The NIH’s View of Bayh-Dole Compliance, 5

We are working NIH’s not so tasty guidance to participants in its SBIR and STTR programs directed at small businesses. We reach the NIH’s account of the “principal features” of Bayh-Dole, at least with regard to “intellectual property” requirements:

Principal Features of the Bayh-Dole Act

Organizations are required to establish a written agreement with all employees to disclose promptly each subject invention made under a federally sponsored program and to execute all papers necessary to file patent applications.

And here’s the first principal feature–of all things–the written agreement requirement. Well, the written agreement requirement is not in the Bayh-Dole Act. It’s not even in the implementing regulations. It shows up in the standard patent rights clause. There’s some logic to why it is there–we have just been through all that–but the written agreement is most certainly not in the Bayh-Dole Act.

Just to be clear: the principal feature of the Bayh-Dole Act is that federal agencies are limited in the conditions under which they can require a contractor to assign to the federal government an invention made with federal support if a contractor has otherwise acquired ownership of that invention. That’s the essence of 35 USC 202(a).


  • pre-empts other laws, but only when the contractor owns;
  • provides protocols for federal agencies to vary from the restrictions;
  • provides default requirements for contractors to retain ownership of inventions;
  • provides requirements for contractors to use inventions for public benefit on reasonable terms, mostly waivable by federal agencies.

The upshot is that Bayh-Dole creates a pipeline of patent monopoly rights in federally supported research to private companies, in secrecy, without public accountability, and with no enforcement of the public protection requirements in the default patent rights clause.

In practice, that pipeline principally exits to serve the pharmaceutical industry and by extension the biotech venture investment industry. The vast majority of federally supported inventions never emerge as products, but are held behind patent paywalls anyway. Money is made trading on speculation of future value of the patents. The rare products that do emerge are then sold at monopoly prices, not on “reasonable terms.” The claim made is that no product would ever emerge from the pipeline without having the benefit of a full patent monopoly.

Thus, the fundamental claim about Bayh-Dole–the principal feature–is that the federal government should subsidize the monopoly interests of speculators on health-related discoveries financed in the public interest by the NIH and other federal agencies.

Look, then, at how the NIH describes the written agreement. First, they double up “subject invention” and “made under a federally sponsored program.” The appearance is that the defining characteristic of a subject invention is that it is one that has been supported by federal funding. But the defining characteristic that matters is that the invention is owned by a contractor. Any invention–even ones made in work receiving federal support–that is not owned by a contractor is not subject to Bayh-Dole. “Subject invention” then includes “made under a federally sponsored program.” The thing that an SBIR recipient might want to know, rather, is that Bayh-Dole doesn’t apply until the recipient owns a given invention made under a federally sponsored program.

Given that small companies especially are not so stupid as to spend prodigious amounts of their money filing patent applications on inventions that have nothing to do with their present or future business, there are good reasons why small company officials would be pleased to discover they can avoid Bayh-Dole altogether by making employees sign a complying (f)(2) written agreement to protect the government’s interest, report to the government what the inventors report to the small company’s patent personnel, and not assume any responsibility to file patent applications until the company has reviewed each invention disclosure for relevance to the company’s business.

Here’s the NIH again:

to establish a written agreement with all employees

Here’s paragraph (f)(2) of the standard patent rights clause:

agrees to require, by written agreement, its employees, other than clerical and nontechnical employees

The requirement is not for all employees. Clerical and non-technical employees are excluded. So are all non-employees–volunteers, visitors, students (in the case of universities).


to disclose promptly each subject invention made under a federally sponsored program and to execute all papers necessary to file patent applications.


to disclose promptly . . . each subject invention made  under contract . . .  and to execute all papers necessary to file patent applications . . . and to establish the government’s rights in the subject inventions.

The NIH leaves out the part that shows the fundamental effect of the (f)(2) written agreement–that inventors, not the organizations, own the subject inventions, and the written agreement cedes to inventors the responsibility to establish the government’s rights–since the inventors have these rights and the organizations do not. Let’s see how long it takes the NIH to find this web page and modify it to add NIST’s strange assignment language, effective May 15, 2018. NIH didn’t update based on the Supreme Court ruling in 2011. Perhaps NIST carries more weight.

By its acceptance of an NIH award, a grantee or contractor organization agrees to obtain written agreements from its employees and:

Here NIH is torn between the language of Bayh-Dole–“contractor,” a singular noun representing the collection of all parties to a given funding agreement–a singular noun representing a plurality of contractors. Like a pod of whales. A herd of administrators.–and the language of grants–“grantee” or “recipient” (now washed away to “non-federal entity” in 2 CFR 200, sigh).

And of course “its” employees should be something like “its technical employees”

  • Promptly report inventions to the NIH.

Yup. Except it’s wrong. Contractors are to report *subject* inventions. That is, patentable inventions that the contractor has acquired. Not simply “inventions” or “all inventions” that are made in work receiving NIH support. NIH overstates the Bayh-Dole requirement and therefore suppresses the structure of Bayh-Dole. Bayh-Dole is a conditional, additional layer of bureaucracy that rides on top of federal laws and executive branch patent policy. Bayh-Dole operates only when a contractor–a party to a federal funding agreement–acquires an invention made in the performance of federal work. In acquiring such an invention, a contractor turns the invention into a subject invention. It is subject inventions that contractors must disclose to the NIH, and then only within a specified period after the subject invention has been properly disclosed to the contractor’s personnel designated for patent matters.

  • Elect, in writing, within two years, whether or not to retain title.

Nope. First the contractor has to obtain title. And obtaining title is *not* a condition of accepting federal funding. This is pure bunk.

