This paper – which can be found here – slooks at whether it is “unfair” (and therefore unlawful) for a SEP holder to license its patents, including both SEPs and non-SEPs, on a portfolio basis.
The basic argument is that “SEP and non-SEP portfolio theories of harm” advanced thus far are not economically sound (this is unsurprising, given that the very first sentence of the article is: “Competition agencies around the world have taken the unwarranted position that antitrust enforcement involving standard-essential patents (SEPs) upon which a patent holder has made an assurance to license on fair, reasonable, and nondiscriminatory (FRAND) terms should be subject to special rules or unique presumptions and burdens of proof” (emphasis is mine)). Instead, for these portfolios to infringe competition law there: “would have to be an exclusionary effect on other sellers because bundling thwarts buyers’ desire to purchase substitutes for one or more of the goods in the bundle from those other sellers to an extent that harms competition in the markets for those products.”
The paper is structured as follows:
- The introduction looks at the logic of applying antitrust law to the licensing of SEPs. This application is premised on the assumption that: “Standard setting involves allowing competitors to get together to pick winners, an act of collusion generally prohibited under most antitrust laws. However, because of the significant procompetitive benefits from standard setting, competition agencies permit this collusion. They instead justify special rules or heightened intervention for SEP holders based on the SEP holders’ ex post unilateral conduct, i.e., conduct following the alleged conferral of market power flowing from the standardization process.” These assumptions are perceived to be flawed on two different grounds: (i) empirical research suggests that standardization does not automatically confer market power, so intervention is not automatically justified; (ii) thus, the issue is whether particular conduct involving SEPs, including evasion of a FRAND assurance, has net anticompetitive effects – which should ae same case-by-case, fact-specific analysis.
The logic of prohibiting the tying and bundling of SEP and non-SEP patents (i.e. patent portfolios) is said to be a result of the flawed logic currently underpinning the antitrust treatment of SEPs. This logic leads to the development of two basic theories of harm, namely that the SEP holder is either: (1) unfairly forcing implementers to license more than they desire or (2) evading its FRAND assurance through package licensing. However, absent exclusion, these theories boil down to a complaint about excessive pricing (or breach of contract).
- Section II rebuts the notion that there should be special rules for antitrust matters involving SEPs, explaining the need for a traditional effects-based analysis in such cases. After reviewing US (lack of) practice and academic debate on this matter, it reiterates the views expressed in the introduction about standardisation not creating market power, and about the need to assess anticompetitive effects on a case-by-case basis. The discussion then moves in Section III into the market power of SEP holders – starting from the somewhat trite observation that holding an IP right is not equivalent to market power for antitrust purposes. More to the point, it is also observed that “even restricting the analysis to truly essential patents, one cannot perfunctorily conclude that an individual SEP or a portfolio of SEPs constitutes a well-defined relevant market or that the owner possesses market power” (…) “Genuinely essential patents are perfect complements (i.e., like nuts and bolts), which creates a connection among patents and patent holders such that SEPs cannot be licensed in isolation.”
- Section IV then provides an overview of US (and some foreign) law on tying and bundling. In short, it explains that antitrust law permits the extraction of surplus by lawful monopolists and seeks to intervene only against exclusionary or predatory conduct that results in the unlawful acquisition or maintenance of market power. As a result, except in very limited cases, “the alleged evasion of a FRAND assurance through tying or bundling (while perhaps a contractual issue) is not an antitrust issue, at least under U.S. law.”
- Section V then looks at situations where trying and bundling can infringe antitrust law. It begins by identifying efficiencies and legitimate business justifications for tying and bundling. In particular: “By offering only a bundle of patents, rather than a la carte patents, patent holders are likely to reduce administrative and monitoring costs associated with identifying patent infringement, since they would not need to track the individual licenses for each licensee. In other words, there are significant costs of prohibiting tying or bundling, including the potentially cost-prohibitive transaction costs to both patent owners and implementers if they were forced to analyze each and every patent in a large portfolio and negotiate a license on each.” Furthermore: “By combining products into a bundle, say a combined SEP and non-SEP portfolio, customers avoid the costs of having to sift through large patent portfolios to identify all relevant patents, determine the essentiality of each patent, and then individually purchase each patent necessary for selling their product.” In short, “because of the widespread procompetitive uses of tying and bundling, differentiating between procompetitive and anticompetitive uses of tying and bundling can be incredibly difficult.”
