We are working through the Faster Cures FAQ on Bayh-Dole. This is the last bit.
7. What have the impacts of Bayh-Dole been?
Bayh-Dole has both passionate supporters and detractors. Advocates argue that it has brought the results of federally funded research out of the laboratory and into the clinic. Critics say it has increased the time and cost of research because it “has negatively affected the practice and norms of science, created “anticommons” problems, contributed to patent hold-ups, and led to unnecessary increases in consumer prices.”xiv
Supporters and detractors are not the issue. The supporters mostly misrepresent the law and make a bungle of the patent rights clauses and ignore compliance and make a virtue out of things that Bayh-Dole does not authorize. Detractors mostly complain about the supporters’ claims, but accept the supporters’ representation of Bayh-Dole. So it’s a love and bitch fest over a shared misrepresentation of the law and wholesale non-compliance with Bayh-Dole’s patent rights clauses. Whatever is happening, it is not Bayh-Dole.
The issue is that there is almost no public evidence that shows that Bayh-Dole has met its stated objectives. Bayh-Dole has no objective to ramp up university bureaucracies. Nothing about creating hoards of patents to prevent broad public access to inventions. Nothing about substituting measures of activity and statements of hopeful intent in place of outcomes. Nothing about conflating Bayh-Dole related work with all university inventions (where Bayh-Dole patents are about 40% of the total university hoard).
The critics recited here are mostly academics. We don’t hear from industry, from inventors, from university faculty. We don’t see actual evidence. The proponents are patent attorneys, and university patent administrators, and university lobbying fronts, and biotech lobbying groups determined to keep the patent monopoly pipeline open to the pharmaceutical industry and to expand the pipeline if possible to other industries and so expand the potential opposition to any change in public policy. There are no public data to support a positive assessment of Bayh-Dole. Bayh-Dole does not require federal agencies to collect data. The GAO reported in 1998 that federal agencies don’t do anything with the data they do collect. And Bayh-Dole purports to exclude all such reported data from FOIA disclosure.
The Association of University Technology Managers (AUTM) reports that in 1979, one year before Bayh-Dole was passed, only 30 universities had a technology transfer office; by 2013, data released by AUTM indicate that more than 200 U.S. institutions with tech transfer offices responded to its annual survey.
This is most strange. Before 1979, most universities used external patent management agents–Research Corporation, Battelle, University Patents, or an affiliated research foundation based on the WARF model. For that matter, Research Corporation advocated “technology transfer” offices. The offices existed to transfer inventions to Research Corporation for possible management. The offices did not necessarily have an independent licensing function. For example, the University of Washington’s Office of Technology Transfer when it was established after Bayh-Dole initially did not do its own licensing but conveyed inventions to the Washington Research Foundation (and, in policy theory at least, other agents) for licensing. The count of university “technology transfer” offices is thus both bogus and irrelevant. Nothing in Bayh-Dole aims to increase the number of technology transfer offices operating at universities. Why should anyone think the growth of an extensive bureaucracy has anything positive to do with Bayh-Dole’s impact?
AUTM data also show that levels of invention disclosures, patent applications, patent issuances, and licensing have increased steadily since Bayh-Dole’s enactment,
AUTM data does not break out federally funded inventions. Patent data shows that universities are now acquiring non-federally funded patents at a faster rate than federally funded patents. And patenting and licensing are functions of activity, not outcomes. One can license a patent and nothing might come from the license. That result is not merely a failed effort to commercialize–it is the exclusion of all other activity with respect to that invention–often a class of compounds or devices or methods–during the term of that exclusive license. Sometimes a failed exclusive license is sufficient to destroy any other opportunity to use the invention. Thus, AUTM’s report of overall activity has little to do with the effect of the Bayh-Dole Act.
and that universities created 10,000 companies since Bayh-Dole was enacted, 4,000 of which are still operating. University licensing income increased from $7.3 million in 1981 to $3.4 billion in 2008, xvii and university patent licensing supported 3 million jobs between 1996 and 2010, an average of 200,000 jobs per year.
The AUTM counts of companies are not audited for accuracy and overcount. Multiple universities may claim the same startup company, universities create startup companies as shell companies that have no operations and exist only to hold rights. At the University of Utah, where they claimed to have started 100 companies in five years, nearly all of the companies were shell companies. A review by the Utah state legislature found a culture of “lies and untruth.” Similar exploits took place at the University of Washington, counting as startups companies long in existence and companies merely for having done a licensing deal (such as obtaining software to use).