  • File a patent application within one year of electing title.

“Electing title” is wrong and misleading. “Electing to retain title” is correct.

  • Acknowledge government support in the patent application and send page of application containing federal support clause.

Sure. But sending the page is just bureaucracy. Does nothing for innovation.

  • Provide the government with a royalty free license to the invention; the confirmatory license should be sent to the Office of Policy for Extramural Research Administration (OPERA), NIH.

The royalty-free license is due when a contractor elects to retain title. 37 CFR 401.14(b). The license is not “confirmatory”–it is a real license that includes a waiver of the contractor’s right to seek compensation for use and manufacturing of the invention by or on behalf of the government in the Court of Federal Claims; further, the license is “to practice and have practiced by or on behalf of the United States”–that is, “the government” is not merely the federal government. Under the executive branch patent policy the license includes the federal government, state governments, and domestic municipal governments. “United States” is not merely “the federal government of the United States.”

  • Make reasonable efforts to attract small business licensees.

Only for nonprofits. It’s an empty requirement. Go read all the prissy qualifications and walkbacks that make it only for show. In the original Bayh-Dole Act, nonprofits could not grant exclusive licenses longer than eight years from the date of license to a non-small company. That requirement was removed three years later–before it could even have been invoked (it takes about three years for a patent to issue). So Bayh-Dole’s small business preference was gutted before the law had a chance to operate.

  • Provide annual reports on the utilization of the invention, including date of first commercial sale or use and gross royalties received.

Notice that “date of first use” does not require any commercial sale. Utilization != commercialization, even for companies. The reports are to be excluded from public disclosure–so there’s no public accountability. One would think the NIH would point this out to small businesses.

  • Agree that exclusive licensee [sic] will manufacture the invention substantially within the US, if it is to be used or sold in the US.

This overstates the requirement. An exclusive licensee to use or to sell in the US must agree that any product embodying or produced through the use of a subject invention will be manufactured substantially in the US. That is, the exclusive licensee to use or to sell may not have any license to manufacture product. 

There is no restriction on exclusive licenses outside the United States, including to manufacture. So a patent owner can grant an exclusive licensee outside the US to manufacture, and so long as any license to use or sell in the US is not exclusive (two distributors, say), the US manufacturing preference does not operate. NIH makes it appear that any exclusive license must require US manufacture for US use or sale. Not so.

However, even the narrow version of the requirement doesn’t matter. The requirement also may be waived by the federal agency, and NIH has so many waiver requests that it set up a web site to process them all.

Invention Reporting though iEdison

All inventions made using NIH supported research must be reported in iEdison (Interagency Edison). You can use iEdison to learn about the law and regulations, and report an invention or patent funded by any of the participating federal agencies.

Nope. Just subject inventions. At least with regard to Bayh-Dole. It’s true that federal agency regulations apply to inventions that are not subject inventions–but the NIH should then make clear the difference between subject inventions–owned by a contractor–and other inventions.

It’s important to remember that invention reporting data is confidential and the government license in no way allows the government to compete with the grantee/contractor for licensing of the invention. All inventions should be reported prior to any publication or presentation of the invention at an open meeting.

Bayh-Dole certainly attempts to exclude invention reporting data from FOIA. It’s not clear that Bayh-Dole meets FOIA’s statutory requirements, however. Bayh-Dole in essence declares that federal agencies must treat invention reporting data as privileged and confidential even if it is not. That’s not FOIA compliant.

But what is this?

the government license in no way allows the government to compete with the grantee/contractor for licensing of the invention

The government license is to practice and have practiced by or on behalf of the United States. Practice = “make, use, and sell.” Anywhere the United States has authority to act, it has the right to authorize the making, using, and selling of subject inventions. There is no restriction that the government will withdraw from its responsibilities merely because a company wishes to expand its potential market to sell products embodying subject inventions. The NIH has absolutely nothing to rely upon in the government license to make the noncompete claim it makes here.

But this is at the very heart of the NIH pipeline that has been in place since at least 1968 with the revival of the IPA program and continued by Bayh-Dole when the IPA program was shut down as ineffective. The NIH in essence asserts that the NIH will withdraw from any responsibility it might have to make a subject invention available or authorize anyone else to make that subject invention available for use by or on behalf of the United States if the grantee/contractor wants to license the invention for such use. Bizarre.

We may also note how bizarre it is that the NIH should be worrying about licensing of subject inventions as something important for SBIR and STTR grants. True, STTR grants can go to nonprofits–but they have to subcontract to a small business, and the nonprofit is forbidden to have an interest in the small business’s subject inventions, and there’s an implication of a compulsory license from the nonprofit to the small company for the nonprofit’s subject inventions.

The premise of the SBIR program is that small companies receive federal support to move through bureaucrat-approved phases of feasibility research and development of a new product, culminating (in theory, not much in practice) in a procurement agreement under which the government buys quantities of the new product and the company sells the new product to everyone else as well. Why would a company that has been presented with this goal worry about licensing the invention? Isn’t that, like, being a patent troll? If the invention isn’t in the company’s line of business, why would the company take ownership of it in the first place? How would the company even have a right to take ownership in such a case?

And why would a small company want to retain ownership of an invention that’s not in its line of business, what with the bother of wasting money on patenting? and the bother of secret reporting? Oh, would that be to troll industry? Not that? Oh, to license to big pharma? Ah, the pipeline. But why would a small company license then at all when it is free under Bayh-Dole to assign the invention–to sell it. Why the heck would anyone license something that’s not in their line of business when they could sell it? NIH is clueless not only about Bayh-Dole but also about IP practices.

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