Michael Eisen has an insightful discussion on his It is NOT Junk site on the CRISPR patent dispute between the University of California and the Broad Institute (Harvard, MIT). It is well worth the read. Behind the discussion lurks the Bayh-Dole Act and its effects on university licensing mindsets. The universities involved in the patent dispute are fighting over who should have the right to troll industry for patent royalties (and, of course, to pocket the money). As Eisen puts it
The academic quest for patents is no longer the side story. Where once technology licensing staff rushed to secure intellectual property before scientists blab about their work, patents now, in many quarters, dominate the game. Experiments are done to stake out claims, new discoveries are held in secrecy and talks and publication are delayed so as not to interfere with patent claims. This is bad enough. But the most worrying trend has been the willingness of some researchers and research institutions to distort history, demean their colleagues and misrepresent the scientific process to support these efforts.
. . .
Encouraged by a small number of patents that made huge sums, universities developed massive infrastructure to profit from their researchers. Not only do they spend millions on patents, they’ve turned every interaction scientists have with each other into an intellectual property transaction. Everything I get from or send to a colleague at another academic institution involves a complex legal agreement whose purpose is not to promote science but to protect the university’s ability to profit from hypothetical inventions that might arise from scientists doing what we’re supposed to do – share our work with each other.
These are concerning observations that show the effect of the universities’ decisions about how to manage inventions after the Bayh-Dole Act. Bayh-Dole permitted universities to own inventions made with federal support (and therefore gave administrators incentives to forcibly take inventions from investigators and deny those investigators the opportunity to decide whether the invention should be made open, assigned to the government (to be made open, in all likelihood), or split up into a public use part and a commercial part for development). Instead of investigator choice, or inventor choice, Bayh-Dole gave us bureaucrat choice, as if this would be the key to unlock American innovation from academic research.
The idea that bureaucrats should decide how research develops is itself an iffy proposition. Here’s how Eisen puts it:
The logic of Bayh-Dole was that, if they owned patents in their work, universities and other grantees would be incentivized to have their inventions turned into products, thereby benefiting the public.
This is a reasonably accurate statement of the thinking. The problem in that thinking comes with the word “their.” There is nothing that would indicate that inventions made with federal support in faculty-proposed projects have anything to do with a university ownership interest. The extent of the university’s involvement is to release faculty to conduct the research. The university is compensated for the research team’s time, for the use of facilities and supplies, and even for the overhead of managing the money. The university has nothing in this game, no reason for an equitable claim on title, nor a license, nor even a shop right. A research finding is not a university finding simply because a university hosts the research. Makes no sense. There is no “their” there.
But here is Norman Latker, NIH patent counsel, testifying before a House subcommittee in 1977 (all caps in Latker’s ms):
IN THIS CONTEXT IT IS APPARENT THAT THE EXISTENCE OF A LICENSABLE PATENT RIGHT IS PROBABLY A PRIMARY FACTOR IN THE SUCCESSFUL TRANSFER OF A UNIVERSITY INNOVATION TO INDUSTRY AND THE MARKETPLACE, AND FAILURE TO PROTECT SUCH RIGHT MAY FATALLY AFFECT A TRANSFER OF A MAJOR HEALTH INNOVATION.
The “university innovation” can be read to mean “an innovation created at a university” or as “an innovation owned by a university.” The slip between the two is no doubt intentional, as Latker’s effort is to enable university ownership. Latker’s use of “probably” gets eliminated by the time we have Bayh-Dole (which was drafted by Latker on the sly while working for the NIH, against the policy directives of his HEW superiors).
The first hook in the scheme is that patents are necessary to “transfer” technology “to industry and the marketplace.” That “probably” carries serious weight–and Latker offers no evidence for the adverb, other than that he argues that penicillin was delayed over a decade because there were no proprietary rights in it to support the profit motives of industry. This, leaving aside that if the compound required “development,” then the proprietary positions would come with finding better ways to produce the compound in quantity and with the right balance of effects and side effects. Latker also argues that having more than one company work on development would, somehow, be a huge waste of resources–leaving the idea that innovation was for monopolists. Perhaps in some way that’s true–Peter Thiel goes there in Zero to One.
