The workshop began at 1:30 with Professor Carlton presenting a paper that recently appeared in volume 9 of the Journal of Competition Law and Economics. The paper goes into somewhat more detail on the topic of this morning's Heath Lecture, so I won't repeat points made earlier.
First point: the abuse of market power in an SSO. Paper by Richard Gilbert in the Antitrust Law Journal discusses coalitions of customers or suppliers that could harm other customers or entrants; goes beyond horizontal agreements and includes some vertical agreements. Prohibition on price discrimination prevents the abuses that can result from coalitions. It doesn't prohibit the abuse of charging an unreasonable royalty rate; that's the "reasonable" component of FRAND.
As Lemley has noted, virtually no SSO has specified what FRAND means. Uncharted territory--rely on courts to sort it out. Carlton was an expert in the Apple v. Motorola matter over which Judge Posner presided. Finally, Posner threw out so many experts he said there's no litigable case. Much of the testimony is self-serving.
Easy to define "reasonable" in the abstract--ex ante value--but in the SEP context, is it conditional on having royalty rates for all the other SEPs? Is the idea that you are negotiating with one entity that owns all (say) 3000 patents, or with 3000 independent patent holders (which raises the Cournot complements problem)?
The paper emphasizes the nondiscriminatory aspects of FRAND, but what does it mean? Discussion this morning suggested two possibilities. If patent lowers cost by $100 (per unit savings), you can define maximum value you would pay to use patent. But not clear how to account for bargaining power. Still, you can ask if firms are similarly situated by discerning whether the patents lowers their cost by the same amount. Eliminate bargaining strength from the definition. Bottom line, firms in different industries may experience different cost savings, can be charged different rates, can still be nondiscriminatory. If we're in the same industry and you get $100 cost savings and I get none, again patentee could charge you more as long as the determination can be objectively made. If it can't, then we could say that if we use the same component to make our device, we pay the same rate.
Subtleties: more difficult if we're talking not about cost savings but demand attributes. Is the patent added to my smartphone as a result of a patent more valuable than the patent added to your smartphone? Conjoint analysis is sometimes used; an econometrician might try to determine incremental value, but could be hard to do. If this is hard to do, go back to the idea of charging the same rate for a common component.
Dynamic issues can be complicated. When you apply an ex ante negotiation analysis, is the ex ante state the state before the patent was established or the state after? Carlton's sense is that the risks of abuse are so great that he would take away the ability of SEP owners to raise the rate. May not always lead to optimal rate, but in general technology causes patents to lose value over time and he is worried about risks of abuse otherwise.
On validity: if the patent has never been litigated, once validity has been determined, the patent is more valuable. You do not want to allow the owner of an SEP to refuse to license, win on patent validity, and then tell the court you aren't obligated to license at the same rate as you licensed people before. You should let the court issue an opinion on what the FRAND rate is on the assumption that validity is probabilistic and has not yet been determined.
On the issue of patent exhaustion: can be used to prevent price discriminate down the chain. Although economists often are not opposed to price discrimination, here it can lead to holdup, so applying the exhaustion doctrine makes sense. Analogy between SEPs and submarine patents.
Predicts you will see SSOs start to determine FRAND. Also will see NPEs amass portfolios to get around Cournot problem.
Moderator (Professor Elizabeth Rowe): A lot of the issues about what is reasonable/fair are better addressed by SSOs. They don't because engineers don't want to get involved. But she doesn't find that persuasive. Carlton thinks that SSOs are reacting more to advice of antitrust counsel that will be sued if they engage in collective negotiations ex ante, even though DOJ has said the rule of reason would be applied. Also, many of them really are engineers who don't want to be bothered with lawyers and economists at this stage. And perhaps the determination process would be biased in favor of one coalition or another. Carlton finds this less persuasive. And that overall, SSOs are in a better position than are courts to resolve this problem as well as the Cournot complements problem.
Bill Page follow-up: you could extend the BMI precedent to characterize the SSO as an integration of productive capacity, pricing ancillary. (Carlton agrees.) Or SSO could require anyone with SSO to join patent pool. (Carlton agrees and notes Tirole and Rocher paper on patent pools.)
Rowe: On the nondiscriminatory issue, I get you could charge different rates to different industries. But from a layperson's definition, that's discrimination. Carlton: the proper interpretation should be not the economist's definition of price discrimination, but rather definition must be made in light of purpose, to avoid coalitions taking advantage of subgroups. Rowe: couldn't you have holdup issue as to that other industry? Carlton: yes, but the nondiscrimination prong gets rid of the coalition-abuse problem. Rowe: then you're really talking about two different standards? Carlton: yes, should be clearer on this in paper.
Professor Rebecca Haw: if you want to extract marginal value from each user (per Chris Yoo paper, to be presented this afternoon), you don't want to forbid price discrimination. Carlton: whether or not patent owner extracts full value, I want to make sure he can't discrimination as to similarly situated. But if you can't determine easily what similarly situated means, use the alternative rule of charging the same for the use of the patent in the same component.
Chris Yoo: Whether a cost reduction will be passed along to consumers depends on price elasticity. The ex ante value is determined not just by cost reduction but also by pass-through. Carlton: partially agree, but not totally. Simple situation: quantity fixed, I know maximum patent owner could charge. SEP owner and manufacturer. Between manufacturer and consumer, value to society could be different, but I don't see how that is taken into account in ex ante negotiations. Consumer surplus that consumer gets won't be taken into account in the ex ante negotiations, but I don't know what I'd do about it. Yoo: the manufacturer could engage in product differentiation. Carlton: is it price discrimination if I you enjoy a $1 cost savings and I enjoy $0, but I charge you $1.50?
My question: in theory, you could have a court determine the probability of validity and decide a FRAND royalty on that basis, but how do you do that in reality?