We are working through an article by Boettiger and Bennett that looks back on Bayh-Dole and wishes things were different in some ways. The actual thrust of the article, though, is about university patent practice. To get there, however, we have to work through their reasoning, history, and context. Vannevar Bush in Science the Endless Frontier argued that the government had an important role to play in supporting two research objectives–to advance the frontiers of science and to conduct research to advance the public welfare by drawing on science, including frontier science. Somehow, it seems, few really want to understand what Bush was getting at–research is research, so patent the results and get someone with money to buy up the patent, repeat. But that is not what Bush was getting at. Let’s consider where patents fit in research to advance the public welfare. The abstraction “public welfare” is part of the problem, because it contains within it all sorts of strange arguments about how greed and self-interest will eventually lead to the best of all possible worlds. And then we are dealing with another abstraction regarding possible worlds. Sigh. It doesn’t much end. Abstractions all the way down. Anyway, let’s look at speculative monopolies and public welfare research, since this is what Boettiger and Bennett are concerned with.
Speculative Monopolies and Public Welfare Research
Why did HEW end the Institutional Patent Agreement program? Two reasons stand out. First, under the IPA program, universities did no better than the federal government in making inventions available. Their “commercialization rate” was about 5%, not the 25% to 33% that universities had claimed for non-federal inventions. Second, private monopoly positions on compounds discovered to advance the practice of medicine weren’t consistent with the federal policy-makers’ idea of research to advance public welfare.
Think of it this way–I pay folks to fight fires and I find out that they’ve been cutting deals to sell their firefighting expertise to speculators who then refuse to fight any fires unless they are paid a premium by the people whose buildings are on fire. And the speculators are sharing profits with my firefighters. Screw that. Screw the firefighters for not fighting the fires themselves. Screw them for selling rights to speculators. Screw the speculators for exploiting crisis situations. Choose any scenario you want–help after a hurricane, EMTs at an accident scene, helping a broken family heal, curing a disease or repairing an acute injury, improving agricultural soil. Any of these can be thought of as humans working to help other humans, or as a market to exploit, with people ready to bet on who will successfully exploit the market and who will fail. As a matter of national policy, which should take top priority? Helping people with new knowledge and practices or generating profits by turning help into a business matter by withholding the new knowledge until someone pays for it? Are we trying to help, or using need as an excuse to commercialize help? Or are we actually just trying to create a betting game over efforts to commercialize help, as a kind of capital-intensive recreation, “fantasy patent licensing” rather than “fantasy football”?
While markets might arise around inventions and discoveries in areas of public need, those markets may well come after government development or might develop in parallel with making research findings available to all. There are historical precedents for such things (the internet, say). The pushback from the pharmaceutical industry has been that without monopolies, they (the pharmaceutical industry) would not develop compounds discovered by the federal government. It was an argument based on the refusal to give up winner-take-all monopoly as a basis for industry competition, even in matters of public welfare.
Under the Kennedy patent policy, if the federal government funded a pharmaceutical company directly to do drug research, then the funding agency could allow the company to retain ownership of any inventions created. But the Kennedy patent policy also set a default of government ownership if the research was to advance public welfare and not just to procure a commercial good for government use. So if the government had wanted to supply a new painkiller to all government employees, then it might retain the services of a company in the business of developing new drugs, and let the company hold the patent rights to supply the painkiller to everyone else at market rates. The government could negotiate its own price or authorize a different supplier of the drug–the company’s upside was profit from doing the development work plus the non-government market. If there was no non-government market, then the profit was in doing the development work.
If the government wants a new painkiller to be made available to the public generally, then the Kennedy patent policy defaulted to government ownership and the public domain. The implication, then, is that private money might supplement government money, but only on terms set by the government–that is, whatever market thinking there might be is limited in scope to permit private investors who refuse to be donors to recover their investments, perhaps with some modest upside incentive. But there’s no reason to sell the situation to investors for the entire term of a patent, nor to permit monopoly pricing the whole time, nor to suppress all other uses, nor to give the investors the ability to extend the monopoly for decades with improvements, all the while preventing others from improving or adapting. Well, actually, apparently there were reasons to do this, because that’s what the IPA program turned out to be doing, and what Bayh-Dole claimed was a virtue to do.
