Patent Fair Use, Commons, and Research Inventions

Madey v. Duke wasn’t a good thing for research. Essentially, it put an end to the idea of a research use exception to patent claims. Whatever the merits of the case, there is no question in my mind that it’s good to have some sort of fair use for research-originated patents. The number one such use is research use, and in particular “evaluation of” and “research on”. I can understand how “research with” could be a more difficult thing, since that means the invention is being used as a systematic tool in research and not just as part of the examination and improvement process concerning the invention. I know, there are issues about what makes something research, and there’s the beauty of unlimited monopoly rights, and there’s even yet further the worries that a fair use statute is somehow a compulsory license rather than a limitation of the statutory monopoly. All these hand wringings can be dealt with. But let’s first get at the rationale why there should be a patent fair use for research inventions, especially those made at universities.

The given purpose of patent law is to promote progress in the useful arts. In exchange for a limited monopoly, the patent owner has the right to exclude others from making, using, and selling the claimed invention. Implicit in patent law is that the inventor will indeed make, use, and sell, or work with one or more people who will do this on the inventor’s behalf. Without activity by the inventor, a patent clearly does nothing to promote progress but rather stands potentially against it.

That is why in some national patent systems, there is a “working” requirement—if the patent owner does not attempt to develop the invention covered by the patent, that patent can be challenged and potentially set aside. In the US, the working requirement shows up only in bits. In general, it is covered by the requirement that the patent owner pay maintenance fees every five years to keep the patent in effect. These fees go up over time. The premise is, a patent owner won’t pay for rights that the owner is not going to exploit. That would be bad business, wouldn’t it? But we also see in this the very wrong assumption that patent owners such as universities are reasonable, diligent folks. (That would be a good statistic–what % of a university’s patent portfolio that is unlicensed does the university pay the 1st, 2nd, or 3rd maintenance fee?) More so, working a patent is difficult compared to sitting on it and then hitting up people who stumble across your rights. That is, assert licensing may be more attractive than trying to develop an invention oneself, or with chosen partners.

A bit of working requirement shows up in the Bayh-Dole march-in provisions. If an invention is made using federal support, then the government has the right to compel licenses if the patent owner or the owner’s licensee is not making the invention available in a timely manner on reasonable terms. The existence of the Bayh-Dole march-in requirements point to a concern on the part of lawmakers that the patent laws have a big hole, and that in terms of research-originated inventions, especially those arising in “basic” research, might be subject to this flip from development to trolling.

There have only been a few requests in 30 years for the government to march-in, and none of those requests have been granted. It’s hard to understand. Is it that patents on research inventions are largely worthless? Or that they are put into play so frequently and so well by universities that there is little reason to complain? (har-har) Or is it that the march-in procedure is so murky and uncertain that there are better alternatives?

My own thought is that it’s mostly that university-held patent rights just don’t matter much. For the most part, research inventions can be designed around, universities do not enforce their rights aggressively, and most of what universities claim by way of patents never means anything to industry anyway. It’s economically dead on arrival. If that’s the case, then it really doesn’t matter that a university places a patent right with a company itself soon to lose interest, become moribund, or attempt, mostly without adequate resources, to develop the invention. As long as the cost to march-in on such a deal is greater than the value, no one will do it. We can take it, then, that there’s not a whole lot of stuff in university patent portfolios that is worth much at all, economically, from the perspective of selling exclusionary rights in exchange for a share of an upside after a lot of risk capital is thrown at the wall.

This is not to say university research for the most part isn’t important, or that inventions, or even patents in such research aren’t important–or valuable. It’s just that these things are not important as primary economic drivers for product development. Their value lies elsewhere. The university emphasis on commercialization of research inventions is largely misplaced. Commercialization happens. It is not something triggered by university administrators going, “I know, I’ll make this happen with patent rights!” What happens maybe once a decade at a larger-sized research institution can’t be made to happen more quickly by accumulating a bunch of patents of dubious direct economic worth and then hiring goodhearted people to bang their heads trying to turn patents into something directly valuable. What they will discover instead is the value of trolling the unsuspecting in industry, who they will have to call thieves rather than developers. As university patent portfolios mature and industry practice follows across these portfolios, there will be increasing temptations by universities to troll industry for patent fees.

