The dogs in the manger, 1

In 1979, R. Tenney Johnson, a career federal attorney, testified before the Senate Subcommittee on Science, Technology, and Space on Senate bill S. 1215, the “Science and Technology Research Development Utilization Policy Act.” This was the bill that was competing with Bayh-Dole to become the primary government-wide regulation regarding federally supported inventions. Bayh-Dole had failed to pass, and now it was Sen. Harrison Schmidt’s turn to try to push through an even more comprehensive bill. The bill would fail, despite having numerous co-sponsors. Bayh-Dole would then succeed by being tagged onto a lame-duck House bill making technical changes in patent law procedures–largely as Sen. Long’s parting gift to Sen. Bayh, who had failed to be re-elected.

Johnson’s testimony is instructive because Johnson helped to draft the Kennedy patent policy. His perspective then helps us understand the attitudes and reasoning behind the bills put forward by Schmidt and by Bayh and Dole. Johnson provides some history for the debates about what to do with patent rights for inventions made under federal contracts.

The great debate at the time was between the “title” argument–the government has paid for the work, it should get title to inventions, and the inventions should be made available to the public–and the “license” argument–the government should have a free license to use what it has paid for but otherwise contractors should be able to patent inventions and do what they want with those inventions, short of being a “dog in the manger” and not doing anything with them. Neither of this arguments is particularly apt and neither has much feck–especially if the idea is that all government contracting must be controlled by one or the other of these arguments.

Thus, the Kennedy patent policy aimed for “uniform flexibility”:

The Kennedy patent policy identified various situations in federal contracting and stipulated uniform approaches to dealing with them. In some situations, contractors should own inventions; in other situations, the federal government should own inventions; and in other situations, federal agencies should decide whether the government should own up front, before the contract is awarded; and otherwise, federal agencies should decide after an invention has been made and a contractor has requested to retain ownership of the invention.

What is continually fascinating about the arguments over inventions made in government contracts is that the legal and administrative representatives ignore entirely inventors, faculty investigators, and the nature of subvention funding–grants in aid–as distinct from procurement contracts. While the patent system is acknowledged to benefit inventors, the government officials and pundits who debated title vs license had nothing to offer inventors. They assumed that the federal government had a right to require inventors supported by federal contracts to assign their inventions to the federal government, and it was only if a contractor obtained ownership ahead of the federal government that there was a question about what the government should do about the contractor’s ownership position.

Here’s where Sean O’Connor might jump in and argue that the officials and pundits assumed all contractors had invention assignment agreements with their employees. But that cannot possibly be correct. First, the basis for the federal government to require assignment from inventors is that inventors were made parties to each contract, along with the institutional or company contractor. The federal government could assert its rights directly through federal contracting regulations, rather than relying on an institutional contractor or company to have a satisfactory private agreement with each employee regarding inventions. Further, anyone could see that universities for the most part did not have such private agreements with their faculty–not in their policies, not in employment agreements. Even up through the 1980s, Stanford, for instance, had a policy not to take ownership of faculty inventions unless legally required to do so. And university officials were present in these debates in Congress–no one had any need to assume universities had corporate-style invention assignment agreements with their faculty, or even that faculty were servants of university administrative masters when it came to federally supported research.

We might say that in federal contracting, what we might call the “regulatory” title to any inventions was, by operation of the regulations that defined federal contracts, vested in the federal government. The “equitable” title to any inventions was, by operation of an employer assigning employees to do work under a federal contract, vested in the employer. And the “formal” title, by operation of federal patent law, vested in the inventors. The fuss then was whether regulatory title should take precedence over the equitable title. The assumption was that one or the other of regulatory title and equitable title would eventually be invoked and the inventors would have to assign formal title.

It’s just that for subvention research at universities, faculty weren’t generally assigned to conduct the research–they proposed it, the government funded it, and universities were asked to handle the money and provide facilities, for which they were also compensated. Thus, universities never had or expected equitable title in such research. That leaves it up to the federal government’s “regulatory” title.