It was this conflict within the Kennedy patent policy that appears to have caused uncertainty in what an agency might decide. To make broadly available? Or to permit private monopolies to develop product? Not for all research, of course, and not in most federal agencies, but just enough bother–and a pharmaceutical industry boycott–to make it appear that the problem was bigger than isolated situations and that the solution was to make federal patent policy arbitrary, and that the new arbitrary policy should make the introduction of private patent monopolies prior to public availability and use a good thing, if not the only thing, rather than a questionable thing. Bayh-Dole was not so much “patent when doing so is the primary means to make an invention available” as it is “patent early and often to prevent anything from becoming available without first extracting payment from monopoly speculators.”
Now think about frontier science for a moment, a related but different situation from that of public welfare research. “Frontier science” is shorthand for the scientific research Vannevar Bush envisioned in Science the Endless Frontier. That science pushes out on the edges of knowledge, looking for new phenomena and challenging existing consensus theories and paradigms. Bush’s idea–often ignored or misunderstood–was that pushing the frontiers of science would create more conceptual tools with which to work to create new things to improve public welfare. Bush did not imagine that the work of science would have to be “commercialized” to produce this effect. If new discoveries were published and available for use, then when a need arose, with a creative spark folks would be able to imagine new approaches, new products, even new industries. And then get about things. Not a linear model.
Bush wanted a bigger tool board in the science shop. That got turned into the idea that selling tools exclusively as commodities should be the necessary end product of federally funded scientific research–“commercialization” (but really what folks meant by “commercialization” was commodification, and eventually what they meant was betting on the likelihood of future commodification).
Notice how so many university licensing office diagrams on the “technology transfer” process end with the license–and it is, typically, a single license. Pay no attention to the thought that the diagrams never point out that the depicted process “works” only about 20% of the time. That is, reaches a license. And that the process beyond the process–that is, commercial commodification of an invention into a product takes place (apparently) only about 1 invention in 200. Here is a common process: “(1) take ownership of an invention; (2) lead inventors to believe that their invention will be licensed to industry and they will make substantial royalties; (3) file a patent application; (4) publish the invention as available for licensing; (4) wait until the patent expires.” Yeah, that’s it. No takers for that exclusive license offered to monopolists willing to share. A second process, one preferred by the Licensing Executives Society, is to replace (4) with “wait until someone infringes the patent and then sue, but call it “monetizing the license.” While patent trolling is about as rare as commercial commodification, it can be just as lucrative.
All this is vastly different from what Vannevar Bush attempted to do by advocating that the federal government had a role in funding basic scientific research. Frontier science opens up new ways of observing and thinking about the world. The job is to discover and to map and to build a basic infrastructure of trails and outposts so that others may follow and figure out what to do. Basically, Lewis and Clark observing and reporting, not Spanish conquistadors claiming ownership of everything they saw, per the Requerimiento.
Whatever the area of science that might be explored, the idea of adding federal funding was to speed the exploration and expand its reach. But if many investigators are now at work rather than just a few, then there’s the prospect that many patents might be issued on bits and pieces of similar stuff, fragmenting the geography of any new area of scientific knowledge. Such fragmentation delays rather than promotes use, as anything interesting lies beyond not one paywall but often five or ten. I recall one nanotechnology startup spent many months securing from various universities the twenty or so licenses it needed to make a single product. There’s a history of how university patenting fragmented an early nanotechnology industry and delayed its development, but it’s difficult to tell a story about what never happened but could have.
