The policies of standard-setting organizations (SSOs) described in the previous chapter address situations in which the SSO develops standards whose use necessarily infringes known intellectual property rights and specify the commitments that participating IPR holders must accept regarding such standards. While the specific language may differ, most SSOs ask rights holders to consent to license their rights on terms that are fair, reasonable and nondiscriminatory (FRAND), with or without a royalty payment. For example, the ANSI patent policy requires assurance that either: a) a license will be made available, without compensation, to the applicants desiring to utilize the license for the purpose of implementing the standard; or b) a license will be made available to applicants under reasonable terms and conditions that are demonstrably free of any unfair discrimination. Otherwise, the standard cannot be approved as an American National Standard (ANSI, rev. 2008).
Neither ANSI nor most other SSOs define “reasonable terms and conditions” or the requirements for a license to be “free of any unfair discrimination.”1 Some SSO policies encourage rights holders who have made assurances to reach bilateral agreements with potential licensees. Disputes may be resolved in court or through arbitration or, in some instances, may involve regulatory agencies if there is an allegation of anticompetitive behavior. Despite a relatively small number of cases in which courts have considered the requirements of FRAND licensing terms, there is as yet no broad consensus on this topic.
In this chapter we take a step back and ask the following basic questions:
- What plausible objectives motivate the adoption of FRAND licensing obligations by members of standard setting organizations?
- Do the objectives of members of SSOs differ from societal and competition law concerns?
- How do differences in private and social objectives inform interpretations of FRAND obligations?
1Special Interest Groups (SIGs) sometimes list explicit licensing terms.
For virtually all SSOs, the minimum goal regarding IPR is to ensure that all known essential claims in patents are available under FRAND license terms. Some SSOs, or discrete working groups within an SSO, may adopt a more stringent set of rules to seek to have all essential claims made available on a FRAND royalty-free basis. FRAND obligations generally provide assurance that licenses are available for technical solutions involving essential IPR. Participants in SSOs generally do not oppose the development of such standards. For example, the 2008 ANSI Patent Policy states “There is no objection in principle to drafting a proposed American National Standard in terms that include the use of a patented item, if it is considered that technical reasons justify this approach.” The FRAND obligations adopted by ANSI and others place limits on the exercise of these patent rights.2
As noted earlier, standards-setting organizations have a diverse set of constituents. Some SSO participants are technology owners and users whose business models are based on the sales of products that implement standards and employ patented technologies. Some of these technology users are not interested in asserting standard-essential patents (SEPs) of their own and desire low – or no – royalties for the SEPs they license from others. While these technology users may not proactively seek compensation from implementers for their own FRAND-encumbered patents, they may want their patent rights to have value in terms of offsetting SEPs held by others in the same standards technology area. Other SSO participants are technology sellers whose business models are based on earning royalties from licensing their SEPs to implementers. These enterprises want high royalties and in some cases may want to use their SEPs to demand that implementers accept a license that includes non-SEPs. Still other participants are both technology users and sellers. There are many other business models that impact the views of the various stakeholders.
Technology owners may also adopt different postures depending upon the individual SSOs in which they participate and their particular strategies and goals as they relate to the standard in question. For example, in cases where promoting rapid adoption of a standard enabling a new technology is important to a patent owner and the patent owner can monetize its technology through the sale of products that implement the standard, it may choose to make its SEPs available for free. A patent owner who believes that his SEP is of particular importance to a standard, or one whose technology may meet the needs of a small-volume product, might seek royalties. At a basic level, many SEPs exist for which no license is sought through negotiations unless a third party seeking royalties approaches their owners and implementers. SSOs inevitably shape their
2We limit the ensuing discussion to patent rights, to which the concept of FRAND obligations uniquely relates. As noted in Chapter 2, IPR policies do typically address copyrights and trademarks as well, but apply a different set of rules that are of limited relevance to this chapter.
IPR policies over time to address the concerns of their existing members and to attract new participants who may be technology users, sellers, or both.
Prevailing practices can also vary from industry to industry, with participants in a number of SSOs in some areas, such as the Internet, preferring to avoid royalty-bearing standards, while those in others, such as consumer electronics, often seek royalties. However, with the convergence of the consumer electronics, information technology, and telecommunications sectors it is becoming more difficult to make distinctions based on the technology area and the impacted industry sectors.
The diversity of actual and potential members of SSOs helps to explain why few of them have developed policies that include detailed definitions of FRAND. Rather, most SSOs rely on general FRAND licensing commitments and certain clarifications with regard to the effect of such commitments as the need arises. SSOs have to govern their IPR policies in an environment of conflicting interests.
To better understand the extent of the limitations imposed by FRAND commitments, it is useful to consider the likely reasons why the members of SSOs arrived at the specific descriptions of their current patent policies. Several concerns are evident.
Ensuring access to patented technologies
Absent a FRAND or other commitment, the owner of a patent has no obligation under the policy to license others to use the patent on any terms. On its face, a FRAND commitment is intended to constrain the freedom that a right holder otherwise has to refuse to license its technology and subsequently enforce its rights. It is understandable that members of a SSO would insist that the organization seek obligations to license patents that are essential to make or use products that comply with a standard. The purpose of an interoperability standard is to coordinate industry activity and take advantage of the economic benefits of scale economies and network effects. These benefits cannot be achieved without widespread licensing of the patented technology that is essential to practice a standard.
