In a series of three articles (here, here, and here), I showed how the University of Michigan intellectual property policy apparatus managed–or didn’t manage–patents and copyrights. One problem is Regents Bylaw 3.10. The Bylaw sets out conditions under which it is appropriate for the university to take ownership of patents and copyrights. In doing so, it brings forward the language of the 1944 patent policy, but then conflates two separate conditions, adds copyright and enigmatic “property rights” in software, and follows with a string of conditions, exceptions, and worries. The basic premise of ownership described by Bylaw 3.10 is this: “When the university specially supports an activity that either needs or produces patents or copyrights or software, the university should own those patents and copyrights and software and enjoy whatever financial benefit might come from such ownership.”
The burden of 3.10 is to give scope to what it means to support an activity. 3.10 focuses on money “administered” by the university. That is, money in university accounts. The worry of the Bylaw is that money from donors or from the state or from research sponsors or from program income might be treated differently. So source of funds is disclaimed. A second worry is that the money could go to support an indirect expense rather than a direct expense. A laboratory could be outfitted with new equipment, for instance. So Bylaw 3.10 calls out indirect support to make clear support is not limited to direct funds.
The distinction between direct and indirect research funding is well established in sponsored research administration. Use of the language here underscores the idea that the purpose of the policy is to deal with when the university itself acts as a research sponsor. This has been a concern in university patent policies from the earliest efforts at policy drafting. If an external sponsor of research can negotiate for rights in patents and copyrights arising in a university research project, why should not the university itself be also able to do that, negotiating with itself, so to speak?
Why should a university negotiate with itself? That’s the interesting question. The reason has to do with the very structure of a university, which exists to support a faculty that carries out the mission of the university. The mission of the university administration is to provide support to the faculty, which in doing what faculty do (teaching, research, public service–professing), accomplish what the founders and supporters of the university desire. The University of Michigan Faculty Handbook (1.C) makes this point, that
Academic freedom is not a term or a condition of employment; rather, it is based in the institutional structure of this and other universities and is fundamental to their common mission of promoting inquiry and advancing the sum of human knowledge and understanding.
The faculty have freedom. That means that what they do and produce is not subject to administrative or corporate control, as a matter of the structure of the university, not as a condition of employment. Throughout the Faculty Handbook, faculty are appointed, not employed. There’s a reason that the Faculty Handbook does not use a term drawn from the rise of corporation-based master-servant relationships with regard to what faculty do for the university. Faculty have freedom. That’s the point of the university. But faculty freedom also means that individual members of the faculty can agree to limit their freedom in return for some opportunity. So in exchange for research funding, a faculty principal investigator may agree to assign patent rights in inventions to the sponsor. Or in exchange for publishing an article, a faculty member might assign copyright in the article to a publisher. Similarly, if the university administration seeks a special service from a faculty member, the faculty member is free to negotiate the terms of that service.
Perhaps the university wishes the faculty member to serve in an administrative capacity for a program, and so produce materials that support that program, and in exchange, the faculty member will receive additional salary and have an account for the program from which to spend. In return, the stuff that gets produced to support the program will be the university’s not personal to the faculty member. It’s a voluntary deal, and so preserves academic freedom. If a faculty member could not negotiate in this way, then her academic freedom would be limited–she could not work anything out with the university.
As faculty became involved in research endeavors, and those endeavors proved capable of producing inventions of value to industry, universities began hiring for research positions–people who would work on directed projects with the aim of producing the next heparin or irradiated milk. When a university hired for such projects, administrators–as well as other members of the university community–felt a need to see some share of any possible financial upside from such dedicated support. So one of the first areas of concern in university patent policies–aside from disavowing patenting, especially in medicine–was to set out expectations and procedures when the university provided special resources, above and beyond those that it made available to faculty as an essential part of the administrative mission–salary, office, secretarial support, library, shop, professional development funds for travel, and the like. These things came with the territory; they were “based in the institutional structure,” so to speak, as part of academic freedom.
