Whistling all the way to the bank, revisited 3

Having established the contracting problem for government-sponsored “basic research,” let’s get into how the patent administration folks got into changing things around from government ownership to institutional ownership, when the institutions didn’t have policies and practices that swept up inventions from their inventors.

In its Institutional Patent Agreement program, the NIH signed a master agreement with each university that required the university to have a patent agreement with each employee involved in the federally funded research that provided that if the university chose to file a patent application, then the employee-inventor agreed to assign the invention to the university. That is, once the university had prepared a patent application, then the employee-inventor promised to assign the invention to the university. The policy was “catch only what you intend to eat” not “catch everything and release only what you are forced to release.” That is, the IPA program required selective logging, not clear cutting.

A university’s management of the invention was subject to the IPA program master agreement. The IPA program was shut down in 1978, just as the NIH was attempting to make the program government wide (and so would permit the pharmaceutical industry to use universities to catch potentially therapeutic compounds in all federal research, not just research supported by the NIH and NSF).

Bayh-Dole attempts to rebuild the IPA program, but now as a matter of patent law. Thus, we have a law that stipulates a “standard patent rights clause” that looks a great deal like the IPA master agreement template–and we might expect this to be the case, given that the IPA master agreement, and Bayh-Dole, and the standard patent rights clause were all drafted, it appears, by the same person, the late Norman Latker, one-time patent counsel at the NIH.

In the IPA program, inventions were deliverables to the government under the authority of the Kennedy executive branch patent policy (later modified by Nixon but kept largely intact). The IPA agreement allowed the NIH to permit a nonprofit contractor to retain “principal” rights prior to an invention being reported, as permitted by the Kennedy policy in “special situations”:


The cleverness of the IPA scheme was to expand “special situations” to be any situation in which a nonprofit presented an approved invention  management program and entered into an IPA program master agreement. Under federal regulations, inventors must assign their inventions made under a federal contract to the federal government unless the contractor met the conditions of the executive branch patent policy for permitting the principal rights to remain with the contractor (who then would have to get the rights from the inventors, in the usual fashion, by assignment).

The NIH IPA scheme, then, expanded beyond section 1(b) inventions (where a contractor had both the technical capability and an established non-governmental commercial position, and the research was intended to “build upon existing knowledge or technology to develop information, products, processes, or methods for use by the government”) to include inventions in research in “fields which directly concern the public health”–for which the Kennedy patent policy provided no such “special situation.” Ballsy.

The IPA program extended things even further, by swapping out the requirement that a contractor “take effective steps within three years after a patent issues on the invention to bring the invention to the point of practical application or has made the invention available for licensing royalty-free or on terms that are reasonable in the circumstances” for one in which the contractor has three years from the date of first commercial sale or eight years from the date of the exclusive license, whichever is earlier. That’s a huge change, not authorized by executive branch policy. Again, this was clever, not compliant. But Latker got away with it for ten years in the IPA program before getting shut down.

Bayh-Dole was an effort to make the non-compliance both legal and permanent. The “special situations” of the Kennedy patent policy became the default, and Bayh-Dole requires “exceptional circumstances” for federal agencies to reverse the default. Thus, in Bayh-Dole we find a law that requires a master template agreement–the IPA, reborn–but now embedded in patent law, not subject to executive branch policy.

But this is where Bayh-Dole spins out of control. Perhaps Latker lost control of changes as the bill worked its way through Congress, perhaps he botched the drafting. Again, in the IPA program, inventions were deliverables to the government as a matter of executive branch patent regulations. The IPA allowed nonprofits to intercept any given invention deliverable, provided that to do so, the nonprofit had a promise from inventors to assign the invention when the nonprofit had decided to file a patent application. The IPA allowed the institutional intercept of federal invention deliverables. 

Here’s the difference for inventors. Without the IPA, the invention goes to the federal government and the government either “dedicates” it to the public or files a patent application and offers the invention to the public on a royalty-free basis. Either way, the inventor is assured of having free access to the invention for further use and development. But with the IPA in place, the institution takes ownership of the invention, with an expectation that it will license the invention exclusively–and it has an indefinite period of time in which to secure such a license, during which time the inventor has no right to practice the invention outside the institution that hosted the research leading to the invention. It is not merely a matter of whether the federal government or the host institution takes ownership of the invention–there’s a huge difference in terms of public access. If the federal government takes title, then the public has access. If the institution takes title, no one does–until and unless an exclusive license is granted. (Yes, there is an apparatus in the IPA master agreement that suggests that an institution will consider non-exclusive license first, but in practice, the institutions almost always licensed exclusively–and a similar but weaker gesture is made in Bayh-Dole for the management of inventions owned by the federal government).

