Sean O’Connor has written a well documented and argued article regarding the history of the Bayh-Dole Act and what he argues is a flaw in government reasoning regarding assignment practices for inventions that arises from a report written in 1947 called the Biddle Report. O’Connor is worth paying attention to. The Supreme Court cited his work in the AIPLA amicus brief at important points in the development of their decision in Stanford v. Roche.
O’Conner argues that, following the Biddle Report, government policy makers assumed that universities like other federal contractors routinely required assignment of inventions made by employees. This assumption then made its way into assumptions behind Bayh-Dole, and from there to university practice, and even into Supreme Court comments in Stanford v. Roche.
It may well be that this is the case, or at least a part of the case. In my twenty years of work in university technology transfer, however, no one ever mentioned the Biddle Report as an authority for IP policy or practice, nor did anyone ever suggest that Bayh-Dole was predicated on an assumption that universities took title to inventions in the corporate fashion of using a patent agreement that required assignment. Nothing of the sort. All the discussions, all the AUTM workshops, all the policy drafting was directed at “implementing” the Bayh-Dole Act. Whatever claim on invention ownership that universities had, administrators believed (or were told) came from Bayh-Dole.
Thus, while it may well be that there was a misunderstanding arising from the Biddle Report, that can hardly be the only misunderstanding in play. Anyone embedded in university IP practice in the 1980s would have known that universities as a matter of practice might request but generally did not demand assignments from their inventors as a condition of employment following disclosure of an invention. Even the federal Institutional Patent Agreement had to insist that universities obtain an agreement by inventors to assign patent rights–clearly the default arrangement was not to require faculty to assign inventions to their employer.
The effect of Bayh-Dole (according to administrative opinions) was to make it a matter of federal law that an inventor must comply with a university’s request for assignment. This was the basic teaching of Bayh-Dole compliance. It did not matter whether one called this a right of first refusal, a vesting of ownership, or federal requirement to assign–the practice and the effect were the same: the university requested a written assignment from the inventor, and the inventor was given to understand that there was no alternative but to assign. Bayh-Dole required assignment when the university elected (to retain) title. Even if title did not pass to the university automatically, once the university “elected title” an inventor had to assign ownership of the invention to the university. It was easy to skip all the details and then simply state (for the sake of clarity) that Bayh-Dole gave universities ownership of inventions made with federal support.
Furthermore, universities did not uniformly claim ownership of inventions. A number of universities directed inventors to assign to an affiliated research foundation. University patent policies did not require assignment of inventions to the university, but named the foundation as the manager of inventions. Thus, it made no sense that Bayh-Dole was written to somehow force a change in this practice. No one changed their practice, as far as I know, on the belief that Bayh-Dole forced a university to take ownership of every invention made with federal support. The general teaching on the street was: Bayh-Dole gives universities a right to title to inventions, but a university still must get a signed assignment from the inventor to secure that title. No one to my knowledge told inventors that the assignment document was merely a paper compliance requirement of the PTO and did not really assign anything, as federal law had already assigned ownership to the university.
A fundamental misunderstanding is that Bayh-Dole applies to universities. Bayh-Dole does not apply to universities; the law applies to federal agencies, which are required to use a uniform patent clause with universities, nonprofits, and small businesses. That uniform patent clause indeed affects universities, but it is a consequence of the law, not the law itself.
The patent clause itself is set forth in the implementing regulations and is not found in Bayh-Dole. While these “standard patent rights clauses” (as they are called in the implementing regulations) must comply with Bayh-Dole, they are not restricted simply to the provisions stated expressly in Bayh-Dole. Bayh-Dole at 35 USC 206 authorizes the Department of Commerce (previously other bits of federal agency) to draft standard patent rights clauses:
The Secretary of Commerce may issue regulations which may be made applicable to Federal agencies implementing the provisions of sections 202 through 204 of this chapter and shall establish standard funding agreement provisions required under this chapter. The regulations and the standard funding agreement shall be subject to public comment before their issuance.
While the regulations implement the law, the standard funding agreement provisions do something rather different: they transmit the effect of the law into federal funding agreement terms and conditions. That is, what reaches to universities is not the law, but rather a federal contract. Because federal agencies are permitted to tailor the standard patent rights clause at 37 CFR 401.14(a), the agreement that controls any particular relationship between the federal government and a university receiving federal support for research is the particular patent rights clause inserted into each funding agreement.