The point of policy was to provide guidance on when it was appropriate for the university to expect a share of the financial success of inventions made in such projects. The general approach to the problem was to create a review panel or committee with significant faculty representation (if not entirely faculty) to review circumstances of special support, compare with policy guidance, and make a determination of what equitable share the university ought to receive–nothing, reimbursement, some share of royalties, all royalties, ownership. The whole range of possibilities was generally available.
In this approach, the university did not withhold special resources until a faculty member caved and agreed to assign all inventions to the university, nor did it hand out resources indifferently to faculty to fuel their personal financial adventures with patents, nor did it claim that any use of the university’s resources or any receipt of pay constituted special resources that gave rise to a claim of university ownership. The equitable review approach was an insightful solution, as it allowed the faculty to review the work of peers and consider personal initiative and insight relative to the special benefits made available by the university–benefits one of the faculty had received, and therefore the others had not.
Two variations on the review method were also developed. One was to call for written agreements to set out the expectations of the university in providing resources and faculty members for accepting them. Such an approach also makes sense, though it also may be time consuming. And how does one decide what is fair, upfront? How does one be consistent when dealing with a junior faculty member one day and a powerful dean the next? Another variation dealt with these issues by setting a threshold in policy for what constituted “significant” resources and place a standing claim on the ownership of inventions developed with support that met the threshold. That saved everyone the bother of review or negotiation. Administrators, however, love avoiding bother, even if the bother is actually at the heart of sound judgment and respect for the idea the university. If a policy can diminish the need for individual judgment while providing an appearance of respect, it all seems so much better–one can hire lower on the evolutionary ladder, spend less time with training, and not worry so much over social niceties.
The University of Michigan 1944 patent policy is a typical effort to express the general statement that specially funded projects that result in patents–whether produced or acquired–ought to carry with them the condition that the university owns those patents. The policy does not attempt to set a minimum threshold for funding and does not include a review committee, but Archie Palmer reports in 1962 that with regard to the 1944 policy “as a general practice, each situation is considered on an individual basis.” The point, then, of the 1944 policy was to establish a principle–university specially funds, resulting patents are the property of the university. How that principle gets interpreted was still case-by-case. Still an equity review for when the university would take ownership, and when not, which makes sense if the policy is not going to state a threshold for funding that triggers the ownership claim.
Regent Bylaw 3.10, which comes along after 1962 and before 1976, keeps the same structure, but with conflation, garble, extensions, and exceptions. The issue still is special funding, but now adding copyright and software, worrying about mixed situations, excluding employment as a form of special funding, and the like. If Bylaw 3.10 had been intended to read as a comprehensive claim on all inventions, copyrights, and software created by anyone receiving funds or using facilities or using resources or working within the scope of employment or working within the scope of one’s field of academic appointment, or working in any way under the “auspices” of the university, there are plenty of direct, clear ways to make the point.
If Bylaw 3.10 had been intended to reach from forms of intellectual property–patents, copyrights, mystery “property rights” in software–to the underlying assets of inventions, works of authorship, code, or anything else, whether it could be owned or not–again, there are ways, not quite so direct and clear, of getting at such things. Bylaw 3.10 does none of this. It takes a tangled mind, or a purposely clever one, to read into Bylaw 3.10 what is not there, and to claim that while the ordinary reader might not be able to recognize a general claim to own all assets produced by anyone at or in or using anything of the university’s, a clairvoyant administrator, properly empowered with the monopoly right of interpretation, may divine the true intent and contort the wording around to this effect.
The Shaw court’s view echoes here: a university’s true intent that is not expressed is irrelevant to the construction of a contract that incorporates a university patent policy. There is a simple lesson in both policy and contracting: words matter. We all recognize that words have limits–one can rarely get exactly what one wants, even when writing fiction. But we can make words do useful work, with clarity, with good sense, with logic, and with reasons. Doing so invites goodwill and common cause, while leaving room on the margins for debate, negotiation, and compromises. If people come to think they are writing sacred text, or have become lawgivers, or even merely have been given the opportunity to lay land mines to protect their livelihoods and opinions, policy and contract both are pretty much doomed.
If Bylaw 3.10 was intended to adopt the wording of prior patent policy but change the meaning wholesale from special projects to any activities at or for or with university support, then it is an easy step to call out the changes, to put everyone on notice, to create a text that any reasonable person might read and reach such an understanding. That is not what Bylaw 3.10 does. It does not state any such general claim.
