August 25, 2004  ·  Richard Posner

Doug Lichtman, a very able IP professor at the University of Chicago Law School, took sharp issue with my brief note on patent fair use, emailing me that my “quick reference to patent fair use…is problematic for the simple reason that, often, the key market for research tools is to sell those tools to other researchers. If a researcher’s use of patented research tool is fair use, that would significantly degrade the incentive to create those research tools inthe first place. Moreover, even if your approach works, it is in sharp conflict with the Bayh-Dole instinct that society might very well be better off in a world where academic researchers patent their work. As you know, that legislation was passed in response to evidence that university breakthroughts were sitting on the shelves both because (a) they could not be owned exclusively under old NIH rules; and (b) universities had too little incentive to bring their work to the attention of industry. Overall, patent fair use and the research exception are an important topic, but your short sentence seems to unfairly duck the many hard issues.”

These are difficult issues, to which I can’t do full justice here. Lichtman and I differ on the importance of patents as motivators of research. The effects of patents on innovation are extremely complex, an important consideration being that when a field becomes blanketed by patents, as is happening with research tools, inventors are forced into what can be costly and protracted negotiations for licenses in order to be able to use and build on previous innovations. So we have to consider carefully what alternatives there are to patents for motivating innovation in pharmaceutical and other research. It turns out that there are many alternatives, including government grants, university grants (universities have their own resources–Harvard has an endowment of $20 billion), the commercial advantages of a headstart, and trademarks.

And are we really better off in a world in which academic researchers can patent their work? Maybe so, but a countervailing factor is that the patentability of academic research deflects academic researchers from basic to applied research, which may have long-run consequences for innovation that are adverse.

  • Polk Wagner

    And are we really better off in a world in which academic researchers can patent their work? Maybe so, but a countervailing factor is that the patentability of academic research deflects academic researchers from basic to applied research, which may have long-run consequences for innovation that are adverse.

    But might your countervalining factor actually point the other way? Clearly, allowing/encouraging patentability of academic research is going to tilt the mix of basic versus applied research in the direction of applied (unless, of course, one thinks that even basic research should be generally patentable, though this is beyond the scope of the topic here).

    This finger on the scale in favor of commercializable inventions might be quite apt — many of the noneconomic incentives for researchers are far stronger in the basic research area (e.g., this is what Nobel prizes are given for). The case for a government role in subsidizing innovation is also much stronger in basic research: while the government arguably has no place in trying to guess the commercial potential of research, it may well be able to invest more wisely for the long-term, without requiring any real expectation of realizable profits. Thus we should expect (and demand) that government subsidies go largely to basic research.

    As you note, these issues are complex, and rife with empirical conjecture. But it certainly seems to me that we want academic research to be a mix of basic and applied, and that allowing patenting (and, to the point of your original post, keeping the rights associated with those patents strong) for many academic inventions can help maintain the balance.

  • Matthew Saroff

    It should be noted that until about 15 years ago, neither software algorithms nor business plans were patentable, and plenty of innovation happened anyway.

    These days, it seems like the patent office will allow you to patent a sunny day.

    I think that legislation specifically removing algorithms and business plans form patent protection would be a good thing ™.

  • Doug Hudson

    The Federal Circuit’s opinion in Madey v. Duke affirms what many in the IP bar already suspected– that there are substantial IP barriers to new entrants in basic research.

    Now any research institution worth its salt must work hard to obtain a reasonable patent portfolio, so that it can cross-license the technologies it needs to do further research. It essentially cartelizes basic research the way private technologies have formed IP cartels in the last few decades.

    Whether this is a good thing (new incentives for investment in developing new research techniques) or a bad thing (restrictions on basic science resulting in retardation of progress or mass export of basic research labs) is up in the air.

    This raises a loosely related question, which is the benefits and drawbacks of the specialization of federal judges as in the Federal Circuit…

  • David B. Woycechowsky

    Speaking as a patent lawyer, I have become very much for field-technology-neutrality in patent law. Any other approach opens the door to lots of rent-seekin’, IMO.

  • Matthew Saroff

    What do you mean by “field-technology-neutrality in patent law”?

    It’s my understanding that until recently, algorityms and business plans were not treated as a process, but something rather more like a recipe, and hence not subject to patent.

    The first software patent that I heard of was for Quarterdeck around 1989.

    Before that they were denied.

    I think that it’s an artifact of establishing a separate patent court.

    Since that is all they get, it’s that situation where if all you have is a hammer, everything looks like a nail.

