Monopoly, Innovation, and Due Process: FTC v. Qualcomm and the Imperative to Destroy

Order a reprint of this 774-page article on Amazon.com

Federal Trade Commission v. Qualcomm Inc., 411 F. Supp. 3d 658 (N.D. Cal. 2019), rev’d, 969 F.3d 974 (9^th Cir. 2020), is the most consequential government monopolization case since Microsoft. Many articles, white papers, and amicus briefs have already been written about FTC v. Qualcomm, as befits a case of such significance. Most of those writings discuss the issues at the level of policy or theory. Abstractions fly back and forth like a shuttlecock in a game of badminton. Virtually none of these commentators digs into the actual evidentiary record of the case. As a consequence, little of the public commentary on FTC v. Qualcomm discerns and engages, at the level of proof rather than theory, the basis for Judge Lucy Koh’s findings of fact and conclusions of law.

My approach in this article is qualitatively different. I conduct in 774 pages a de novo review of the publicly available trial record in FTC v. Qualcomm. I have done so in part to demonstrate to any jurist interested in the question that the task is entirely possible to accomplish, and in part to show why due process of law requires it. I have read in their entirety the 11 volumes of the public (redacted) trial transcript containing the testimony of the 27 witnesses who testified at trial, as well as the public (redacted) deposition designations of the 22 other witnesses whose testimony the parties replayed by video at trial. My analysis reveals Judge Koh’s errors of fact concerning monopoly, innovation, and due process—as well as her errors of law, many but not all of which the Ninth Circuit identified.

On May 21, 2019, Judge Koh issued her findings of fact and conclusions of law in the FTC’s monopolization case against Qualcomm, a leading U.S contributor to cellular standards and a leading developer and supplier of modems used in smartphones and other mobile devices. Judge Koh found that Qualcomm had engaged in a series of interrelated practices between 2006 and 2016 that collectively harmed competition in what she defined as two relevant antitrust markets for different versions of such modems.

Upon finding a violation of both section 1 and section 2 of the Sherman Act, Judge Koh ordered Qualcomm to submit to a permanent, worldwide injunction that would require, among other things, that the company renegotiate its patent-license agreements (which number in the hundreds and span several decades) and offer exhaustive licenses to its portfolio of cellular standard-essential patents (SEPs) to rival modem manufacturers (rather than solely to original equipment manufacturers (OEMs), which make mobile devices such as smartphones). Judge Koh also prohibited Qualcomm from limiting the sale of its modems only to OEMs that have first taken a license to Qualcomm’s cellular SEPs. Judge Koh’s permanent, worldwide injunction would run at least seven years. Yet, even if one assumes for sake of argument (but contrary to fact) that the FTC’s theories of liability were well founded, Judge Koh’s findings of fact and conclusions of law did not support her imposition of a permanent, worldwide injunction.

In broad terms, three substantive errors invalidated Judge Koh’s findings of fact and conclusions of law. Those errors necessarily invalidated the reasoning for the permanent, worldwide injunction that she imposed on Qualcomm.

First, Judge Koh’s exercise of discretion over fact finding in the bench trial in FTC v. Qualcomm denied Qualcomm due process of law. Judge Koh seemed to adopt, without attribution, the opinions of the FTC’s expert economic witness, Professor Carl Shapiro. He testified at trial that Qualcomm had market power in two relevant antitrust markets and that a theoretical bargaining model revealed how Qualcomm’s challenged practices supposedly harmed the competitive process. By presenting answers to complex economic questions as being based solely on her own (undefined) analysis (rather than having been influenced in any way by the voluminous expert testimony of Professor Shapiro), Judge Koh avoided having to analyze any of Professor Shapiro’s testimony in her findings of fact and conclusions of law. Most significantly, Judge Koh avoided the duty to reconcile her findings and conclusions with the testimony of Professor Shapiro that materially tended to exculpate Qualcomm because his testimony was shown on cross examination to expose the implausibility of the FTC’s theory of competitive harm. Consequently, Judge Koh also impaired the ability of the Ninth Circuit and the Supreme Court to examine complex questions of economic fact addressed in Professor Shapiro’s testimony. In violation of Rule 52(a)(1) of the Federal Rules of Civil Procedure, Judge Koh failed to make specific findings of fact—in particular, findings concerning the plausibility or implausibility of Professor Shapiro’s theoretical bargaining model, which appeared, from his public testimony and from the portions of his expert reports that are in the public domain, to resemble the model upon which he had predicated testimony that the D.C. Circuit had found unreliable several months earlier in the litigation over AT&T’s acquisition of TimeWarner, United States v. AT&T. Finally, Judge Koh gave no weight to the testimony of many of Qualcomm’s witnesses (including all of its expert economic witnesses), yet her reasons for doing so were unpersuasive and at times contradicted by substantial evidence in the record.

