Universities participating in the NSF’s IPA program operated two technology transfer programs. One program–the one endorsed by the IPA program–focused on patents and licensing. Of the 645 inventions reported by universities (and other nonprofits) made in work receiving NSF support between 1974 and 1978, universities claimed ownership of 223 and obtained 106 patents. Those 106 inventions, then, become the subject of the IPA program’s success metrics. As we have seen, of those 106 patents, 28 were licensed and 13 generated “royalties”–but nothing that approaches royalties from actual sales of commercial product for any significant time or scale. Instead, the royalties appear to be mostly patenting cost reimbursements and license issue fees–things that run up the cost of acquiring rights and have little to do (other than cost) with developing inventions into commercial products.
But that’s only the overt technology transfer program endorsed by the IPA program. The other program, the tacit one, involves the other 422 inventions that were not claimed by universities as inventions–but nonetheless, universities were aware of these inventions (and reported them in detail to the NSF) and personnel at these universities were in possession of the inventions (even if the NSF claimed ownership of them and released the rights to them open access). This tacit program of technology transfer makes use of a wide range of opportunities, including the use of the patent system. It differs from the IPA program in that the institution hosting the research does not demand an ownership position in patents for the purpose of dealing in exclusive licenses for a share of the money that might be made in suppressing all other practice of the invention. An inventor in the tacit system may use patenting resources made available by the host institution, but that is a choice of the inventor, on the inventor’s terms–which is, in practice, a huge difference.
These two programs are in competition. The IPA program aims to suppress inventor use of the patent system (the host institution must have this opportunity), presents inventor publication of inventions as the destruction of valuable public property (the patent right), and argues that inventor consulting to assist others in the practice of an invention denies the host institution its “right” to money from licensing the patent–in which case, the consulting is included at no charge (and therefore for no personal income to the inventor) as part of the licensing transaction. Any income beyond salary for the inventor comes when the license produces payments greater than the costs of obtaining the patent and enforcing the license and patent rights.
The IPA program, then, is the antagonist to the tacit program of technology transfer, denying inventors and research teams the use of open access and the patent system both, while suppressing accounts of how the tacit system has operated. The tacit program then operates covertly, without institutional resources, and often with institutional opposition. One might say that prior to the IPA program–which required institutional ownership and patenting of IPA inventions–institutional patenting resources co-existed with the tacit program of technology transfer. While the federal government might choose as policy the suppression of nonprofit tacit technology transfer programs, it is not clear at all why university administrators chose to align their institutions with such a policy. It does not result in “more” technology transfer, nor “more rapid” technology development, nor greater “commercialization” of inventions, nor more significant “financial return” to universities (a handful of lucrative exclusive patent license deals per decade not withstanding).
Beginning in 1974, the NSF ran an Institutional Patent Agreement (IPA) program until IPAs were shut down in 1978 as ineffective and counter to public policy. Bayh-Dole, one among a number of attempts, replaced the lost IPA programs in 1981.
How effective was the NSF IPA program? Here’s a summary document:
Here’s a table showing patent activity. Of 645 reported inventions, over half are “inactive”–meaning, nothing happened. About a third of the inventions, 223, ended up under the control of contractors.
Now for the licensing metrics for the NSF IPA program:
Of those 223 inventions, 106 received patents (assuming one patent per reported invention)–just over half of the inventions on which patent applications were filed. That means that for just under half of the inventions claimed by contractors, the contractors gave up on the patenting, or all claims were rejected–meaning that the inventions reported weren’t patentable inventions after all. We don’t have here any way of knowing why patent applications did not result in issued patents, but it would sure be helpful to know whether “inventions” were being over-reported or whether contractors (mostly university-affiliated patent development firms) did not have the budget or did not get the broad claims they hoped for).
