In February 2019, ILR Review published “Academic Entrepreneurship: The Bayh-Dole Act versus the Professor’s Privilege.” The article is paywalled, but a slide version of the paper by the authors is available at the National Academies website.
The authors compare U.S. university startup activity after Bayh-Dole with various European countries, with particular interest in Sweden. European countries, it is observed, envy the success of the U.S. after Bayh-Dole. This is as expectable as it is dismaying. The U.S. metrics for Bayh-Dole’s “success” are, like accounts of the law and its history, largely political spin–a kind of noble cause bullshit (that is, statements that do not have any interest in the truth and instead assert what one aims to make true, or what is made to sound true, or even what one sincerely believes is true but without the effort to validate).
Thus, the question facing European countries’ research invention policies is whether there’s a PhD entrepreneurial “gap” between the U.S. and their university research activities. The authors worry the data, but they don’t worry it nearly enough. Why?
1) There isn’t any Bayh-Dole data. No one publishes their subject invention activity separate from their other IP activity. Bayh-Dole keeps reports of subject invention use and development a government secret. AUTM does not ask for or publish subject invention metrics. Instead, the authors look at personal income data of academic STEM PhDs.
2) The data that matter are the outcomes. Bayh-Dole’s standard is utilization with benefits of that use available to the public on reasonable terms. Anything else–patenting, licensing, startups, amount of investment–is throwing sticks in the air if the outcomes aren’t there. A patent or a license or a startup is not an outcome. Personal income of academic STEM PhDs is not an outcome.
But what’s most odd is that the authors approach the issue by way of the income made by university STEM PhDs excluding those over 60. Whether university employees make money from non-university work seems way far afield from outcomes having to do with invention use or public benefit or reasonable terms. They apparently ignore such things as the value of stock that faculty hold in startups (and which they may receive for serving on advisory boards and not leaving the university) and look only at reported financial income in their study period. Makes no sense to go at it this way.
The authors conclude, after messing around with statistics which can be used to do pretty much anything one wants, that if Sweden adopted a Bayh-Dole-like approach (i.e., a faux Bayh-Dole approach, giving ownership of IP to universities), there would be less faculty entrepreneurship–at least as the authors define it. Again, what does it matter if there’s no attention to the outcomes? They end with this point:
Do not expect great economic outcomes from academic entrepreneurs
That’s a strange point! It may be that their assessment (based on personal income of selected STEM academics) that the money from entrepreneurship is not all that great. But that’s only part of the compensation, and having a startup on one’s academic resume may contribute significantly to a pay increase if one moves to another university position.
But an “economic” outcome is not merely a “personal financial” outcome. The authors make a jump–that somehow if faculty entrepreneurship does not pay them as well–taken as an aggregated group–as being professors, then we should not expect the results of their activity to have a “great economic outcome.” That’s just weird. Given that entrepreneurial speculation can be boom or bust, even one huge startup success (like, say, Google or Genentech) would swamp out all the other activity. But we aren’t to expect even any more of those?
And further. In an account posted at Sage, the authors comment on their study (BDA is Bayh-Dole, PP is Professor’s Privilege):
The paper shows that as long as they are stable and supported by an appropriate institutional framework, both types of legislations — BDA and PP — can generate fruitful outcomes in terms of invention commercializations by their creators.
The conclusion is thus: don’t rock the boat. Make small changes if necessary.
The authors might consider whether entrepreneurial activity in the context of government-supported research comes about *despite* government and institutional practices rather than is somehow *enabled* by it. Otherwise, we are looking at a post hoc fallacy couched in the form of an academic study.
Consider, for instance, these claims made by the authors:
Making money is an important motivator for employment
Professors respond to economic incentives
For academics, entrepreneurship does not pay well
There is a higher relative rate of entry into entrepreneurship by
academics in Sweden than in the U.S.
Why then do Swedish PhD’s bother with entrepreneurship? It doesn’t make any sense as a money proposition, other than as speculative entertainment, like playing the stock market or getting involved in a marketing scam just this side of legal.
The authors find that lower paid academics are more likely to leave for startups–that is, post-docs and adjunct positions, not tenured and not tenure-track faculty. Well, is that any surprise? And is it even meaningful? We aren’t even talking here about academic inventors leaving to start companies around their inventions. We are just talking about people slipping out of academics and then (often) back in, with the hypothesis that they do it for the money and then find out the money is not so good. A STEM training does not prepare one for the range of motives that might stir someone to act. Money or stability might have little to do with it.
When Scott Shane studied U.S. entrepreneurs, he found that a common theme was freedom from employment requirements (see The Illusions of Entrepreneurship):
The typical entrepreneur isn’t a Silicon Valley tycoon, but rather a white man, married and in his forties, who started his business because he didn’t want to work for someone else and who is just trying to make a living.
