This working paper, which can be found hereseeks to understand what happens when standard essential patents (SEPs) are litigated in court. The authors find that, contrary to expectations, courts are more likely to find that SEPs are valid patents than a matched set of litigated non-SEP patents. However, courts are also significantly less likely to find that SEPs were infringed. One of the reason for this seems to be that many SEPs are asserted in court by non-practicing entities (NPEs), and NPEs do much worse in court than other patent holders.

This has interesting implications for policy debates about both SEPs and NPEs. Standard-essential patents may not be so essential after all, perhaps because companies tend to err on the size of over-disclosing patents to standard-setting organisations. On the other hand, the failure of NPEs to win cases concerning the validity of what are, prima facie, a strong set of patents raises interesting questions about the role that NPEs play in patent law and about the policy efforts to curb patent litigation abuse.

PAtent Licensing

The paper is structured as follows:

Part I discusses the background to SEPs. I have described this multiple times now, so you may want to skip this as there is really nothing new in this section.

Industry standards are critical to major sectors of the market economy. The computer, Internet, and telecommunications industries in particular depend on standards to ensure that different products work with each other. When a patent is truly essential to a standard, there is no way to design around it and still comply with that standard. As a result, ownership of an essential patent by a company can interfere with the standard’s widespread adoption. Standard Setting Organisations (SSOs) have responded by limiting the use of patents that cover the standards they adopt. While some SSOs require royalty-free licensing of patents that cover a standard, others simply require disclosure of the existence of those patents. Most commonly, SSOs require patent owners to license their SEPs on Fair, Reasonable and Non-Discriminatory (FRAND) terms to anyone who adopts the standards.

Owning the right to be paid a royalty every time a standard is used – even a fair, reasonable, and non-discriminatory price – can be quite valuable for SEP holders. At the same time, if the patentee does not agree to (or seeks to evade) a FRAND commitment, the possibility of an injunction against a technology everyone has to adopt can be quite powerful. Licensing SEPs is made more complicated by the fact that products can incorporate many standards, and most standards are covered by multiple SEPs. If each of those patents is truly essential, there is a risk of double-marginalization or “royalty stacking” because each patent owner can demand a disproportionate share of the revenue from the product. A number of scholars have worried about the risk of patent holdup that could result.

Given the value of SEPs, it is no surprise that SEPs are much more likely to be enforced in court than other kinds of patents – SEPs are five times as likely to be litigated as comparable non-SEPs. When those patents are enforced, however, virtually everything about the FRAND commitment has proven to be controversial. Litigants and scholars have fought about whether a FRAND commitment prevents a patentee from getting an injunction, whether the fact that a patent is standard-essential should bar an injunction even if there is no FRAND commitment, whether a patentee that makes a FRAND commitment must offer it to everyone or only willing licensees, who is a willing licensee, whether the FRAND commitment is an enforceable contract, who decides what royalties are FRAND, what royalties are FRAND, and what are the consequences of reneging on a FRAND commitment. Given this outpouring of litigation and scholarship, we actually know surprisingly little about the enforcement of SEPs.

Part II explains the method adopted in this study.

The authors set out to understand how SEPs fared in court. To evaluate this hypothesis, the authors collected data on lawsuits concerning essential patents accepted by thirteen SSOs. The authors identified 422 patents that have been asserted in at least one case, but focused on the 355 patents where there was complete data on the outcome of at least one case. Those 355 patents were litigated in 537 unique cases. Since in many of those cases more than one SEP was asserted, they arrived at a total of 1,446 SEP assertions. The overwhelming majority of the SEPs (95%) in the sample were subject to a FRAND commitment. The authors then matched each SEP patent-case pair to a randomly-selected non-SEP patent from the same patent class that was filed in the same year and first-litigated in the same year as its SEP “twin.” The authors also determined, for each patent, whether it was asserted by a practicing entity or an NPE, based on the entity status coded by the Stanford NPE Licensing Database.

Part III presents the results.

27.2% of all patent assertions were made by practicing entities, and 72.8% by NPEs. NPE shares of patent assertions drop dramatically if one counts only unique patents: NPEs assert less than half of the patents, and only 37.6% of the SEPs in the sample. In other words, NPEs make intensive use of the patents they acquire, asserting them in more than three times as many cases as product companies do.

