Julie Preston, writing in MedCity News (“Why does anybody own CRISPR? An argument against academic IP“), reports on a talk given by Michael Eisen a few days ago. Eisen has published an article at his blog on the patent fight between UC Berkeley and The Broad Institute (and its backers MIT and Harvard). “The soul of academic science is being destroyed, one patent at a time,” writes Eisen. According to Preston, in his talk, Eisen argued against “intellectual property in academia”:
Intellectual property in academia is a drain on the system. It’s a model that was ushered in decades ago with an aim to encourage innovation. Instead, it stifles the academic process with licensing costs and intellectual secrecy. The incentives it creates, Eisen argued, run counter to the pursuit of knowledge.
It may be worth disentangling some things. Eisen is discussing a patent fight, not all IP. Copyright is a different matter. Heck, the basis of open source software is copyright. And it makes sense that an author might limit how people can change his or her work and redistribute that work–call it “curatorial” copyright. It’s not “intellectual property” that is a “drain on the system.” It is something else, that Eisen doesn’t name, but I will give it a shot. The issue here is institutional control of intellectual property, and in particular institutional control of patent rights–not just patents, but the right to decide to patent, and the right to decide how to license what has been patented.
That is, the problem is university administrators who have the idea that no innovation should be made available to anyone who doesn’t pay the university for the right. In short, their vision is “No innovation pie without a bureaucrat’s thumb.”
And even here, it is not simply that institutions control inventions and the patents on these inventions and the licensing of these inventions–the problem lies in how institutional administrators write patent policy and how they choose to go about their work. It is not intellectual property on its own that stifles “academic process”–it is administrators who have made decisions regarding intellectual property. I know what I talk about. I’ve been one of those IP administrators for nearly twenty years, and worked at the University of California for six of those years (I’m not there now).
There is nothing in University of California patent policy that requires university administrators to acquire title to any invention–policy allows administrators to waive any invention within the scope of the patent policy’s claims. And there’s nothing in UC patent policy that requires administrators to license a patent exclusively, or to require payment in the form of royalties for any license. Indeed, in sponsored research agreements, UC may agree up front to license any inventions that might be made in the research royalty-free. And that’s exactly what happens with any federally supported research inventions–UC grants the federal government a royalty-free license “to practice and have practiced” any such invention. Patent administration does not have to be a “drain on the system.”
But Eisen is right–patent administration the way it is mostly practiced at universities is a drain on the system, in more ways than one. Not just the bother of reporting inventions, filing patent applications, and dealing with the problems of publishing before filing applications–but also with the consequences of granting exclusive licenses that are so broad in scope they encumber research and block the use of new findings that build on what has been patented and exclusively licensed. And that’s not all–any patent that’s not licensed but which is held for exclusive licensing is a huge disincentive for anyone to use anything published. Why use such stuff, when the university administrators won’t grant a simple non-exclusive license for use? Why even study such stuff, other than to find a way to design around it?–that is, not only not to use it, but to render it useless. I’ve talked to industry folks who routinely instruct their technical staff not to read the academic literature. It’s poison to bring in–treble damages for willful infringement–and bitterness to realize there might be something useful that one is unlikely to be able to obtain.
But Eisen is even more right. Perhaps there could be an argument that if patent licensing made enough money, then all the bother and bitterness would somehow be justified. It’s a sketchy argument. If in a time of drought I make enough money to provide clothing for orphans, then am I justified in withholding water from everyone in the city but for my favorites? You see the logic. But the problem with the money argument is that most university patent licensing offices don’t make enough money to matter–many don’t break even and have the, um, balls to argue that they should continue to do their thing because it is a public service. As if “patent licensing” is an unqualified public good, just to exist. Sure, there may be a public service in patent licensing–I argue there can be a public service in patent licensing–but not in the way most university administrators do it. Not the double monopoly approach of demanding ownership and then seeking exclusive deals. Not the non-selective portfolio approach that claims to own everything–thousands of assets per year at the University of California–and then claims success on the basis of a handful of deals per decade. CRISPR is supposed to be one of those deals. No wonder administrators are ahold of its patents with white knuckles.
