We are working through an article by Famiya Masood published March 11, 2020 in The Nation, a Pakistan newspaper. Masood takes up an important issue–how to make Pakistani research supported by the government more productive for things that people are thought to want from it. Masood focuses on what she takes to be the problem that Pakistani research “hardly ever translates into inventions that are eventually patented.”
Perhaps it is difficult to see why the public would care whether research produces inventions or that the inventions are patented. The implicit argument goes something like this: without patents there can be no commercialization, and without commercialization the public does not get the benefit of the research. Buried in this implicit argument is the idea that the primary public benefit of university research should be direct commercialization of the results and not, say, new knowledge that helps anyone, including those with commercialization thoughts, or training students to think clearly about the world and its unknowns using repeatable experiment as one means of doing so, or new working standards so that whatever does get developed works consistently across manufacturers, products, and interfaces. Or that a desirable result of research might be better research (better understandings, better research tools) or clearer choices among existing alternatives (better testing) or results that people can use immediately without having to bother with expensive, high-risk (meaning uncertain, not validated) “development.”
Imagine, for instance, an agricultural research station conducting research. One research team discovers a way to treat salty soils to improve growing conditions using a readily available compound. They teach farmers how to use this compound, and the farmers immediately begin improving their salty soils. No patents are needed, and it just doesn’t matter whether the discovery was inventive or just timely and helpful. Farmers benefit, the public benefits.
A second research team creates a soil filtering machine that requires farmland to be dug up and run through a massive machine that removes excess salts along with beneficial organisms such as worms, which then have to be replenished after the reconditioned soil is re-deposited. This second research team has invented all sorts of things–a process to recondition salty soils, a machine to do it, and new methods of adjusting this new machine to meet different soil conditions. They obtain a number of patents, making certain government officials joyous. Research has become more “productive” because there are now inventions and patents.
But the cost to build the machine is great, and the cost to farmers to use the machine is even greater–they lose the use of their fields for an entire year. So some company has to put up the money to build a prototype machine at scale and rip up some salty farmland to demonstrate that the invention will indeed work. But even then, the economics won’t likely work out. It will cost more to build the commercial prototype than a company can ever expect to recover from farmers. But because government officials believe that inventions are a sign of research productivity, and mass commercial production is the way in which the public must benefit from that “productive” research, the officials push funding to patent licensing offices.
Those licensing offices, in turn, are under great pressure to show commercial licenses and investment in building salty soil clearing machines. Technology transfer is hard, they complain. It’s “high risk, high return.” The rationalize that they have a portfolio of patents, and only one invention out of 100 or even 1,000 is all that is needed for their patent licensing office is a financial success. Maybe, then, they don’t need to license the salty soil clearing machine. Or, if they want to pump up their licensing numbers, they will do what US university licensing offices do–create their own startup for the salty soil clearing machine, and turn the problem into one of getting investors for the company rather than finding an existing, capable company willing to take the risk and build a commercial version of the machine. In this way, the licensing office gets the inventions off their “to do” list and gets to count both a license and a startup as if these are “productivity” and indications of “innovation” and “economic development.” But clearly they are not–they are just making look good an idea that was poorly suited to any use from the outset.
You might see then how one could push for more research to be done with an eye for patenting the results. More research using technology to produce inventions, no matter how complicated, expensive, damaging, or unworkable. Directed at recruiting private agents willing to act in isolation to “commercialize” the research results. Inventors (and perhaps their universities) then ought to be incentivized to change their research interests to be like those of the second agricultural research team, to invent and patent, hoping their organization will then license patents to investors, making all those involved money and making it the problem for those investors to create a profitable product. A virtuous circle, as it were, where investigators now are motivated by the allure of making money beyond their salaries. To do drive this virtuous circle, the patent is presented as crucial. Without a patent, so the argument goes, there can be no monopoly over an invention, and without a monopoly, no one will use the invention, develop the invention for use by others, or mass produce it. On top of promoting the invention of complex, expensive things, this approach also proposes complex, expensive patenting and licensing, and complex, expensive development undertaken by speculative investors because most no one else is willing to make such an effort. It is not so much that such efforts are complex and expensive–it is that they are foolish.
This “virtuous circle” argument is not rooted in practice reality. It is more like a cognitive illusion. In practice, most inventions don’t go anywhere, and patents don’t improve their chances. In practice, people make money from research results all sorts of ways, if making money is their thing. They can consult, hold workshops, write books, get jobs in industry, serve on boards, start companies–and for any of these things, a patent might help a little bit or might be a huge bother.
As one computer science faculty member put it to me–I paraphrase–“I can announce a workshop in my area of specialty any time I want and get 50 people to attend, paying $2,000 a piece. I can clear $80,000 a workshop for a few weeks of preparation and two days of work. Now, how soon can you get me $80,000 from licensing a patent on any invention I may make? And how likely?” I do a quick calculation–$80K as a 30% share of net royalties after patenting expenses is like maybe $250K of university licensing income which represents maybe 2% of net sales of a licensed product or $12.5M. So maybe five years, with one chance in 100.
You can see the problem. She could get that $80K in three months, with a 90% certainty. Or wait around for it in five years with a 1% certainty. Patent licensing income appeals primarily to those who like to roll the dice and are prepared to wait a long time to see if there’s any result. Patent licensing income is not an immediate incentive for any research, because there rarely is any immediate patent licensing income, and there is often no patent licensing income at all, ever.