In the absence of FRAND licensing obligations, there is a risk that implementers will accede to demands to accept discriminatory licensing terms in order to adopt fundamental ICT standards embodying essential patents that cannot be worked around. The owner of a SEP may choose to license the proprietary right with terms that are less favorable to some licensees. For example, without FRAND, a patent owner might license its own customers and partners under terms that differ from those offered to a significant rival, either within or outside
the bounds of antitrust law and other legal constraints.3 Thus, an objective of a FRAND licensing commitment is to avoid discriminatory licensing terms that disadvantage some licensees by imposing on them substantially larger royalties or more restrictive conditions in comparison with others. Indeed, this assurance of non-discrimination is fundamental to the weighing of benefits and costs that an industry participant makes when deciding whether to participate in an SSO or adopt a standard.
Avoiding ex post hold-up
The economic alternatives available to licensees often differ before and after the adoption of a standard. In the process of developing a standard, several alternative technological solutions may be available that have similar cost and performance characteristics. “Ex post,” after firms and consumers make investments that are specific to the standard, the economic choices are far more limited if adopting an alternative technology for the standard would impose substantial additional costs and delays. Furthermore, in markets with large network externalities it may not be feasible to coordinate the actions necessary to switch the market to a different standard. Ex post hold-up can occur if the owner of a SEP chooses royalty terms that reflect the high cost of switching to an alternative technology after firms and consumers have made specific investments, rather than the value of the claimed invention.
Ex post patent hold-up imposes costs on licensees and consumers. It also rewards the patent holder via windfall profits reflecting the costs of switching to an alternative technology rather than the economic merit of the selected standard. Because such hold-up is potentially costly to members of SSOs, it is reasonable for their IPR policies to seek to guard against it by requiring FRAND commitments. This concern is greatest in ICT sectors where standards aspire to global adoption, switching to alternative standards can be costly, and patent filings have proliferated in many countries.
Competition authorities in both the EU and the United States see the potential for costly hold-up, as noted in the following joint statement by officials of the Department of Justice Antitrust Division, Federal Trade Commission, and Competition Directorate General of the European Commission:
SSOs constrain the license terms for SEPs because of the substantial market power necessarily enjoyed by the owner of an SEP in a successful standard. Moreover, this market power is achieved through the joint action of entities—the SSO members—that might be in competition with each other outside the SSO (Kühn, et al., 2013).
3Georgia-Pacific Corp v. United States Plywood Corp., 318 F. Supp.1116 (S.D.N.Y. 1970). The well-known Georgia-Pacific case, in fact, recognizes the relationship between the patent owner and the infringer as a factor to consider in assessing “reasonableness” in a patent damages case.
Managing royalty allocation and stacking
Interoperability standards often have numerous patents declared essential to their use. Economic valuation of a particular patent becomes difficult and even arbitrary in this circumstance. If several patents are essential to make or use a product that complies with a standard, then each patent has a claim on the value of the product. There is a concern that, acting independently, the individual holders of patents essential to a standard will demand royalties that, in the aggregate, are so high as to impede the adoption and use of products that implement the standard. Such high aggregate royalties can be detrimental to rights holders as well as to the consumers of products that implement the standard and can be a drag on future innovation. Economists use the term “royalty stacking” (Lemley and Shapiro, 2007) to describe outcomes in which the cumulative effects of individually rational royalty demands result in aggregate royalties that harm consumers, rights holders and innovators.
Royalty stacking is different from standard-related hold-up. When a standard-using product requires numerous SEPs, a single patentee can demand a disproportionate share of product value even if licensees do not incur costs to switch to a different product. Because all the SEPs are necessary, licensees and customers would switch to a different product only if the total royalty is excessive, which leaves room for individual patent holders to demand a disproportionate share of total royalties.
Several important ICT standards illustrate the potential for royalty stacking. GSM is a second-generation standard for mobile telecommunications and W-CDMA (also called UMTS) is a third-generation standard for networks based on the GSM standard. One study identified more than 50 entities that each disclosed patents or patent applications as essential to the GSM or W-CDMA standards.4 Together these entities disclosed over 23,500 patents, belonging to at least 1729 patent families.5 Other mobile telephony technologies have similar characteristics. Numerous firms in total declared more than 750 unique patent families as essential to the second-generation GSM standard and 500 unique patent families as essential to the fourth-generation LTE standard.6
A U.S. district court concluded that 92 companies identified patents as essential to the 802.11 (“Wifi”) wireless local access network family of standards and 59 companies filed blanket declarations without identifying specific patents. The court accepted testimony that there are possibly thousands of patents declared essential to the 802.11 family of standards.7 The same court concluded that approximately 33 U.S. companies declared patents essential to the H.264
4See Bekkers and Martinelli (2012), Table 6, p. 1205. (The table includes GSM as well as W-CDMA patent disclosures).
5Bekkers and Martinelli (2012), pp. 1203-1205.
6See Blind et al. (2011), Table 3-3, p. 36.
7Microsoft Corp. v. Motorola, Inc., Findings of Fact and Conclusions of Law, 2013 U.S. Dist. LEXIS 60233 at paragraph 335 (W.D. Wash., Apr. 25, 2013).
advanced video coding standard and 19 additional companies provided blanket declarations to the ITU (one of the developers of the standard) without identifying specific patents.8
These figures could both overstate and understate the potential for royalty stacking in different respects. The numbers could overstate the potential for royalty stacking because patent owners that participate in SSOs have incentives to declare, as essential, patents that may not in fact be valid or infringed by a product that implements the standard. Evaluating the scope and validity of every patent is a costly exercise and patent holders may prefer to declare patents as essential without subjecting them to careful scrutiny. Further, SEP owners may err on the side of disclosing patents that are not essential to the standard because failure to disclose a SEP for which royalties are demanded, where it is required by an SSO, may expose the owner to future litigation. Finally, ownership of a large stock of patents that are declared essential to a standard can be a valuable asset for a firm that seeks licensing royalties or cross-licenses at favorable terms. A larger quantity of essential patents is also more valuable if the patents are contributed to a patent pool that seeks royalty income and allocates that income in proportion to the number of patents in the pool.
