There is a vast public literature–academic, legal, and popular–that claims Bayh-Dole is a wild success. That literature claims that compliance with Bayh-Dole requires commercialization, and that it is by commercialization that public benefit from federal research comes about. That argument reduces to the following:
- innovation requires bureaucrats everywhere
- federally supported research requires profit-taking before the public may benefit
- federal agencies as a matter of law may not challenge these claims
Consider the role of bureaucrats.
Here’s Bremer:
Here’s Latker:
Here’s Bayh:
Here’s Bremer again:
The central thrust of these drafters of Bayh-Dole is that the ultimate value of federal funding for university research is to create jobs for middlemen, for bureaucrats tasked with trying to extract profits from patent positions in the name of commercialization.
It would not be so bad if Bayh-Dole merely allowed such monopolies to happen from time to time (say, every 1,000th invention). But instead, we are told Bayh-Dole mandates it, that this is the grail, the best possible action. This ideology is then made a matter of law across all federal research contracting. That is the true horror. Not only that, but Bayh-Dole appears to be 100x less effective than the private network of patent agents that Bayh-Dole destroyed.
Even the current fuss about drug prices is focused on improving Bayh-Dole, not replacing it for drug research. That’s the tyranny of “uniform” policy that is really *compulsory* policy. Why should federal research policy regarding drug discovery have to involve all other areas of federal support for science?
In 1958, the National Science Foundation published a report on the effect of federal funding on universities, Government-University Relationships in Federally Sponsored Scientific Research and Development. There, the NSF recommends that funding come from a “diversity of sources” within the federal government:
The report goes on to identify a variety of funding mechanisms and advocates that federal agencies use the appropriate mechanism for the purpose:
The NSF identifies three forms of project, grants to individuals, and three forms of grant to universities, to be distributed to support internal research. The report concludes that these forms provide the necessary flexibility:
Five years later, Kennedy issues his Statement on government patent policy, which follows these recommendations. The Kennedy patent policy provides flexibility when it comes to contracting for patent rights. The premise of the Kennedy patent policy is that there should be consistency “to avoid difficulties caused by different approaches by the agencies when dealing with the same class of organizations in comparable patent situations.” For this, the Kennedy memorandum that prefaces the policy is clear:
“A single presumption of ownership” is not the thing. Consistency and flexibility are. Uniformity is not:
Common principles, but subordinate to the missions of the various agencies, with flexibility for special situations. This is a uniform policy. It’s just not one that decrees a single compulsory approach to patent ownership. It favors government ownership of patents, with a presumption that such ownership provides “full public access to resulting inventions.” Private ownership is allowed for a contractor who has
This much is clear. Let a contractor have rights when it has (a) a commercial position already and (b) there is a greater likelihood that an invention will be worked than if rights are “more freely available.” That’s the judgment call, in two steps.
Bayh-Dole does away with all this, riding on a claim that a “uniform” policy requires not merely consistency but that there must be a change in the “presumption of ownership.” There are plenty of defects in the argument–not the least of which is that in 35 years, Bayh-Dole has proved to be a disaster. But another key one is that the Kennedy policy disavowed a singular presumption of ownership. The spin put on government policy in the run up to Bayh-Dole was just the opposite, that the government presumed ownership. Then, cleverly, Bayh-Dole folks claimed to change this presumption in favor of middlemen. That is, they relied on a fake claim about government patent policy, and then used this fakiness to transfer the presumption to contractors–to university middlemen, to patent brokers. Universities own in place of the federal government owning, with the same status of claim that the government has in contracting, even though the university has no stake whatsoever in the contract from the government’s side. The university is paid its direct and indirect costs–it brings nothing to the table but the release of its personnel and provision of its facilities, for which it is compensated. It even gets to keep most any equipment purchased with federal funds.
Bayh-Dole eliminates the role for federal judgment at every chance. Universities don’t have to have any technology transfer capability. Their IP programs can be great or incompetent. It doesn’t matter. They can own patents and sit on them or license them to folks who sit on them or to folks who speculate on them or trolls who shake down industry and even to folks who develop an invention for public use. It doesn’t matter. There’s virtually no oversight, no public reporting, no accountability, no way for the government to intervene. This is the shape of the federal policy crafted by Bayh-Dole: bureaucrats seeking profits must be allowed to intervene without supervision on any discovery that might have public value. Before the public may benefit, bureaucrats must first be given the opportunity to make money from monopoly patent positions. This, then, becomes the de facto purpose of federal funding: welfare for patent attorneys and patent administrators. Happy times for bureaucrats, who make money even if only 1 invention in 1,000 becomes a successful product.
The Kennedy patent policy sets out a two-step decision process. Does the contractor have an established commercial position? Will the contractor be more likely to work the invention than if the invention is made available to all via the government? The first question is pretty easy to answer. The second one might take more consideration. But the Kennedy patent policy makes the work easier by requiring the invention get used within three years of the patent issuing–about six years from the time an invention is first reported. So a wrong decision is only a 3-year problem, not a 17-year (at the time) problem.
The Kennedy patent policy does not call out non-profit contractors. They are in group 1(c)–not having an established commercial position. There, decision requires taking into account “the intentions of the contractor to bring the invention to the point of practical application.” An agency is allowed to create special regulations to permit a contractor “to acquire greater than non-exclusive rights” in any invention. That would no doubt be where universities would have their place.
Instead, Bayh-Dole breaks up this ontology of rights–government, commercial contractors, others based on their intentions, and agency discretion to create special programs–and replaces it with a law of middlemen, regardless of their capabilities, intentions, or practices.
Remember those 28,000 federally owned patents that Bayh-Dole advocates complained were being “commercialized” only about as frequently (5%) as were inventions owned by universities under the IPA program? Here’s another view of it: 95% of federally owned inventions were placed in a federal commons for all Americans to use, and only a few thousand were allowed to be used to form limited monopolies, and only then to acknowledge established commercial positions and the use of private risk capital.
I’m not seeing in this approach a big problem. If there was a problem in 1978, it was that the patent counsel for NIH did not agree with the director of HEW over how big pharma got access to work done with federal funding. There did appear to be problems. There were uncertainties in decisions, and boycotts, and all the like. But those are local issues. The resolution did not have to be a law that swept flexibility aside in favor of compulsion or disregard for contractor exploitation of patent positions.
Consider, then, that the problem with drug prices based on compounds discovered in federally funded research is now stuck in a debate that has to involve all federal contracting, whether it is mission-directed or for basic research, whether it is in weapons research or organic chemistry, whether it is with nonprofits or small companies or large companies, whether the companies have commercial positions or not, whether they operate as contract research organizations or as venture-backed startups. All that matters under Bayh-Dole is that middlemen get a first crack at exploiting monopoly positions for profit, and they can take as long as they like to do it, and they can use any patent cleverness they want to (at least, in the faux Bayh-Dole version of things).
How this is a meaningful federal innovation policy is beyond me. The reality is, Bayh-Dole is a clever ruse. It was built on political frauds–that’s to be expected. But it has been maintained by disregard for even its own terms and billed as a success using flaky proxy data in place of the data that were indicated (such as date of first commercial sale or use–never, ever reported). And the best we can make of the outcomes is a massive sequestering of research results under patent claims, a terrible licensing and commercialization record, and a huge bloat of bureaucracy. There are “success stories” of course. But at what expense? Ah, a record year for the plantation cotton.
Bayh-Dole is welfare for patent attorneys. It’s no way to run federal research innovation policy. The Kennedy policy was much better. And it would still be worth returning to Vannevar Bush’s insights before we rush off in some other direction.