The use of the patent system for federal research results, 13: The failed middle ground

We’ve looked at the early Federal Security Administration policy on inventions made in federally contracted work–FSA order 110-1, issued in 1952. The government’s policy as set forth in David Lloyd Kreeger’s report for the Attorney General in 1947 was that inventions owned by the federal government should be made available condition-free and payment-free to all citizens, and that where “development” of an invention was needed to meet a public purpose, the government should fund that work. As for federal ownership of inventions made by federal employees, the Supreme Court in Dubilier in 1933 said that this was a matter for Congress. President Truman, however, used an executive order, 10096, in 1950, to establish executive branch patent policy for federal employees.

But where was the policy mandate for the government to acquire inventions made in federally contracted work? Should the government take title to such inventions and make them open access, following the Attorney General’s recommendation? Or is just a license enough? In Science the Endless Frontier (1945) Vannevar Bush devoted a chapter to patent policy for his proposed National Research Foundation, arguing that a government license was sufficient for inventions arising in federally supported work intended to advance the frontiers of science. But what about other work? What about work that asks a company to modify product it already has for a government-specified requirement? Why should the government obtain any inventions made in that work as deliverables, forcing open a line of development that the company might otherwise have done anyway and had rights to? Why would a company do anything for the federal government in such circumstances?

Thus, the issue of “participation.” Senator Kilgore, in a sense Bush’s political competitor, wanted to spread federal government research funding around and so avoid the “concentration of economic power” that federal funding might otherwise provide to companies favored with federal contracts. Vannevar Bush thought that if you wanted to get something done, done right, and done soon, then go with the best and the brightest and skip wasting resources on the rest. Administrators who wanted things both ways then worried about contracting policies that might exclude the best and the brightest because they declined to participate if participation meant giving up their prospects of control over lines of technology development. The administrative agenda then became how to condition innovation research so that it could be conveniently and efficiently administrated. And within this agenda, it then became important to make administrated research produce outcomes that would warrant continued if not expanded funding, leading to continued and expanded administration, progressively improving its efficiency and effectiveness.

There were two readily available perspectives from which to build an administrative framework for invention rights. The federal government could declare some area to be of public importance, and then demand ownership of inventions made in work receiving federal funding in that area. In effect, the claim relied on the idea that the work was public and companies that chose to participate were patriotically contributing their effort to that public work, and were being compensated for doing so. Think of it as a war effort. That’s how Vannevar Bush characterized the move from military development to medical development: a public “war” on disease, with companies recruited to help in their areas of specialty, enabled by lots of new science generated by the NRF and lots of recent science released to the public from military classification.

There’s another way to look at it, however: companies were already working in their own way (with profit motives to guide them) on matters of public health, but treating their activity as a private enterprise, not a public one. The federal government, if it wanted things to move more quickly, in more areas, could subsidize company activity–contract for services, purchase at a premium, or invest like a private investor. By doing so, the government could make companies working public health more profitable, attracting even more investors, and displacing other efforts to address public health (such as education and prevention–where there was not only less profit but worse an active effort to reduce the size of the market and the financial importance of finding remedies). In this way of thinking, companies got subsidies to offset the risks taken by other investors. Companies did not step into public works, dedicating their future discoveries to the public and to their competitors–especially to free-riding, price-cutting “generic” producers of medicines–simply for accepting federal money to do specified research work that they would do anyway, when they got around to it.

Do companies receiving federal money join public projects? Or does federal money augment private initiatives so that there are more of them, done faster and better than otherwise? There’s your dilemma–one that a federal administrator is not likely to have the wherewithall to resolve. All or none doesn’t feel right, and anything in between requires exercise of personal judgment unless a determination can be made into a process that produces the proper results as if there were no need for judgment, just a gathering of facts and a formula for deciding whether companies become deputy federal agencies or federal agencies turned into private investors and customers.

FSA 110-1 tries to navigate this dilemma by proposing a middle ground with a rough administrative (read: fantasy) set of procedures for deciding whether a contractor (company or nonprofit) ought to deliver invention rights along with other contract deliverables or whether the federal government could get along with a non-exclusive license for government purposes. (The point of the government license was two-fold. First, it freed the government from financial claims made by patent holders on inventions made in federally supported work for practice of claimed inventions. Second, it allowed the government to choose any contractor to practice the invention in providing goods and services to the government, so federal agencies were not stuck with sole source contracts with patent holders for the duration of their patent rights. Ownership of the invention gets to the same place, but also allows the government to grant access to anyone, not just those performing government-authorized work.)

Here’s a bit of diagram to show the flow of 110:

First, if the form of federal support is a grant-in-aid, then the agency unit head may decide, or may delegate the decision under what would be called “institutional patent agreements” under which nonprofits would make the invention available on reasonable terms (without unreasonable restrictions or excessive royalties). If not a grant-in-aid, then the agency unit head may allow a contractor to retain rights if in doing so the invention will be released open access or the unit head may allow a contractor to deal in patent exclusivity. Otherwise, the invention is assigned to the government, unless the unit head decides the invention is of doubtful importance or the government’s “equity” in it is minor, in which cases the government makes no claim and places no conditions on contractor ownership of the invention. While the policy starts with a default of open access, in the logic of things, open access is only one outcome among a number, subject to choices made the unit heads. There’s really nothing that constrains a unit head one way or another. Default open access is actually deep in the logic cascade of policy 110. In the way of getting there are determinations–implying procedures, fact-gathering, plans, balancing of interests, anticipating public benefit, consideration of anticipated speed of development and widest use.

