[9/20/18–See this discussion of Rev Proc 2007-47 by the University Industry Demonstration Project–“Public Policy Regarding Industry-Sponsored Research at U.S. Universities.“]
Many public universities use tax-free bonds to construct their research buildings, and when they do, they run afoul of the Tax Reform Act of 1986. That law places a strange set of restrictions on research conducted in these buildings and on the disposition of inventions and “other products” produced in that research. The problem is what are called, generally, “private business use” issues. If a company sponsors research, and expects deliverables, then the whole deal may be classified as private business use. Exceed the private use limit (typically around 10% of the total bond issue) and the tax-free status of the bonds may be revoked. Although the TRA86 is heavily amended with exceptions (so that, for instance, public schools can have vending machines run by private industry), university research administrators have not had any interest in seeking an exception for industry-sponsored research. Why? They want the private business use issues–at least the patent licensing administrators do. Those private business use issues go hand in hand with the misrepresentation of Bayh-Dole.
The IRS regulations on private use of tax-free bonds for research are explained in a “revenue procedure” document, Rev Proc 2007-47 (previously 97-14), which has this charming provision:
If a research agreement is described in either section 6.02 or 6.03 of this revenue procedure, the research agreement itself does not result in private business use.
That is–there is a “safe harbor” from a finding of private business use if:
.02 Corporate-sponsored research. A research agreement relating to property used for basic research supported or sponsored by a sponsor is described in this section 6.02 if any license or other use of resulting technology by the sponsor is permitted only on the same terms as the recipient would permit that use by any unrelated, non-sponsoring party (that is, the sponsor must pay a competitive price for its use), and the price paid for that use must be determined at the time the license or other resulting technology is available for use. Although the recipient need not permit persons other than the sponsor to use any license or other resulting technology, the price paid by the sponsor must be no less than the price that would be paid by any non-sponsoring party for those same rights.
This provision sounds like it was written after consulting university technology licensing managers–“basic” research (why not “fundamental” research?), “competitive price” (what does competition mean here, except that one cannot fund research expecting that funding is itself also payment for the deliverables and that one cannot make the results “open”), and the “price” determined after there’s been an invention disclosure (so, no negotiations at the time of the research grant, even if a sponsor specifies what is to be developed). And the IRS regulations put the federal government in this same category as a non-qualified user, so something must be done with regard to Bayh-Dole, which does set the terms for government license up-front–the federal government does not have to pay a “competitive price” for any inventions made in a project receiving federal support. One effect of Bayh-Dole’s license is to exempt the federal government from being exposed to a claim for “reasonable and entire compensation” for the manufacture and use of an invention subject to patent under 28 USC 1498. One might think, then, that “competitive price” in Rev Proc 47-2007 must mean “reasonable and entire compensation”–the amount that the owner of a patent would be expected to be awarded were it to pursue a claim of infringement against the sponsor of the research.
This one poorly conceived provision in the tax regulations has done more damage than anything I can think of to university-industry collaborations. More even than ITAR and EAR (which have done their own substantial share of damage).
This is plainly inept stuff. It doesn’t get at research collaborations; it is restricted to “basic” research (what about “applied” research?); it is expansive beyond merely inventions (“technology”); it fails to consider “technology” already subject to other claims (such as open source software or technology provided by the sponsor and improved); it creates the fiction that the essence of avoiding a private business use is that the sponsor has to pay a market price for new technology–but for new technology there is generally not a market, and the main market that is likely to form is one in which it is worth paying to *keep the technology away from the sponsor* for competitive reasons. There is simply no way of ascertaining the “price that would be paid” by anyone for new technology. It’s inept. It’s stupid. It’s a few attorneys making things up in the privacy of their own regulatory cocoon.
The problem they perhaps want to get at is the use of tax-free bonds that end up creating facilities that are merely leased to commercial concerns, which then have a lower cost, tax-avoiding way of financing new construction. For-profit hospitals could do something like that, say. Universities could create research parks for companies and together they would collude to avoid taxes. While no citizen has any legal obligation to find ways to throw themselves in the way of paying taxes, it’s clearly an IRS interest in expanding rather than narrowing the scope of its taxing authority. So push the law as far as possible and carve out a tiny “safe harbor.” Leave the rest in uncertainty, and use uncertainty to scare the poor bond counsel into behaviors that force governments to use taxable bonds “just to be safe.” The upshot is that the whole tax-grubbingness limits the range of research collaborations companies and universities can have.