The typical entrepreneur doesn’t start a business because of a desire to make money, for the thrill of starting a business, to support their families, or to become well known; the typical entrepreneur starts a business because he doesn’t like working for someone else.
Shane also points out that entrepreneurs tend to start businesses in areas where companies fail often, and in areas where it is easy to start companies, and where they know something about what they are doing. People who are unemployed or employed part-time, change jobs often, or are paid poorly are typically the ones who will start companies. If the federal government wants merely to get more academics to start companies, then there’s a really simple answer: lower their pay, especially for conducting research, make their positions less stable, and control their work with enough regulation and sucky management that they will want to leave. Then you will get startups, especially in industries prone to startup failures, or where it’s easy to start companies. Big whoop.
But perhaps what the policy people want is an atypical entrepreneur. They want someone who is well-paid, in a stable job, being productive as a researcher, producing inventions–they want those people to start companies around their inventions, around inventions that other people would not know how to use or develop, but which used and developed would become important national assets, benefiting the public, creating value for investors, and stimulating the economy.
But what are these inventions–that can’t be used but for the personal involvement of the inventor, and for which use requires development that can be undertaken only with capital made available to a company, and for which it is more important that the inventor leave academic research and pursue a commercial venture despite the prospect that the venture will fail and the money will not be so good as academic pay? “For the public good I must leave the research position for which I have trained for decades and attempt to raise money to develop this invention by appealing to venture investors, and I must do so even with the likelihood of failure and the prospect of making less money.” This surely is not a typical entrepreneur. This is a very mixed up, conflicted person. Why would a national research policy aim to select for this kind of researcher–or create this kind of researcher so that more of these folks might be amenable to launching out as academic entrepreneurs? Would it be to make the bean-counting statistics look better? To have more rooster farms than there are in the U.S.?
Here’s the big problem: what are those inventions that these atypical academic entrepreneurs ought to be recruited to “spin out”? Do they even exist? Is the intersection of these inventions, should they exist, empty when crossed with the mixed-up, conflicted academics that appear to be the beneficiaries of federal incentives to entrepreneurp? Suggestion: it’s all fantasy.
The dichotomy over ownership of inventions (or IP or whatever anyone can claim)–institutional vs individual–also misses all the activity that supports a robust public domain, cumulative technology, research tools, depositories of materials, software and software code libraries, shared data and experimental setups, platforms, interoperability, collaborations, open innovation practices, reciprocity, and standards. These areas of activity support rapid development of supply chains and distributors, speed adoption decisions by infrastructure buyers, and expands post-acquisition support. These areas of activity create entrepreneurial opportunities by providing others with the freedom to practice, whether other academics, professionals working in a given field, company scientists and engineers with access to the same resources for their work and to evaluate claims made by academics, as well as by companies and investors seeking to make a new commercial product. Often these open assets, whether owned or not, do not require PhD academics to leave their positions to become “entrepreneurs” for there to be public benefits from the use and development of these sorts of outcomes from doing research.
It’s not a matter of patents being “good” or “bad.” It’s not a matter of capitalism (those with money should own in order to make more money than others would make if they owned) or communism (no one should own personally anything so that government or the government’s chosen favorite can serve the people most efficiently and effectively). And it’s not a matter of ownership vs no control at all. Ownership is important. Personal interest is important. Institutions and governments may influence what happens. But no one single policy–call it uniform or arbitrary–can anticipate innovation. Just can’t.
Start with the outcomes. Ask how each outcome *really* happened. Look at warfarin, the blood thinner. How did that come about? Oh, it was a farmer hauling a dead cow and a pot of uncoagulated blood through a Wisconsin snowstorm on a Saturday to try to find someone at the university who could help him learn why his cows were dying. And a creative faculty member who was able to recruit talented students to work on the project. They isolate the compound produced by a mold growing on sweet clover hay. They create 100 variations on the compound. They pick one to try as a rat poison, which WARF pays to patent and license as a commercial product. Later, a sailor tries to commit suicide by taking the product, which works slowly over a matter of days. He regrets his decision and doctors figure out an antidote so he doesn’t bleed to death like the farmer’s cows. Then the President has a heart attack, and they need a blood thinner, and here’s this rat poison that can be managed to do just that. Where’s the policy for farmers with trucks? Or faculty able to find and inspire talent? Or the invention management organization with money and ready to assist when asked? Where’s the suicidal sailor or the president with a gimpy heart?