Estimates of essential patents and the number of SEP owners could understate the potential for royalty stacking in other respects. Multiple SSOs participate in the development of some standards and not all SSOs maintain easily accessible databases of patent declarations. Some participants in SSOs make blanket disclosures and commit to license patents deemed essential to a standard at FRAND terms without identifying the specific patents in advance. Also, some firms might hold essential IPR without having an obligation to disclose these patents, for instance because they are not participating in the Working Group in question or because the participating individuals were not aware of those patents. Non-participant firms are not required to disclose the existence of these patents and would not appear in a review of licensing commitments at the relevant standard-setting organizations.
Regardless of whether observed statistics overstate or understate the number of owners of essential patents, it is certain that numerous distinct entities own patents that are essential to make or use products that implement many common ICT standards. Given the large number of SEP owners, efforts to obtain royalty income for these patents can result in heavy monetary burdens for those who make or use products that implement the standard. In theory, the total royalty stack can be so large that it would suppress the adoption or use of standardized technologies and impose excessive costs on all segments of the industry that implement a standard.
Despite this concern, the committee has found no empirical evidence showing that royalty stacking currently suppresses the adoption or use of standard-compliant products. Firms can mitigate the burden from aggregate royalties
in several ways.9 Some rights holders enter into cross-licensing arrangements with zero royalties or with royalties equal to the difference between the fees charged by each party. Patent pools exist for some standards, which reduce transaction costs and mitigate royalty stacking by setting a single fee for a portfolio license. Firms that own patents and sell products covered by those patents have incentives to charge low or zero royalties to promote the commercialization of their products. In addition, firms have strategic incentives to refrain from charging high royalties. Indeed, product prices have been dropping for devices such as mobile phones and laptop computers that support multiple standards for which there are thousands of declared SEPs owned by hundreds of entities. Furthermore, not all standards, even in the ICT area, invoke large numbers of patents with widely distributed ownership.
Nonetheless, the committee cautions that the costs from royalty stacking could increase in the future if more patent owners choose to monetize their patent rights. At some point the cumulative burden of making multiple royalty payments to distinct entities could become so large that adoption or utilization of standard-compliant products would be suppressed and the resulting higher costs of developing and producing these products may become a drag on future innovative efforts.
Courts play a crucial role in this area since judicial decisions both directly set norms for FRAND terms and settle disputes over whether royalty offers comply with FRAND commitments. Thus, courts can contain the risk of royalty stacking and hold-up by ensuring that awards for patent infringement are reasonable, taking into account the contribution of the patented inventions and the costs of obtaining all other intellectual and physical inputs that are necessary to make or sell the infringing product.10 For patents that are essential to a standard and allegedly infringed by an implementing product, this requires allocation of the value contributed by the standard as opposed to the contributions of others, including patents that are essential to other standards.
An allocation of the value of a standard to its essential patents based on simple numeric proportionality, under which a patent owner’s share of value is equal to its share of patents that are essential to make or use the products at issue, has the virtue of simplicity and ease of administration. However, it also raises the potential for imprecision and strategic manipulation. Some patents are more valuable than others in terms of their contributions to the standard, either because their validity has been firmly established through litigation or because
9See Damien Geradin, Anne Layne-Farrar, and A. Jorge Padilla, The Complements Problem Within Standard Setting: Assessing The Evidence On Royalty Stacking, B.U. J. Sci. & Tech. L., Vol. 14:144-176.
10Additional claims on the value of the product could come from patents covering other standards implemented in the product and patents on proprietary technologies that contribute to the overall value of the product.
they are more central to implementing products.11 Furthermore, a numeric proportionality rule encourages patent owners to file separate patents for essential claims instead of including multiple claims in a single patent. Nevertheless, the fact that many owners of SEPs participate in patent pools that allocate royalties based on numeric proportionality suggests that such rules, however imperfect, are a plausible starting point for the valuation of individual SEPs.
On April 25, 2013, Judge James L. Robart in the Western District of Washington issued a detailed decision calculating FRAND royalty rates for products that involve multiple standard-essential patent rights.12 The decision relied on the factors laid out in the earlier Georgia-Pacific case13 but modified them to account for the FRAND context. In particular, Judge Robart noted that in considering a hypothetical negotiation to arrive at a reasonable royalty “the hypothetical negotiation almost certainly will not take place in a vacuum: the implementer of a standard will understand that it must take a license from many SEP owners, not just one, before it will be in compliance with its licensing obligations and able to fully implement the standard.”14 In particular, Judge Robart concluded that “a proper methodology for determining a [F]RAND royalty should address the risk of royalty stacking by considering the aggregate royalties that would apply if other SEP holders made royalty demands of the implementer.”15
At issue in the case was the value of Motorola’s patents for Microsoft products that implemented the ITU H.264 standard for video processing and the IEEE 802.11 family of standards for wireless communications. In constructing a reasonable royalty, Judge Robart focused on royalties that Motorola would have earned for its patents if they had been contributed to existing patent pools relating to those standards. Although the judge listed a number of reasons why pool rates should not be determinative of the FRAND rates for all SEPs for a standard, he found them to be useful indicators based on the facts of the case.
