Research Enterprise

Let’s be clear. Bayh-Dole doesn’t transfer ownership to grant recipients. Grant recipients obtain ownership when they exercise protocols available to them under Bayh-Dole, and only then when they choose to become the owners. The transfer of ownership in inventions is from the inventors to the grant recipients or to whomever the grant recipients may designate. That happens by assignment. Bayh-Dole allows grant recipients to step ahead of the Federal agency and have standing to require the assignment of patent rights from the inventors. Huge difference. It’s not the law that does this, it is an administrative choice. Universities do this to themselves.

Furthermore, the effort of the Act is not to move research technology to the public domain, but rather to use the patent system to promote the use of federally supported inventions. That’s not a policy of the public domain but of the patent system. Perhaps the authors meant by public domain “public benefit”. But if so, then it’s sloppy, since “public domain” is a term of art in patent work.

In the case of medical technology, which the authors raise as an instance, contra the assertion they make, those doing the development may be especially well trained and equipped to teach and enable the use of what they develop. How could one say otherwise in general? That the surgeon who develops a new stent isn’t qualified to continue to work toward widespread use? No, there’s a difference between mass production and regulatory approvals and being ill equipped to provide the public with access. But these are nits, perhaps. I don’t see in Bayh-Dole a goal of *increasing* public access. Check for yourself: here. It may be that an argument used to pass the law was that there would be more public use, but it’s not in the law. There’s a difference between arguments used to pass a law, and what the law itself sets out as the law. Hmpf.

The upshot of this paper is that federal contractors have an implied duty to partner with industry to commercialize “promising federally funded research”. They don’t. They have a choice about it, not a duty. They don’t have to commercialize at all. They don’t even have to collaborate with industry. They can hand stuff off to others that do this. They can grant a royalty-free public license. That’s not commercializing. Companies can just make and use the inventions, not making products at all. That is, use by commercial entities is not use in commerce. Methods to improve safety or production. Not commercialization. Products? Not required. It’s a choice.

The paper has many flaws. It’s not my place to work through them all. The central flaw, however, is the displacement of practical application with commercialization, and then a selective slog through the law to find support for this reading. Badly reasoned, poorly annotated, not persuasive. I’m sorry.

Bayh-Dole lists a number of objectives. Commercialization is mentioned but it is given little room in the law or its implementing regulations. Thus, I guess the need to squeeze something implied out instead. Practical application, however, plays a much greater role, is given a definition, and tracks the overall goal of promoting use. The law also encourages nonprofit-industry collaboration, especially through small companies, and support for US manufacturing. But commercialization is not the same as practical application. It cannot be because both are present in the law, and there is no indication they are intended to be used interchangeably or synonymously. Just as practical application is not simply or secretly commercialization, so also collaboration isn’t secretly and implicitly licensing patent rights for product development and sales. Practical application encompasses all sorts of beneficial uses, and does not require that products be made and sold. A company might self-implement an invention, realize benefits, and pass these along to the public, without any product being made whatsoever. And companies might collaborate with nonprofits to do this. The internet is an example, with many inventive technologies federally supported and deployed without commercialization. That’s because they became standards, not products.

We see the authors of our paper conflating commercialization and practical application: “In addition, the Act mandates that contractors take necessary steps to commercialize any discoveries or inventions resulting from federally-funded research….” This mandate is not in the Act. It is an objective but the mandate does not fall on the contractors to commercialize, but rather makes a condition of taking title to use the patent system to promote use.
The Act: “to promote the commercialization and public availability of inventions made in the United States by United States industry and labor”. This is one of eight or so objectives along with, for instance: “to promote the utilization of inventions arising from federally supported research” and “to ensure inventions made by nonprofit organizations and small business firms are used in a manner to promote free competition and enterprise without unduly encumbering future research and discovery” and this: “to…protect the public against nonuse or unreasonable use of inventions.”

There is an express duty to meet the conditions of the law. The law is implemented by means of the standard patent clause at 37 CFR 401.14. It is placed in the context of federal research policy by OMB Circular A-110. A university comes within the scope of the law when it accepts federal funding carrying the obligations of 37 CFR 401.14 as a federal contract. At that point, the university has a good faith obligation to comply with the contract to achieve its objectives–not only 37 CFR 401.14 but also Circular A-110. The law creates the foundation for regulations that establish a standard Federal patent clause. If a contractor elects to retain title in a subject invention, then the contractor necessarily accepts the additional stated duties with regard to reporting on utilization, filing patent applications, and the like. Apparently the authors want it to be that regardless of the express statements of the Act, the implied duty is that only one of the objectives of the Act really matters objectives–to “commercialize” inventions. One can assert this. One can insist it matters. But it’s not an implied duty. It’s an asserted one.

The authors want march-in rights to be more fodder for the implied duty to commercialize, but even in their development they use “practical application” which is a different, broader thing, of which commercialization is only one way. But march-in rights are much more directly dealt with in the Act, to protect the public from non-use or unreasonable use. If there is use, and it is reasonable, we are done. But not if there’s an implied duty to commercialize, or, that the law hints at something it doesn’t come out and say but that anyone reading this paper would realize is also part of the law. The authors call this “use it or lose it” policy. Yes, but they never can connect up use and commercialization.

