I have been trying to figure out what happened to Research Corporation and why AUTM is so antagonistic to the idea of faculty inventors having a choice of who they work with to deploy their research findings. I call it inventor-loathing. It will be a happy day when AUTM renounces its inventor-loathing and rediscovers how productive relationships for creative work can be developed without the need for slave labor, autocratic control, and an ungainly and often poorly done institutional apparatus fitted over every possible opportunity.
There are various histories of university patent stuff and Bayh-Dole. Rebecca Eisenberg and David Mowery and their respective co-authors have done a lot to contribute to this area. Their research challenges commonly made claims, such as that the government was doing a poor job with its patents, that Bayh-Dole sparked the university technology transfer revolution, and that the metrics put out by AUTM show that the present system of compulsory institutional ownership of inventions is a “success.”
The challenge in all of this is that the winners tend to tell a different story than others. Thus, there’s plenty of material around written by folks who are proud of getting Bayh-Dole passed, and who associate the growth of university technology licensing operations with the “success” of Bayh-Dole. But there appear to be factors at play, and as I have talked with folks who were in a position to observe, a different picture gradually emerges. Thus, this provisional history. I expect there are still things I don’t have nuanced or pinned in the right part of the bug collection, but I’m thinking it is getting close.
University faculty created the non-profit external agent approach to patent management, starting with Frederick Cottrell at the University of California who in 1912 set up Research Corporation as a non-profit foundation, one of the first such foundations in the country, to manage a portion of his patent rights in an environmental invention, the electrostatic precipitator. The Wisconsin Alumni Research Foundation (WARF) was created in 1925 by Harry Steenbock and others affiliated with the University of Wisconsin to manage patents related to the irradiation of milk to produce vitamin D, and was a variation on the Research Corporation theme. Rather than providing services nationally, both in terms of assisting faculty with their inventions and providing funding for further research, WARF chose to focus on University of Wisconsin only.
WARF sparked a number of such research foundations, including those at Purdue, Washington State University, and Kansas State. By 1962, there were over 50 such foundations, each focused on an individual campus. In addition to Research Corporation, Battelle Development Corporation began to work with university faculty and their institutions. Other agents, such as for-profit University Patents (now Competitive Technologies), were also supporting faculty work. It was this activity, led mostly by external agents, some national and some focused on a single university, that provided the licensing statistics that underscored the arguments for passage of the Bayh-Dole Act in 1980.
Starting in the mid-1960s, Research Corporation, which had contracts for management with many universities and their affiliated research foundations, had advocated that universities create “technology transfer” offices (Mowrey, et al., 73-77). The purpose of such offices was to assist faculty in identifying inventions, prepare the necessary information for evaluation for management by Research Corporation, and assist in obtaining the necessary signatures should the agent accept an invention for management. In this way, what was transferred was the responsibility to manage patent rights, independent of whether that management took the form of a non-exclusive licensing program (as WARF’s programs had largely been) or an effort to seek a single commercialization partner, as was indicated in some biomedical licensing situations.
These technology transfer offices had no licensing function. At the time Bayh-Dole came into effect, only a few major universities had an in-house patent licensing office, including MIT, the University of California, and Stanford. A minority of universities had formal patent policies. Of those, many did not claim ownership of faculty inventions but rather referred faculty to external agents. Thus, for about 70 years, external agent, non-profit management of faculty inventions was the norm, was successful, worked well with university values, and laid the groundwork for the federal legislation.
When ownership of inventions was claimed by a university, the claim was typically based in hiring faculty with specific job functions to lead institutionally funded programs of research (“official duties”). Otherwise, a grant of ownership arose from a review of what equity the institution and others may have in a given invention, given the circumstances. Such reviews were typically led by faculty committees rather than administrators. A typical university policy (such as Cornell, 1952) provided that if a faculty inventor assigned to Research Corporation or another agent with which the university had a contract for services, then the equity issue was satisfied, since Research Corporation would share a portion of its income, after costs and paying royalties to the inventor, with the contracted university that had hosted the work. University administrations typically did not claim inventions made with the support of the “normal academic environment.” The University of Southern California 1961 patent policy went so far as to mention that “prestige accruing to the University as being the source of useful knowledge” was sufficient consideration.
