Professor Rebecca Haw begins with an overview of the facts of the Microsoft v. Motorola breach of contract case. Motorola wanted a 2.25% royalty; question was whether this was FRAND. Judge Robart noted that a FRAND rate should avoid holdup and royalty stacking. Correct measure, according to Kevin Murphy, is that it should be the ex ante incremental value, but Judge Robart rejects this (sort of) as being impractical. Instead, you conduct a hypothetical bilateral negotiation using a modified version of the Georgia-Pacific factors. (Bill Page interjects: parties would have conducted this negotiation so as to avoid royalty stacking, but this demonstrates that it's not a real negotiation.) He goes awry by setting the FRAND negotiation in the context of the Georgia-Pacific factors. Focus on comparable deals. But by only comparing to FRAND agreements, FRAND is only what parties think FRAND is. He sees certain patent pools are being comparable. What would Motorola have gotten if it had used the pool, and if other SEP patents were all in the pool? Judge Robart then multiplies this by 3.
What is the economically correct definition of a FRAND rate? What Murphy said: the isolated incremental value before the standard was set. This was avoid holdup. Can approximate this by positing a hypothetical negotiation. Legal scholars have pointed to Georgia-Pacific as being a proxy for this. But crucially, it must take into account the next-best noninfringing alternative, the incremental value. Problem is there are so many unknowns: perhaps such factors as bargaining power, ambitions of the two parties, the alternatives for the SSO (what other standards could they have considered, and at what cost).
Haw's criticism of Robart: he sets the time frame at the wrong time. Looking to comparables doesn't necessarily focus on the ex ante incremental value. Could we try to estimate what the patent owner and the SSO itself would have negotiated, had the SSO been a party? This would involve a cost-benefit analysis.
Professor Jeff Harrison: Two things to mention. First is a comment about Judge Posner and the use of hypothetical bargains, which he uses in other contexts (e.g., contract law). Can you use that model and not flesh it out to the full Posner model that considers what the user would have done in the event of other contingencies, etc. Second is: I don't understand why there's a negotiation here at all. We know that the Patent and Copyright Clause has a utilitarian purpose. There's a contract between the public, as represented by the government, and the patent holder. If so, why does the patent owner have any say at all on what the rate should be? Here's an analogy. In tort law, we consider negative externalities. We want the best cost avoider to avoid the accident. In IP, we want the patent owner to create positive externalities. So why pay them any more than is necessary to achieve that? Why turn it over to negotiations between these two private entities? Why not award the patentee a "fair rate"--the long-run reservation price, lowest price it would charge to bring the product to market?
My comment: This reminds me of a point Ted Sichelman makes in his forthcoming paper in the Texas Law Review. But how do we know what the reward should be that would have provided the proper incentive?
Carlton: some have proposed a prize system instead. Or don't give the first inventor exclusive rights.
Harrison: I agree it's impractical, but it reveals a different philosophy of why we should have patents.
Yoo: This is in many ways a product of an antitrust perspective. It's not about long-run marginal cost but about reservation prices. If we allow the SSO to calibrate FRAND, no one has any more or less bargaining power than they would have had without the standard. Antitrust doesn't necessarily ask whether this is good or bad, but IP scholars do.
Haw: This raises the question whether we really need so much patent protection in the IT space.
Carlton: The notion that you have a hypothetical negotiation is meaningless if you have the court influencing the negotiation. The only reason you use hypothetical negotiation is because it takes into account threat points. eBay influences the threat points. Second point: if there is a Cournot complements problem, do you consider the royalty the SEP owner would have sought in light of the existence of the other SEPs? Here's what is more complicated: you could have adopted a completely different standard using different patents. So is the incremental value the value of the next-best standard? That's a very hard problem.