From provider to predator: University of Texas patent policy, Part 4

In 2005, BethLynn Maxwell, a patent attorney then in the Intellectual Property Section of the Office of General Counsel for the University of Texas System, published a brief article on the Bayh-Dole Act, “Twenty-Five Years After Bayh-Dole” in the Office of General Counsel newsletter.  We know more now about Bayh-Dole than this article evidences, since we have Stanford v Roche to set matters straight, but it’s useful to take a look the article to get a sense of what University of Texas System attorneys were thinking about Bayh-Dole and how such thoughts may have contributed to changes in Texas patent policy.

The general drift of this brief article is that Bayh-Dole has been wildly successful in getting university technology transferred to industry as the basis for commercial products. Bayh-Dole has done this by giving universities outright ownership of inventions made with federal support. The claim is, then, that university ownership and licensing of patents, especially exclusively, is a good thing, endorsed by federal policy and proven by all the reports of success.

The problem with this line of reasoning is that there’s little to show that Bayh-Dole has been wildly successful as advertised, and Bayh-Dole does not give universities ownership of inventions. The metrics that are published are indicators of activity and expenditure, but universities do not publish the key metrics indicated by Bayh-Dole, including for each subject invention the licensing status and the date of first commercial use or sale. Universities do not even report subject inventions separately from other patentable inventions, and do not report the licensing of patentable inventions separately from other things called inventions.

One would think that if Bayh-Dole really was that successful, people would publish the data that supports the claim, not proxy data that shows how large technology transfer offices have become, how much they spend, and how many licenses they have granted and startups they have licensed to–none of which gives any information about whether inventions made with federal support have achieved (1) practical application with (2) public benefits (3) on reasonable terms–what Bayh-Dole expects. One would also think that if a research university was going to reverse a former policy, it would do so with good data and reasoning to show why the reversal was indicated.

Let’s work through this article, then, not so much to pick at it–it’s over ten years old and written before Stanford v Roche, after all–but rather to examine the frame of mind and practice care that administrators and faculty alike might have encountered from an IP attorney at a leading American research university.

Bayh-Dole: What Metrics?

The article begins by repeating the mantra, once just a comment in The Economist, that Bayh-Dole is “probably the most inspired piece of legislation to be enacted in America over the past half-century” (“Innovation’s Golden Goose,” 2002). (Maxwell will note and dismiss out of hand criticisms of Bayh-Dole at the end of her article.) She may not have seen the article three years later in The Economist that reports all the problems (“Bayhing for blood or Doling for cash?,” 2005) :

Many scientists, economists and lawyers believe the act distorts the mission of universities, diverting them from the pursuit of basic knowledge, which is freely disseminated, to a focused search for results that have practical and industrial purposes.

…

Moreover, there is ample evidence that scientific research is being delayed, deterred or abandoned due to the presence of patents and proprietary technologies.

…

Even industry is starting to complain about a gold-digger mentality among academic administrators.

These quotes from The Economist are not so quick off the tongues and pens of advocates for Bayh-Dole.

Continuing her account of the Bayh-Dole Act, Maxwell states as fact that Bayh-Dole has “increased” technology transfer. She provides no evidence for her claim. That’s understandable. Not only is this just a newsletter piece and not a law review article, but also there is no data whatsoever available to support this position–then or now–unless all she means is that there are more “university technology transfer offices with more personnel spending more money” than before. “Technology transfer” generally means the movement of technology from one group to another–from one industry to another industry (computer science to biology, say); from one country to another (as in the developed/developing meme); or from a lab to industry (in the commercialization meme). Such movement may happen any number of ways–direct instruction, publication, transfer of technicians to new companies or cities, open source/hardware/architecture programs, development of standards, exchange of technology (cross-licensing), and reverse engineering (inspect and recreate). As technology moves, it may need to be adapted to new situations, or the social and technical worlds around it may have to change to accommodate what’s new–creating other opportunities for change.

