Guest Post by Jeremy Sheff. In view of the pending Monsanto case, I asked Professor Sheff to provide some of his thoughts on the patenting of self-replicating technologies – DC. This essay is cross-posted on PrawfsBlawg.
Self-replicating technologies, once the subject of theory and fantasy, are now upon us. The original self-replicating machine—the living organism—has already been harnessed by biotechnology engineers and, more to the point, their lawyers. The next wave of self-replicating technologies, be they nanomedical robots or organic computers, are not far behind. Rather than triggering a “grey goo” apocalypse, these technologies are, at present, raising far more prosaic issues of intellectual property and antitrust law.
Those issues have now apparently caught the attention of the Supreme Court. A few weeks ago, the Court called for the views of the solicitor general on the certiorari petition in the case of Bowman v. Monsanto. This is the latest in a series of cases in which the Federal Circuit has addressed the application of the doctrine of patent exhaustion to the genetic engineering technology embodied in Monsanto's "Roundup-Ready" herbicide-resistant seeds. Seeds are the prototypical self-replicating technology, and a number of similar herbicide-resistant crops are in the pipeline of the largest agribusiness concerns. In each of the Roundup-Ready cases, a farmer has argued that Monsanto's patent rights do not extend to the second generation of soybeans grown from a patented first-generation seed. In each case, the Federal Circuit found for Monsanto and against the farmers.
Patent exhaustion (or "first sale") doctrine serves as a limit on patent rights, and provides that once a patentee has made an authorized sale of an embodiment of its patented invention, its patent rights with respect to that embodiment are exhausted, and the purchaser is free to use or re-sell the embodiment as it sees fit. Like analogous doctrines in copyright and trademark, it is motivated by competition concerns. Its aim is to enable the creation of downstream or secondary markets in patented articles, and to prevent patentees from using their intellectual property rights to gain market power in markets other than the market for the patented technology. When the Supreme Court last spoke on the issue, it rebuked the Federal Circuit for giving these pro-competitive policies insufficient weight. It seems to be considering an encore in the Roundup-Ready cases. For reasons I'll explain after the jump, I think that would be a mistake.
The Federal Circuit's analysis of patent exhaustion in the Roundup-Ready cases is admittedly not a model of the judicial craft. Framing the issue as a formal question whether a second-generation soybean is a different "article" than the first generation seed from which it grew, the court's main justification for its result was the bare assertion that any alternative result would "eviscerate" Monsanto's patent. But this is a question-begging explanation, and there are other, better reasons why a patentee's sale of a single embodiment of its self-replicating technology ought not to exhaust patent rights with respect to the second, third, or nth generation of the technology that is propagated from that first embodiment. Moreover, these reasons are consistent not only with the reasons for granting patent rights in the first place, but with the pro-competitive principles that justify limiting those rights through exhaustion doctrine.
To get at these reasons, I propose a thought exercise. Let's imagine that the Roundup-Ready cases came out the other way–that purchasers of Roundup-Ready seed from Monsanto were free, as a matter of patent law, to use all subsequent generations of soybeans grown from those first purchased seeds however they saw fit. What would we expect the Monsantos of the world to do? How do we believe their behavior might be influenced by this new legal framework?
One possible answer to this question is: not at all. It may be that the additional revenues to be derived from selling additional embodiments of a self-replicating technology to the same customer are trivial (perhaps due to the structure of demand), and that the prospect of any one customer re-selling a subsequent generation of the technology to another potential customer of the patentee is remote. Nanomedicine, particularly personalized nanomedicine, may one day prove that this is a possible result. But in the agriculture context, it strikes me as unlikely.
Where the technology at issue is an input for the production of a commodity, and the demand for that technology is broad and essentially undifferentiated, I would expect that the possibility of re-sale of nth generation seeds by the patentee's customers would significantly eat into the patentee's revenue stream, potentially making it impossible for the patentee to recoup the investment in research and development required to develop the technology in the first place. This is the classic free-rider problem that patent law is supposed to prevent: we preserve the incentive to engage in costly research and development by giving the inventor a limited-time monopoly. Other scholars have noted that this free-rider rationale is particularly salient for inherently self-disclosing inventions (inventions that are easy to copy once they have been introduced to the public). I would add that self-replication exacerbates the problem of self-disclosure: the patentee selling an embodiment of its invention would not only be teaching competitors how to practice the invention, it would in essence be building their factories as well.
So there are sound justifications grounded in the innovation policies underlying patent law for the Federal Circuit's rulings in the Roundup-Ready cases. But of course, patent exhaustion doctrine is concerned not only with innovation policy, but also with competition policy. This brings me back to my earlier question: how would we expect the Monsantos of the world to react to the free-rider problem if patent law did not protect them against competition from nth generation copies of their own first-generation products? I can imagine two possible strategies a technologist might pursue to circumvent the free-rider problem: contract and secrecy. And I think both of these alternatives are inferior to the patent solution crafted by the Federal Circuit on competition grounds.