For 645 inventions marking key results in NSF-funded research over six years, we have 106 patents, or 16%. That means that 84% of inventions made in NSF-supported work became available open access, but as a result of the IPA program, open access was delayed by failed patenting efforts for 97 inventions, or 15%, that were thought to be important enough to patent, but no patent issued. For the two or three years that, on average, patent applications were pending in the PTO, nonprofit contractors–universities and their research foundations–withheld these inventions from open access. Given that the time immediately after an announcement of a new discovery or invention or research finding is when there is an opportunity of great interest, an effect of the IPA program is to delay open access at exactly the time of potential great interest in transferring the technology0–locking out all potential recipients in the search for a single commercialization investor at this crucial time in the life of a new technology. There’s a metric for you: number of failed patent applications for which open access at research announcement was denied as a measure of the delaying (if not destructive) effect of an IPA-like program (such as Bayh-Dole).
Let’s look at the “licensing” part. 28 licenses issued. But we don’t know how many of the patented inventions were licensed. It could have been one invention licensed 28 times or 28 inventions, each licensed once (and perhaps exclusively). Without exclusive licensing, the premise for the IPA programs falls apart, just as it does for Bayh-Dole. Without exclusive licensing, the only issue is that of how to manage open access, and then the only concern surrounds federally owned patents. If a federal agency must grant a license for every citizen’s use of a government-owned invention, what a bother. Perhaps a university could do a better job at that. But if licenses are made “public”–upon notice, then there’s really nothing any contractor could do better than a federal agency with regard to the disposition of rights in an invention. How an invention and its attendant cloud of knowingness gets transferred then is a matter of instruction, availability, and assistance, not *licensing of rights*.
For that–who is in the best position to teach, make data and experimental setup and prototypes available, and to provide assistance to others seeking to practice the invention? Maybe the inventors. Maybe the research team–others, not including the inventors may understand the invention and its uses as well or better than the inventors. Maybe technically minded people at other universities–in which case the initial transfer that matters is from the originating researchers to those better prepared technically minded people at some other place. When a university takes a patent position, this initial transfer is effectively shut down. Those technically minded people elsewhere might learn all they want about the new invention but to teach others without a license would be to contribute to infringement. Their opportunity, then, once they find out about some university invention made elsewhere, is to design around it, or invent applications or variations that block the initial university’s line of development. The result–the university’s patent position itself motivates others to make that patented invention irrelevant or the patent position irrelevant. The university’s action to patent for exclusive licensing itself changes the circumstances under which anyone would adopt the patented invention. The IPA programs–and Bayh-Dole–create a broadbased, active effort to undermine IPA inventions. The big companies often don’t even care about the university patent positions. If the university sues for infringement, so be it. It crushes the university’s goodwill, and it generally forces a non-exclusive license on court-approved terms. So long as the company is not found to be willfully infringing, the settlement is for actual damages. Thus, big companies also have a strong incentive not to read academic publications and rather do their own development. The patent positions universities take, then, create huge disincentives for technology transfer–nothing that renaming of the technology transfer office with something perky like “Technology Ventures” or “Innovation Partners” will ever overcome.
The university’s “effort” to “transfer technology” to “industry” for “commercialization” then cuts off what may be the most crucial bits of transfer opportunity at the outset and creates in their place huge disincentives for others to take up the new technology. Not only the transfer to interested companies (for their own research, for evaluation without any need or plan to make or sell commercial product, or to make such product) but also the transfer to other research programs well positioned for dissemination.
Now look at the results of the NSF IPA program. Of the 28 licenses granted, 13 produce “royalties.” That’s 13 licenses produce payments to universities and their licensing agents. We are to think that there’s one license per patent, given that the 5.8% is 13/223. The licensing rate is 12.6% (28/223), again assuming one license per invention. But the actual rates those on the total reported inventions–645–not just the ones that were appropriated by contractors. There we have a licensing rate of 4.3% (28/645) and a payment rate of 2% (13/645). The “royalties” designation is ambiguous. Any payment in consideration for a patent license is technically a “royalty.” Thus, payments to reimburse a patenting firm for its patent prosecution costs can be booked as “royalties”–as can upfront fees, license issue fees, milestone development fees, and equity or equity cash equivalents–none of which reflects the practice of a licensed invention. In fact, payments other than those based on sales of licensed product reflect a cost to the licensee greater than that of open access–a net drain by the bureaucracy of dealing in patents on the potential for technology transfer.