AUTM uses an almost cartoon method of calculating “jobs” supported. It takes the reporting licensing income, assumes this money represents a royalty based on earned sales, and from that comes up with a number of what those “sales” must be by multiplying the income by the reciprocal of an assumed royalty rate. So an assumed royalty rate of 2% means multiply the licensing income by 50. Now divide by an average annual “wage” and one gets, quite magically, “jobs.” Idiotic, really. First, not all licensing income is earned royalties from sales. Most licensees pay not from sales but from their pockets–to reimburse patenting expenses, to pay license issue fees and milestone payments, or in the form of stock that later is sold by a university. None of that reflects sales. And as for royalties on sales, say from pharmaceutical companies, there’s nothing to indicate that the company’s income is used entirely to support “jobs” at the company. Profits, for instance, go elsewhere. AUTM creates nonsense, pitches it as fact, and folks at places like Faster Cures repeat it.
Individual universities can generate large revenue streams in the event they own the rights to a patent covering a commercially successful product.
The universities don’t “generate large revenue.” They position themselves with patent filings to require companies to pay–either cash or equity. The “large revenue” events happen at best once a decade in places like Stanford (three big hits in 36 years) and a lot longer at others (one big hit in 35 years at the University of Washington, for instance). And here’s the bitter reality: a university patent licensing operation can appear financially successful with just one big hit every two decades, even though it holds hundreds of patents that keep inventions from being used that it is unable to license or if licensed do not result in any commercial product. The Bayh-Dole Act’s public protections do not permit such practices, but federal agencies do not enforce the default patent rights clauses and thus we may blame Bayh-Dole for destroying a large swath of the research commons and ready access by companies and organizations to the use of inventions made with federal support, quite apart from any attempt at development of these inventions as commercial products.
A 2013 study released by the Brookings Institution, however, showed that most universities do not break even on their technology transfers; on average, 87 percent have finished in the red over the last 20 years.
That a university does not make money on licensing is a meaningless observation. A university could operate a patent licensing office as a public resource, obtaining patents and making inventions available to everyone non-exclusively, royalty-free, and withholding licenses only from those that assert infringement against the university or any of its licensees. That office would also “lose money.” The premise that a university must manage inventions for the money runs against public policy, too. The university ought to operate in the public interest, not institutional financial self-interest. That a university seeks to make money from patents undermines its role in supervising research for integrity, in working with all companies and individuals that seek its assistance, with reporting without spin to the public on the research efforts it hosts.
The Brookings Institution report skews the numbers anyway. I conducted an extensive review of university licensing office revenues a number of years ago. Licensing offices that are relatively new (under 20 years) are much less likely to show significant income–because (we might conjecture) they usually have not had time for the one big hit deal to show up that would make them financially set for the life of some patent–for 15 to 20 years. That lack of a big hit deal, in turn, is a consequence of the failure of university administrators to recognize that most of the immediately successful university licensing operations were started to manage a big hit invention they already had rather than in the hope for such an invention in the maybe distant future. It’s a fine line, I know. But if one works at it, the difference begins to dawn. “Pillage, then burn…. pillage, then burn.” It’s a basic lesson, but lost on folks who think the field of dreams applies to patent monetization. As Thucydides put it, “It is a habit of mankind to entrust to careless hope what they long for, and to use sovereign reason to thrust aside what they do not desire.”
For all that, the licensing money is not the issue for Bayh-Dole. There is no mandate in the standard patent rights clause that a university must make money through its licensing of subject inventions, or that money making is a standard of success. What matters is that the practice of the licensing office results in i) practical application–use of inventions with benefits available to the public on reasonable (not monopolistic) terms; use of inventions takes place more rapidly than if no patents were obtained by the university (i.e., “promoting” the use of inventions); and the university’s administration of inventions results in more opportunities for companies and other organizations to exploit the inventions, to compete in their use, and even compete with the university’s own research programs and startups to create improvements and applications (i.e, “promoting” free competition and enterprise). No one tracks such outcomes. They don’t matter to university administrators, apparently. And no doubt such outcomes mostly aren’t happening.
