Sunday, May 7, 2023

Disclaimers, excuses, evasive answers: EU SEP Regulation lacks evidentiary basis, Brussels event primarily about GROWing IPlytics' business

The proposed SEP Regulation that the EU's internal market commissioner Thierry Breton presented a little over a week ago continues to disappoint in many ways. The official presentation of two "empirical studies" on Thursday (May 4) was no exception. The meeting room in Brussels was sparsely filled, with panelists probably outnumbering the audience, and they claimed to have about 100 online participants, while the webinars held by the European Commission two years ago to discuss SEP policy were typically watched by approximately 300 people (each).

If there is any company whose business the DG GROW initiative is really designed to grow, it's neither an innovator nor an implementer of standards (neither side likes the proposal): it's just IPlytics, a service provider. A study paid for by the European Commission should be presented in a company-neutral setting, but that is apparently too much to ask for. What DG GROW did here is an example of socializing costs and privatizing profits. They used taxpayers' money to fund a (low-quality) study, but then allowed IPlytics and its parent company (LexisNexis) to convert the official presentation--which was closely related to the Commission's presentation of the legislative proposal--into a customer acquisition event.

Not only was IPlytics allowed to organize the presentation as a corporate event, but in order to register (even if only for online access), one had to provide far more data than a governmental entity would request: "organization type" and answers to the questions of whether one currently has "access to SEPs, technical standard and contribution data" as well as one's "biggest challenge when it comes to SEPs, technical standard and contribution data" had to be entered in mandatory input fields. Those answers have nothing to do with whether one has a legitimate reason for watching the presentation of two studies funded by European taxpayers. They're simply about maximizing the value of the sales leads IPlytics' salespersons can then follow up on.

Until I officially complained about it (which complaint has not been satisfactorily resolved and which I will soon refer to the European Ombudsman; even the Court of Auditors may ultimately be interested in this), the website even featured the European Commission logo next to that of the two corporate hosts (click on the image to enlarge):

The fact that the EC logo was removed on the business day following my complaint is insufficient. The input fields that served no other purpose than to optimize data for IPlytics' customer acquisition purposes remained mandatory. And there is no indication of the EC having taken action against an unauthorized use of its logo by LexisNexis and IPlytics.

The more fundamental problem is the advice that the EC apparently received. For IPlytics' business, it's a boon if the EU mandates top-down valuations of SEP portfolios based on purely statistical evaluations of data (coupled with individual low-reliability assessments of essentiality). They don't care if it means that quantity is prioritized over quality, numbers over value. "Quality" is not just a question of dubious essentiality checks or essentiality ratio estimates, but of the specific value of the technologies contributed by a given patent holder.

A top-down mandate is even more beneficial to IPlytics if every time a SEP holder is preparing or merely contemplating litigation, they firstly have to start a "FRAND conciliation" proceeding. And what data source would be used in those proceedings? IPlytics would tell them all that IPlytics itself advised the Commission on this legislative proposal, and that IPlytics is the data source referenced in the Commission's own impact assessment and was awarded the contract:

The contract for the study was awarded to iplytics GmbH after a call for proposals, based upon a proposal by a consortium consisting of lead researcher Dr. Justus Baron (Northwestern University), Dr. Tim Pohlmann (iplytics and TU Berlin), Dr. Pere Arque-Castells (University of Groningen), Dr. Amandine Leonard (University of Edinburgh), and Dr. Eric Sergheraert (University of Lille and darts-ip).

While the essentiality checks study names only Justus Baron as its author, it acknowledges that "[t]he five members of the consortium have participated in discussions regarding the definition of policy options." This means IPlytics founder Tim Pohlmann can't disclaim his involvement with one of the studies. His company was awarded the contract; all consortium members defined the options; and the paper on essentiality checks cannot be separated from the "impact assessment" study that builds on it. But Mr. Pohlmann also likes to speak out of both sides of his mouth when his company releases "rankings" of top patent holders in a field of technology: when responding to criticism, he says it's all just about showcasing what one can do with the IPlytics tool, but the headlines and the announcements are all about leveraging the rankings to generate attention and, as always, sales leads.

