Having looked at the various topics Research Enterprise has covered over the past four years, it’s also good to look at where we are in terms of university innovation management.
Prior to Bayh-Dole’s passage in 1980, university innovation practice was diverse, highly selective, and led by faculty decisions regarding what to hold on to, and what to let go to sponsors, and especially to the federal government. Most universities operated on an external agent model, with many recommending inventions to Research Corporation and a number to an affiliated research foundation. A very few universities operated their own licensing offices. For federally supported inventions made at universities, faculty inventors typically made the decision when to seek to retain invention ownership rather than assign to the government. For institutions in which such retention of rights was claimed regularly, the government worked out institutional patent agreements, or IPAs. For others, an agency and inventors negotiated things out, with the agency responsible for making sure that whatever private arrangements were proposed would meet agency objectives and look out for a broader public interest.
Bayh-Dole normalized the IPA process for federal agencies while preserving the fundamentals of the approach that was already in place. Indeed, it was the success of the external agent model that made the compelling case for the passage of Bayh-Dole: a diversity of institutional approaches matched to local conditions, availability of resources, and academic mission; use of external agents as the preferred method of developing faculty-generated (and student, and staff) inventions; and faculty-led decisions on what to retain, and what to release to a sponsor, and what to dedicate to the public. The purpose of university patent policy was generally to ensure that commercial interests did not interfere with instruction and research, and that where there was use of university resources, the institution shared in any financial benefits that faculty inventors might receive.
Over time, following the passage of the Bayh-Dole Act, this approach was transformed into something very different. University patent administrators conflated bits and pieces of Bayh-Dole’s language, failed to implement key parts of Bayh-Dole’s implementing regulations and its standard patent rights clauses, and created an institutional shorthand which they substituted for the language of the Act, and which they came to believe was the law of the land. In this substitute Bayh-Dole, “elect to retain title” came to mean “elect title” and that in turn meant “take title”. A requirement of a patent rights clause that established when a federal agency could require under a funding agreement assignment of inventions from inventors as deliverables became, in the restatement, a federal law vesting title in inventions with universities.
This restatement was subtle. University patent administrators claimed to know the law as experts, but they slanted the law to their own purposes. That’s natural, perhaps, but it also played on the trust placed in them both by senior administration and by faculty. University administrators then used this restatement of Bayh-Dole–that it was federal law that universities owned faculty inventions made in federally funded research–to take ownership of thousands of inventions made by faculty, staff, and students.
They have used various versions of the vesting argument: Bayh-Dole compliance requires universities to take ownership (it doesn’t–Bayh-Dole applies to federal agency invention procurement practices; the restatement makes it appear that Bayh-Dole applies to universities); Bayh-Dole vests ownership directly with universities (it doesn’t; Bayh-Dole makes no change in where ownership vests, and certainly does not vest ownership against established federal patent law in universities rather than inventors); Bayh-Dole gives universities first right to take ownership of inventions by choice, by “electing” ownership (it doesn’t; Bayh-Dole’s priority ordering has only to do with whether inventors make an assignment to a pre-approved agent, or do not make such an assignment). Even the variations among these arguments indicates that something is wrong. If the law vests title, then it doesn’t give universities first choice. If universities have merely a choice, then they don’t have to take ownership to comply with the law.
University patent administrators then changed university patent policies to conform to their restatement of Bayh-Dole. At the same time, they broadened the policies to include a claim of ownership to more than patentable inventions. Again, being subtle, they swapped in “intellectual property” for “patents” and swapped in “whether or not patentable” for “may be patentable”, extending institutional claims to copyrights and trademarks, to data, software, technical information, and know-how, among other things. In many policies, what was a condition of employment was that inventions were to be reported, and in a separate step, the inventors and university would decide on the disposition of the invention. Once the patent policies had been altered, however, to move from a review requirement to an ownership claim, the ownership claim became the condition of employment. It was quite the trick to play on university inventors, but it served the self-interest of the patent licensing office to do so. The result of these changes was the confident claim that Bayh-Dole authorized universities to take title to inventions made with federal support as a condition of employment, that Congress expected universities to implement policy changes to comply with this authorization, and that universities had to create technology licensing offices to manage the inventions thus claimed. All this was packaged up as a good thing, since the process of patenting and licensing would lead to royalties from new products, and the sales of those products would add jobs and create wealth. It was a rosy picture, but it wasn’t true, and it has never performed in the manner promised.