The same study shows that most of the revenue accrues to a small subset of universities—the top 10 percent of earners accumulate nearly 75 percent of the revenue. Licensing revenue is also correlated with the number of employees in a university’s technology transfer office and university research expenses.
These are random correlations that don’t have anything to do with causes or outcomes. Big hit revenue is a power function. A few big hits, and long tail of small amounts of revenue. The institutions may change but the structure of the curve will remain stable, just as the 80-20 rule suggests for power functions. Licensing revenue is correlated with employees because when a university hits a big licensing deal, administrators expand the technology transfer office in the hopes that more people will lead to more big deals faster. That generally is not the case. More people means more expenditures, a greater volume of activity, and nothing much more. When the big hit deals run out, the university lays off technology transfer staff. That happened at Stanford after the Cohen-Boyer patents expired, and again at the University of Washington after the Hall patents expired, among other places.
What any of this has to do with the impact of Bayh-Dole is beyond me. Bayh-Dole’s default patent rights clauses have never been complied with by universities, never enforced by federal agencies. Federal agencies do not act on the rights reserved for them by Bayh-Dole–they do not act on their broad government license to practice and have practiced; they do not march in to protect the public from nonuse and unreasonable use; they do not act when there are exceptional circumstances to establish alternative patent rights clauses. Bayh-Dole, we might argue, has had almost no impact other than to greatly increase the paperwork involved in reporting subject inventions and federal agencies collecting useless invention utilization reports and grant close out invention reports.
What has had an effect is the faux Bayh-Dole Act, the misrepresentation of Bayh-Dole that has caused universities to repudiate faculty freedoms of research and publication in favor of an overreaching demand for institutional ownership of inventions. Faux Bayh-Dole induced many universities to alter their patent policies to forms unfavorable to innovation, collaboration, cumulative technology development, rapid uptake, and standards formation. Faux Bayh-Dole has induced university administrations to create unmanaged (and unmanagable) institutional conflicts of interest between the administration of research and the search for lucrative patent licensing deals. Faux Bayh-Dole has caused universities to adopt “commercialization” as a goal rather than “practical application” or “non-discriminatory access.” And with “commercialization” has come the assignment of inventions under the cover of exclusive patent licenses–so that university administrators will refuse to grant non-exclusive licenses in the hope of later finding an exclusive licensee (assignee) willing to pay a substantial amount to speculate on the future value of a given patent or patents. These are results of faux Bayh-Dole, not reported by universities or AUTM, as doing so would destroy the claim that Bayh-Dole practices have been virtuous, successful, and beneficial to the public.
The impact of Bayh-Dole has been to allow the emergence of faux Bayh-Dole. The Supreme Court in Stanford v Roche beat down a major element of faux Bayh-Dole, but the misrepresentation of the law persists in almost all of its substantive provisions–the definition of subject invention, the right of the university to take title, the imperative to commercialize, to grant exclusive licenses, the disregard for American manufacture, the disregard for the small business preference, the disregard for practical application, the disregard for compliance with the nonprofit-specific requirements on assignment, on use of income with respect to subject inventions. The federal agencies, too, have been derelict, as if they do not care, or as if they prefer as public policy faux Bayh-Dole. Since Bayh-Dole itself is punchless and unenforced and federal agencies decline to act on the rights the law offers to them, we can conclude that Bayh-Dole has
- permitted the faux Bayh-Dole practices to flourish
- provided a pipeline for pharmaceutical companies to acquire monopoly positions in health-related inventions made with public support
- made universities complicit in monopoly pricing and the suppression of all use of an invention but for the portion that results in a “commercially successful” product
- induced university administrators to adopt predatory patent policies that tie up in institutional bureaucracy inventions that should have followed different, and better, lines of management
- led to universities holding tens of thousands of patents representing millions of individual claims that block the rapid use of inventions in favor of holding back such use in the hopes of finding one patent deal every two decades that results in substantial revenue for the institution
It’s odd that the proponents of Bayh-Dole–not that Faster Cures is–seem particularly uninterested in getting Bayh-Dole right. It’s a shame, however, that Faster Cures does not provide a better FAQ that holds both federal agencies and universities accountable for the outcomes promised by Bayh-Dole and not delivered.