Herbert Hovenkamp writes that the First Court’s recent ruling against American Airlines and JetBlue for coordinating operations in New York City and Boston exemplifies the correct application of antitrust’s rule of reason, which has troubled courts and plaintiffs and led to underenforcement for decades.
Antitrust’s rule of reason is a mess. The problem is not how the rule is articulated but rather how it is applied. However, the First Circuit’s recent decision against American Airlines and JetBlue for forming a joint venture to coordinate operations in Boston and New York City presents a step forward toward reforming the application of the rule of reason.
Briefly, antitrust agreements fall into two categories. The per se rule is reserved for “naked” restraints, whose profitability depends on the exercise of market power. These include things like price fixing and naked market division. These agreements do not involve any joint production, research, or integration of operations, but only an agreement to fix prices or divide markets. When the per se rule governs, market power ordinarily need not be proven and the range of permissible defenses is limited. The fact that power and defenses need not be shown, however, makes it critical that per se offenses be defined carefully and narrowly. Once conduct is placed under the per se rule, the court is unlikely to evaluate power and effects in that particular case. As a result, courts need to be sure that the possibilities of competitive improvement are so low that investigating them is not worthwhile. Naked agreements among competitors that fix prices, divide market or boycott rivals are currently the only conduct falling into that category.
By contrast, “ancillary” restraints can be profitable even in the absence of market power. Their gains can come from lower costs, product improvement, or innovation. For example, if five computer technicians in Philadelphia should form and do business under the name “Computer Associates,” they almost certainly could not be fixing prices because their market share is too low. Their gains come from members sharing office space and costly diagnostic equipment and staffing and schedule coordination, including taking turns having one member always on call for emergency service. That is, they profit from cost savings and quality improvements that accrue to cooperation, and they can do this even if they have no market power. They may even have to agree on prices so that they can quote a common price no matter which technician is assigned to a particular job.
This does not mean that such agreements get a free pass, but only that they fall under antitrust’s rule of reason. The plaintiff must prove market power and at least one anticompetitive restraint—i.e., something that realistically causes higher prices, reduced output, or restrained innovation. Further, the court will not permit a defense whose rationale depends on the elimination of competition. For example, in the first NCAA case (1984) the Supreme Court rejected the NCAA’s defense that a limit on televised football games served to protect live game attendance. That defense admitted to the fact that the challenged restraint reduced output, thus forcing viewers to go elsewhere. It was like arguing that we need a coffee cartel in order to get people to drink more tea. In the 1978 Professional Engineers decision, the Supreme Court concluded that a professional organization’s ban on members from submitting competitive bids for work could not be justified with the argument that higher prices were needed to prevent engineers from cutting corners. As Justice John Paul Stevens wrote in the majority opinion, “the Rule of Reason does not support a defense based on the assumption that competition itself is unreasonable.”
The question whether a particular practice is unlawful per se or under the rule of reason is entirely one of law for the judge. The courts commonly say that prolonged “judicial experience” determines which rule is to be applied, and of course juries do not have any judicial experience at all. One consequence is that expert testimony on the question is not permitted, because experts can testify only about facts.
Once the courts establish that the rule of reason governs a case, how do they decide it? The worst way is to attempt to “balance” harms and benefits, a task that is far beyond a court’s capacity in any but the most obvious cases. Justice Louis Brandeis suggested such balancing in the 1918 Chicago Board of Trade case, holding that the court needed to balance a large number of vague historical and operational factors related to the history and purpose of the restraint. The result was large, meandering and often pointless records in rule of reason cases.
What is often overlooked is that Brandeis rehabilitated himself fifteen years later in the 1931 Standard Oil of Indiana decision. His opinion for the Court in that case refused to condemn a patent cross-licensing venture, citing two reasons. First, the government had made no attempt to show that the venture led to a reduction in market output. Second, the defendant’s low market share indicated that it did not have the power to pull off such a reduction in any event. Those two questions, as Brandeis showed, should be central in a rule of reason case.
