extension

On February 18, 1983, President Reagan issued a memorandum on government patent policy to the heads of executive departments and agencies. The memorandum directed federal agencies to extend the terms and provisions of the Bayh-Dole Act to all government contractors, regardless of size. In practice that meant a contractor performing federally funded research and development could be granted title to patents arising from the funded work, large company or small, in exchange for royalty-free use of the patented invention by or on behalf of the government. The directive was an executive act rather than a statute, and it landed on top of a law that had been deliberately drawn narrower.

As enacted in 1980, the Bayh-Dole Act gave the election-of-title right to only two classes of performers: nonprofit organizations and small business firms. Every other contractor, meaning companies of ordinary commercial size, remained governed by the patchwork of agency patent policies that the 1980 Act had been enacted to escape. Those policies traced to the statement of government patent policy issued in August 1971, implemented through agency regulations, contract clauses, and case-by-case waiver decisions. A large company performing federally funded research might eventually obtain rights in its inventions, but only through that fragmented, agency-by-agency machinery. The 1980 Act itself gave it no statutory election right. The choice to begin with nonprofits and small businesses reflected the coalition that carried the bill through the Senate in 1980 and the judgment, recorded in the Senate report, that universities and small firms lacked the bargaining position to secure rights under the old waiver system; large contractors were expected to continue operating under the existing agency policies.

The Reagan memorandum announced the administration’s decision to end that two-track arrangement. Congress then supplied the statutory footing. Title V of the Trademark Clarification Act of 1984, Public Law 98-620, signed November 8, 1984 and headed “Government Research and Development Patent Policy,” amended chapter 18 of title 35. Section 501 of that title rewrote several Bayh-Dole provisions, and together the changes did two things at once: they brought large businesses inside the Bayh-Dole regime and they strengthened the government’s oversight of the statute’s exceptions.

The codification of the large-business extension ran through section 210(c). That subsection had always provided that nothing in the chapter limited agency authority to agree to the disposition of rights in inventions made under funding agreements with persons other than nonprofit organizations or small business firms, in accordance with the statement of government patent policy. The 1984 amendment updated the date reference in that provision from August 23, 1971 to February 18, 1983, the date of the Reagan memorandum, and it added a new mandatory floor: all funding agreements, including those with other than small business firms and nonprofit organizations, had to include the requirements established in paragraph 202(c)(4) and section 203. The effect was to extend the Bayh-Dole disposition regime to large contractors while attaching the statute’s core public protections to every such agreement. The government-use license in 202(c)(4), the nonexclusive, nontransferable, irrevocable, paid-up license to practice or have practiced any subject invention throughout the world on behalf of the United States, and the march-in provisions of section 203 became nonnegotiable terms even where the contractor retained title under the memorandum’s policy rather than under section 202 itself.

The amendments also reworked the circumstances in which a funding agency could take title up front under section 202(a), and here the change ran in the opposite direction, narrowing rather than expanding government authority. The original 1980 Act provided that a funding agreement could provide otherwise when the funding agreement was for the operation of a Government-owned research or production facility. That clause gave agencies broad authority to withhold the election right at any government-owned, contractor-operated facility. The 1984 Act struck that language from clause (i) and replaced it with a different circumstance entirely: when the contractor is not located in the United States or does not have a place of business located in the United States or is subject to the control of a foreign government. It then added a new clause (iv) covering the government-owned facility situation in far narrower terms: when the funding agreement includes the operation of a Government-owned, contractor-operated facility of the Department of Energy primarily dedicated to that Department’s naval nuclear propulsion or weapons related programs, and all funding agreement limitations under this subparagraph on the contractor’s right to elect title to a subject invention are limited to inventions occurring under the above two programs of the Department of Energy.

The evolution is worth stating exactly, because it shows what the amendment took away as well as what it added. The 1980 phrasing reached every government-owned research or production facility, whichever agency operated it. The amended text reaches only Department of Energy facilities, only those primarily dedicated to naval nuclear propulsion or weapons-related programs, and only inventions occurring under those two programs. For the Energy Department’s other facilities and its other programs, the broad 1980 title-taking authority disappeared, and the regular election right applied.

The 1984 Act also moved oversight of the election-right exceptions out of the agencies and toward the Department of Commerce. Section 501(4) rewrote paragraphs (1) and (2) of section 202(b). An agency exercising its rights under subsection (a) first had to determine that at least one of the conditions in clauses (i) through (iv) existed, and except in the case of subsection (a)(iii), the intelligence and counterintelligence exception, it had to file a copy of the determination with the Secretary of Commerce within thirty days after the award of the funding agreement. In the case of a determination under the exceptional-circumstances clause, subsection (a)(ii), the filing had to include an analysis justifying the determination. If the Secretary of Commerce believed that any individual determination or pattern of determinations was contrary to the policies and objectives of the chapter or otherwise not in conformance with it, the Secretary was to so advise the head of the agency concerned and the Administrator of the Office of Federal Procurement Policy, and recommend corrective actions. That same paragraph authorized the Administrator of the Office of Federal Procurement Policy to issue regulations describing classes of situations in which agencies could not exercise the authorities of clauses (i) or (ii), whenever the Administrator determined that agencies were using that authority contrary to the chapter’s policies and objectives. A new paragraph (4) of section 202(b) added a contractor remedy: if the contractor believed a determination was contrary to the policies and objectives of the chapter or constituted an abuse of discretion, the determination was subject to the appeals procedure of the last paragraph of section 203.

The Act also amended section 203 itself, but not by adding a new march-in ground. Section 501(9) of the 1984 Act added “(1.” before the word “With” in the first line of section 203 and appended a new paragraph (2) at the end of the section: determinations under section 203 or section 202(b)(4) were not subject to the Contract Disputes Act, an administrative appeals procedure was to be established by regulations under section 206, and an adversely affected contractor, inventor, assignee, or exclusive licensee could petition the United States Claims Court within sixty days to affirm, reverse, remand, or modify the agency’s determination. The four substantive march-in grounds, including the fourth ground concerning the section 204 manufacturing agreement, were not added by the 1984 amendments. The amendment history of section 203 records no substantive change to the grounds between the original Act of December 12, 1980 and the 1984 amendments; the section 204 ground dates to the original 1980 Act.

Implementation followed the statute with some lag. The Department of Commerce, newly designated as the chapter’s overseeing agency, issued a consolidated rulemaking on March 18, 1987, published at 37 C.F.R. Part 401, that brought the 1980 Act, the 1984 amendments, OMB Circular A-124, and the Reagan memorandum into a single regulatory framework. That consolidation matters for understanding what the 1984 amendments meant in practice: the memorandum’s directive that all contractors be treated under Bayh-Dole principles was no longer an administrative instruction resting on presidential authority alone, but a policy the statute itself recognized and fenced. The amendment to section 210(c) replaced the August 1971 Nixon statement of government patent policy, the document that had governed the old agency-by-agency waiver system, with the February 18, 1983 memorandum as the operative statement of policy. Symbolically as well as legally, the era in which each agency ran its own patent policy, deciding contractor rights through bespoke regulations and individual waiver petitions, gave way to a uniform national regime in which large contractors enjoyed the same election framework as universities and small firms, subject to the same statutory safeguards.

Two further 1984 changes deserve mention for what they reveal about the bargain Congress was striking. The Act removed the limitations the original statute had placed on exclusive licenses, including the provision that had barred exclusive licensing by contractors not themselves engaged in manufacturing or holding substantial interests in manufacturing organizations, and it replaced the royalty-use rules for facility-operating contractors with a formula directing royalties first to research, development, and education consistent with the facility’s mission. The through line of the 1984 package was consistent: widen the circle of performers who could commercialize federally funded inventions, and simultaneously tighten the statutory machinery that kept those performers accountable.

The policy declaration and the exceptional-circumstances valve

Section 200 of title 35 declares the policy and objective of Congress in a single long sentence, and it is unusual among federal statutes for wearing its intentions so openly. It states the policy and objective to use the patent system to promote the utilization of inventions arising from federally supported research or development; to encourage maximum participation of small business firms in federally supported research and development efforts; to promote collaboration between commercial concerns and nonprofit organizations, including universities; to ensure that inventions made by nonprofit organizations and small business firms are used in a manner to promote free competition and enterprise without unduly encumbering future research and discovery; to promote the commercialization and public availability of inventions made in the United States by United States industry and labor; to ensure that the government obtains sufficient rights in federally supported inventions to meet the needs of the government and protect the public against nonuse or unreasonable use of inventions; and to minimize the costs of administering policies in this area. Courts and agencies have treated these seven objectives as the interpretive lens for everything that follows in the chapter, the yardstick against which each discretionary decision under the statute is measured.