And Eisen continues to be right, when he argues (as reported by Preston) that
The incentives it creates . . . run counter to the pursuit of knowledge.
Here the antecedent of “it” is “intellectual property.” But intellectual property, like a rifle, does not on its own “create incentives” that are “counter to the pursuit of knowledge.” Intellectual property offers “affordances”–places that suggest certain actions. There are a number of such affordances–the right to exclude others, the right to trade a promise not to exclude for payments. But there are other affordances, too–the opportunity to contribute to a standard, a defense against someone seeking to block broad access to improvements, a detailed public notice of an invention that raises the bar for any subsequent efforts to patent in the same area. The problem is that university administrators routinely choose to emphasize patent affordances that indeed do “run counter to the pursuit of knowledge.”
University administrators are in denial about this, of course. They will point out that a patent does not prevent anyone from studying anything. Research can still go on, and new things discovered, regardless of any patent. And that is so true. Most university faculty don’t bother to study the patent literature before deciding what to investigate. They just propose projects, get funding, and do the work. It may be that they are infringing stuff here and there, but the problem with patent infringement, at least, is that one is exposed to actual damages–and the cost of suing for patent infringement is often way, way more than any possible damages. There’s no “research exception” in patent law, but in practice there’s a limit to what patent owners looking to “monetize” their patents are willing to fight about.
But Eisen (I’ll venture) doesn’t mean merely that patents “run counter” to conducting research. He uses “pursuit of knowledge” to stand for the whole academic endeavor–to conduct research, to publish findings, to build on the work of others, to find ways to apply that work for the benefit of one’s society, and even to find ways to benefit from being involved in those applications, if that’s what one wants to do. Monopoly-based patent management destroys academic choices. A patent application is a publication, but when university administrators demand to own an invention and insist on seeking a patent, they then force faculty to publish their inventions in the patent literature. That’s not one’s academic freedom to publish. It is administrator-dictated publication. And academics don’t have a choice whether to publish openly or take a patent position. Again, that’s an administrator’s decision.
Once there is a patent, the academic investigator has no say in whether the patent is made available non-exclusively and royalty-free (har, har) or licensed exclusively or assigned (often licenses labeled “exclusive” are in fact assignments). There’s further little choice in the term of the license, the scope of the license, the diligence required to make something into a product, or the structure of payment. Public universities even go so far as to claim that inventor efforts to influence university licensing practices are unethical–the inventor has a conflict of interest because the inventor might be paid a share of any royalties. No one bothers with the case in which the inventor does not want a share of royalties and does not want there to ever be any royalties. That, too, gets dismissed as unethical–taxpayers deserve a “return on their investment” and that “return” must (according to patent administrators) take the form of money from licenses. Not that any of that money ever gets to the taxpayers. Bizzare.
To show just how little the money from patent licensing matters, no university I know of publishes a regular account of how licensing money is actually used. Yes, we know that much of it goes to cover the livelihoods of all the patent administrators, and more goes to pay a “share” to “inventors” (but not to investigators who didn’t invent, and not to investigators who published the work that made the invention possible, and not to research staff who washed all the glassware, and not to the graduate students who were assigned a different role in the project). But the extra–the amounts that Bayh-Dole’s standard patent rights clause says must be used for “scientific research or education”–we never see where that money goes, or what results from its use. That money is so uninteresting that it goes unreported. No one cares about it. Thus, it is difficult to maintain an argument that the money in patent licensing justifies the suspension of academic freedom and the disruption of the pursuit of knowledge–where the point of that pursuit is the use of that knowledge without having to wait to pay a university administrator aiming a patent rifle at you.