We can see how different use and commercialization are in the Act by looking at the chain of decisions that result in the inventors retaining title to their inventions. If the contractor waives the right to retain title, and the federal agency chooses not to require assignment of title, then title stays where it always was, with the inventors. Under the Act, the inventors are required to elect title and file, including in patent applications a statement of federal funding, report on utilization, prefer US industry in exclusive licenses, and accept march-in provisions in the event of non-use, unmet health or safety needs, public use specified in federal regulations, or to deal with exclusive licensees not manufacturing in the United States. If the inventors fail use, the remedy is a compulsory license required by the government with “terms that are reasonable under the circumstances”, not a loss of title. If the inventors fail to file, that’s a different matter, but that is not “use it or lose it”–it is “file it or waive it”. Failure to commercialization does not cause a loss of title, but may create an obligation to accept a government imposed license (with reasonable terms, including one might infer consideration). The difference is pretty clear.

The authors argue that that “ultimate goal” of Congress is that “federally funded inventions are made available to the public, for the public’s benefit.” By “ultimate” we must infer “not stated in the law, but sure helps our argument if you pretend it’s there.” Again: the Act says that the objective is to use the patent system to promote the use of inventions made with federal funds, with some things that follow from that. If the government wanted the public to have access to the inventions, it would have required publication of all invention reports and access to the inventive labs for a period of time so the public could come and gawk and fill their pockets with as much access as they could carry. No need for patents. No, patents serve a national interest–inventions made in the US result in US jobs. That’s what’s clear in the language of the Act, whatever purposes we wish to assign to the Congress that passed it into law.

The authors conflate invention, technology, patents, and products. Making it simple like this gives them a plausible argument to discover implications. But the conflation does a disservice to the Act and to technology transfer practice. Where it may take a lot of specialized resources to get a new drug approved, it may take very little to implement a new disease assay, and in fact, we find that university lab medicine personnel are just as good if not better at doing so than their industry counterparts. While the lab medicine doctors might not be able to come up with branding, or restrict the invention to a particular product form easy to manufacture and designed to maximize profits, they are way better at practical application and are much the better route to public benefit.

I know this is all sort of slant from the endpoint of the paper, which is to argue that Bayh-Dole represents a societal statement encouraging universities to work closely with industry despite cultural differences and the prospect for conflict of interest, and that with thoughtfulness these challenges can be met so that federally supported inventions get commercialized. One might even think the implied duty to commercialize is somehow intended as a federal pre-emption to concerns about conflict of interest or an indifference to commercialization, especially among biomedical researchers.

I think it is quite fine to construct a societal argument for the importance of commercialization. This can be directed at the role of industry in deploying medical knowledge through technology. One can look at the rise of new technology with the potential for undertaking medical interventions. One can ask how markets and investors create value that might focus development work to move from bench to bedside. These are decent things, and worthy. It’s a choice to do these things, not an implied duty under the law.

There’s just no reason to squeeze implied duties out of the law to get there.

Looking at this paper from 2002: “Academia, Industry, and the Bayh-Dole Act: An Implied Duty to Commercialize”. I’ve seen it before and always shied away from commenting, but with all the buzz about Bayh-Dole around things like Stanford v. Roche, maybe there needs to be some close look at this kind of article. In general, I have to say that a lot of this stuff is sloppy. The peers reviewing this kind of work need to be something other than sloppy themselves. Indeed, they need extra vigilance. Until the peers are experts, and therefore are competent judges of their own expertise and that of others, peer review is another property of a bozonet. That is, peers of an author, given standing but lacking expertise, just as the author lacks expertise, are just as likely to let through badness and demand the removal of goodness as offer comments for improvement of the article. And they won’t know it. I saw this all the time teaching writing, where in group work students asked to edit were routinely handing out really bad advice about other students’ work. So in the absence of a working peer review function, reviews have to be done in the wild and later. Here goes.

The first sentence [with my annotations]: The Bayh-Dole Act “fundamentally altered [standardized] the ownership paradigm [policies] of [regarding] intellectual property [patents] developed [on inventions made] with [within the planned and committed activities supported by] federal research dollars, transferring [giving a standard first option to elect title, with conditions] that ownership [that first option, absent a patent administration agreement] from the federal government to grant recipients (grantees) and organizations that are parties to government funding agreements (contractors) [the Act does not distinguish these--all are contractors], in an effort to enhance [normalize] the public’s [contractors'] access to technology [patent rights arising from inventions] developed with federal funds.”

That should be enough. The rest of it is wrong. There’s not an implied duty to commercialize in Bayh-Dole. Universities when they take ownership under protocols of Bayh-Dole undertake an express duty to use the patent system to promote use of the invention. There’s no secret implied meaning that promoting use means commercializing. If a university asserts this is their goal, then fine, that’s an assertion. But don’t pin it on the law, and certainly not as the secret unstated desire of the law. Hmpf.

Let’s be clear. Bayh-Dole doesn’t transfer ownership to grant recipients. Grant recipients obtain ownership when they exercise protocols available to them under Bayh-Dole, and only then when they choose to become the owners. The transfer of ownership in inventions is from the inventors to the grant recipients or to whomever the grant recipients may designate. That happens by assignment. Bayh-Dole allows grant recipients to step ahead of the Federal agency and have standing to require the assignment of patent rights from the inventors. Huge difference. It’s not the law that does this, it is an administrative choice.

