We are done with Reimers’s op/eds. Reimers makes things up. Reimers grossly distorts Bayh-Dole. It’s politics, so that’s what people often do. Reimers doesn’t expect any consequences if he is caught out. But we don’t have to believe his fakery. That’s on us, if we do. But Reimers is a bug in the grass, as far as the Rabbit King is concerned.
Let’s look at the big design things that Bayh-Dole has got wrong. It’s not that Bayh-Dole could never work–it might work, if federal agencies complied with the law and enforced the patent rights clauses, and if NIST had a friggin’ clue about how to interpret and implement the law, and university patent administrators were not fixated on exclusive rights or nothing.
1. Bayh-Dole relieves federal contractors from having to make a case for how their exclusionary control over an invention made in work pitched to the government for funding as in the public interest better serves the public than does open access.
The argument here appears to be that the public interest will never be served unless a university puts each invention behind a paywall and makes the invention available to only a single company, if at all. This is a nonsense argument. It doesn’t hold up to practice. Companies all the time work with inventions that are not under their exclusive control. Roche made a commercial product without exclusive rights from Stanford. All Stanford did was nothing with its share of rights but sue Roche after Roche turned down a Stanford license.
What best serves the public? It is a challenging question. But what’s interesting is that this is not a question to be fussed over by patent policy wonks. What best serves the public in a very real sense is a question be debated by the public. It’s difficult to believe that the public would settle on the result that they only want access to publicly supported inventions in the form of commercial products under a patent monopoly ceded by the government and enabled by a university, sold at whatever price a company wants to dictate, with whatever supply of product is convenient to the company, without multiple sources, without competition, and without any right of appeal. Making the public case is crucial.
2. Bayh-Dole makes everything about an invention and its development a government secret.
Truth needs sunlight, and so do good IP practices undertaken in the public interest. Bayh-Dole as originally passed permitted federal agencies to withhold information that companies were required to provide if that information was otherwise “privileged” or “confidential.” But in 1984, Congress amended Bayh-Dole to change a “may” to “shall” and that changed the law to mandate that all information received was to be treated as except from FOIA, even if it wasn’t. Well, that’s a terrible thing, as there then is no way to determine whether a federal agency is failing to act to protect the public from nonuse or unreasonable use of a subject invention. At the very least, the date of first commercial sale or public availability should be public, and with it documentation of the terms under which the product or availability is offered to the public. The public should be able to access a list of subject inventions patented by federal contractors, and for each see if there’s been documented use, and as of what date. This should be easy. Really. And ought to be done.
3. Bayh-Dole provides no mechanism for public oversight or public appeal.
The Supreme Court in Stanford v Roche found the lack of protections for inventors in Bayh-Dole potentially “deeply troubling”–saved only by interpreting the law narrowly in order to save it. There’s no mechanism for the public to request march-in. Any information gathered by a federal agency regarding invention use, development, licensing, and terms becomes a state secret–so no public discussion of what’s going on, no way to see whether the law is being enforced, or whether federal agencies are complying with the law, let alone federal contractors.
4. Bayh-Dole fragments IP ownership in distributed research work.
Maybe if there were only one federal contractor and only one company, everything would be peachy. But in any given area of research, the federal government spreads the work around. Multiple federal agencies give grants and contracts to multiple research contractors, companies and nonprofits alike. If every contractor has the right to keep exclusive patent control over every bit of inventive work they acquire in that research, the rights in the overall project get fragmented across multiple organizations. It gets worse when the universities involved all demand an exclusive license or nothing. If they each license to their own paper startup, then those startups all face infringement claims by all the other startups for anything that requires more than a single patent–and most things do rely on multiple inventions.
The thing that enables startups to work quickly and effectively is access to technology platforms, standards, open data, a robust public domain, and a robust trade in cross-licensing. I once worked with a nanotechnology company. The CEO was proud of the fact that over two years he had obtained 20 licenses from universities that allowed the company to make product. His competitive advantage was that no other company would be able to get licenses, even non-exclusively, in much less time. The ineffectual licensing programs at universities created the barrier to entry–not actually having exclusive patent rights. In nanotech, universities patented up every aspect of the carbon nanotube that they could imagine. Scores of patents. No one has been able to commercialize anything–we are. in the long wait for all the patents to expire, and then we will see what sort of commons there is, and whether any of that research is still useful. It probably is, but the inventors have mostly moved on, and any of that work will have to be redone.
Fragmented ownership of IP is a disaster for innovation if all or most or even just many of the owners of inventions are fixated on extracting rents from industry by means of exclusive licenses. If the owners of inventions go non-exclusive, especially on fair, reasonable, and non-discriminatory terms, then everything can work, even with the bother of paperwork for licenses. Bayh-Dole enables the disaster. University administrators don’t have to default to exclusive licenses–and every so often, one doesn’t–but by and large university administrators believe that without exclusive patent rights, no company will develop a commercial product based on the bit of invention each university controls. Maybe it’s self-delusion at this point. These folks listen to Reimers. Reimers set up the Stanford OTL, then he reworked University of California’s OTL, and then MIT’s. These folks push the monopoly meme. But they don’t tell you how the monopoly meme works when research in a given area is distributed to fifty universities across five funding agencies over twenty years. Well, it doesn’t work. It screws everything up. But all the OTLs can say is how hard technology transfer is. They draw diagrams of the technology transfer “process” that shows one invention, one patent, one marketing effort, one “licensee,” and then–whoa–money! Rarely do these diagrams show a product, and never a product available on reasonable terms. And never, never do these diagrams ever show scores of inventions at scores of universities all entangled and overlapping for use in any given happy innovative new life transforming product, or even a better spam filter or a clever new digital watch face. Reimers is wrong about Bayh-Dole providing an “easy way.” Bayh-Dole, the way it is represented, is a hard way to go. An impossible way.
