The failed Bayh-Dole bargain, 2

We are working through an article about the possibility that the federal government might use march-in to address unreasonable pricing for Xtandi. The case for march-in is as strong as any could be. But to get at this issue, our authors choose to frame things as a matter of what’s “uniform” about Bayh-Dole. The implication is that Bayh-Dole is supposed to make federal policy “uniform” and since march-in for Xtandi would not be uniform, well, then, there’s your implicit answer.

Now let’s look at Bayh-Dole. Is Bayh-Dole “uniform”? Hardly. The law suffers from a host of defects with regard to “uniform.” First, exceptions are baked into the law. The primary exception is for DOE nuclear research and development. The DOE fought to have its “non-standard” default reflected in the Bayh-Dole’s implementing regulations. There were then two default patent rights clauses–at the old 37 CFR 401.14(a) and (b). The design idea in presenting these clauses as (a) and (b) was that as agencies identified other “exceptional circumstances” under 35 USC 202(a), these could be codified at 401.14 as (c), (d), and the like. Exceptions were anticipated. Non-uniformity was anticipated. Second, Bayh-Dole goes out of its way to pre-empt any federal statute that dictates an invention ownership policy at odds with Bayh-Dole. Except that Bayh-Dole does not pre-empt the Stevenson-Wydler Act. Again, not uniform. Furthermore, Bayh-Dole provides that Congress can pass laws that also pre-empt Bayh-Dole, provided such a law cites Bayh-Dole and expressly indicates that the new law takes precedence. Not uniform.

We can go on, and will. Bayh-Dole applies only to nonprofits and small companies. Senator Dole cut a deal with Senator Long that Bayh-Dole would not be extended to large companies. There’s your “bi-partisan” deal. So instead of amending Bayh-Dole to cover all contractors, Senator Dole and Norman Latker got President Reagan to change executive branch patent policy through a 1983 Memorandum and a 1987 Executive Order. But this executive action could not pre-empt federal laws (and so they still applied, where they had not been subsequently amended) and in any event was drafted so that federal agencies were authorized to follow Bayh-Dole in contracting with large for-profits, but did not have to. Again, not at all uniform.

And that’s not all, though I wish it were. Federal agencies have found ways around Bayh-Dole altogether. One way is by means of “Other Transaction Authority.” Essentially, the ruse is, if an agency does not call the transaction a grant, contract, or cooperative agreement, then it is not within the definition of “funding agreement” at 35 USC 201(b) and therefore outside of Bayh-Dole. Another is to change the purpose of the contracting, so that it is not directed at research or development but rather is a “loan” or “investment.” That the recipient uses the money for research is beside the point, since the “official” purpose is to loan money or invest. Yet another is to provide additional money on the indirect costs or “Facilities & Administration” side of a contract. In this scheme, a contractor is paid money on the direct funding side of the contract, under Bayh-Dole, but is also paid money on the indirect side. Since the direct side research is often subject to competitive bidding, that side is relatively fixed. But companies–unlike nonprofits–have much more flexibility in negotiating the indirect cost payments. So they can propose work to be undertaken with indirect cost funding, and this work is outside (so the explanation goes) of Bayh-Dole. It’s called “Independent Research and Development” or “IR&D.” See for instance 48 CFR 205-18. Not uniform. Got it? Yet another workaround is to label the transaction “procurement.” The government then is just buying something, maybe for a lot of money, and if a company has to spend some of that money figuring out how to make what the government is buying, that’s outside of Bayh-Dole.

