Stanford v. Roche: the U.S. Bayh-Dole Act and Inventors’ Rights

Professor Sean O’Connor is a Research Affiliate to IP Osgoode, Director of the Law, Technology & Arts Group, Faculty Director of the Entrepreneurial Law Clinic, and Professor of Law at the University of Washington School of Law in Seattle.


Traditionally, debates over the ownership and use of government funded inventions revolve around the interests of: the institutions that receive funding, the government, and the public. Of late, however, attention is being paid to the interests of inventors themselves. What rights do they have to compensation for their inventions, or to take the inventions with them if they leave the institution, or to assign or license the inventions to third parties? The latest installment of this fast growing area of interest is the case of Board of Trustees of Leland Stanford Junior University v. Roche Molecular Systems, Inc., for which the U.S. Supreme Court recently granted certiorari. In this blog post, I will discuss how a key assumption from the early days of U.S. Federal R&D patent policy led to a vexing conundrum for contemporary technology transfer law and policy that must now be resolved by the Supreme Court.

As U.S. Federal R&D spending increased dramatically in World War II, President Roosevelt charged the Attorney General to investigate the patent practices and policies of the various Federal agencies and their contractors. A report was released in 1947 and is mainly known for its recommendation of a “title policy” in which Federal agencies would be required to obtain title to inventions arising under government funding. This would apply equally to intramural research conducted by Federal employees and extramural research conducted by outside contractors. But the Attorney General also investigated the extent to which contractors were themselves securing title from their employees. Because Federal R&D contracts were generally between Federal agencies and institutions or firms, if those organizations did not secure title from their employees, then the organization would not be able to assign title to the government. The results showed that most private companies and nonprofit research institutions routinely secured IP assignments from inventive employees, but that many universities did not, especially in the case of faculty members. Notwithstanding this, the Report concluded that contractor employees were almost always obligated to assign their inventions to their employers. While this may have been true in the aggregate, it obscured the uncertain nature of university IP assignments (perhaps excusably because most practical development contracts were awarded to firms or research institutes). This then became the fundamental—but flawed—assumption that allowed the Attorney General expressly to limit his recommendations to allocating rights and obligations as between agencies and contractors only. Inventors’ rights and obligations would be left to contractors and their employees/independent contractors.

It was not until 1963, however, that any president acted on the Report’s recommendations with regard to extramural research (President Truman established formal policies with regard to intramural research patent ownership in 1950). But, President Kennedy’s “uniform” patent policy memo to departments and agencies was hindered by a number of ad hoc statutes Congress had passed allocating IP title in the case of particular technology fields—such as atomic energy—or with regard to specific agencies—including NASA. The Kennedy policy largely tracked the Report’s recommendations, albeit with a more nuanced title policy that allowed, on a case-by-case basis, for contractors to retain title or exclusive rights when needed to incentivize practical application of the research results. Following the Report, the Kennedy policy also restricted itself to allocation principles as between agencies and contractors (other than where dictated by the ad hoc statutes), and left inventors’ rights to the contractors. The policy did not affect contractor rights by operation of law, but rather had to be used to secure contractual rights by agencies in executing contractor funding agreements. It thus placed these issues squarely within the Federal procurement system.

Nixon largely reaffirmed these policies in 1972 and ordered that formal regulations be promulgated. Meanwhile, the National Institutes of Health and the National Science Foundation had already developed the institutional patent agreement (IPA) system that provided master agreements between those agencies and certain of their contractors to eliminate case-by-case title determinations. The IPAs established a default rule of title remaining with the contractor, with a license, march-in rights, and certain other protective rights granted to the government (all as originating in the Attorney General’s Report and adopted by both Kennedy and Nixon). However, the IPAs expressly required the contractor institution to secure patent title rights from employee inventors—perhaps because of an increased sense that the Attorney General’s assumption may indeed have been flawed. The government-wide regulations that issued also included such an express requirement.

When Bayh-Dole was passed in 1980, it likewise focused primarily on the allocation of title and rights as between Federal agencies and contractors (flipping the default title policy from the government to the contractor). Contractor employee inventors were addressed in two ways, both derivative of contractors’ rights. First, contractors must share revenues with employee inventors. Second, if contractors do not elect to take title to an invention, then the inventors may petition the funding agency to grant them title (rather than having it transfer to the agency). Bayh-Dole’s implementing regulations largely tracked the existing Nixon policy regulations, albeit with the default title rule change. This included the requirement that contractors must secure the government’s rights under Bayh-Dole by appropriate patent agreements with employee inventors.

Despite all this, a belief arose among many university officials that Bayh-Dole provided some sort of backstop title allocation mechanism as between a university and its employees, including faculty. The requirement to secure patent assignment agreements apparently came to be seen as one only to provide a formal path for perfecting assigned titles with the U.S. PTO. Thus, a failure to secure the assignment contractually from the inventor increasingly became viewed as a small matter—the inventor would “have” to assign the patent anyway under Bayh-Dole’s statutory provisions. Accordingly, many universities adopted a policy of having inventive employees sign future contingent assignment agreements based around the phrase “agree to assign,” rather than the immediate present assignment agreements favored by industry, which use the phrase “hereby assign.”