Furthermore, the effort of the Act is not to move research technology to the public domain, but rather to use the patent system to promote the use of federally supported inventions. That’s not a policy of the public domain but of the patent system. In the case of medical technology, those doing the development may be especially well trained and equipped to teach and enable the use of what they develop. How could one say otherwise in general? That the surgeon who develops a new stent isn’t qualified? No, there’s a difference between mass production and regulatory approvals and being ill equipped to provide the public with access. But these are nits, perhaps. I don’t see in Bayh-Dole a goal of *increasing* public access. Check for yourself: here. It may be that an argument used to pass the law was that there would be more public use, but it’s not in the law. There’s a difference between arguments used to pass a law, and what the law itself sets out as the law. Hmpf.

The upshot of this paper is that federal contractors have an implied duty to partner with industry to commercialize “promising federally funded research”. They don’t. They have a choice about it, not a duty. They don’t have to commercialize at all. They can hand stuff off to others that do this. That’s not commercializing, it’s protecting the government’s interest. Or they can use the patent system to promote utilization. That’s not commercializing. Companies can just use the inventions, making not products at all. Methods to improve safety or production. Not commercialization. Use by companies. Not the same thing. Not required. It’s a choice.

The paper has many flaws. It’s not my place to work through them all. The central flaw, however, is the displacement of practical application with commercialization, and then a selective slog through the law to find support for this reading. Badly reasoned, poorly annotated, not persuasive. I’m sorry.

The objectives of Bayh-Dole list a number of objectives. Commercialization is mentioned but it is given little room in the law or its implementing regulations. Practical application plays a much greater role, is given a definition, and tracks the overall goal of promoting use. The law also encourages nonprofit-industry collaboration, especially through small companies, and support for US manufacturing. But commercialization is not the same as practical application. And collaboration isn’t secretly and implicitly licensing for product development. Practical application encompasses all sorts of beneficial uses, and does not require that products be made and sold. A company might self-implement an invention, realize benefits, and pass these along to the public, without any product being made whatsoever. And companies might collaborate with nonprofits to do this. The internet is an example, with many inventive technologies federally supported and deployed without commercialization. That’s because they became standards, not products.

We see the authors of our paper at hand conflating commercialization and practical application: “In addition, the Act mandates that contractors take necessary steps to commercialize any discoveries or inventions resulting from federally-funded research….” This mandate is not in the Act. It is an objective but the mandate does not fall on the contractors to commercialize, but rather makes a condition of taking title to use the patent system to promote use.
The Act: “to promote the commercialization and public availability of inventions made in the United States by United States industry and labor”. This is one of eight or so objectives, along side: “to promote the utilization of inventions arising from federally supported research” and “to ensure inventions made by nonprofit organizations and small business firms are used in a manner to promote free competition and enterprise without unduly encumbering future research and discovery” and this: “to…protect the public against nonuse or unreasonable use of inventions.”

There is an express duty to meet the conditions of the law. The law is implemented by means of the standard patent clause at 37 CFR 401.14. It is placed in the context of federal research policy by OMB Circular A-110. A university comes within the scope of the law when it accepts federal funding carrying the obligations of 37 CFR 401.14 as a federal contract. At that point, the university has a good faith obligation to comply with the contract to achieve its objectives. The law creates the foundation for regulations that establish standard federal contract clauses. Still not finding something implied. Forward. If the contractor elects to retain title in a subject invention, then the contractor as a condition accepts additional, express duties with regard to reporting on utilization, filing patent applications, and the like. Apparently the authors want it to be that regardless of the express statements of the Act, including a set of objectives, the implied duty is that only one of these matters objectives really matters–to “commercialize” inventions.

The authors want march-in rights to be more fodder for the implied duty to commercialize, but even in their development they use “practical application” which is a different, broader thing, of which commercialization is only one way. But march-in rights are much more directly dealt with in the Act, to protect the public from non-use or unreasonable use. If there is use, and it is reasonable, we are done. But not if there’s an implied duty to commercialize, or, that the law hints at something it doesn’t come out and say but that anyone reading this paper would realize is also part of the law. The authors call this “use it or lose it” policy. Yes, but they never can connect up use and commercialization.

We can see how different use and commercialization are in the Act by looking at the chain of decisions that result in the inventors retaining title to their inventions. If the contractor waives the right to retain title, and the federal agency chooses not to require assignment of title, then title stays where it always was, with the inventors. Under the Act, the inventors are required to elect title and file, including in patent applications a statement of federal funding, report on utilization, prefer US industry in exclusive licenses, and accept march-in provisions in the event of non-use, unmet health or safety needs, public use specified in federal regulations, or to deal with exclusive licensees not manufacturing in the United States. If the inventors fail use, the remedy is a compulsory license required by the government with “terms that are reasonable under the circumstances”, not a loss of title. If the inventors fail to file, that’s a different matter, but that is not “use it or lose it”–it is “file it or waive it”. Failure to commercialization does not cause a loss of title, but may create an obligation to accept a government imposed license (with reasonable terms, including one might infer consideration). The difference is pretty clear.

The authors argue that that “ultimate goal” of Congress is that “federally funded inventions are made available to the public, for the public’s benefit.” By “ultimate” we must infer “not stated in the law, but sure helps our argument if you pretend it’s there.” Again: the Act says that the objective is to use the patent system to promote the use of inventions made with federal funds, with some things that follow from that. If the government wanted the public to have access to the inventions, it would have required publication of all invention reports and access to the inventive labs for a period of time so the public could come and gawk and fill their pockets with as much access as they could carry. No need for patents. No, patents serve a national interest–inventions made in the US result in US jobs. That’s what’s clear in the language of the Act, whatever purposes we wish to assign to the Congress that passed it into law.

