This, then, is the “policy” of Bayh-Dole that Allen champions–that nonprofits can and should deal in patent monopolies. Some historical bluffing from Allen’s policy advice (his emphasis):
At that time the federal government funded about half of the R&D in the country (it’s now a third) and the vast majority of basic research where breakthrough products are discovered. We found that 28,000 inventions had been taken by the government with less than 5% ever being licensed. We also discovered that not a single new drug had ever been developed under these policies. That was a tragic waste of billions of taxpayer’s dollars spent annually on government R&D.
Most of the 28,000 patents were defense related. For those, the contractors declined to take title and patent the inventions. A federal official testified that the patenting was done for “defensive” purposes and probably did not need to have been done. In all of those patents, according to Rebecca Eisenberg, there were 325 presumptively biomedical inventions, and the licensing rate for these was 23%:
For example, 325 of the 28,000 patents in the government’s
portfolio were from HEW, and seventy-five (or twenty-three percent) of these HEW patents were licensed as of the end of fiscal year 1976.
In the decade that the revived IPA program was in operation, universities reported 4 commercial products from 96 inventions. One can’t quite compare the 4% university rate resulting in commercial product (of whatever sort) with the 23% federal licensing rate (meaning, apparently, just granting an exclusive license without regard for outcome), but there’s a suggestion that the federal government was doing a way better job with inventions that it sponsored than the universities were doing.
“Licensed” in all these statistics is of course nearly meaningless. What matters is that something comes of the license–use, contributions to the technology, product development, new products. That a license has been granted and counted is meaningless. We can’t even tell from these accounts whether the licenses were exclusive or non-exclusive. If exclusive, then the federal licensing activity and the university licensing activity apparently advocated by Allen have opposite effects. We cannot compare the two activities by licensing rates. In the federal licensing practice, the “licensing rate” would indicate the number of inventions *no longer available to all.* In the Allen-advocated scheme, the “licensing rate” would indicate the number of inventions *available to only one company* with the rest *not available to anyone* pending some one company coming forward to take advantage of the “incentives” of the patent system. These are huge differences.
Allen continues, by citing nonprofit obligations imposed by Bayh-Dole. The impression is that these obligations somehow shape the incentives of the patent system for public good by constraining how nonprofits deal in patents on federally supported inventions. Think about it–if these constraints result in better outcomes than unconstrained licensing, then the patent system in general is flawed. But if these constraints make development more difficult, then Bayh-Dole is a burden on, not a stimulus to, commercial development of inventions.
Here is the start of his depiction of Bayh-Dole’s requirements on nonprofit owners of patents on subject inventions:
The law says that universities must give preferences in licensing their discoveries to small companies and those who agree to manufacture resulting products substantially in the U.S.
The substantive small business preference in the law was eliminated in the 1984 revisions, leaving a stub of a requirement in the form a gesture to small business preferences. Look for yourself (37 CFR 401.14(a)(k)(4)):
It will make efforts that are reasonable under the circumstances to attract licensees of subject invention that are small business firms
That is, nonprofits have to try to attract small businesses as licensees. So far so good. No federal agency to my knowledge enforces this requirement. No agency asks a university to demonstrate that it has made efforts to attract small business licensees. But then there are the usual Bayh-Dole walk-backs:
and that it will give a preference to a small business firm when licensing a subject invention if the contractor determines that the small business firm has a plan or proposal for marketing the invention which, if executed, is equally as likely to bring the invention to practical application as any plans or proposals from applicants that are not small business firms
The implication is that the preference for small businesses applies only to exclusive licensing. Otherwise, there would be no particular reason to have to prefer one licensee to another, such as a small company with a big company, or compare their marketing plans. The rest of the constraints are just super-silly: a nonprofit must prefer a small business licensee when it presents a marketing plan that is as good or better than any non-small business’s marketing plan. Ho there! The preference implies there are multiple companies competing for a license. Otherwise the preference reduces to “don’t be stupid–choose an exclusive licensee that has the best chance of achieving “practical application.” More walk-back:
provided, that the contractor is also satisfied that the small business firm has the capability and resources to carry out its plan or proposal.