  • Joe O

    Field-technology-neutrality is not necessarily a good idea for patents. Patents are important incentives for some fields like drug development. It’s much easier to copy a valuable drug than to design a valuable drug. Without patents the incentives for new drug development are severly reduced. In other fields, such as computer software (which is also covered by copyright), patents may reduce net productivity. Idealy, the strength of patent system protection for a given field should be selected based on the economics.

  • David B. Woycechowsky

    By field-of-technology neutrality I mean that the substance of the technology, be it electronics, mechanics, physics, chemistry, product, process, matter field, energy field, 17th dimension string field, should not matter (so to speak). The ultimate touchstone should be whether anybody was going to give us the wisdom within a year or five, assuming the “inventor” had never been born. If somebody was, no monopoly. If not, let the monopoly range across the “commensaurate” (see In re Fisher) scope of the magical, early-disclosed knowledge.

    This should be the touchstone of obviousness. Reading old Posner opinion from 1983, you can already see this disbelief that anybody can patent a sunny day “these days.” A more meaningful test of obviousness would easily cut accross these debates about whether software patents versus surgery patents versus business method patents versus patents on barbed wire hurt or help in the economic sense.

  • Doug Hudson

    On one hand, field specificity (such as patent term extensions for FDA approvals) may be necessary where a specific arena as a high investement barrier or regulatory barrier to innovation, and some statutory provision is needed to balance the playing field.

    On the other hand, field specific statutes have a history of non-use or non-effectiveness. See the boat hull protection act, the semiconductor protection act, the second plant protection act, specific provisions under title 35 that apply to DNA or biotechnology issues, etc.

    There was a paper by some MIT students some time ago (available on the internet somewhere or other) showing how patents properly compensate for innovation investment in some industries, but in other industries, such as software, patents tend to create a “free ride” effect under some circumstances. There may be no right answer in the short term, but field-specific legislation may not be the best answer.

  • Branko Collin

    “The ultimate touchstone should be whether anybody was going to give us the wisdom within a year or five, assuming the ‘inventor’ had never been born.”

    That’s an interesting notion. That could mean that high-demand inventions, such as drugs against AIDS, could not be patented, because somebody would come up with them, regardless the cost or the inventiveness it takes to think them up.

    On the other hand, things that nobody would reasonably need, and therefore not readily think of, such as a new toy, could be patented quietly and worded broadly, and then be used as a litigation hammer.

  • David B. Woycechowsky

    I am serious aboiut what I said. Under my vision of obviousness law, patent law would work like a self-centering screw. If someone is really going to give us, without monopoly rights, an AIDS drug independently and within a year of someone enjoying a 20 year monopoly, then that is better for AIDS sufferers.

    Of course, the key is showing that this-or-that advance was really forthcoming. But in the case that the advance really were independently forthcoming, then I don’t see why monopoly rights amount to more than “rent.”

    If that shifts duplicative research fron AIDS drugs to toys, then let the good times roll!

  • David B. Woycechowsky

    Of course, once all the AIDS researchers re-allocate and pursue toy design: (1) toy design will become much less susceptible to patent; and (2) AIDS reseach will become much more susceptible to patent. That is the gist of my self-centering screw analogy. Thought it might not be clear from my previous post/

  • Branko Collin

    “I am serious about what I said.”

    I did not claim otherwise. :-)

    Your idea is so new to me (would you be able to patent it?), that I wanted to make sure I understood you correctly.

  • David B. Woycechowsky

    I wanted to write a book about it, but during the year I was unemployed I was too depressed to write. Hopefuly now that I am working again, I can expalin at length.

  • Tracy

    I would like to point out an important difference between grants and patents. Grants pay out for research done, while patents actually cost money to get, and the monetary benefits come from actually selling the product to end-users. Thus patenting, on first principles, is a better tool than grants for actually ensuring the research is used and does not just get published and sit in a journal on university shelves.

    This has been a major concern in NZ science policy – that NZ scientists are great at basic science, but not at moving them through to market, and that that is probably related to that traditionally scientists were just paid for doing the research.

    There are other methods of transferring science, e.g. Coase’s Theorem didn’t need patenting to really transform economists’ ideas about property rights. And model farms have been a good way in NZ of transferring knowledge about farming research to farmers. And patent-owners can, in at least some circumstances, gain money more from suing infringers than from selling the goods themselves. But patents and grants are not perfect substitutes, they place different incentives to researchers.