In all of this, however, runs the assumption that every invention allowed to be claimed by a contractor will be used and developed by that contractor or licensed to one or more companies that will use and develop the invention. It’s as if contractor ownership and licensing assures use and development and therefore licensing and public benefit can be equated. Licensing then becomes not only a proxy for public benefit (when obviously it can’t be) but also a metric of “success” of the policy. The great problem with this way of thinking is that it misses entirely the actual outcomes of practice under the policy. People may trade on patents for years and not produce any product for public benefit based on actual use of a claimed invention. The profit motive for deriving “value” from a patent is not necessarily aligned with the profit motive of making a profitable product and selling it widely without direct competition. In this policy, accountability falls to institutional contractors–universities, mostly–and to federal agencies. Any policy, in practice, then depends on a readiness of institutions to enforce contracting terms and licensing terms. Without enforcement, the policy is a gesture. But if the policy leaves the unit head responsible for enforcing the choices the unit head has made, there’s a conflict of interest–to enforce requirements is to admit that the choice was poor. That’s tough to do in an administrative bureaucracy.

FSA 110 does not worry that a claimed invention–held by a contractor or a federal agency (inventors are ignored)–will not be productized. Nothing about fumbling, bumbling, wasteful, indifferent, grasping, complacent, ignorant, foolish, bureaucratic, risk-averse contractors gaining patent rights and doing nothing (other than maybe making a show of trying) to provide the public with something better than it had before, on reasonable terms (more muddle).

Thus, the FSA policy declares default open access but allows agency unit heads to opt out of open access. Does this sound reasonable? Sure. Is it workable? No, not. It is practice fantasy cast as rationalized process.

The FSA middle view of patents is that there may be some benefit for universities and their nonprofit patent management firms to hold exclusive rights in inventions made in federally funded public health research just long enough to distribute those rights to qualified companies. That’s the premise of the first generation Public Health Service IPA program. The implicit idea is that this distribution of rights from a private patent holder would more effectively transfer federally supported inventions than would condition-free federal open access for those same inventions. The right to exclude in such a case would be temporary–only long enough to establish patent rights and engage companies willing to develop and distribute commercial versions of the inventions under management. Those companies that show early interest, then, would have an advantage over other companies that declined to get involved. In this middle ground, patent exclusion is expected to be limited to the wanna-be companies that arrive late to the party, do shoddy work, and don’t contribute the development effort. Only in rare cases would it be necessary for a single company to receive an exclusive license because it was necessary (not that a company demanded it) that a single company fund the entirety of commercial development. Fine. But it’s a muddle to think this way. Try it.

The FSA policy then covers a final set of edge cases–if the federal government’s “equity” in a given invention is “so minor,” then the agency can release any claim to such invention. So the thinking goes, where the federal government contributes research funding, but there’s other funding also involved, then it would be inequitable for the government to claim an undivided interest in a given invention. This sounds reasonable if you don’t think about it. If you do think about it, there’s nothing in the abstract that ties the act of inventing to the spending of money, especially in research. And there’s nothing that ties the amount of funding to the importance of the invention for public, governmental purposes. The spending of money often comes along later, once there is an invention, to build and test prototypes, to ramp up production facilities, and the like. An invention can be made with the realization that all the money spent on research to that point has been wasted–that there’s another way to address the public’s health problem. Or the invention, the discovery, has nothing to do with the direction and purpose of the research–it’s entirely a side thing, not specified, not paid for, not studies, not a deliverable, nada–but perhaps that side thing is something central to some other federal agency public purpose need. You see–muddle. The money spent by either the government or whatever other sources of funding might have nothing to do with such an invention. Or everything. The idea sounds reasonable but it just doesn’t work in practice.

And even more so, there’s the weirdness that no-one bothers with–that inventors need not be paid anything by a company or university or federal agency for inventing what they invent. They could be thinking on their own, not directed to think such thoughts, and not paid by anyone specifically to think. The government funding that might support their research then might appear to dominate the funding, but not recognize at all the time (without compensation) an inventor might spend thinking and working out ideas that prove to be inventive. The idea is strangely bureaucratic that money spent by organizations is an equitable way for administrators to divide up claims to inventions. The idea is even further out of place when the government’s purpose is to support “basic” research to advance the frontiers of science rather than something “mission-directed” to address a pressing public health need.

The FSA 110 policy then amounts to “open access, but not if a unit head decides otherwise, but by procedure.” The grammar suggests open access all the time and a unit head does not have to bother with anything else. But the logic does the opposite–everything is in play, mediated by procedures, and a unit head could leave rights with contractors (inventors don’t figure) or determine that most inventions are either of doubtful importance or have received less than minor funding from other sources.

