By Richard A. Epstein
The recent Supreme Court decision in Comcast v. Behrend is not likely to attract much popular press. The case is worlds apart from the Court's highly publicized class-action decision in Wal-Mart v. Dukes, which addressed burning issues of workplace parity between men and women. In contrast, Behrend reads like a quintessential technical case reserved for class action gurus and antitrust professionals. But on closer look, it may well turn out to be much more.
The Factual Background In Behrend, the plaintiffs allege that the cable company Comcast is violating the Sherman Act through its "clustering" program. Under that program, the company swaps its facilities in areas where it has a low concentration of customers to other cable TV companies, in exchange for those companies' facilities in regions where Comcast has a higher customer concentration. One such area was the Philadelphia Metropolitan Region, where, as Justice Scalia reports in his five-member majority opinion:
In 2001, [Comcast] obtained Adelphia Communications' cable systems in the Philadelphia DMA, along with its 464,000 subscribers; in exchange, petitioners sold to Adelphia their systems in Palm Beach, Florida, and Los Angeles, California. As a result of nine clustering transactions, petitioners' share of subscribers in the region allegedly increased from 23.9 percent in 1998 to 69.5 percent in 2007.
These numbers suggest that Comcast had acquired a dominant position in the geographically discrete Philadelphia market, which under orthodox theory should allow it to raise prices above the competitive level, holding service quality constant. On the other side of the scale is the prospect that Comcast generated various kinds of operating efficiencies that could offset, either in whole or in part, the social welfare loss from higher market concentration.
Even this brief summary reveals that any individual plaintiff in the Philadelphia DMA could face a unique set of considerations to the extent that they subscribe to different systems in different geographical markets. Nothing about this overall pattern indicates that the Comcast clustering strategy should have the same antitrust effect across various submarkets. In some submarkets, the concentrations may not rise up to dangerous levels. In others they will be more severe. But the overall 46 percent increase makes it highly likely that some real negative effects occurred in at least some, and perhaps many, of these submarkets. The fact that swaps instead of direct purchases were used to achieve these concentration levels is irrelevant to the Sherman antitrust issues. The short statement of fact shows that something is afoot. The question is what to do about it.
The Majority Response It is here that the plot thickens because of how the evidence was presented. The plaintiff introduced four different theories as to how the higher concentration hurt consumers. To quote Justice Scalia again:
First, Comcast's clustering made it profitable for Comcast to withhold local sports programming from its competitors, resulting in decreased market penetration by direct broadcast satellite providers. Second, Comcast's activities reduced the level of competition from "overbuilders," companies that build competing cable networks in areas where an incumbent cable company already operates. Third, Comcast reduced the level of "benchmark" competition on which cable customers rely to compare prices. Fourth, clustering increased Comcast's bargaining power relative to content providers. Each of these forms of impact, respondents alleged, increased cable subscription rates throughout the Philadelphia DMA.
None of these so-called theories leap out from the others. From this morass, however, the District Court held that only the overbuilding theory made any sense. Once that was done, the plaintiff introduced an expert analysis by Dr. James McClave that presented a "regression model comparing actual cable prices in the Philadelphia DMA with hypothetical prices that would have prevailed but for petitioners' allegedly anticompetitive activities," and which yielded a tidy sum of about $876 million in damages for the entire class. Justice Scalia speaking for the five conservative justices held that this procedure did not meet the exacting standards for class certification that apply uniformly to all issues under Rules 23(a) & (b), in this instance on the predominance requirement, by showing that the proof of individual antitrust injury could be established by evidence that was common to all class members. To Justice Scalia, the great sin of the plaintiff's proof is that it attempted to show the McClave model bore on the question of class certification just because it was relevant to establishing the damages in the case if the merits had been reached on the overbuiding theory. In his view, the fatal mismatch at the certification stage took place because the regression analysis did not relate exclusively to the single overbuilding theory that the District Judge had allowed into evidence.
An Alternative View A stinging dissent by Justices Ginsburg and Breyer claimed that the writ of certiorari was improvidently granted because of a set of procedural wrangles on the question presented that have been ably analyzed elsewhere, and need not be discussed again here. But the larger question that they raise is whether class actions can ever be brought in this fashion if the type of regression prepared by McClave is insufficient to meet a class-certification threshold. What follows is how I would defend this approach.
The District Court was not concerned that three of the plaintiff's theories for antitrust damage were struck because the regression in question picked up all the losses by comparing the price movements in areas where Comcast had a large concentration with those where it did not. In my mind, the regression asked just the right question for estimating these damages. Unfortunately, the decision of the District Court raised an unnecessary damages kerfuffle by treating the four different lines of proof as though they were separate and distinct, such that the plaintiffs had to win on one or all of them. I regard that as a mistake. Properly understood, this case should have been able to go forward on an antitrust theory even if the District Judge had concluded that none of the four mechanisms identified by the plaintiff drove up the price of services to monopoly level.
The ultimate question in these cases is whether the price increase was attributable to the added concentration, and for that question the regressions have to be admitted because they apply to the class as a whole. The information on the four possible sources of the increase should not be looked at in the alternative; if examined at all, the theories should be treated at most as cumulative descriptive evidence that is weaker in kind than the quantitative evidence in the regression itself. It is therefore a plus that the regression is not tied to the overbuilding theory. If this analysis is correct, it is mistaken to insist that the harms suffered by the plaintiff class do not derive from the distinctive overbuilding theory put forward by the plaintiff. Instead, the numbers tell the key story, as each of the four theories mentioned could offer a partial explanation as to the subsidiary question of how the antitrust injury came to pass.
At this point, the relevant choices should be stark, given the limitations of regression analysis. No matter what regression is used, it is still the case that all the individual members of a given class will suffer somewhat different injuries that could never be picked up or measured if each person were to bring his own separate lawsuit. But in this instance, the class action offers a better vehicle for analysis because it attempts to measure aggregate social harm. That calculation in turn sets the stage for determining optimal deterrence against a defendant, by taking the total amount of antitrust injury that their actions caused and dividing it among the plaintiffs in a form that is certain not to reflect the exact injuries that each member of the class sustained. Yet at the same time, these errors do not systematically favor any identifiable class members and thus tend to cancel out. Allowing averaging across the plaintiffs, therefore, does improve the position of every member of the class, for each does far better off with a pro rata recovery than with nothing at all.
Why Class Actions Anyhow? These observations only go to show that the class action in the context of many smallish harms has a hidden advantage over individual law suits when precise estimates of individual harms are not possible. It still leaves open the prior question as to whether the proper antitrust response is through any kind of damage action. The transactions between Comcast and its trading partners were all matter of public record, and it could have been possible to vet the transaction for its positive and negative effects by some kind of pre-clearance procedure that sought to address any net consumer welfare losses that derived from the swaps. But if that is not done, then it seems odd to kill the private right of action by forcing the plaintiff to take what surely seems an unnecessary step in these cases, namely pleading the particular type of evidence that they use to establish the injury in question. So at this juncture, it is hard to predict what will be made of Behrend. Will it be treated as a misadventure in pleading or a major revolution in the proof of damages in consumer class actions? Only time will tell.