We have worked through the set up to a review article on the Bayh-Dole Act from the perspective of a couple of university licensing office practitioners. It’s a fascinating exercise in repeating the conventional vocabulary of university licensing offices but ignoring the Act itself in setting up what needs to change. Clearly university licensing practices are (in general, but not everywhere) a mess, and Bayh-Dole has created the opportunity for that mess to grow. Changing Bayh-Dole–tweaking the law as it twerks–won’t do much to get at that university licensing practice, which is now reinforced by new university patent and research policies.
But our authors Boettiger and Bennett have some suggestions for changes in Bayh-Dole, and these come despite the efforts of stalwart Bayh-Dole advocates to control the changes to the law. While it was, apparently, according to Bayh-Dole advocates, good to make secret all reports of subject invention use, and to remove the restrictions on exclusive licensing, and to remove the conflict of interest provisions for nonprofit assignments of subject inventions to invention management organizations, and would make it acceptable if patent positions “encumber” future discovery and research–just so long as the encumberment isn’t “undue”–most anything that would actually promote the use of subject inventions has been resisted, if not openly mocked by AUTM and other organizations.
Let’s look then at these suggestions for changes to Bayh-Dole. There are four. Let’s take them one at a time.
The need for a practical research exemption to permit the ongoing use of technologies developed with public funds for noncommercial research
The authors appear to think that at some point federal patent law contained a “research exception.” There was a practice arising in dicta concerning such a thing, but never in patent law. The authors cite the Madey v Duke decision as finding that “any university research may be viewed as advancing the business interests of the institution and thus of a fundamentally commercial, rather than philosophical, nature.” If a university has a technology transfer program that insists on ownership of inventions and insists that the purpose of such ownership is commercialization for profit, then a court might well indeed feel free to view research as business and therefore not free from claims of patent infringement. But this finding arises from how a university manages research results, not from something intrinsic in research or patent law. Universities brought the Madey decision on themselves by adopting a patent management model that makes such a decision reasonable.
[University ownership of inventions with an insistence on “commercialization” is the Jonah on the ship. Given the court’s observation regarding university “commercial” interest, even if there were a new research exemption in patent law (including Bayh-Dole) for “noncommercial” research, it would not apply to universities that owned inventions and sought to commercialize them or make money from them. You see why. So it is fantasy for university administrators to insist on ownership and commercialization of inventions and then beg to be released from obligations to other patent owners.
Take it a step further. If university administrators really wanted a research exemption for their faculty and students, then they would grant that research exemption to all for every invention the university owned. This never happens. It’s not made formal policy. It is not the subject of a general public patent license that is attached immediately to every invention a university owns. At best, one finds a grudging reservation of rights in exclusive licenses for “educational or non-commercial use”–but that is a reservation of rights, not a grant of rights, and the reservation takes place only when an exclusive license is granted and never before, and then only for those inventions for which an exclusive license is granted. University administrators don’t deserve a research exemption–their own actions betray any standing to argue the point.
And one more thing. A research exemption, as it were, makes little sense for technology transfer. What’s the point of having the freedom to conduct research using an invention patented at another university, if there’s nothing that can be done with the result of that research but for the permission of the patent-holding university? And that university, as standard practice, is looking for an exclusive license. So why would anyone think that this university would grant anyone else a non-exclusive license to practice anything–inventive or otherwise–that comes about from such “exempt” research use? One is merely conducting research that inures to the benefit of another university as patent owner, or if the university has licensed exclusively, to the benefit of that university and its commercial partner (if not assignee of the invention in question). The only research that makes any sense in that situation is research that seeks to design around, undermine, disprove, or make obsolete the patented invention that is subject to the exemption.
You do see the problem. To exempt the research use, one is encouraging others to ruin the patented invention–no wonder university administrators won’t grant a general research exemption–if they hold onto their mania default for exclusive licensing (the monopoly meme hysteria), then granting a general research exemption won’t happen. When at the conference that produced the Nine Points to Consider document, I proposed just such a general exemption, not just for universities and not just for nonprofits but for all who might use a university-owned invention. The licensing administrators of multiple universities rejected the idea out of hand. They wouldn’t even discuss it. Dark lords.
