In 2001, Eva Pell, Penn State’s Vice President for Research prepared a report for the Trustees on “technology transfer.” In the discussion about why universities should be involved in technology transfer, Pell includes the following account of Bayh-Dole:
Until 1980, the federal government owned about 30,000 patents, many of which had resulted directly from research conducted at universities; and only a small percentage ever became commercialized.
The figure cited was 28,000 patents. But most of these patents arose in defense industry contracting, not university contracting. And there, most of these patents were on inventions that the defense contractors could have owned under federal contracting agreements but chose not to. So this 30,000 figure is inflated. And the “many of which” is simply not true. According to claims made at the time, 5% of these federal patents were licensed for commercial use. However, for biotech inventions, the figure was 23%. For university-managed inventions made with federal support, in the period 1968-78, the rate was about 5%. The government’s licensing rate for biotech inventions was over four times better than the university licensing agents’ rate.
As the for the “only a small percentage ever became commercialized”–this is bombast. First, the overall commercial use rate for patented inventions is about 5%. But Pell’s bombast is designed to mislead the trustees (even if Pell merely repeats stuff that she’s been told is true). The point of the federal patent ownership program was not “commercializing” inventions but rather making them available for public use and benefit. Thus, the focus of executive branch patent policy was in using patents to bring inventions “to the point of practical application” and then releasing the invention for all to use. In some federal programs, such as those at the Department of Agriculture, all inventions reviewed by Harbridge House in 1968 (for two study years) achieved practical application and became commercial products. It is no surprise that many defense, space, and atomic energy inventions did not become commercial products. The “market” for these inventions was the federal government. To “commercialize” these inventions is a meaningless claim. Bombast.
On December 12, 1980, the “Patent and Trademark Act Amendments of 1980”, also known as the Bayh-Dole Act, were [sic] enacted to create a uniform patent policy among the many federal agencies that fund research, and to create an environment more likely to foster commercialization of the patents the federal government had seeded through its financial support.
The Kennedy patent policy of 1963, updated by Nixon, was uniform. Unlike Bayh-Dole, however, the Kennedy patent policy was also flexible rather than arbitrary. And from 1968 to 1978, the NIH and NSF operated an Institutional Patent Agreement program under the Kennedy patent policy. The IPA program permitted participating nonprofits to own inventions arising with federal support, subject to various requirements for managing patents on these inventions. Given that the NIH and NSF were the primary federal agencies funding university research–and the DoD allowed contractors to own inventions made in research–universities already had “an environment…likely to foster commercialization.” Bayh-Dole did little to change this environment for universities.
Furthermore, there’s nothing express in Bayh-Dole that focuses on “commercialization”: the standard is “utilization” or “practical application” of inventions made with federal support. While “commercialization” is one way in which utilization or practical application may be achieved, it is not the only way and not the only way recognized by Bayh-Dole. Even in the use reporting provision, the datum to be reported is “first commercial sale or use.”
Finally, there is no evidence to demonstrate that university ownership of inventions made with federal support creates an environment “more favorable” to “commercialization” than federal ownership or the private ownership of inventions. Harbridge House found that the best commercialization rates (practical application, commercial use, or sale) came about when an invention made with federal support was owned by a contractor with experience in the market. The worst rates were those involving contractors without experience (such as nonprofits) and involving licensing rather than ownership. Thus, as far as the Harbridge House study was concerned, Bayh-Dole presented the worst possible combination: ownership by a contractor without experience, who then must grant patent licenses in order for an invention to achieve practical application.
Bayh-Dole enables small businesses and non-profit organizations, including universities, to retain title to materials and products they invent under federal funding.
This, too, is full of misstatement. Bayh-Dole requires federal agencies to use a patent rights clause that permits nonprofits to “retain” title to inventions that the nonprofits have acquired outside the conditions of the federal funding agreement. Bayh-Dole gives universities no special privilege to obtain title to inventions made with federal support. It simply requires federal agencies to use a contracting provision that prevents the agencies from requiring assignment of title by universities if the universities obtain title. “Retain” does not mean “own.” It means “do not have to sign over to the federal government.”
And Bayh-Dole does not deal with “materials and products”–it deals with inventions that are or may be patentable, when owned by a contractor. The phrase “materials and products” makes it appear to the trustees that Bayh-Dole is broader than it is. Non-bombast version: “If a university obtains title to an invention made with federal support, then Bayh-Dole’s standard patent rights clause permits the university to retain title to that invention.”
One more bit of bombast: “they invent.” The antecedent of “they” in Pell’s sentence is “small businesses and non-profit organizations, including universities.” But universities don’t invent. Inventors invent, and federal common law of inventions provides that inventors own their inventions. So Pell’s presentation assumes what isn’t the case–that universities somehow invent and can retain title to these inventions. Penn State cuts inventors entirely out of this discussion of federal patent policy, as if this was the purpose of Bayh-Dole. The Supreme Court ruled that’s not the case.
By placing few restrictions on the universities’ licensing activities, Congress left the success or failure of patent licensing up to the institutions themselves.