The apparatus of technology transfer policy, copyright policy, and present assignments is built on the authority of Bylaw 3.10. That apparatus spends a great deal of time muddying the water with grossly bad drafting, extending university claims into regions that Bylaw 3.10 provides no authority, and loading the policy with the ability to change the policy at will, while demanding that as a condition of employment (that is, for faculty, as an attack on tenure), everyone at the university must accept whatever changes the Office of the Vice President for Research imposes, or whatever any unit of the university “in good faith” bargains differently in a contract. The fundamental claim that gets made is the one voiced to the Regents in 1985 by Dr. Sussman, that “the university owns all intellectual properties of the faculty.” Wherever this claim comes from, it does not come from Bylaw 3.10 and it does not come about as a result of federal patent or copyright law, and it does not come about from the Faculty Handbook, and it does not appear to be based in any express employment contracts or patent agreements from 1985 that the university entered into with each member of the faculty.
As for the university patent policy in 1985, I don’t have a copy. But the Proceedings of the Regents does have the text of the Policy on Intellectual Properties introduced in 1987, which according to the Proceedings is revised to “bring it up to contemporary status and make it comparable to those in effect at peer institutions.” Given the lack of clear drafting in many post-Bayh-Dole efforts at revising university patent policies, it is worth wondering just what those at the University of Michigan thought those peer patent policies actually did. The 1987 policy recites Bylaw 3.10 and states that the “first, and primary” objective of the the policy is to “facilitate” the efforts of faculty and staff to “carry out the University’s mission.” Two other objectives are mentioned–“efficient transfer of knowledge and technology” and “attract resources for the support of University programs.”
There is no claim of ownership in the policy. Instead, the policy requires disclosure of “intellectual property”:
University employees have an obligation to disclose promptly and completely to the Intellectual Properties Office any intellectual property developed or discovered as described in Regents’ Bylaw 3.10.
It is not clear at all where the employees have got this obligation from–the language here makes it appear that the obligation comes from some other source, not the present policy, and not from Bylaw 3.10, which is invoked only for the scope of the obligated disclosure. There is no special definition of “intellectual property” in the new policy statement, so it would appear that patents, copyrights, and any property rights in software (which by 1987 would have been patents and copyrights, as forms of intellectual property, though in some odd way, perhaps trademarks might come within the scope of “property rights” pertaining to software) are the assets in question. But disclosure typically is concerned with reporting inventions, not in reporting the existence of a patent. Further, one develops inventions or discovers new things, one does not develop a patent or discover a patent. So intellectual property here is meant to cover not only statutory forms but also the works on which the legal rights are based.
The words “developed or discovered” are a strange twist on the language of Bylaw 3.10 and a needless restriction. Bylaw 3.10 concerns patents and copyrights that are “acquired or issued as a result of or in connection with” educational activities specially funded by the university (to paraphrase the last part for clarity). The wording in the 1987 policy ought to use “as a result of or in connection with” and not introduce new terms. While “discovery” has a connection with patent law, being referenced in the US Constitution (“secure for limited times to inventors … the exclusive right … to their discoveries”), the subject of a patent is an invention and the act that matters is the one of discovering the invention, not discovery in general. “Discoveries” is therefore much more sweeping than “invention,” which in turn is more general than what is really looked for, “an invention which is or may be patentable.” “Developed” is generic and it is difficult to understand what scope is intended here. Perhaps it is used to mean “made” or “come into being,” but “developed” may also mean “augment” or “advance”–which again would be much require a scope broader than the language of Bylaw 3.10 supports. Perhaps “developed” is merely bad usage for what authors do when they fix original works of authorship in a tangible medium of expression. In that case, why not use “prepared or invented”–or better yet, why not state the conditions for disclosing patentable inventions separate from those for works of authorship and avoid the conflation altogether?
It is possible that putting the tag “as described in Regents’ Bylaw 3.10” at the end of the sentence reduces all the wording to whatever it is that Bylaw 3.10 does describe–in which case, why not just cite Bylaw 3.10 accurately? In any event, the policy requires disclosure only of the work that Bylaw 3.10 expects to be the property of the university.