Judge Robart’s approach provides less quantitative utility to assess reasonable royalties for multiple standard-essential patents where there are no existing patent pools of SEPs related to the standard in question. Nonetheless, his efforts to distinguish comparable royalty negotiations, identify the risks of royalty stacking, and take into account the potential aggregate burden of licensing
11For illustrations of royalty allocations under different assumptions, see Layne-Farrar et al. (2007) and Salant (2007). Although several techniques are available to allocate value to SEPs or to patents that are complementary to the value of products that implement a standard, they require more information about technology characteristics than is typically available.
12Microsoft Corp. v. Motorola, Inc., Findings of Fact and Conclusions of Law, 2013 U.S. Dist. LEXIS 60233 (W.D. Wash., Apr. 25, 2013) [hereinafter Microsoft v. Motorola FFCL].
13Georgia-Pacific Corp. v. United States Plywood Corp. (318 F. Supp. 1116 [S.D.N.Y. 1970]).
14Microsoft Corp. v. Motorola FFCL, op. cit. at 11.
demands from many patent owners provide guidance that may be useful to assess lower and upper bounds for FRAND royalties when many patents are essential to make or use products that comply with a standard.
In general, determining whether a given licensor is making excessive demands that contribute to either royalty stacking or standard-related hold-up can be very challenging. If a standard requires the patented contributions of many different licensors, each of which is essential to implement the standard none of which have close substitutes ex ante or ex post, then the division of royalties among patented contributions is essentially arbitrary. In other instances of standard setting, some features of a standard may be more important than others and some features (and the technologies that implement them) may have feasible substitutes prior to the standard being adopted. In these substitutions some licensors may argue, with some justification, that their technology is worth more than others. Further, in the presence of switching costs it can be difficult or even impossible to distinguish excessive demands from hold-up. A large royalty demand could reflect the effort of a patent owner to appropriate a share of the switching costs that lead to ex post hold-up. But it also could be an effort by a particular patent owner to capture a disproportionate share of an ex ante reasonable aggregate royalty.
There is little evidence that the existing IPR polices of most large SSOs effectively limit the ability of individual patent owners to negotiate for a disproportionate share of product value in their royalty demands. Still, some SSOs, including various consortia, IEEE, Wi-Fi, and ETSI, have explored the idea of using various licensing disclosure arrangements to attempt to avoid royalty stacking, although the effort and expense of putting such arrangements together is justified only in some circumstances.
One approach that may address concerns about royalty stacking is for SSOs to require patent owners that intend to assert their patents to post a maximum royalty before the standard is adopted. The posting of royalties would allow potential licensees to detect potential ex ante hold-up and possible royalty stacking by considering the implications of individual posted maximum royalties for the aggregate royalties required to make or use products that comply with the standard. SSOs might also establish mechanisms for effectively avoiding or resolving disputes, such as alternative dispute resolution (ADR).
That said, policies requiring ex ante disclosure of licensing terms have been suggested and rejected at several SSOs. Such policies are unpopular with patent holders that prefer to negotiate royalty rates with potential licensees and not to be pressed into royalty rates and terms before the standard is finalized and relevant SEPs can be determined.
Another objection to ex ante licensing disclosure policies is that they could lead to anticompetitive conduct by licensees if royalty rates were jointly negotiated—a scenario referred to as oligopsony or a buyers’ cartel. Given this concern, both VITA and IEEE obtained written assurance from competition au-
thorities before adopting ex ante licensing disclosure policies.16 Opponents of ex ante disclosure of licensing terms also point to practical difficulties, such as fixing the timing of disclosures and designing policies that would accommodate blanket disclosures and avoid the need for costly patent searches.17 The latter concern is especially relevant for companies that generally do not proactively seek licenses from implementers, but rather use their SEPs largely for defensive purposes. In addition, many SEPs arise not because a patent holder contributed the technology to the SSO, but rather as a result of the collective drafting exercise. Mandatory ex ante disclosure could also disrupt technical committee work, if participants were asked to review all possible SEPs and related licensing terms, particularly for ICT standards that can reach hundreds of pages in length and implicate dozens or hundreds of patents.
Because few SSOs have adopted policies with regard to ex ante disclosure of licensing terms, empirical evidence on these questions remains quite limited. However, one study found that the ex ante disclosure policies adopted by VITA, where disclosure is mandatory, and IEEE, where disclosure is optional and seldom used, had no measureable impact on their standard-setting processes (Contreras, 2013).
Another proposal to address royalty stacking involves an SSO’s establishment of an aggregate cap on royalties that can be charged with respect to patents essential to a particular standard. This approach was proposed within ETSI as early as 2005, but its consideration was terminated after an unfavorable reaction by the European Commission’s Competition Directorate-General.18 Yet another possible approach would be for SSOs or courts to make an effort to consider the landscape of patents that are essential for use of a standard. That recognition would help avoid some of the most egregious errors (such as 5% royalty per patent). Equal apportionment of values, though sometimes flawed, could be a starting point from which to argue that some patents are worth more than others, while still recognizing that one or a few patents may not account for all or most of the product value when there are many other essential patents.
16See U.S. Department of Justice, Business Review Letter to VMEbus International Trade Association (Oct. 20, 2006), U.S. Department of Justice, Business Review Letter to Institute of Electrical and Electronics Engineers (Apr. 30, 2007), and Deborah Platt Majoras, Recognizing the Procompetitive Potential of Royalty Discussions in Standard-Setting Remarks prepared for “Standardization and the Law: Developing the Golden Mean for Global Trade,” Stanford Law School (Sept. 23, 2005).
17These difficulties apply generally to SSO policies regarding ex ante disclosure of SEPs, not just to licensing terms, as discussed in Section 4.