I know, this is really difficult. Both approaches involve inventions, patents, and licenses. How can they be so different? One involves developing inventions, using patents and licenses, and the other involves developing patents, using inventions and licenses. Of course, these are day and night obvious to anyone in practice. One is the inventor building something new for the world, and the other is the troll waiting for others to do this, and then nailing them for money when they have the misfortune to surface with some degree of success. University tech transfer as Whac-A-Mole. Anyone who thinks Bayh-Dole was written to create a market built by and for university patent trolls is smoking something.

If we kick back to this idea of progress in the useful arts, there is a way to construe trolls as contributing. You see, if trolls make it dangerous to develop stuff without consulting the patent literature (and even then, there is 18 months of unpublished uncertainty), and raise the incentive to design around, then there is indeed progress of a sort—closer attention to the patent literature, more filings of patents to create defensive positions (but of course this is meaningless because one’s own patents do not defend against trolls, who have no operations of their own), and more research to work around the trolls’ patents–even though no one has bothered to develop the trolls’ underlying inventions! Progress by not trying!

This is a pathetic view of progress. Yet it is what is implicit in university management of research patents on their self-stated commercialization model. It isn’t in the rhetoric of intentions of course—that is all about public good, economic development, and curing the ailments of children. Rather, it is in the actions of patent administrators and the consequences of those actions. The universities don’t license the majority of their patents, and of those they do license, their licensees do not develop the inventions in any meaningful way, and yet universities set up their licensing deals so that the licensees can troll industry as a legitimate way of making money. While universities might create a protocol that limits infringement suits, in general they do not do this. If a licensee were able to bring a suit for infringement but itself had no obviously diligent operations involving the invention, this ought to trigger a cancellation of the license for lack of diligence or an automatic Bayh-Dole march-in.

The benefit of a research patent should be obtained by practice that meets the objectives of patent law and in the case of federal support, federal patent procurement law. That is, research inventions ought to have a stronger working requirement for there to be any enforcement (that is, any patent!), and the benefits of the patent should only be realized when that requirement is met, and to create incentives to stay with the diligence, there should be more research access, if not competition. That is, in research, the bar to patentability is not new, useful, and non-obvious, but rather, all that AND practical application.

One might think of music licensing mechanical rights. If you release a recording of a song, then others can also “cover” it, paying a mechanical license fee per copy set by a royalty tribunal. One might go so far as to say: in a research environment, to maintain a full patent right, a university has to have partners developing *the invention* within a limited period of time or else anyone can “cover” the invention within 36 months of publication of the patent upon notice, for a standard, nominal payment set by a royalty tribunal. Free for internal use. Payment for sales.

The problem with all this is that research patents arise, typically, in settings where the organization that is good with figuring out what to study isn’t so good with having the manufacturing and development and investment resources to make a go at stuff that is invented. Even in industry, it’s a bear to get something from research into product. This is especially true where existing products are successful. How could anything new and untested compete with the present profits and the critical need to defend successful products from competitive threats? So there is a bunch of “spillover.” More so at universities, where nearly everything, with the exception of some stuff that might get used in medical centers or engineering labs, is spillover. It’s just one big mess. It should be a wealth of opportunity, but for the efforts of IP policies and technology transfer offices to grab onto it and try to make sense of it. And this much is true: it isn’t any better for research inventors to hold the IP personally, in the context of their research projects, or in the context of being tied to research organizations making public commitments.

Thus, back to the working requirement. The flip side of the working requirement is that if you are not going to work the research invention, then you ought not get to work the patent rights to the invention. That’s what really gets in the way of research invention. It doesn’t really matter much if one takes out a patent on a research invention, if one then doesn’t do anything to enforce the patent. All the patent does is serve as a broad public marker that something of merit has been done, taught to the world and demarcated as to scope. Good stuff! If it’s not going to be immediately developed—and we already know this is hard for industry, much harder for universities—then it also should not be enforced, at some level! The issue is just how much shouldn’t be enforced.

