The “Rambus litigation” in all its many permutations — Justice Department investigation, FTC proceedings and multiple civil cases — has been documented and commented upon widely. For a recap see Andy Updegrove’s article here. At the heart of the legal controversy is the allegation that during the 1990s Rambus, the owner of key DRAM patents or pending patents that solved the CPU-memory chip “bottleneck” problem, failed to disclose these patents to JEDEC, an important standards-setting organization (“SSO”) to which Rambus belonged. JEDEC, uninformed of the existence of these patents, incorporated the Rambus technology in its standards, which were then widely adopted in the memory chip market.
Because Rambus withheld disclosure of its patents, JEDEC did not have the opportunity to exercise either of the two options open to it when a member disclosed proprietary technology: either choose another technology or negotiate industry-wide favorable licensing terms as a condition of adoption of the standard (so-called “reasonable and non-discriminatory” license fees, or”RAND” royalties). RAND royalties are negotiated and agreed-upon ex ante, that is, before the technology owner’s IP is adopted, and therefore before the technology owner acquires market power by reason of the adoption.
By the time Rambus announced its patents and began demanding royalties (and filing patent infringement suits against companies that refused to pay royalties), Rambus had achieved a technical “lock-in” that made it difficult for the memory chip industry to move to a different technology. Rambus’s lock-in allowed it to obtain a 90% market-share, and demand supracompetitive royalties from companies that were producing JEDEC-compliant memory devices. Rambus has earned several billion dollars in licensing fees to date, and by some estimates its total royalties could reach as high as $11 billion.
The FTC Decision
The Federal Trade Commission brought proceedings against Rambus, and issued a 120 page decision in 2006 holding that Rambus was guilty of monopolization in violation of Section 5 of the FTC Act. The FTC issued a remedial order mandating the maximum royalties that Rambus would be permitted to charge.
The FTC’s key conclusions were:
- In the early 1990s Rambus engaged in a multi-year, intentional campaign to conceal its pending patents from JEDEC, in violation of JEDEC policies that required disclosure from members. Rambus’s actions constituted a “deliberate course of deceptive conduct” and were calculated to mislead JEDEC. Rambus even went to far as to tailor its patent claims to cover parts of the proposed standards. As a result, JEDEC was unaware of the patents and pending applications when it adopted the standard. The FTC emphasized that Rambus’s conduct occurred in the context of an industry standard-setting process, where members had a legitimate expectation of good faith and candor.
- Rambus’s deception was material, and led to JEDEC adopting the Rambus IP in its standards.
- “But for” Rambus’s deception, JEDEC would have either excluded Rambus’s technologies from the DRAM standards, or would have demanded assurances of reasonable royalties before adopting the technologies. The FTC specifically found that alternative technologies were available that would have provided an alternative to the Rambus technologies, had Rambus been unwilling to negotiate RAND royalties.
- Once the standards were adopted and implemented by chip manufacturers the cost and technical obstacles to switching technologies were significant, and as a result the industry was “locked-in” to the Rambus technology.
- As a result of Rambus’s actions, JEDEC’s adoption of the Rambus technology, and the lock-in effect, Rambus acquired monopoly power – a more than 90% market share in the relevant market. The FTC found that the legal requirement of a “causal link” between Rambus’s conduct and its achievement of monopoly power had been established.
- Rambus’s actions allowed it to charge supracompetitive patent royalties, unconstrained by competition. This harm to competition led to reduced output and decreased overall social welfare.
The FTC rejected Rambus’s “inevitability/superiority” argument: that even in the light of full disclosure JEDEC would have standardized on Rambus’s technologies due to their superiority. The FTC also rejected Rambus’s argument that Rambus’s monopoly power was not enduring because there were no barriers to entry to rivals wishing to challenge its monopoly position. The FTC held that the DRAM industry was locked into the JEDEC/Rambus standards due to switching costs and issues of backward compatibility.
Rambus’s Appeal to the D.C. Circuit
Rambus appealed, and the D.C. Circuit reversed the FTC decision on April 22, 2008. However, the D.C. Circuit was severely constrained in its ability to review the FTC decision. While a federal appeals court has the authority, on appeal, to review, de novo, purely legal issues decided by the FTC, the FTC’s factual findings are conclusive if supported by “substantial evidence.” The FTC’s factual findings were extensive; therefore, the path of least resistance for the D.C. Circuit was to focus on “legal” issues.
Rambus and “Antitrust Injury”
The D.C. Circuit took a legally complex and circuitous path to get to what it concluded was the core issue in the case. That process is itself highly questionable — the D.C. Circuit appears to have misapplied the standard of proof for causation in a case of this sort. In essence, the FTC drew legal inferences in favor of Rambus based on the FTC’s inability to analyze Rambus’s conduct within the context of a hypothetical marketplace that never existed because of Rambus’s fraudulent actions.