At this point, the authors identify some forth economically-sound theories of harm concerning tying and bundling, namely:
- Leveraging – the authors review the literature and conclude that this can happen and have anticompetitive effects. However, “The intuition that the objective of tying is to leverage market power from one good to another usually fails to withstand rigorous economic scrutiny.” Furthermore, they find that the two conditions under which the ICN considers that leveraging can expand a firm’s market power (i.e. (1) “if the tied product has uses unrelated to the tying product” and (2) “if the tying product is durable and upgrades are important”)) are unlikely to apply to SEPs. Even if these conditions are met, however, they find that even exclusionary conduct arising from this can have an uncertain effect on welfare.
- Monopoly maintenance – which posits that tying may be used to preserve an insecure monopoly in the tying product by committing to the bundle and the implicitly low price. Such conduct can deter a competing producer of the complementary product from obtaining enough sales to think that entering the tying (primary) market is worthwhile, thereby preserving an insecure monopoly.[NB: this is basically the theory underpinning the Microsoft internet browser case]. The authors hold, however, that: “Even if the offer of the vertically integrated SEP holder could de facto be its conduct will not lead to the foreclosure of the component market if (i) the vertically integrated SEP holder does not assert its patents at the component level, and (ii) it licenses its SEP portfolio to end-device manufacturers on FRAND terms irrespective of whether they source components from its own subsidiary or from the non-integrated rival.”
- Section VI concludes that: “Courts and competition enforcers should reject pleas to impose special rules or burdens of proof for conduct involving SEPs as based on the fundamentally flawed notion that standardization necessarily confers market power, which must be constrained through imposing competition law sanctions for evasion of a FRAND assurance even in the absence of evidence of net harm to the competitive process.” Instead, we should just apply straightforward, time-tested antitrust principles
This is a balanced and thought-through paper, with extensive discussions of a number of applicable theories of harm. I enjoyed it quite a lot – and I agree that bundling SEP and non-SEP patents should not be a per se infraction. It is clear that portfolio licensing is industry practice and there is a risk of form-based prohibitions preventing this potentially pro-competitive practice.
However, I have some issues with the arguments advanced in this paper. First, the argument/assertion that antitrust should be applied as in every other sector based on a case-by-case analysis seems to miss the point. Without wanting to endorse any of the approaches currently adopted in different jurisdictions: (i) the existence of “essential” standards could very well provide a basis for a presumption that SEP holder’s possess some type of market power. If so, it could be reasonable to elaborate enforcement rules (e.g. burdens of proof, etc.) based on that presumption; (ii) similarly, if it is indeed found that the SEP holder has market power (which, as the authors correctly point out, is an issue complicated by the fact that essentiality is usually self-reported to a standards-setting entity), it may make sense to set some standard of behaviour which breach would infringe upon antitrust rules unless objectively justified. In other words, it may make sense to apply some specific rules to the antitrust assessment of SEP / FRAND cases, even when applying traditional antitrust concepts.
I also think that the authors fail to address the main problem when applying antitrust to SEP licensing – namely, at which point does non-FRAND behaviour become exclusionary (or, in certain exceptional circumstances, anticompetitive despite not being exclusionary)? In short, it is not necessarily true that the portfolio theories of harm amount to excessive pricing and, hence, should not be subject to antitrust scrutiny. Instead, while adequate theories of harm should be applied to identify antitrust harms, it is possible for non-FRAND negotiations and terms to have de facto exclusionary effects in the downstream market – something that the authors’ themselves recognise as having the potential to be anticompetitive .
A last point is that I am slightly disappointed by the authors’ failure to engage with the traditional theories of harm regarding FRAND licensing (namely, patent holdup and holdout), and how they may apply in the context of portfolios of patents.
Having said all this, and as should be obvious by the long analysis, this is a very interesting paper that is well worth a read.