Second, Judge Koh’s findings of fact (irrespective of whether they were based on a complete or truncated evidentiary record) had virtually no support in the evidence that she identified. Judge Koh’s findings of fact that Qualcomm was charging OEMs “unreasonably high” royalties for an exhaustive license to its cellular SEPs, that Qualcomm squeezed its rivals’ margins on modems, and that Qualcomm’s practices reduced the sales of those rivals all lacked a factual foundation. She chose to place heavy weight on documentary evidence from the past and discount trial testimony (particularly when Qualcomm witnesses testified). Yet, the recurring problem with Judge Koh’s analysis of documentary evidence was her failure to comprehend, and then explain in her findings of fact and conclusions of law, the proper context for interpreting that documentary evidence. Judge Koh misconstrued the transcript of Qualcomm’s meeting with the Internal Revenue Service in July 2012, which explained why Qualcomm’s device-level licensing of its cellular SEPs was substantially more profitable than component-level licensing. She also misconstrued the evolution of the doctrine of patent exhaustion in the Federal Circuit and the Supreme Court since the 2000s. She did not ask basic questions about the provenance and use of a purported smoking gun—an expurgated slide from November 2009 containing draft strategy recommendations emailed by a non‑executive-level Qualcomm employee. Of Qualcomm’s hundreds of patent-licensing agreements, Judge Koh focused on a single agreement, between Qualcomm and LGE in 2004, in which Qualcomm had charged asymmetric royalties for devices equipped with Qualcomm modems versus non-Qualcomm modems. Yet, she made that solitary occurrence the ostensible rule rather than the exception for Qualcomm’s supposed course of dealing with OEMs. Judge Koh rejected as “pretextual” that Qualcomm made relationship-specific investments to design and supply thin modems for Apple, yet she concluded that an entry barrier exists in the modem industry because of the need for entrants to make large sunk investments to supply modems to smartphone OEMs. From her specious analysis of these key pieces of evidence, which she repeatedly emphasized, Judge Koh inferred anticompetitive intent, disparaged Qualcomm witnesses, and made legally erroneous findings concerning patent exhaustion, among other issues. Judge Koh’s conclusion that Qualcomm’s licensing practices had harmed the competitive process was not a proven fact; it was, instead, strictly her a priori conjecture that violated established economic principles used by courts in antitrust cases to analyze markets and competition. Judge Koh never analyzed actual empirical evidence about competition in what she defined as the two relevant product markets for modems. Publicly available evidence contradicted her conclusion that Qualcomm’s conduct had harmed competition in those product markets. Judge Koh also rejected the proposition that the FTC bore the burden of production and the burden of persuasion to prove a reasonably proximate causal nexus between Qualcomm’s supposedly anticompetitive acts and the supposed harm that its rivals experienced. She disregarded preponderant evidence establishing the technical superiority of Qualcomm’s modems and relied instead on the self-interested testimony of rival modem manufacturers to conclude that Qualcomm’s acts “promoted” the “exit” of certain rivals from the industry or produced some other supposedly anticompetitive effect. Judge Koh’s findings of fact and conclusions of law enforced what is best termed the “Intel-welfare” standard—instead of the consumer-welfare standard—for interpreting the Sherman Act. In short, Judge Koh’s findings of fact were unsupported by substantial evidence and in many instances were demonstrably incorrect.

Third, Judge Koh misapplied or ignored controlling Supreme Court precedent. She found Qualcomm’s practices unlawful because they reduced rivals’ margins, even though the Supreme Court held in linkLine that a firm has no duty to price its products so as to preserve its rivals’ profit margins. Judge Koh ignored the Supreme Court’s call for caution in condemning low prices when she concluded that the reduction in the price of modems that Qualcomm offered to Apple in exchange for (alleged) exclusivity was anticompetitive. She reached that conclusion without performing any of the analysis that Tampa Electric requires to determine whether Qualcomm’s payments to customers (which is to say, Qualcomm’s price reductions to Apple and other smartphone manufacturers) foreclosed so many sales within the relevant market for modems in question as to be capable of harming competition—as opposed to harming merely individual modem manufacturers that failed to make the sale. And Judge Koh implausibly turned Aspen Skiing from an ugly duckling into a swan. What was by the Supreme Court’s own description a narrow exception to the rule that even a monopolist may decline to deal with its smaller competitors was transformed by Judge Koh into an expansive rule of compulsory licensing available to every corporate giant, including Apple, the largest seller of smartphones in the United States and the new owner of Intel’s former modem business. Judge Koh’s error was highlighted by the FTC, which disavowed her findings of fact and conclusions of law regarding Aspen Skiing during oral argument in Qualcomm’s appeal of FTC v. Qualcomm to the Ninth Circuit.

The quality of justice manifested in Judge Koh’s bench trial in FTC v. Qualcomm invites concern for the following reasons, which are distinct from the three specific grounds that I have identified above for reversing her judgment and vacating her permanent, worldwide injunction and her grant to the FTC of partial summary judgment:

1.  As noted above, Judge Koh adjudicated a claim, predicated on Aspen Skiing, that the FTC claims it never argued. Article III judges may not issue advisory opinions. They are not inquisitors.