Research Corporation used the remainder of licensing income after expenses to support faculty research at universities nationally. Early in Research Corporation’s existence, such research support was tremendously valuable to university faculty. Indeed, even after the creation of the NSF in 1950, university faculty treated government funding of research with suspicion, as a political intervention in independent inquiry. In his farewell address to the nation in 1961, President Eisenhower worried the effect of federal funding on the independent researcher. How times have changed. The rapid expansion of federal funding into university work shifted more work to federal agencies, which varied in their approach to inventions. Their approach was based on a series of presidential executive orders recommending flexibility in contracting. Some agencies, such as NSF and the Department of Defense, allowed contractors to own inventions made with federal support. Others, notably DOE and NIH, did not. The NIH situation apparently was dictated by legal counsel for the HEW, now HHS, its parent agency, in the early 1960s.
NIH had worked out a scheme of “Institutional Patent Agreements” with approved universities and their agents that could shake loose federally supported inventions for non-federal management. For each arrangement, the university’s patent policy and licensing practices were reviewed and approved, and a standing agreement signed. The scheme did not extend to most other federal agencies and for those institutions not under an IPA, case-by-case review was slow to execute, and left inventions in limbo as to whether an agent could manage the patent, or whether the federal agency would decide to request assignment. That’s where the “title certainty” argument arose. Title certainty had nothing to do with taking title expediently and potentially against their will from faculty inventors, but rather it had to do with retaining that title when the inventors had already decided to assign it to an agent, and the agent agreed to take management. While NIH IPAs required “Grantees” to have a patent agreement with technical personnel under which they agreed to assign inventions “to Grantee or its approved patent management organization,” that was not the basis for the title certainty issue.
Obviously, if no agent wanted an invention, there was no need to worry anything with the government. This was essentially the University of Arizona’s policy in 1962: the inventor selected what agents to send an invention to for management. If agents wanted the invention, then the inventor picked one to work with. If none of the agents wanted the invention, then disposition was the inventor’s choice. Title certainty was only an issue when an invention was made with federal support from an agency that did not have a routine practice in place that allowed an agent to retain title once assigned.
NIH apparently could not negotiate internally a change in HHS patent policy, so its patent counsel worked to set up federal legislation that would normalize an approach to agency claims to inventions made with federal support. That legislation became the Bayh-Dole Act, and portions of the Act are almost but not quite entirely like the old IPA language. Much of the confusion put out about Bayh-Dole appears to arise from folks claiming that the IPA language and the Bayh-Dole language are one and the same, and whatever was in the IPA was “intended” to be in Bayh-Dole. Perhaps. But Bayh-Dole is what it is, and the text necessary to get the law passed is not then simply to be interpreted as if it is the text of something else, no matter how desired.
In developing the language of the bill, another motivation appears to have been in play. This was the competition between the national agents represented by Research Corporation and the university offices and affiliated agents, represented by MIT, Stanford, and WARF. The university-affiliated agents competed not only for access to inventions made by faculty but also for space in university policy statements. If university administrations were made the first point of control for faculty inventions, then the local foundation would have a better chance at managing a “big hit” invention. The IPA approach was good this way, as it forced the university accepting funds into taking ownership or designating an “approved patent management organization.” The competition was over whether that organization would be local or national. The local foundations that advocated for Bayh-Dole, it appears, wanted the university, not faculty investigators, to make that choice. As Mowery and Sampat put it:
Indeed, a growing fraction of the Research Corporation’s clients bypassed the Corporation during the 1960s and 1970s, “cherrypicking” their invention portfolios and choosing to independently manage the patenting and licensing of those inventions expected to be especially profitable. As we noted above, this “cherrypicking” reflected the success of the Research Corporation’s efforts to train university personnel in the management of invention patenting and licensing, as well as a growing willingness of more universities, especially private universities, to manage these patents themselves.
Again, these inventions were few and far between. In 36 years, Stanford saw 3. In 80 years, WARF saw “a handful” (Mowery et al., 56 citing Blumenthal, et al.). Most universities that have had one big hit have not had another. Bayh-Dole was shaped so that it gave visibility to university administrator control over faculty inventor choice for inventions made with federal support, even though faculty inventor choice was still operative. WARF, UC, and MIT advocated for Bayh-Dole. (Although Mowery et al. state that Research Corporation also supported Bayh-Dole, my understanding is that Research Corporation’s board of directors asked its leadership to take no position.)