It is not that a research exemption undermines “technology transfer.” It is that a default exclusive license on offer undermines “technology transfer.” Oh, yes, there is that seductive rationale involving the monopoly meme–what is available to all will be used by none–but it is just a stupid meme, not a general truth. If it were true, then no one would ask for a research exemption unless they wanted to be the free helpers of some monopolist interest. And that would put a pretty turn on research in the public interest, wouldn’t it? All technology transfer would have to go through whatever monopolist interest had secured rights to any university invention that had been licensed–or wouldn’t happen at all. It is non-exclusive licensing–fair, reasonable, non-discriminatory, non-exclusive licensing–that enables university technology transfer generally, at least technology transfer that must find a way to survive in the presence of patents.
If university research gets to some public interest result before companies and speculators do, then university administrators, if they were to have any helpful role, would make patents available to all–for research use, a general license; for commercial making and using, a standard deal; for sale, a licensee-exercised option on standard terms. The role of a university getting involved in research discoveries is to expand the early commons against the interests of patent speculators and monopolist interests. Exclusive positions should come later, after a technology has accumulated and become a platform; exclusive positions should involve non-essential claims, not claims essential to the nascent technology. Thus, a better manufacturing process or a follow-on application, not anything core. When a university seeks to derive its income from dealing with a speculator–that the value of the patent is in excluding researchers at other universities from applying for grants or being involved with technology transfer themselves–the university engages in a form of autoimmune disorder, an administrative lupus that attacks university technology transfer while all the while putting on the appearance of being university technology transfer.
If you have followed the argument so far, then consider the problem created by Bayh-Dole. Bayh-Dole permits each federal contractor to preempt the public purposes for which any given allocation of research funding has been made and substitute its own institutional self-interest. In doing so, Bayh-Dole enables the fragmentation of any cumulative technology before it has a chance to accumulate. Carbon nanotubes got split up among scores of universities, each holding on to its petty bit of patent right, ensuring nothing much could happen for twenty years. But the research funding side of the federal apparatus continued on as if nothing had happened–spreads the work around, justifies it as in the public interest as if the results would be published and inventions shared. Mindless. Useless. Living in a fantasy land without accountability. Same for the university researchers taking the money. It doesn’t have to matter to them whether the results are patented–so long as they can publish and get the next grant, they don’t have to care. In many universities–especially public universities–now, there are conflict of interest policies that announce that it is unethical for a university inventor to care. To care would mean attempting to influence a contracting involving the university (if not the state itself) for a personal reason.
Unless a university commits to non-exclusive licensing of any invention it claims from the moment the university claims it, the whole idea of university technology transfer is blown up and replaced with an administrative lupus–an attack on technology transfer for the money or out of the seductive delight of serving the monopoly meme.]
The Bayh-Dole Act requires the grant of a nonexclusive license to the government and so provides an effective research exemption for federal government entities to practice the invention, but this license does not extend to universities or nonprofit research institutions.
The grant to the government in the standard patent rights clause is not research-only. It is a general grant to the government to “practice” the subject invention: “to practice or have practiced for or on behalf of the United States” (37 CFR 401.14(b)). There is no definition for “practice” but one would assume that means to make, have made, and use (and permit others to use as necessary for government purposes). Previous definitions of “practice” in federal patent policy also include “sold” and “have sold” but there’s no reference to “import.” Certainly nothing would prevent the federal government from distributing product without a sale–by gift, say, or loan or bailment–so long as the action is within government authority.
At any rate, there is no restriction to research in government licenses to subject inventions. Clearly, the federal government has a general license to practice any subject invention. That the government operates without a profit motive and without the need to be in the business of selling product (“commercial”) does not mean that the government’s license is for “noncommercial” uses only. There is no such limitation, and clearly the federal government could permit commercial concerns, operating under federal contracts if necessary, to practice a subject invention on behalf of the federal government. Thus, while the federal government’s use of a given subject invention may be “non-commercial,” the federal government has the right under its license to commission others who may well be commercial to practice the invention.
The authors therefore aren’t accurate regarding the scope of the federal license. The federal government can commission universities and nonprofits to practice a subject invention, if for federal purposes, just as easily as it might for-profit concerns. Those federal purposes could be as generic as “to conduct research to expand scientific knowledge.” At any time the federal government could authorize others to practice any subject invention. In the case of generic drug manufacturers, the federal government could authorize the making of a patented drug (assuming that no non-federal patents were required) and could pay the manufacturer to distribute the drug at no charge. There would be no sale of product. That more than anything would force down the prices of monopoly prescription drugs–the price over time to the government could be no higher than the cost to the government to bother to commission the manufacture and distribution of its own generic version.