Take the contract approach, which has been explicitly advocated by Yee Wah Chin, one of the attorneys representing the interests of Monsanto's farmer customers. To avoid the problem of free-riders Monsanto might, for example, restrict sales of its seeds to customers who sign a license agreement in which the customers undertake to monitor the uses of nth generation embodiments. So, a farmer might have to agree to sell his soybean crop only to buyers who have their own license agreement with Monsanto, or to Monsanto itself. Or Monsanto could include field-of-use restrictions in its licenses, as Ms. Chin proposes: "Monsanto could have licensed seedmakers to sell seed embodying Monsanto technology on condition that the second-generation seed be either consumed or sold to buyers who agree to either consume the seed or isolate that seed from other seed and sell the seed only for consumption."
This does not strike me as a pro-competitive result, for a few reasons. First, it incentivizes Monsanto to extend its influence into downstream markets–such as the market for commodity soybeans and their derivative products–in ways that it would have little incentive for under the Federal Circuit's approach. This downstream market creep is precisely the type of expansion of patent rights that exhaustion doctrine is supposed to prevent, out of fear that the patentee's interests are not likely to be consistent with the efficient functioning of those downstream markets. Second, and perhaps more importantly, forcing Monsanto to look to contract rights to protect its investment in research and development shifts the costs of monitoring and enforcing the Roundup-Ready patents from Monsanto itself onto its customers, who are likely to face higher monitoring costs.
We must remember, Monsanto's customers are largely farmers, who lack Monsanto's economies of scale, its greater expertise with its own technology, and its understanding of the functioning of the markets for that technology. Moreover, shifting enforcement responsibility from the patentee to its customers is likely to create agency costs where they would not otherwise exist. A farmer who is paying Monsanto a premium for Roundup-Ready seeds probably has far weaker incentives to vigorously monitor for violation of Monsanto's license terms than does Monsanto itself, which is reaping the premium. Finally, in the event that a customer breaches these monitoring obligations, either maliciously or negligently, Monsanto's technology could fall into the hands of a competitor who is not in privity of contract with Monsanto and thus (absent any unfair competition type of claim) would be free to use the nth generation seed (in which Monsanto's patent rights are exhausted) to compete with Monsanto. An individual farmer is likely to be judgment-proof in the face of the claims Monsanto might make should such a competitive threat emerge outside the reach of its licensing provisions, which once again leads us to the original problem: how would we expect Monsanto to respond to this risk of free-riding?
This brings me to the last alternative to the Federal Circuit's solution in the Roundup-Ready cases: secrecy. Monsanto might seek to prevent free-riding by refusing to release its technology to public view, and relying on trade secret protection to protect against free-riding. But in order to preserve its secret (a prerequisite of trade secret protection), Monsanto would have to ensure that nothing it released into the market disclosed its genetic technology. As I noted above, self-replication can be seen as a heightened form of self-disclosure, and so this type of secrecy would be fairly hard to maintain. Indeed, I think the only plausible way of doing so would be to pursue a course of comprehensive vertical integration. Monsanto would not only have to be in the business of propagating seeds, but also in the business of cultivating and harvesting soybeans, and processing them into useful products (oil, animal feed, industrial adhesives, tofu, you name it) that do not reveal the genetic material at the core of Monsanto's invention. Even if this were technically possible (a big if), the effect on all sorts of markets, both for inputs and outputs of the soybean market, is likely to be catastrophically anti-competitive. Where the alternative is such drastic shocks to competition in the market for, e.g., miso paste, soy-fed livestock, and arable land, the Federal Circuit's decisions in the Roundup-Ready cases start to look surprisingly pro-competitive.
The big question in my mind, then, is not whether the Federal Circuit's reached the right result in the Roundup-Ready cases. Given the factual setting of those cases, I think the answer to that question is a relatively uncontroversial yes. The real question, to me, is whether the same holds true for self-replicating technologies other than seeds for agricultural commodities. I already noted above one type of self-replicating technology–personalized nanomedicine–that may not present the same incentives for patentees, their customers, and their competitors, as do herbicide-resistant soybeans. Given how little we can presume to know about the future development of other self-replicating technologies, it is likely unwise to try to set a rule today to govern the rights of downstream users for all such technologies that may arise tomorrow. And for this reason alone, it may be worth getting some discussion of the issue from the Supreme Court, which seems particularly sensitive (almost to a fault) to the hazards of establishing brittle legal rules to govern the unknown future of technology. If the analysis that emerges is more substantive and functionally-minded than the under-argued, formalist analysis of the Federal Circuit (admittedly, another big if), I would be happy to see the Court take the case, if only to put the type of issues I've discussed in this post on the table.