Mr. Pohlmann's welcome address on Thursday was just a disclaimer. He said that their "empirical" study had to rely on publicly available data, but acknowledged that SEP licensing and litigation are largely non-public, only to then tell stakeholders they could "contribute further evidence to the Commission now." Seriously? Contribute evidence after the legislative proposal has been made? At a time when the EC will still play a coordinating role, but the actual decision rests with the two institutions that are the EU's co-legislators (Council and Parliament)? That makes no sense.

Qualcomm's Fabian Gonell later asked what companies like his could do next time in order to ensure that their input would be considered while still of use. Obviously, it was a diplomatic means of questioning the process that led to the April 27 proposal, but the panelists were merely being evasive even when he underscored his question was just about future situations.

The quality of the IPlytics-led consortium's work leaves a lot to be desired.

Mr. Pohlmann said they had been working on their studies for 1.5 years. Apparently that was not enough time to fix countless typos, one of the funniest ones is this (click on the image to enlarge):

It's also sloppy to switch between a Times Roman and a Sans Serif font within a paragraph or between subsections (click on the image to enlarge):

On those documents, by the way, the European Commission logo is still found. But then the official legislative proposal and the EC's impact assessment also come with typos--and they didn't even fix obvious factual errors such as the ones I discussed in a post on what the proposal and the impact assessment say about patent pools. Whoever put out that proposal either set a low quality standard for their own work and that of the "researchers" and/or was under incredible pressure to rush out something half-baked and ill-conceived. This is not the right approach to policy making in the absence of a crisis--and in fact, the documents that the Commission put out during the pandemic were of far higher quality though there must have been huge pressure and objective time constraints.

There are serious substantive issues that call into question the "consortium's" collective competence.

I'll discuss the various errors and shortcomings in different posts. There are so many issues that I'll content myself with only one salient example here, the first paragraph of subsection 5.2.2 of the "impact assessment study":

"Although this has not yet been discussed by courts, the approach of distinguishing between FRAND licensing and pure questions of patent law could also apply to the question of the determination of essentiality. Whether a patent is actually essential to a standard is a technical question. It is not a legal determination coming from patent law."

That's simply wrong. As I explained in my previous post on the proposed EU SEP Regulation, an essentiality determination is an infringement analysis based on a specification of a standard as opposed to an accused product, and it does depend on legal questions.

In the United States as well as in Germany, claim construction is resolved by courts as a question of law (not fact). In rare circumstances, courts may need expert testimony on how a person of ordinary skill in the art would have understood a certain term at the time the patent application was filed. Even then, the ultimate determination is a legal one.

The answer of whether a patent claim reads on a standard is not just one of whether the claim language at first sight covers a part of a standard. The claim language has to be interpreted in accordance with applicable laws, and in light of the specification. I've attended numerous SEP infringement trials and analyzed lots of documents from U.S. infringement lawsuits and ITC investigations. Rarely does the language of a patent claim mirror that of the specification of the standard. The patent applications must be filed before a specification or a proposal for one is published. There are almost always some deviations that result from the work done in the standard-setting groups.

How could that IPlytics-led "consortium" of five "researchers" get things so wrong?

It's instructive to look at the background of the members--which DG GROW should have viewed far more critically.

I'm sorry for Justus Baron, who used to enjoy a very good reputation for papers he had previously authored. He should never have joined that group, but now it's too late.

Mr. Baron is one of three economists. The others are Mr. Pohlmann--who is simply pursuing his company's business development interests and did not state an affiliation with a university (but just the IPlytics brand) at the Thursday event--and Pere Arque-Castells (University of Groningen). But why didn't the two jurists then explain to everyone that essentiality determinations do involve questions of law?