Read as a set, the objectives embody two pulls that the statute never fully resolves. The first several objectives push toward decentralization: inventions should be utilized, commercialized, and made publicly available, and the engine for that work should be the performers themselves, small firms collaborating with universities and commercial concerns, rather than the government. The later objectives push back toward public protection: inventions must be used in a manner that promotes free competition without unduly encumbering future research, and the government must obtain sufficient rights to meet its own needs and to protect the public against nonuse or unreasonable use. That last phrase, the protection of the public against nonuse or unreasonable use, is the textual home of the public-protection argument throughout Bayh-Dole commentary. March-in rights under section 203 are the enforcement mechanism the phrase points to, and the preference for United States industry in section 204, with its requirement of domestic manufacture for exclusive licensees, is the companion safeguard on the commercialization side. The free-competition objective, meanwhile, anchors the licensing conditions Congress imposed on nonprofit contractors in section 202(c)(7), including the preference for licensing to small business firms. Nothing in the statute commands that the decentralization objectives always win or that the public-protection objectives always win; the declaration is a balance, and each application of the statute is an occasion for striking it. That balancing function is not decorative. When an agency prepares an exceptional-circumstances determination, the written analysis required by section 202(b)(1) must articulate, in terms of these objectives, why withholding the election right better promotes the chapter’s policies than granting it. When an agency weighs a march-in petition, the decision is tested against the same list, most directly against the objective of protecting the public against nonuse or unreasonable use. And when a funding agency shapes the licensing conditions it negotiates under section 202(c)(7), the objective of promoting free competition and enterprise without unduly encumbering future research supplies the standard. The objectives thus do double duty: they are the substantive criteria the agency must satisfy before it may depart from the default of contractor ownership, and they are the vocabulary in which that departure must be justified for the record. The exceptional-circumstances determination in section 202(a)(ii) is the statute’s front-end safety valve for that balance. The clause permits a funding agency to restrict or eliminate the contractor’s right to elect title to a subject invention where the agency determines that such restriction or elimination will better promote the policy and objectives of the chapter. The mechanism operates at the contracting stage, before any invention exists and before any rights vest: the agency examines the contemplated funding agreement, measures the standard election right against the section 200 objectives, and concludes that withholding the right would serve the chapter’s policies better than granting it. It is therefore the mirror image of march-in. March-in is the back-end valve, invoked after title has vested and after a problem has emerged, with the agency forcing a license to a responsible applicant because the titleholder has not achieved practical application, has left health or safety needs unsatisfied, has failed to meet public-use requirements, or has breached the domestic-manufacturing commitment. The exceptional-circumstances determination is the front-end valve, invoked before any of that can happen, to keep the election right itself out of a funding agreement where the agency believes the chapter’s objectives demand it.

Congress wrapped that front-end valve in procedure rather than leaving it to agency discretion. As amended in 1984, section 202(b)(1) requires the agency to make a formal determination that one of the subsection (a) conditions exists, and except for the intelligence and counterintelligence exception in clause (iii), to file a copy of that determination with the Secretary of Commerce within thirty days after the award of the funding agreement. For a determination under the exceptional-circumstances clause, the filing must include an analysis justifying the determination. The implementing regulations at 37 C.F.R. 401.3 carry this forward: before utilizing any of the exceptions, the agency must prepare a written determination including a statement of facts supporting it. The Secretary of Commerce has an oversight function that the 1984 amendments created and that the original Act lacked. If the Secretary believes that any individual determination, or any pattern of determinations, is contrary to the policies and objectives of the chapter or otherwise not in conformance with it, the Secretary is to advise the head of the agency concerned and the Administrator of the Office of Federal Procurement Policy and recommend corrective actions. The Office of Federal Procurement Policy Administrator, for his part, may issue regulations describing classes of situations in which agencies may not exercise the authorities of clauses (i) or (ii) when agencies are found to be using them contrary to the chapter’s policies and objectives. And since 1984, a contractor that believes a determination is contrary to the policies and objectives of the chapter or constitutes an abuse of discretion may appeal under section 202(b)(4). The legislative history anticipated that the exceptional-circumstances exception would be used sparingly and on a contract-by-contract basis, not as a class device, and the procedural architecture was built to enforce that expectation.

In practice, the agency most closely associated with actual use of the mechanism is the Department of Energy. The Department has issued a published series of Determinations of Exceptional Circumstances covering fields where it has concluded that the standard Bayh-Dole framework is insufficient for the national interest at stake. Among the fields covered in the Department’s published determinations are uranium enrichment technology, storage and disposal of civilian high-level nuclear waste and spent fuel, and national security technologies classified or sensitive under section 148 of the Atomic Energy Act, as well as program-specific determinations for initiatives including the Steel and Metals Initiative and the Solid State Lighting program. The pattern is consistent with the valve’s design: the determinations cluster in areas implicating national security, the nuclear fuel cycle, and domestic industrial policy, the fields where the Department concluded that contractor ownership under the default rules would not better promote the chapter’s objectives. The statute’s other front-end valves are more narrowly drawn and standing rather than case-by-case: clause (i) permits withholding where the contractor is foreign-located or foreign-controlled, clause (iii) where a federal intelligence or counterintelligence authority finds withholding necessary to protect the security of its activities, and clause (iv) for Department of Energy government-owned, contractor-operated facilities dedicated to naval nuclear propulsion or weapons-related programs, the narrowed successor to the original Act’s broad government-facility clause. The exceptional-circumstances clause is the only one that asks the agency itself to weigh the chapter’s objectives and justify the answer in writing, and the Commerce Department oversight machinery is the only external check on whether the agency’s weighing was honest. The valve has never been a routine tool; its significance lies less in frequency than in what its existence says about the statute’s design, which pairs a strong presumption in favor of contractor ownership with a procedural conscience that can be consulted when the objectives point the other way.

The 1986 companion: the Federal Technology Transfer Act

Bayh-Dole addressed only one half of the federal research enterprise. Its election right belonged to the contractor, the extramural performer, the university or small firm that performed federally funded research and development off campus. The laboratories operated by the government itself, where federal employees conducted research inside the walls of federal facilities, were governed by a separate statute: the Stevenson-Wydler Technology Innovation Act of 1980, Public Law 96-480. Stevenson-Wydler was the intramural counterpart to Bayh-Dole, and its ambitions were modest. It required federal laboratories to establish Offices of Research and Technology Applications to identify technologies with potential for use beyond the laboratory, and it directed that laboratories with annual budgets exceeding twenty million dollars staff those offices with at least one full time employee. It articulated a duty to make technology available to industry and to state and local governments, but it gave the laboratories little commercial machinery with which to do so.

The Federal Technology Transfer Act of 1986, Public Law 99-502, signed by President Ronald Reagan on October 20, 1986, supplied that machinery. It amended Stevenson-Wydler and did four things of consequence. First, it authorized federal laboratories to enter cooperative research and development agreements, the CRADAs, with universities, private industry, state and local governments, and other parties. The statute permitted the director of a government operated federal laboratory to negotiate these agreements on behalf of the agency, with the express purpose of decentralizing and expediting the technology transfer process. Second, it authorized laboratories to negotiate licenses for patented inventions made at the laboratory, including exclusive licenses. The collaborating party was guaranteed, at a minimum, an option to an exclusive license in a prenegotiated field of use for inventions made in whole or in part by a federal laboratory employee under a CRADA, while the government retained a nonexclusive, nontransferable, irrevocable, paid-up license to practice the invention or have it practiced on the government’s behalf. Third, it created royalty sharing with laboratory inventors. Agencies were required to share royalties with their inventors: the regulations were to guarantee a fixed minimum payment to each inventor in each year that the agency received royalties from that inventor’s invention, to provide a percentage royalty share above a threshold amount, and to ensure that total payments to all inventors exceeded fifteen percent of total agency royalties in any given fiscal year. The balance of royalties was transferred by the agency to its government operated laboratories, with the majority share going to the laboratory where the invention occurred. Fourth, it established the Federal Laboratory Consortium for Technology Transfer, to coordinate and assist the transfer work of the federal laboratories.