And then there are the regulations implementing Bayh-Dole’s public protections. A federal agency can march-in for nonuse or unreasonable use of an invention. Except bizarrely there’s a bit of regulation (37 CFR 401.6(h)) that says a federal agency doesn’t have to march-in, even if it has found nonuse or unreasonable use. A federal agency can waive the requirement for US manufacturing in the case of an exclusive license in the US to use or to sell. And a federal agency can choose to enforce or not enforce pretty much anything else–such as whether to acquire an invention in the case of a contractor failing to comply with the conditions for retaining title, such as filing a patent application. Nothing here is uniform. And these are the same sort of agency decisions that created “uncertainty” for the sad university-affiliated patent administrators who just could not handle not knowing if a federal agency would be persuaded that their plan for exclusive control better served the public interest than would open access. Think about it–if the government acquired the invention, then the invention management firm would have–most assuredly and certainly–access to the invention, and could transfer it, with its technical information, any prototypes, and with inventors attached to explain every detail. All it would not have is the right to exclude all others (even the inventors) from practicing the invention, but for their permission (and, generally, payments). The government not acquiring title changes the form of “technology transfer” undertaken by contractors and inventors. There’s your policy change to consider–technology transfer mediated by exclusive licenses granted (or not) by organizations or any other of the ways that technology change comes about. Bayh-Dole practice suppresses all those other ways and drives away potential adopters who do not want, or need, or could not obtain a license from an organization holding rights.

It’s nonsense to claim that Bayh-Dole establishes a uniform federal patent policy. It does no such thing. It just changes where the variations are. What it does do is remove the requirement that companies and nonprofits must justify how holding patents on health inventions made in federal work better serves the public interest. That’s the fundamental change Bayh-Dole introduces in the contracting side of the law. What’s made uniform is not federal patent policy but a uniform restriction on the authority of federal agencies to, as a default, require contractors to explain how their ownership of patent rights (to exclude all others, by default) better serves the public than would open, competitive (and collaborative) access. There’s the use of “uniform” in a proper context. That’s it, that’s the fight, hidden out of view, with various different straw man fights introduced to distract the debate, such as “price controls” and the need for private “investment” and therefore for monopoly pricing to give “incentive” for such use of private capital.

Next sentence from our article:

The only consistent rules were that the government retained title to inventions and would only license inventions nonexclusively.

This too is garble. The government did not “retain” title. The government asserted the right to *acquire* title. The Federal Procurement Regulation, implemented in 1975 based on the Nixon amended executive branch patent policy, expressly required contractors to have patent agreements with research and development employees that would “effectuate” government acquisition unless the government decided otherwise. The inventor would then either assign to the contractor or assign to the government. The FPR didn’t care how it happened. If the inventor didn’t assign to the contractor, then the inventor could petition to hold rights. If the inventor did assign, then it was up to the contractor to petition. Easy peasy.

Let’s take a break and retrace how this pattern of revisions to laws and federal regulations screwed up everything. First, we have the FPR, requiring contractors require employees to either assign to the government upon request or assign to the contractor, who then might have to assign to the government upon request. That’s a standing requirement across all government contracting. Now introduce Bayh-Dole, which is limited to small companies and nonprofits. Bayh-Dole supersedes this standing requirement in the FPR, so that small companies and nonprofits, if they obtain ownership of an invention arising from federal work from their inventor-employeres, may choose to “retain” title to what they have acquired. If they don’t obtain title, then the FPR default applies, and the inventor should have been required by the contractor to agree to assign to the government upon request. It’s just that the agreement specified in Bayh-Dole’s standard patent rights clause (now at 37 CFR 4o1.14(f)(2)) does not stipulate assignment to the contractor of any invention–just subject inventions, and subject inventions are inventions, by definition 35 USC 201(e), that the contractor has acquired. So far, so good, if not a bit hairy. But if your mind is disciplined, you will see the logic at work, and it gets easier, after a fashion. But it’s not all good. The FPR was replaced by the Federal Acquisition Regulation in 1984. The FAR presumes Bayh-Dole applies to all federal contracting and so leaves out the FPR requirement regarding a patent agreement under which inventors must agree to assign to the federal government upon request if they haven’t assigned to the contractor.

When the government obtained rights to an invention, it did not have to try to obtain a patent from itself. It could release the invention, “dedicate” it as the Kennedy patent policy has it. Patenting, then, played a different role, to promote the progress (or dissemination) of the useful arts by publishing the invention together with sufficient detail to permit one with ordinary skill in the art to practice the invention. That’s the policy theory, anyway. The government did not need exclusivity (or to obtain for itself, from itself, the authority to sue citizens for practicing the invention, when providing them with the opportunity to practice the invention was the very thing) nor did the government need to take a money interest, or align itself with the interests of those taking a money interest, nor audit licensees for the money they had made, nor sue others to suppress their practice of inventions, to bolster the commercial positions of those companies favored with government licenses. So the government would not “only license inventions nonexclusively”)–the government did not have to license at all. It could “dedicate.” So much nonsense, so little time.

