Let’s look at three antecedents for Bayh-Dole’s definition of “subject invention”: The Institutional Patent Agreement master, 1968, that allowed participating non-profits to end-run DWEW contracting policy and take ownership of inventions made with NIH funding (Latker said that Bayh-Dole was a codification of the IPA); the Federal Procurement Regulation, 1975, that codified the Nixon version of executive branch patent policy for contracting (Latker also worked on the FPR); and the Thornton bill (HR 8596), 1977, which proposed a “uniform” federal patent policy (Anker-Johnson said that Bayh-Dole was copped from Thornton). I’ll highlight bits for discussion.
Here’s the IPA (Article II):
(a) The term “subject invention” as used in this Agreement means any process, machine, manufacture, composition of matter or design, or any new or useful improvement thereof, and any variety of plant which is or may be patentable under the Patent Laws of the United States made in the course of or under research supported by grants and awards from the Department of Health, Education, and Welfare.
(b) The term “made” when used in relation to any invention or discovery means its conception or first actual reduction to practice.
No “of the contractor.” No plant varieties (prior to the PVPA). Uses “in the course of or under” (rather than “in the performance of”–conventional language for “as specified and what you did”). “under research” is garble. Restricted to “grants and awards” (IPA restricted to nonprofits–suggests that grants and awards are distinct, and that there may be other forms of funding not covered).
Here’s the FPR (41 CFR 1-9.107-5(a)(1)):
“Subject Invention” means any invention or discovery of the Contractor conceived or first actually reduced to practice in the course of or under this contract, and includes any art, method, process, machine, manufacture, design, or composition of matter, or any new and useful improvement thereof, or any variety of plant, which is or may be patentable under the Patent Laws of the United States of America or any foreign country.
Compact, all in one definition. Adds “of the Contractor” (there is context that makes it work). No extension to PVPA plant varieties. Specific to patent laws. Adds patent laws of foreign countries. Adds “any art” to list. Uses “in the course of or under.”
Here’s Thornton (Section 511):
(e) The term “invention” means any invention or discovery and includes any art, method, process, machine, manufacture, design, or composition of matter, or any new and useful improvement thereof, or any variety of plant, which is or may be patentable or otherwise protectable under the laws of the United States.
(f) The term “subject invention” means any invention or discovery of the contractor conceived or first actually reduced to practice in the course of or under a contract.
(c) The term “contract” means any contract, grant, or agreement entered into between any Federal agency and any person for the performance of experimental, developmental, or research work substantially funded by the Federal Government. Such term includes any assignment, substitution of parties, or subcontract of any type entered into for the performance of experimental, developmental, or research work under a contract.
(d) The term “contractor” means any person (as defined in section 1 of title 1, United States Code) that is a party to the contract.
Longer cascade of dependent definitions (has to be general for all contracts, not any specific agreement). Has “of the contractor.” Retains “in the course of or under.” “Laws of United States”–not restricted to patent law (adds “otherwise protectable”–then might implicate copyright or trade secret). Contractor is “any person” with 1 USC 1 definition. Contract is broad (not limited as in Bayh-Dole definition of funding agreement).
What do we make of these antecedents? First, that Bayh-Dole does owe much here to Thornton. In particular, Thornton’s use of “otherwise protectable”–which is tied to any laws of the United States, not just patent laws, ends up as garble in Bayh-Dole, which is part of federal patent law: “patentable or otherwise protectable” by federal patent law. Someone was in a hurry and didn’t get a lot of editorial help. Second, that Bayh-Dole shares Thornton’s approach to stacked, dependent definitions. Really, all this could be much easier:
“Subject invention” means anything that is or may be patentable arising in the course of or under work receiving federal support.
There are two purposes at work here: identify invention with patent law; scope such inventions to federally supported work under contract.