The CRADA differed from a Bayh-Dole funding agreement in a fundamental respect. Under Bayh-Dole, the government is a funder: it pays a performer to do research, and the statute allocates the resulting patent rights between them. Under a CRADA, the federal laboratory is a research partner. The statute specified that the laboratory could accept funds, personnel, services, and property from the collaborating party and could provide personnel, services, and property to the collaborating party, while the government was expressly prohibited from providing direct funding to the industrial partner and the work performed under the agreement had to be consistent with the laboratory’s mission. The CRADA was defined in the legislative history as a specific legal document, not a procurement contract, developed at the laboratory level with limited agency review. In the original 1986 act, the CRADA authority applied to government owned, government operated laboratories; it was extended to government owned, contractor operated laboratories, principally the Department of Energy laboratories, by the fiscal year 1990 Defense Authorization Act, Public Law 101-189.

The 1980 and 1986 statutes together completed the technology transfer architecture that Bayh-Dole had begun. Bayh-Dole governed the extramural relationship: the government pays, the performer may elect title, and the conditions of section 202(c) follow the title. Stevenson-Wydler as amended by the 1986 act governed the intramural relationship: the government performs the research in its own laboratories, and the statute builds the commercial bridge through CRADAs, laboratory licensing, inventor incentives, and the Laboratory Consortium. Both statutes imposed the same policy DNA on both halves of the system: preferences for small businesses and for domestic manufacture, royalty sharing with inventors, and a requirement that commercialization serve the public purpose that justified the federal expenditure. The pair reflected a single legislative conclusion, reached across two Congresses, that federally funded research was a public asset whose value was realized only when it moved, by deliberate institutional machinery, from the laboratory into practical use.

March-in rights

The name is the first misleading thing about the provision. “March-in” suggests federal officials striding in and seizing a patent. The statute authorizes nothing of the sort. Under 35 U.S.C. section 203, what the government can do is compel licensing. Title to the invention stays with the contractor or its assignee. The remedy is an additional license, on terms reasonable under the circumstances, granted to a party the law calls a responsible applicant. The government becomes a broker of access, not an owner of the invention. Every argument about the provision ever since has been colored by that gap between the dramatic name and the modest remedy.

The statutory grant is framed as a right held by the federal agency under whose funding agreement the subject invention was made. Where a small business firm or nonprofit organization has taken title to a subject invention, the agency may require the contractor, an assignee, or an exclusive licensee to grant a nonexclusive, partially exclusive, or exclusive license in any field of use to a responsible applicant or applicants, upon terms that are reasonable under the circumstances. If the contractor, assignee, or exclusive licensee refuses the request, the agency may grant the license itself. The entire mechanism turns, however, on four conditions, and the agency may act only if it determines that action is necessary on one of them. Stated in the statute’s own terms, they are these.

First, the contractor or assignee has not taken, or is not expected to take within a reasonable time, effective steps to achieve practical application of the subject invention in the field of use at issue. Second, action is necessary to alleviate health or safety needs which are not reasonably satisfied by the contractor, assignee, or their licensees. Third, action is necessary to meet requirements for public use specified by federal regulations, and those requirements are not reasonably satisfied by the contractor, assignee, or licensees. Fourth, action is necessary because the agreement required by section 204 has not been obtained or waived, or because a licensee of the exclusive right to use or sell any subject invention in the United States is in breach of its agreement obtained pursuant to section 204. That fourth ground, tied to the section 204 requirement that exclusive licensees manufacture substantially in the United States, is part of the statutory text and completes the set.

The first condition depends on a defined term, and the definition carries much of the provision’s weight. Section 201(f) defines “practical application” as manufacture, in the case of a composition or product; practice, in the case of a process or method; or operation, in the case of a machine or system; and, in each case, under conditions establishing that the invention is being utilized and that its benefits are, to the extent permitted by law or government regulations, available to the public on reasonable terms. A product sitting on a laboratory shelf, however promising, is not practically applied. A product on the market may still fail the test if its benefits are not reaching the public on reasonable terms. That phrase, “reasonable terms,” would become the hinge of every later dispute about whether price belongs in the analysis at all.

The implementing procedure sits at 37 CFR 401.6. It is a deliberately slow instrument. A march-in proceeding begins with a written notice from the agency to the contractor and any known assignee or exclusive licensee, stating that the agency is considering the exercise of march-in rights, giving the reasons in terms sufficient to put the contractor on notice of the facts, and specifying the field or fields of use in which licensing is being considered. Within thirty days of receiving that notice, the contractor may submit information or argument in opposition, in person, in writing, or through a representative, including anything raising a genuine dispute over the material facts on which the proposed march-in is based. If such a dispute is raised, the head of the agency or a designee undertakes or refers the matter for fact-finding, conducted in accordance with procedures the agency establishes. After fact-finding, the contractor and agency representatives receive thirty days to submit written arguments to the head of the agency or designee, and oral argument is held before that decision-maker if the contractor requests it. The head of the agency or designee then bases the determination on the facts found, the administrative record, and any written or oral arguments, taking into account the consistency of exercising march-in rights with the policy and objectives of section 200. Written notice of the determination goes out by certified or registered mail within ninety days after fact-finding ends or after oral arguments, whichever is later; if it does not, the proceeding is deemed terminated, and no march-in based on those facts and reasons may be exercised afterward. An agency may also terminate a proceeding at any time if it decides it does not wish to exercise the rights.

The contractor’s defenses are layered. Beyond the administrative process, section 203(b) provides that a contractor, inventor, assignee, or exclusive licensee adversely affected by a determination may, within sixty days after it issues, file a petition in the United States Court of Federal Claims, which decides the appeal on the record and may affirm, reverse, remand, or modify the agency’s determination as appropriate. In cases brought under the first and third statutory grounds, the agency’s determination is held in abeyance pending the exhaustion of those appeals. An administrative appeals procedure is also to be established by regulation. The structure tells its own story: Congress built march-in as an extraordinary remedy surrounded by procedural armor, available in principle but difficult to deploy in practice.

And the practice bears that out. Despite repeated petitions across four decades, no agency has ever exercised march-in authority. The procedure has been invoked, considered, and denied. It has never once been completed in the affirmative. That unbroken record is the fact that drives every modern argument about the statute, because it forces the same question from opposite directions: is the lever unused because it was built to be nearly unusable, or because agencies have simply refused to pull it.

Four petitions, four denials

The first petition came from a competitor seeking to turn the provision against a patent holder in active litigation. In 1997, CellPro, Inc. asked the National Institutes of Health to initiate march-in proceedings against patents owned by Johns Hopkins University and licensed to Baxter Healthcare Corporation. The patents covered a stem cell separation technology, based on work by a Johns Hopkins researcher under NIH grants, that CellPro’s own Ceprate SC Stem Cell Concentration System had been found to infringe. A federal district court in Delaware had held the infringement willful, and CellPro faced an injunction on sales of its device once Baxter’s competing Isolex system cleared the Food and Drug Administration. CellPro argued that march-in was warranted both because Johns Hopkins and Baxter had not taken effective steps to achieve practical application and because government action was necessary to alleviate health or safety needs, given that CellPro had an FDA approved product on the market and patients needed continued access during bone marrow transplantation. The petition had originally been sent to Health and Human Services Secretary Donna Shalala, but march-in authority was delegated to the head of the funding agency. On August 1, 1997, NIH Director Harold Varmus denied the petition. His reasoning, set out in an NIH statement, was that patient care needs would be met so long as one or the other cell separation device remained available under the terms of the court’s order, that NIH would continue to monitor the situation until a comparable alternative reached the market, and that the underlying patent dispute was for the companies to resolve between themselves.

