A WIPO Economist Gets Bayh-Dole Wrong

Here’s an article by Mario Cervantes, an economist at OECD, “Academic Patenting: How universities and public research organizations are using their intellectual property to boost research and spur innovation start-ups.” Cervantes claims that universities “protecting their inventions” somehow increases their research funding and startups. It is an economic fantasy. Patenting of inventions does little to attract research funding. The only funding that comes in is from exclusive licensees/assignees as part of transfer deals. Since those deals are rare–a few a year per institution–the research funding from royalties or from direct research sponsorship is a drop, other than a maybe once-every-thirty year windfall. But patenting does deter other companies–the ones that don’t get the exclusive license and thus are denied access to the invention from providing funding of any sort.

Patenting of inventions also fragments cumulative technology and research tools across institutions, especially where government funding has spread around research in a given area. Each institution claims the inventions of its researchers, seeks exclusive licensees, and as a result it is virtually impossible to create commons, standards, and interoperable systems. Given that standard terms in university exclusive licenses poison the opportunity to cross-license or grant royalty-free sublicenses, a university exclusive license actively works against widespread access to a university-hosted invention.

Here’s Cervantes on Bayh-Dole:

Indeed, in 1980, the United States passed what is widely considered landmark legislation, the Bayh-Dole Act,

Passive voice. Who does this considering? But more so, it’s fake history. Before Bayh-Dole, the NIH and NSF ran IPA programs, which did pretty much what Bayh-Dole does, and just as badly. While Bayh-Dole is legislative where the Nixon patent policy was merely codified (in 1975 as the Federal Procurement Regulation) to legitimize the IPA programs, it is difficult to see how Bayh-Dole is landmark.

which granted recipients of federal R&D funds the right to patent inventions and license them to firms.

Fake law. Recipients gained (in the absence of federal agency policies that already allowed it, and overriding federal laws that required a federal claim unless an agency determined otherwise) the right to retain inventions made in federally supported work that they had otherwise acquired. The law did not grant recipients the right to patent–recipients gained the right to patent when they acquired inventions. The law instead prevented federal agencies from disturbing the right to patent–provided that the recipients indeed then attempted to patent. So, oddly, Bayh-Dole does mess with a recipient’s right to patent by forcing the recipient to use that right, even if public release of the invention might be somehow advantageous–the recipient does not get to retain the right to release an invention. If a recipient fails to seek a patent–or fails anywhere along the line in patenting and maintaining and defending any issued patent–then the federal agency providing the funding can demand ownership and license any patent exclusively, effectively cutting the recipient off from collaborations with regard to the invention.

The main motivation for this legislation was to facilitate the exploitation of government-funded research results by transferring ownership from the government to universities and other contractors

The Supreme Court in Stanford v Roche made clear that Bayh-Dole does not transfer ownership of any inventions to contractors. Not from the inventors, not from the federal government. So this continues the misrepresentation of the law, even if it is arguable that some people clearly did think that Bayh-Dole was designed to vest invention rights with contractors rather than with inventors. Senator Bayh argued in an amicus brief that inventors were intended to be last in line, after contractors and the government, when it came to ownership of the inventions they made. It’s just that this motivation failed to be implemented in the law that Congress approved.

While the phrase “facilitate the exploitation of government-funded research results” sounds wonderfully bland, it is, word by word, counter to Bayh-Dole. Bayh-Dole states its policy and objectives at 35 USC 200. Its focus is inventions, not results. Inventions made in work arising from federally supported research or development. The government funds some part of research or development work, and any invention made in that work, whether the invention was funded by government funds or not, is within scope. Bayh-Dole’s stated policy includes using the patent system to promote–not facilitate–the utilization–not exploitation–of inventions within scope of the law. To promote the utilization means to make that utilization more likely, accessible, available, attractive. To facilitate exploitation means, in contrast, to make it easier, cheaper, faster to deal in the control of “results.”

Simplify, coarsely. To promote use means to get more and better uses for inventions. To facilitate exploitation means to make it easier for organizations to seek money from patent licenses. Bayh-Dole introduces a working requirement into federal patent law. Its policy to use the patent system to promote utilization of inventions is a restriction on the rights of a patent holder–the patent system cannot be used to suppress utilization of inventions. That policy then acts as a restriction on the remedies available to a patent holder for infringement. One can enforce a patent to prevent a utilizer from taking actions that would suppress broader utilization, but not for utilizing. If you get this point, then you start to understand the law Congress passed and have a chance of seeing through the bogusness promoted by political spinnners and sloppy academics.

who could then license the IP to firms.

Well, it isn’t really “the IP”–Bayh-Dole is limited to its definition of invention–is or may be patentable plus (through a sleight of hand) plant varieties. Using “IP” here makes it sound like Bayh-Dole is comprehensive and takes in all forms of IP and underlying “results.” Bayh-Dole is much narrower in focus, which leads to the obvious question of how these Bayh-Dole inventions play out with regard to data, to works of authorship, to know how, to collected stuff, otherwise unpatentable research tools. Bayh-Dole selects out of all possible research outputs those that might be suppressed by patent and demands that the patent system be used to promote use, not suppress it. How this policy then coordinates with all the other possible research outputs ought to be the focus of discussion. But it is not. Instead, fake political spin is taken as fact and the discussion veers off into a fantasy land.