Considering all the patent-case pairs in the sample, patentees won 41.7% of cases, with no significant difference between SEP and non-SEP win rates. Even varying the analysis to control for potential skews in the sample, differences between SEPs and non-SEPs were not statistically significant: SEPs are no more likely to be found infringed than non-SEPs. That itself is a very interesting finding. One would expect SEPs to be stronger than non-SEPs, but that does not turn out to be true in cases litigated to judgment. At the same time, the SEP validity win rate was 83.7%, significantly higher than the non-SEP validity win rate of 60.8%. Thus, surprisingly, SEPs do no better in infringement cases than their matched non-SEP counterparts, but are less likely to be invalid.

The most significant finding of this study is that there are substantial differences in the rate of success of patent assertions depending on whether the patent holder is a practicing or a non-practicing entity. Practising entities in this study win their cases at more than twice the rate of NPEs. That is consistent with other work that shows that NPEs fare worse in litigation than other plaintiffs generally. Furthermore, while practicing entities win virtually the same percentage of their SEP and non-SEP cases, NPEs win many fewer cases and are particularly unlikely to win their SEP cases, winning only 6% of them.

In short, this study has produced at least three interesting findings. First, despite their name, SEPs do not seem to be all that essential or, at least, are not often found to have been infringed by standard implementation. Second, when SEPs go to court they do not fare significantly differently than other patents of similar age and type. Third, NPEs do very poorly even when they assert SEP patents.

Part IV discusses some implications of these results.

One implication is that over-disclosure of SEPs seems to be rampant: when SEPs are asserted in court, most of them turn out not to have been infringed. This indicates that many SEPs are not in fact essential – which would be in line with prior evidence that suggests that over-disclosure of SEPS is common.

Over-disclosure increases the cost for implementers of figuring out what licenses they need. Over-disclosure might also distort the true cost of a standard, making it appear more encumbered than it is. This risk is compounded if courts use counts of declared essential patents to apportion royalties in damages calculations, as some have done. Another issue raised by strategic nondisclosure is that it permits a patent owner to lure an SSO into adopting a standard without understanding the full costs of that standard – thereby facilitating patent ambushes once implementers are locked into an industry standard. On the other hand, over-disclosure of patents can be beneficial, since most SEPs are subject to generous licensing requirements.

A last implication of this study is that who owns the patent matters, in surprising ways. NPEs that do not participate in SSOs have the advantage that their patents are not burdened by FRAND licensing requirements. As a result, when NPEs buy patents subject to a FRAND requirement, one might expect the NPE to take on that burden because it thinks it is getting better patents. However, NPEs do poorly even with what seem like they should be strong patents – and they do even worse with SPEs. Why do NPEs do so poorly? The authors do not know the answer, but can envision several possible explanations. First, it might be that the quality of the patents NPEs asserted in the paper was dramatically worse than the patents asserted by practicing entities – but this is unlikely, as the authors tried to control for this. Second, perhaps only NPEs systematically over-disclose their patents to SSOs. The authors think that this is, at best, only a partial explanation – after all, many patents have been accepted to be SEPs, and they are supposed to be essential. Third, perhaps there is something about the nature of NPEs that makes them less likely to win. That could conceivably be a matter of business strategy. If a NPE wants to settle its case for money, perhaps it settles the good cases and only ends up going to judgment if its case is so bad that no one will pay it. Alternatively, maybe their lawyers are not as good or have as many resources – NPEs are more likely to be represented by solo practitioners or small law firms. Lastly, the authors venture that maybe courts just do not like NPEs.


I found this working paper – and I must emphasise that it is only that – quite interesting. From a competition standpoint, I take two main points. Firstly, the mechanism for selecting essential patents is not perfect, and can add to the risk of patent holdup (and, thus, to competition concerns regarding standard licensing). Secondly, non-practising entities seem to impose a burden on the justice system that exceeds that of practising entities, and hence may be detrimental to consumer welfare.

This, of course, says nothing about whether antitrust intervention may be appropriate, particularly against NPEs – a matter which is discussed in greater detail in the paper reviewed below.

Author Socials A weekly email with competition/antitrust updates. All opinions are mine

What do you think?

Note: Your email address will not be published