Not only must the plan be just as likely, the small company has to have the capability and resources to execute the plan. It’s not that there’s a reasonable expectation that the small company can obtain the capabilities and resources that the plan requires–but that the small company already have these. But that’s not the end of the walk-back:
The decision whether to give a preference in any specific case will be at the discretion of the contractor.
The nonprofit decides whatever it will. There is effectively no particular preference for small businesses. What is the government’s remedy?
However, the contractor agrees that the Secretary may review the contractor’s licensing program and decisions regarding small business applicants, and the contractor will negotiate changes to its licensing policies, procedures, or practices with the Secretary when the Secretary’s review discloses that the contractor could take reasonable steps to implement more effectively the requirements of this paragraph (k)(4).
The federal government cannot force changes to past practices that ignore the small business preference–it can “negotiate” changes in future practices to be “more effective.” The federal government has never bothered with this provision, as far as I know.
By contrast, in the original version of Bayh-Dole, nonprofits were forbidden from granting exclusive licenses to non-small companies:
(B) a prohibition against the granting of exclusive licenses under United States Patents or Patent Applications in a subject invention by the contractor to persons other than small business firms for a period in excess of the earlier of five years from first commercial sale or use of the invention or eight years from the date of the exclusive license
With a typical Bayh-Dole walk-back:
excepting that time before regulatory agencies necessary to obtain premarket clearance
And another walk-back, to allow federal agencies to waive the requirement (which, we might expect, that the NIH would be more than willing to do):
unless, on a case-by-case basis, the Federal agency approves a longer exclusive license.
So there’s effectively no small business preference in Bayh-Dole. Allen wants his audience to believe something, but the law doesn’t support him.
Allen, repeated for convenient reference:
The law says that universities must give preferences in licensing their discoveries to small companies and those who agree to manufacture resulting products substantially in the U.S.
Despite Allen’s syntax, Bayh-Dole does not require a preference for U.S. manufacturing. Although 35 USC 204 is labeled “Preference for United States industry”–the provision itself is stated as a requirement, unless the preference is coded into the adverb “likely.” Go see for yourself. In any event, the U.S. manufacturing “preference” in 35 USC 204 is meaningless, too.
The U.S. manufacturing “preference” applies only to exclusive licenses to use or to sell in the U.S.–and thus does not apply in the case of assigning an invention to a company (as any small business contractor might do with inventions it acquires subject to Bayh-Dole’s small business patent rights clause). Further, the 204 requirement is easily worked around–grant a single non-exclusive license in the U.S. rather than an exclusive license. Or, grant non-exclusive licenses to use or sell in the U.S., but grant an exclusive license to manufacture to a foreign company. That frustrates the 204 requirement, but it’s perfectly compliant with 204’s requirement. Not that any of this matters, because federal agencies can waive the requirement on the condition that a contractor has (paraphrased) “tried and failed” to find an exclusive licensee to use or sell willing to acquire product made in the U.S. or “hasn’t bothered to try at all.” The NIH, to make the waiver process go smoothly, has a web site devoted to fielding applications for waivers. If a federal agency actually wanted to enforce section 204’s U.S. manufacturing “preference,” it would have to use the march-in procedures, which are a tangled mess of determinations, appeals, and further appeals that Howard Bremer surely was correct that march-in was designed not to operate.
Finally, Allen:
Government agencies are allowed to use inventions they fund to meet their needs (generally advancing science or meeting mission requirements).
Bayh-Dole (35 USC 202(c)(4)) requires federal agencies to use a standard patent rights clause that requires contractors to grant to a non-exclusive license to “practice and have practiced by or on behalf of the United States.” “Practice” means in executive branch patent policy “to make, use, and sell.” That’s the same meaning given to “practice” in the IPA program, which Latker (who drafted Bayh-Dole) claimed was the basis for Bayh-Dole. We don’t even have to go to the difference between “the United States” and “the federal government of the United States.” In executive branch patent policy (and the IPA program that somewhat complied with that policy) the non-exclusive license included state governments and domestic municipal governments. That’s a much broader right than Allen’s depiction of “to meet their needs.”
We could go on about Allen’s depiction of march-in.
The government can also march in if the developer is not able to meet the production needs of a health or national security emergency.