[One more point, here. The license granted under Bayh-Dole to the government is broad in another way–not just in terms of “to practice.” Here’s the license grant as implemented in the standard patent rights clause at 37 CFR 401.14:
the Federal government shall have a nonexclusive, nontransferable, irrevocable, paid-up license to practice or have practiced for or on behalf of the United States the subject invention throughout the world.
Here’s the text of Bayh-Dole, proper (35 USC 202(c)(4)):
the Federal agency shall have a nonexclusive, nontransferrable, irrevocable, paid-up license to practice or have practiced for or on behalf of the United States any subject invention throughout the world
The scope issue is not what entity within federal government receives the license, but the extant of the scope–“for or on behalf of the United States” is not merely “the Federal government” or “the Federal agency.” In the Kennedy executive branch patent policy, and replicated in the Institutional Patent Agreement program, “Government” is defined to include “any agency thereof, state, or domestic municipal government). The Nixon revisions to this policy break out the various elements: “the Government of the United States (including any Government agency) and the States and domestic municipal governments,” with the provisio that the federal government might suspend the grant to States and municipal governments if decides to do so “in the public interest.”
When Bayh-Dole extends the government’s license to “on or behalf of the United States,” the grant is not restricted to the federal government only–the grant is expressly on behalf of the united States: both the federal government and the united States. Thus, Boettiger and Bennett already have the scope of license they desire for public universities–there is no need for the federal government to grant such a license–it has already obtained that license on behalf of the states and their various agencies, including universities. What’s missing is that universities have the intelligence and courage to stand up to claims of infringement–coming from nonprofits, universities, and their loser proxy licensees.]
The lack of a research exemption has created the unusual situation where a university invention, if licensed exclusively, may be unavailable for use in fundamental, noncommercial research, even in the very laboratory where it was made.
The problem is a big one, but it has nothing to do with Bayh-Dole and everything to do with university licensing practices. A university may reserve in an exclusive license the right to license an invention for research uses. There actually are multiple such uses–evaluation of an invention (to review published claims, for instance); research on an invention (to determine how it works); research with an invention (to conduct research using the invention); and research involving an invention (to consider applications and improvements and extensions of the invention). Universities could authorize any and all of these uses up front. They could put a legend in every patent application announcing a public license. They could do these things, but most don’t, and won’t. They reserve a right for themselves, and maybe for non-commercial university uses, but won’t go even so far as to license non-profits.
There are more basic problems to a non-commercial research only license. A non-commercial only license is discriminatory. Why should the purpose of the research have anything to do with the exception? Any research use ought to be permitted. At best, one might restrict research use to those uses that publish the results. Don’t want to publish, then get a license for proprietary research. There’s nothing that Bayh-Dole needs to do to create such a “research exception.” Universities could implement immediately. They could make their public grant of rights conditional: the grant only applies to organizations that make the same grant across their entire patent portfolio. Companies that don’t like that can take a license on other terms, if available. But Bayh-Dole is part of federal patent law, so adding a research exception is not one of changing the licensing terms but of limiting the property right in a patent on a subject invention. That’s something I proposed recently in a series on a new Dole/Bayh Act, one for the 21st century. One that knows now what folks didn’t want anyone to know back then.
Even deeper into the basic problems is that an infringement case depends on a finding of actual damages. If a research use is putt-putting along, there’s no way that a patent owner can sue and come out ahead. At best, one can get the research use to stop–it costs real money to defend an infringement suit, and that is often sufficient threat to send university legal counsel running for the gun cabinet to shut down the infringing lab. And for public universities, they can’t rightly be sued in federal court for infringement unless they permit themselves to be sued. That makes it difficult to bring suit in the first place.