The second petition, and the first to make drug pricing the explicit target, concerned the HIV drug Norvir. Abbott Laboratories had developed ritonavir with partial federal support, roughly three and a half million dollars in NIH funds against more than three hundred million Abbott spent bringing the drug to market. In December 2003, Abbott raised the price of Norvir by about four hundred percent for United States patients, from a single dose price of one dollar and seventy one cents to eight dollars and fifty seven cents, while holding the price steady elsewhere. The increase fell hardest on the drug’s widespread off label use at low doses to boost the effectiveness of other protease inhibitors, and it provoked protests at Abbott’s headquarters, a physician boycott of the company’s products, and antitrust lawsuits. Essential Inventions, the nonprofit founded by James Love, petitioned NIH to march in, arguing that the price made the drug unavailable on reasonable terms in violation of the statute. NIH held a public hearing on May 25, 2004, at which Senator Birch Bayh, the law’s coauthor, told the agency that the clear intent of the law was to bring inventions to market and that march-in was meant to ensure health and safety needs were met, not to set prices. On July 29, 2004, NIH Director Elias Zerhouni denied the petition, finding that Norvir had been publicly available to patients for eight years and that there was no evidence march-in could alleviate health or safety needs not already reasonably satisfied by Abbott. Zerhouni wrote that the agency agreed with public testimony that the extraordinary remedy of march-in was not an appropriate means of controlling prices, that the issue of drug pricing was appropriately left for Congress to address legislatively, and that questions of anticompetitive behavior belonged to the Federal Trade Commission.

The third petition arose from scarcity rather than price, though price hovered at the edges. In August 2010, three people with Fabry disease petitioned NIH to march in on patents covering Fabrazyme, the enzyme replacement therapy agalsidase beta. The patents were owned by the Mount Sinai School of Medicine in New York, based on NIH funded inventions, and exclusively licensed to Genzyme. Viral contamination at Genzyme’s Allston, Massachusetts plant in 2009 had shut down production, and for more than a year the only approved Fabry treatment in the United States had been rationed, with patients receiving reduced doses or skipping infusions. The petitioners invoked the health and safety ground, arguing that a government compelled license would let other manufacturers supply the drug. On December 1, 2010, NIH Director Francis Collins denied the request. The agency’s determination held that a march-in proceeding would not increase the supply of Fabrazyme in the short term, since the complex biologic could not be brought to market by a new manufacturer quickly enough to relieve a temporary shortage, and that Genzyme expected to restore normal dosing during the first half of 2011. NIH asked Genzyme for monthly updates on the supply situation and continued to monitor it, effectively choosing surveillance over compulsion.

The fourth petition returned to price, and to a drug at the center of the modern debate. On January 14, 2016, Knowledge Ecology International and the Union for Affordable Cancer Treatment asked NIH and the Department of Defense to exercise march-in rights, or the government’s royalty free license, on patents covering Xtandi, the prostate cancer drug enzalutamide. The drug had been developed at the University of California, Los Angeles with NIH and Defense Department grants, licensed to Medivation, and marketed in the United States by Astellas Pharma. The petitioners cited a United States list price of about one hundred twenty nine thousand dollars per year, roughly eighty eight dollars and forty eight cents per pill, two to four times the price in other high income countries, and pointed to an offer from the Canadian manufacturer Biolyse Pharma to supply a generic version to federal programs for three dollars per pill. On June 20, 2016, Director Francis Collins denied the petition in a brief letter, finding no basis to authorize generic production because there was no shortage of Xtandi. Knowledge Ecology International responded that the letter had not engaged with its evidence on international price disparities, restricted formularies, or the burden of copayments on patients, arguing that charging United States residents multiples of the foreign price was inconsistent with making the product available to the public on reasonable terms.

The record from 1997 to 2016 is unbroken: four petitions, four denials, zero exercises of the authority. Three competing explanations circulate for why the lever has never moved. The first is that the conditions are narrow by design, drafted so that ordinary commercial behavior, even aggressive pricing or a temporary shortage, falls outside them. The second is that agencies have read the conditions narrowly as a matter of institutional choice, with three successive NIH directors reaching the same result across the Clinton, Bush, and Obama administrations. The third is that the political cost of a first use has been prohibitive, since marching in on a patent would send a signal through the entire technology transfer system that Congress and the executive branch have been unwilling to send. James Love, whose organizations filed two of the four petitions, has argued that the failure to grant a single request has told patent holders the agency will permit almost anything. Defenders of the denials, including Joseph Allen of the National Technology Transfer Center, have countered that the statute is a balancing of interests and that disturbing the balance would put the whole technology transfer enterprise out of kilter. The dispute is inseparable from the broader architecture of pharmaceutical patent policy under the Hatch-Waxman framework, where drug patents enter the story through an entirely separate statutory scheme. Readers working through the petitions in sequence may find it useful to track the statutory conditions against each denial in a legislation study notebook, which keeps the four grounds and their dispositions side by side for comparison.

Was march-in meant to be a price control?

The most heated question about section 203 has nothing to do with what the statute says on its face and everything to do with what one phrase inside it might mean. The practical application definition requires that an invention’s benefits be available to the public on reasonable terms. The health and safety ground asks whether needs are reasonably satisfied. Petitioners in the Norvir and Xtandi cases built their entire arguments on those words. If a drug developed with taxpayer money sells in the United States for two to four times its price abroad, is it available on reasonable terms. If patients face copayments that limit access or formularies that restrict it, are health needs reasonably satisfied. Knowledge Ecology International put the point directly in the Xtandi matter, writing that price discrimination against United States residents was not consistent with the reasonable terms requirement, and that the high price limited patient access, strained health budgets, and imposed unreasonable cost sharing. The Norvir petitioners had made the parallel claim a dozen years earlier, arguing that Abbott’s four hundred percent increase violated the statute’s availability condition. On this reading, price is not an alien concept imported into the law. It is already inside it, sitting in the definition of practical application, waiting to be given content.

The opposing reading starts from the same text and reaches the opposite conclusion. The conditions name practical application, health and safety needs, and public use requirements. None of them names price. A drug that is manufactured, marketed, prescribed, and obtainable, even at a price its critics call excessive, has been practically applied in the ordinary sense of the definition’s first clause. The health and safety ground asks whether needs are reasonably satisfied by the contractor or its licensees, and in the Norvir case NIH found they were, since the drug had been publicly available for eight years with no gap in supply. Agencies have read the conditions this way with striking consistency: Varmus in 1997, Zerhouni in 2004, Collins in 2010 and 2016, across three administrations and both political parties, each concluding that availability meant presence on the market rather than affordability. The law’s coauthor reinforced the reading in person. At the 2004 NIH hearing, Birch Bayh stated that the clear intent of the law was to bring inventions to market and that march-in rights were intended to ensure health and safety needs were met, not to set prices. Duke law professor Arti Rai, commenting on the Norvir decision for Chemical and Engineering News, noted that Bayh’s view, while not legally binding, appeared to weigh heavily in the agency’s determination. The structural argument runs alongside the textual one: Congress built the statute to encourage private firms to invest in commercializing federally funded research, and NIH’s own 2001 report to Congress on its abandoned reasonable pricing clause experience concluded that price conditions discouraged the very commercialization the law was designed to promote. On this view, turning march-in into a price control would invert the incentive structure the 1980 Act created.

Both readings can claim a lineage in the record, which is what makes the question genuinely unsettled. The petitioners can point to the reasonable terms language and to the policy statement in section 200 that the government should protect the public against nonuse or unreasonable use of inventions, and ask what “unreasonable use” could mean if not, in part, pricing that puts an invention out of reach. The agencies can point to four decades of consistent administrative interpretation, to the coauthor’s disavowal, and to the absence of any court decision saying otherwise. No agency has ever exercised the authority, so no reviewing court has ever been asked to say what “reasonable terms” includes. The dispute is not a matter of one side misunderstanding the statute. It is a real interpretive fork, with text, history, and institutional practice plausibly supporting each branch, and it remains open because the one event that would force a resolution, an actual march-in, has never occurred.

The reasonable pricing clause and its repeal

Nine years after Bayh-Dole became law, the National Institutes of Health tried a price condition of its own, and the experiment ended in retreat. The clause belonged to the intramural world. The Federal Technology Transfer Act of 1986 had authorized federal laboratories to enter cooperative research and development agreements with private firms, under which companies could contribute funds, personnel, services, and equipment while government scientists contributed their expertise, and collaborators could receive exclusive license options on resulting inventions. The CRADA was Congress’s answer to the complaint that federal laboratories produced discoveries that industry never developed. NIH, as the lead technology transfer office for the Public Health Service, built an active CRADA program around it. In 1989 the Public Health Service Patent Policy Board, acting through NIH, adopted a policy statement and model provisions for the model CRADA and the model exclusive license agreement, requiring what came to be called the reasonable pricing clause. The vehicle was not university licensing under Bayh-Dole but the intramural side of NIH: inventions made by federal scientists and licensed to industry, and collaborations under the CRADAs.

