ABSTRACT

Mark Lemley and Carl Shapiro propose that standard-setting organizations (SSOs) mandate that their members henceforth submit to binding, final-offer arbitration (commonly called "baseball arbitration") to set fair, reasonable, and nondiscriminatory (FRAND) royalties in licensing disputes concerning standard-essential patents (SEPs). SSOs should reject this proposal. It does not rest on sufficient facts or data, nor does it apply intellectually rigorous principles and methods of law and economics in a reliable manner. This is not to say that the voluntary use of arbitration to resolve FRAND licensing disputes is inherently problematic. However, the incremental efficiency that Lemley and Shapiro claim that their proposal would achieve over litigation or conventional commercial arbitration is illusory. For one, it is much harder to value a portfolio of SEPs over the span of five years than to value an individual baseball player for a single season. The Lemley-Shapiro version of mandatory baseball arbitration would not shed light on the question of what constitutes a FRAND offer. To the contrary, Lemley-Shapiro arbitration by design collapses questions of validity, infringement, and essentiality of the patent to the standard into a single damage calculation in which the arbitrator's sole responsibility is to choose one of two disparate estimates of reasonable royalties. Yet, a FRAND offer contains not only a price, but also terms and conditions that (because they are nuanced and possibly tailored to the unique needs of an individual licensee) do not lend themselves to being easily standardized, let alone summarized in a single number, as the description of Lemley-Shapiro arbitration might incorrectly lead some to assume. Lemley-Shapiro arbitration would not say whether a royalty offer was fair, reasonable, and nondiscriminatory. Lemley and Shapiro claim that their arbitration proposal offers "best practices" for SSOs. That label is unsupported and misleading. The package that Lemley and Shapiro call "best practices" is in fact not a narrow proposal for binding baseball arbitration but rather a roadmap to redefine patent rights in a manner that would transfer wealth from inventors to infringers. Embedded within Lemley-Shapiro arbitration are normative changes in patent law and policy that Lemley and Shapiro have previously advocated but that SSOs and courts have not adopted. An SSO that adopted Lemley-Shapiro arbitration could expect its members to commercialize their next generation of inventions outside that particular SSO, if not outside an open standard altogether.

I. INTRODUCTION

Final-offer arbitration, commonly called "baseball arbitration" because of its use in Major League Baseball disputes over player salaries, requires an arbitrator to pick exclusively one of the two offers made by the opposing parties to a negotiation.1 Mark Lemley and Carl Shapiro have proposed binding baseball arbitration as a mandatory procedure to determine royalties in disputes over the licensing of standard-essential patents (SEPs) on fair, reasonable, and nondiscriminatory (FRAND) terms and conditions. Specifically, Lemley and Shapiro urge standard-setting organizations (SSOs) to adopt rules- purportedly constituting "best practices"-that would mandate binding baseball arbitration for resolving FRAND disputes.2 Presumably, an SSO would need to adopt these "best practices" by voluntarily amending its bylaws.

FRAND disputes are important because manufacturers of standard-compliant products may need to procure licenses to patented technologies that are incorporated into the standard. To facilitate this process, prospective licensors (and frequently prospective licensees) participate in an SSO, which develops "agreements containing technical specifications or other criteria," promotes "efficient resource allocation and production by facilitating interoperability among complementary products," and, in general, participates in the advancement of the standard and associated technology within an industry.3

An SSO typically requires one of its members to disclose or declare any patent that the member believes is potentially essential to a proposed standard. A patent that claims an invention that is necessary to practice a technical standard is a standard-essential patent. The declarant agrees to offer to license its SEPs to third parties on FRAND terms. Scholars in law and economics actively debate the meaning of "fair, reasonable, and nondiscriminatory" licensing terms.4

Lemley and Shapiro propose that, when an SEP holder and an implementer cannot agree "over what is FRAND,"5 the parties shall enter into binding baseball arbitration.6 Each party must present its final offer as its first offer, leaving the arbitrator the limited discretion to award one of the two parties' offers as the final binding FRAND royalty. In conventional arbitration, the arbitrator receives the parties' final offers and then determines the award based on the arbitrator's own judgment and independent evaluation of the parties' arguments and offers of proof.7 In determining the FRAND royalty, the arbitrator ultimately may set a royalty that differs from both parties' proposals. In contrast, under baseball arbitration, the arbitrator is limited to choosing between the respective royalties proposed by the SEP holder and the implementer.

Lemley and Shapiro champion mandatory baseball arbitration as an alternative to traditional district court litigation of SEP-licensing disputes. Such arbitration, they argue, may induce the parties to produce offers and counteroffers closer to the patent's "true" value, may reduce transaction and error costs of dispute resolution, and may provide greater incentives for the parties to settle.8 I disagree. Arbitration is not inherently problematic as a means to resolve FRAND licensing disputes. Indeed, in certain circumstances, parties to a licensing negotiation may find that arbitration offers distinct advantages over traditional litigation. However, Lemley-Shapiro arbitration is ill-suited to resolve disputes over FRAND royalties for SEPs. The proposal would address neither the objectives of setting open standards nor what makes FRAND commitments distinctive, and it would prevent the parties from agreeing on the meaning of FRAND prices, terms, and conditions.