The status of this restated and altered Bayh-Dole came to an end in June 2011 with the decision in Stanford v Roche. Stanford, supported by 70 universities, AUTM and a cluster of related university organizations, had argued that Bayh-Dole was indeed a vesting statute. The Supreme Court was unpersuaded. It worked through the law and found nothing that overturned federal patent law that vested title of inventions with their inventors. Bayh-Dole was a law of administrative procedure, not vesting of invention ownership. “Elect to retain title” applied when one was the owner of title. Electing to retain title did not establish that title, mandate taking of title, or imply that the law had somehow already vested title so that it could then be conveniently retained. The stunning result of this decision, at least for a great many university administrators, is that universities have built thirty years of IP policy and licensing practice on clearly faulty, if not intentionally wrong, claims about Bayh-Dole.
Oddly, however, university administrators have not, generally, withdrawn policy statements, changed their guidance documents, or acknowledged that their arguments were simply wrong with regard to Bayh-Dole, that anchor of most university patent licensing practices. Instead, university administrators have re-asserted the argument that the real object of their work was to secure all faculty intellectual property (and other intangible assets as well) as expeditiously as possible, and if Bayh-Dole did not do that, then they would rely on their modified IP policy statements to do it, augmenting them as they could.
Thus, administrators used the Stanford v Roche decision to make the university claims to ownership of faculty scholarship even more extreme. Universities introduced “present assignments” to their policies, employment agreements, and even requests to undertake outside professional work such as personal consulting. A present assignment is an assignment made before an invention is made, so that when the invention comes into existence, the assignment operates to transfer immediately ownership to the assignee. University patent administrators claimed that with a present assignment, Stanford would have been able to double-cross its prior arrangement with Cetus (now owned by Roche) and have complete ownership of the disputed invention, allowing Stanford to sue Roche for a quarter of a billion dollars for infringement. But even this argument was bogus, since the issue in the Stanford v Roche case never was whether the Cetus agreement was superseded by a prior obligation to Stanford, but what happens after Stanford waives its claim to an invention made at Cetus, and then regrets that waiver when it later is involved in making an invention that is also obligated by one co-inventor, to Cetus, with its approval.
Undeterred, university patent administrators have exploited the Stanford v Roche case, which they lost rather conclusively, to demand ownership of faculty (and others’) inventions, using arguments such as a condition of employment (fails for faculty and students, who are not hired to invent in federally funded research), use of resources (fails since resources are provided as a condition of accepting federal awards, and paid for by indirect cost allocations by the government), and participation in extramural research (again, fails as faculty are released from other duties to conduct extramural research, so such a release is not a basis for a claim of ownership). For public universities, the claim to inventions is compounded by eminent domain issues: the taking of private property (which inventions most certainly are under US patent law) for public purposes requires due process and just compensation (which university royalty sharing policies are in general not sufficient, since most of the time there never is any royalty to share, but the taking has indeed taken place, and therefore has taken place without any compensation whatsoever, let alone “just” compensation).
For universities both public and private in a number of states with laws protecting inventors from overreaching employers, university invention ownership claims also appear to violate those statutes, setting up precisely the abuse of power that the laws were intended to prevent–that an employer can simply claim every invention in the hopes of making money by doing so. And that, in essence, is what a university technology licensing office is, once university ownership of inventions is made compulsory. All this, however, is apparently beyond the interest of many university patent administrators, and certainly beyond their professional organization, which is adamant that compulsory institutional ownership is the best way to advance innovation, economic development, and institutional profit-making.
One may reach the conclusion that in addition to being largely ineffective in its primary purpose, which is to license patents to speculative investors to create products producing royalty income, and quietly and pervasively damaging to scholarly collaboration and emerging opportunities of all sorts, the dominant approach to university-originated inventions is fundamentally irrational. It is not that the people involved are irrational, but that they lack sound reasons in law, contract, and policy for what they are doing. Arguably, they also lack sound reasons pertaining to innovation, creative class culture, and the like. Stanford v Roche should have thrown university invention management wide open again. It should be the start of a great new renaissance in innovation practices. Indeed, it can be. But things have to change for this to happen, starting with senior leadership in universities–both administrators and faculty governing bodies have to decide to repudiate the present course, based as it is on sly twists of law and policy to create a uniform policy for institutional ownership and control of faculty scholarship.