Now repeat that for all the other outcomes, whether success or failure. Can any single government research policy could anticipate and reproduce that spread of outcome pathways? Would it magically turn failures into successes? And preserve the successes or make them even greater successes? Is the Sabine polio vaccine a failure because Sabine never gave investors a chance to buy in and speculate on the financial success of his work? Is that what government research invention policy ought to enforce–that regardless of what some investigator might want, the government should intervene to make sure that investors make money, or universities make money, or even that faculty must become entrepreneurs that make money. What a screwed up way of thinking, that.
I’ve worked in university technology transfer for nearly 30 years. My experience is that research invention use comes about mostly *despite* federal government policies and institutional IP management practices attributed to those policies.
Prior to Bayh-Dole, the federal government presumed ownership of inventions made in federally supported work would remain with contractors (if the contractors acquired ownership) except in five areas (federal development for release to industry, public health or safety, where the government is the primary user or developer and private ownership would create a preferred or dominant position, supervising the work of others, work done by nonprofits or contract research organizations) as well as where federal statute stipulated government ownership (space technology, nuclear weapons and energy). Bayh-Dole preempts executive branch policy and federal statutes when a contractor acquires ownership of an invention made in work receiving federal support. The law does not vest ownership, does not give any special privilege to obtain ownership, does not require a contractor to acquire ownership, does not require commercialization. See Stanford v Roche (2011).
Bayh-Dole (1980) replaced the Institutional Patent Agreement Program of the NIH and the NSF (which covered many of the nonprofits receiving their funding–1968-1978) and the Federal Procurement Regulation (1975) that implemented the Nixon patent policy (1971, based on Kennedy 1963). Both the IPA program and the FPR provided for contractors to keep ownership (when they obtained it) of inventions made under federal contract. The FPR also allowed inventors to keep ownership (if they didn’t assign their invention to their employer or another party). The Department of Defense also had a long-standing policy of allowing contractors to keep ownership of such inventions. The Kennedy patent policy (followed by Nixon) identified five conditions under which the federal government should require assignment of title, , including when the research pertained to public health. Congress added a number of other laws that made a government claim of ownership pertaining to specific research and development funding situations, but in all these cases, a federal agency could release the government’s claim if a contractor could show that its ownership would result in a greater public benefit than if rights were acquired by the federal government and then released to the public.
Whenever a contractor acquires ownership of a federally supported invention, Bayh-Dole preempts executive branch patent policy and the various federal statutes that require federal ownership. In essence, Bayh-Dole suspends the public interest justifications under which federal funding is allocated for private research and development so that nonprofits do not have to show how their ownership and management of inventions will benefit the public more than would open access, especially in matters of public health, where regulations require use, or where a federal agency intends to fund the development itself prior to release to the public.
AUTM’s licensing survey, for instance, does not distinguish between federally supported inventions (about 40% of university patents) and other inventions, nor between patented inventions and other technology that’s not patented. The survey, furthermore, does not report outcomes–such as Bayh-Dole’s standard of utilization or practical application–use with benefits available to the public on reasonable terms. For Bayh-Dole it is neither here nor there whether a university or a faculty member starts a company or deals in exclusive licenses. The outcome that matters is the public benefit–and it’s a stretch to argue that the primary public benefit is that companies have been started for investors to speculate over.
Bayh-Dole was created by patent counsel within the NIH to end-run Public Health Service patent policy and reestablish a pipeline of patent monopolies from federal funding to pharma. Everything else is distraction, counter-measures, and the rationalizations of interested pilers-on. As it is, while the federal government funds roughly 1/3 of biomedical research and development, only about 7% to 10% of commercially manufactured drugs are based on federally supported inventions. Of these, an even smaller portion pass the test of being available to the public on “reasonable terms.” And in all, the economic activity around anything Bayh-Dole does not amount to a rounding interval in the national economy. While the federal government provides about 60% of U.S. university funding, only about 40% of university patents cite federal funding. Even the presence of federal funding, along with a compulsory patenting requirement, does not increase university patenting and what university patenting does happen appears to block rather than enable entrepreneurial opportunity.
In the barnyard version of technology transfer, U.S. universities build chicken coops and then populate them with roosters and call it a new chicken farm. All we get is a rooster farm with mess and noise. They claim Bayh-Dole makes them do it. And they claim they do it very well. But it’s not the number of rooster coops built that matter, regardless of whether a university administrator does the building or a faculty member does. And it’s not how much money anyone makes for building rooster coops. We just don’t have a big pressing interest or need in more rooster farms.
If either prong of the authors’ dichotomy of arbitrary government policies appears to “work,” so long as change goes slowly, then perhaps the policy advice on offer is that if changes are made slowly, then those gaming a university system for personal or institutional advantage (or even in the public interest or for the good of research) can make the necessary adjustments without significant protest or bother.