Bayh-Dole is a law governing federal contracts, not inventions. This is a fundamental misunderstanding that haunts even in law journals the analysis of Bayh-Dole. While it is general practice to ignore the regulations and speak to the law, in the case of Bayh-Dole, this is fatal when the discussion has to do with university practice. The authority for university practice is the patent rights clause in a given federal funding agreement: that is what the university agrees to, that is what controls the relationship, that is what matters for compliance. Bayh-Dole merely sets up the possibility of the relationship, like a clock-maker god, and does not concern itself with the particulars of the relationship. The meeting of minds that forms a federal contract is in the patent rights clause in a given relationship, not over the meaning of the Bayh-Dole Act in the abstract.
Thus, clauses such as 37 CFR 401.14(a)(f)(2), which requires contractors such as universities to require their research employees to make a written agreement to protect the government’s rights in inventions made with federal support, is critical to an understanding of invention assignment compliance. The (f)(2) clause is not in Bayh-Dole, but it is necessary to implement delivery of patent rights to the government in those situations in which a contractor or inventor fails to take certain actions. Even if one assumed that a university would seek to obtain ownership of every invention, the (f)(2) clause would still be prudent, covering those cases in which a university or another agent for some reason did not obtain ownership.
Look carefully at the (f)(2) requirement:
(2) The contractor agrees to require, by written agreement, its employees, other than clerical and nontechnical employees, to disclose promptly in writing to personnel identified as responsible for the administration of patent matters and in a format suggested by the contractor each subject invention made under contract in order that the contractor can comply with the disclosure provisions of paragraph (c), above, and to execute all papers necessary to file patent applications on subject inventions and to establish the government’s rights in the subject inventions.
If the (f)(2) agreement merely required a university to have a patent agreement with each employee to secure for the university ownership of inventions, then (f)(2) would have been drafted rather differently. Not “agrees to require by written agreement employees to execute all papers to establish the government’s rights” but rather “agrees to require assignment to the university by written agreement so that the university may establish the government’s rights.” Here is how an Institutional Patent Agreement template did it:
“Obtain patent agreements requiring all persons to assign all subject inventions to Grantee or its approved patent management organization.” If folks were following the IPA to draft the standard patent rights clause authorized by Bayh-Dole, then they couldn’t have bungled the job any worse. The IPA is crystal clear. The scope is “all persons who perform any part of the work.” Compare with the clumsy “employees, other than clerical and nontechnical employees,” which depends on “employee” rather than “performance of work” and creates a strange exclusion for “clerical and nontechnical” persons, as if these people cannot possibly invent. Consider as well that in the IPA version, the Grantor is required to get an agreement directed at the Grantor, while in the SPRC version, the contractor is required to get an agreement directed at the Government. Finally, note that in the IPA version, the assignee may be either the Grantee or its “approved patent management organization”–namely, an affiliated foundation, such as Research Corporation or WARF (in the case of University of Wisconsin). The IPA is crisp and clear, while the SPRC is a mangled mess, at least if the effort was to port the IPA language into the SPRC.
My view is that there was a big fight over the SPRC, perhaps the pivotal fight in the whole Bayh-Dole affair. This fight was over whether the power of federal funding would be used to force universities to institute corporate controls on inventions (such as demanding ownership of all inventions, now as a requirement of receiving federal support), and whether the university practice of designating one or more invention management agents for use by faculty inventors would have to change, making all inventions first become owned by the university, and then assigned, or licensed, to an agent for disposition. There is nothing in Bayh-Dole that requires either, and one might imagine that in the debate over the SPRC, there must have been advocates who argued that the IPA language was inappropriate and not authorized by Bayh-Dole. That is, while a federal agency and a university might enter voluntarily into a master agreement for a bundle of research tasks, making such an arrangement a pre-condition of all federal support to all universities is an entirely different thing. What is good for the goose might be good for the gander, but it may not be good for every living thing on the face of the earth.
I do not believe the (f)(2) language is the result of a misunderstanding. It is the result of a fight, where the policy issues were perfectly clear to the disputants. The IPA advocates lost this fight. Perhaps as a consolation prize they were allowed to gut public accountability, which they no doubt hated more than how they got ownership of inventions. In any event, (f)(2) requires a delegation of authority by the university to certain employees so that they protect the government’s interest, not a promise to assign inventions to the university. The “action of the contractor” required by (f)(2) is to delegate, not to demand. That delegation in turn is crucial in enabling the definition of “subject invention” to work across all possible outcomes of the disposition of ownership. When an inventor invents, supported by federal funds, that’s a “subject invention” if the inventor is, by delegation under (f)(2), a party to the federal funding agreement through its patent rights clause for special performance of disclosure, assistance, and licensing or assignment to the government. If a party, then a contractor; if a contractor, then the invention is “of the contractor”–meaning the inventor as minor contractor, not the university as general contractor.