Neither purpose is really that simple, and the problem changes depending on whether the invention is a contract deliverable. If an invention is patented, then clearly it is patentable. That takes care of the “is” in “is or may be” patentable. But “may be” patentable is wildly unconstrained. Who gets to think the thought that something is patentable? An inventor? A university administrator? A patent attorney? A federal contract officer? Is it a matter of going, “we have to assume that someone might think this thing is patentable, and once we make that assumption, then well isn’t it the case that this thing “may be” patentable in some obscure sense because we haven’t ruled out absolutely that the invention is not patentable? On that standard, just not doing a prior art search may be enough to trigger the “may be.” If you keep yourself in doubt, then it follows that “may be” continues to be possible.
There is an easier criterion: if you file a patent application, then you think the thing is patentable. You could, of course, think a thing is patentable and not prepare a patent application, or prepare a patent application and then decide the thing isn’t patentable. But if you file–there you have it. If the federal government cares only that it has rights to deliverables, then only the inventions that matter are the ones for which a contractor or a contractor’s employee-inventor obtains a patent. Everything else ends up in the public domain, so the government and all its potential contractors have open access. We are left with a residuum of those inventions that the contractor (or its inventors) don’t seek patents for but which are patentable and the government would obtain patents for itself in order to publish the inventions in the patent literature and thus make easier searches for prior art. The government generally did not license exclusively (as that would carry the implicit threat that the government would get involved in suing citizens for practicing inventions made with government support), so the government demand to own would be for the rather empty need to publish in the patent literature stuff for which the government would not exploit the exclusionary power of the patents.
It is right at this point that the rhetorical argument about patented inventions “sitting on the shelf” in government labs gets its legs. The travesty of waste, according to this argument, is that the government refuses to allow these patents to be used to exclude others, which exclusion, so the argument goes, would attract investment that otherwise would not be made in the inventions, to “develop” them for public use. If the government left inventions that contractors did not want (of the 26,000 federally owned patents in the mid-1970s, the vast majority were in this category) in the public domain, at least there would be fewer “zombie” patents that were issued just to enhance the patent literature.
That leaves the other prong of the rhetorical argument about patentable inventions that go unpatented, whether by the contractor (not suited to its business purposes), inventors (indifference, lack of resources), or government (no use for the power to exclude or financial interest in practice)–even if none of these want a patent with the power to exclude, somebody might and therefore patents ought to be obtained anyway, and held in reserve for the eventual arrival of someone who cares about the power of patents to exclude and uses that exclusion to justify spending the money to “develop” the inventions as products so that the public may benefit from the federal expenditure of public research dollars. Thus, nonprofits and the government ought to obtain patents and wait for the right somebodies to come along, and then as it were, re-issue the patents to these somebodies with the requirement that they use their new exclusivity to make commercially successful products.
Perhaps you see the problems with this line of reasoning. One of these problems then is that nonprofits and government agencies must exclude all those who would practice any given invention without exclusive rights in favor of those that demand exclusive rights, even if those that demand exclusive rights don’t show up right away. The more interest there is in practicing an invention without exclusivity, the more “value” a patent has to exclude that practice. The patent does not make the invention more attractive to develop; it makes the patent more valuable because it prevents all competing practice. In the deployment of exclusivity for things that start with a premise of public funding and open, pre-competitive access, separating out those things that would never be practiced but for a substantial private investment to “develop” the invention (because the government and publicly minded organizations and individuals decline to pay for such development) would be at the top of the list. If an invention does not need to be “developed” to be practiced (many methods, basic improvements, simple but clever devices), or if the cost to develop is not substantial and within the means of most anyone, then there’s no public need to offer the power to exclude to attract non-inventor investors. The investors that desire to be attracted would of course beg to differ–in their view, the public needs them to invest, to realize the benefit of any invention, all the more so for those that anyone could practice or develop without a patent. Not using patents when patents are possible means that the opportunity to profit from an otherwise unnecessary patent position itself is wasted.