The shaping was done by omission. Bayh-Dole does not address how assignment of a federally supported invention gets from the inventors to anyone else. It’s not there. That was left to drafting the implementing regulations, which was a contentious effort among the agencies to agree on common language for a standard patent rights clause, which had to deal with ownership matters. It is therefore remarkable that the drafters did not simply adopt the NIH requirement that universities had to have a patent agreement forcing assignment to the Grantee or its approved patent management organization. If the drafters had done so, however, the university offices would not have been “cherrypicking” any longer, but rather drinking from the end of a fire hose. The aim of university folks was to get a law that would let the universities continue to cherrypick inventions, and leave the rest to the national agents.
Thus, Bayh-Dole did not address a change in the vesting of inventions under federal patent law, and the patent rights clause was part of a federal contract that did not reach directly to each university employee who may work on a grant. The structure of the patent rights clause, as well, was fundamentally different from the IPAs. The IPAs were a kind of master agreement with each university that was approved. The patent rights clause authorized by Bayh-Dole went into each funding agreement. Unlike the IPAs, Bayh-Dole contemplated that the disposition of invention rights would proceed agreement by agreement.
Further, most universities did not require ownership, a majority had no formal patent policy, and it would be a huge administrative burden if they were all required to adopt a policy that changed substantially the arrangements under which faculty were engaged to do research. So the solution that was devised was to require the university to flow down the first obligations with regard to any invention to potential employee inventors. Rather than require a patent policy at each university and require every university to adopt a claim to ownership of inventions as a condition of federal funding, the approach taken in drafting the standard patent rights clause was to require the universities to give the inventors a choice–either assign to an agent or deal with the government’s interest directly with the agency. That is the essence of the (f)(2) provision of the standard patent rights clause, combined with the requirement for direct dealing with inventors under 35 USC 202(d), reprised with variations in 37 CFR 401.9. There would be no need for direct dealing with inventors if Bayh-Dole vested ownership of inventions directly with universities-as-employers. Clearly, there’s more to the drafting of the standard patent rights clause than boffins churning out mechanical prose.
The resulting standard patent rights clause worked for everyone. It was up to the universities individually to decide what sort of invention management program to pursue, based on their local policy, or lack thereof, and contracting preferences. This worked for everyone. Internal licensing shops could get inventions by asking for them or demanding them; external agents could get inventions by having an arrangement with the host university or soliciting inventions directly from inventors. And inventors could still deal directly with the funding agencies if they did not assign to either their host university or to an external agent. It’s actually brilliant, compared with something as dull and petrifyingly lock-step as requiring all universities to demand that their creative workers best able to compete for government grants must assign all their inventions to the university, for access by folks fixed on seeking profit from patents.
What happened next, however, was not anticipated. Bayh-Dole was represented not simply as a means to assure title once title was assigned by limiting agency claims to title, but that the law somehow vested title outright in the university, or that federal law required universities to take title as a matter of compliance, or prevented inventors from assigning to anyone but the university, or at least required inventors to give universities right of first refusal. None of this is in Bayh-Dole or in the standard patent rights clause. But this interpretation was given vocal support by organizations like AUTM. Thus, “elect to retain title” was shortened to “elect title” meaning “take title by notifying the government.”
As a result, universities that had “technology transfer” offices were pushed to transform them, along the lines of UC, Stanford, and MIT, into patent licensing offices. University-affiliated research foundations with significant licensing income, like WARF, could hold their own, but shell foundations were replaced by university-based TLOs. The losers in all this were the national patent management agents, and especially Research Corporation, which sought to reposition itself with a focus on startups with Research Corporation Technologies. The drivers, beyond the misreading of Bayh-Dole, or perhaps which made the misreading attractive, were local control, access to an increasing stream of inventions supported by the government as government funding came to dominant university research, and especially the claim that there would be a better return on investment if Research Corporation (or any local research foundation without a big hit deal) was not in the middle taking a share of income for nationally sourced research. University administrators were attracted by the idea that they could retain all the proceeds after expenses for local university use and share with no-one. Patent licensing officials were attracted to the idea that university funds, rather than private funds, might be used to support the basic operations of a licensing office.