The misunderstanding of all this lies in misreadings: “written agreement” is misread to mean “patent agreement between the university and its employees.” That in turn is misread to mean “employee or anyone else” or “employee including for work that employees are released from their duties to the university in order to perform for the federal government.” There are problems in these misreadings having to do with the status of faculty doing work for the federal government: faculty may be in some sense “employees” of the university, but in their work for the federal government, it is not at all clear that universities have standing to claim that such work, when proposed by the faculty members, and which stands outside assigned work under the control of the university-as-employer, permits a university to claim a right to own such work, simply as a condition of employment. It’s not any better if a federal agreement requires a university to make such a demand of its employees, when their work exceeds the scope of the university’s legitimate interest.
It is clear that the (f)(2) language is not directed at requiring a university to have a written agreement that binds certain of its employees to assign inventions to the university. The Supreme Court got this bit right without ever considering (f)(2)–they saw that Bayh-Dole did not address ownership as between a university and an inventor, so the law, clearly could not effectuate a transfer of ownership and could not void a transfer of ownership. Nor could the SPRC demand a transfer of ownership, since there was no authority in Bayh-Dole for doing so. The SPRC could, however, demand a delegation of authority so that inventors became limited parties to the funding agreement, to protect the government’s rights first, before any other action might be taken by inventors, by universities, or invention management agents, or by the federal agencies. And that is what the SPRC does.
Finally, it is worth addressing the continued misreading of the Stanford fact set in Stanford v Roche. Stanford’s patent agreement’s problem, if one wants to call it that, was not that it was a promise to assign rather than a present assignment of future rights, but rather that it did not require assignment when it was possible not to require assignment. Stanford made a big deal about this limitation in its own interest. For Stanford to require assignment, Stanford had to show that it was not possible for Stanford not to require assignment–that Stanford was constrained by law or contract to require assignment. A present assignment would have made absolutely no difference in this. If an invention was out of scope of the patent agreement, then it would not be subject either to a request to assign or notice that assignment had been made automatically. Actually, a present assignment in such a case would be a huge mess. How would anyone know what was automatically assigned until the scope of Stanford’s obligation under law or contract was resolved. Imagine a contract in which a sponsor has a right to an exclusive license from Stanford–well that looks like Stanford is required to obtain assignment. But if the sponsor does not exercise this right, then it looks like Stanford has no requirement to obtain assignment. Until the condition is resolved, how is it possible to understand the status of a present assignment clause?
Beyond all this, there is also the fact that Stanford approved the consulting, and in doing so appears to have waived any later claim Stanford might have wanted to make. The consulting was outside Stanford’s scope of claim. Promise to assign or present assignment did not matter. The consulting was off-campus, not funded by Stanford, not funded by any contract involving Stanford. Stanford had no interest in inventions–though clearly Stanford had an interest in learning PCR technology and applying what it learned to its own research programs. And that’s what the patent agreement required by Cetus was all about–to manage what interest, if any, Stanford would have in inventions made by its employee on loan to Cetus. There, justice is manifest: if Stanford had succeeded in its effort to void the patent agreement between its employee and Cetus, it would also have destroyed the basis for any company working with any university and relying on a written agreement as to disposition of inventions. All a university would have to do to void such an agreement would be to shift federal funds into the work. Clever, but evil.
Stanford of course could change the scope of claim in its patent policy, which it has done, making its claim more general and therefore dulling and antagonistic to academic freedom. That’s a different discussion, and does not involve the facts under Stanford v Roche. Universities adopting a general claim or using present assignment language are not in a better position, other than that most universities have way more money to spend pounding on faculty inventors than most faculty inventors have attempting to explain the law, and equity, and freedom, and innovation to university administrators.
The misunderstandings present in discussions over Bayh-Dole and Stanford v Roche are numerous. That they exist at such a fundamental level indicates the utter lack of professional practice that has sprung up around federally supported research at universities. AUTM may crow about the growth of technology transfer operations, collecting detailed metrics of activity such as patents filed and licenses granted, but policy makers should stop to wonder whether the growth has been poison oak or blueberries–something damaging or something beneficial. My take is that there has been more fundamental, long-term damage than there has been big-hit monetary benefit. University technology transfer is like smoking–it may appear cool, and it might not have any noticeable effect other than being expensive and causing occasional coughing spells, but in the long term, the odds are that it is doing damage, and there will be hell to pay.