Laughable. Bluff. Here’s the provision, in 35 USC 203:
. . . the Federal agency . . . shall have the right . . . to require the contractor . . . to grant a . . . license . . . if the Federal agency determines that such
(2) action is necessary to alleviate health or safety needs which are not reasonably satisfied by the contractor, assignee, or their licensees
No mention of emergencies. In an emergency, march-in could not possibly operate–a contractor–or assignee–or exclusive licensee–could drag out the march-in process for months, if not years. Bayh-Dole’s march-in is more aptly called “creep-in.” The march-in is not restricted to “production.” Anything that does “not reasonably satisfy” health or safety needs may trigger a march-in. Availability, rather than production. Price, rather than production. Alternative formulations, rather than production. Alternative compounds covered by the same patent, rather than production.
It is easy to play fast and loose with history and law to leave an impression. That’s politics, of course, and bluffing is to be expected. We need not be shocked. But we also don’t have to pay any further attention to bluster. And we certainly should never mistake bluffing for fact.
I’ve had about enough, but here’s one more assertion:
Universities license about 70% of their patents to small companies. Even in high risk, high cost industries like drug development, small companies are prominent players. Half of our new drugs originate in small companies.
The first statement is unsupportable. Stanford reported in 2009 that it licenses only about 20% of its inventions. No university I know of licenses 70% of its patents, unless it has only, say, five patents, or it only obtains patents after it has a company that has agreed to take the license. And of course licensing a patent says nothing about making an invention available for public benefit. An exclusive patent license sequesters an invention in private hands. That’s the “incentive” of the patent system that Allen alludes to. Otherwise, we are back to the non-exclusive Cohen-Boyer approach that forgoes a trade in the patent monopoly and which tracks the past federal approach, which Allen disparages and in its place presents Bayh-Dole as the proper policy alternative.
Furthermore, the quoted argument above shifts from Bayh-Dole inventions to university patents generally. Universities have obtained about 50,000 U.S. utility patents on Bayh-Dole subject inventions, and another 120,000 patents on non-Bayh-Dole inventions. No one breaks out licensing of Bayh-Dole patents. I don’t see how Allen would know the university patent licensing rate. AUTM doesn’t ask. The federal government might request this information, but Bayh-Dole makes the information a government secret. And of course, a licensing rate is meaningless without an account of outcomes in the form of practical application–use with benefits available to the public on reasonable terms. Or, rather, without an account of outcomes, a licensing rate for an exclusive licensing practice (one taking advantage of the “incentives” of the patent system) recites the number of inventions that are made not available for broad public access. The figure 70% is presented to sound impressive. If it points to exclusive licensing practice, without a report of the outcomes for those exclusive licenses, it is just the opposite. It should be consternating, horrifying, an absolute disaster. If 70% figure is correct, then the argument for Bayh-Dole favors disaster as federal policy. If the figure is incorrect, of course, then no one should bother with the argument.
Perhaps what is meant is that of the licenses that universities grant, 70% go to small companies. But then we are dealing with AUTM’s definition of a license–which for some time has been any license for anything that results in a payment of $1,000 or more. That is, one can license biomaterials or software for end use and count that as “commercial license”–and it has no relevance to Bayh-Dole or to the argument for the “incentives” of the patent system. Similarly, universities routinely grant non-exclusive licenses to research sponsors–often now upfront and royalty-free. Heck, on that argument, a university grants a license to every Bayh-Dole invention (and thus to every patent on every Bayh-Dole invention)–to the federal government. That’s also typical practice in research consortia.
Again, such licensing has nothing to do with the Bayh-Dole small business “preference” and the metric (whatever it might be) does not show that Bayh-Dole has any influence on small business licensing. Finally, universities have adopted the practice of licensing to their own startups–often nothing more than a shell company without operations or the repackaging of on-going research to horn in on SBIR funding that would otherwise go to small companies formed independently of any university. The small company licensing claim is another bluff. There’s no documentary evidence to connect it to Bayh-Dole. And it’s irrelevant whether in general small companies are “prominent players” or that they produce new drugs. What’s the connection between these assertions (however true) and Bayh-Dole inspired practices? More bluff.
Bluffing is part of politics. But when the politics makes claims about how research results might produce public benefits, but really the claims are about extracting money from science using public subsidies, it’s worth giving anyone pause in deciding whether to rely on bluffs for policy advice.