The trigger was the price of AZT. When the drug was launched in 1987 at eight thousand to ten thousand dollars per patient per year, the price drew sharp congressional and public criticism, and the involvement of National Cancer Institute scientists in the compound’s development created a public expectation that the medication would reach patients at a more accessible price. No CRADA or exclusive license governed AZT itself, but the Patent Policy Board responded by writing pricing clauses into the model agreements going forward.

The clause required that there be a reasonable relationship between the pricing of a licensed product, the public investment in the product, and the health and safety needs of the public, supported by reasonable evidence. It applied in two circumstances: when new intellectual property arising under a CRADA was then licensed exclusively to the collaborator, and when NIH developed new intellectual property through its intramural research and licensed it exclusively to a company for commercialization. The model CRADA and the model exclusive license each carried similar provisions. The clause was carefully hedged. The model exclusive license stated that it would not restrict the licensee’s right to price a product so as to obtain reasonable profit, and that it did not permit the Public Health Service or any other government agency to set or dictate prices. No law or regulation required the clause; the Department of Health and Human Services was the only federal agency to use it, apart from the Bureau of Mines.

Industry treated the clause as a warning sign. Companies grew reluctant to sign new CRADAs while continuing to collaborate informally with NIH scientists, a pattern that complicated the resulting ownership of intellectual property. A Brookings Institution analysis later summarized the episode as NIH abandoning its 1989 through 1995 fair pricing effort because drug companies refused to sign the new agreements while still working with NIH scientists outside them. The Pharmaceutical Research and Manufacturers of America and the Biotechnology Industry Organization went on record as strongly opposing the model pricing clauses. Numerous NIH intramural scientists reported difficulty entering into CRADAs at all, and some companies became reluctant to supply research materials or share expertise even outside the context of an agreement. NIH’s own review found that the vast majority of CRADAs produced new scientific knowledge rather than new products. Between fiscal years 1987 and 1994 the agency executed 237 CRADAs, and no product developed under a CRADA had ever been brought to market under an exclusive license. Meanwhile the review pointed to competition as the better safeguard: two AIDS therapeutics co-developed by NIH competed with AZT in the marketplace, and a cancer drug analog was being developed to compete with Taxol.

In 1994 NIH held two public meetings on the clause, on July 21 and September 8, drawing patient advocacy groups, industry, academia, and the scientific community. The clause had already come under congressional scrutiny in 1992 during the development of Taxol, the cancer drug developed with National Cancer Institute involvement. Attendees generally supported the purpose of the clause but few believed it functioned as intended, and many said it was driving industry away from beneficial collaborations without a compensating public benefit. The agency’s advisory committees, including the NIH Directors’ Advisory Committee and the PHS Technology Transfer Policy Board, concluded that the clauses should not impair NIH’s ability to conduct collaborative research and that the agency lacked the legislative mandate and the expertise to regulate prices, a role they said would conflict with its mission of transferring technology to the private sector.

In April 1995 NIH Director Harold Varmus removed the reasonable pricing clause from the model CRADA and the model exclusive license agreement. An extensive review, he said, had shown that the clause drove industry away from potentially beneficial scientific collaborations without providing an offsetting benefit to the public, and eliminating it would promote research that could enhance the health of the American people. Varmus noted that no law or regulation required or expressly authorized the clause, and that apart from the Bureau of Mines no other federal agency used one. The review’s reasoning was blunt: the clause addressed the rare breakthrough product at the expense of a more open research environment, and one had to have a product to price before one could worry about how to price it. NIH’s primary mission, Varmus said, was biomedical research, not product pricing.

The repeal drew fire in Congress. Representative Ron Wyden, Representative Bernie Sanders, and Senator Paul Wellstone pushed legislation to reinstate a pricing requirement, but those initiatives largely failed. In the conference report for H.R. 4577 dated December 15, 2000, Congress instead requested that NIH prepare, by July 2001, a plan to ensure that taxpayers’ interests were protected, based on a list of FDA-approved therapeutic drugs with annual United States sales above five hundred million dollars that had received NIH funding. NIH issued the plan. A summary by the State Science and Technology Institute reported that the agency found only four of forty-seven such drugs had been partly developed with NIH-funded technologies, and that the plan called for little or no modification of the agency’s existing efforts.

The episode matters because it is the closest the federal government ever came to imposing a price condition on the commercialization of federally funded inventions, and the government retreated on the evidence. The reasonable pricing clause was milder than the march-in petitions that price advocates later filed: it asked only for a reasonable relationship supported by evidence, explicitly disclaimed any power to set prices, and still it suppressed the collaborations it touched. The experiment ran six years, from 1989 to 1995, and ended without a single product covered by the clause reaching the market under its terms. That history is why the clause keeps surfacing in the march-in price debate. When advocates argue that Bayh-Dole’s march-in rights should be used to police drug prices, the pricing clause stands as the executive branch’s own experiment with that logic, conducted over six years and abandoned by the very agency that imposed it.

Stanford v. Roche and the ownership question

Thirty years after the Bayh-Dole Act took effect, the Supreme Court was asked to answer the question the statute had quietly carried since 1980: who owns a federally funded invention first, the inventor who conceived it or the institution that received the grant? The answer arrived on June 6, 2011, in Board of Trustees of the Leland Stanford Junior University v. Roche Molecular Systems, Inc., 563 U.S. 776, decided 7 to 2. The majority held that the Act does not displace the longstanding norm of patent law that an inventor owns his invention at the moment of conception, and that a university’s rights under the statute depend entirely on whether the inventor has validly assigned those rights to it. The decision turned on two pieces of paper signed by one researcher, and its consequences reached into the employment agreements of nearly every research university in the country.

The researcher was Mark Holodniy, a research fellow at Stanford. He undertook to develop an improved method for quantifying HIV levels in patient blood samples, using the polymerase chain reaction technique. Because Holodniy was largely unfamiliar with PCR, his supervisor arranged for him to conduct research at Cetus, a company where PCR techniques had matured. When Holodniy joined the Stanford research program in 1988, he signed a Copyright and Patent Agreement, the CPA, in which he “agree[d] to assign or confirm in writing to Stanford and/or Sponsors that right, title and interest in” such inventions as required by contracts or grants. The following year, as a condition of gaining access to Cetus, he signed a Visitor’s Confidentiality Agreement, the VCA. That agreement stated that Holodniy “will assign and do[es] hereby assign” to Cetus his “right, title and interest in each of the ideas, inventions and improvements” made “as a consequence of [his] access” to Cetus.

The operative difference between the two documents was a matter of tense. For the next nine months, Holodniy conducted research at Cetus. Working with Cetus employees, he devised a PCR-based procedure for calculating the amount of HIV in a patient’s blood, a technique that allowed doctors to determine whether a patient was benefiting from HIV therapy. Holodniy then returned to Stanford, where he and other university employees tested the HIV measurement technique. Over the next several years, Stanford obtained written assignments of rights from the Stanford employees involved in refining the technique, including Holodniy, and filed several patent applications related to the procedure. Stanford ultimately secured three patents: U.S. Patent Nos. 5,968,730, 6,503,705, and 7,129,041. Some of the research had been supported by a National Institutes of Health grant, which subjected Stanford to the Bayh-Dole Act. Stanford disclosed the invention to the government, formally elected to retain title, and conferred on the government a nonexclusive, nontransferable, paid-up license. Meanwhile, in 1991, Roche Molecular Systems acquired Cetus’s PCR-related assets, including all rights Cetus had obtained through agreements like the VCA signed by Holodniy. After clinical trials, Roche commercialized the HIV quantification procedure, and its test kits were used in hospitals and AIDS clinics worldwide.