This is strange. Congress places the fewest restrictions on inventors, when they retain title. That’s 35 USC 202(d) and 37 CFR 401.9. Next in line is small businesses. That’s 35 USC 202(c)(1)-(6), (8); 37 CFR 401.14(a)(a)-(j), (l). Then we have nonprofits–they have the restrictions at 202(c)(7) and 401.14(a)(k). Those restrictions include prohibiting assignments of inventions except to patent management organizations (and even then the nonprofit obligations must follow the assignment) and restricting the use of licensing income (share with inventors, deduct only expenses for managing federally supported inventions, and use the rest for scientific research or education).
But Bayh-Dole also is part of federal patent law, not federal procurement law. In federal patent law, Bayh-Dole establishes a new category of invention, the “subject” invention, and specifies as patent law a policy that restricts the use of the patent system for subject inventions. Thus, the licensing program of any owner of a patent on a subject invention must conform to the public covenant that runs with any subject invention.
Yet, there’s deeper bombast here. Congress also required contractors to grant to the Government a license to “practice and have practiced” any subject invention–and “practice” means “make, use, or sell.” That is, the public use of any subject invention does not depend only on “universities’ licensing activities” but also on the government’s activities–including “sell” and “have sold.” Furthermore, Congress included “march-in provisions” that aim to limit the nonuse or unreasonable use of inventions by universities and other contractors. Thus, the “success or failure” is not left to “the institutions themselves.” In practice, federal agencies have indeed left universities to their own devices–but Congress did not do so. Deep bombast.
The success of Bayh-Dole in expediting the commercialization of federally funded university patents is reflected in the statistics.
The Kennedy patent policy was focused on “expediting” bringing inventions to the point of practical application. It did this by limiting a contractor’s patent monopoly to three years from the date a patent issued. Any advantage a contractor gained would arise because the contractor got things done within the six years from date of invention. Harbridge House found that for commercial contractors, 77% of inventions made with federal support by commercial contractors were in use by the time the patent issued.
Prior to 1981, fewer than 250 patents were issued to universities per year. By 1993, almost 1,600 were being issued each year. Of those, nearly eighty percent stemmed from federally funded research.
Bombast. Prior to 1981, patents were largely issued to patent management agents, not to universities. Only a handful of universities took ownership of patents directly–including MIT, UC, Stanford. The rest used agents such as Research Corporation or an affiliated university foundation. Even after Bayh-Dole came into effect, there was still a lag of time for research contracts to be awarded, research to be done, inventions to be made, and three years or so later patents issued–so, perhaps four or five years beyond 1981–so, maybe 1985 or 1986 would show the first patents managed under the Bayh-Dole regime. “Commercialization” of these inventions would take even more time–time to find licensees, to negotiate deals, and time for those licensees to create product. The first outcomes of Bayh-Dole licensing might not be meaningfully realized until about 1990.
I ran numbers at the USPTO web site for patents issued to universities, foundations, and institutes. Let’s talk 1981, since that will be before Bayh-Dole has any effect. There were 548 patents issued to universities, foundations, and institutes. Of these, 146 cited government interest. In 1993, there were 2029 patents issued to universities et al. 585 cited government interest. That’s just under 30% of university et al. patents, not nearly 80%. At this point, Pell is just pulling numbers from her posterior cortex. There’s nothing to indicate the government share of university patenting that Pell cites to the trustees. But the trustees would have nothing to go on except trusting Pell to have her numbers right. If she makes things up, how would they know? So she writes things that sound good, and leads on the trustees. Perhaps she didn’t know any better. Perhaps she was given this information by someone she thought knew something about Bayh-Dole. In any case, it’s bombast.
What is Pell’s goal? Here’s what she had to say:
A final question then, is how will the University benefit from a successful Tech Transfer enterprise? If we look at The University of Wisconsin, we can see that their foundation with its $1 billion endowment got its start over 70 years ago and has been fueled by many successes starting with the discovery of lWarfarin [sic] as I have already said, and synthetic vitamin D.
Lots going on here. First, WARF was started to manage the “synthetic vitamin D.” That is, WARF was started with a meaningful invention already in hand, just as Research Corporation was with the electrostatic precipitator. But Penn State apparently starts its research foundation before it has that first meaningful invention. Thus, it has to subsidize the search for a paying invention rather than have a paying invention at the start. This is a fundamental difference in practice. If one has a successful invention, then one has a good chance of creating a helpful patent management practice around it. But it is not the case that if one creates a patent management practice, that practice will somehow speed up the search for a successful invention. It’s a Bugblatter Beast of Traal kind of moment.
Second, Pell makes it sound like WARF had “many successes” to create its “$1 billion endowment.” But that’s not the case. As Blumenthal, Epstein, and Maxwell point out in a 1986 article, only a few biomedical inventions account for WARF’s licensing “successes.” Here’s an excerpt from a history of the University of Wisconsin by Edmund Cronin and John Jenkins:
So, 42 profitable inventions, or about a 17% commercialization rate. Keep in mind generating money from a license does not mean that the licensed invention has been made into a commercial product or has been used commercially–a company can pay for a license and fail to make or use the invention. The important bit is that WARF had three big hit inventions in its first 50 years–about 1% of its patents–and 12 modest successes (including the big hits) or about 5%. The two “biggest winners” account for $19m in revenue.