Of the payments reported in the NSF IPA program, none of them reflect significant “commercialization,” especially if we assume that all the licenses granted were “royalty-bearing.” That’s reasonable. Nonprofit contractors are not likely to spend money on patenting only to give away free access to inventions. Only the federal government would be fool enough to do that (if it is indeed foolish–maybe not, but that’s another article). Look at the average payment per paying license–about $20,700. A patent in the mid 1970s still might run $5,000. So these licenses are barely covering patent costs, and then only for no more than 13 of 203 inventions for which patenting costs were incurred.
Given that the year 1977 reports $0 in “Royalties Produced,” it appears that none of the inventions licensed in previous years was the basis for a product in the marketplace for more than three years. The “Royalties Produced” for 1976–$4,500 for a single license–sure looks like a patent cost reimbursement. Maybe later some of these patent inventions resulted in commercial products. Development times can be long. But the Harbridge House report (1968) found that federally supported inventions, when owned by company contractors, were brought to practical application within three years of patent issue:
About a third of the inventions had been used commercially by the time a patent application was filed, and assuming three years for patent issue, about two-thirds had been used by the time a patent was received. (I-26)
Think about this finding. Companies that invent with federal support were using one third of the inventions before they even filed a patent application. For the NSF IPA program, and for Bayh-Dole, nonprofit owned inventions often aren’t getting used commercially–when at all–until after a patent issues, or at least claims are allowed. Think more. University “technology transfer” runs years behind comparable company uptake of new technology that the company produces. If the public deserves the benefit of the use of inventions made in work with federal funding, the clear policy response is to shift federal research funding from universities to companies. University “technology transfer” programs based on patent rights run six years or more behind an originating company’s own practice. And that’s assuming a university licenses much at all of what it claims for patenting. In the case of the NSF IPA program, six years out, nothing is happening. Licensing activity for the NSF IPA is half of what universities (and their licensing firms) claimed was their licensing rate of 25% to 33%. The income reported amounts to patenting cost reimbursements and upfront fees. If there was ever any commercial product, it had a short, uninteresting run.
Sure, something might have happened after 1978 for some of these 106 patented inventions in the NSF IPA program–but the fundamental point is that whatever it was happened years later. Clearly, nothing about the IPA program matched the rapidity of utilization when a company hosting the invention puts it into practice. And it’s clear that the IPA programs lead universities to cut off access to new technology at a key moment when access may be crucial for transfer to ever happen. Wait six years and the technology has moved on. People have designed around the university’s patent positions. The IPA program does, in practice, the opposite of what it claims is its purpose. So like bureaucrats to then ask what might be done to build on their, er, success.
Bayh-Dole states its policy and objectives at 35 USC 200. Here there are, with a more readable layout:
It is the policy and objective of the Congress 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.
These statements are presented as “policy” and not merely as a rationale for the law. The clear intent is that these statements, and the law that follows, replaces executive branch patent policy for the areas covered by Bayh-Dole–invention disposition in contracting, once a contractor has acquired rights, and the licensing of federally owned inventions. These policy statements may look bland, but they are anything but. Them is fighting words. But to see the fight, you have to go back in time, Mr. Peabody style. Sherman, set the Wayback Machine for 1977 and the Thornton bill that tried to do what Bayh-Dole eventually did.
The Thornton bill has a statement of “purposes” at section 102. Let’s work through the Thornton “purpose,” compare with Bayh-Dole, and see the outlines of the fight that was, and ought to be, and will be again. Continue reading →
We had a look at antecedents to Bayh-Dole’s strange definition of “subject invention.” We saw that the definition has two purposes–
(1) to identify the use of the term “invention” with patent law, so Bayh-Dole works with a defined term not just the common usage of “invention”: “invention” now means inventions that are “protectable” as a matter of patent law (and, um, plant variety protection that’s not patent law). The problem introduced by this new definition becomes what is meant by “protectable,” which is not defined; and
(2) to place such inventions within the scope of federal interest (which is no longer a default claim to own but rather a conditional option to request ownership in limited circumstances).