Patent licensing income may be part of a virtuous circle, but it is only so for university administrators, who can distribute their risk of return over a portfolio of inventions. Only one invention every 20 years making lots of money is sufficient to drive the model financially. In the US, that’s like 1 invention in 1,000 at elite universities, and 1 lucky invention in a lifetime per 50 other universities. The inventors’ “incentives” in a portfolio model end up “helping the university look good” or “getting a patent on my c.v.” rather than making money from commercialization. While in theory university inventors may change their research to make money from patents, in practice most don’t. And that would appear to be a good thing, especially for basic research. Otherwise, researchers tend to run toward wherever there was obvious money to be made with patents, and that would greatly reduce the range of research. It would be like kids playing soccer–all rushing to the ball rather than working their positions on the field. Worse, money can be made with patents all sorts of ways that do not involve creating beneficial products for public use.
We have looked at the problem of using patents to block research and internal uses of inventions to make a market for mass production look better, in the hopes of landing a company as an exclusive licensee.
There’s a second problem with the general case of patenting publicly funded research–distinguishing commercialization investment from patent exploitation. An exclusive patent license does not imply commercialization and public benefit. An exclusive patent license moves rights to exclude from one entity to another. The receiving entity–a company, the exclusive licensee, the assignee of the invention–may have any number of motives to acquire an invention. One motive might be to invest and mass produce. But others, equally possible, include acquiring an invention to prevent its use by competitors, to control how the invention is used relative to the company’s own products, investments, and directions, to use to trade via cross-license with competitors to gain access to technology otherwise not readily available, or to hold and develop or divest later, depending on how other research and markets play out.
Or, a company may acquire purely to flip the patent for profit. Buy low, sell less low. That’s a basic approach behind some university startups. Acquire a patent from a university at a low cost–small upfront, a royalty on future sales–do some follow-on development, and then sell the company (with the patent license) to another investor or a big company. The value to the speculator lies in the equity value of the startup, not necessarily the future value of the patent license. For the price of the patenting and some risk money to create a prototype or get some further data, a speculator can flip the company for 100x profit. That’s a pretty good living, and I’m not knocking it, but that is not at all the same as working diligently to mass produce a beneficial product in less time and at a more favorable price than if things were left on their own without government incentives for university administrators to patent–or even for university researchers to patent.
Put another way, a university startup may be part of a pyramid scheme–do just enough work to flip the company and its liabilities to someone who wants it more and is willing to pay to get it. Again, it’s a good living, but is it the sort of thing that deserves a national law, government subsidies, and the attention of university administrators. “We don’t have as many pyramid schemes exploiting university research inventions as do the US and India!” Is that any way to go about the public policy discussion? No, not hardly.
When patent-based startup companies fail, their investors treat the patent rights as assets to “monetize.” That is, failed startups tend toward patent trolling–another form of speculation. Rather than grant non-exclusive licenses up front, when an invention has been made and is prime for early adoption by those who are technologically capable, universities exclusively license to startups, those startups fail, and in failing they then aim to sue companies that have used some aspect of the claimed invention anyway. The startup or its investors cry “foul” and seek compensation for infringement, but that infringement plus compensation ends up being another form of non-exclusive license, just now after the fact, catching companies up and using a patent position to make them pay. Now, if public policy is that domestic companies must pay to use inventions made with public support, and patent licensing is just the thing rather than, say, corporate taxes or not paying at all (beneficial use being the goal rather than paying what amounts to a tax on adopting something new from public research), then patent trolling ends up being way more expensive than just granting non-exclusive licenses on fair, reasonable, and non-discriminatory terms.
One exclusive license might be worth two hundred non-exclusive licenses. But getting two hundred non-exclusive licenses for 1/200th the cost of an exclusive license may be way, way easier to accomplish. If we are going to be general about it, then at least let’s do the general math. If an exclusive license is valued at $500M and the likelihood of securing it is 1 in 1,000 (for a net present value of $500K), then two hundred non-exclusive licenses at $100K each over a ten-year life, or $10K per year with a likelihood of 1 in 5 (for a net present value of $4M) ends up being a much better deal.
It’s not the value of any one transaction that matters, but rather the odds that any given transaction will get done and will provide the expected benefit. It’s not even that a license gets done, but that the full expected value of the licensed invention will be realized. That there will be, in the case of an exclusive license, successful development and mass production, that the company won’t fail or the invention won’t fail or be replaced by a non-licensed work-around or that the company will pay up and not seek to renegotiate the deal or sue for patent invalidity or any other black swan that could wreck the deal. By contrast, non-exclusive licenses are relatively easy to deal, don’t demand much money, and so aren’t likely to motivate work-arounds or disputes over diligence, accounting, and the like.
When the cost of a liberal license is less than the cost of lawyers to fuss over it, then it is likely there won’t be fusses that require lawyers. That’s almost always not the case with exclusive licenses, and thus by basing public use of research conducted with public support at universities on the prospect of mass producing some new product using a patent to exclude all others dramatically–maybe fatally–raises the risk of failure as it raises the appearance of great value. In doing so, this approach attracts speculators who feed on that risk while putting off those who would use or develop but without an interest in mass production or without any need to exclude others or without the means to pay an exorbitant price to buy up the entire patent monopoly. That’s some public policy purpose. I suppose one could go there. I just don’t see any *good* reason to do so.