Now we come to another fake bit of futzed up history:

Federal agencies had collectively obtained over 28,000 patents, but out-licensed less than 5% of that number. In contrast, in cases in which private companies were allowed to retain title, companies successfully secured licenses for about 25% – 30% of those patents.

Yes, the government held those patents. But as Admiral Rickover made clear in his testimony before a Congressional committee, most of those patents were defense-related, there was no “market” for them, and they were made available to everyone mostly without keeping records that required formal licensing. That only a very few were “licensed” has little to do with which ones were practiced. Given that historically only about 5% of the inventions covered by US patents are ever practiced commercially, a 5% figure is not low–it’s reasonably expectable if it were any indication at all about uptake of inventions. Where there were records kept by federal agencies, such as for health inventions, the licensing rates were consistent with the best licensing rates reported by universities (not companies, as our authors have it) for their non-federal inventions (not where companies retained title to federal inventions)–around 25 to 30%, and that’s after the university-affiliated patent managers had rejected 85% to 90% of the inventions submitted to them. The licensing figure–25% or whatever–is in its own way rather useless. Licensing does not equate to practice, and certainly not to public benefit. Licensing just tracks deployment of rights, not the development of products. And what’s licensed exclusively is still not available to anyone else to practice, so until there are actual sales, the value of the patent, and what the licensee pays the nonprofit licensor for, is the right to suppress all other practice of the invention.

The authors here repeat uncritically a propaganda statistic, treating it as a fact and asking us to exercise self-delusion to set up their discussion of Bayh-Dole.

The Act established a new framework by which small businesses and non-profit institutions could retain title to inventions developed under federally funded research programs and could commercialize federally funded inventions, and by which a private business could become an exclusive license[e] for an agency’s invention.

It wasn’t a new framework. It was Latker’s IPA framework. Latker gave public talks in which he said Bayh-Dole was based on the IPA program for nonprofits. The contractors don’t “retain title” in the sense of having any right to take title, or have title vest in them rather than in inventors–see Stanford v Roche–they have a right to keep title they have acquired without making a case for how doing so will better serve the public than would open access. It’s that plan which has dropped out, and with it the purpose of serving the public interest. As for “commercialize”–that’s not what nonprofits generally do. They don’t “commercialize” inventions. They instead put them behind patents and look for licensees that will “commercialize.” But Bayh-Dole does not mandate commercialization–making product for sale, mass production, and the like. Bayh-Dole’s express objective is utilization of inventions–and commercialization is but one way to do so. For instance, use of an invention in a standard, or use in research, or use by professionals without the need for a mass-produced product version. Given that maybe only 1 in 2,000 inventions (Stanford–3 in 6,400; University of California, maybe “1 in 2,000”) held by a university is ever “commercialized” in a manner that has any public significance, that 5% figure for federally owned inventions, undocumented and goofus as it is, starts to look really nice compared to university licensing practice under Bayh-Dole.

But the big point in these sentences is that Bayh-Dole authorizes federal agencies to deal in exclusive licenses, and to take a money interest in those licenses. These are actions that the 1947 AG’s report recommended that federal agencies not do. The conflicts of interest are huge and not manageable. The government plays favorites. The government pits its interest in the success of the license against its mandate to serve the public generally and not suppress a whole industry so that one company might hold rights to (and pay for that right) in a federally owned invention.

The Act has been assessed as “the most inspired piece of legislation to be enacted in America over the past half-century.”

This was not an assessment. It was an anonymous piece labled “opinion.” No data to support the claim. Who wrote it? Latker? Dole? Allen? Here’s the actual quote from The Economist Technology Quarterly:

Possibly the most inspired piece of legislation to be enacted in America over the past half-century was the Bayh-Dole act of 1980.

Love the omission of that “possibly.” The Economist proper ran a story a few years later that walked back the opinion piece–“Bayhing for blood or doling out cash.” And not labeled “opinion.”

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