We are working through a recent “webinar” on the Bayh-Dole government license to practice and have practiced. In part, the webinar provides the opportunity to set some things right about Bayh-Dole and to resist the machinations of NIST to try to rewrite Bayh-Dole through regulatory shenanigans and therefore somehow preserve a patent monopoly pipeline from federal funding to pharma that feeds a minor parasitic patent licensing industry that has managed to persuade university administrators that gambling with patents is a sure way to a “new source of revenue” to save universities from their financial troubles. The webinar also provides an opportunity to show how the conventional reasoning–and the panel is very conventional–runs against Bayh-Dole and even against their own stated interests in recovering patenting costs and collecting beans of impact.
In a scenario in which the federal government presents as the only purchaser of product made under a patent on a subject invention, a webinar panelist justifies patenting in order to
“demonstrate that we are having impact”
There can be no licensing strategy with regard to the federal government. Or, licensing is not a good way to try to track impact for the federal government. Licensing in such a circumstance involves getting first in the way of impact (patenting) and then getting out of the way as expeditiously as possible (licensing). The absurdity of it all lies in trying to make this process of obstruction and removal of obstruction happen as efficiently as possible, and when it doesn’t, complaining about how complicated and expensive it is and that it is companies’ faults for not helping expeditiously to remove the obstruction. This is what NIST means in its Green Paper by “unleashing innovation.” The aim is to make obstruction and removal of obstruction more efficient. Bizarre? Yes. Another reason NIST should be relieved of its delegation to have anything to do with Bayh-Dole.
Under Bayh-Dole, the federal government gets its license when a contractor, having obtained title (and thereby coming within the scope of Bayh-Dole) elects to retain title. One can then try to game what counts as a subject invention–to avoid disclosure, electing to retain title, and thereby triggering the government license. But the scenario first presented posits a subject invention and that the only user will be the government and then worries about the mechanics of sales to the government and how to recover patenting costs from any other licensee and how to collect impact beans.
If we consider the two paths by which the government may access products based on inventions made in work it has supported, however, much of the complexity disappears. If the government chooses sales, then when it purchases product from a contractor licensee, the contractor gets whatever the license agreement with the seller provides. Because the government has chosen to purchase, it also has agreed that a portion of its payment might go to the contractor in the form of a royalty.
There’s no basis for the government to insist on a discount based on the seller’s obligation to pay a royalty, and there’s generally no good way–given upfront fees, milestone payments, equity, options to buy down royalty rates, and royalty tiers based on sales or time from license–to even calculate such a discount. It would be foolish for the government even to try, especially since the discounts are more like 1% than 5%. If the government chooses the “have made” path, then it can authorize product from whomever it pleases, with or without a license from the contractor, and owes nothing and those doing the “have making” owe nothing to the contractor, under license or otherwise.
If the “have made” path is not worth the 1% savings, then the government won’t bother with it and will purchase product at whatever price it negotiates. Even if it gets its 1% discount claiming its government license (a bogus claim, as we have seen, since the government license involves license rights, not a right to a better purchase price), that does not relieve the licensee/seller from paying a royalty to the holder of the patent on the subject invention. It is absolutely no matter to the government what the backside deal looks like between the seller and the patent holder, so long as the government agrees to a purchase price. The Bayh-Dole government license has nothing to do with purchases.
The question, then, on a broader licensing strategy is then how to engage companies that might sell to the government so that they would prefer to sell than to be authorized by the government to do some “have making.” One thought is to use a royalty-free license. Then the choice between sale and have-making is neutral, and if the university offers assistance in using the invention (and related stuff), then the value in the relationship will recognize assistance and services rather than the patent license (which the company doesn’t rightly need for its federal work–it will just do “have made” work if the university runs up the costs and the bother).
The panel raises an important point–that a company doing work for the government may later decide to do work for and sell to non-governmental customers. For that, the company might need a broader license than “have making.” We need to be clear, however, that the company selling product to the federal government has nothing to do with the federal government’s Bayh-Dole license. It just doesn’t. Having made does. Selling doesn’t. So if a university is going to grant a license to the company, it might make the “field” of sales to the government royalty-free and not trigger a cost accounting on whether to sell or do some “have making.” But more importantly, the university ought to offer the company at least an option to make, use, and sell for non-governmental customers, including its own operations). Bayh-Dole’s objective is to use the patent system to promote utilization of subject inventions. Why then would anyone use patents to create a barrier or disincentive or threat so that a company might not want to use the invention?
The patent system publishes inventions in a national registry. That in itself promotes use of an invention by establishing that the invention is new, useful, and non-obvious and thus something more than just happy hopeful hype. The minimum licensing conditions that go with publication in the patent literature are just what they are for the government license–royalty-free, irrevocable. For the government license, the consideration (to the extent this isn’t an “I’ve altered the deal, pray I don’t alter it further“) is the government allowing the contractor to elect to retain title. That’s the exchange in 35 USC 202(c)(4) and in 37 CFR 401.14(b). The problem in offering a public license of the same sort is establishing consideration for the license that makes the license part of a binding contract. In short, the prospective licensee *wants to pay something* to ensure that the license cannot be canceled at will by the university. So perhaps the university requests $1 (or $500 because it costs more to collect $1 than it is worth for either party to the deal).