Under the NIH’s IPA scheme from 1968 to 1978, the NIH master agreement required participating universities to obtain from faculty an agreement to assign to the university any inventions the university administration decided to attempt to patent. There was no equitable title involved–universities simply signed the NIH’s master IPA agreement and turned around and told faculty that their federal contract now required them to require faculty to promise to assign inventions, regardless of what the university’s patent policy might say. It’s worth pointing out that this scheme was hatched with the patent counsel at the Wisconsin Alumni Research Foundation–and the University of Wisconsin at the time had no patent policy and made no claim on faculty inventions. There was no equitable claim based on employment or even use of resources. One can see then how WARF might really want Wisconsin to sign up for the IPA program–so there would be a pipeline of inventions obligated to Wisconsin (who then would direct faculty to assign rights to WARF).

Again, be clear: the IPA program required universities to require faculty inventors to assign inventions when the university decided to seek a patent on a given invention made with federal support.

What had to happen? First, the university had to know enough about the invention that it could decide whether to file a patent application. Second, the university had to decide to file a patent application. Third, the university had to obtain assignment of the invention–something that faculty inventors had promised to do as a condition of access to the federal funding under the IPA master agreement. The promise to assign was conditional upon the university agreeing to file a patent application. Hold that thought.

If a university did not choose to file a patent application, then inventors had no obligation to assign to the university. Invention rights then were controlled by the regulatory requirements on federal contracts otherwise in place at the federal agency providing the funds–generally, the federal agency would claim its regulatory title, obtain assignment, and dedicate the invention to the public, sometimes filing a patent application but under no obligation to do so. Most patents were licensed non-exclusively, if the agency even cared about matters of formal licensing.

Compare, then, with Bayh-Dole. Bayh-Dole applies in its contracting part only to subject inventions–inventions of the contractor, inventions already owned by a party to the funding agreement. Thus, under Bayh-Dole, first the contractor has to obtain ownership of a given invention, then has to do various things to keep that ownership–such as filing a patent application. But where the IPA program required a university to make the decision to file a patent application first, and then obtain assignment, Bayh-Dole reverses things–first a university has to acquire ownership and then to keep that ownership (after up to three years of delay) must file a patent application. If a university does not acquire a given invention, then that invention is not a subject invention, does not have to be disclosed under Bayh-Dole, and the invention is subject to whatever regulatory requirements run with inventions made with federal support that aren’t subject inventions.

Actually, it’s one step more clever than that. When the feds came to create the standard patent rights clause authorized by Bayh-Dole, they threw in a few requirements not in Bayh-Dole. Under one of these non-statutory requirements, universities (and other contractors) are required to subcontract to make employees individually parties to the funding agreement. Thus, when they invent, they are contractors, they own their inventions, and their inventions then are also subject inventions, even if not assigned to the university. But then Bayh-Dole provides a separate patent rights clause, treating inventors as if they are small business contractors, but with fewer obligations than either small businesses or nonprofits. Look, it’s a scheme designed to snooker folks. You aren’t expected to follow the logic. That’s the beauty of the scheme.

In any event, the IPA program required assignment of inventions, but only after a university had decided to file a patent application. Bayh-Dole doesn’t require assignment of inventions and only operates after a university has obtained ownership, and then gives a university up to three years to get around to filing a patent application. Pretty substantial differences abound, if you care about this sort of thing.

For all that–you can see how Johnson’s testimony illustrates the administrator mindset with regard to inventions made in federally supported research. Things are dichotomies–title or license; contractor owns or the government owns–and all that the Kennedy patent policy did was rationalize when the dichotomy should fall one way or another.

Here, by “uniform” Johnson means “arbitrary” or “without regard for purpose or circumstances.” The “balance” involved, however, is administrative–how competing institutional administrators (and their lawyers) might navigate various contracting situations. From an inventor point of view, all this is an institutional fight over rights of the form “who gets the first opportunity to strip the inventor of rights?” “Balance” in a policy on inventions doesn’t stimulate research, and doesn’t attract contractors, avoid monopolization, or anything else. Balance just postpones decisions.