While I don’t have direct access to Michael Eisen’s thought processes, I will venture that the problem is not “intellectual property” but rather “university administrators misbehaving with intellectual property.” Or, another way, intellectual property does not by itself run counter to the pursuit of knowledge, but university administrators use intellectual property to disrupt the pursuit of knowledge. What is the point of pursuing knowledge if only administrators’ favorites get to use that knowledge for twenty years? It’s a strange argument for scientific “progress” that a patent serves as incentive for everyone who is not the chosen favorite to design around the invention, undermine it, dismiss it, and otherwise hope it fails.
We come now to our obligatory Bayh-Dole section:
The model traces back to the Bayh-Dole Act of 1980, he said. Prior to this, the federal government owned discoveries made through the research it funded.
Prior to Bayh-Dole the federal government had flexibility in ownership of inventions under the Kennedy/Nixon patent policies. Companies with active businesses often had the option to own. Nonprofits that didn’t have any business could still own if they had an Institutional Patent Agreement with a federal agency (HEW or NSF). Others could own if they made a persuasive case to the federal agency that funded the work. There was a uniform policy that argued for flexibility. Flexibility, in turn, depended on administrative judgment, and in government, the use of judgment means long processes, slow decisions, and variations depending on circumstances and prevailing political winds.
Government-owned inventions were to be “dedicated or licensed” to the public. The government, then, decided whether to seek patents (and therefore license to the public, generally without a profit motive) or not (and so dedicate the invention to the public domain). Thus, the federal government’s “ownership” of inventions was not in anyway ordinary–certainly not like the “ownership” of patent that gives the owner the right to exclude others. The federal government policy was to take ownership of inventions to preclude the exclusion of others (except in certain areas, such as atomic power, nuclear weapons, and space craft–where there might be reasons to exclude others, but owning a patent was not the only way an agency had to do that). Thus, patent “ownership” for the federal government did not mean anything close to what patent “ownership” means in general.
This is one of the clever bits in university administrator’s misuse of language. When universities were permitted under the IPA program to “own” inventions, their ownership was restricted (they acquired what was termed “principal rights”) and their use of that ownership was stipulated by federal contract: non-exclusive licensing by default, show cause for any exclusive license, limit the term of exclusive licenses, show cause for any extension of that term, allow the federal government to step in at any point if the university is not doing a better job than the government or isn’t doing what the government needs to do. Even dark-hearted Bayh-Dole retains some bit of this goodness in its statement of policy and objective, its definition of practical application, limits on exclusive licensing and assignment, and march-in procedures (though these procedures were subsequently in the implementing regulations designed not to function).
But when university administrators use patent “ownership,” they want people to think of ordinary patent ownership, where the patent owner gets to try to make money, can play favorites, can exclude everyone from using an invention, can sue anyone who does use the invention. That’s not what the federal government’s ownership was, and it wasn’t even what university ownership practice was, for federally supported inventions. More Eisen, via Preston:
The problem was that the government wasn’t good at commercializing and using those inventions.
This was the argument made by the patent brokers. It doesn’t hold up. First, the government’s purpose was not to “commercialize” inventions, but to make them available to everyone. If a contractor had the market position to “commercialize” an invention, then the executive branch policy was to allow the contractor to have “principal rights” to the invention and get on with it. Many of the inventions the government ended up with patenting were weapons-related inventions that the company contractors had declined to own. There was no point in seeking to “commercialize” such inventions–the contractors had decided that federal government dedication or royalty-free licensing was preferable to taking out an exclusive position.
The point: the federal government was not attempting to “commercialize” its patents. It was attempting to make the underlying inventions available broadly to the public. Anyone could “commercialize” anything. The government’s position was that it should not play favorites with patent rights. What’s odd is that the government’s reported commercialization rate was about 5%. Universities argued their rate was 25% or better. But what the universities didn’t bother to point out (and so doesn’t get quoted by those hunting for convenient references) was that the federal government’s commercialization rate for biomedical inventions was 23%–comparable to the universities’ claimed rate (which was really a research foundation rate, because most universities didn’t have licensing offices of their own).