  • David B. Woycechowsky

    The Coase Theorem is an interesting example to ponder, but how does the history of the academic development of this theorem really cut?

    My torts professor made a point of telling us that the theorem (1959) was slow, not fast to catch on. If this is true, maybe the slowness was the result of Coase’s employer perceiving itself as unable to stake out a Kitchian patent claim. Maybe if Coase’s employers had a money incentive, it wouldn’t have taken the theorem 14 years to show up in a legal text.

    Of course, this stakeholder argument is often given to support the patent system in other contexts. Why would that not be valid in the market for books and articles? Transaction costs? Why would an author’s employers incur a greater helping of transaction costs than any other manufacturer or service provider?

    I think academic publishing is an area where the traditional field-of-technology-non-neutrality has made us miss an important opportunity to optimize the patent laws as a whole. Of cousre, now that State Street has called the “printed matter” exception into question in the US, we may be getting this particular opportunity back.

  • David B. Woycechowsky

    Further to the previous: It is also worth noting that Coase’s two most famous articles relating to transaction costs where spaced about 20 years apart. Maybe patent rights would have provided the right kind of incentives ($$) to speed up the man himself.

  • Greg Wolff

    Isn’t the underlying question how to enhance the exchange value of research tools rather than patent policy per se?

    Doug Lichtman rightly points out that incentives should be provided for developers to make their work available. However, the underlying assumption that patents will provide that incentive in our current economic system appears to be unsupported. At both the individual and institutional level, the practice of patenting often increases the cost and overhead of exchange thereby actually work against the intended outcome.

    At the macro level, approximately 1 in 100 patents have economic value of any sort (e.g. used in licensing or litigation). After the famous examples of lucrative wins at a few top research universities (PCR at Berkeley) many universities try to emulate the success by setting up their own “office of technology liaison” (they go by various names) to handle IP issues including convincing researchers to patent their work (with rights assigned to the University) as well as licensing IP assets to commercial entities. The last numbers I saw indicate that very few if any of these offices actually generate enough licensing revenue to cover the costs of their (legal) staff.

    Worse yet, at the micro level these offices create policies which expressly prevent researchers from sharing their (unpatented) work with others in order to protect the patent rights of the University. For an associate professor looking to get tenure, filing a patent looks like a barrier rather than an incentive. A costly, unfamiliar, and time-consuming process, filing for a patent just detracts from their objectives (publishing) with highly unlikely and uncertain rewards. Furthermore IP rights become real sticking points in collaborative work across institutions. In this environment, most research tools do sit on the shelf or at best follow the graduate students to their next institution without too much concern for IP issues.

    Yes the issue of providing appropriate incentives is too complex to handle here. The answers probably lie more in the realm of markets that radically reduce transaction costs and other barriers to exchange rather than discussions around the current patent system.

  • David B. Woycechowsky

    Dear Mr. Wolff:

    Of course, your arguments about the ways the current patent system provides a barrier for researchers are powerful and may well be true.

    My point is that these same arguments apply with equal force against all developers of all products and all processes with the same force.

    By having academic types, who tend to be articulate and self-aware, exempted from the patent system in their own careers, we fail to really appreciate the damage the patent system is doing to the less articulate and policy-driven.

    The whole patent system needs fixing (I think primarily by a major revamp of obviousness law –perhaps by being eliminated altogether — perhaps creative ways of reducing transaction costs). If Prof. Posner had had to negotiate licensing fees back in 1973 to use the Coase theorem in his commercial product (a textbook), then you better betcha boots the patent system would be in much better shape (or perhaps gone) now. He would have made it so. He is a very capable man.

    Alternatively, maybe Coase’s licensor would have commissioned the book (in substance) back in 1963. That is how the current patent system would have you believe things work — maybe that hypothesis is valid or can be made valid with some systemic tweaks.

    Thought is thought, accelerating thought is accelerating thought. Field of technology distinctions act as if there is no Newtonian-like universal law of intellectual inertia — as if eggheads like torts professors respond to ecomic incentives and transaction costs differently than the nameless drones who create new designs for washing up fluid bottles. Worse yet, distortions caused by uneven application of the patent system obscure any universals and thereby prevent the patent law from developing in ways that give all people who live by their wits the push on those wits that Jefferson insisted it should.

  • David B. Woycechowsky

    In my previous post I made some cryptic references to tweaking the patent system and especially its obviousness law. For those interested, I have provided a two page summary of a specific proposal at:

    Hopefully, this document at least demonstrates that there are other options besides either the current patent system or no patent system at all.