Now Johnson gets to the crux of the matter. He believes that the key issue is who pays for what we might call “development” of an invention–the work necessary to turn an invention into a commercial product ready for sale. Keep clear: inventions to be practically applied do not have to become commercial products. They may be used (and improved) directly, by professionals not only in research but also in the provision of services. Thus, doctors may use patented techniques; researchers may use patented research tools; companies may use patented methods (as with software). Further, inventions may become used as standards–as is the case with internet communication protocols. The general case is not that any invention to be used must first become a commercial product, especially for research-originated inventions. The general case, rather, is that inventions may be used first, and later also become commercial products. For public interest research, in particular, where there is no meaningful argument for suppressing competition, the patent right that matters is “sale,” not “make” or “use.”

Here’s Johnson:

Again, this is an attorney dividing the matter. If the government funds development to the “point of commercial application,” then there’s no need for any contractor to claim title to any invention made along the way. That’s a huge point. It drives a couple of ways.

First, in matters of public health, why has the government given up the effort to develop inventions to the point of “application”–regardless of whether companies selling goods and services are involved? If the NIH has $30B or so each year to throw to university research, why not throw only $28B each year and reserve $2B for development of promising drug candidates? That might fully develop four or five new drugs per year, each of which could be made available non-exclusively for generic manufacture and sale, no need for patent monopoly pricing.

Second, why should not federal agencies have an easy default patent rights clause when they do intend to develop inventions directly rather than throwing the public good to the decision making of companies that depend on patent monopolies? From Johnson’s point of view, march-in would work just as well as the government taking title. Even if the government wanted to develop a public health invention to the point of application, commercial, professional, or otherwise, Johnson’s argument is that it should not be given that chance–instead, it has to allow a private company the opportunity exploit a monopoly patent position to protect its own invested money.

Even if you were prone to agree with Johnson that government title and march-in ought to do about the same thing (and therefore create a “uniform” appearance to the policy by choosing march-in, look what Johnson says about march-in:

Under the Kennedy patent policy, march-in “never needed to be used.” No government contractor was ever a “dog in the manger”–a company preventing others from using an invention but not using the invention itself. On the face of it, one cannot help but believing that Johnson has no evidence. In fact, that’s what Johnson asserts–“no one ever came up with any instance.” Johnson doesn’t say there never were instances in which inventions were not used because one company held a patent–he just says that “no one ever came up with any instance” of that. Someone gullible (or wanting to cover over a situation) might surmise that if no one had come up with an instance, that there never were such instances. But one would also have to surmise that someone had looked diligently for such instances and not found them, and that companies would be forthcoming about what they had been doing, and the like.

Consider Bayh-Dole. Bayh-Dole makes it optional for federal agencies even to request reports on the use of inventions owned by contractors. And if an agency does request a report, it must keep that report secret–not even share it within government without the contractor’s permission. That’s akin to not bothering to look, and not being able to “come up with instances.” Thus, the absence of information regarding nonuse is a “sign of success.” Isn’t political rhetoric fun?

The basic fact is that a patent permits the owner of an invention to exclude all others. If others then choose not to attempt to infringe (or to seek a license), then the basic fact is that they do without or design around the federally supported work. The basic fact, then, is that an arbitrary title policy creates incentives for all other companies to avoid, undermine, or design around the federally supported invention vested with a single contractor. That’s a queer sort of innovation policy. But there won’t be any obvious dogs in the manger preventing use–other than university folks like Caltech suing Broadcom or Stanford suing Roche–but even in those instances of university nonuse, the dogs don’t want to prevent the horse from eating the hay, they just want a piece of the horse’s benefit from doing so. Think parasite, not dog. There might not be dogs in the manager–but there could also be ticks.

This entry was posted in Bayh-Dole and tagged , , , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.