It is this difference in public access (and inventor access, and principal investigator access, and research access) to inventions made with federal support that do not directly concern public health that is at the heart of the insistent cleverness of Bayh-Dole. The advocates for Bayh-Dole argued that public access destroyed the commercial potential for inventions (pay no attention to the digital computer), and that universities could do a better job finding licensees than could the federal government (the data showed this had not happened, but no matter), and that somehow preventing public access to publicly supported inventions would restore America’s global technology leadership. It was pretty much crap. University folks admitted in Congressional testimony that they hadn’t done a very good job under the IPA program–but they could get better. They admitted that moving inventions directly to companies without an intervening nonprofit licensing step was more effective (as the Harbridge House report indicated), but “the public isn’t ready for that yet.”

The IPA, and then Bayh-Dole, aimed to prevent public (and inventor, and investigator) access to inventions made at nonprofits, and to translate the value of exclusion into profit-taking for institutions, investors, and companies exploiting patent monopolies to prevent competition and maximize prices. That’s the policy debate in a nutshell. Are we better off routing our inventions to nonprofit institutions in a largely failed attempt to share in the profits from the exploitation of our health issues as a speculative commercial market? Are we better off with bureaucrats deciding how to deploy research inventions than with the inventors, or with the federal government expanding the public domain with publicly supported research? It is a rancid nut in that nutshell: extract maximum profits from the suffering–that’s where universities will get more money for research. When Bayh-Dole’s advocates argue that patent monopolies are the only way that the public will benefit from publicly supported research, they are not stating a fact–they are stating a purpose, an objective: no public benefit without profit-taking enabled by patent monopolies.

Ask yourself: is that the public policy you want? It didn’t work for the IPA program. It hasn’t worked for Bayh-Dole, other than that it appears to be true, that the profit-takers have taken their share from the inventions that have survived the Bayh-Dole process of institutional management.

So now we come to how the apparatus of Bayh-Dole was supposed to work but got cocked up. Bayh-Dole was supposed to enable the same intercept of invention deliverables as the IPA program, but without the need for a master agreement–the master agreement would be a standard patent rights clause attached to every funding agreement. If an institution was worthy to receive a federal grant, it was worthy to intercept the invention deliverables. That’s why we have the “elect to retain title” language in Bayh-Dole. There had to be a standing federal requirement that inventors assign title to the federal government for the intercept to work. But Bayh-Dole asserts precedence, and voids any such federal requirement. That’s the first cock up.

Otherwise, Bayh-Dole’s scheme would have been the IPA’s scheme. “You inventor are required by federal regulation to assign all inventions to the federal government, except your university has an option to receive the government’s right to receive your assignment. By notifying a federal agency of its decision to “elect to retain title,” a university received the federal government’s right to require assignment of title as a condition of the federal award. The university stands in for the federal government with regard to invention deliverables, takes title (because the inventor is required by federal regulation to assign title), and passes through to the federal government a non-exclusive license to practice and have practiced. The standard patent rights clause only has to have a requirement that inventors agree to establish the government’s rights inventions (the (f)(2) requirement) because the “elect to retain title” notice intercepts this government right. The government, as it were, through this notice designates the university to stand in for the government. The university gets the government’s rights, and then grants its own license to the government, and the government cannot get more rights unless the university fails in specific ways–to disclose, to file patent applications, and the like. But Bayh-Dole cocks up and takes control of the government’s rights–the government gets only what Bayh-Dole allows it to get, and that’s not much. But worse, doing so disables the intercept of invention deliveries. There’s nothing to intercept. 

The advocates for Bayh-Dole must have had a dark night of the soul when they realized they had got their bill passed, but it lacked the central driver to make the scheme work. Rather than try to fix it (and call attention to exactly what they didn’t want to call attention to), they just announced that the law worked by cleverly meaning that “elect to retain title” meant “elect to take title.” That held them for thirty years, until Stanford v Roche, when they thought they had a chance to make their claim legal, and failed. The second cock up.