Of course, technology transfer can also involve licensing patent rights, followed by instruction. Or, alternatively, one can learn a new technology and then discover that patents cover its use and get sued. In such a case, the technology gets transferred and then there’s a question about whether it can continue to be practiced. A number of universities are exploiting this last approach. Thus, when Apple uses (apparently) technology patented by Caltech, the technology has transferred. The question is whether Caltech will confirm the transfer with a license, for which Apple might acknowledge the priority of Caltech research, or whether Caltech will demand payment–it’s potentially infringement, after all, and a patent gives one the right to exclude (in general, though it is unclear whether Bayh-Dole gives one the right to exclude what has met the objectives of the Act–practical application with public benefit on reasonable terms. I argue Bayh-Dole exhausts a patent holder’s right to sue at that point).

To track the “level” of “technology transfer,” then, one would need to establish a baseline of flows of new technical information and prototypes from university labs to industry settings. One would have to see that these flows result in adoption of new practices in industry, not just new products covered by university-owned patents.  For all that, technology might flow from a university lab to graduate students, who when they finish their programs leave for industry where they develop something new based on their training, without any trail of university-owned patents.

Technology transfer then may involve new uses for an existing technology, the development of altered or improved technology, and even technology that obsolesces what was brought in. And technology transfer might result in the development of new products, too. But it does not have to–internal use is sufficient. So is the development of common tools or standards–none of which need rise to the level of commercial product offered for sale. To limit technology transfer to a process of “commercialization” damages the idea and creates ambiguity–there may be plenty of technology transfer in the broad sense, and much less in the sense of a patent license that comes before and is the basis for a n new commercial product developed and sold in a market.

There are arguments that universities taking patent ownership positions on faculty research have impeded technology transfer in this broad sense. The argument is that they are now much less effective in moving stuff “from the laboratory to the market place” because stuff cannot get so readily from the lab. The University of Texas intellectual property policy, for instance, forbids all publication and disclosure of most anything that an administrator could label “an intellectual creation” prior to its review for System ownership and possible “protection.” It is not at all clear–we are reasoning here–how such a bottleneck requirement makes it easier to move new information from the lab to industry. But all we have to go on are stories from the field and reasoning because Bayh-Dole makes the data on utilization exempt from public disclosure (see 35 USC 202(c)(5)). And universities suppress their own data.

For all this, nowhere in the objectives for the Act given by Congress (35 USC 200) is there anything about increasing technology transfer. Even if technology transfer takes place, Bayh-Dole is not about how much transfer happens. It is about making available a particular tool that may be used for some such transfers. Commercialization, too, is not a primary objective. Commercialization is given a mention, but only in the same phrase as “public availability” of inventions and all of that is qualified by the primary interest of the clause, that inventions are made in the US, by US companies and workers.

The Objectives of Bayh-Dole

The first stated objective of the Act is

to use the patent system to promote the utilization of inventions arising from federally supported research or development.

Use patenting to promote use–commercialization by contrast is mentioned as a matter of creating jobs, not a legislated goal for university licensing operations.

I will attempt here a subtle but critical point. The law says, “use the patent system to achieve a desired effect”; it does not say, “use the patent system, and by doing so you will achieve the desired effect.” It does not say, “first use the patent system; then try to achieve the desired effect.” The Act’s first objective is promoting the use of subject inventions, those made with federal support. The patent system, says the Act, should be used when doing so will accomplish this purpose. The law does not have as an objective that more subject inventions should be patented. Increased patent counts–more ownership claims–is not a stated objective. Higher patent counts has absolutely nothing to do with any “success” of Bayh-Dole, however proud university administrators might be about making more of something.

Thus, a focus on universities owning  subject inventions is also not an objective of the Act–though it might be an objective of administrators and their legal counsel for other reasons. Anyone who sets up a university patent policy and licensing program on the premise that Bayh-Dole expects patenting and licensing misses the point. Bayh-Dole expects such things when and only when they promote the practical application of subject inventions. To own first and then as a separate step to fuss with promoting practical application through licensing completely misses the point.