You might see, however, how someone working with the Kennedy patent policy would see that research to expand scientific knowledge ought to expand the public domain for new science or create a federal patent commons for new science–licensed non-exclusively, typically royalty free. At least, that should be the default until someone makes a clear case, in a given circumstance, for something better. Does this approach seem unreasonable? Vannevar Bush didn’t think so. But the Bayh-Dole folks clearly think it is unreasonable.
And that’s why it is a bit mind-bending to characterize Bayh-Dole as “shifting” the “incentive structure.” The federal government had no financial incentive to obtain or license patents. The U.S. patent system was not set up to generate money for the federal government, but rather to encourage publication of inventions in exchange for the right for limited times (two sets of apprentices plus three years) to exclude others. Whatever “incentive structure” the federal government had, it was to get more scientific knowledge at the boundaries of science and make that knowledge broadly available. A patent might do that, but publication did just as well. It’s clear that the federal government’s “incentive structure” did not “shift”–it was supplanted by one in which institutions did assert a financial interest in patents that were licensed, and then asserted a financial interest in all inventions with the claim that the university (and its patent broker partners) would patent most everything and license what they could of it. Sort of like ocean trawling–catch everything, keep what’s valuable and ruin the rest. That’s the mindset smiling at you from university IP policies these days. What’s there to like about that?
We might do better, then, to say that Bayh-Dole created an incentive structure for universities to seek to profit from patent licensing. This, so the argument went (and still goes), is more effective at meeting the public interest than the Kennedy patent policy default of public domain or federal patent commons. Except that the evidence from practice (secret as most of it is) does not point to patent licensing practice under Bayh-Dole doing better than the public domain or federal patent commons. The evidence is that under Bayh-Dole, things are much worse–with commercialization rates one tenth that of the IPA program, which in turn was considered unproductive. Private patent monopolies on research to advance science or promote public welfare create just enough income to preserve the licensing practices and create sufficient volatility that from time to time a monopoly license “hits it big” either the introduction of a new product under license, or a transaction involving the company that holds the license, or through litigation for infringement when commercial development has failed but other companies are using the invention anyway.
Thus, to say that Bayh-Dole “shifted the incentive structure” makes little sense. Bayh-Dole created the opportunity for universities to abandon their role–at least since 1947–of providing the resources to enable federally supported research and turn that research into an engine by which to create private monopolies prior to the development of a robust public domain or patent commons or early adopting use by capable technologists and companies. The motivation for Bayh-Dole was to end-run HEW policy on research in medicinal chemistry so pharmaceutical companies could enjoy private monopolies and charge prices accordingly. To cover their tracks, those involved sought to make the policy arbitrary and government-wide, as if private monopolies were the only, best, primary, first engine of innovation. They also extended the law from universities and nonprofits to include small businesses–and by doing so conflated the expectations of grants (subventions) with the expectations of procurements (purchases).
Trying to sort out this conflation also bends the mind. Grants became another form of procurement, and that in turn suppressed the idea that the funding agency had a purpose other than procurement–agencies bought reports of science rather than aimed to expand the frontiers of science for use by everyone. But if the government were merely procuring, then it should be fine with non-exclusive rights, leaving nonprofits and small companies to develop inventions for public use, however which way they might. Small companies, however, were justified in participating because they “needed help”–so procurement became, in a way, subvention–that the purpose of government funding to small companies was to make them into bigger companies, and the like. Or, another way, to subsidize the investment in such companies, and to reduce the risk of investors in pursing product development.
That’s an interesting outcome, but not one identified as the policy and objective of Bayh-Dole. It’s rather odd that allowing small companies to own patents on inventions made in federal contracts should be a strong inducement for small companies to “to encourage maximum participation of small business firms in federally supported research and development efforts.” The “commercialization” rates for SBIR grants appears to be dismal as well. There, however, it appears to be the case that small companies that get federal contracts like the contracts more than they do developing products for non-government markets.
Pingback: Ten Years After 25 Years After Bayh-Dole, Part 1 | Research Enterprise