Although patenting in US universities did occur prior to the passage of Bayh-Dole Act, it was far from systematic.

This, too, is unsupported by history. Mostly, US universities did not patent because they directed invention matters to external agents, and especially to Research Corporation. The agents did the review for patentability, spent the money on patenting, and worked to license inventions to industry. Universities did not have to do patenting work, and for the most part did not want to do patenting work. Three universities especially went against this trend. The University of California in 1960 created its own patent and licensing office (but inventions were reported there voluntarily). Stanford and MIT also created their own licensing offices as well. Stanford’s was mostly voluntary, MIT’s was compulsory. David Mowery (and his coauthors) has shown that Research Corporation was working to systematize the identification and reporting of inventions in the 1970s. The very name “technology transfer office” derives from Research Corporation’s efforts to establish such offices on campuses to facilitate transfer of inventions from inventors to Research Corporation for management. Similarly the NIH and NSF IPA programs systematized nonprofit reporting and patenting of inventions made in work funded by these agencies.

Perhaps Cervantes has a special secret idea of what “systematic” means, or what “far from” means, but there’s a good case that by the 1970s, a number of US universities had built systems for dealing with inventions. They did not do the patenting, but they most certainly had in place the infrastructure to pass inventions to the agents that did do the patenting. It may be then that what Cervantes might have meant, had he given himself the chance, was that before Bayh-Dole, universities relied on external agents to consider patenting, and those agents were much more selective in what they chose to patent than are US university administrators now. The effect of Bayh-Dole–though no where in the law–is that US universities have taken patenting in-house and have become much less selective in what they patent. They patent “just in case” or “to create the possibility that someone will pay” or “because more patents looks good.”

Prior to Bayh-Dole, external agents patented because they believed they could license certain inventions to industry–often non-exclusively, even for biopharma inventions. After Bayh-Dole, universities patented because they hoped they could license inventions exclusively to companies. They built portfolios of patents. They adopted a warped sales funnel rationalization that the more they built the portfolio, the more that some small percentage would make money to pay for the entire operation. Patent 1000 inventions, license 200, maybe one of those inventions hits it big every thirty years and everyone thinks the program is wildly successful.

But this is not Bayh-Dole’s policy, which is specific to each invention claimed, not to the results of a portfolio funnel in which 999 inventions may be suppressed by patenting so that maybe one invention becomes a lucrative license deal. Cervantes ought to deal with the shift from an external agent utilization approach, invention by invention, to an internal office portfolio approach that leads to patent speculation and failing that patent trolling.

At the end of the 1990s, emulating the US policy change, many other OECD countries reformed research funding regulations and/or employment laws to allow research institutions to file, own and license the IP generated with government research funds.

I was involved here and there with countries intent on “emulating” the US. Here, I confess I was not effective, because these countries emulated the fake version of Bayh-Dole, not what the law states and what the Supreme Court has confirmed. So OECD countries created a bureaucratic law of patents on research inventions that places ownership of inventions with “research institutions” rather than with inventors or the professors that lead the research. Not only have the OECD countries got Bayh-Dole all wrong, they also have used their new thinking to disrupt the very different conditions that prevailed in European countries with regard to inventions made in academic work. It may be that handing inventions to academic administrators in preference to individual inventors or faculty research directors is, for these countries, a really good thing, but that’s neither the context nor the outcome that emulates the US.

The right to ownership has now been transferred to the universities while academic inventors are given a share of royalty revenue in exchange.

This is not the US situation. US universities must assert their own claims to inventions without any aid or mandate from federal law. They routinely do so. And they claim it is a good thing for them to do so. They just refuse to provide the evidence to show what has actually taken place. And Bayh-Dole, for inventions made in federally supported work, makes all reports of invention use a government secret.

Cervantes represents the OECD approach as an exchange–universities get ownership and inventors get a share of royalty revenue. There is a huge asymmetry at work. The universities get ownership outright, every time. Inventors get the right to a share of royalties, but that’s often nothing at all. There’s no exchange in practice. Inventors get nothing. A university may fail to license (normal), and may eat up any licensing income with its costs (patenting, litigation). The exchange is more like an eminent domain taking, but with the creative variation that the former owner of the property may never receive compensation at all, let alone reasonable compensation. One can see how such an exchange would shimmer in the minds of bureaucrats. Even a rare upside looks attractive if there’s basically no downside. Inventors can’t negotiate value upfront, can’t hold back inventions, can’t seek competent patent services. Clearly, these OECD laws do “facilitate exploitation” of inventions–but the facilitation is directed at the bureaucracy obtaining those inventions to for portfolio machinations, not at industry using those inventions to advance production and products. You see the advantages of abstractions in economics papers.


This entry was posted in Bayh-Dole and tagged , , , . Bookmark the permalink.