It is ironic that institutional ownership of federally supported inventions was the government’s standard position prior to Bayh-Dole. The government asserted that it contracted for inventions as deliverables in its funding agreements. It was the government’s accumulation of such inventions, and its inability to license them with any consistency, that was a compelling argument in favor of Bayh-Dole. Bayh-Dole was not a law that aimed to shift a government monopoly on inventions to universities, but with less public oversight, greater fragmentation, and a propensity for granting exclusive licenses. No, Bayh-Dole moved agencies from an institutional ownership model to a private agent model as a first point of management.
Stated simply, Bayh-Dole provides that if an inventor supported by federal funds assigns inventions to a pre-approved agent, then the federal agency must allow it under the patent clause of the relevant funding agreement, subject to standard performance requirements.
Two classes of agents are pre-approved by Bayh-Dole.
1) agents having invention management as a primary function
2) the inventor’s employer.
In any other situation, the agency may claim the right of approval–including approval of inventors retaining ownership and not assigning to anyone (for which see 37 CFR 401.9).
In this way, Bayh-Dole is a workload reduction law for federal agencies: they did not have to work out custom invention management agreements for each institution, and they did not have to conduct specific reviews of each invention that an inventor had assigned for management. Instead, Bayh-Dole authorized the creation of standard patent rights clauses with uniform procurement provisions to be accepted by any agent that sought to retain title to inventions assigned to it–with Bayh-Dole directing its most demanding requirements at universities and other nonprofits (such as university affiliated research foundations). Universities are not first in line under Bayh-Dole. But they are singled out as the most in need of regulation to protect the interests of inventors, agencies, and the public alike (for which, see 37 CFR 401.14(a)(k)).
There is nothing, therefore, in the claim that is being made these days that Bayh-Dole anticipated that universities would create technology licensing offices with IP expertise and therefore ownership should vest with universities as an expedient first step on the road to commercialization and university profit-seeking. Some proponents of Bayh-Dole may have had this idea. But as far as Bayh-Dole is concerned, universities as the chosen agents can be clueless and resource-starved, so long as they have personnel designated as responsible for patent matters and they educate their personnel on the importance of reporting inventions. University-as-contractor does not have to have invention management as a primary function. At the time Bayh-Dole was enacted, most university technology transfer offices did not have licensing expertise, and many had no licensing function whatsoever. They existed to transfer technology from faculty to external agents. When the University of Washington created its technology transfer office in the early 1980s in response to Bayh-Dole, one could assign inventions to four organizations–the university, the Washington Research Foundation (a private nonprofit foundation affiliated with the university), Research Corporation, and Battelle. Later for a time there was a fifth agent, the Washington Technology Center (a state agency operating on the university campus). For a number of universities, those agents were either an affiliated research foundation or Research Corporation.
It is therefore simply not the case that Bayh-Dole expected university technology licensing offices to come into existence to take back licensing functions from external agents. It is true that Bayh-Dole has permitted such a thing to take place–and it has–but the motive for doing so rests with university administrators, not with federal innovation policy. There is no statement in Bayh-Dole that university ownership of inventions or their operating licensing programs is an objective; there is no requirement in Bayh-Dole that a university even have an IP policy or a technology transfer office–none of this. Not at all. Everything needed to implement Bayh-Dole is found in the standard patent rights clauses, and the obligations for universities as contractors are trivial compared to the pre-approved obligations for agents that accept assignment of federally supported inventions.
The universities’ ubiquitous claims to ownership and control of faculty inventive scholarship are suspect on multiple counts. For starters, there is the argument the federal agencies had used to justify their institutional claims to inventions–uniformity, governmental control, protection of the public from unregulated private interests, and the like. The university position merely argues that university administrators are better at patent management than are federal administrators. I don’t see any evidence that this is true in general.
Furthermore, there is a body of literature that argues that outright institutionalization of inventions is not generally an effective way to advance innovation or economic development. Yes, the UK does have a patent law that vests employee inventions with employers, and various European countries have implemented similar statutes in the last decade, apparently to harmonize with the misstated university version of Bayh-Dole. But the arguments made in favor of doing so are paper arguments–based on institutional self-interest–and not directed at the observed dynamics of creative class culture–including the conditions that existed when Bayh-Dole was being drafted and debated.