Bayh-Dole does not set out whether inventions made with federal funding should be patented, or whether universities own or do not own, or whether inventors are allowed to retain rights to their inventions or assign them to federal agencies. Bayh-Dole does not concern itself with whether universities should try to make money from licensing–and that in itself is a defect.

The stated objectives of Congress as set forth in the Act then are (I paraphrase–read them in full at 35 USC 200) promote use with the patent system, involve small businesses in federal research, promote collaboration between nonprofits and industry, promote free competition, use US labor, protect the public from nonuse and unreasonable use, and lower administrative costs.

One might think from the law’s statement of objectives that Congress intended universities to be selective in choosing things to patent, respectful of faculty inventors’ decisions of whether to pursue public availability or commercialization or allow rights to go to the federal agency, and to focus on free competition rather than on creating monopolies. But we don’t find university lawyers much repeating anything like this, even though these things, too, are supported by a reading of the law.

Certainty of Title

The article then turns to “certainty of title” as the great benefit for commercialization:

Certainty of title to inventions made under federal funding is perhaps the most important incentive for commercialization.

This was another of the rhetorical arguments that was pushed to get Bayh-Dole passed, and has been repeated by advocates since. There was nothing particularly uncertain about title to inventions before Bayh-Dole. Federal agencies contracted for research and had policies with regard to whether they allowed contractors to retain title to inventions or required assignment of title to the federal government. What was uncertain was, in any given case, whether an agency that did not allow a contractor to retain title would grant an exception. In those rare cases where two federal agencies with different policies on contractor ownership of inventions both supported the same invention, there was still the uncertainty that one agency might not allow the contractor to retain title.

The uncertainty of title that got the university-affiliated research foundations upset was what happened if they got ownership of an invention they wanted to try to license but could not get the federal agency that provided the funding to agree to allow them to keep title. That situation was indeed uncertain. Somehow this limited case of getting out ahead of the federal agencies on wanting to patent something got turned into the idea that companies just could not accept new inventions unless they were sure about the title to inventions. True, in the sense that it’s a rare company that is willing to pay for a license for a patent from someone who does not have proper right to own that patent. But if there weren’t a patent in the first place, there would be no title uncertainty, and if the invention was assigned following federal agency policies and approvals, there also is no uncertainty. It’s only if one is in a rush and skips steps that one creates uncertainty. That’s what the private invention management foundations were doing. They created the problem, and then blamed federal agencies for it, and then got a law passed that created even more problems, but they got what they wanted–the right to resist claims by federal agencies that public aims would be better served by the government owning any patent rights.

Implementation of uniform patenting and licensing procedures, however, combined with the ability of universities to grant exclusive licenses, are also significant ingredients for success.

Bayh-Dole authorized the creation of standard patent rights clauses to be placed in federal funding agreements, and required federal agencies to use these clauses or show good reason for exceptions. Uniform patenting and licensing procedures at universities is an entirely different thing. Different industries use patents in different ways, and diverse technologies benefit (or don’t) from various patent licensing strategies–and some technologies prosper best when there are no patents–as is the case, often, with open standards and rapidly changing technologies, especially early on when adoption even for research purposes is an iffy thing.

Universities always had the ability to grant exclusive licenses–but most university policies warned against it. Moreover, most universities did not manage patents. Other than a few scattered licensing offices–MIT, University of California, Stanford–licensing was done by research foundations and national patent licensing agents such as Research Corporation. The big change that came after Bayh-Dole was the abandonment of the external agent model and the move for universities to deal directly in patents.

This combination of factors led to a tremendous introduction of new products through university technology transfer activities. The licensing of new technologies has led to the creation of new companies, thousands of jobs and educational opportunities and the development of entirely new industries.