To charge anything more than a token amount, however, must be squared with Bayh-Dole for how the added charge promotes utilization of the invention. How exactly does it work to demand a stiff upfront fee or royalties on sales? Perhaps there’s a reason in the grant of an exclusive license (but for the government license), such as calling forth private investment without which the invention cannot be used, but even then one has to grapple with whether an exclusive license is even necessary to promote utilization (and if so, how about maximizing the participation of small businesses and promoting free competition–how do such things work out with an exclusive license?).
There’s a huge difference between offering an exclusive license because no one is willing to take a non-exclusive license for a token amount, and offering an exclusive license because someone is willing to pay to keep the invention out of the hands of other companies who *would* pay a token amount for a non-exclusive license. Again, if someone cannot see this huge difference, or does not want to see it, then they should not be discussing Bayh-Dole licensing strategies. They have moved on, and don’t actually care about Bayh-Dole compliance, even as they talk around the subject.
The objectives stated by Bayh-Dole are utilization, participation, collaboration, free competition, US manufacture, sufficient government rights, minimize administrative costs. There’s nothing there about getting more patents to show the research has been impactful. Anyway, a patent on its own does not show impact. It does not even reflect productivity in research. As the panelist points out, the number of patents is driven by budget not inventions–the lab patents about half the inventions it sees. Absent the patent fetish, one might ask what constitutes a meaningful lab research output worth making a hullabaloo about. As Richard Feyman has it, anyone with any expertise at all can think up ideas that are not generally obvious all day long, and it’s not a matter of productivity.
Similarly, a patent license does not show impact. It does not even show potential for impact. It shows, instead, that someone valued the threat to exclude practice as sufficiently legitimate that it was worth removing the threat via contract. Even if a licensee puts money into “development,” it takes additional work to show that this money becomes available as a result of a patent license and would not be allocated anyway, as in the presence of competitive investments by others.
For instance, if a contractor does not get a patent, companies may well move even more quickly to put money into development, especially if doing so quickly increases the chance of a sale to the government. In such a case, a patent, and the need to negotiate for a license, and to have to pay money for that license before even starting work on “development” all work against adoption of the invention and its impact. Even the prospect that a university or other contractor will hold out for an exclusive license can turn companies away from bothering. Worse if the university gets the idea to start a company and hand the exclusive license to its own startup. Essentially, the university turns aside all existing companies. Why would any of them even bother? More likely: they must work to circumvent the invention, block it, undermine it, talk it down, or invalidate the patent. If the invention is really all that important, companies will be motivated to ask the government to choose the “have made” route and cut past the university and its exclusive licensee.
There are two typical responses to an invention showing up in emerging industry practice. One is to file as many patent applications around that patent to cover anything that the company has–applications, methods, combinations of the patented invention with the company’s own products. The aim is to set up a countersuit or cross-license if the patent holder ever shows up looking for royalties or to try to suppress use. The countersuit/cross-license strategy won’t work against patent trolls (like universities) but will within industry, where it quells the impulse to use patents as competitive bombs rather than to reserve seats at the table to work out standards and interoperability while still maintaining competitive independence.
The second typical response to an invention showing up is to get out in front of development. Get to the juicy applications before anyone else. If the invention is patented and made available non-exclusively (as, say, a standard, or in open innovation, or because some university licensing officer is suddenly and perhaps quite by accident enlightened), then the IP rights that will create an advantage (and get one a seat at the standards/cross-licensing table) will be the ones made in development. So the typical response is to do enough development immediately to justify filing patent applications on improvements–that’s much more likely (and less costly) than attempting to negotiate an exclusive license. In some industries, companies with market power have to take some care in obtaining exclusive licenses–if they don’t work the inventions, that may appear to be a suppression of competition. Thus, the non-exclusive license plus proprietary rights in developments is often preferred.
Any company that cannot set up a defensive cross-license or proprietary IP in developments is left with avoiding the technology, working to keep it out of standards, and generally undermining the perception of its utility. In this way a patent may–especially if licensed exclusively or offered by default for an exclusive license–work immediately against adoption by most companies active in a given industry, Sony Betamax style, and therefore work against the collection of impact beans.
One can track “impact” without taking a patent position, or if taking that position offering a public license without requiring payment. One could, for instance, track requests for assistance, or track downloads of associated data, or track attendees at seminars or workshops that teach the invention to companies. Getting a patent for $10,000 or more to restrict who can have access to an invention is an expensive way to track research impact, while at the same time delaying if not suppressing that impact. Think about it. The issue is not “what can we do with a patent other than promote utilization?” Rather, it is, “what can we do to promote utilization, and if we are going to spend $10,000 to do so, what should we spend it on?”