I analyze the Lemley-Shapiro proposal, as well as the critique of it by Pierre Larouche, Jorge Padilla, and Richard Taffet.9 Baseball arbitration may work to resolve a dispute over a player's salary in Major League Baseball, but it is ill suited to resolve a dispute over FRAND terms for standard-essential patents for complex consumer products such as smartphones. Moreover, the errors of Lemley-Shapiro arbitration would not be randomly distributed. Larouche, Padilla, and Taffet show analytically what should be intuitively obvious and compelling-that baseball arbitration as envisioned by Lemley and Shapiro would consistently undercompensate SEP holders, a result that would destabilize the standardization process.10

In addition, Lemley and Shapiro prescribe rigid "best practice" rules for SSOs to adopt in lieu of relying on the more flexible and reasoned standards that courts would develop incrementally, case by case. The Lemley-Shapiro prescriptions proceed from the false premise that one can identify-and then should pledge obedience to-"best practices" to address all foreseeable contingencies. Assuming (unrealistically) that complete-contingency "best practices" are possible to identify for a technologically dynamic industry in which many firms cooperate to create an open standard, who first determines that a given practice is "best?" Lemley and Shapiro answer that question by declaring their own prior prescriptions to be "best practices," regardless of whether anyone has actually adopted those practices and regardless of whether they are biased toward the infringer. Analysis of those "Lemley-Shapiro best practices" reveals that they are in fact riddled with errors of legal and economic reasoning. FRAND disputes cannot be encapsulated in one arbitrary number, as Lemley and Shapiro claim. In addition, the Lemley-Shapiro proposal conflates the SEP holder's duty to make a FRAND offer and the duty to enter into a license. Lemley-Shapiro arbitration also ignores the question of whether an SEP is truly essential, which is determinative for whether the FRAND obligations apply. Lemley and Shapiro claim that their arbitration procedure is free from bias, yet they provide no evidence supporting this claim. Moreover, the hypothetical negotiation framework used in Lemley-Shapiro arbitration is one-sidedly biased against the SEP holder.

An additional factor that would reduce compensation to SEP holders is the requirement in Lemley-Shapiro arbitration that the arbitrator's chosen royalty be based on an ex ante probability of validity of the SEP holder's entire portfolio.11 Lemley and Shapiro further argue that a court's subsequent confirmation that the patent in suit is actually valid should not change the FRAND royalty determination, which had been predicated on the Lemley-Shapiro assumption that the patent in suit was only possibly valid. 12 The Lemley-Shapiro notion of probabilistic patent validity is problematic. Lemley and Shapiro fail to explain how an arbitrator could rigorously assess the probabilistic value of the licensed patents in choosing which of two proposed rates is FRAND-a factual determination that typically requires an entire court case. Evaluating the probabilistic validity of an entire patent portfolio would require the patent-by-patent review of hundreds or thousands of patents, which renders Lemley-Shapiro arbitration infeasible in the real world.13

A given arbitral procedure can affect the incentives and bargaining positions of the parties to the dispute. Key factors affecting the parties' bargaining strategies include the degree of information revealed during the arbitration and the binding nature of the arbitration. The restrictions of Lemley-Shapiro arbitration may bias royalty awards in favor of net implementers and thereby reduce expected returns to SEP holders. Any unintentional cognitive bias among arbitrators that would favor smaller royalty payments would systematically undercompensate SEP holders. This bias would harm innovation and investment in new SEPs. Over time, Lemley-Shapiro arbitration would cause SEP holders to reduce or withhold participation in SSOs and would reduce incentives for innovation.

What is the implication of Lemley-Shapiro arbitration for the public's understanding of the meaning of FRAND? Lemley-Shapiro arbitration would not inform the current debate over what constitutes fair, reasonable, and nondiscriminatory terms for SEPs. Because the arbitrator would be constrained to pick one of the parties' two opposing offers and could not determine an intermediate FRAND rate of her own, the final award could not result from the arbitrator's informed assessment of what constitutes a FRAND royalty. The arbitrator would be constrained to choose one of the parties' two offers, even if neither offer was in the FRAND range. The additional Lemley-Shapiro requirement that arbitration decisions be disclosed to willing licensees in future negotiations14 would not provide useful benchmarks for future negotiations. To the contrary, doing so would undermine future negotiations, reduce the incentives of SEP holders to participate in SSOs, and promote buyer collusion among implementers.15

The Lemley-Shapiro proposal claims to resolve FRAND royalty disputes more efficiently than court litigation. However, the proposal contains errors of legal and economic reasoning that tend to reduce compensation to SEP holders. Lemley-Shapiro arbitration would not provide guidance on what constitutes a FRAND royalty, and in the long run it would reduce incentives for innovation and participation in the setting of open standards.

II. WOULD LEMLEY-SHAPIRO ARBITRATION BE PLAUSIBLE AND EFFICACIOUS?

Read entire article (PDF).