18Claudia Tapia, Industrial Property Rights, Technical Standards and Licensing Practices (Frand) in the Telecommunications Industry 165-66 (2010). A modified form of this proposal has recently been made by Contreras, who analogizes the aggregate royalty on SEPs covering a particular standard to the collective royalty charged by a patent pool. See Jorge L. Contreras, Fixing FRAND: A Pseudo-Pool Approach to Standards-Based Patent Licensing, 79 Antitrust L.J. (2013).
Competition agencies in the United States and the European Union have proposed interpretations of FRAND licensing obligations in an attempt to “fill in the gap” created by the lack of clarity in SSO IPR policies.
Competition authorities and a number of scholars have endorsed the principle that a “fair and reasonable” royalty should reflect its incremental value relative to the next-best alternative assessed before firms and consumers make investments specific to the technology. For example, the Federal Trade Commission concludes that “Courts should cap the royalty at the incremental value of the patented technology over alternatives available at the time the standard was chosen.” (U.S. Federal Trade Commission, 2011) Guidelines issued by the European Commission state that “whether fees imposed for patents in the standard-setting context are unfair or unreasonable will be based on whether the fees bear a reasonable relationship to the economic value of the patents.” In making this assessment, the Commission notes that “it may be possible to compare the licensing fees charged by the undertaking in question for the relevant patents in a competitive environment before the industry has been locked into the standard (ex ante) with those charged after the industry has been locked in (ex post).” (European Commission, 2010)
Many SSOs emphasize that it is desirable to have early disclosure of relevant patents and some have provided for ex ante commitments to specific licensing terms and conditions. However, absent further clarification of the meaning of FRAND, it is not clear whether members of SSOs intend that FRAND royalty commitments should reflect incremental values or some other notion of fair and reasonable pricing. This topic is currently under discussion at some prominent SSOs. If the term does not reflect incremental value one might question whether norms such as economic efficiency should determine the interpretation of fair and reasonable license terms.
Incremental value provides a means to assess the ex ante contribution of a patent that covers a discrete technology whose value can be assessed independently from the contributions of other technologies. For example, a patent on a technology to increase the signal-to-noise ratio of transmissions using a wireless communications standard can be assessed independently from the contributions of other patents that are essential to the standard. Farrell et al. (2007) provide a formula for the incremental value of a technology. Swanson and Baumol (2005) suggest that SSOs should conduct an auction to determine the incremental value of the best technology. These approaches may provide feasible means to estimate the value of a patent that is essential to a standard that involves a single patented technology and if alternative technologies can be compared based on a single attribute such as cost. However, the technologies incremental
approach is not sufficient to determine the appropriate royalty for a particular patent when there are many SEPs, as is typically the case for interoperability standards.
When multiple patents are essential to make or use products that comply with a standard, neither incremental value nor auction approaches provide a practical means to allocate the economic value of the technology to different necessary patents within the standard. When patented technologies in a standard have different attributes, valuation requires aggregation of the attributes into a metric that can be used to rank alternatives. This is a complex problem and technology users are likely to disagree about the appropriate weights for the different attributes.19
Practically speaking, in assessing “reasonableness,” courts have resorted to measures other than incremental value. Recently, a long-time rule of thumb gauging royalties at 25% of the infringer’s profits was rejected by the U.S. Court of Appeals for the Federal Circuit.20 In accordance with a statement by the European Commission, the royalties a patent has realized in the past before a standard is approved can shed light on value.21 What royalties the patent has realized in the past in other license negotiations after or (in accordance with the European Commission statement) before a standard is approved, can shed light on value. Royalties for comparable patents or licenses that are encumbered by FRAND commitments may also inform the evaluation of reasonable royalties, although recent cases have raised the bar on validating the circumstances and expert testimony on which cases are comparable.22
At the most basic level, a commitment to license at RAND terms should not permit a patent holder to obtain a royalty that reflects standardization effects rather
19Patent pools have developed “rough and ready” approaches to this apportionment problem, such as allocating royalties in proportion to each firm’s count of essential patents. However, each approach bears its own problems, especially where attributes must be weighted.
20Uniloc USA Inc. v. Microsoft Corp. 10 1035 (U.S. Court of Appeals for the Federal Circuit, 2011).
21See Communication from the Commission: Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal cooperation agreements, at http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:C:2011:011:0001:0072:EN:PDF. As an economic matter, it is questionable whether ex post royalties should inform FRAND royalties.
22As noted by Judge Robart in Microsoft Corp. v. Motorola, Inc., No. C10-1823JLR (U.S. District Court, 2013), not all royalties are comparable or based on similar considerations. For example, “[t]he court concludes that where multiple technologies (including both standard essential and non-essential patents) are licensed within the same agreement, it is necessary to apportion the value of [the SEPs at issue] from the other licensed properties. Such apportionment would be difficult.” (Order 137-138). See also Uniloc USA, Inc. v. Microsoft Corp., 632 F. 3d 1292, 1317 (Fed. Cir. 2011) (“The meaning of these cases is clear: there must be a basis in fact to associate the royalty rates used in prior licenses to the particular hypothetical negotiation at issue in the case.”).
than the value of the claimed invention; nor should a patent holder automatically realize less than the rate based on the ex ante contribution of the invention merely because s/he participates in a standards effort. In Microsoft v. Motorola, Judge Robart articulates two high-level principles for determining FRAND rates. First, the value of the SEPs should be assessed separately from their inclusion in the standard. Second, a FRAND rate should be based on the importance of the patents to the standard and the importance of the standard and the patents to the products at issue. As noted above, one corollary of this second principle, the ruling suggests, is that when there are multiple holders of FRAND-encumbered SEPs for the same standard, it would be necessary to consider cumulative royalty rates.
U.S. courts frequently cite the 15 factors enumerated in Georgia-Pacific Corp. v. United States Plywood Corp.23 that are relevant to a hypothetical negotiation between a patentee and a licensee to assess monetary compensation for patent infringement. A key issue now is whether these factors are informative for assessing FRAND royalties for standard-essential patents. Judge Robart concluded that they are informative, albeit with important qualifications. He noted that the hypothetical negotiation under a FRAND obligation differs from the typical Georgia-Pacific analysis conducted by courts in a patent infringement action because, among other reasons, the owner of a SEP is under the obligation to license its patents on FRAND terms.
For example, the first Georgia-Pacific factor is “The royalties received by the patentee for the licensing of the patent in suit, proving or tending to prove an established royalty” and the second factor is “The rates paid by the licensee for the use of other patents comparable to the patent in suit.” Judge Robart concluded that to be comparable, past royalty rates for a litigated SEP or another similar patent must be negotiated under a clearly understood FRAND obligation.
Several Georgia-Pacific factors address the technical characteristics of the patented technology and its value to the licensee. Factors 6 and 8 relate to the value of the patent in promoting sales of the licensee’s products and factor 9 is “The utility and advantages of the patent property over the old modes or devices, if any, that had been used for working out similar results.” Factor 10 is “The portion of the realizable profit that should be credited to the invention as distinguished from non-patented elements, the manufacturing process, business risks, or significant features or improvements added by the infringer.” In interpreting the implications of these factors for a hypothetical FRAND royalty negotiation, Judge Robart concluded that a reasonable royalty for a SEP should reflect the contribution of the patent to the standard and to the value of the implementer’s products, but should not take into account the value to the licensee created by the standard itself.
Factor 15 is “The amount that a licensor, such as the patentee, and a licensee (such as the infringer), would have agreed upon at the time the infringement began if both had been reasonably and voluntarily trying to reach an agreement …” In evaluating the implications of this summary factor, Judge Robart noted
23318 FSupp 1116, 6 USPQ 235 (SD NY 1970).
that “the SEP owner would have been obligated to license its SEPs on RAND terms which necessarily must abide by the purpose of the RAND commitment of widespread adoption of the standard through avoidance of holdup and stacking.”
It remains to be seen whether other courts will adopt the FRAND framework described by Judge Robart in Microsoft v. Motorola. Nonetheless, the court’s application of the Georgia-Pacific factors to a hypothetical negotiation over royalties for a FRAND-encumbered patent emphasizes certain points that are widely discussed in the literature and have achieved some consensus within the standards community and among antitrust authorities. Specifically, the court determined that the participants in a “hypothetical negotiation would set [F]RAND royalty rates by looking at the importance of the SEPs to the standard and the importance of the standard and the SEPs to the products at issue.”24 It is reasonable to apply the Georgia-Pacific factors to help determine FRAND licensing rates, with appropriate modifications to reflect the commitments made by SEP owners and the characteristics of industries within which SEP owners and implementers operate.
Some notion of non-discrimination must be central to any meaningful FRAND concept.
In his review of the Georgia-Pacific factors to assess FRAND royalties, Judge Robart noted that the fifth factor—”The commercial relationship between the licensor and licensee, such as, whether they are competitors in the same territory in the same line of business; or whether they are inventor and promoter”— does not apply in the FRAND context. This is because having committed to license on FRAND terms, the patentee is obligated to license all implementers on reasonable terms and may not discriminate against its competitors in terms of licensing agreements.
Although the non-discrimination requirement of the FRAND commitment is clear, it is much less clear what that implies in practice. A patentee can structure royalties in different ways. One might imagine a licensor asking for a unitary fixed charge independent of the licensee’s sales volume, a given fee per unit of sales, or a mixed royalty. Each of these may be considered nondiscriminatory on its face but generate wide variations in effective licensing terms across licensees. For example, a royalty rate that is a fixed percentage of a using product’s market value implies a much higher per-unit dollar royalty on higher value goods than on less expensive goods, even though the same technology is in play.
In general, neither a single fixed fee nor a single per-unit fee is likely to promote the most efficient utilization of a technology. Economic output may increase if licensees are free to choose among different fee schedules. For example, a licensee may choose a royalty schedule that has a low initial fee and a
24Ibid., p. 4.
high per-unit fee or a different royalty schedule that has a higher initial fee and a lower per-unit-fee.
One interpretation of the non-discrimination requirement is that patent owners should offer licensees the same choices of royalty schedules. Under this interpretation, the alternatives of a high initial fee/low per-unit fee or a low initial fee/high per-unit fee generally would not be discriminatory if these choices are available to all licensees. It is not clear, however, whether this interpretation is consistent with the non-discrimination requirement of FRAND or whether non-discrimination requires such menus. Moreover, there are circumstances in which such choices likely should be considered discriminatory. For example, the choices could be structured so that only very large users of a technology could profitably accept a license.
Portfolio licensing and cross-licensing
The ability to determine whether the price and other terms relating to a given SEP or essential claim are reasonable and have been applied on a consistent basis is further complicated by the fact that those claims may be part of a bundle of other claims. The latter may not be essential to the implementation of the standard but could be needed in connection with the manufacture of any product incorporating the feature that complies with the standard. Typically, if both parties agree, all of these claims will be included under a single license that includes a single set of terms applicable to all of the referenced patents or claims. Because the owner of these claims is under no obligation to license the non-essential, but desirable, claims included in the bundle subject to FRAND terms, it will be difficult to determine whether a FRAND obligation has in fact been met.