The authors conflate invention, technology, patents, and products. Making it simple like this gives them a plausible argument to discover implications. But the conflation does a disservice to the Act and to technology transfer practice. Where it may take a lot of specialized resources to get a new drug approved, it may take very little to implement a new disease assay, and in fact, we find that university lab medicine personnel are just as good if not better at doing so than their industry counterparts. While the lab medicine doctors might not be able to come up with branding, or restrict the invention to a particular product form easy to manufacture and designed to maximize profits, they are way better at practical application and are much the better route to public benefit.

There is a simple IP policy no university will implement:

It is our policy to take no position on any IP arising in the university unless the university commissions its creation or the proprietors of the IP request our involvement.

For one thing, such a policy is too short. IP policies need to run to at least a page to evidence heft and seriousness, and to make enough bad choices in phrasing that there is a continuing role for administrators and oversight committees to work on trying to fix things.

Here is another IP policy gesture:

It is our policy to support creative insights when they intermittently arise, with the resources we have available at the time, as best we can, and to do those things that contribute to an environment in which the intermittency of discovery is recognized and valued.

Urk! no one would do that! Management is about just the opposite–regularity, volume business, process, efficiency. Management is about supervision and being the head not the durn hands, compliance and consistency, and making stuff so stupidly easy technicians in fear of the jobs can be made to do it, even if that means turning the activity itself into something so stupid that even management can understand it!

The idea of IP “management” simply cannot get away from the desire for regularity, process. Compound that with “policy”. The Idea of Order at Key West. It’s just that innovation, for all the focus and discipline and coordination that it takes to make anything new happen, doesn’t thrive in a bureaucratic fishbowl of regular, process-bound anything.

We are playing for the rare event. The insight or discovery or epiphany or development that matters is outside the mainstream. It is outside the volume reports, even of insights and discoveries and epiphanies and developments. It is not typical. It does not conform to what makes work tractable for administrators. It does not benefit from fairness or the assumption that everything should be treated the same until a triage sheet shows it to be different, and then to have a policy applied that in all pragmatism asks people to treat their insights as if they are typical. Fill out this form. Oh gosh, how far folks have fallen in their quest to “manage” research IP. Time to apply unmanagement ointment to the swollen portfolios in university offices. Doing so will create shining, intermittent opportunity. And bureaukleptic IP management darkness will comprehend it not.

A friend sent me a link to this article by Steve Blank that shows how venture backed start ups have moved from IPO to acquisition as the primary exit. If the primary purpose of starting a company is selling it to big companies, then one is in the business of making krill not new whales. Sure, there is money in making krill, too, but as Steve Blank points out, life after an acquisition is not likely to be the same as after an IPO. Management changes, the culture of the start up changes. The start up may simply be shut down to get it out of the way. That was the value of its purchase price. The value it had was the bother factor to the status quo.

We can take this state of affairs to the current university effort to start companies. A decade ago, starting companies out of a university was an iffy thing. A lot of universities couldn’t take an equity position, and when they did, they didn’t have a good way of handling the stock, the possible board seat, the conflict of interest with regard to research sponsored by the start up or the conflict of interest with regard to whether to enforce the contract to develop the licensed invention or to play for the success of the company and its equity value, regardless of the fate of the invention. Now, there’s a lot of talk about start ups, and talk of the need for gap funds to cross the “valley of death”, to go where private investors are generally unwilling to go (and that in itself should give folks pause, but it does not). Universities that claim a lot of start ups are the object of envy. And it truly is impressive to see a concerted effort to start companies using university resources, assets, and talent. Impressive like Cirque du soleil. It’s a great show.

But I can’t help feeling like the universities are getting it badly wrong. That they are offering incentives to start companies that never should be. And even if they still wanted to do that, they are a decade late on the mark (or maybe a decade early). At any rate, it’s terrible timing to be throwing everything but the kitchen sink at start ups. That these start ups serve as shells for research IP exclusively licensed with only tiny bits of reserved rights become moribund vessels that restrict opportunity and a broader commons of access to research findings. Where collaboration across organizations and fields is necessary to create a new platform for technology, placing each invention into a patent-holding start up is like encasing a chewy nougat in cement. That appears to be what has happened in nanotech. Research spread out everywhere, and every university rushed the bowl to file its patents, creating such fragmentation that few can afford to try to gather up sufficient rights to practice and not be trolled–and stopped by the sheer bureaucracy of licensing–by the rest. It’s like a mountain road controlled by tribal warlords. Every few miles, a new warlord, and a new demand for payment for safe passage. At some point, it is not worth being on the road in those mountains, because one cannot make it through without being picked clean. The warlords by their nature–they are warlords, after all–don’t get along with each other. So it’s not like one can pay off the first warlord, and get safe passage the rest of the way, with the first warlord sharing in a warm and fluffy way with all the other warlords along the route. No warlord is going to go for that!

So we have university warlords setting up their patent munitions to guard the road to innovation. That’s essentially what determining “commercial value” of an “early stage invention” arising from “fundamental research” is all about. Are we near the road? Can we block it? Will someone pay to go past? It’s like a neurotic passive aggressive patent assert program ironed into abstract rationality by bureaucrats. A organized process-bound transparent troll with warm fluffy public mission feelings that get easily hurt if corporate travelers won’t pay. Starting companies rather than offering licenses is a new development. Instead of blocking the road for others, the strategy is to block the road with one’s own travelers, at least for your part of the road. It’s like a denial of service attack in the internet world, but instead it’s a denial of research access attack. An exclusive license for research says to everyone else: you can know but cannot act. What’s the point of that? For the university warlords, it’s a leverage point–if you won’t license research IP to make commercial products for a royalty, then we will license our technology to our own start up companies, compete for venture funding, and make the products ourselves. Or, do the paperwork that makes it appear there will be products, but everything good anyway so long as we get paid. It’s a road that you can’t even get on unless you do business with, infringe the rights of, or buy out our start ups.