Using this reasoning, the court narrowed the case to what it considered to be the key “legal” issue: whether Rambus’s deception caused “antitrust injury” by preventing JEDEC from negotiating RAND royalties before adopting the Rambus technology. The answer to this question is critical, since the Supreme Court has repeatedly warned that conduct evaluated under the rule of reason that harms competitors is not enough, alone, to violate the antitrust laws – the conduct must harm competition. However, what constitutes harm to competition, or “antitrust injury” is a perplexing question. Often, this mantra is simplified to mean harm to consumers, who are the beneficiaries of competition. However, this is an oversimplification — the distinction between harm to competitors and harm to competition is as much a question of economic theory as of law. As George Bernard Shaw famously noted, if all economists were laid end to end, they would not reach a conclusion. The idea of federal judges, who are self-taught in economics at best, applying economics wrapped in the suffocating folds of stare decisis would have left Shaw (and probably even Oscar Wilde) without a bon mot.
Here is the Court’s conclusion on this issue:
JEDEC lost only an opportunity to secure a RAND commitment from Rambus. But loss of such a commitment is not a harm to competition from alternative technologies in the relevant markets. . . . Indeed, had JEDEC limited Rambus to reasonable royalties and required it to provide licenses on a nondiscriminatory basis, we would expect less competition from alternative technologies, not more; high prices and constrained output tend to attract competitors, not to repel them.
The D.C. Circuit went on to reject the arguments that (a) “price raising deception” of the sort alleged against Rambus resulting in higher royalty payments to Rambus was competitive harm, and (b) Rambus’ patents put it in a position of monopoly power, and any conduct that permitted a monopolist to avoid restraints on that power must be anticompetitive. As the D.C. Circuit put it, “an otherwise lawful monopolist’s end-run around price constraints, even when deceptive or fraudulent, does not alone present a harm to competition in the monopolized marketplace.”
D.C. Circuit’s Misplaced Reliance on NYNEX v. Discon
In reaching the conclusion that Rambus’s actions did not present a “harm to competition,” the D.C. Circuit relied almost entirely on the Supreme Court’s 1998 decision in NYNEX Corp. v. Discon. However, the NYNEX decision was largely concerned with whether so-called “two firm boycotts,” which is a vertical arrangement (as opposed to a horizontal “group boycott”), should be reviewed under the rule of reason or put in the forbidden “per se” category of antitrust violations. Boycotts – two firm or multifirm — were not at issue in FTC/Rambus, and in its 90 page appellate brief opposing Rambus’ appeal, the FTC didn’t mention the Discon case at all. Indeed, Rambus itself touched on NYNEX v. Discon only in passing in its appeal brief and its reply brief.
The D.C. Circuit’s seized on NYNEX as if it were a drowning man reaching for a life preserver. A secondary issue in NYNEX was the allegation that the defendants were engaged in a complex kickback scheme, a “regulatory fraud,” that allowed them to perpetuate market power. The Supreme Court found that this conduct, although fraudulent and deceptive, did not create a basis upon which to apply the per se rule to an alleged boycott with no horizontal elements. Additionally, the Supreme Court noted that there was competition in the affected market, indicating that although there was harm to the plaintiff, Discon, there was no injury to competition, the sine qua non of an antitrust violation. The Supreme Court noted that “the complaint itself … suggests the presence of other or potential competitors, which fact … could argue against the likelihood of competitive harm . . . entry was easy, perhaps to the point where other firms . . . might have entered that business almost at will.”
The ease of entry by competitors that caused the Supreme Court to suggest the absence of antitrust injury in Discon was not present in FTC/Rambus. As noted above, the FTC found that the industry was “locked-in” to the Rambus technology, and therefore there were no alternatives that could act as a competitive restraint on Rambus’s ability to overprice its patent licenses. Also, as noted, the FTC expressly rejected Rambus’s argument that its monopoly was not “enduring” due to the potential entry of competitors. These “findings of fact” by the FTC were not challenged by the D.C. Circuit, and they distinguish the analysis of the “harm to competition” in NYNEX from the “harm to competition” in the Rambus DRAM market completely.
Finally, the D.C. Circuit’s finding that Rambus’s “price raising” conduct did not give rise to the sort of competitive harm addressed by the antitrust laws was presented without any economic basis or discussion, and appears faulty. By charging supracompetitive royalty rates Rambus raised the cost of a key component of the products the DRAM chips were used in. Under any rationale view of economic theory or antitrust precedent, the higher prices charged by Rambus, given the absence of entrants that could drive down prices, constituted harm to consumers, and therefore “antitrust injury.”
Appeal or Defeat?
The FTC has a number of options open to it: a request for rehearing en bank to the D.C. Circuit, a retrial of some issues at the FTC, a Supreme Court appeal, or no action at all, which would be an admission of defeat.
It appears that this case would be ripe for an appeal (recognizing, of course, that the Supreme Court has discretion to take appeals, and takes very few each year). An appeal, if granted, would allow the Supreme Court to clarify its views on antitrust injury in the context of patent disclosures to a standard-setting organization, something it has never done. An appeal might resolve the many questions that are now outstanding following the D.C. Circuit’s in Rambus. The importance of this to the health of the standards-setting process is almost impossible to overstate.
However, the FTC’s decision to appeal to the U.S. Supreme Court must be informed by the reality that the Supreme Court has overwhelmingly favored antitrust defendants since 2004, and therefore the mathematical odds alone disfavor success on an appeal. However, one would hope that a Supreme Court appeal would allow the Court to make it clear that a deceptive failure to disclose its technology to a standard setting organization, enabling the patentholder, as a result, to charge supracompetitive royalty rates and obtain monopoly power, is exactly the kind of economic behavior that the antitrust laws are designed to prevent.