Friday, November 29, 2019

40 law and economics professors supporting FTC against Qualcomm's appeal contradict themselves just two pages apart

Last week, the Federal Trade Commission (FTC) filed its answering brief (prior coverage and commentary: 1, 2, 3) to Qualcomm's Ninth Circuit appeal of the agency's antitrust victory in the Northern District of California. This week, amicus curiae briefs in support of the FTC are due, with industry sources expecting a dozen or more submissions, and Professor Jorge Contreras (University of Utah) was first to file (this post continues below the document):

19-11-26 Jorge Contreras Acb by Florian Mueller on Scribd

Professor Contreras is not only a lawyer but also understands technology very well. He's cited all the time, including by some other amicus briefs that have meanwhile been filed in the same case. The first part of Professor Contreras's brief discusses what is also my #1 priority here: chipset-level licensing. After explaining the history of FRAND, which starts with competition law, and other legal aspects, Professor Contreras says that baseband chips are "highly complex" and "embody the principle technical features of the standard." That is, by the way, consistent with what Qualcomm's German outside counsel (then defending Daimler against Nokia) told the Munich I Regional Court last month. In the related footnote (#6), Professor Contreras notes:

"Moreover, it is not clear that a smartphone implements the entirety of the relevant standards either, as Qualcomm seems to argue, given that some functionality described in those standards is implemented in base stations and other central facilities."

At least one other amicus brief I've downloaded by now makes that point as well, and it's too important for a mere footnote. Those monetization-focused SEP holders who refuse to license component makers--Qualcomm, Nokia, Ericsson, and various trolls (though there's only a floating border between former handset makers and trolls)--come up with arbitrary and shifting-sand-style positions on what hardware components are needed in order to implement a standard. For an example, in its German infringement actions against Daimler, Nokia argues that only an end product--in that case, a car--implements a standard, but car makers purchase telematics control units (TCUs) that, in turn, come with connectivity modules (often called network access devices, or NADs), and the NADs actually are like a complete phone, just without a screen, and cars add absolutely nothing that is required to practice the standard.

Even if the debate is about chipsets used in smartphones (as in FTC v. Qualcomm), the phone is an arbitrary choice: as the footnote quoted above notes, only the combination of an entire network (with all its base stations) and the end-user devices would implement the standard if one followed Qualcomm's (or Nokia's or Ericsson's) logic. As a result, no company other than a Huawei or Samsung (which make base stations as well as end user devices) would be entitled to a cellular SEP license--or maybe telcos that operate networks and resell phones could obtain a license, too. Such a nonsensical result would be an invalid outcome that would make it impossible to give any remotely reasonable interpretation to FRAND licensing pledges.

The second part of Professor Contreras's amicus brief explains that Qualcomm's reference to its own past license agreements as a point of reference for determining reasonable (the "R" in "FRAND") royalties is "circular logic." Here, Professor Contreras cites to a publication by Professor Thomas Cotter (University of Minnesota and author of the highly recommended Comparative Patent Remedies blog): Reasonable Royalties, in Patent Remedies and Complex Products: Toward a Global Consensus.

In the third and final part, Professor Contreras takes aim at Qualcomm's "national security" argument. He provides examples of comparable companies that are doing well without Qualcomm-like misconduct:

"[B]ased on publicly-reported 2018 financial information, Intel achieved a profit margin of approximately 62% on net revenue of $70.8 billion, and Broadcom achieved a profit margin of approximately 52% on net revenue of $20.8 billion. [...] Qualcomm, by comparison, reported a profit margin of 55% on revenue of $22.7 billion."

Related to this--and also very interesting--is the comparison of R&D investments:

"In 2018, Intel invested $13.5 billion in R&D (19% of revenue) and Broadcom invested $3.7 billion in R&D (18% of revenue). [...] Qualcomm, by comparison, invested $5.6 billion in R&D (25% of revenue)."

The brief also shows that "Qualcomm is not the global leader in 5G standards or technology development, nor does the U.S. lead in this technology sector." Pointing to an IAM article by IPLytics founder Tim Pohlmann, a table is shown according to which Qualcomm is just #7 in the world in terms of 5G patent families held (with Intel, the only other U.S. company among the top 10, following closely, and those patents were acquired by Apple this summer).

Whatever the reason may be, Professor Contreras filed a separate brief from the one submitted by 40 (precisely twice as many as their colleagues who supported Qualcomm in August) law and economics professors (this post continues below the document):

19-11-27 Law & Econ Pro... by Florian Mueller on Scribd

The passage of the "40 profs" brief I like best is the one that explains the economic dynamics resulting from Qualcomm charging a patent royalty separately from selling its chipsets and making it harder for others to compete with them in the chipset business.

In connection with Qualcomm's "No License-No Chips" policy, the 40 professors address the Supreme Court's linkLine decision which said that a margin squeeze is not enough to prove an antitrust violation: instead you either need to show a duty to deal at the wholesale level or exclusionary conduct (predatory pricing) at the retail level. The FTC's explanations as to why linkLine is inapposite here are very strong, though it obviously doesn't hurt if amici address the same (critical) question, too.

Unfortunately, the O'Melveny & Myers and Hausfeld lawyers who represent the 40 professors made a mistake that the signatories--many of whom I know (a few of them I've even met) and respect, but all of whom are presumably extremely busy--didn't notice...

The brief first argues (on page 14 based on the numbering at the bottom of each page; or page 20 of the PDF) that Qualcomm can't engage in a price squeeze affecting chipset manufacturers because it doesn't license or assert patents against them:

"As to form, the input here is the license to Qualcomm’s SEPs, and the non-integrated competitors are the rival chipset manufacturers. Because Qualcomm refuses to license chipset manufacturers, it is not squeezing them with a higher license fee."

But then, only two pages later, comes footnote #14, which flatly contradicts that formalistic approach:

"Qualcomm argues that the [No License-No Chips] policy is not anticompetitive because the cost is not levied directly on the competing chipset makers. [...] But as demonstrated by the Microsoft "per processor" royalty cases, there is no requirement that a monopolist impose costs directly on its competitor. [...] What is significant is that the monopolist imposes a charge on the transaction involving the competitor." (emphasis added)

One mistake doesn't devalue an entire (otherwise very strong) amicus brief. I agree with the 40 professors that the FTC's win should be affirmed; I largely agree with their reasoning. And again, the professors themselves presumably just lacked the time to identify the flaw highlighted above before they signed. It changes nothing about my respect for all of them. While I do want Qualcomm to be required to license rival chipset makers and to stop its "No License-No Chips" policy, there are cases where Qualcomm is right and/or Qualcomm's adversaries are wrong. I commented favorably on parts of Qualcomm's motion to dismiss (2007), I criticized the methodology used by one of the FTC's experts, and I pointed out months ago that it isn't easy to "shoehorn" Qualcomm's refusal to license rival chipset makers into the Aspen Skiing pattern. Now I believe the professors' footnote #14 is right (yes, the focus should be on the economic effect of a charge as opposed to just form), but then Qualcomm, too, is entitled to the benefit of that approach in the linkLine context.

No matter how hard I try to find an element that reasonably sets one context apart from the other (the wider context is actually the same: No License-No Chips), I can't find one. Either one wants to argue that indirectly-imposed costs count, or one doesn't.

Expect more posts on FTC v. Qualcomm amicus briefs in the days ahead.

Share with other professionals via LinkedIn: