I've now had a chance to read through the Joint Research Centre (JRC) Science for Policy Report titled Licensing Terms of Standard Essential Patents: A Comprehensive Analysis of Cases, which I mentioned on this blog last month (here). (The JRC is "the European Commission’s science and knowledge service," and I have previously blogged on its recent report on patent assertion entities in the EU, here.) The authors are Chryssoula Pentheroudakis and Justus A. Baron, and the editor is Nikolaus Thumm. The report, which is fairly lengthy, provides a terrific summary of all of the FRAND/SEP cases from the U.S., the E.U., China, Japan, and India, and thus should be of immense interest to anyone involved with these cases. (It's a long report, but if you're pressed for time there is a concise executive summary at the beginning; and, as noted, the report is a good reference to have on hand for the details of all the FRAND cases to date.) The report also covers the various issues relating to holdup and holdout, royalty stacking, calculating FRAND royalties, whether or not to grant injunctive relief, and it concludes with some specific policy recommendations. The principal normative conclusions, in my reading, is that (1) FRAND is a range, not a specific number; (2) courts face immense difficulties in calculating FRAND royalties; and (3) at least within the E.U., the emphasis should be on encouraging voluntary negotiations by establishing some boundary conditions (e.g., the steps that SEP owners and implementers must take to avoid or invoke, as the case may be, the competition law defense, in accordance with the CJEU's Huawei v. ZTE decision). On this last point in particular, the authors write, under the heading "There are limits to what courts can do or should be expected to do," the following:
Evidentiary rules and sophisticated methodologies developed by the US courts for the calculation of FRAND royalties are not particularly useful in the European context. These tools are designed to assist the US courts in determining a single FRAND rate. In contrast, in the context of injunctions, European courts have focused on defining the conditions under which the conduct of the negotiating parties is incompatible with their FRAND obligations. The increased reliance of firms on the judicial system for the determination of FRAND rates risks undermining their incentives to agree the price of intellectual property through bilateral negotiations. The judicially defined rates are generally based on 1) the prices of infringing components, which may bear little information on the value of the technology, and 2) comparable licenses that reflect the parties’ assumptions regarding the outcome of litigation rather than their valuation of the patented technology. It’s hard to imagine that substantial methodological progress could be made starting from these premises. An economically sound approach is only possible once it is recognized that the ex ante-driven methodological challenges that courts need to overcome to determine ex post an appropriate royalty rate are simply overwhelming.
Against this background, policies that support market mechanisms and conditions conducive to bilateral negotiations and their proper conduct as early on as possible can enhance clarity around the definition of FRAND and restore legal certainty in the field of SEPs (p.161).
See also p.124 (stating that "In order to determine a single royalty rate deriving from a hypothetical agreement of this kind, US courts are methodologically sophisticated when they approach FRAND. In contrast, European courts are more reluctant to define a single royalty rate. Instead, they focus on the conduct of the parties during the bilateral negotiations and assess whether it complies with the specific FRAND commitments made prior to awarding injunctions.").
Among the other notable aspects of the report are its discussions of the literature on the ex ante negotiation benchmark (e.g., when is the negotiation assumed to take place, the lack of a clear answer thus far as to how to take into account patented alternatives, etc.); the distinction between the stand-alone value of the patented technology and its value by virtue of being included in the standard (see pp. 129-31, in particular); the use of comparable licenses as benchmarks (pp. 147-50); and a proposed four-step framework for choosing a royalty base (pp. 142-47). On this last issue in particular, the authors argue that "it is necessary to move beyond the concepts of EMVR [entire market value rule] and SSPPU [smallest saleable patent practicing unit]," and that in some instances "appropriate compensation may be a very large share of the price of a component implementing the patented feature . . . not restricted to values below 100%," as Nicolas Petit has also observed. The authors' proposed four steps, which they explain at some length, are (1) "Examine whether the suggested base (end product or component) satisfies the EMVR. Examine whether the suggested base (end product or component) is the SSPPU."; (2) "Examine whether the price of the chosen base accounts for the value of the technology."; (3) "If this is the case, examine whether a method exists to reliably isolate and identify the value from this price."; and (4) "Assess whether the reference to the base nevertheless invites for confusion and may mislead the jury in its decision." The authors also suggest in passing that conjoint analysis can be helpful in carrying out the necessary apportionment. Finally, I'm intrigued by the authors' discussion (pp. 156-57) of Judge Thomas Kühnen's proposed methodology for calculating FRAND royalties, which apparently is set out in the new (ninth) edition of his Handbuch der Patentverletzung (which I'll have to acquire; I don't think the English translation of this edition by Frank Peterreins is out yet.), which is reminiscent of some remarks by Judges Kühnen and Maimann as recorded in this article (and which I also cite in this article at p.45). From the JRC Report:
The purpose of determining a FRAND royalty is not to provide an adequate compensation to the SEP owner for the use of its patents, but to achieve a balance of interests. The court may use different methodologies for the royalty calculation such as cost-benefit analysis or comparable licenses. In the absence of comparable licenses, the court has to rely on the data points and market-related information provided by the parties in support of the proposed rates. Judge Kühnen cites concrete examples for the calculation of royalty rates as well as the apportionment of value for an SEP portfolio; factors relevant for this “undocumented” (“vorlagenfrei”) calculation are the number of SEPs/non-SEPs, the various degrees to which the underlying technology drives the sales of the end-product (“first class”, “second class”, non-practiced bundles with defensive use), essentiality and sales data. Albeit not binding, the calculation examples include a ceiling cap of about 1/3 of the net selling price of the highest-priced standard-compliant product and apportion a higher share of the total royalties for “first-class” SEPs as opposed to the merely nominal royalties apportioned for “second-class” SEPs. Irrespective of the chosen methodology, the ultimate purpose of the royalty calculation is not to achieve mathematical accuracy, but an approximation based on certain values and estimates for the sake of procedural efficiency.
The authors go on to note, however, that "most scholars believe it fairly improbable that German courts would imminently adopt novel arguments or develop certain methodologies on FRAND . . . ."