2.  Judge Koh revealed by her courtroom remarks during trial that she inadequately understood how the doctrine of patent exhaustion informed the antitrust questions presented. Consequently, she also was incognizant of the enormous relevance of the evolution of that doctrine in the Supreme Court to her earlier order granting the FTC’s motion for partial summary judgment, in which she had found that Qualcomm owed contractual duties to two standard-setting organizations (SSOs) to offer exhaustive licenses to its cellular SEPs at the component level. If Judge Koh did not understand patent exhaustion at trial, she certainly did not understand patent exhaustion two months earlier, when she granted the FTC’s motion for partial summary judgment.

3.  Judge Koh impugned the integrity of Qualcomm’s founder and executives, its outside counsel, and its expert economic witnesses (as well as executives from Ericsson and Nokia whom Qualcomm called as percipient witnesses to testify in its defense). Yet, she credulously accepted as truthful, reliable, and relevant the self-interested testimony of competitors or customers, who might be expected to benefit from a permanent, worldwide injunction that would destroy a large portion of the value of Qualcomm’s uniquely innovative, successful, and productive business model.

4.  Judge Koh suppressed all findings of fact concerning the testimony of the FTC’s expert economic witness, Professor Carl Shapiro. Most notably, she suppressed his admissions that would have materially exculpated Qualcomm under the FTC’s theory of liability positing a royalty “surcharge” and modem “tax” levied by Qualcomm for an exhaustive license to its cellular SEPs. This judicial conduct by Judge Koh violated Federal Rule of Civil Procedure 52(a)(1) and the Due Process Clause of the Fifth Amendment as interpreted by the Supreme Court in Mathews v. Eldridge. The Supreme Court should make clear that due process of law requires that a judge in a bench trial faithfully discharge the obligation, akin to the duty that the Brady rule imposes on prosecutors, to report materially exculpatory evidence in the judge’s findings of fact.

5.  Judge Koh displayed naïveté about economics in the most important government monopolization case since Microsoft. Her findings of fact and conclusions of law misapplied undergraduate economics. Judge Koh could have avoided that embarrassment if she had either weighed the opposing expert economic testimony from the parties or exercised her prerogative under Federal Rule of Evidence 706 to appoint her own economic expert. Similarly, Judge Koh could have accepted the Antitrust Division’s recommendation that she hold an evidentiary hearing on remedies before issuing her permanent, worldwide injunction.

6.  Judge Koh seemed not to comprehend the economic consequences of FTC v. Qualcomm, either for Qualcomm or for other innovative firms. Because Judge Koh appeared not to understand what was at stake, she necessarily appeared not to understand the tradeoff to be made on the margin between the cost of, and the expected benefit from, her allowance of additional increments of procedural safeguards. To paraphrase Judge Richard Posner’s explanation of the economic logic of Mathews v. Eldridge, in giving Qualcomm the amount of process that it was due under the Fifth Amendment in the face of the government’s attempt to deprive the company of a large portion of the value of its business model, Judge Koh should have considered the value of Qualcomm’s property at risk of deprivation, the probability of erroneous deprivation if particular procedural safeguards were withheld, and the cost of supplying those additional safeguards. In various ways documented in detail in this article, Judge Koh denied Qualcomm procedural safeguards whose marginal cost of provision would have been minimal or nil. Two conspicuous examples were proper adherence to Federal Rule of Civil Procedure 52(a)(1) and to the controlling burden of production and burden of persuasion under American Express. The expected marginal benefit to Qualcomm from Judge Koh’s provision of these (denied) procedural safeguards was enormous. The expected marginal benefit to Qualcomm from preventing the improvident imposition of Judge Koh’s permanent, worldwide injunction amounted to nothing less than averting the destruction of a large portion of Qualcomm’s uniquely innovative, successful, and productive business model. That destruction would be measured in the tens of billions of dollars. The additional increments of process necessary on the margin to reduce the probability of the government’s erroneous deprivation of Qualcomm’s property would have a cost many orders of magnitude less than the harm to be averted. That probability of erroneous deprivation of property was obviously dangerously high: after all, the Ninth Circuit unanimously reversed Judge Koh’s findings of law.

The Ninth Circuit’s three-judge panel hearing Qualcomm’s appeal in FTC v. Qualcomm properly reversed Judge Koh’s judgment, vacated her permanent, worldwide injunction, and vacated her grant to the FTC of partial summary judgment. But the panel might not have recognized, because it did not conduct a de novo review, that Judge Koh’s decision in FTC v. Qualcomm also was deficient in its respect for those appearing in her courtroom, in its sophistication about the proper application of economic analysis to complex commercial litigation in technologically dynamic industries, and in its fidelity to due process of law.