Patent law (Bayh-Dole is part of patent law), it is worth pointing out, also does not otherwise define invention. The point of patent law is that the individual who invents or discovers is entitled to a patent (35 USC 101):
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
It’s all verbs and adjectives that matter. No definition of “invention”–except here in the part of federal patent law that is Bayh-Dole. The Bayh-Dole variation really should have been more like this:
Whoever in the course of or under a federal funding agreementinvents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title, including the additional conditions and requirements of Chapter 18.
Let’s look at three antecedents for Bayh-Dole’s definition of “subject invention”: The Institutional Patent Agreement master, 1968, that allowed participating non-profits to end-run DWEW contracting policy and take ownership of inventions made with NIH funding (Latker said that Bayh-Dole was a codification of the IPA); the Federal Procurement Regulation, 1975, that codified the Nixon version of executive branch patent policy for contracting (Latker also worked on the FPR); and the Thornton bill (HR 8596), 1977, which proposed a “uniform” federal patent policy (Anker-Johnson said that Bayh-Dole was copped from Thornton). I’ll highlight bits for discussion.
Here’s the IPA (Article II):
(a) The term “subject invention” as used in this Agreement means any process, machine, manufacture, composition of matter or design, or any new or useful improvement thereof, and any variety of plant which is or may be patentable under the Patent Laws of the United States made in the course of or under research supported by grants and awards from the Department of Health, Education, and Welfare.
(b) The term “made” when used in relation to any invention or discovery means its conception or first actual reduction to practice.
No “of the contractor.” No plant varieties (prior to the PVPA). Uses “in the course of or under” (rather than “in the performance of”–conventional language for “as specified and what you did”). “under research” is garble. Restricted to “grants and awards” (IPA restricted to nonprofits–suggests that grants and awards are distinct, and that there may be other forms of funding not covered). Continue reading →
Most university patent policies don’t specify exactly when to disclose an invention. “Promptly” is the recurrent–and meaningless–requirement. Here’s Northwestern University:
to protect academic priority as well as commercial priority, any Inventor making any Invention or Discovery subject to this policy is encouraged to report it promptly in writing and in reasonable detail to the Innovation and New Ventures Office (INVO) via the University’s disclosure web site preferablywithin 30 days of making the Invention or Discovery.
This requirement is made in the context of potential loss of patent rights in foreign countries for public disclosure prior to filing a patent application–though the policy does not bother to make that clear. The policy also makes the absurd claim that patenting is necessary to “protect academic priority.” That’s just nuts. Publication “protects” academic priority. In the U.S. and most other countries, being first to file “protects” patent rights. While there may be an implied requirement to report inventions, the wording “is encouraged to report it promptly” makes the timing discretionary. Report whenever, but report. The “is encouraged” weakens everything that follows: is encouraged to report, is encouraged to report promptly, is encouraged to report in reasonable detail, is encouraged to report to the INVO, preferably within 30 days. And what does it mean to “make” an invention? For obsolete patent purposes, an invention is “made” when it is both conceived and reduced to practice either by a patent application or by testing to demonstrate that it performs in every element as conceived.
So no guidance on when to disclose, but sometime. The express preference is for after the invention has been made–and so, after filing a patent application or completing testing to demonstrate everything works as conceived. Not, apparently: when you have an idea, or have made a sketch of something that hasn’t been built, or prior to completion of full testing. Continue reading →
“Technology transfer” is not so obvious an idea as it may seem. There’s technology transfer from developed nations to “developing” nations. There’s technology transfer from one industry to another. There’s technology transfer from applications in the military to civilian uses. There’s technology transfer from lab to manufacturing and from lab to other labs. There’s even technology transfer from universities to patent licensing firms–as Research Corporation used the term. Despite all these various usages, let’s ask what is involved in effective university technology transfer.