How then didWARF get to $1b in its “endowment”? Investing in common stocks and emerging growth companies. Here’s what Cronin and Jenkins have to say:
The early royalties for vitamin D synthesis (it wasn’t synthetic vitamin D, it was vitamin D created by UV radiation, to get around federal regulations regarding adding things to milk) provided funds to invest in the stock market, and it was those investments that built WARF’s “endowment.” If Penn State wanted to follow WARF’s path, the university would aim to hit it rich in the stock market. Funny, but Bayh-Dole prevents such an effort for inventions made with federal support. Nonprofits are forbidden from using income from licensing for stock investments–the money has to be used for “scientific research or education.” More Bugblatter Beast thinking, here.
But here’s Pell:
The royalties from those first discoveries, and others since that time have generated the endowment, which is now being used to seed research and fund graduate education. This certainly is our long-term goal as well.
She seems to think that WARF’s endowment is made from royalties, not from stock investments. Thus, she leads the trustees to think that WARF’s “success” (many lucrative licenses, amounting to a billion dollars in 70 years) can be replicated as a “long-term goal.” Here’s what David Blumenthal et al. say about the WARF situation in “Commercializing University Research“:
A close examination of the experience of this foundation reveals that, with good fortune and good management, a patenting and licensing organization can enjoy financial success. It also indicates, however, that the success experienced by WARF would be difficult to achieve today and that such efforts may yield lower financial returns than university administrators expect.
That would be, a 1% big hit licensing rate and 5% any sort of profit from licensing rate would be difficult to repeat now–and that was Blumenthal in 1986. The University of California recently reported that its commercialization rate–anything becoming a commercial product–was 1 in 200, or 0.5%.
Pell argues that there is a “gap” between university inventions and licensing–a version of the “funding gap” or “valley of death” claim to explain why university patent licensing is so horribly ineffective. The patents it turns out, don’t provide sufficient incentive for a single company or investor to step in and “develop” inventions as products. Thus, money from somewhere else must be “invested” in the inventions to put them in a condition to be licensed as a monopoly to support commercial development and sales. This is all very strange, though it is blithely repeated by university licensing folks. It appears that the patents on university inventions preclude further research development rather than provide incentives that speed that development. To mitigate the problem, administrators then create a secondary market of investment on the prospect that eventually some monopolist will take an exclusive license and the secondary investors who provide the “gap funds” will see a return on their investment when the university makes money. In essence, that’s what the paper startup companies created by universities do–create a betting pool for investors hoping that they can push an invention toward commercial viability and make good money doing that.
We might pause to consider the huge difference in approach between the Kennedy patent policy focus on getting an invention “to the point of practical application” and current university licensing practice, enabled by Bayh-Dole. The Kennedy patent policy allowed federal agencies to use patents to give contractors an incentive to develop inventions for public use quickly–within three years of a patent issuing. Once an invention had been developed to the point of practical application, plus three years of patent monopoly, the invention was to be released via non-exclusive licensing. The focus, then, was on using the patent system to speed the development of an invention to this “point of practical application.” By contrast, university “gap” practice aims to develop an invention sufficiently that it can be assigned to a monopolist for the remaining life of the patent, to extract as much financial value from the patent position as possible. Utterly unlike the Kennedy practice, which limited the duration of the patent monopoly as an incentive to get things developed quickly, the Bayh-Dole enabled practice sells off a financial interest in maintaining the patent monopoly for its entire term.
For Penn State, in 2001, patent licensing under Bayh-Dole wasn’t much working. Inventions had to be developed on the university side rather than by a licensee (and certainly not by multiple licensees). Investment funds here allow “investors” to buy in to the future “commercialization” which likely will never happen, given the reported outcomes at places like Stanford and UC.
What Pell leaves to handwaving is how anything the university does by patenting with a fixation on commercialization will benefit the public or industry. The idea of creating an “endowment” like WARF’s is sketchy enough, starting without a big hit, and relying apparently on federally supported inventions and so cannot be used to speculate in the stock market to build the endowment.
What would Pell’s argument to the trustees have been if she had reported Bayh-Dole accurately and had a clue about WARF? Instead of claiming there was a “federal mandate” to “transfer technology,” where would she find that mandate? Instead of claiming that Bayh-Dole had done anything beneficial, it would be just another federal bungle of regulation to be navigated. Instead of claiming WARF had many successes, she would point out that WARF had just a handful (Stanford’s experience is similar). Rather than trying to create the appearance of a system to manufacture licenseable inventions, Pell might have discussed how to plan for finding a “big hit” invention–by directing inventors to invention management resources, by operating in a voluntary invention ownership policy, by being selective in what inventions to take on, and by working to license non-exclusively–all things that characterized the WARF experience. Instead, Pell uses a fantasy and bombast to justify what Penn State is doing.
Perhaps there is some sense in what Penn State was trying in 2001. But what then would be the non-fantasy justification for the Penn State technology transfer program?