Nothing to support these claims but heresay, spin, and cherry-picked accounts. Many of the new companies claimed were shell companies created by universities to out-compete existing small businesses for the newly created SBIR and STTR funds. The universities don’t report when their startups fail. No tombstone marks the grave. No press releases get updated. There’s nothing to indicate, further, that there has been a “tremendous introduction of new products” as a result of university patent licensing. There have been some new products, of course, but the University of California estimates only 1 in 200 inventions has ever led to a commercial product–and that’s without considering whether those commercial products are profitable, and furthermore without considering which of those commercial products arose from federally supported, university-patented inventions.

I have commented in detail on Maxwell’s next paragraph in an earlier article in this series, so I will only quote it here to maintain continuity for what comes next. Nothing about this paragraph of claims makes logical or practice sense. Patents owned by universities do not protect scientists from anything and universities owning inventions is not the only way that a university can bring itself to accept research funding from sponsors. Maxwell, a patent attorney, surely knew this could not be so. Perhaps it’s just a lapse, or a moment of carelessness. But when the writer is IP counsel to the University of Texas System, the opinion even if a lapse carries weight:

A university owning the inventions it creates with Federal funding has another significant benefit — it protects the right of scientists to continue to use and to build on a specific line of inquiry.  This is very important to research-focused institutions because of the way research is typically funded — with multiple funding sources. Title to inventions to the university/institution is the only way to ensure that the institution will be able to accept funding from interested research partners in the future.

Practical Application, Not Commercialization

Now we get an account of the “Congressional intent of Bayh-Dole”:

Increase American innovation
Promote commercialization of inventions
Encourage participation of small businesses

This is a selective list–I laid out the full set of objectives above. But it’s also a misleading list. The participation of small business objective has to do with getting federal research funds to small businesses–the SBIR program was started at about the same time as Bayh-Dole. The Congressional objective was not to somehow stimulate university-affiliated startup companies to “commercialize” university inventions.

As for the second objective the article lists, the full statement of the intent is to “promote the commercialization and public availability” of inventions made by US industry and workers. Here “commercialization and public availability” do not represent a time sequence of commercialization first and then public availability of products. Public availability and commercialization are presented in parallel, just as they are in other sections of the law. “Practical application” is given a definition (“commercialization” is not):

The term “practical application” means to manufacture in the case of a composition or product, to practice in the case of a process or method, or to operate in the case of a machine or system; and, in each case, under such conditions as to establish that the invention is being utilized and that its benefits are to the extent permitted by law or Government regulations available to the public on reasonable terms.

The aim of practical application is to make and use, with the benefits of doing so available to the public “on reasonable terms.” There is nothing here that requires commercialization or even for-profit companies to be involved. What matters, as far as the definition is concerned, is that the benefits of use are available–there need not be any products, unless the patent claims a product. For instance, a standard would be just as important as a product, and a class of products produced by artisans would be as welcome as one product sold by a company holding a monopoly position as an inducement to invest.

One can impose on this definition of practical application and on the statement of the Act’s objectives that the “secret, true” intent of the law is to encourage universities to file patents and license them exclusively to companies willing to try to make commercial products, but there’s nothing in the Act to support such a claim. There’s just nothing that expressly excludes such practice–which perhaps is a defect of the law.

Getting Right the Major Provisions of Bayh-Dole

We now reach in the newsletter article a bullet list of the “major provisions” of Bayh-Dole. I’ll comment on each item in turn:

Inventions conceived or reduced to practice using federal funds are owned by the University – not the Federal Government

Incorrect. Patentable inventions are owned by the inventors, as always. Disposition of initial ownersip of subject inventions is not in the law, not in the implementing regulations. No excuse for such a reading, other than someone who has not read the law carefully. But then, this reading is what most everyone in university patent administration was repeating. It would take the Supreme Court in Stanford v Roche in 2011 to explain to these folks that they really ought to read the law.

Universities are encouraged to collaborate with commercial concerns to promote the utilization of inventions arising from federal funding

The objective as stated in the Act is simply “to promote collaboration” between industry and nonprofits. There is nothing that restricts that collaboration to the first stated objective, of using the patent system to promote invention use. Oddly, there is nothing obvious in the law that corresponds to anything addressing such collaborations. One might think licensing of patent rights as collaboration–but that’s often the source of anti-collaboration. A more interesting comment on such collaboration is given in the implementing regulations, at 37 CFR 401.1–

An example of such related but separate projects would be a government sponsored project having research objectives to expand scientific understanding in a field and a closely related industry sponsored project having as its objectives the application of such new knowledge to develop usable new technology.