But research administrators just love it, as do technology transfer officers. Look at what the Rev Proc says: to avoid a finding of private business use, the university has to own, the company has pay as if it had not sponsored the work (i.e., treat the company as if it were making a donation), and it cannot even know what it will have to pay until something is created and is “available for use” and some mythical “market” has been imagined in which it competes to establish a price. University folks have used this clause to argue that they cannot even agree to make “technology” broadly available at no cost, because, well, the technology transfer office is dedicated to “commercializing” everything, and thus, setting a price at the outset of $0 would also be setting a price and it can’t be done. And, of course, to permit use of inventions will require that the university own the inventions, and that means filing patents, and that means running up legal expenses, and the company will be at least on the hook for those costs, whatever they may be. It’s sweet. And stupid.
The new Rev Proc 2007-47 revises a second provision having to do with collaborative research, and appends a strange Bayh-Dole discussion. Clearly here one can see the hand of the universities advocating for their misreading of Bayh-Dole, and they get it into the tax code. Now, post Stanford v Roche, someone has to go in undo their damage, here, too. Here’s the first of these provisions:
.03 Industry or federally-sponsored research agreements. A research agreement relating to property used pursuant to an industry or federally-sponsored research arrangement is described in this section 6.03 if the following requirements are met, taking into account the special rules set forth in section 6.04 of this revenue procedure in the case of federally sponsored research —
- A single sponsor agrees, or multiple sponsors agree, to fund governmentally performed basic research;
- The qualified user determines the research to be performed and the manner in which it is to be performed (for example, selection of the personnel to perform the research);
- Title to any patent or other product incidentally resulting from the basic research lies exclusively with the qualified user; and
- The sponsor or sponsors are entitled to no more than a nonexclusive, royalty-free license to use the product of any of that research.
This is not much of a safe harbor. It applies, as does .02, only to “basic” research. Furthermore, it requires the “qualified user” to “determine” the research to be performed. What does that mean? That the university administration has to decide what the research will be, and some company shows up to say, yeah, what a great idea, we’ll help fund? In almost all industry collaborative situations, the projects to be funded arise by mutual discussion. The university might decline to take on a project, but it isn’t going to do a project that the company doesn’t want, but still with company funding. In practice, then, it is simply not what happens. It’s a foolish condition. If the effort is to say: you cannot lease space to a company so that it decides what to do, nor can you simply assign research personnel to do what the company directs–fine, don’t be a landlord or contract research operation. But this reaches much further and interferes with how university faculty and company scientists and engineers do collaboration. It overreaches. It makes a mess of things. All done in the name of officious tax collecting. I feel a George Harrison lyric coming on.
This provision for collaborative research adds the requirement that ownership of everything (“patent or other product incidentally arising”) must be exclusively with the university. This is also sloppy drafting. The actual issue is whether the sponsor can own anything in a collaborative project. In .02, there is no requirement that the qualified user own anything, just that the licensing is done with the goofy fiction that the sponsor hasn’t been sponsoring. Here, however, .03 distinguishes between the “corporate-sponsored” research of .02 and “industry-sponsored” research. What? What is the difference between “corporate” sponsored and “industry” sponsored? Nothing. It’s just sloppy drafting. Section .03 used to be about “collaborative” research, back in Rev Proc 97-14, but now that’s gone. The “exclusive” ownership in a collaborative agreement would mean that the sponsor could not claim joint ownership, say, in new technology merely by collaborating in the research. It need not say anything about the disposition of the ownership by those involved other than the sponsor–researchers, say, or the university, or a research foundation, or a nonprofit that was pooling rights to coordinate research. Here, “exclusive” now appears to read that the university has to take title from the poor researchers. It also appears to mean that even in the case of joint inventions or joint works of authorship, where federal law would provide for joint ownership, it’s not within the safe harbor for there to be joint ownership.
We may add that the safe harbor only applies to a limited license to “use” the “product” of the research. Note that patent rights typically consist of the rights to make, have made, use, sell, offer for sale, and import. Thus, it would appear that any license that grants more than the right to “use” inventions or data or software (open source!) is outside the safe harbor. This is a really tiny, useless safe harbor.