But we might also consider that competitive development does not have to be duplicative–companies can develop an invention in different ways, apply the invention in different ways, combine the invention with different product sets, for different applications. That creates diversity and creates incentives for standardization, which in turn leads to cooperation of a sort.
Further, companies do not have to be bitter rivals in the development of platform technologies–the sort of thing that it’s claimed university “basic” research is well suited to. Companies can cooperate, like the bike racing teams in the peloton in the Tour de France. They cooperate to set up their competitive end games. That same is often true with new technology–before there can be competition, a platform needs to develop, with variations, interfaces, performance tests, and all the rest. And for platforms to develop, commons of various sorts are clearly important, whether the public domain or an open standard or technology available for cross-licensing (you won’t sue me and I won’t sue you stuff).
A primary mindset that ought to be present in university patent licensing, and isn’t, is that of “crossing the commons”: help to create a commons that supports platform development, and from this platform various folks can then compete to provide services, products, and the like, keeping the platform itself free to develop as well. That’s how we have the digital computer and the internet. That’s how we got the aircraft industry going in the United States. That’s why nanotechnology has languished here for two decades–universities patented every little variation and held onto each with white knuckles, refusing to form commons, and so industry for the most part has to wait twenty to thirty years to gain access to all those research findings and sort them out for utility.
Of course, if Latker was at all right with his “probably,” then the university-produced patent gridlock approach, leaving technology for twenty years until it hits the public domain, means that the universities have essentially destroyed the commercial opportunity. If industry wouldn’t touch penicillin without a monopoly opportunity, then why on earth would industry bother with two decades of university nanotechnology findings? Of course, I think Latker was totally bonkers, but in that clever political way which if not contested can carry the day, which he managed to do, and so we have Bayh-Dole, which burns his bonkeredness into federal statute.
Here’s Howard Bremer, patent counsel for the Wisconsin Alumni Research Foundation, testifying a before a Senate subcommittee in 1978:
Despite all of the difficulties attendant upon technology transfer, universities are in a unique position to objectively seek the best qualified industrial developer and, under appropriate licensing arrangements, to monitor the diligence of development efforts by such developer. Such arrangements can, of course, be made only if the university can furnish some incentive to the industrial developer, the innovator. This is best provided in the form of a limited exclusive license under appropriate patent coverage.
Bremer argues that universities (note the abstraction–a university cannot possibly do any seeking or have an “objective” perspective because a university is an abstraction and has no mind whatsoever) should be the ones deciding how to place research findings with industry. There’s no indication that university investigators, by publishing or teaching or consulting, might also make the transfer of research findings–or that inventors might, without the involvement of university administrators, work with any of a number of patent brokers, such as Research Corporation, set up to introduce new technology to industry. Somehow, without meaningful consideration, universities become the ideal champion for “objectively” deciding technology winners and losers when it comes to research inventions.
For Bremer’s scheme to work, of course, he needs a law that strips faculty investigators of the opportunity to decide among the various options–publish openly and blow patent rights; or secure patent rights to defend research from blocking monopolist speculators; or secure patent rights as a negotiating tool to open up proprietary technologies to standards; or assign to the government, which will decide between dedicating the invention to the public domain or licensing it non-exclusively; or work with a patent broker to license the invention to industry for development. Investigators can make such decisions, too, and have.
Put another way, Bremer must attack academic freedom. A patent is a publication, and Bremer’s scheme works if university administrators dictate where and how a faculty inventor may publish his or her work, and control the effect of that publication. If a discovery is made open, anyone can read the published account and practice the invention. If there’s a university patent behind it, then the publication is an announcement but not a teaching document. No one can practice the invention until a university bureaucrat approves. And according to Bremer’s scheme, a university bureaucrat will decide “objectively” on the “best qualified industrial developer” and everyone else can go take a swirly. Missing from Bremer’s scheme is the likelihood that no industrial developer will step up and take on the burden of a university exclusive license, with university bureaucrats holding whips to make sure there’s “success” (or, as Bremer puts it, “to monitor the diligence of development efforts”). Bremer was as bonkers as Latker, but clearly knew who to work a crowd that didn’t have a mind for the details.