What the universities also did not point out was that in the IPA program, where they managed federally supported inventions, their commercialization rate was pretty much the same as the overall federal government’s rate. That is, the universities did about 5x worse with federally supported inventions than they did with inventions that were brought to them by faculty inventors who had made the decision to pursue a patent and wanted the university to be involved (usually via Research Corporation or a university-affiliated foundation). And about 5x worse than the federal government for biomedical inventions. When asked about their relatively lousy results in a Congressional hearing, university officials admitted that they weren’t doing so good, but that there was great potential to do better, if only they were given the chance. It’s that great potential argument that gets repeated, and as Dave Henderson, the baseball outfielder, once quipped, “Potential means you haven’t done anything yet.”
The group that objected to the federal policy of “dedication or non-exclusive licensing” was pharma. As research in medicinal chemistry discovered compounds with potential therapeutic effect to treat infections (vaccines), deficiency diseases (vitamins), and cancer, big pharma boycotted compounds discovered with federal government support. Monopoly or nothing was their demand. They got something of what they wanted via the IPA program, which was designed by patent counsel at the NIH to end-run HEW policy on non-exclusive rights for everyone. NIH could circumvent HEW policy by outsourcing medicinal chemistry research to nonprofits, and then via the IPA program allow the nonprofits to license exclusively to pharma companies, who then could enjoy monopoly pricing and profits for the term of the patent (in the IPA program, of course, non-exclusive licenses were the default, and there were term limits on exclusive licensing, but when folks at HEW looked at the data, they saw that the nonprofits almost always granted exclusive licenses and advocated for those licenses to run the full term of the patent–Cisplatin-like).
A Senate committee beat back an effort by the NIH to make the IPA program government-wide. HEW then shut down their IPA program (in 1978) as (as best I can tell) a rogue operation that undermined HEW policy. Bayh-Dole was the rogue response by NIH patent counsel (who had someone else type up the final draft of the law to prevent the draft from being traceable to his NIH typewriter–to give some sense of the degree of covertness in all of this). Bayh-Dole was a work-around by big pharma to a flexible government policy that started with the idea that inventions made in subvention funding ought to be available to the public, and if any nonprofit could do a better job than the government in making the benefits of inventions available (to the public, on reasonable terms), then the nonprofit could put forward a plan and acquire principal rights. Bayh-Dole blew up this approach and replaced it with what pharma wanted–monopoly positions, for which it was willing to pay a royalty of 0.5% to 2% to the university middlemen who secured patents on the inventions, thereby preventing them from entering the public domain through academic publication.
Everything else about Bayh-Dole is pure bluster–that it is government-wide (“uniform”) is just to cover that the only industry that matters is pharma. That it will stimulate innovation is also pure bluster–the innovation was happening as a result of the academic pursuit of knowledge. What wasn’t happening was the preservation of monopoly pricing and profits from the results of that pursuit. The devil’s bargain between university patent brokers and pharma–now into its fourth decade–was that university administrators would take ownership of faculty inventions in return for their share of royalties–$500m here or there, once every ten or twenty or thirty years per research university. And that was enough to buy out the administrators, have them change patent policies, and bring in patent brokers to make their vision of using the threat of patent infringement to force industry to pay up–not just pharma, but any industry, anywhere, regardless of its structure or management of innovation.
With the Bayh-Dole Act, ownership rights were transferred to the institutions in the hope that it would encourage innovation and the practical use of scientific breakthroughs.
Here we get the standard misrepresentation of Bayh-Dole. Bayh-Dole transferred no ownership rights to institutions. That’s the Stanford v. Roche decision. (For what it’s worth, I was involved in drafting amicus briefs that beat back the university claim that Bayh-Dole vested ownership of inventions with the institution). All Bayh-Dole did (and does) is restrict federal agency contractual claims on any invention made with federal support if an institution acquires ownership of that invention. And those “ownership rights” came with stipulations that made it clear (despite efforts to make it muddy) that patents on federally supported inventions were not ordinary patents, and universities could not do just anything once they came to “own” those abynormal patents.