For many companies the objective of taking a license is to obtain the freedom to operate in a technology space. These companies may not be interested in a license to a single patent if their activities may infringe other patents owned by the licensor. Instead, they may want access to the portfolio of patents owned by the licensor, including future patents. This can be achieved with an appropriately structured portfolio license. Similarly, the licensor often may want the freedom to operate without the risk of liability for infringing the licensees’ current or future patents. This freedom can be obtained with an appropriately structured cross-license between the two companies.
Portfolio licenses and cross-licenses raise challenges for evaluating whether a particular patent license is consistent with a FRAND obligation. The royalty paid for a patent portfolio covers many patent licenses, and it can be difficult, or even impossible, to allocate the portfolio royalty to individual patents in a meaningful way. Cross-licenses add a further complication because parties to the arrangements often net out the payments for patents they entail. Parties to a cross-license may not pay any royalties if they agree that their respective portfolios have equal values. But that does not mean that the portfolios—or the individual patents in the portfolios—have no values.
Portfolio licenses and cross-licenses raise issues of transparency for patent royalties. It is difficult to know what the royalty may be for a single patent in a portfolio license or a cross-license. If the SEP owner insists that the SEP be licensed as a component of a portfolio that includes other, non-essential patents, the result can be a demand for a royalty in excess of the FRAND rate for the SEP. A similar outcome may ensue if the licensor insists on a cross-license to patents that have a total value in excess of the FRAND royalty on the SEP. Of course, licensees may not agree to such demands in practice.
Mutually agreed cross-license arrangements are certainly acceptable under SSO policies. However, one-sided demands by SEP owners that the licensee accept patents other than the SEPs in the standard, as a condition of access to the SEPs, would violate the terms of many IPR policies. These expressly limit cross-license terms, typically under a “reciprocity” provision, to other claims that are essential to the implementation of the same standard. An owner of SEPs in this situation would not be permitted to demand a broader cross license without violating its FRAND obligation. Even absent such an explicit rule, trade usage might indicate that the party demanding the broad cross-license would be found in court to have violated its FRAND commitment. The competition authorities have suggested that it likely would be anti-competitive for a holder of FRAND-encumbered SEPs to demand a cross-license to non-SEPs unless both parties voluntarily agree to a broader cross-license deal.
The Department of Justice has suggested that SSOs should prohibit the mandatory cross-licensing of patents that are not essential to the standard or a related family of standards, while permitting voluntary cross-licensing of all patents. Further, in its Decision Regarding Google’s Acquisition of Motorola Mobility DG Competition of the European Commission wrote that “Another concern would be that the SEP holder may force a holder of non-SEPs to cross-license those non-SEPs to it in return for a license of the SEPs.”25
The problems of low transparency and potential discrimination in licensing terms could be addressed by requiring owners of patents with FRAND obligations to provide to an implementer, upon request, a royalty or royalty schedule at which they are willing to license their patents (Gilbert, 2011).26 Portfolio licensing and cross-licensing would remain permissible choices. However, parties to a portfolio or cross-license would have the alternative of accepting a license for a single patent at the posted rate.27
Other licensing terms
Licenses typically include many terms negotiated between parties. In addition to fixed and per unit royalties, the license may have a royalty cap, payments
25See European Commission (2012).
26The committee cautions that there is little systematic evidence of how such a requirement would affect incentives of inventors to contribute IP to a developing standard.
27See for example, American Bar Association (2007).
that are conditioned on the licensees’ sales, and discounts for prompt payment. The royalties comprise only one component of licensing terms. The license may be worldwide or restricted to a country or region. The license may be for a particular field of use or for any economic activity by the licensee. Royalty terms may differ depending on the products made using the licensed technology.
In addition, license agreements may seek to include grant-back provisions in which the licensee agrees to provide the licensor with a license to certain future patents issued to the licensee. In fact, most SSOs’ IPR policies are silent on the topic of grant-backs, which may or may not be within the FRAND framework. Where it exists, the relevant IPR policy may limit the scope of the grant-back to patents that are essential to the same standard and may address whether the grant-back is exclusive or non-exclusive and whether it has a royalty obligation.28 The grant-back license may require that the licensee agree either to license its SEPs under FRAND terms or not to assert its own SEPs under the same standard against the licensor or against other licensees implementing the same standard.
This complexity of licensing terms greatly complicates the determination of “fair and reasonable” royalty terms and any assessment of unfair discrimination.29 A license with a given per-unit royalty that is limited to production in Europe and requires a royalty-free grant-back of future patents is not comparable to an unrestricted worldwide license with the same per-unit royalty. While the IPR policies of SSOs could theoretically define a standard set of licensing terms to facilitate compliance with FRAND licensing obligations, this option has rarely been pursued. One reason is that SSO members prefer to sell technology under their own licenses.
The most common exception arises where the founders of the standardization effort seek to develop only one or a few closely related standards. In this case, the founders will often enter into a cross-license among each other, commonly referred to as a “Promoters’ Agreement,” under which the SEPs of all promoters are pooled on a FRAND or FRAND-RF basis. Under some Promoters’ Agreements, each promoter in turn may be free to license the pool of claims to third parties for purposes of implementing the standard in question. These rights are contained in “Adopter Agreements,” which may include royalty-free grant-back terms similar to those in place between the Promoters. Under such an arrangement, extraneous licensing terms are effectively excluded.
As Judge Robart concluded, FRAND obligations may be informed when a formal patent pool is formed, under which implementers pay a single fee referenced in a common fee schedule and sign virtually the same license as all other pool licensees. Because such pools are difficult and expensive to create and of
28Grant-backs are generally non-exclusive since competition questions could arise where the licensor precludes its licensee from transacting with other firms.
