Royalty stacking in FRAND litigation – part two

The first part of this article gave the economic definition of the ‘royalty stacking problem’ (for further information, please see ‘Royalty stacking in FRAND litigation – part one’). This second part discusses the reasons why this issue is difficult, if not impossible, to prove or remedy.

The previous article discussed two hypothetical licensors, each of whom charges a 5% royalty on $1,000 in sales and might coordinate their pricing to increase their royalties by reducing the licensee’s costs, charging 9% on $1,200 of sales – the joint optimum. The inability of uncoordinated licensors to achieve this coordinated outcome constitutes the true royalty stacking problem.

In addition to the necessary supply and demand conditions, this result depends on the licensed inputs being economic complements, rather than substitutes. SEPs are economic complements as their use is required by definition. In the much more common case where inputs are substitutes, coordination (ie, horizontal price-fixing, which is illegal) implies raising the joint price. For this reason, antitrust agencies will approve horizontal price coordination (eg, patent pools) only when the pool members show that their patents are complements (it is difficult for individual owners of “SEP portfolios” to show this, because such portfolios typically comprise patents that may not be actually essential and that therefore may be substitutes for one another).

Coordination among licensors is also contrary to the practice envisioned by standard-development organisations (SDOs). For example, the European Telecommunications Standards Institute (ETSI) states that royalty terms and conditions should be established as the result of “friendly bilateral negotiations” (“Interim Report of the UMTS IPR Working Group”, September 1998; see ETSI’s IP rights policy (Annex 6 of the ETSI Rules of Procedure)). Two parties who cannot agree have resort to a national court, whose decision is binding on them alone (perhaps within that jurisdiction only). But if SDO policy specifies that royalty terms and conditions can be set for individual licensor-licensee pairs only, taking other contracts as given, the best possible outcome of that policy is an uncoordinated one. Nothing in it contemplates a systemic inquiry into what might constitute a theoretically superior set of royalty terms and conditions across the entire industry.

Given the typical focus on innovators’ alleged failures to uphold their FRAND commitments, it is easy to overlook the requirement that this theoretically superior set of royalty terms necessarily encompasses the mandatory payment of the optimal royalty stack by all implementers. In other words, coordination of royalty rates among licensors must imply coordination of royalty payments from licensees to licensors. Since holdout by implementers (ie, the failure to pay royalties on time, or at all) is widespread, a regime that enforced an optimal royalty stack would increase the total royalty payments made by most standards implementers.

Even if some entity (eg, a national court) attempted to determine the joint optimum royalty terms, it would invariably lack the information, the analytical framework and the jurisdiction with which to carry out any such determination. To begin with, the coordinating entity would need information, including the existing royalty terms and conditions negotiated by every industry licensor and licensee (in the FTC’s recent case against Qualcomm, the parties subpoenaed all industry SEP licences, which appears to be the only time that this highly confidential information has been assembled in one place). Similarly, the entity would require an analytical framework by which to interpret dissimilar agreements (eg, cross-licences) and fixed or running-royalty licence structures and then to re-form such contracts to achieve the theoretical joint optimum, for which there is currently no economic model. Finally, the entity must enforce the joint optimum on all licensors and licensees, not just those subject to its jurisdiction. No such information-gathering process, analytical framework or enforcement mechanism exists in the telecommunication (or any other) industry.

Given these limitations, implementers cannot prove, even in theory, that the uncoordinated licensing system contemplated by SDO IP rights policies produces a royalty stack that causes harm, relative to a known and provable joint optimum benchmark. Much less is there any empirical evidence for such harm. Absent such evidence, factfinders should not consider the royalty stacking problem at all (Ericsson v D-Link Systems, Inc, 773 F3d 1201). But even if it were shown to exist, the royalty stacking problem is not caused by an individual licensor. It is caused by SDO policies that favour uncoordinated bargaining, which is a social not private problem.

On the other hand, given that nearly all SEP portfolios are licensed by a sub-set of implementers only, innovators as a group are demonstrably harmed by the systematic non-payment of royalties. In general, the social problem is not that implementers’ royalty stacks are too high, but that they are too low.