If the government funds the basic work, and a company can step in and fund the applied work, then a university can use the federal funding and the prospect of managing inventions from it, in coordinating with companies for related work–but work also outside the scope of Bayh-Dole.

The list continues:

Universities are expected to file patents on inventions they elect to own

The universities have the right to elect to retain title to inventions they already own. The law gives a university no right to “elect to own” inventions and thereby somehow come to own them. As the Supreme Court put it, how universities come to own inventions is a matter of conventional patent agreements and assignments. The law limits how federal agencies may require assignment of any invention to the federal government. And anyway, one files patent applications, not patents.

Universities are expected to give licensing preference to small businesses

Correct. The decision whether to give preference is “at the discretion of the contractor” according to the standard patent rights clause. In the original version of Bayh-Dole, before such things were amended away, the licensing preference to small business included allowing exclusive licenses to run for the full term of the patent, while exclusive licenses to non-small companies were restricted to no more than eight years.

Manufacturing should be in the United States

Only with regard to exclusive licenses to use or sell.

The government retains a non-exclusive license to practice the patent throughout the world

No, the government does not “retain” the right. It does not have the right until that right is convey to it following a title chain extending back to the inventors. The Act requires federal agencies to use a standard patent rights clause that requires contractors to take actions to protect the government’s interest and to grant licenses to the government if they elect to retain title to inventions assigned to them. And the license is “to practice and have practiced” the patent, and “practice” meant, in past government patent policy, “to make, to use, and to sell.” That’s a broad right.

Institutional Control of Patents

So, a bit sloppy, and incorrect on the major point regarding ownership. It’s understandable that in 2005, with advocates for Bayh-Dole claiming that Bayh-Dole vested ownership with universities. Eventually a set of university front organizations and forty-five universities, led by their legal counsels and including the University of Texas at Austin, signed onto an amicus brief in Stanford v Roche claiming that Bayh-Dole was a vesting statute and universities owned inventions outright whenever federal money touched research involving the invention.

Bayh-Dole provides for a “[d]isposition of rights” in federally funded inventions, 35 U.S.C. § 202 (title), and establishes a presumption that ownership is allocated to the university or other nonprofit, 35 U.S.C. § 202(a). Individual researchers are allocated rights of ownership only if both the university and the federal agency providing the funding decline to take title.

It may well be that university administrators presumed they had title, but that presumption was not established by the Act. The Act, whether by politics or lapse, is silent on ownership. This gap was recognized by the federal agency (now the Department of Commerce) that drafted the standard patent rights clauses authorized by Bayh-Dole. There, they included the (f)(2) requirement–that universities were to require research employees to protect the government’s interest, including conveying rights to the government. The (f)(2) requirement does not involve inventors assigning title to universities or patent management organizations. University policies and assignment instruments that require assignment to the university do not meet the obligation of (f)(2). Instead, they attempt–often badly–to circumvent the requirement.

The attorneys, having established a faulty premise, felt at liberty to reason from it to consequences:

By the same reasoning, the federal government’s
own rights and interests are subject to circumvention
at the whim or by the inadvertence of an individual
inventor. The court of appeals’ refusal to enforce
Bayh-Dole places at risk, for instance, the government’s right under the statute to reserve title to itself in certain instances, to take title from a nonprofit
that violates the Act’s provisions, to receive a license
to practice the invention, and to require the nonprofit
to grant a license to a third party under circumstances
specified in the Act.

The regulation that protects the government’s interest is not in Bayh-Dole. It is in the standard patent rights clause authorized by Bayh-Dole. Remember: whatever Bayh-Dole requires has to do with federal agencies, not citizens. Anything the agencies must do flows to universities through funding agreements. Anything that flows to inventors is a result of actions taken by the universities.