Here’s a decent working definition of technology transfer from a Senate subcommittee report from 1968: “the conscious process by which new knowledge is made available to others than those who generated it” (The Prospects for Technology Transfer: Report of the Subcommittee on Science and Technology to the Select Committee on Small Business United States Senate). Notice that this definition ends with “made available” rather than with “is adopted and used.” It’s one thing to advertise “new knowledge” and it is quite another to find people who have a use for that new knowledge and go on to use it. So we can push this definition a bit to include a better endpoint–when a technology is transferred, it is used by those obtaining it. We can also push on the idea of “process”–there’s no reason why technology transfer must involve a process. Sure, we can limit transfer to “conscious” (as distinct from unconscious, or more seriously, from accidental or spontaneous transfer) but “process” is also too limiting, too happily bureaucratic. We might say, then, that technology transfer is “an intentional activity by which new knowledge is made available to others than those who generated it and is adopted and used by those others.”
We should also ask what is involved in “new knowledge.” How long is new knowledge new? Does it matter? If you “generate” new knowledge today, and someone shows up needing that new knowledge next year, is it still new knowledge? If it is new knowledge to the receiving party, isn’t that enough? Consider, for instance, Everett Rogers’s definition of “innovation”–something perceived as new by an adopting group. For that matter, then, why should technology transfer be restricted to “those who generated it”? Does that mean that inventors must transfer their inventions, and if university patent administrators attempt to do so, that’s not technology transfer because the administrators clearly are not the ones “generating” the “new knowledge”? How about if the administrators transfer rights to a research foundation? The foundation clearly didn’t “generate” the “new knowledge.” You would think that a more workable definition of technology transfer would be the movement of “knowledge” from people who know it to people who don’t know it but use it when they get it. If they don’t use it, then we just have teaching of knowledge–all very well and good, like the crap Paul Simon learned back in high school or what James Taylor can’t remember much of at all. Continue reading →
Here’s a recent search on Research Enterprise: “bayh-dole provisions only apply to subject inventions.” Is it a question? Is it an assertion?
Does Bayh-Dole apply only to subject inventions? No. Some Bayh-Dole provisions do apply to subject inventions, but much of the law does not concern subject inventions. 35 USC 200, for instance, applies to all inventions arising from federally supported research or development. 35 USC 207-209 applies to the licensing of all inventions owned by the federal government and 35 USC 202(c) provisions expressly apply to subject inventions but not to inventions made under a funding agreement but that a contractor does not acquire.
There is no preference for US manufacturing under Bayh-Dole.
Section 204 is titled “Preference for United States industry.” Sounds nice. 204 says it takes precedence over any thing else in Bayh-Dole, and Bayh-Dole says it takes precedence over anything that has gone before but for Stevenson-Wydler. Still sounds nice. But it’s an empty gesture.
Section 204 is more accurately a preference for United States manufacture in limited exclusive licensing to use or to sell in the United States. Continue reading →
A university patent policy designed to promote effective technology transfer will have these key provisions:
Default institutional non-exclusive FRAND offer
These are key elements. FRAND is “fair, reasonable, and non-discriminatory.” We will work through the reasons why these are key provisions for policy. But first, let’s be clear. We are talking university policy for effective technology transfer.
We are not talking about university administrators trying to make a lot of money from patents. The current approach to licensing is no better at making money than other approaches, and in many cases is much, much worse despite all the glossy licensing office reports. The present approach involves a gambling ethos–speculate often and get lucky once a decade. Or, “high risk, high return.” Or, winners pay for losers. Or, this stuff is all “early stage.” All the same fluff talk. It’s just that the patent speculation involves (i) trading on people’s career research; (ii) dealing mostly with other speculators and gamblers; (iii) excluding almost everyone else, from the general public to researchers to small companies to entrepreneurs to big companies that don’t see the point of a university or federal agency demanding an investment in trying to develop a commercial product as a precondition of access or use. Other approaches have made money just as readily, if money is the thing. Other approaches, too, have made better money–money from better sources, for better uses, with better relationships. We can go down this road and explore what “better” means–so long as we don’t end up with what bureaucrats think of as “better.” Institutional pie in the sky is often shit pie, but to bureaucrats it sure looks like chocolate cream shimmering up there. Continue reading →