Finally consider the purpose of a research-only use. For research tools, that’s enough. There’s no need to sell what one uses–just use it and get on with discovery. But for innovation research–research on, research about–the primary reason to research is to design around, unless one can get ready access to a license that permits broader use or sale. Think about it–even at a university–why do research on something patented elsewhere if no-one could use anything that practice the invention without the permission of the patent owner? And if that owner is a university, then getting a non-exclusive license on reasonable terms is pretty much impossible. So the primary viable thing left is to do research to figure out how to render the invention obsolete and the patent worthless. Once universities have adopted an exclusive-license-first approach, a research exception is itself mostly useless, as it provides an incentive to design around. If that was Bayh-Dole’s purpose–to use the patent system to promote the design-around of inventions made with federal support, then a research exception is just the thing. But without a broader make-use commons, a research exception is next to worthless, if not counter to federal patent policy objectives.
the lack of a systemic approach to provide broad access to publicly funded upstream inventions and research tools
The research tools issue is broader than patent rights, and so moves beyond Bayh-Dole into copyright, data, and the like. NIH’s guidelines for research tools can’t step over into patent territory without running afoul of Bayh-Dole’s requirements for declaring exceptional circumstances in order to compel non-exclusive licensing. Essentially, the NIH guidelines on tools attempts to recover what Bayh-Dole destroyed when it made patent matters arbitrary and independent of federal agency purposes. Proposing a change to Bayh-Dole at this point amounts to repudiating the law.
Our authors propose a government-wide “voluntary” adoption of the NIH tools guidelines. This hasn’t happened, and really it’s quite meaningless. As a “voluntary” thing, it is easily ignored. Where patents are involved, university administrators are happy to ignore guidelines when it serves their interest. The NIH tools requirements include a model material transfer agreement that creates all sorts of demands and guidance on exclusive licensing that has little meaning:
Exclusive licenses for research tools (where no further research and development is needed to realize the invention’s usefulness as a tool) should generally be avoided except in cases where the licensee undertakes to make the research tool widely available to researchers through unrestricted sale, or the licensor retains rights to make the research tool widely available. When an exclusive license is necessary to promote investment in commercial applications of a subject invention that is also a research tool, the Recipient should ordinarily limit the exclusive license to the commercial field of use, retaining rights regarding use and distribution as a research tool.
“Should generally be avoided except…” where the licensing is going to sell something or the licensor can make the tool available–and so exactly what would the exclusivity entail, if not making and using and selling? Would it be that the licensee has the exclusive right to sell and the university gives the tool away? I’m not following such guidance when it pertains to patents.
But read the NIH stipulation:
Under the Bayh-Dole Act and its implementing regulations, agreements to acquire materials for use in NIH-funded projects cannot require that title to resulting inventions be assigned to the provider. For this reason, definitions of ‘‘materials’’ that include all derivatives or modifications are unacceptable. Other unacceptable variations include definitions of ‘‘materials’’ that include any improvements, or any other materials that could not have been made without the provided material.
This is not accurate. Even before Stanford v Roche, it is not accurate. Bayh-Dole pertains only to patentable inventions that are made within the scope of a federal funding agreement. Thus, any invention that has not been specified within the scope of a funding agreement, even if closely related, is not a subject invention under Bayh-Dole. It might be the case that a federal contractor cannot agree to assign rights to an improvement or derivative of a material that is the subject of the “planned and committed” activities of a research project (“We propose to improve on this material that someone else owns and we will have to license in”) but then the owner of the material has no obligation to provide the material and accept such requirement. If resulting materials are not patentable, then Bayh-Dole doesn’t apply.
Further, even if the resulting materials are patentable, then the contractor can agree to assign invention rights in resulting materials to the provider, and can agree to deliver the changed materials back to the provider and agree not to otherwise distribute them, subject to a non-exclusive license in the practice of the invention to the federal government. Without actual possession of the altered material, it’s likely that the government gains no benefit from having a non-exclusive license to practice the material it doesn’t have.
After Stanford v Roche, it’s not even clear that a university is prevented from assigning subject inventions, provided that the university does not obtain ownership of those inventions in the first place. There’s no obligation in Bayh-Dole for assignment of inventions to the research host (and still won’t be after NIST tries to make the standard patent rights clause require assignment). If the university does not acquire title, then the invention is not “subject” with regard to the university–and not “subject” at all unless the university has implemented the written agreement at (f)(2) of the standard patent rights clause, and I’ve yet to find a university that has done so. Since Bayh-Dole applies only to patentable “inventions of the contractor” and federal agencies don’t enforce (f)(2) of the standard patent rights clause, there’s nothing that makes patent rights in modifications in received improvements “subject” under Bayh-Dole. It’s a strange outcome, but Bayh-Dole is a do WTF you want sort of law. Thus, it shows up as powerful strange when either a federal agency or a university gets all tipsy about complying with Bayh-Dole–agencies don’t enforce the law or standard patent rights clause, and universities willfully ignore what they don’t want to bother with.