If university legal counsels had bothered to read the standard funding agreement–the first, most basic rule of contract interpretation–they would have seen that the federal government’s interests do require university action, but not what the universities (and their legal advisors) presumed. Universities were to delegate responsibility to potential inventors, not demand assignment from them. Universities could, of course, demand assignment for other reasons–but not on the claim that federal law, or regulations, or funding agreements required assignment. University attorneys didn’t bother to work through the difference. That many attorneys ought not be wrong, but they were.

The Choices of Administrators and Their Legal Advisors

Maxwell ends with an account of criticisms of Bayh-Dole, but then blithely dismisses them because, apparently, there’s a consensus that the law is successful, so the criticisms must be wrong:

But, the Bayh-Dole Act is not without its critics; many of whom argue that it does little more than give industry carte blanche to acquire intellectual property rights on technologies supported by taxpayers, thereby giving away taxpayer rights to the inventions. They have also argued that the commingling of academia and commercial enterprises has turned the American university into a sort-of-school-corporation entity that stifles innovation in pursuit of licensing revenue — which could lead to bias in scientific findings and undermine public trust. And, its critics also say that it slows and hampers research because data that would otherwise be openly shared, is now tied up in patent rights that prevent other researchers from making use of it.

In spite of its critics, most would agree the Bayh-Dole Act continues to be a national success story, representing a successful merger between government, universities, and industry.

Here’s the bottom line. It’s not Bayh-Dole. Bayh-Dole has its problems. It is not a well written law. It may not even be a well conceived law. But the problems we see in universities are a failure of administration and of legal counsel. The problems involve incompetent drafting of policy, failure to implement the requirements of the standard patent rights clause authorized by Bayh-Dole, an incapacity to explain to university inventors their rights, and an unwavering belief in the idea that a bureaucrat’s thumb should be in every innovation pie, as a matter of national research policy. That university attorneys would cater to this belief is a cause for concern.

It may well be that patent management is a complicated thing requiring the skill of experts. Universities, luckily, are not the country’s only repository of such experts, and all too often it is obvious that the folks at the IP wheel at universities have little awareness of their trade. Institutional ownership of inventions (and whatever else) can make patent management even more difficult. Just contracting with a university is difficult, not to mention trying to obtain patent rights. Exclusive licenses make things even more difficult, involving background rights, sublicensing, right to sue, and diligence, among other things. Claiming everything makes things more difficult. More and more noise, less and less signal. University policy requirements and administrative management choices make an already complex thing way more difficult than it ought to be. (I’ve seen IP attorneys at two public universities fighting as a matter of principle over which state’s law will govern the license agreement–“our state will not submit to the laws of your state.” Can we just stop now?) Changing patent policy statements to assist overwhelmed administrators rather than enable inventors and principal investigators to manage their affairs, including invention rights, actually makes the whole thing even worse.

It may be that patent attorneys see Bayh-Dole as a golden goose, not because it is wildly successful in transferring technology from labs to companies, but because it creates a tremendous amount of new business for patent attorneys. It could be that simple. When one’s hand is in the cookie jar, it’s hard to talk about health or moderation. I suggest that the predatory aspects of the University of Texas System intellectual property policy may well be primarily an outgrowth of administrators and legal counsel who see career opportunities advance with increased volume, complexity, and liability. As a result, they advocate for policies that demand more volume of work, create more problems, carry more risk, are more difficult to manage, and give them more power and better pay–all for which they then have a basis to request more budget (and often, that budget comes from licensing revenue–so universities should report their full technology transfer expenditures–absolute numbers and as a percentage of their licensing income).

What has happened to university technology transfer after the passage of Bayh-Dole may have much less to do with insight into how innovation happens and much more to do with being blinded by institutional incentives. Folks have made a career from a chronically failed idea about research-involved innovation. It is convenient to hold to this model, if for no other reason than they think their careers depend on it. Anything else is a threat, er, would be innovation.