5. Bayh-Dole authorizes federal agencies to grant exclusive licenses (to the point of assigning federally owned inventions to private firms).
Permitting federal agencies to deal in patent monopolies is a huge breach of public trust and destroys much of the safe harbor and robust public domain created by federal open access under the 1947 Attorney General’s report recommendations. When federal agencies acquired inventions, they then released those inventions for public access–mostly without licensing requirements. Not even paperwork. It all made sense. Inventors could be assured they would always have access to their work. They could move to industry or start a company and still make, use, and sell their invention. People with knowledge moving to new situations to practice that knowledge is one of the *easier* ways to transfer technology.
But Bayh-Dole screws all that up by authorizing federal agencies to deal in patent monopolies. When the government acquires patent rights now, under Bayh-Dole, it can put those rights behind its own paywall and play favorites with what company it chooses to exclusively license to, if the agency licenses at all. Stuff that would have been open, whether patented or not, becomes off-limits. Talk about “uncertainty” in dealing with federal agencies. Will the NIH license non-exclusively on FRAND terms or will it play favorites with some company and cut everyone else out? There’s no guidance in Bayh-Dole to make such decisions “uniform.” And so they are not. Can you see any way that, in the general case, our dearly desired innovation will come about better and faster when federal agencies play favorites over exclusive licenses, especially given the distributed nature of federal research enterprise and the fragmentation of rights among agencies and contractors–not to mention all the stuff going on outside the scope of federal funding and patent fiddling? I sure can’t.
The threat that a federal agency could assert exclusive control over any given subject invention and prevent open access creates a weird incentive for inventors to beg universities to take the invention and do the patent work. But even then the inventors are stuck with university administrators asserting control. Many universities have policies that expressly exclude inventors from attempting to influence licensing decisions–a conflict of interest, because the inventor has a personal interest in a share of royalties, and rather than such an interest giving the inventor something of privity in the deal gets twisted to imply that inventors have a conflict of interest and must be recuse themeselves–even when they are trying to license back into their startup what the university has taken from them. It’s so very messed up. There’s no open option in Bayh-Dole unless a federal agency gets up the courage to determine exceptional circumstances, for a given grant or grants program, and stipulate open access. No agency has done so. Bayh-Dole does not require any agency to do so. No agency will. And so Bayh-Dole effectively has plugged up many inventions made in federally supported work from open access. When universities take ownership of such inventions, they don’t default to open access, and if after three or five years of trying to get an exclusive license and have failed, they don’t then open the invention up for access. They would rather sit on the patent until it expires than let the rights go.
In one situation, a university was sitting on a set of patents that they hadn’t licensed in fifteen years. The patents turned out to be important. A who’s who of tech companies had patents that cited the university’s patents. Lots of citations. These were fundamental patents, and likely lots of companies were practicing under the claims, or would if they patents were opened up. Usually if a university suspects folks are using inventions unlicensed, they boil up into patent trolls and go after them. That’s pretty much what the Washington Research Foundation did with its bluetooth patent. But I proposed a different approach–for every company that has citied the university’s patents in its own patent filings, send the company a royalty-free non-exclusive license with assurance that the license covers any and all past practice of the university’s inventions. Along with the license, offer companies access to the inventors, their expertise and data and prototypes. Use the license to get the companies’ attention, and propose working collaborations that might result in service contracts for the university, or commons-based research. But the university’s TLO would have none of it, and so it lost the opportunity to collaborate, lost the opportunity for its inventions to come clear of covert use, and lost the prospect for making money–I mean, developing new sources of support.
In another situation I’ve been working on now for more than two years, a nonprofit’s TLO refuses to allow some of its doctors share ideas with doctors at other institutions and with companies to try to find treatments for their patients suffering from a fatal disease for which no effective drugs (or other interventions) exist and the normal drug development pathways simply don’t work–animal models fail, and there’s this blood-brain barrier problem. Why does the TLO refuse? Because it hopes to make money from exclusive patent licenses on its doctors’ ideas, and sharing those ideas would (so the TLO believes) ruin the chance to make millions. Meanwhile, by blocking collaboration, the TLO suppresses the likelihood that doctors working together to better diagnose their patients and find treatments for just those patients in their care (for the months the patients have left). The TLO’s position is deeply unethical, running against AMA ethics policies and against the ethos of doctors sharing ideas and information and technology to figure out what commercial companies have not been able to do, scratching around for general, all-purpose, profitable products. It may be cynical but when the nonprofit is making a heady business offering $100K surgeries to extend life by six months or so, and companies can make hundreds of millions on a chemical that might delay disease progression for even three or four months, there does not appear to be any particular rush to let doctors collaborate across organizations.
Bayh-Dole IP practice’s fixation on exclusive licensing–a mindlessly simplistic one patent to one company or nothing at all–works against initiative, collaboration, competition, entrepreneurship, public and equitable access. Works against all those wonderful things that Niels Reimers says is happening with Bayh-Dole. Reimers has it backwards. I think he knows. But instead we get an op/ed arguing just the opposite of what’s happening in university IP practice.