Innovation Policy
The Bayh-Dole Act
One law, passed in 1980, created the modern biotechnology industry, gave Stanford $336 million from PageRank, and transformed American university research into a commercialization engine. Here is how it works.
Before Bayh-Dole: a broken system
Before 1980, when a university made a discovery using federal grant money, the federal government claimed ownership of the resulting patents. This seemed logical — taxpayers funded the research, the public should own the results.
In practice, it was a disaster for commercialization. The government owned thousands of patents but had neither the resources nor the incentive to license them effectively. Agencies had their own incompatible licensing programs. Private companies were reluctant to invest in developing government-owned patents because they couldn't get exclusive licenses — and without exclusivity, there was no incentive to bear the risk of commercialization.
By 1980, the federal government owned approximately 28,000 patents. Fewer than 5% had been licensed to industry. The research sat unutilized. Meanwhile, America's industrial competitiveness was declining relative to Japan and Europe.
What the Act does
The Bayh-Dole Act (Patent and Trademark Law Amendments Act, 1980) allows universities, small businesses, and non-profit organizations that receive federal research funding to elect to retain ownership of inventions made with that funding. In exchange, the institution must: report the invention to the federal agency that funded it; file patent applications to protect it; try to commercialize it; give preference to small businesses in licensing; and manufacture products in the US (where practical).
The federal government retains a 'license' — it can use the invention for its own purposes. It also retains 'march-in rights' allowing it to compel licensing to others in certain public-interest circumstances. But day-to-day commercialization is handled by the university.
Universities are required to share a portion of royalty income with the inventors — at least a percentage specified by the institution's own policy. Most universities share 30–50% of net royalties with inventors, after deducting patent costs and administrative fees.
Tech transfer offices: the mechanism
Every major research university has a technology transfer office (TTO) — sometimes called the Office of Technology Licensing or the Innovation office. Its job is to identify inventions from university researchers, evaluate their commercial potential, file patents, and license the technology to companies.
The TTO process: a faculty member or student files an invention disclosure form. The TTO evaluates whether to pursue patent protection (considering commercial potential, patentability, and cost). If yes, the TTO files a patent application. The TTO then markets the technology to companies — existing companies interested in licensing, or startup spinouts founded by the inventors.
Famous TTOs: Stanford's Office of Technology Licensing licensed PageRank to Google for 1.8 million shares (later sold for ~$336 million). MIT's TTO has generated billions in royalty income. Tech Launch Arizona (University of Arizona) and Skysong Innovations (Arizona State) are active TTOs for the Southwest.
What Bayh-Dole built: the biotech industry
The most significant Bayh-Dole outcome was the biotechnology industry. Before 1980, basic genetic research was academic. After Bayh-Dole, universities could patent recombinant DNA techniques. The Cohen-Boyer patent (US 4,237,224) — filed by Stanford University on a Bayh-Dole-eligible recombinant DNA process — generated $255 million in licensing revenue and enabled Genentech, Amgen, and hundreds of other biotechnology companies.
Every major pharmaceutical company today licenses technology originally developed in university labs under Bayh-Dole. The mRNA vaccine technology behind COVID-19 vaccines traces partly to Bayh-Dole-enabled university research. HIV antiretroviral drugs were developed from NIH-funded research that universities patented and licensed.
An AUTM (Association of University Technology Managers) survey found that Bayh-Dole-enabled university licensing generated over $1.9 trillion in GDP contributions and helped create over 700,000 jobs between 1996 and 2020.
March-in rights: the controversy
March-in rights (35 USC § 203) are the federal government's most powerful Bayh-Dole tool. If an invention made with federal funding is not being commercialized or made available on reasonable terms, the government can require the university or licensee to grant additional licenses — or can grant the license itself.
The drug pricing debate: patient advocacy groups and members of Congress have repeatedly petitioned NIH to exercise march-in rights on pharmaceutical products to mandate generic licensing and lower prices. NIH has consistently declined, ruling that high prices are not sufficient grounds for march-in rights under the statute.
As of 2024, march-in rights have never been successfully exercised in a contested case. The government's consistent position is that Bayh-Dole's purpose is to promote commercialization — and exercising march-in rights would discourage the private investment that commercialization requires.
Licensing from a university: what startups need to know
Many successful startups are founded on university-licensed technology. The typical structure: the founder(s), usually the academic inventors, form a startup and negotiate a license from the university TTO.
License terms vary widely. Typical components: a modest upfront cash fee; equity in the startup (1–5% is common, higher for very early-stage spinouts); development milestones tied to the technology field (IND filing for drugs, first commercial sale for tech); annual minimum royalties; and running royalties of 1–5% on net sales.
The exclusivity question is critical. An exclusive license means the university cannot license the same technology to competitors — giving the startup a monopoly on the technology. A non-exclusive license means anyone can license it, which is generally worth less but comes with lower fees. For biotech therapeutics, exclusive licenses are standard (the capital required to develop a drug demands exclusivity). For general-purpose software tools, non-exclusive is more common.