What puts all of this in perspective is an asterisk note at the end of the 1987 policy statement (there is no asterisk in the text corresponding to it):
The University will continue the tradition of not claiming ownership or a share of the proceeds from scholarly works and textbooks. Software and works expressly commissioned by the University will continue to come within the purview of the policy.
Here we find that “scholarly works” does not include “textbooks.” Both forms of works are not “expressly commissioned” by the university–at least a gesture to the thrust of Bylaw 3.10’s special funding for educational activities, though expressly commissioning a work is only one way in which special funding might be authorized. It is an odd expression, however, to write that the university “will continue the tradition of not claiming” when there is nothing in policy that establishes a claim to scholarly works or textbooks unless there is special funding associated with it. That’s not a “tradition”–it’s either a function of policy that does not authorize such claims, or a pattern of setting aside a claim in policy (as, when a scholarly work or text meets the scope of arising in or connected with a specially funded educational activity). There is nothing about such behaviors that rises to the level of a “tradition”–it’s policy and established practice under the policy. To call something a “tradition” here is, essentially, to designate it for change at a whim, as if the policy does claim ownership of scholarly materials and textbooks, but the university administration chooses not to have the fight about it at this time. “Regardless of what you may think, I own all your stuff, but since you will object, I will bide my time until the opportunity presents to take what I want.”
This is all complicated further, however, by yet another statement in the policy:
Inventors and authors of intellectual properties resulting from specifically assigned tasks will not normally participate in the royalty/equity revenue distribution plan described herein.
So even though scholarly works and textbooks are not even claimed by the university and lie outside the policy altogether, if one is “specifically assigned” a task then the university apparently owns the work and owes no royalties. Or is the condition triggered when one is “expressly commissioned” to create a work?–is it the task that is assigned or the work product that is desired by the university? In any event, if the task that is specifically assigned involves federal funding, then this policy provision violates the standard patent rights clause in the funding agreement.
The 1987 policy, in an interesting move, gives management options to those disclosing their developments and discoveries. (The policy introduces the term “inventors” and promptly defines it, conflatingly, to mean “inventors or authors.” That is, “inventor” means “inventor” and “not-inventor.” There must be some disease in Ann Arbor that has prevented clear drafting practices for decades.) All three options assume “commercialization”–there is no indication that any other forms of deployment of intellectual property exist or might be considered, such as broad public access and use, standards development, or dedication to a commons, nor even mixed use, where for research and internal uses there are no controls, while commercial implementations must provide some value added. The three options are:
- Licensing Third Parties
- Licensing Employee-Inventor Owned Companies
- Reassignment of Ownership to Inventors
The options imply that there is an assignment step by which the university comes to own the intellectual properties prior to the “inventors” making an election. If the election is with the “inventors,” why not require disclosure, make a determination that the university won’t take ownership, and then waive the obligation to assign, and let the thing rest? Instead, the reassignment option really isn’t–it’s a deal that demands reimbursement of any university patent and licensing expenses plus a 15% royalty for the transaction. If university administrators wanted to prevent this third option, a simple practice would be to run up a big patenting and licensing bill immediately, before the “inventors” have time to elect this third option. As one university patent administrator on the West Coast told me once (I paraphrase), “we don’t care how much we spend on patenting, because the licensee will have to reimburse us.” That university’s patenting costs were easily 30% higher than appropriate–a feeding frenzy at the bowl of state money for the patent attorneys involved. The third option is unsettled anyway, as the policy gives the Vice Provost for Research the unilateral power to make the royalty rate even higher.
The second option is akin to the third, but instead of assigning back to the “inventors,” the university licenses to a company in which the “inventors” have a financial interest. The guidance here dances around the idea that licensing agreements with inventor startups can’t have terms different than for other licensees (as that would be, no doubt, an institutional conflict of interest as an instrument of the state relative to its licensing behavior toward other companies). But “helping the company become viable” sounds like offering favorable terms, as does the cloudy language about terms reflecting the “inventors” “increased acceptance of responsibility.” In any case, two of the three options available to “inventors” cause the university to assign or license the property back to the “inventors.” The effect of the assignment and transfer is for the university to spend money on the asset–filing patent applications, essentially–and harvesting a financial return as a matter of contracting.