It gets worse. Look at .04, where the universities appear to have gotten into “fixing” the Rev Proc to take care of the problem that the federal government as a sponsor fell outside the safe harbor of Rev Proc 97-14:
.04 Federal Government rights under the Bayh-Dole Act. In applying the operating guidelines on industry and federally-sponsored research agreements under section 6.03 of this revenue procedure to federally sponsored research, the rights of the Federal Government and its agencies mandated by the Bayh-Dole Act will not cause a research agreement to fail to meet the requirements of section 6.03, provided that the requirements of sections 6.03(2), and (3) are met, and the license granted to any party other than the qualified user to use the product of the research is no more than a nonexclusive, royalty-free license.
How bad can it get? All the Rev Proc had to say was that it isn’t a private business use for the federal government to have rights in inventions made in tax-free, bond-supported facilities. Is that too difficult? No. But apparently it’s not possible. Instead, we have the IRS arguing that the federal government being (potentially) a private business user for facilities for which it has waived tax obligations for the bonds that built them. Further, the safe harbor appears to be framed so that the university has to own the resulting “patent or other product.” That’s not what the Bayh-Dole Act provides for–it’s just one possibility among a number of options. Bayh-Dole (35 USC 202(b)) also provides for the federal government to own inventions made in federally sponsored research, such as in declarations of “exceptional circumstances.” The effect of Rev Proc 47-2007 is to say that if the federal government owns such inventions, the research is outside the safe harbor for private use and could jeopardize the tax-free standing of the bonds.
But that’s not the worst of it. The safe harbor doesn’t operate if “the license granted to any party other than the qualified user to use the product is no more than a nonexclusive, royalty-free license.” It’s not clear why “other than the qualified user” is here at all, since the qualified user has to own the property per 6.02(3), so why would the owner need a license at all? But the bigger wonderment is that no other organization can receive other than a nonexclusive, royalty-free license.
Let’s reset. The federal government funds research. The university obtains ownership of an invention. The government per Bayh-Dole allows the university to retain that title, provided the university uses the patent system to promote practical application and free competition. But lo! the IRS says, it’s not within the safe harbor if the university grants licenses other than for “use” or other than “nonexclusive” and “royalty-free.” This isn’t to the sponsors of the research but to “any other party.” Any “commercialization” deal that’s exclusive or royalty-bearing is outside the scope of this safe harbor. Bond counsel should be suffering a massive reflux event at this point. This little sentence throws out the whole scheme by which universities expect to license inventions made with federal support to industry and make some money on the deal. Why would this be? Why would it be a private business use for the university to obtain property created in research funded by the federal government and license it to other parties for a royalty? or exclusively? And it would not be a private business use if the same work were sponsored by a “corporate” sponsor, per section .02? No, it’s got to be sloppy drafting or a fundamental failure to comprehend what is going on–or both.
But wait, there’s more. The Rev Proc offers an explanation:
Thus, to illustrate, the existence of march-in rights or other special rights of the Federal Government or the sponsoring Federal agency mandated by the Bayh-Dole Act will not cause a research agreement to fail to meet the requirements of section 6.03 of this revenue procedure, provided that the qualified user determines the subject and manner of the research in accordance with section 6.03(2), the qualified user retains exclusive title to any patent or other product of the research in accordance with section 6.03(3), and the nature of any license granted to the Federal Government or the sponsoring Federal agency (or to any third party nongovernmental person) to use the product of the research is no more than a nonexclusive, royalty-free license.
The profundity here is breathtaking. Bayh-Dole does not require a university to take ownership of any invention made in federally supported projects. A university does not own outright any such invention unless it has created the conditions that would give rise to a finding of equitable title–for instance, that the university assigned employees to invent or to engage in experimentation, or that those doing the inventing are officers of the university and have an implied duty to assign. Section 6.02(3)–requiring title with the qualified user–amounts to an implied obligation (to meet the IRS safe harbor) for a university to demand ownership of all inventions made in a facility supported by tax-free bonds. In this scenario, all university faculty inventors become non-qualified users. This implied obligation to take ownership to stay within a strange safe harbor runs against the idea that a university provides resources for faculty to pursue their own scholarly interests, formally assured by university policy of the freedom of research and freedom to publish.