29This statement should be qualified by noting that the patent holder’s commitment to license on fair and reasonable terms does not prohibit it from entering voluntarily into more complex but mutually beneficial licensing arrangements.
limited value if they do not attract very wide participation either by SEP holders or interested implementers, they are rarely created. But they can serve a valuable purpose in setting FRAND rates, providing guidance in structuring FRAND licenses, and potentially addressing patent-stacking issues.30
There are many respects in which the private concerns of members of standards-setting organizations are congruent with the objective of creating standards that enhance economic welfare. SSO members want to develop standards that are effective technical solutions. They want to avoid capture of a standard by one or a few entities, and they want a fair opportunity to participate in setting the standard. These are all consistent with enhancing economic welfare.
However, SSO members do not necessarily internalize all of the relevant effects of their standards choices and IPR policies on economic welfare. SSO members who are implementers may not decisively oppose high royalty payments, even for patents of questionable technical value in the standard, if they can pass these costs onto consumers. Moreover, they may face the threat of the SEP holder obtaining injunctive relief if they do not accede to the latter’s licensing demands.31 High royalties can reduce output and raise product prices, which harm consumers. To be sure, these costs should be set against the dynamic benefits of royalty income. Patents are an important means of realizing returns on R&D investment. Royalties are often used to fund further research that develops better and more efficient products and improved methods of achieving interoperability.
Despite the lack of clear guidance to date in SSO IPR policies, we should not understate the importance of a FRAND commitment and its ability to deal with potential strategic concerns. SSOs are largely effective by virtue of the credibility of the voluntary consensus-based standards process in delivering high quality standards that can be implemented while avoiding serious IPR-related consequences. SSOs remain important and the fact that more are created all the time suggests that market participants know they have much to gain by expending resources in helping develop and implement standards. Indeed, alleged abuses of FRAND commitments have been relatively infrequent in the past relative to the large number of standards and the much larger number of SEPs that standards incorporate. This suggests that most rights holders have not pushed the limits of a FRAND interpretation through the strategic enforcement of their patents.
Within the past few years, however, there have been an increasing number of lawsuits alleging that SEP holders have demanded non-FRAND terms from implementers and used the threat of injunctive relief to try to force these implementers to accept such terms or risk having their product sales halted. This may
30These points were addressed also in Judge Robart’s opinion, cited above at footnote 37.
signal that attempts at patent hold-up are increasing, or it may be that in the past implementers have chosen to accede to unreasonable terms instead of challenging them in courts.32 The current debates at a number of SSOs about whether to further clarify the effect of a FRAND licensing commitment in their IPR policies suggest that industry players have diverse views as to what FRAND means and what constraints it actually places on the SEP holder.
Because of the strategic concerns listed earlier, competition authorities and a number of companies now recommend that SSOs clarify the various effects of a FRAND commitment by formulating certain principles. These principles may include, among other conditions for compliance with FRAND, guidance related to royalty demands that could be a disproportionate share of product value when many patents are necessary to comply with a standard.
The issues discussed in this chapter are highly complex, leaving relatively little room for broadly applicable recommendations. However, there are areas in which the committee sees the potential for improvement relating to SSO policy regarding licensing commitments. Underlying these recommendations is a fundamental principle: The committee believes that a FRAND licensing commitment represents more than the patent owner offering a license on its own terms. A FRAND commitment is also mutual in the sense that both the SEP holder and any prospective licensee are expected to negotiate in good faith towards a license on reasonable terms and conditions that reflect the economic value of the patented technology.
The committee urges SSOs to become more explicit in their IPR policies regarding their understanding of and expectations about FRAND licensing commitments. SSOs should clarify the various effects of a FRAND commitment by formulating certain statements of principle. These principles could include, among other conditions for compliance with FRAND, guidance regarding royalty demands that could be a disproportionate share of product value when many patents are necessary to comply with a standard and the relevant product includes multiple technologies.
The committee recommends that SSOs include in their policies statements that implementers and the consumers of their products and services are the in-
32A further factor in mobile communications may be that a number of large participants contribute relatively little to standards development and may not want to pay for the cost of the innovation underlying those standards.
tended third party beneficiaries of licensing commitments made by SSO participants. Although the enforceability in all courts of such a term may not be guaranteed as the law in this regard is still evolving, inclusion of such statements would inform courts of the intent of SEP owners participating in SSO working groups. It would also provide greater confidence to potential implementers and promote greater certainty in the event of a dispute.
Several recommendations are aimed at improving clarity within SSOs regarding the bundling of licensing commitments.
SSOs should clarify in their policies that prospective licensees may request a license to some or all FRAND-encumbered SEPs owned or controlled by a patent holder. Licensors may not tie the FRAND commitment and the availability of the requested SEPs to a demand that a licensee accept a package or portfolio license that includes non-SEPs or SEPs for unrelated standards. Nor may the licensors tie the FRAND commitment and SEPs availability to a requirement that the licensee agree to license back unrelated SEPs or non-SEPs.
SSOs should clarify in their policies that a holder of FRAND-encumbered SEPs may require a licensee to grant a license in return under FRAND terms to the SEPs it owns or controls (and those of its affiliates as specified in the SSO’s policy) covering the same standard or, as specified by the SSO, related standards.
It should be understood that SSOs’ IPR policies do not affect the freedom of parties to voluntarily enter portfolio or cross licenses beyond the scope of the standard. This includes situations where prospective licensors offer to license SEPs in a package, such as a fixed pool.