Where does this get to? Right now, it gets to feeding krill to the status quo for money. That’s where the current school of university start ups is headed. Sold to venture to sell to the bigs. The bigs, as likely as not, strip off the talent and toss the licensed technology. Well, not toss it exactly but rather keep the license and shift resources to other things. Since the university gets paid, it keeps mum–money=success in warlord shakedown thinking. This doesn’t happen all the time, but as the article we led off with indicates, it’s happening an awful lot of the time now.

The effect of the heavy emphasis on university start ups to “commercialize” research technology is to dry up available investment resources for entrepreneurs not taking licenses from universities (the universities rely on their reputation and access to “gap funds” to competitively “de-risk” possible investments). Then these start ups are fed to the status quo based on their bother, salvage, and uncertainty value. A service might be retained, or some data, but for the most part the intangible assets that make the start up generally valuable–brand, customer base, development team, management team, technology, vision, even relationship with university–are the things that get shed in an acquisition with a major player. It’s a fact of life. I’m not bothered particularly. But if this is the case, then the current university start ups are not an index into innovation but much more directly a measure of the cleverness universities are coming to to capture government money for their operations, and to make money feeding university inventions to the status quo. University tech transfer folks are entrenching the status quo through research innovation. Not what they say, or say they are doing, and not everywhere, but it’s the consequence of preponderance of their actions.

Companies going to IPO, private investment, or mergers among small companies to fuel growth and market reach, even sales relationships between a start up company and one or more large companies–these are measures of innovation vitality. Big companies swallowing small ones is a measure, more often than not, of an entrenching of the status quo, not a transformation of it. It’s an odd thought, isn’t it? That all the university talk is about how research fuels innovation which changes the world for the better and creates jobs. Yet the result of the actions taken by university bureaucrats amounts to deepening the hold the status quo has on the economy.

Now, I dearly love the status quo. It greets me when I get up and the sound of its distant train whistles and ferry boat horns lull me to sleep at night. In between, there is wonderful work done by big companies and small, governments and non-profits, and with all the natural people that work the formalities to animate these fictional corporate “people” that carry the status quo. Without the status quo, why, it would be chaos and night and monsters eating their children and then the next moment not even being the same monsters. Shudder. But the status quo also needs change agents. And one thing that status quo is absolutely no good at is proposing significant change and the executing on it. Jon Stewart’s bit on presidential calls for oil independence comes to mind. It’s the chamber of commerce dilemma. All the companies in the chamber are for progress, economic development, and prosperity, just that they would like it most if it came about without changing their positions at the table. It’s not like they are going to advocate for bringing in a new company so popular and successful it puts half of them out of business. No, they want change that is socially invariant.

Here’s the thing. The universities depend on the status quo and contribute to it. They are remarkably conservative things. They are like cockroaches. They survive in the harshest of environments. But their extramural research activity is expected to be progressive. This is the vision of Vannevar Bush in Science the Endless Frontier. Even more so in As We May Think. The applications of science have built us a “well-supplied house”. The challenge now is to avoid becoming “bogged down” by the complexity of things or “perish in conflict” by creating weapons rather than things for good. The role for universities, as a primary seat of science (and other good things), is to offer the status quo–no poke the status quo–with what the status quo cannot propose for itself, and that is transformational change, change that creates new social and economic opportunities outside the immediate reach of the powers that be.

Let’s repeat: the greatest economic thing the management of university research assets can do is to poke the status quo with the possibility of transformational change. Unexpected, unwanted, unplanned transformational change. Cures rather than making a business of turning acute conditions into paying chronic ones.

This poke is sometimes mistaken for confrontation, a mere antagonism, a lulz, by which university harbors activists who derive pleasure disrupting things others value. But we don’t need to conflate things this way. Research discovery opens up new things in the world. It turns out that the choice is not so much between weapons and fluffy comforts but rather between things that feed and entrench the status quo and things that would extend and transform it.

In this, it is not that the university creates companies, lots of companies, krill in a cold sea to keep whales happy–but what role these companies have with regard to the potential to change lives, to provide something new in the world that can transform the status quo. Like the telephone did, or the airplane. Instead, universities don’t report that kind of thing. It’s enough that they have started companies and they are making money doing so. But they appear, increasingly, to be selling to the edges of the status quo, so that things get more entrenched rather than more ready to change. The status quo, especially a status quo served by a collection of large companies, like an informal–even unplanned–monopoly, has its own ideas about when and how to change, so that change is properly timed, efficient, signaled to others, and taken on only after the return on past investments is adequately covered to justify moving on.

If universities are starting companies just to sell them out, for the most part, to the status quo, then what’s the point? Having no technology transfer function is just as fine to do that–maybe even better. The whale eats the krill, and the krill conveniently and inconsequentially disappear. The only issue is the cost to the whale. We don’t need any help leaving things for the status quo, and there is no point trying to make a buck doing that, especially in universities where for the vast majority they have *no meaningful plan for why they want to make money licensing patents*. I’ve read a ton of university IP policies. Not there. No meaningful reason to make money. At best, it’s public benefit=more money for us. It’s simply terrier behavior. They chew the sofa because they chew. The warlord demands protection money to travel the trail because that’s the life of the warlord. They try to make money because money is supposed to be made when licensing patents. Making money makes them look good. It doesn’t matter what the money comes from–it is as fine to sell out to the status quo as to challenge it. Doesn’t matter where it goes. Presently, it appears there is a good business in selling to the status quo, so that is what university technology transfer offices are doing. That’s what “commercialization” means.