We might list a few of the concerns raised by institutionalization of faculty inventions:
- State control of scholarship (in the case of public universities)
- Institutional capture of assets without productive follow-on activity
- Institutional conflicts of interest between research support and integrity, and profit-seeking
- Loss of academic freedom to publish, collaborate, and teach
- Eminent domain taking (in the case of public universities)
- Disenfranchising faculty and other inventors as independent scholars and leaders
- Forcing innovation pathways through narrow, inflexible institutional contracts
In short, the present situation is a fine mess. University IP policies are generally badly conceived and ineptly drafted and implemented. University innovation practice has descended into a monoculture of patent licensing practice, with virtually no reporting of substantive activity or metrics. Many universities lack the resources or expertise to manage the portfolio of inventions that their policies now claim. For all the success stories that are produced in annual reports each year, there is no reporting of the suppression of opportunities, the litigation and threatened action against inventors, the lack of diligence in developing inventions. In short, the present situation is worse than a fine mess, it is a smartly marketed disaster.
What should be done?
First, universities should relax their institutional claims on faculty inventions. This is actually very easy to do. Doing so restores creative class activities and opens the door for a wide range of opportunities. To comply with standard patent rights clauses in federal funding agreements, universities should implement the (f)(2) requirement that delegates commitments regarding inventions to research personnel. Doing these two things restores the external agent model baked so wonderfully into Bayh-Dole.
Second, separate matters of ownership and control. Even when a university comes to own an invention (whether by compulsion or voluntary assignment), there should be a delegation of control that includes the principal investigators and the inventors in any decision with regard to disposition of the invention. That delegation should allow the investigators (before they invent) and inventors (who have invented) to choose their agent–which could be the university, but also could be a research sponsor (as in the case of the Myelin Repair Foundation and others), or a patent management firm (like, say, Foresight, Research Corporation, or Innovaro). A university may require inventors to choose an agent if they aim to pursue commercial licensing, but a university may not require that agent to be the university’s own technology licensing office, any more than it can require all faculty to publish through the university’s press. If inventors refuse to choose, then it is appropriate for the university to request ownership, or direct ownership to an external agent.
Finally, many universities must review and correct their guidance concerning Bayh-Dole, inventions, and innovation dynamics. They must acknowledge the decision in Stanford v. Roche, and if they signed onto the various amicus briefs arguing that Bayh-Dole was a vesting statute, they need to acknowledge publicly that they were wrong and are correcting the record as well as their policies. These universities also need to acknowledge and respect faculty leadership in matters of innovation. The appropriate forum for addressing academic issues pertaining to invention is the faculty governing body, such as a faculty senate. Yes, of course, patent matters are quickly complex, but so are discussions over faculty parking privileges. Faculty are pretty good with the uptake on these matters. For the most part, however, the details of patenting are not the issue: the issue has to do with how innovation pathways are formed, and the consequences for the faculty, for students, and the public in doing so.
With the move to faculty leadership, it is still vital that universities provide supporting resources. Technology transfer offices provide a tremendous service for a challenging area of activity. At one point, the Vice Provost I worked for called technology transfer the second most challenging job in the university, after the dean of medicine. I tend to believe him. Faculty and other university inventors need all the assistance they can get, and what they need most is thoughtful advice, willing to be off script and responsive to the needs of research and opportunities as they arise. Innovation does not need institutional processing, and does not prosper when it is subordinated to the convenience of administrators.
Finally, universities must be willing to negotiate their approaches and obligations with regard to particular inventions that faculty and others are willing to place with their university for management. Such a negotiation is not to be avoided, but rather made the centerpiece of establishing a productive working relationship, where the challenges to success can be acknowledged by both the inventors and the licensing officers.
Doing these things–actually relatively simple snips in the current policy and practice regime–will do a great deal to open up opportunities for research-based innovation.
These changes:
- Are native to faculty scholarship and independent status
- Encourage a great deal more informal collaboration–at the heart of opportunity creation
- Provide greater selectivity of important work for attention
- Build stronger working relationships between inventors and agents
- Allow universities to serve as stewards, looking out for the interests of all
- Create better expectations and commitments for desired results
- Makes more private resources available for use in invention management
- Creates much improved support for multi-institution and open innovation projects
By following these recommendations, universities are actually in a better position to make money, while at the same time reducing their costs and their liabilities. And university licensing offices will still get more work than they can accept–but now they will be in a position to accept only what they really are good at.
This discussion really does reduce to a simple choice:
Adopt methods of creative freedom and work out how to be supportive
OR
Adopt methods of institutional captivity and work out how to mitigate the consequences of exploitation.