  • Eric Sergheraert points to his affiliation with the University of Lille, but primarily he's the "Global Patent Manager" of Darts-ip, a litigation data service. The benefits to Darts-ip from the EC's proposal are less clear than the ones to IPlytics. But Darts-ip is an IPlytics partner, so at least there is an indirect benefit if IPlytics' business grows thanks to DG GROW.

    It's hard to find any academic write-ups by Mr. Sergheraert. He's not present on SSRN, and on I just found papers on life sciences patents, none of them related to claim construction or infringement analysis.

  • Then there's Amandine Leonard, an "Early Career Fellow in Intellectual Property Law" and lecturer at Edinburgh University. She was not present at the Thursday event. The research output I found on her university's website suggests she has more of a policy focus ("trolls", compulsory licensing, FRAND, abuse of rights).

There are gaping holes in DG GROW's justification of the proposal.

The "researchers" (in no small part: business developers) gave input to the EC on potential measures, but then wrote their two "studies" within the parameters given to them by the Commission. However, ultimately the Commission has to take responsibility, and it can't hide behind "studies" that don't look at the issues holistically.

Mr. Baron can certainly take the position that "random errors" in essentiality checks are of no concern to him--but every single error that stigmatizes a patent is an injustice that cannot be accepted from a legal point of view. The right of appeal cannot be negated purely based on cost considerations. If Mr. Baron wants to conduct a study that arrives at an overall essentiality ratio for SEPs across all standards, and if they don't publish the results of individual essentiality checks, then they may decide to just accept a high error rate. No particular company--or group of companies (such as all contributors to a specific standard such as 5G)--would be prejudiced.

But every single patent is a property right, and the EU can't just diminish the commercial value of such a property right without due process. Keeping the names of the "assessors" confidential even after they've done their work, and not allowing full-blown appeals including the right of recourse to a court of law, is unacceptable.

Mr. Baron's paper on essentiality checks--contrary to what DG GROW makes of it--even mentions "the analysis by [Keith] Mallinson (2022), who calculates that sample sizes of several thousand patents may be needed to produce reliable and precise estimates of essentiality rates in different portfolios." But the proposal (including the impact assessment and the "studies") does not envision such sample sizes. So there is built-in injustice that cannot be reconciled with fundamental rights in the EU, under the European Human Rights Convention, and the national constitutions of EU member states that protect intellectual property rights like other property.

Let's look at it this way: it would be a severe violation of individual rights if the EU created and maintained a "land contamination" register and declared various pieces of property contaminated just based on an error-prone rapid test. The value of the false positives would be diminished massively.

In his study, Mr. Baron mentions standard confidence levels (the most common one, which is 95%, as well as 90% and 99%). But that is not a determination for him and the business developers on his consortium to decide. If the value of a patent portfolio held by a major SEP holders like Qualcomm, Ericsson, or Nokia was diminished by only 1%, the commercial damage would be significant.

DG GROW allowed a company with clear business interests to put together a "consortium" and make proposals that will benefit that company while potentially harming many others. Economists can make any efficiency argument they want, but ultimately we're talking about justice in each invididual case, with respect to every single intellectual property right.

The Commission limited its policy options beforehand. And the "business developers" had no incentive to explain to DG GROW that the range of choices was too limited.

The Commission could have decided to enshrine Huawei v. ZTE in an EU directive, with greater clarity. In the end, what they are concerned about is the leverage patentees get from (particularly German) injunctions. If that's the concern, then the answer would be a clearer set of rules, as opposed to bureaucracy that offers no tangible benefit but has the potential to do serious harm to individual patent holders.

I suspect that DG GROW just didn't dare to codify Huawei v. ZTE for lack of competence. It would obviously be much more difficult to understand the related options and to craft legislation than what they're now proposing instead (and they didn't even get that one right for lack of understanding how patent litigation and licensing work).

The proposal that is on the table now is ill-conceived, and part of the reason is that the Commission relied on a "consortium" led by a company with business development interests instead of proper consultation with stakeholders on specific ideas.