The current tech transfer view of “commercialization” has nothing, really, to do with the innovation that universities, distinctively, should be about. This is why there’s academic freedom and tenure. This is why research is mostly extramural–outside the management control of bureaucrats and “managers”. This is why governments have a reason to seek out individual faculty to lead efforts outside of government labs and controls and direction. This is why there should be faculty investigator and faculty inventor choice in management agents. To keep transformation innovation from being sold off to the status quo simply for money. All this, to do what on behalf of the status quo what it cannot do for itself. Propose and execute on transformations led by technology that are not socially or corporately invariant.

Inside Higher Ed is running a story that faculty members have been approached to serve as consultants to BP with regard to the oil spill. See here.

It’s pretty typical in consulting agreements that data rights, publication, and IP can be locked up for the consultant’s employer. If funding went through sponsored research, however, then a sponsor might be allowed pre-publication review (but not approval) and data would be made publicly available.

“You’re working for a side with a financial interest [either way],” he says. “The federal government is trying to maximize the damage assessment for obvious reasons, and the oil companies are trying to minimize it.”

D’Elia says his preference would be for the federal government to provide a pool of money to scientists for the purposes of studying the spill’s impact. Absent that, research becomes part of a legal process — not necessarily a scientific one, D’Elia says.

The issue at the core is whether scientific practice can withstand the mercenary and personal pressures to be on one side or another of such a bi-polar issue. Add in a university or company claim to IP, and it’s pretty clear there is no neutral ground available–at least not without a lot of candor, respect, and discipline. It’s like being a referee in a World Cup match. If the referee takes sides, then it’s all up. And if there are no referees, what keeps it civil?

Regardless of who is paying–the feds or BP–it comes down to a huge potential for buying off the science. That is, taking money from either sponsor with a legal or political agenda makes things mercenary. One has to be really vigilant. It’s all too easy to game the data, the experimental set up, the interpretation of results.

More concern is that once the buying spree is done and there are government-sponsored faculty scientists and oil company-consulting faculty scientists, then what faculty researchers are left with a voice in the matter? Where are the scientists focused on science, not just on amassing and quantifying “data”? Science doesn’t exist in a vacuum, but also science is only as good as the candor with which it is done. That is, one can go through the motions of actions we would attribute to scientists–making observations, proposing hypotheses, collecting data, running analyses, using technical language, and relying on academic credentials–and it can still not be science.

In Against Method, a work that isn’t intended to be even handed about things, Paul Feyerabend argues that the separation of science and state is more vital than the separation of church and state. Yet in the United States at least government funding has come to dominate many areas of science, technology, and this unnamed activity that has many of the motions of science, but isn’t. Sort of a golem science. It may be such work serves a purpose. Clearly it serves a master. It may even seek to stand up to that master–and that may be a good thing, or not. But if we can’t tell–if mercenary science is just as good as any other–then what we need to ensure is that mercenary science isn’t the only science. And perhaps in key areas of policy, it should not even be the dominant form of science.

In The Survival Game, David Barash discusses the prisoner’s dilemma as a instance of where the payoff for defecting on collaborators is better than playing nice. When such situations repeat, there are huge problems for collaborators in responding to attempts to defect. In simulations, defecting back every other time is the best available response. Not defecting back means the defector confirms the payoff, and defecting back every time means baiting and predictability.

One observation that comes to mind: it may be future defectors that most want everyone else collaborating. That is, the defector’s payoff is not something folks just stumble into–it can be constructed that way. One might think, therefore, that all the present calls for collaboration–and virtually no calls for research competition (except in “grand challenges setting”)–might set up defection payoffs.

More broadly, I’m thinking about intermittency. Might intermittent responses disrupt the defector’s incentives and expectations? Might intermittent efforts at innovation be better than systematic ones, consensus ones? Might the policy framework for innovation actually aim to set up a condition where the defector is the innovator, is the cheat on cooperation? If so, then for university IP policy, enforcing things to prevent local initiative simply because it violates the policy would utterly miss the point of setting up innovation conditions. Yes, the crass sloppy mediocre stuff–stomp it out. But the insightful, clever, opportunistic stuff–well, maybe it is from defectors on the status quo of university administration that research innovation is going to arise. That is, the technology transfer office with its insistence on process and policy and cooperation sets up the defectors’ payoff. Just that the compliance and ethics officers don’t understand this.

If this makes any sense, then one might also begin to see why policy mandating a process or system or office or “principles” (unbending requirements lofted to categorical imperatives: you will serve the policy as if it were sacred text) doesn’t make much sense for innovation relative to the status quo. If it’s the intermittency of attack that matters, then whatever the baseline processes are, the practice has to be opportunistic enough to allow local, unplanned, sporadic shots at the status quo. It’s really hard to write bureaucratic policy that condones such stuff. Consistency, regularity, and the like is so much easier to imagine than intermittent mobilization for a purpose outside the anticipated bounds of committee approved processes.

A little housekeeping. Spent the weekend trying to fix the RSS feed. Appears to be a Firefox problem. Things are working for the comments feed, and in IE, I think. Sorry about the glitches.