Bremer then moves from the premise that universities must play the role of choosing winners and losers in industry to the need for them to use patents to do so, citing the Institutional Patent Agreement (IPA) program as an instance:
The certainty of the university having ownership of any patents is essential to the transfer of the technology. Without that certainty, timely patenting activities cannot be engaged in and the inventions are less likely to be developed to the point of marketability. This is equally true whether federal or other funds are involved in the research effort leading to the conception or reduction to practice of an invention, and is the principal reason why the IPA is so important to the universities. Most universities though they rarely make any sizeable income from inventions, would largely lose the incentive to seek licensees if they did not hold the patent rights.
This is the canard about “title certainty.” If a university cannot strip title from inventors with impunity and then prevent the federal government from getting it back, then how can universities create the monopolies in industry needed for there to be innovation? Work it slowly–Bremer argues 1) that universities, not inventors or the federal government, should own inventions; 2) that investigators should have no role in managing the research findings that arise in the projects they lead; 3) that bureaucrats should have the lead role in deciding who in industry should benefit; and 4) that universities must have some “incentive” to get involved, and that incentive should be financial.
Bremer argues that somehow this scheme will better “involve the inventor” in the commercialization effort, agian, apparently because there is money to be made by the university (and shared with inventors). Thus, “commercialization” comes to mean not just the development of research findings into useful products but also the reshaping of university administrative values around generating income from patent positions and trying to change research “culture” to look out for profits ahead of findings–just what Eisen decries.
It may not be that getting rid of all patents in university research is the thing–but getting universities out of the patent business, other than perhaps as voluntarily chosen escrow agents to build platforms and commons would appear to be the way forward out of the mess created by Bayh-Dole.
A Wired article by Megan Molteni takes up Eisen’s concerns (ad block warning) and melds them with James Love’s efforts to petition the NIH for march-in to deal with the UC-Broad Institute fuss. Wired doesn’t do so well with Bayh-Dole. Here’s a bit from the next to last paragraph:
A director at the NIH declined to comment on Love’s petition. But the agency does have an interest in not exercising its legal leverage. If it were to start granting marching-in requests, NIH grants would look less appealing to researchers. Taking federal money would mean accepting weaker patent rights. And that could impact the agency’s ability to draw high-profile collaborators.
What do you make of the logic here? Where are the researchers who wouldn’t seek NIH funding because if they did, the NIH might “march in” to protect the public interest, and that would be, what, bad?
Taking federal money already means accepting weaker patent rights–that’s what the line of federal government reasoning has been since 1947 to the present. That’s what Bayh-Dole originally implemented and what is still present in the law, though chopped up by amendments and reduced to trash in the implementing procedures–Bremer, for instance, took credit for making sure the march-in provisions would never operate. It’s difficult to believe that the NIH worries about whether its grants program “draws high-profile collaborators.” Perhaps Molteni has some inside information from NIH, but really, it looks like she is just making something up that to her mind sounds good. Given Eisen’s comments regarding patents and my own experience working with university researchers, I expect 95% or more researchers couldn’t give a rat’s ass about patent rights but they do get bothered and even upset when a university strips them of their personal and academic rights and choices, and limits their involvement in the release of their findings to a share of the money (if ever and usually never).
It is time for Bayh-Dole to go. Enough of bonkers approaches to university research innovation. Bayh-Dole clearly doesn’t work–its commercialization rate is unreported but apparently under 1% and perhaps much less, two orders of magnitude below the rates claimed by universities for non-federal inventions in the 1970s–when most patent work was voluntary, agent-based, and highly selective. Bayh-Dole fragments research inventions, prevents the formation of commons and platforms, and pits university against university. A lot of wasted effort, technologies delayed for decades, but with livelihoods for bureaucrats. Worst of all, $40b or so each year of public money without public oversight or accountability.