My argument in Research Enterprise has been for freedom. The history of innovation bears out the case for freedom. Institutional capture advances the status quo and leads to an emphasis on incremental improvement, mitigation of problems created by the status quo, and a need for efficiency. These are of course areas in which a kind of innovation does come about, but these are not particularly in need of the findings of basic research from universities. These sorts of innovation actually work to entrench status quo positions. As Neal Stephenson has quipped, we may look back on this period and regret that our best minds were employed writing spam filters.
The role of faculty-led university research is much greater than that of entrenching the status quo, even if there is good money in the licensing associated with doing so. Basic research has the potential to challenge the status quo, not merely with rhetoric, but with discoveries and inventions. The neat thing about an invention is that, if it works, it is not merely rhetoric but something new in the world that does something that creates new opportunities that may challenge the status quo, along with its leaders, its experts, its blue ribbon panels, and its best practices. That’s the neat thing about transformative forms of innovation–innovation from the outside, disruptive innovation from below, shanzhai/System D innovation, even high-tech innovation.
These same issues also apply to innovation in research enterprise. The status quo of university patent administrators have created a nice nest with their move to institutional ownership leveraged by a mis-statement of Bayh-Dole. It is understandable that they do not want any challenge to their position, their practice, their investment in developing a compulsory infrastructure to feed a speculative monopoly licensing program. They have shown that they intend to hold their position as long as possible, despite Stanford v Roche, adverse effects on collaboration and opportunity, and despite costs and liabilities far outweighing benefits. Here in the state of Washington, the University of Washington has spent $50m over the past four years on its technology licensing operation, which promised so much income from startup companies that the university would be able to transform its overall financial model. Instead, the results have been flat or falling, staff turnover has been 50%, and the licensing office has stopped publishing annual reports and most of its activity metrics.
Agent choice also challenges the research enterprise status quo. That it does so is truly odd, because agent choice was the model for 70 years or more in university innovation practice. The success of faculty agent choice was the foundation for Bayh-Dole. And it is that choice that has been displaced by a migration of the institutional status quo that was present in federal research contracting prior to Bayh-Dole. Bayh-Dole has been used to expand institutional controls, rather than to encourage the broader application of faculty-led agent choice, which was to be freed of federal agency overhead of review, approval, and conflicting requirements. Bayh-Dole has never had the chance to operate as it was set up to do. Whatever has happened in university licensing and innovation, really, very little of it has had to do with Bayh-Dole. Restoring an agent choice approach would give us a chance to see what Bayh-Dole can really accomplish if given a chance.
Again, it is ironic that some of those so defensive of Bayh-Dole while being so deeply committed to institutional control are actually determined to prevent Bayh-Dole from operating as it was drafted. The agent model is stronger than the institutional capture model, as university licensing before Bayh-Dole demonstrated. For all that, the agent model provides universities with way more opportunities for impact, for collaboration, and even for making money, if that’s what they most want out of all this.
Here at Research Enterprise, I have advocated for freedom over capture. It’s not just an opinion. The argument for freedom is rooted in law, in history, in the dynamics of innovation, in public policy. My advocacy for freedom is also rooted in my twenty years of work in university licensing and research administration environments. I have worked a wide range of university technology licensing practice–from developing forms and databases to policy review to negotiating and drafting license agreements to developing new strategies responsive to the needs of faculty and of industry. The work my colleagues and I have done has been remarkably successful, in terms of working relationships, creating conditions for new enterprise, impact, and even in generating licensing income. I am aware of what goes on in university licensing shops. I know the issues, from the big picture down to the details of whether one should outsource maintenance fee payments. I know what people have said in their unguarded moments, what the stats are that no one is willing to make public. AUTM folks may disagree with me, but there is no way they can say I’m uninformed, haven’t thought these things through, or am insensitive to the nuances and complexities of IP management. There is plenty of room for differences of perspective in this business. The point, however, is that there should be an active debate, there should be a diversity of practice, there should be local programs adapted to local needs, without everyone having to trot in line to whatever a few big university licensing offices happen to think suits them just fine.
This sort of diversity is what Research Enterprise has been advocating. As universities break from their thirty year fascination with a version of Bayh-Dole that has never existed in law, we have the opportunity to build new practice, develop new forms of collaboration and contracting, new ways of deploying intellectual property as a tool for public innovation–and to recover past effective strategies, such as agent choice, that have been overwhelmed by self-interested administrative policies.