Stanford sued Roche for patent infringement. Roche’s defense was that it could not be sued for infringing patents it partly owned, by virtue of the VCA. Stanford countered that any assignment by Holodniy to Cetus was inoperative because the Bayh-Dole Act automatically allocated all rights in the inventions to Stanford, the federally funded contractor. The district court ruled for Stanford, but the Federal Circuit reversed: it held that Cetus, and therefore Roche, had acquired rights in the patents, and that Stanford lacked standing to sue without joining all co-owners. The appeals court’s reasoning on the two agreements was decisive. The CPA’s “agree to assign” language, it held, was merely a promise to assign rights at some undetermined future point. The VCA’s “will assign and do[es] hereby assign” language was a present assignment of Holodniy’s future inventions, giving Cetus immediate rights that vested no later than the filing of the common parent application. By the time Holodniy executed a later assignment to Stanford, his rights had already transferred to Cetus, and the subsequent assignment was void.

The Supreme Court granted certiorari on a single question: whether the Bayh-Dole Act reorders the normal priority of rights in an invention conceived or first reduced to practice with federal support, placing ownership first with the contractor rather than with the inventor. Stanford argued that it did. A majority of the Court rejected the argument. The opinion’s logic unfolded in three moves. First, the Court grounded itself in the Patent Act’s background rule: although much in intellectual property law had changed since the first Patent Act, the basic principle that an inventor owns the rights to his invention at the moment of invention had not. The Bayh-Dole Act contained no language displacing that norm, and Congress does not displace foundational background principles of law by implication. Second, the Court examined the Act’s text and found that it presupposes, rather than creates, contractor ownership. The statute defines a “subject invention” as an invention “of the contractor” conceived or first actually reduced to practice under a funding agreement, a phrase that assumes the invention already belongs to the contractor, which is to say that the inventor has assigned it. Likewise, section 202(a) gives the contractor the right to “elect to retain title,” and retention is a word for keeping what one already has, not for receiving what the statute hands over. Third, the Court observed that the Act imposes its requirements on contractors, not on the contractor’s employee inventors. The disclosure, election, filing, reporting, and utilization obligations all attach to the party to the funding agreement. The election right is meaningful only if there is an owner in a position to elect. Put together, the holding was compact: the Act “does not displace the basic principle that an inventor owns the rights to his invention,” and “[o]nly when an invention belongs to the contractor does the Bayh-Dole Act come into play.”

Justices Breyer and Ginsburg dissented. Their position was that the majority’s reading defeated the statute’s central purpose. The Act had been enacted because federally funded inventions were languishing uncommercialized, and its mechanism, the dissent argued, was to place title reliably in the hands of the contractor, the party positioned to develop and license the invention. On this view, Stanford’s reading was the one that made the election scheme function: if an inventor could assign his rights to a third party before the contractor’s rights attached, the entire Bayh-Dole framework could be defeated by the happenstance of an employee’s side agreement. Breyer’s dissent warned that the majority’s rule turned ownership into what one commentator quoted him as calling “a technical drafting trap for the unwary,” where the fate of a university’s patent portfolio depended on whether its forms used promise language or present-tense assignment language. The dissent treated the question as practical rather than textual: the universities needed to be able to rely on a stable chain of title, and Congress, in the dissent’s account, had meant to give it to them.

The practical consequence was immediate and national. The holding changed university paperwork across the country. The Federal Circuit’s distinction between “agree to assign” and “do hereby assign” had already existed in case law, but the Supreme Court’s decision made it the governing architecture for every technology transfer office in America. Universities that had relied for decades on employment agreements containing mere promises to assign discovered that those agreements could be outranked by any later present-tense assignment an inventor signed elsewhere, even one signed casually as a condition of visiting another company’s laboratory. Counsel nationwide rewrote researcher employment agreements to use present assignment language, commonly in the dual formulation “will assign and do hereby assign,” so that rights transferred at the moment of conception rather than at the moment of a later written instrument. Policies governing outside consulting, visiting arrangements, and confidentiality agreements were tightened, and many institutions began requiring legal review of third-party agreements before researchers signed them. The lesson for institutions was not that the Act was weak but that it was conditional: it protects the contractor’s election only after the contractor has secured what the statute assumes, an assignment from the inventor. The broader pattern of how university research practices adapted to the Act’s machinery is traced in the university research companion piece.

The unused lever

The Bayh-Dole Act answered a problem measured in warehoused patents. Before 1980, the federal government took title to inventions arising from the research it funded and then attempted to license them itself. The record of that system was the one statistic that carried the bill: roughly 28,000 patents sat in government control, and fewer than 5 percent of them were ever licensed or put to use. In biotechnology, the field where federally funded basic research mattered most, the sponsors’ accounting held that not a single new drug had ever been commercialized while the government owned the patent. The government’s own patent portfolio functioned as a warehouse, not a marketplace, and critics of the old system argued that no company would invest in developing a drug, a diagnostic, or a device that the government might license to a competitor the next year.

The statute’s response was a bargain rather than a giveaway. Universities and small businesses, the parties to federal funding agreements, received the right to elect title to the inventions their researchers produced. The election came with obligations: disclose the invention to the agency within two months, elect title within two years, file patent applications before statutory bars, report utilization, include the government-support statement in the patent, prefer American manufacturing, and accept that the government kept a worldwide, nonexclusive, paid-up license to practice the invention. Behind those obligations stood the lever. If the contractor failed to take effective steps to achieve practical application, or if health or safety needs went unsatisfied, the funding agency could march in and require the contractor to grant licenses to responsible applicants, on terms reasonable under the circumstances. Congress was not handing federally funded inventions away unconditionally; it was entrusting them to the parties best positioned to develop them, while keeping a retained interest and an intervention power in reserve.

That is the answer the statute gives to the single question this article set out to address: who should control a federally funded invention, and on what terms? The control rests with the contractor that elects title, the terms are the obligations of sections 200 through 212, and the government’s backstop is a license it holds in every subject invention plus the march-in authority it has never exercised. The unused lever remains the provision that generates nearly all of the public argument about the Act. Petitioners have asked the National Institutes of Health to march in over drug pricing, over supply shortages, and over access, and the agency has declined every request, most visibly in the CellPro petition of 1997, the Norvir petition of 2004, the Fabrazyme petition of 2010, and the Xtandi petition denied in June 2016. The declinations followed a pattern: where the invention was manufactured, practiced, and available to the public, the agency found practical application achieved, and it treated price alone as insufficient to trigger intervention.

Three explanations compete for why the lever has stayed in its holster, and the honest account includes all three. The statutory conditions are narrow, written to catch nonuse, not pricing. The agency has read them narrowly, measuring “reasonable terms” by availability rather than by affordability. And the political cost of the first march-in has always been high, since the government that uses the lever for the first time will be testing not only the statute but the willingness of industry to accept federal funding with strings attached. Whether those readings are faithful or timid is the dispute that never ends, because every controversy over a publicly funded medicine runs through the same unanswered question. The statute has worked as its drafters intended, generating thousands of licenses and a university technology transfer industry that did not exist before 1980, while its most famous provision has never once been used. That is the profile this law presents: a rarely used provision generating most of the public argument, a lever kept visible precisely so that it never has to be pulled, situated within the broader survey of American science policy legislation.

The rights and obligations table

Element	Statutory section	What it provides
Contractor’s election right	35 U.S.C. 202(a)	A nonprofit organization or small business firm that is party to a funding agreement may, within a reasonable time, elect to retain title to any subject invention.
Agency title-taking circumstances	35 U.S.C. 202(a)	The funding agreement may provide for agency title instead of contractor title where the contractor is foreign-located or foreign-controlled, in exceptional circumstances, for intelligence or counterintelligence activities, or where the agreement covers operation of a government-owned research or production facility; the 1984 amendments narrowed the facility clause to Department of Energy naval nuclear propulsion and weapons programs.
Disclosure and election timeframes	35 U.S.C. 202(b); 37 CFR 401.14(c)	The contractor must disclose the subject invention to the agency within two months after the inventor discloses it to the contractor, elect title within two years of disclosure, and file a patent application within one year of election or at least 60 days before the end of any statutory bar period; missed deadlines can lead the agency to request conveyance of title.
Patent filing and utilization reporting	35 U.S.C. 202(c)(5); 37 CFR 401.14(f)	The contractor must file patent applications before statutory bars, continue prosecution or protection of elected inventions, and report periodically on utilization or efforts at obtaining utilization of licensed or assigned subject inventions.
Government-support statement	35 U.S.C. 202(c)(6)	Every patent or patent application for a subject invention must contain a statement that the invention was made with government support and that the government has certain rights in the invention.
Government’s retained licence	35 U.S.C. 202(c)(4)	Each funding agreement must reserve to the government a nonexclusive, nontransferable, irrevocable, paid-up license to practice the subject invention, or have it practiced, throughout the world.
March-in conditions and procedure	35 U.S.C. 203; 37 CFR 401.6	The agency may require licensing, or grant licenses itself, when the contractor has not achieved practical application, health or safety needs are unmet, federal public-use requirements are unmet, or the U.S. manufacturing preference is unmet, following written notice, contractor response, factfinding, and a determination by the agency head.
Manufacturing requirement and waiver	35 U.S.C. 204	Exclusive rights to use or sell a subject invention in the United States may be granted only to licensees who agree that products will be manufactured substantially in the United States; the agency may waive the requirement where the contractor made reasonable but unsuccessful efforts to find such a licensee or where domestic manufacture is not commercially feasible.
Inventor ownership under Stanford v. Roche	563 U.S. 776 (2011)	The Act does not displace the norm that an inventor owns his invention; the contractor’s election right presupposes an owner to elect, so the contractor’s rights under the Act depend on a valid assignment of the inventor’s rights, and present-tense assignment language controls priority among competing claims.