For option 2, licensing to the inventors’ startup, the policy authorizes the university to withhold any royalty payments to the inventors. “The inventor’s potential for revenue under the University’s standard distribution plan outlined above is exchanged for the potential rewards from the license or assignment.” If the university ever followed this procedure with an invention made with federal support, the university would have violated the Bayh-Dole Act, which requires federal agencies to require the university to share licensing proceeds after expenses with inventors. It is not the inventors’ fault that the university puts them in a position to benefit from both a university royalty stream and whatever financial stake they may have in a startup taking a license from the university. And the policy does not begin to comprehend the situation. An inventor might choose option 1, and have the university license the invention to a company, which then brings the inventor into a financial relationship (for instance, as a consultant, a line officer, or an equity holder) and the situation would be exactly the same as in option 2, except that the inventor in option 2 is excluded from a share of the university’s licensing income.
These are the hazards of creating a program based on demanding ownership of a broad swath of intangible assets up front, spending money on those assets outright, and then trying to sort out how to deal with the consequences. If a university pursued an equitable sharing approach, then following disclosure of a property that was acquired or issued as a result of or in connection with a university-funded educational activity, the review would be to consider the extent of university support. Then if the “inventors” wished to pursue commercial exploitation, the university–by policy–would assert a financial interest in any proceeds, rather than take ownership, spend money, and then seek to base the value of its claim on how badly the inventors want the property back, or how well funded their startup is, or how much the university has gone off and spent. That is, the entire apparatus of review for commercial potential, patenting, negotiating, and licensing management would only take place where the inventors requested university involvement and had made the choice not to pursue their own commercialization–and perhaps had chosen to publish the results for use without requiring a commercial product first be created.
The 1987 Policy on Intellectual Properties assumes assignment of patents and copyrights will be requested but provides no guidance on when or how. Bylaw 3.10 draws a narrow scope on the property rights that the university is to have the right to claim–patents and copyrights that are acquired or issue as a result of or in connection with specially funded educational activities. The policy statement, however, can be read as if anticipating that Bylaw 3.10 will be read to mean that the university claims ownership of all faculty work, and that these various annotations and options are there to relax the university’s grip where it would be contested, such as over academic freedom or entrepreneurial faculty who are ready to start companies and don’t need the facilitation of a university office to do so.
The IP practice at the University of Michigan is what it is. But the policy apparatus is a mess. Given the narrow authority of Bylaw 3.10, the unparseable nature of the technology transfer policy, the garbled mess of work made for hire by the copyright policy, and the further indefinite and garbled present assignment required on the deceptively titled Supplemental Appointment Information form, it would appear that the university has no basis to claim that its policies form a contract that requires assignment of inventions (or anything else). Policies work best when they constrain the institution, and therefore constrain the officers of the institution who are authorized to take actions on behalf of the institution. Here, at the University of Michigan, the policy is used as a threat. That faculty or anyone else actually assign inventions to the university must be the result of their voluntary choice and not as constrained by either contract or policy. It may be that they choose to assign because they want the benefits of the technology transfer program–filing patent applications and paying for them, representing the invention in licensing negotiations, and handling the money. All that is possible and good. But it may well also be that some faculty assign because they don’t want the hassle, the threats to their jobs and reputation, the cost of battling an unreasonable administrative claim on the meaning of a policy’s words, given that the administration has a nearly unlimited legal budget to defend its reading, as contestable as it is. As the late Thomas Roach–for fifteen years a regent at the University of Michigan–put it in 1983 in the context of a debate on another aspect of research policy, “It appears the proposed policy was supported by the faculty because it could have been worse.”
If the reason faculty opt to participate in the technology transfer program is that it otherwise “might be worse” for them, then we have a better understanding of what the administration means by “facilitating” their choices. Make the policy a mess, threaten to make it even worse, and the faculty will come around. Such a way to inspire innovation from faculty inquiry and initiative!