The Bayh-Dole Act provides that an agency may obtain ownership of inventions (if provided for by statutes otherwise preempted by Bayh-Dole) when a university (or other contractor) does not obtain ownership of an invention or having obtained title elects not to retain title to the invention, or screws up in any number of ways in handling the invention and the federal agency has the right to request title. If the agency requests title, Bayh-Dole’s standard patent rights clause leaves the university (but for non-disclosure) with a non-exclusive license with limited rights to sublicense (see 37 CFR 401.14(a)(e)). If the agency does obtain title, however, this blows up the safe harbor set out, badly, in section 0.3. Further, if the agency exercises march-in rights, it can force a license to a “responsible applicant” on “terms that are reasonable under the circumstances.” The license can be nonexclusive, partially exclusive, or exclusive. If the agency forces anything but a royalty-free nonexclusive license, however, its actions cause the university (the qualified user) to fall outside the safe harbor. The explanation reads: the existence of a march-in right doesn’t cause the use to fall outside the safe harbor, but if the march-in right is exercised in any of a number of ways outside the control of the university, then the use falls outside the safe harbor!
It is clear that the IRS has no working understanding of Bayh-Dole. It doesn’t help that a lot of university patent administrators also have shown they don’t know Bayh-Dole, which has become especially clear in light of what they have put in writing in their briefs to the Supreme Court in Stanford v. Roche. And it doesn’t help that sponsored research administrators think it’s really keen that this safe harbor business forces ownership of research products to the university and prevents offering licenses for any pre-established value upfront to sponsors–as if royalty-free licenses to “use” are really keen.
Malcolm Gladwell in Outliers writes that it takes about seven to go wrong, on average, to cause the crash of a commercial aircraft. Here, it appears that all we need is maybe three total screw ups to get it really wrong. As if it weren’t enough for university bond counsel to worry about, keep in mind that universities allow faculty to own copyrights in their scholarly works, which potentially could be “products” of sponsored research. Further, universities routinely do not claim title to inventions made in clinical trial work where the protocol has been provided by the sponsor. For that matter, clinical trials are not “basic” research. And universities do a great business in clinical trials–and should–as their efforts are an important contribution to the public safety of new drugs and other medical interventions. A few universities allow industrial sponsors to own inventions and software arising in research they fund. They might have to make sure that research doesn’t happen very often in tax-free bond financed facilities. A number of universities assign all inventions to their affiliated research foundations. This would appear to push them out of the safe harbor set out in the Rev Proc, too.
The point is–this whole safe harbor thing is totally screwed up–the federal government’s use of sponsored research results is apparently a private business use, so are clinical trials, so is any licensing of any rights to inventions sponsored by the federal government other than as royalty-free nonexclusive licenses, so is assigning inventions to a research foundation. The limited worry that businesses could scheme to get their buildings created with tax-free bonds turns into a farce of IRS opinions about basic research, federal research, the Bayh-Dole Act, ownership, licenses, “market value” and “use.” None of which makes any sense. Luckily, it does not appear any tax-free bonds have lost their tax-freeness as a result of any of this. But collaboration with industry has taken a dive. One may want to think that university culture is just so different from industry culture and therefore negotiations over research contracts are bound to be a challenge–but it’s really not that–it’s that the IRS hasn’t got a clue regarding research collaborations, has created a hugely problematic body of “guidance,” and there is a passel of university research and patent administrators who think this “guidance” is pretty keen stuff.
Thus, I posit that a material way to enhance the Bayh-Dole Act is to get the IRS out of university research altogether. Amend the Tax Reform Act of 1986 to allow all kinds of research collaboration, allow the licensing of inventions to any party on any terms that meet the objectives of Bayh-Dole, allow inventors to own (as Bayh-Dole does) and allow universities to assign inventions they come to own to research foundations and other agents. Replace the foolish language of Rev Proc 2007-47 with something that actually addresses the narrow concerns of circumventing the idea behind tax-free bonds. The government will get its taxes from economic growth arising from research collaborations, not worrying about whether universities here and there should be turning away from bond financing of new research facilities. Get the priorities straight, and we can straighten out university-industry research collaborations. Dealing with TRA86 and Rev Proc 2007-47 would be, as they say, a good start.