Making voluntary guidelines government-wide is an empty gesture. Universities do not need federal agency regulations to implement sharing policies. Clearly, they won’t when given the change. And making the guidelines mandatory all but repudiates Bayh-Dole and returns federal agencies to practices under the Kennedy patent policy–contractors should not have exclusive rights to inventions except when they already have a commercial position or when they make a case better than the federal default of dedication to the public or non-exclusive licensing, royalty free.
the need for reform to address the ‘anticommons’ effects that result from fragmented ownership of publicly funded research inventions
Here our authors are onto one of the primary outcomes of the Bayh-Dole Act. Bayh-Dole did not, as our authors claim, establish “incentives for universities to develop independent technology transfer programs.” There’s absolutely nothing in Bayh-Dole that addresses whether a university develops its own licensing program, uses an affiliated research foundation’s program, or an independent regional or national invention management organization. And of course a university can stay out of all of this by refusing to accept title to inventions made with federal support–all it need do is comply with the written agreement to protect the government’s interest (the (f)(2) agreement), provide education on the importance of reporting inventions, designate personnel responsible for patent matters, and report inventions to the government.
The fragmentation of rights arises from university practices, not from Bayh-Dole, and from federal non-enforcement of Bayh-Dole. Without timely licensing, every patent held by a university is subject to a finding of nonuse and non-availability. The march-in provisions should be used hundreds of times per year per university, forcing inventions into the federal patent commons (if already patented) and to the public domain otherwise. Under the Kennedy patent policy, three years from patent issuance was sufficient time to find an exclusive licensee or unclench the fist and make the invention available non-exclusively. The IPA program ignored the three year limit and made it three years of exclusivity from first commercial sale or eight years total exclusivity, whichever came first. Bayh-Dole originally limited exclusive licenses no longer than five years, but that was quickly removed by amendment four years later. Wherever there has been an opportunity for federal restrictions on the term of monopoly for subject inventions, university patent administrators have worked to remove those requirements. That Congress has gone along with the changes means that fragmentation is an intended result of federal patent policy. Universities would have to work against this policy in order to implement sharing practices–non-exclusive licensing, dedication to commons, and the like.
There are remedies to fragmentation. The first is to restore the five-year limit on exclusive licenses that was in the original Bayh-Dole Act. The second is to restore the three-year limit on granting exclusive licenses. Get a license done and make it perform, or offer rights non-exclusively, or let the government do that.
Commons–in various forms–play an important role in the diffusion and adoption of new technology. A patent system can be used to support diffusion as well as to position patent owners for maximum income–but it’s up to university administrators to decide the balance. For research intended to advance scientific knowledge or address matters of public welfare, that university administrators choose, repeatedly, to try to make more money is difficult to accept, especially when it is apparent that they do not make more money but they do sequester and fragment the rights to many research inventions.
The examples of patent commons offered by our authors–PIPIRA and BiOS–are both fine examples. We developed similar practices for software, data, digital media, and the like at the University of Washington and University of California. In those cases, however, the practices were attacked by both faculty and especially “boosters” of the university for failing to seek what was claimed as “maximal” financial reward by offering exclusive licenses.
Our authors cannot figure out how Bayh-Dole might be modified to limit the adverse effects of patent fragmentation:
There is no readily apparent amendment to the Bayh-Dole Act that might provide incentives for such approaches. However, a starting point may be to require, as a condition of some of these large-scale research projects, the development of multi-institutional strategies of IP management that go beyond the normal acknowledgement of joint inventorship.
Joint inventorship results in the dreaded “interinstitutional agreement” negotiation, which can run to pages and pages on who files, who pays for filing, who has a say in filing, who markets, who licenses, with what right of review, who can enforce, and how costs are accounted, and how revenues are shared. But there are ways–see, for instance, either of the two forms of alternative Dole/Bayh laws I’ve proposed. In one, there is a broad commons for making and using. In the other, protections to get institutions out of predatory mode. Let inventors decide–and make their case to agencies and to the organizations that would manage their patents as agents. Academic inventors show a propensity to create commons, to recognize the collaborative nature of their work, and bear the consequences of defecting on that collaboration in ways that institutional agents can ignore.
At the core, however, the creation of anticommons, of fragmentation of rights works against using the patent system to promote use of subject inventions. To change Bayh-Dole at this point is to repudiate it and return to the Kennedy-Nixon patent policies.