It’s procedure that makes a muddle of things. It’s the procedure that Bayh-Dole advocates then, later, attacked–not to improve such procedures, but to swap them out for a different set of unworkable procedures, so that private patent exclusion become the new default, and unit heads have to follow an impossible process in advance of contracting if they feel the need to change that default. One muddle swapped for another, none of it responsive to the various ways in which the federal government might go about research–doing that research itself or contracting with others to do it, and among the others contracting with nonprofits, or with universities to gain access to and support faculty and students, or with for-profit contract research companies, doing things professionally and perhaps with greater direction and speed, or with companies operating in non-governmental markets (not military, space, nuclear, perhaps public health) which have their own products and futures to take care of.

Policy administrators dream in abstractions that make a nice, neat divisions of what may be chaotic, opportunistic, and changing set of events. They seek to impose order on what appears as disorder, to train others to behave so as to make their imposed order “work” or at least appear to “work.”

At the University of California, I once thought it would be good to write up some of the procedures we had used to make decisions, so that if others wanted us to do the same for them, there was a record of how we had gone about things. It wasn’t even policy–just a record of what we had done within the constraints of policy. But I was not allowed to release this write up until I had reviewed it against every possible University of California policy at the campus level and the system level for possible implications. It wasn’t worth the time. I’d be at it for months, and everywhere there might be an implication, I would have to negotiate with the owner of that policy–with little prospect for success. Policy consistency destroys practice consistency.

In another situation at the University of California, I was working with a company that wanted access to some IP we had. I asked if they had worked with UC before and they said, yes, we did a deal for a similar bit of IP a few years ago with another UC campus. Did you like that arrangement?, I asked. Yes, it was fine. How about then I get that document from the other campus and add this IP in addition? Company was elated. But when I go to get the new deal confirmed, I’m block by UC system officials. Can’t approve your deal. Why not? The first deal involved an exception to policy and if we did the deal again the same way it would imply that the first deal was a precedent not an exception. If we let you do the same deal again, it would implicitly change the policy and we can’t have that. So I could get a different exception to policy for this deal, just not the same exception that worked so well for the company last time. Yes, you got it.

Without policy, institutional actions are left to personal judgment (a horrifying idea, at least in public, for administrators) and to the random walks and self-organizing behaviors (more horrifying ideas, even in private, for administrators) of administrators, inventors, and those that become aware of their inventions. Somehow, personal judgment of people enforcing policy compliance (or granting waivers) is better than the personal judgment of people working under general principles to carry out their bit of the institutional mission. The University of California goes so far as to declare, by policy, that it is unethical for its personnel to do something that would be better than what University policy provides. No “higher purpose” is allowed than what is specified by policy, such as doing new business with a company on the same terms as prior business.

FSA 110 is an early attempt to bring research activities and outcomes within an administrative rationalization, where the virtues are that the policy is “uniform” and there are procedures leading to determinations to allocate institutionally held patent rights, but with limits placed on those patent rights. Inventors drop out of the equation, like hens when one is looking at a recipe for souffle. Vannevar Bush’s recommendation (government needs only a license) get ignored. The Attorney General’s recommendation (government should own but with open access and should pay for any needed development) get ignored. Even Truman’s Executive Order gets ignored as a template for what the government should expect of contractors in their dealing with their employee inventors working on federally funded work.

The FSA 110 middle ground sounds nice if one is administrative about it. Otherwise, it’s a muddle that could not possibly work, other than human beings are adaptable to exploit it or work around it. Bush’s proposal would work. The Attorney General’s recommendations would work. But there’s no way to combine the two into a single “uniform” policy in the abstract with a middle ground of procedures and determinations, and there’s no way to properly apply administrative trappings to innovation research without going abstract and dropping out many of the chaotic, opportunistic, and changing things that make for advancing the use and development of inventive things made in publicly funded work in areas such a public health.

In 1963, FSA 110’s general strategy was included in the Kennedy patent policy with a specific call out for inventions made in public health directed research to be owned by the federal government and made available open access (unless a unit head granted an exception). The idea was to rationalize federal agency practices while preserving flexibility. In 1968 Harbridge House, commissioned to do a study of government patent policy and practice, endorsed policy flexibility in its final report. Also in 1968, the NIH restarted its IPA program, citing the Harbridge House report, preserving the same gesture toward open access but waiving the requirement if a nonprofit tried open access and it didn’t work out or decided it wouldn’t work out in advance and didn’t bother to try. In 1971, Nixon’s revision of Kennedy had introduced the idea that federal agencies could take rights and then license exclusively rather than use open access. The NSF added its own IPA program in 1974. By the time of Bayh-Dole, the FSA 110 idea that agency heads could decide between open access and patent exclusivity, between contractor ownership with conditions and federal ownership, was in full rout. In that, the Bayh-Dole advocates were spot on. The FSA middle ground was unworkable.  But so would be Bayh-Dole.

This entry was posted in Bayh-Dole, History, Policy, Vannever Bush and tagged , , , . Bookmark the permalink.