This leads us back to the idea of a research exception. From research, more research. But that’s also how progress is made in technology. Futzing around until someone makes a break for a market implementation. The challenge of course is not simply to do research but also to have the right to act on it, at some point. So if research team A comes up with an invention and patents it, and research team B has a right to study team A’s invention and comes up with an improvement and patents that, then B can’t develop their improvement without A, and A can’t implement the improvement without B. Before jumping to solutions, multiply this by the US government’s pleasure in distributing work all around, encouraging multi-institutional centers, consortia, and the like. Okay, now we have team A and its patents, and teams B..Z with theirs. How does anyone put this together again, other than waiting out the 20 year term on A, and perhaps on a lot of B…Z? A research exception only gets folks not to fight over research, but it does not get folks doing business with one another over their tiny bits of monopoly…monopoly that prevents anything from happening at scale.

This is all really great for the status quo. I can’t think of anything the status quo could like more than to have transformational, game-changing research tied up in fragmented research patent rights claimed by organizations without a clue as to what ought to be done, held with the determination to recover sunk costs, unable to let anyone else win out. That is, all this research serves not to challenge the status quo, or to lead it, or to obsolesce it, but to strengthen it, entrench it.

So a research exception is good as it strikes a balance between inventors working their inventions and inventors working their patents but not their inventions. There ought to be a standard non-assert for research patents obtained by inventors at research institutions—no litigation based on others doing research. No licensing with rights that allow licensees to bring such litigation based on research uses. Has to be something more, like sales.

And this is just the problem. For research inventions, Sales = Success. It’s the grail of the whole effort. Spillover research technology doesn’t get to sales very readily. In fact, the host organization is generally the worst at getting things to sales. In industry, because of the preoccupation with existing products and their improvements. In universities, because of the preoccupation with getting more research, the desire for strategic partnerships with industry for research and gifts, and the lack of infrastructure to support investment in anything like manufacturing or development. Universities are now all about frequent, sweaty inventive conception and not at all about gestation, delivery, family, or child support. Technology transfer offices are like midwiferies where most everything dies. It’s really very sad. What happens is that the inventions die but the patent rights can be made to live on. That’s what so screwed up in university implementations of Bayh-Dole.

This state of affairs is why efforts to create commons and exchanges and open innovation are so important. These social and economic structures—pre-markets, if you wish—aim to bring back together working the invention and working the patent. Neither abandonment nor trolling. Putting in place a framework for multiple organizations to participate. Moving value from reciprocity (the bilateral license agreement) to externalities (behaviors that collectively scale and contribute to a favorable value position). In this, we see the tyranny of the contract when used as an instrument that has to capture fully the value exchange between two organizations. In an open innovation environment such as open source software, the value exchange is between the organization and a community. The bits of license that contribute code to a community require a minimal of consideration, typically in the form of behavior, but the organization realizes the benefit of others adopting, supporting, and also contributing to the community. That is, the externalities are more valuable than the direct reciprocities.

In this, one can see where a university model committed to using patents on research inventions to create commercial products that pay a royalty can have such a seductive feel, such wonderful instances of success, and yet be so far off base from what is actually happening with 99% of research inventions. If you are only interested in gold you might miss friendship. And even if you are interested in gold, it may be that friendship first will get you to gold later, too.

The starting point then is a gesture to friendship, that is, a research exception for patents, especially for research organizations like universities. Imagine that, introducing friendship into business settings. That is, research universities should offer such an exception as a standard non-assert for their patent portfolios. Second, research inventions that are not worked should within 60 months of disclosure be placed in a patent exchange or commons, where making and using the inventions is subject to a standard, semi-structured license such as the ones Science Commons is working on. These commons should at the minimum permit making and using the invention. Selling, importing, and sublicensing can be dealt with in the same framework, or as something bilateral. But give inventions enough room to be developed, especially if the host institution is not going to be doing that itself, or with a partner that is relatively quickly identified. For the rest—the great majority of research inventions–establishing commons is a necessary means to recover from the patent fragmentation that otherwise haunts research-based inventions. It’s essential to making the connection between research findings and economic vitality for most of the work that’s done in research settings.

This entry was posted in Bayh-Dole, Commons, IP, Technology Transfer. Bookmark the permalink.