Frequently Asked Questions

Q: What did the Bayh-Dole Act change about federal patents?

Before the Act, inventions arising from federally funded research generally became the property of the government, which then attempted to license them to industry. The system performed badly: tens of thousands of patents accumulated in federal hands, and only a small fraction were ever licensed or commercialized, because companies were reluctant to invest in developing products the government could license to competitors. The Bayh-Dole Act, signed December 12, 1980, reversed the default. Nonprofit organizations and small business firms that were parties to federal funding agreements gained the right to elect title to inventions arising from that research, subject to disclosure, reporting, and licensing obligations. The government retained a worldwide, nonexclusive, paid-up license and gained march-in authority as a backstop. The change converted the federal patent portfolio from a warehouse into a pipeline.

Q: Who owns a patent from federally funded research under Bayh-Dole?

Ownership follows a two-step rule clarified by the Supreme Court in Stanford v. Roche (2011). First, the inventor owns the invention at the moment of conception, because the Act does not displace the background principle that an inventor owns his invention. Second, the funded institution owns it only if the inventor has validly assigned those rights to the institution. Once the institution holds title by assignment, the Bayh-Dole framework applies: the institution may elect to retain title under 35 U.S.C. 202(a), subject to the statute’s disclosure, filing, reporting, and licensing obligations, and the government holds its retained license under 35 U.S.C. 202(c)(4). Stanford’s loss in that case illustrated the sequence precisely. The university held no title to convey because the inventor’s present-tense assignment to Cetus had transferred his rights first, so the Act’s election machinery never came into play.

Q: What are march-in rights under Bayh-Dole?

March-in rights are the federal funding agency’s authority, under 35 U.S.C. 203, to require the contractor, or its assignee or exclusive licensee, to grant licenses in a subject invention to responsible applicants, or to grant licenses itself, when any of four conditions is met: the contractor has not taken effective steps to achieve practical application of the invention; action is necessary to alleviate health or safety needs that are not reasonably satisfied; action is necessary to meet requirements for public use specified by federal regulation; or the U.S. manufacturing preference of 35 U.S.C. 204 has not been met. The implementing procedure appears at 37 CFR 401.6. The provision was designed as Congress’s assurance that the new contractor ownership would not leave publicly funded inventions unused or inaccessible. In the decades after 1980, no federal agency exercised the authority.

Q: Has the government ever used Bayh-Dole march-in rights?

No federal agency has ever exercised march-in rights. The National Institutes of Health considered petitions in 1997 (CellPro, concerning a stem cell separation device licensed by Johns Hopkins to Baxter), in 2004 (Norvir, concerning the price of Abbott’s ritonavir), in 2010 (Fabrazyme, concerning a supply shortage of Genzyme’s Fabry disease therapy), and in 2016 (Xtandi, concerning the price of Astellas’s prostate cancer drug), and declined each one. The pattern of the declinations was consistent: where the product was manufactured, practiced, and available to the public, the agency found the statutory standard of practical application satisfied, and it treated high price alone as insufficient to trigger the health or safety condition. The unbroken record of nonuse is what makes march-in the most discussed and least exercised authority in American innovation law.

Q: What did Stanford v. Roche decide about Bayh-Dole?

The Supreme Court held, 7 to 2, that the Bayh-Dole Act does not automatically vest title to a subject invention in the university contractor. The majority reasoned that the Act presupposes rather than creates contractor ownership: a “subject invention” is defined as an invention “of the contractor,” and section 202(a) gives the contractor the right to “elect to retain title,” language that assumes an owner already in a position to elect. The Act imposes its obligations on the party to the funding agreement, not on the party’s employee inventors. Because Stanford researcher Mark Holodniy had signed a present-tense “will assign and do[es] hereby assign” agreement with Cetus, his rights transferred to Cetus before his later assignment to Stanford, which was therefore void. Stanford never held title, so it had no standing to sue Roche for infringement. The dissent argued that Congress had meant the Act to place title reliably in the contractor.

Q: Why did Bayh-Dole pass in a lame duck session?

The bill reached the floor in the closing weeks of the 96th Congress, after the 1980 election that swept Ronald Reagan into the presidency and ended the Democratic majorities in their existing form. President Jimmy Carter signed the bill on December 12, 1980, during the lame duck session. The timing worked in the bill’s favor: the measure had bipartisan sponsorship from Indiana Democrat Birch Bayh and Kansas Republican Bob Dole, drawing co-sponsors from across the ideological spectrum, and the outgoing Congress was willing to act on a consensus innovation measure before the change of administrations. The December passage meant the law took effect under Carter while its implementation, including the 1983 Reagan memorandum extending its principles to large businesses, unfolded under his successor. The lame duck setting gave the Act a bipartisan birth that neither party later disowned.

Q: Does Bayh-Dole require US manufacturing?

The Act contains a U.S. industry preference, at 35 U.S.C. 204, rather than an absolute manufacturing mandate. Neither the contractor, nor any assignee, nor any licensee may grant to any person the exclusive right to use or sell a subject invention in the United States unless that person agrees that any products embodying the invention, or produced through its use, will be manufactured substantially in the United States. The funding agency may waive the requirement in individual cases if the contractor or assignee shows that it made reasonable but unsuccessful efforts to grant licenses on similar terms to potential licensees willing to manufacture substantially in the United States, or if domestic manufacture is not commercially feasible under the circumstances. Failure to meet the preference is also one of the four statutory triggers for march-in under 35 U.S.C. 203.

Q: What is the government license retained under Bayh-Dole?

Under 35 U.S.C. 202(c)(4), every funding agreement must reserve to the United States a nonexclusive, nontransferable, irrevocable, paid-up license to practice the subject invention, or have it practiced for or on its behalf, throughout the world. The license attaches automatically when the contractor elects title; it does not depend on any march-in proceeding or finding of nonuse. Its practical meaning is that the federal government, and any contractor acting on its behalf, may use the patented invention without paying royalties and without facing an infringement claim from the patent owner. The license is nontransferable, so the government cannot assign it to a private party as a substitute for march-in, and it is paid-up, so no compensation flows from the government to the patent holder. It is the quieter half of the government’s retained interest, overshadowed in public debate by march-in but present in every elected subject invention.

Q: What is a subject invention under Bayh-Dole?

A subject invention is defined at 35 U.S.C. 201(e) as any invention of the contractor that is conceived or first actually reduced to practice in the performance of work under a funding agreement. Three elements matter. The invention must be the contractor’s, meaning it belongs to the party to the funding agreement, a point the Supreme Court underscored in Stanford v. Roche when it held that the definition presupposes a valid assignment from the inventor. It must have been conceived, the formation of a definite idea of the complete invention in the inventor’s mind, or first actually reduced to practice, the physical construction and testing of the invention, during the funded work. And that work must be performed under a funding agreement, which covers contracts, grants, and cooperative agreements entered into between a federal agency and any person for the performance of experimental, developmental, or research work funded in whole or in part by the federal government.

Q: What does practical application mean under Bayh-Dole?