It is understandable, then, that AUTM reports only the activities of its membership, not the broader technology transfer efforts that might be supported by external licensing agents, or by faculty inventors working independently. The AUTM licensing survey charts the growth of the in-house or affiliated office and the demise of the external agent selected by faculty inventors. That growth is held as proxy for university technology transfer and the importance of federal research funding, even though AUTM does not report Bayh-Dole subject inventions separately, and does not report actual use of inventions, but substitutes instead an activity summary to show how busy internal and affiliated offices have become.
It has only been through the US Supreme Court’s ruling in Stanford v Roche that the apparatus of the Bayh-Dole Act has been recovered. The Supreme Court rejected the contention of university patent administrators that Bayh-Dole vested ownership of federally supported inventions with the universities that hosted the research. Administrators were stunned by the decision, and even now most universities have refused to change policy statements and guidance documents that claim federal law requires assignment to the university. Indeed, administrators have moved to change the premise on which the university demands ownership from that of federal law to employment or use of resources, even when it is clear that faculty request a release from official duties to participate in federally funded research and universities agree under funding agreements to provide the resources the faculty use in that research, and the universities are compensated for doing so through indirect costs. Other premises appear to be those of “because we can” or “because we intend to preserve the system we have constructed.”
None of these rationales addresses, however, either innovation or academic freedom. Vannevar Bush valued university faculty for their independence, not because they might be a source for institutional seeking of profits from patent licensing. Those advocating for compulsory university invention ownership policies would do well to consider the foundations of creating a national policy for university research innovation that requires a university administrator to approve, subject to a payment expectation, of each use an inventor or anyone else might make of an invention.
There are other serious, even critical, problems with university internal licensing offices, especially at public universities, but these problems have taken a complete cycle of a patent portfolio–about 25 years–to become evident. These include administrative inefficiencies in public university offices caused by laws and regulations, institutional failure to reinvest in patenting and licensing resources, a policy preference for process over judgment, adoption of an exclusive licensing approach to the detriment of non-exclusive programs (even though Cohen-Boyer and Axel patents, both pre-Bayh-Dole, had been non-exclusive), and starting without a significant “big hit” patent that ensures an initial revenue stream. For those institutions with one big hit deal, it has been nearly impossible to replicate the “success” with more. In part, this is simply a failure to recognize luck rather than process. In part, this lack of repeat success also is the result of the fixation of licensing practices in university policies on a particular version of biotech licensing that followed the first wave of biotech discoveries. These licenses were exclusive to venture-backed biotech companies and to big pharma at a time when there was a spectacular investment window–speculative, well funded, aggressive.
As that window closed in the mid 1990s, rather than look for new speculative windows, and hiring licensing officers who knew these emerging areas of innovation, and adopting licensing practices that would be compatible with practices in these areas, university licensing offices for the most part attempted to force the biotech model on everything new. For those with biotech backgrounds, this was all they knew. For better or worse, their livelihoods were at stake, and every biotech hammer sought to make new inventions into biotech nails. Best practice was ubiquitous practice, and ubiquitous practice was trying to license biomedical inventions to biotech firms, and if existing firms did not want the inventions, then to biotech startups with venture, then angel, then seed, then government backing, as deals dwindled.
Meanwhile, investment windows opened and closed in areas such as semiconductors, software, internet, nanotech, web, synthetic biology, on-line learning, energy, and mobile communications, with university TLOs making little effort to hire appropriate expertise, change licensing models, and focus on the areas receiving speculative investment. In these areas, university arbitrary claims to invention ownership, combined with a demand for an exclusive license to develop product meant that areas in which platforms preceded products were forced into the biotech model, in which product came first, at great expense, followed when the patents expired by a generic platform. University TLOs as result were seen as obstructions to platform development, innovation, and even commercialization because they fragmented rights, blocked the assembly of common platforms and standards, and favored increasingly speculator investors ready to turn to patent assert to recover their investments. All of this increased the overhead of dealing with university TLOs, heightened distrust and uncertainty, and shifted industry away from university collaborations, especially collaborations with smaller companies unable to enlist the advocacy of a dean to protect desired relationships. The recent spate of university litigation is but the playing out of the patterns of TLO badly aligned licensing undertaken for the past decade in non-biotech areas.