Put it another way. In basic federal patent law practice, the inventor has the choice of whether or not to use the patent system. An inventor decides whether to seek a patent. The bureaucrats attracted to federal funding of research then insist that contractors and federal agencies should require inventors to give up this right to decide. If the contractors decide not to seek patents, then the federal agencies should seek those patents, and since the contractors have opted out and federal agencies releasing the patents is a waste, then the federal agencies should hold the patents for exclusive licensing to those presumed somebody investors who value the power of the patent to exclude all others (but for the original contractor, in the Bayh-Dole implementation, but then only in a revocable but appealable way–see the unused bits in 37 CFR 401.14(e), not authorized by Bayh-Dole but copped from the FPR).
In this work up, inventors are excluded but for their continued employment by the original contractor. Bayh-Dole gives them a place in line for the exclusive license–and that shows up as 35 USC 202(d)–as if the inventors never had assigned their invention rights in the first place, but really, what 202(d) means is that once a contractor has waived its interest and before the federal government goes off to find an exclusive licensee that’s not the contractor, the inventors can beg the federal government to let them have their invention. They will play the role of the presumptive investor, or will find that presumptive investor. Or–and this is the part the Bayh-Dole people hate but is there in the implementing regulations–the inventor then gets to decide whether to use the patent system or release the invention to the public domain.
In the arguments surrounding the Dubilier decision in 1933–in which the Supreme Court found that two federal employees, having invented new electronics using federal lab resources and time, owned their inventions too bad for the federal government–one was that a review of what inventors did when they did keep patent rights was that they sold these rights off to companies. The fall out then is: to protect the public, nonprofits and federal agencies should take possession of these rights and choose reputable companies, which (so the argument goes) they can do better than mere inventors who do not have such a strong mandate to serve the public interest.
We are at the core clockworks, here, of the rhetoric that rationalizes Bayh-Dole. Lift control of inventions from inventors–so, preempt the fundamental of the U.S. patent system. If a company contractor does not choose to exploit the patent system, then the federal government should hold the invention in favor of someone that will. As for nonprofits, if they cannot find someone to exploit the power of exclusivity, then they should hand the patent to the federal government to do so. What is fascinating about the Stanford v Roche decision is that the Supreme Court determined that Bayh-Dole does not lift control of inventions from inventors, despite scores of university lawyers and even former senator Bayh arguing otherwise. Federal agencies ought then to be precluded from making efforts to lift control of inventions from inventors.
But that’s not how NIST has gone about things–NIST rather defies the Supreme Court and adds an assignment requirement to the non-Bayh-Dole written agreement requirement at 37 CFR 401.14(f)(2). But that assignment provision is specific to subject inventions–ones a contractor must already own–and so it too fails to force inventors to give up their rights because the rights they would have to give up under NIST’s nonsense must be the ones they have already given up some other way. The disposition of inventions made in federally funded work is so unimportant that we can afford to have clowns in charge of the regulations. Well, now, I guess we ought not get too worked up about it.
Lift control from inventors, if company contractors don’t exploit patent rights, then nonprofits and federal agencies must obtain and hold those patent rights to the exclusion of all those that would practice and develop inventions without patent rights in the hope that speculative investors will one day arrive and take exclusive licenses.
In this way, so the Bayh-Dole clockwork provides, the public receives the “value” of the power of patents to exclude all others, in the form of a share retained by nonprofits and federal agencies of 1% or 2% of net sales income from the maybe as much as 4% or 5% of such inventions that are indeed developed to become commercial products. There are over 50,000 such Bayh-Dole patented inventions held by universities and their affiliates. Most unlicensed, many of those licensed undeveloped, and of the few that have become commercial products, many of those do not meet the Bayh-Dole standard for continued exclusivity because they are not made available for a price comparable to what the price would be were there competition. This holding of patents is necessary to preserve the opportunities for even a few inventions to become commercial products through this bureaucratically constructed process. The present efforts to “unleash innovation” by regulatory changes to Bayh-Dole focus on trying to make this bureaucratically constructed process somehow perform less worse than it has.