Practical application is defined at 35 U.S.C. 201(f): to manufacture, in the case of a composition or product; to practice, in the case of a process or method; or to operate, in the case of a machine or system, under conditions that establish the invention is being utilized and that its benefits are, to the extent permitted by law or government regulations, available to the public on reasonable terms. The definition does the heaviest lifting in march-in disputes, because the first statutory trigger for march-in is the contractor’s failure to take effective steps to achieve practical application. In the petition record, the NIH treated manufacture, practice, and operation, combined with availability to the public, as sufficient evidence of practical application, and it declined to read the “reasonable terms” phrase as a price control. That reading is why products on the market have consistently survived march-in petitions.

Q: How does Bayh-Dole differ from the Stevenson-Wydler Act?

The two 1980 laws divide federal research into two domains. The Bayh-Dole Act governs extramural research: inventions made by outside parties, universities, small businesses, and later large businesses, under funding agreements with federal agencies. The Stevenson-Wydler Technology Innovation Act of 1980 governs intramural research: technology developed inside federal laboratories, and it created Offices of Research and Technology Applications at those laboratories to move laboratory technology toward private use. Where Bayh-Dole’s mechanism is contractor election of title, Stevenson-Wydler’s mechanism is cooperative research and development agreements, authorized by a 1986 amendment, that let federal laboratories share personnel, equipment, and research results with private partners. The two statutes are companion pieces, but they answer different questions: Bayh-Dole asks who should own an invention made with federal money by an outside researcher, while Stevenson-Wydler asks how federal laboratories should transfer the technology their own scientists produce.

Q: Can large businesses use Bayh-Dole, or only small businesses and nonprofits?

As enacted in 1980, the Act’s election right ran to small business firms and nonprofit organizations, reflecting the congressional judgment that universities were the trusted vehicle for moving federally funded inventions into use. In February 1983, President Reagan issued a memorandum on government patent policy extending the principles of Bayh-Dole to contractors of all sizes. Congress codified the change in the 1984 amendments, which made the election provisions available to large businesses as well. The extension did not alter the underlying bargain: every contractor, regardless of size, takes the disclosure, election, filing, reporting, government-license, manufacturing-preference, and march-in obligations with the title. The policy rationale was that excluding large firms left some of the most capable developers, particularly in pharmaceuticals and electronics, outside the commercialization framework the Act had created.

Q: What happens if a university misses the disclosure deadline?

Under 37 CFR 401.14(d), the funding agency may request conveyance of title to the subject invention when the contractor fails to disclose the invention or elect title within the times specified in the patent rights clause. In practice, agencies have applied the rule with restraint. The NIH and other agencies routinely grant extensions of the disclosure and election periods, and retroactive approvals of late disclosures were commonplace in the early decades of the Act’s administration, since the agency’s interest lies in an orderly patent process rather than in seizing title from the institutions it funds. The consequence is not automatic forfeiture: the agency must request conveyance, and the contractor has procedural protections. The provision functions mainly as leverage, reminding contractors that the election right is conditioned on timely disclosure, and it is one of the four circumstances in which the government can take title outright rather than merely licensing the invention.

Q: What is the two-month disclosure rule?

The patent rights clause at 37 CFR 401.14(c) requires the contractor to disclose each subject invention to the federal agency within two months after the inventor discloses it in writing to the contractor’s patent personnel. The clock starts when the inventor reports the invention internally, so the rule has two practical parts: the contractor must maintain internal procedures that get invention disclosures from researchers to the technology transfer office promptly, and the office must then report them to the funding agency within the two-month window. Election of title must follow within two years of the disclosure to the agency, subject to extension, and patent applications must be filed within one year of election or at least 60 days before the end of any statutory bar period. The two-month rule is the opening obligation in the chain that conditions the contractor’s ownership.

Q: Can the government give a federally funded invention to someone else?

The government cannot simply reassign a subject invention to a party of its choosing, but it has two related powers that approach that result. Under the retained license of 35 U.S.C. 202(c)(4), the government may practice the invention itself or have it practiced on its behalf anywhere in the world, without royalties, which in practice lets federal agencies procure the invention’s benefits from whomever they select as a government contractor. Under the march-in authority of 35 U.S.C. 203, the funding agency may require the patent owner to grant licenses to responsible applicants on reasonable terms, or grant such licenses itself, but only after the procedures of 37 CFR 401.6 are completed and only when one of the four statutory conditions is found. Both powers are constrained: the license is nontransferable, and march-in requires notice, a contractor response, factfinding, and a determination by the agency head, with judicial review available.

Q: What was the CellPro march-in petition about?

In 1997, CellPro, Inc. petitioned the NIH to march in against patents covering a stem cell separation device. The patents had arisen from research at Johns Hopkins University and were licensed to Baxter Healthcare Corporation. CellPro argued that Baxter had failed to take effective steps to achieve practical application of the invention, and it filed the petition after a federal district court had found that CellPro’s competing device willfully infringed the licensed patents. The NIH declined to initiate march-in proceedings, finding that Baxter met the statutory and regulatory standard for practical application through its manufacture, practice, and operation of the invention and the invention’s availability to and use by the public. The decision was the first significant test of the march-in mechanism and established the template the agency followed in later cases: demonstrated commercialization and public availability satisfy the statute, even when a competitor stands ready to license.

Q: What was the Fabrazyme march-in petition about?

In August 2010, three patients with Fabry disease petitioned the NIH to march in against patents covering Fabrazyme, Genzyme’s enzyme replacement therapy. The patents were owned by the Mount Sinai School of Medicine and arose from NIH-funded research, and they were exclusively licensed to Genzyme. The petitioners argued that a viral contamination at Genzyme’s Allston, Massachusetts, plant had curtailed supplies for more than a year, forcing patients onto reduced doses of the only approved treatment for their disorder. In December 2010, NIH Director Francis Collins denied the request, determining that a march-in proceeding would not increase the supply of Fabrazyme in the short term, since bringing another manufacturer of a complex biologic through regulatory approval would take too long to relieve the shortage. The agency requested monthly supply updates from Genzyme instead. The petitioners appealed, but the decision stood.

Q: What happened with the Xtandi march-in petition?

In January 2016, Knowledge Ecology International and the Union for Affordable Cancer Treatment petitioned the NIH to march in against the patents covering Xtandi, a prostate cancer drug developed at the University of California, Los Angeles, with NIH and Department of Defense funding, and marketed by Astellas Pharma. The petitioners argued that the drug’s United States price, more than $129,000 per year at the time, was two to four times the price in other high-income countries, and that such pricing meant the invention was not available to the public on reasonable terms. In June 2016, NIH Director Francis Collins denied the petition, finding no evidence that Xtandi was or would be in short supply and concluding that the drug’s widespread availability as a prescription medicine satisfied the practical application standard. The denial reaffirmed the agency’s longstanding position that price, without a supply failure, does not trigger march-in.

Q: How did universities change their paperwork after Stanford v. Roche?

After the 2011 decision, university counsel across the country audited the assignment language in researcher employment agreements and found that many, like Stanford’s Copyright and Patent Agreement, contained only a promise to assign future inventions rather than a present assignment. Because the Federal Circuit had held that “agree to assign” transfers nothing until a later written instrument is executed, while “will assign and do hereby assign” transfers rights at the moment of conception, any researcher who signed a present-tense agreement with an outside company could convey title out from under the university. Institutions rewrote their forms to use present assignment language, frequently the dual formulation “will assign and do hereby assign,” and extended the review to consulting agreements, visiting-scientist agreements, and confidentiality agreements that researchers signed with industry collaborators. The paperwork change was the decision’s most tangible legacy: the Act’s election machinery was only as strong as the assignment that fed it.

Q: What does reasonable terms mean in the march-in statute?

The phrase appears in the definition of practical application at 35 U.S.C. 201(f): an invention reaches practical application when it is manufactured, practiced, or operated under conditions establishing that it is being utilized and that its benefits are available to the public on reasonable terms. The statute does not define reasonable terms further, and the meaning has been fixed largely by agency practice rather than by court decision. In the march-in petitions decided through 2016, the NIH treated the phrase as an availability standard rather than a price standard: a product manufactured and sold to the public through normal commercial channels met the test, and the agency declined to treat a high list price, even a price far above foreign prices, as making the invention unavailable on reasonable terms. Petitioners in the Xtandi case argued the opposite reading, that price discrimination against American patients violates the phrase. No agency has adopted that reading.