Patent law is all social convention anyway. It is something we make up. We then task courts with enforcing our made up stuff as laws. Practices and habits build up. People and people in companies get used to the habits, including cheating, being nasty, trying to do the right thing, getting confused, holding out, exploiting loopholes, and trying to change things so they are better suited to one’s advantage, or are more orderly, or sound better. Okay. Complicated. The point is, there’s no natural history of patents. It’s not something that has to be any one way. It’s not that an inventor has to own his or her invention. No one has to own an invention. An invention doesn’t have to be defined as anything to own. So it’s entirely possible to set things up so universities own what faculty researchers come up with. And it’s entirely possible that universities can set things up so that when they own something, bureaucrats rather than researchers decide what to do with those inventions. This all is possible. I say, it’s bad policy for research and innovation to do so. It’s bad policy to be making. It may look good and orderly: “I know, we will have administrators manage all the faculty inventions for them”. “Specialists in IP and licensing will work diligently to commercialize all inventions they take under management.” If one likes order on paper, then one will love this sort of thing. But the history of stuff doesn’t support the idea that host university management does a better job with research innovation than anyone else. And where there is management at all involved, I say it’s better to let the researchers choose what management they want to work with.

That is, it’s good university policy to say, if you want to do commercial things with your invention, take leave and go do that, otherwise, choose management to do that for you and stick to teaching and research. It’s bad policy to say, that management will always be university administrators (or affiliated foundation administrators). It’s even worse policy to say, don’t bother, we already own everything you might invent. Just tell us what you have done and we will take it from there. And, oh, don’t bother showing anyone else what you know until we license the patent rights, because they will just be infringers. It is possible to construct a law providing for invent-for-hire. One could even insert that in Bayh-Dole. It is not there now. Even if it is possible, it’s really bad for university administrators to be pushing for it. Patent Lupus. The mad attack of management on the creative class. The corporatization of university research. Rather than making university research excel, become stronger, be a force, the administrative urge is to make it orderly, bland, and profitable. We already have that with industry research. University research has other roles to play in an innovation economy. The last thing we need is an inventor-loathing monoculture of inventive practice run by bureaucrats. Ugh.

I belabor this because I had an attorney visiting Seattle schedule time for coffee so I could be reamed out on exactly this point. Yes, Bayh-Dole is OMG automagical. WTF?! No, the law doesn’t have to say anything about it elsewhere. No, the 30 years of doing it via assignments was extra. But now we see that the law really did create automagical vesting of inventions with bureaucrats, simply by providing notice of election to retain title to the feds. When I started to present counter arguments, the attorney huffed off. Mercenary instance. There are real folks who *want the bureaucrats to run university innovation practice.*

Now, I’ve been doing this work for 20 years. I have been in these sorts of roles. I like to think I’ve been a valuable adjunct to people getting their work into circulation. I am also confident that if the university had no IP requirements, people would still come to me with their opportunities. I’m not making an argument against management, or even university management, of IP. I’m making an argument that compelling creative folks to use only their employer’s management, and when that compulsion gets harder to do, and in response trying to make title automagical and trying to cut off opportunities to consult and collaborate to get there–that’s bad doodoo. That’s bad innovation policy. In terms of Teece, that’s the triumph of infrastructure over innovators. It’s inventor loathing. It’s a will to monoculture. It’s the work of a bozonet–a majority consensus dedicated to making their work easier for themselves without getting into what makes their work worth doing or how to do it really well.

It is as if Bayh-Dole creates a force-field that prevents title from attaching to anything until the university chooses to “retain” it. The law does not say this, of course. It does not point out that by electing to retain title, something else substantive happens at that moment–that title is transferred from individuals–their private property–to the university. It does not say, title does not attach to subject inventions until the university decides what to do. Even though this would be a remarkable action in patent law, though not without precedent, it need not be noted. Not being noted–now, that has no precedent. No, it is not obvious that there is a magical force-field.

I suspect there also may be something else going on.

Under the Fifth Amendment to the Constitution, the government cannot take private property without due process and just compensation. The contract between the government and the university forms that process and provides for compensation. For the Federal agency, the compensation is in the form of the grant funding. For the university–especially if it is a public university–the law provides for sharing with the inventors. That is, if there were no sharing, and the law provided for the state to acquire the private property of the inventors–their patent rights–without compensation, then there would be a huge problem under the Fifth Amendment. There would not be “just compensation”. For private universities, there would be a similar, but less direct, argument: the Federal government was taking patent rights (under contract) and then agreeing to the transfer of these valuable rights to a private entity (not the inventors). Essentially, a taking without compensation. It would be even worse if Bayh-Dole were automagical, but we have dealt with that elsewhere.

For universities, where extramural research is not assigned nor controlled as to the performance of work and reporting of results, it is simply not possible to construct, within the current policy structure at most universities, anything like invent-for-hire, construing the university as inventor, or accepting that the inventor is an individual but that the property created, and the patent rights that attach to it, are stripped off by the force-field of federal funding. To get there, then the royalty sharing with inventors, which is often construed as a fluffy, warm-hearted gesture of niceness in university policy, really is not. It is a duty under federal contract, implemented to make that contract constitutionally sound. Bayh-Dole fails if the compensation offered inventors is not “just”. That means, royalty schedules set in university concrete policy may not be “just”. Just because they are in policy, or uniform, or look good to whomever had the committee assignment to figure something out doesn’t mean they are “just”. This is entirely different from an employment relationship involving a private company–or private university–and an employee-inventor.