As a result, Bayh-Dole stipulates a patent rights clause, but that patent rights clause has nothing to say about the initial ownership of inventions–not with regard to inventors nor the federal government nor institutions hosting research. And given the rules of precedence, the only way the federal government can make any change in the disposition of ownership is by declaring exceptional circumstances. There is no authority in Bayh-Dole to require institutions that host federal research to require assignment to them of inventions. It would violate Bayh-Dole for NIST to change the (f)(2) written agreement provision to become an institutional assignment requirement. Doing so is outside the authority of Bayh-Dole, turns Bayh-Dole back into a vesting statute against the interpretation of the Supreme Court and what it would find “deeply troubling,” and besides, is a thoroughly rotten idea for which the data (what of it there is) is overwhelming against the Bayh-Dole institutional invention intercept practice.

Instead of an intercept of inventions otherwise obligated by a federal funding agreement to the federal government, the Supreme Court examined Bayh-Dole and found nothing, nothing, nothing at all. The “elect to retain title” happens only after an institution has obtained title through ordinary means. If the institution does not obtain title, there’s nothing to elect to retain. If the institution obtains title and choose not to elect to retain that title, then the federal government can acquire title. This, in essence, is the primary method permitted by Bayh-Dole for the federal government to acquire ownership of inventions made with federal support. If universities and other contractors did not get assignments, Bayh-Dole requires the federal government to deal with the inventors, and can get title to inventions only if an inventor fails to report the invention or chooses not to file a patent application, or files a patent application later than required. This is, from a Bayh-Dole drafting point of view, yet another cock-up.

So why couldn’t Bayh-Dole just require the standard patent rights clause to include a requirement that inventors must assign to the institution that hosts the federally supported research, just like the IPA program? Well, one problem is that Bayh-Dole is part of federal patent law, and federal patent law is based on the common law principle, reaffirmed by the Supreme Court in Stanford v Roche, that inventors own their inventions. The federal government has constitutional authority to reserve for those inventors exclusive rights to their inventions–and those exclusive rights take the form of federal patent law. It is not possible for federal patent law to reserve those exclusive rights for non-inventors, for institutions that happen to host research and don’t even control or direct the research–the research is not done for their benefit. To do so lies beyond the federal government’s constitutional authority to reserve rights for inventors, not some predatory nonprofit. If Bayh-Dole wasn’t a part of patent law–say, a matter of regulation of federal research contracting–then it could require inventors to assign their inventions to the government and work the intercept scheme–the inventors assign to their host institution because the host institution is permitted to stand in for the federal government. when it chooses to do so. So, more cock-up.

We might put it simply, after all this, that Bayh-Dole aimed to implement the intercept strategy used by the IPA program, but because it was placed in patent law and asserted precedence over executive branch regulations that required federal ownership of inventions made under federal funding agreements, Bayh-Dole failed to implement the intercept, was left with a strained appearance of vesting, and that it did no vesting was made clear by the Supreme Court in 2011. For the intercept strategy to work, Bayh-Dole would have had to leave in place the executive branch policy requiring delivery of inventions made in federally supported research to the federal government unless it agreed otherwise. But Bayh-Dole was intended to disrupt this policy because the idea was that the “unless it agreed otherwise” should be turned on its head. But the logic fails.

One gets to: “the government cannot require delivery of inventions made in federally supported research unless an institutional host of that research agrees otherwise.” That doesn’t enable the intercept strategy because now the federal government has no delivery requirement to option off to whomever it pleases (such as a research host). And one wonders if it is legal (or proper) for the federal government to, in effect, privatize federal rights with such an intercept. The greater wonder, of course, is whether the effect of introducing billions of federal money into expanding university basic research should be the occasion to also expand private patent speculating that fragments shared research enterprises and prevents access to research results for everyone but the chosen (if ever) exclusive commercial developer (who may never develop a commercial product anyway, but may try). One wonders, too, if such federal largesse of funding has fundamentally altered who university faculty seek out to make a case for funding their next research project–members of the community, or industry, or a federal funding agency? At some point, does providing a pleasing proposal to a federal agency review committee translate so well into research findings that will be useful to any community–including, even, the research community itself?

As it is, the cleverly crafted scheme worked for the Bayh-Dole advocates. They were able to reverse the defaults, so federal agencies were forbidden by law from requiring inventions as deliverables in university research agreements if the invention became a subject invention–the first principle of Bayh-Dole. The rest is pretty much a cock up, spun with a glitzy story, and we have survived it, mostly, but at a cost we have not had the heart to consider. Thus, the data on invention use remains a government secret, per Bayh-Dole.

This entry was posted in Bayh-Dole, History, Sponsored Research, Technology Transfer and tagged , , , . Bookmark the permalink.

Leave a Reply