The misunderstanding at the root of all of this has to do with the relationship of research to innovation, and with the premises on which universities, as hosts and trustees for such research, come to appoint themselves as unregulated speculators on rights, attackers of industry, and advocates for financial monopolies. Where university administrators should be looking out for community interests, they have abandoned that role to take whatever they can get, if not by speculative monopoly, then by troll-like infringement litigation. Their claim to public purpose in doing so wears so thin it is laughable to see what lies underneath. Sadly, however, tens of thousands of creative ideas, opportunities, and initiatives have been destroyed without notice by this pattern of dealing, all for the hoped-for benefit of money-grubbing administrators. The present great rhetoric of university technology transfer is that faculty are invited to participate in the institutional greed of administrators as they seek their fortune.
Many things in research enterprise are mysteries scabbed over by superficial rhetoric. It may be that we cannot unfold the mystery of how someone thinks to look in an unexpected place, or comes to see differently what otherwise is obvious, nor do we know much at all about how the next innovation might happen, or the next black swan. But we can rule some things out, or at least limit their claims to what they can justify. University ownership of faculty inventions by fiat, for speculative money-seeking is one of these. Historically, freedom was the university watchword for “success”–and historically, that’s how research work was disseminated. Historically, this approach to inventions laid the foundation for the arguments that pushed Bayh-Dole through Congress. What happened after Bayh-Dole was passed, however, was not simple misunderstanding: it was deliberate. It was a matter of taking what others did not know to resist, a way of exploiting an administrative claim against faculty freedom to get invention rights closer to speculators and further from industry.
The ownership issue lying behind Bayh-Dole is caught up in federal agency legal politics: must an agency demand, as HEW had done, ownership of all inventions that federal funding has in some way supported? The pharmaceutical industry instituted a boycott of federally supported compounds in the early 1960s in response to the HEW invention ownership policy. The workaround in Bayh-Dole was that it would be okay for industry to get rights to federal inventions–even without paying the government for those rights–if industry obtained the rights from a university-as-owner. Universities were set up as the proxy for industry rights in federally supported research. Perhaps the motivation was to circumvent GATT limitations on government subsidies to industry, as a matter of fair trade. Maybe the motivation was just to find resolution to legal bickering between federal agencies. In any case, the idea was that inventors working with their universities or other invention management agents might find industrial developers for the application of federally supported inventions better than could the federal government. What was misunderstood was the expectation that industry would out-compete speculators for access to these inventions. Worse, it was not anticipated that universities themselves would come to prefer speculation to broad dissemination, and aspire to be the speculators themselves, if given the chance.
The present calls for universities to be directly involved in startup companies are one symptom of this turn to speculation as the primary outcome of Bayh-Dole. Places like the University of Washington have drafted formal business plans under which speculation in “early stage” company equity is presented as offering the prospect of changing the financial model of the university. UCLA has proposed a “Newco” that will take over decisions about IP, apparently without carrying any of the risk, and somehow creating huge income streams for both UCLA and the speculators behind Newco. The University of Utah, which made much of the idea of startups, used five years of shell companies with pretty web sites to snooker millions in economic development aid out of state government–aid that went to build more buildings at the University and did next to nothing for innovation, jobs, or economic development.
There is nothing at all wrong with startups based on patented research findings, just as there is nothing wrong with startups based on anything but research findings. For every Google, there are hundreds of Facebooks and Apples. What is wrong is the university administrative involvement, 1) in taking ownership of what it should manage on behalf of others; 2) by seeking to profit for itself in that ownership; 3) by preferring speculators to practitioners; and 4) by turning to litigation to raise the financial value of its patent positions. These positions within the university administration motivate the development of misreadings of Bayh-Dole, of Stanford v Roche, and of economic development metrics. One may call them misunderstandings, but if bullshit is statements made without regard for the truth, then these are a kind of complementary bullshit–misunderstandings made without regard for the truth. We are not faced with a “gosh, we were wrong but only on a complex technicality” moment here. It is not simply a problem with the Biddle Report from 1947. It is a “My god, what have I done” kind of realization that ought to sweep through university administrator souls. The sooner the better.