Bayh-Dole, made a part of federal patent law, states as its first “policy and objective” (35 USC 200) that the patent system should be used “to promote the utilization of inventions arising in federally supported research or development.” That “utilization” is formalized in the definition of “practical application”–(I paraphrase here) to practice an invention such that it can be established that the invention is being used and its benefits are available to the public on reasonable terms. There was more than “hope” involved–Bayh-Dole stipulates that a patent on a subject invention (a “posi,” if you will, both singular and plural) be used in certain ways–to promote use of the invention, to promote free competition, to use American labor–and not in other ways, which would work against use or competition or American labor. It’s just that university administrators ignore these provisions, treat posi as ordinary patents, and federal agencies turn a blind eye to everything. There is little to show that most posi meet the standard set by the definition of “practical application”–the benefits of their use are not available to the public on reasonable terms.
Thus, even the argument that somehow despite all the administrative bother and disruption of the pursuit of knowledge and lack of royalty income that makes enough difference that anyone much cares, we don’t even have an argument that society benefits, despite all these awful hurts inflicted by university administrators (and their legal advisors, and their patent brokers, and their university innovation “boosters” and speculators) in their patent management efforts.
Though I’d put it differently, then, Michael Eisen is spot on. What many–not all–university administrators are doing with intellectual property does run counter to the pursuit of knowledge. Universities have been made by these administrators to work against the very core of their public research mission. It’s a sad day, year, and three decades of administrative abuse of intellectual property and research, all to have a livelihood being the petty tyrants of patent rights obtained under the color of law. The CRISPR patent dispute just illustrates how bad things have got–even when the inventive work is hugely important.
in the long run, science would be accelerated and made more efficient if the IP barriers were removed.
Not sure about this–“in the long run” is a needless qualification. What Vannevar Bush wanted (and Bush, in many ways, started the federal government down this path, though he advocated for something rather different) was to expand the “frontiers” of science–to put more scientific discoveries on the tool board where they could be used by folks working outside of established orders to develop new things and industries that those established orders could not imagine, would not specify, and would be unwilling to support until proven out. So maybe what Eisen means is that research would go faster and more efficiently if university administrators did not interpose so much legal fussiness regarding patents–that is, the barriers the administrators create using IP under their control. There’s a bunch written about “accelerating” company formation and product development and not a whole lot of work in how to “accelerate” research. That’s something that ought to be studied (we started a study of research acceleration as part of my Kauffman Foundation grant, back in 2010).
Preston ends her article with an insightful comment:
Whether Eisen is correct or not, the overarching idea is is that we haven’t been interrogating the current model to determine whether or not it works. There is very little data and, as a result, very little debate.
The dismissive opening clause isn’t needed–Eisen gives voice to a situation many experience. But the main clause lays out the problem: there’s no data, no debate–university administrators get away with what they are doing, and few folks challenge them. When I was active on Techno-L and involved in AUTM, I found that university administrators would run and hide from any debate on these issues. Everything was devoted to spinning up a story of Bayh-Dole’s success, the public benefits of patent licensing for commercialization, and the need to make the acquisition of patents (and their disposition) more efficient–and hence the patent policy changes, ending up with the grotesque efforts to include “present assignments” of future, as yet unmade inventions in patent policy statements. (For which NIST will soon make a matter of federal contracting by turning Bayh-Dole into a vesting statute by twiddling with the standard patent rights clause–sigh, the more things change, the more they get insane).