If conducting research is voluntary, and if the public university in its policies claims ownership of inventions, then to survive a taking of private property for public use, there has to be just compensation, and it sure as heck isn’t one’s salary, which is established *without regard* to inventions made in extramural research, and it’s hard to see that access to research facilities is on its own “just” compensation for any inventions so made.

Let’s recap, top down.

The US Constitution provides for the Federal government to provide for patents and copyrights. The stated purpose is to promote progress of useful arts and sciences by allowing inventors and authors exclusive rights to their inventions and writings. The states invest this right in the Federal government on behalf of inventors and authors, on behalf of promoting progress. It is a personal right. The properties that result–patents and copyrights–are personal until circumstances change that outcome.

The Fifth Amendment says, the government cannot take private property for public use without due process and just compensation.

Bayh-Dole comes along and says, use the patent system to promote progress via practical application, better relationships, jobs all around, less administration, competition (but not so intense it screws up research and discovery), and yeah, some commercialization too. Bayh-Dole says, if the university takes the rights, it has to compensate inventors by sharing royalties. It says, if the university doesn’t take the rights, then the government can take those rights for itself, like the old days, where the compensation is the provision of funding. The university *doesn’t provide the funding*. So it can’t make the same argument within Bayh-Dole.

Now we have the bureaukleptic lobby rushing the food bowl barking: take the property outright, automagically OMG!

You see how it doesn’t work. The offer of a royalty share in a bureaukleptic policy is not “just compensation” for the taking unless *something gets licensed for a royalty*. It’s not sufficient to say, since we didn’t license it, or didn’t license it for more than our costs, that it *wasn’t worth anything*. No, the presumption is, it was worth something–that’s why it was taken–and now it isn’t worth anything, and therefore there is harm, without compensation.

Bayh-Dole says: if universities are going to stand in and do the taking, then they have to have a deal with the inventors for “just compensation.” That deal happens at assignment. Without assignment (as in force-field automagic), there’s no deal. At that point it’s entirely open if there is just compensation. At that point, the constitutionality of Bayh-Dole is undermined.

Just a thought.

Bayh-Dole 1) normalizes government agency approaches to claims on inventions made in government funded research; 2) places research institutions in a voluntary position to direct the disposition of claims on invention ahead of government agencies, provided the research institutions use the patent system to carry forward stated government objectives.

Bayh-Dole lists the US government’s objectives, including university-industry collaboration (not reported in any AUTM survey to my knowledge), support for small companies (treated as start ups, for the most part, in AUTM reporting), support for American manufacturing, practical application (use broadly, not simply creating products to be sold in lucrative markets paying a royalty), lowering administrative costs, protecting research from restrictions, and protecting the public from non-use and unreasonable use. Commercialization is in there, but not the star attraction. It’s practical application that gets the definition, and everyone one might look for commercialization words, there are also practical application and use words. Products and sales are part of it, but not the whole picture.

The big challenge is that Bayh-Dole provides no funding for university-led patent interventions. If it takes typically 10 to 20 years to build a sufficient patent portfolio, expertise, and relationships to receive significant income from efforts toward practical application, then we’re looking at an initial investment by the research institution of between $5m and $20m over 10 to 20 years, and more like the upper end of that, before one is in a position to be breakeven on an annual basis. Recovering that initial investment takes significant relationship–an important technology with one or more industry partners happy to support the research institution via royalty or research payments. If the institution is not worthy of support, if the partners are not happy, or if the technology is not important–well, then, darn.

This is one reason why there are efforts to argue other ways of recognizing research institution investment in technology transfer. That the initial investment is recoverable through other forms of income, not just licensing. So, donations or industry sponsored research or favorable placement to compete for federal grants, or state support. I worked in a program where the metric was all income from industry sources, whether gift, grant, license, or equity. There was no mandate to maximize license at the expense of the others, and the vision was long term–10 years–to the chagrin of some other administrators, who looked at cash each year as their measure of success–making sure they spent all of it, trying to find more of it.

If one wanted to vary Bayh-Dole, one could ask whether it makes sense to treat all government research the same with regard to use of the patent system. Even Bayh-Dole has break outs for naval nuclear propulsion, DOE weapons related programs, and exceptional circumstances.

One could also ask whether appropriate incentives and protections are in place to keep research institutions–especially administrators–focused on the broad objectives of Bayh-Dole, and not a narrow one, such as exclusive licensing for product development, even while recognizing that in some circumstances, this is indeed the expected practice. For that, one might audit each of the stated objectives of Bayh-Dole against actual use of the patent system by research institutions, not just looking at “success stories” but at all the stories and non-stories, the fights with inventors, the bitternesses and sadnesses with industry partners, the failed or moribund startups, the research teams that fought, the research teams that were indifferent, the research teams that were secretive, the research teams that were so focused on the next grant that they could not afford to be distracted to deal with the outcomes of the current one.

Unfortunately, Bayh-Dole also puts utilization reports under FOIA protections. I don’t know of any university that releases its utilization reports, even in redacted form. So we have more darkness on the matter. I find it odd that universities roundly resist efforts by companies to control publication, on the concern that the company would allow only positive results to be published and not adverse ones, to the detriment of the public, but then do just this thing in their reporting of their IP practices–never an anti-portfolio, never a plain account of all the efforts that are presently inconsequential and unremarkable, nothing much about disputes. No wonder the academics studying tech transfer practice have such a difficult time getting into practices and so resort to reasoning from assumptions and inferences and partial statistics.