This has come to a head now in Stanford v. Roche. Stanford included the “agree to assign” language in its Copyright and Patent Agreement (“CPA”) that researcher Mark Holodniy signed in 1988. The following year, with Stanford’s knowledge (and apparently encouragement), he engaged in research at Cetus, a Stanford research partner whose PCR business and assets were later purchased by Roche. Holodniy signed Cetus’ Visitor Confidentiality Agreement (“VCA”), which used the present conveyance “hereby assign” language to effect the transfer title to all of Holodniy’s present and future inventions arising “as a consequence” of his work at Cetus. When the inventions were later developed, Holodniy and his co-inventors disclosed them to Stanford, which in turn filed a patent application resulting in the three patents-in-suit. Because Federal funding had been used in the Stanford lab, Bayh-Dole was deemed to apply and the university added in the required Government rights language to each patent. Along the way, Stanford duly required the inventors to assign the inventions, which each did. The issued patents list Stanford as the assignee.

When Roche began selling HIV assay kits based on the inventions, Stanford disclosed the patents-in-suit to it and offered a license. Base on Holodniy’s VCA and the fact that the techniques he learned and applied in developing the assay came from Cetus, Roche claimed that it had co-ownership, or at least a shop rights license (Stanford held joint title at least through the other Stanford researcher inventors), of the patents-in-suit and refused to take the license. Stanford sued and the trial court found, in part, that Holodniy could not have transferred any title to Cetus/Roche, because title had transferred to Stanford automatically, or by operation of law, under Bayh-Dole once the inventions arose. Roche appealed and prevailed, with the Federal Circuit ruling that the Bayh-Dole Act only regulates the legal relationships between Federal agencies and their contractors, and not as between inventors and the Federal contractor. The Federal Circuit held that because the Cetus VCA executed an immediate assignment of future inventions, it vested equitable title in Cetus before any title vested in Stanford. While Holodniy had signed Stanford’s CPA before signing the VCA, it only imposed a contractual obligation on him to assign inventions when they arose and at Stanford’s request. Thus, Holodniy essentially breached the CPA when he signed the VCA, because he assigned away the then future inventions and would not be able to fulfill his obligation to Stanford when the inventions later arose.

Stanford has presented the issue to the Supreme Court from a reverse perspective through a single question for certiorari:

Whether a federal contractor university’s statutory right under the Bayh-Dole Act, 35 U.S.C. §§ 200-212, in inventions arising from federally funded research can be terminated unilaterally by an individual inventor through a separate agreement purporting to assign the inventor’s rights to a third party.

But this question presumes that the “statutory right” under Bayh-Dole allows contractors to take title from any inventor who happened to develop an invention under the contractor’s Federal funding, or, alternately, that the Act automatically transfers such title by operation of law. From this presumption, Stanford asks whether the inventor can “terminate unilaterally” Stanford’s “statutory right” to the inventions. Bayh-Dole does not appear to work this way, however, and thus Stanford never acquired Holodniy’s title and nothing is being “terminated.”

It is true that Bayh-Dole envisions a system in which title allocation decisions are first made as between the contractor and the funding agency, with the inventor only being able to retain title afterwards in certain circumstances. And this system is jeopardized if inventors can simply assign their inventions to third parties before title vests in either the contractor or the government. But the system relies on contractors to secure title rights from inventors in the first place. At the dawn of Federal patent policies for extramural inventions, the exhaustive Attorney General’s Report concluded that contractors were already doing this and thus the government did not need to directly address that part of the equation. This allowed the Attorney General and subsequent presidential administrations from having to worry about constitutional and practical implementation questions with regard to a wide and complex field of contractor-inventor relationships (including students, staff, faculty, independent researchers, etc. at universities). Even amidst the extensive debate over Bayh-Dole and similar bills in Congress during the later 1970s and 1980, little to no mention was made of whether or how such a massive “automatic” title transfer from inventors to contractors would be authorized. By contrast, the extensive debates and detailed provisions of Bayh-Dole and its implementing regulations focus on the minutiae of how title and rights are to be allocated between contractors and the funding agencies. Bayh-Dole only provides two minimum requirements for how contractors should treat inventors. Its implementing regulations clearly require contractors to secure title from inventors by written agreement.

Congress likely has the power to pass a law effecting the automatic transfer of patent title from one party to another for all Federally funded inventions. The issue is that Congress does not appear to have done so in Bayh-Dole. The problem is that the practice of whether and how contractors secure title from inventors does not comply with the practice as found to generally exist by the Attorney General in 1947. This makes Bayh-Dole a potentially flawed law in its current state. It is beyond the scope of this post to outline how the law could be changed. The Supreme Court will soon let us know whether Bayh-Dole does in fact effect a massive, nearly unbounded transfer of title from inventors—who are essentially only third party beneficiaries of Federal funding agreements and thus not directly in privity with the government—to contractors. Universities complain that if it does not rule this way, then they will not be able to ensure that they can secure title to Federally funded inventions. But this is a disturbing claim as it means that they must not be able to ensure this for any other of their inventions either. Their licensees and assignees will surely be unhappy to hear this. At the same time, it simply seems to be an overblown claim. Universities are perfectly well able to secure the proper written agreements to protect their, and the government’s, rights, and they must do so under Bayh-Dole.

More details and analysis of this topic are available in both a report I co-authored for the U.S. National Academy of Science, Legal Context of University Intellectual Property and Technology Transfer, and Supreme Court amicus brief I co-authored for the American IP Law Association (AIPLA).

  1. It will be interesting to find out to what extent the Supreme Court’s decision in the Stanford v. Roche patent litigation resolves some of these longstanding questions regarding ownership of IP rights by university employees and affiliates.

Comments are closed.