Bayh-Dole requires that federal agencies contract to have the right to request reports on the use of subject inventions. But Bayh-Dole does not require agencies to request those reports and Bayh-Dole also requires agencies to keep secret any reports they do receive (well, technically Bayh-Dole doesn’t do this, but who gives a rat’s ass what Bayh-Dole technically requires these days–it’s a do WTF you want sort of law, for the convenience of university patent administrators and the patent attorneys who make their livings from filing the thousands of useless, obstructing patents that university patent administrators order up each year, creating opportunities for the patent fights, such as we see with CRISPR). So we have no data. And federal agencies have to keep mum about it.
Whatever the Bayh-Dole advocates say goes generally unchallenged. Sure, some academics have spoken out–Mowery, Eisenberg, and Rai among them. And of course, Research Enterprise has been working for nearly nine years to get a discussion going–it’s just that there are few within the university technology transfer community willing to challenge publicly the prevailing view of the Bayh-Dole advocates. I worked for nearly two decades in that community. Many of those folks fear for their jobs, their reputations, their careers if they come out in favor of a change. Their own legal advisors work against them and push nasty IP policies to senior administration. It’s time to have a discussion that gets at the realities–that looks at the data, that reports what is happening, not merely some spin about the potential of what’s happening.
Michael Eisen gets at some of what is happening. He argues that IP is the problem. I think it is university patent administrators that are the problem–uniformed, or befuddled, or fooled, or swallowing a dream about money happily bound up with public purpose, or filled with a petty desire to be obeyed regardless, or throttling those that would have the discussion.
I hold out hope that IP management can accelerate research, can make science more efficient, and can help the public realize benefits from academic research. But to get there, we have to do things way, way different than we are now. We might start with allowing faculty choice regarding IP–and that means investigators as well as inventors, based on the commitments they make to each other and to public science. We could follow by creating make/use commons as the first order of business. We could then within those commons develop the platforms and relationships that lead to opportunities to apply findings and make benefits available all sorts of ways–whether commercially or through user communities or both.
To get there, we need forums in which actual information on university IP management is made available. Not echo chambers that repeat the same truisms. Holy cow. Even if Bayh-Dole was the “most inspired” law in the past “half century” (according to an anonymous quip in The Economist), that half century was from 1952 to 2002. By 2005, even The Economist was walking back on its assessment. Bayh-Dole has not been a wild success, that much is clear. Most university inventions go unlicensed–80% or so. Only 0.5% or so apparently become commercial products at all. We are talking commercialization rates 10x lower than the federal government’s rate prior to Bayh-Dole, and the federal government wasn’t even trying. University patents are held for exclusive licensing and thus administrators deny the public broad access to key research findings. It’s worse than the dam in Madagascar 2 and not at all funny.
The starting point is discussion. Then real, actual data. Then make the changes. Bayh-Dole doesn’t have a thing to do with these changes. Bayh-Dole doesn’t prevent anything. Repealed, reformed, or left to rot, Bayh-Dole isn’t the problem. It’s the mindset among university administrators and especially their legal advisors that refuses to let go of a vision of money from patents, spun as a public good. That’s a seductive vision. And the way it has played out has been clueless, wrong, bankrupt (often literally), and corrupt (at times).
Let investigators choose whether to be open or proprietary. Allow them to make and honor commitments to public science and to the public. Create commons and platforms and only rarely patent monopolies. Build relationships. Follow opportunities. Make academic researchers’ future federal funding depend in part on how they have managed the IP arising in past research. Did they enable broad use of tools? Did they help people learn to use discoveries? Did they publish all their data? Did they create commons or contribute to standards? Did companies spin out from these commons? Did new science get done as a result? Did the public enjoy some benefits on reasonable terms? Is all this too difficult? Perhaps, for the plantation owner.
University administrators have roles to play in all this, as they control access to institutional resources, and those resources count for something. Absconding with IP for monopoly games is not one of those roles. The CRISPR patent fight is just one more way–a helpfully visible way–that the prevailing approach by university administrators to IP management degrades research and damages the position of academic research in the vision of how we improve our society by the pursuit of knowledge.