Last month, Judge Victor Marrero of the District Court of the Southern District of New York issued a decision allowing the proposed merger between Sprint and T-Mobile to proceed, rejecting a suit brought by a coalition of state Attorneys General. This decision is deeply disappointing—Public Knowledge and other groups fought hard to stop the merger, which we believe will raise prices for mobile phone service, particularly in low-income communities. Unfortunately, some important features of our court system, and the strategic way that Judge Marero wrote the decision, mean that appealing the decision would have risked making future antitrust enforcement harder without much chance of overturning the decision. That is why it was prudent for the Attorneys General in New York, California, and other states that had originally challenged the merger to decide not to appeal Judge Marrero’s ruling.
Now is the time for focusing our energy on other monopolistic behavior antitrust enforcers need to challenge, improving the antitrust laws through statute, and making sure the Federal Communications Commission strictly enforces the settlement agreement to achieve the maximum competition it can hope to achieve in this market. We argued to the FCC and in court filings that the FCC/Department of Justice plan to bring back four nationwide carriers eventually, losing one in the meantime, was a risk that wasn’t worth taking. But now the best thing for consumers would be for that plan to work, and for enforcers and advocates to hold companies to their commitments.
While it is always good to want to challenge a bad court decision and fight on in the face of adversity, we must weigh costs versus benefits. First, the judge’s ruling was based on a misguided view of the facts and appellate courts almost never question a district judge’s fact judgments. Part of what was so frustrating about this decision was that Judge Marrero made the wrong factual determinations. Fact-finding is a key role of district court judges like Marrero, and just like in jury trials, findings of fact by a judge are not subject to appeal—only mistakes of law are. Judge Marrero deferred to the DOJ and FCC, and credited CEO testimony over rigorous economic analysis. CEOs are often very appealing witnesses to judges. From my many years of experience in pushing for strong antitrust enforcement, I have seen so many cases like this. The skills CEOs are selected for overlap strongly with the skills of a good witness: charismatic speakers who can wow a judge with their business vision and earnest pursuit of better products and services for their customers. They also may even benefit from judges’ unconscious biases.
In antitrust cases, judges are supposed to look beyond the witness personalities to the underlying incentives and opportunities profit-maximizing companies will invariably follow. Judge Marrero just didn’t do that. Litigation is inherently uncertain, and whether the judge gets along well with your witness can unfortunately have a big impact.
This problem is exacerbated by the fact that a district judge’s determination of facts is incredibly hard to overturn on appeal. On appeal, a judge’s legal determinations are reviewed de novo, meaning the appeals court gets to start from scratch in making legal determinations. But their factual determinations are only reviewed for clear error, a hard-to-meet standard. When deciding whether to appeal any court decision, litigants are much more likely to appeal based on legal arguments. They recognize the uphill battle of appealing a factual determination is unlikely to succeed.
Second, since the problems with this decision were mostly factual, there’s not a good chance of success if the State AGs had decided to appeal. However, some have argued that the State AGs should appeal based on Marrero’s determinations of law, in particular, his decision to credit efficiencies claimed by Sprint and T-Mobile. It’s important to be clear about the fact versus law distinction here. I believe strongly that many of the claimed efficiencies 1) are unlikely to materialize, 2) will not outweigh the costs of the merger, and 3) don’t require a merger to accomplish them (in the language of antitrust law we’d say they are not “merger-specific”). However, each of these is a factual determination. The legal question is whether efficiencies are currently permissible as a way to justify a merger. I think it’s worth having a discussion about how to change how the law treats efficiencies, since they are often misused by parties attempting to merge. But that doesn’t change the current law.
The argument for appeal is an interesting one, but it asks a lot of the reviewing court. It hinges on the idea that even though the antitrust enforcement agencies and courts routinely credit efficiencies as a defense to a presumption that a merger is anticompetitive—and the Horizontal Merger Guidelines that parties have relied on for years say the same,—the 1967 Supreme Court precedent of FTC v. Procter & Gamble[1] seems to say the opposite. Courts have discussed this situation, but none has explicitly rejected efficiencies as a potential defense.
The closest a court has come to rejecting efficiencies as a defense was in U.S. v. Anthem[2]. However, in that case, then Judge, now Justice, Brett Kavanaugh wrote a dissent detailing the legal precedents for an efficiencies defense in the DC Circuit (citing U.S. v. General Dynamics Corp.[3], U.S. v. Baker Hughes[4], and FTC v. H.J. Heinz Co.[5], as well as the DOJ and FTC’s Horizontal Merger Guidelines). Even if the states wanted to argue the risky claim that efficiencies defenses are not recognized in modern antitrust law, it’s clear that such a theory does not have the votes to win at the Supreme Court.
This isn’t to say that parties could not try to argue at court that these earlier precedents ought to win out, even though courts have not been following them for years. But certainly an argument that efficiencies are already not credited in merger enforcement would be novel. To win on this appeal, we would need the appeals court to depart from widely accepted practice.
Finally, and most importantly, the Second Circuit—the appeals court that the State AGs would find themselves in if they had decided to appeal the decision—does not seem likely to look favorably on this type of argument. An appeal in the Second Circuit is extremely dangerous at this moment. That Circuit overturned an Obama Administration case against American Express, affirmed by the Supreme Court, that locked in horribly inappropriate antitrust precedent making it probably much more difficult to challenge dominant platforms. On top of that, the Supreme Court is highly unlikely to be friendly to an appeal based on rejecting a business-friendly ruling crediting efficiencies. By appealing on this weak legal argument, we would risk locking in horrible single-case reasoning as the supreme law of the land.
For example, Judge Marrero’s misunderstanding of key facts could be locked into legal precedents for decades to come. His treatment of what “facts” are enough to rebut the structural presumption against a merger in a highly concentrated market could lead to an appellate ruling that virtually eviscerates the presumption. Marrero’s description of how traditional antitrust doesn’t apply well in “complex” and “dynamic” markets like wireless service could be chiseled in legal stone to limit antitrust enforcement across the technology sector and similar markets. These are legal risks simply not worth taking.
Unlike some, I disagree that the judge clearly got the law wrong. He certainly didn’t apply the law in the way I’d like it to work, but he wrote his decision carefully. He covered his reasoning with existing legal precedents, not claiming to make new law or leaving much room to even assert that he was doing so. This, together with the risky courts we face right now, counsels against an appeal by the states. But rather than just give up, we need to move on to push for strong enforcement against other monopolistic practices, try to fix the law in Congress and pressure the FCC to make this settlement work as well as it can.
[1] 386 U.S. 568, 580 (1967).
[2] 855 F.3d 345, 353 (D.C. Cir. 2017).
[3] 415 U.S. 486, 498-504 (1974).
[4] 908 F.2d 981, 989-90 (D.C. Cir. 1990).
[5] 246 F.3d 708, 720 (D.C. Cir. 2001).
Photo by Mika Baumeister on Unsplash