Instead of balancing, the courts have developed a multi-step process under which the plaintiff must first show market power plus the existence of a competitively suspicious restraint. The operative word is “suspicious.” The requirement operates somewhat like the probable cause needed to justify a police search, but not final proof of wrongdoing. Second, if the plaintiff meets those requirements the burden of proof shifts to the defendant to provide a procompetitive or harmless explanation for the challenged restraint. Third, if the defendant succeeds in making that showing the burden again shifts back to the plaintiff to show that these goals could have been attained by a less restrictive alternative. Fourth, and failing all of that, the court may have no choice but to engage in some balancing. However, the entire object of this series of inquiries is to avoid the need to balance whenever possible.
However, the multi-step application of the rule of reason has never met expectations. The primary problem, as the Supreme Court noted in its 2020 Alston decision, is that plaintiffs almost never get beyond step one, the prima facie case. The courts have essentially required the plaintiff to pack its entire case into the first step, including those elements that should go into the defendant’s rebuttal. Here, the main culprit is the Supreme Court’s 1999 California Dental decision. The Court held that a suspicious agreement by dentists that eliminated quality advertising and severely restricted price advertising could have been justified by information imbalances or other market failures in the dental services market, and the plaintiff (the Federal Trade Commission) did not account for these in its prima facie case.
But there are good reasons for requiring the defendant to prove these justifications. The dentists were the creators of their own restraints. They were in by far the best position to provide an explanation for them. One purpose of the rule of reason’s burden-shifting framework is to place the obligation to provide proof on the person most likely to know about it. The impact of the Supreme Court’s failure to see this has been devastating. Plaintiffs lose more than 90% of their rule of reason cases, most of them at the very first stage. One effect is that harmful restraints go unchecked. Another is that plaintiffs go to extreme lengths to place restraints into the per se category in order to avoid real confrontation with the facts that a restraint presents. Indeed, this state of affairs has led to a phenomenon found in the NCAA cases: the only restraints ever found unlawful under the rule of reason are naked restraints that probably should have been unlawful per se. That outcome is unfortunately dictated by the Supreme Court’s conclusion in the first NCAA case that the rule of reason must be applied whenever cooperation among firms is necessary to create the product. Operating under that rule, the Court quite candidly characterized the challenged restriction on televised NCAA football games as “naked,” even while applying the rule of reason. Even price fixing of stadium hot dogs would have to be addressed under the rule of reason.
The rule of reason after American Airlines/JetBlue
In 2020, American Airlines and JetBlue formed an alliance that came close to turning their operations in New York City and Boston into a single airline. The Department of Justice and several states sued the two carriers in 2021, alleging that the alliance violated the Sherman Act’s prohibition on unreasonable restraints on competition. Earlier this month, the First Circuit ruled against the defendants on appeal. Importantly, the court applied the rule of reason to arrive at its decision in favor of the plaintiffs. The application of the rule of reason to condemn a joint-venture agreement that was clearly not a naked restraint suggests it may no longer be the obstacle to plaintiffs’ lawsuits it once was.
Looking more closely at the details of the case, this particular joint venture raised issues that often occur in cases involving mergers and joint ventures of transportation networks. When the venture partners are operating competing services on the same line, their venture generally serves to limit competition, with increased prices or reduced services. However, they tend to produce lower prices and better service when they apply to connecting lines, transfer points, or other situations in which the firms operate as complements instead of substitutes. In a recent merger case that involved JetBlue and Spirit Airlines, a court also noted that the merger eliminated competition on some lines, but greatly improved connections and scheduling on others. If the merger were permitted, the court concluded, travel would “become more widely available to more consumers….”
The idea is hardly new. The Interstate Commerce Commission and the Eighth Circuit had already observed it in the first antitrust merits decision to reach the Supreme Court—the Trans-Missouri railroad decision of 1897. Two lower tribunals had rejected the government’s challenge to a joint venture of railroads that controlled both parallel and connecting routes. Except for the technology, it was in many ways similar to the American/JetBlue venture. In approving the venture, the Eighth Circuit relied on the ICC’s report:
The fact that the business of railway companies is irretrievably interwoven, that they interchange cars and traffic, that they act as agents for each other in the delivery and receipt of freight and in paying and collecting freight charges, and that commodities received for transportation generally pass through the hands of several carriers, renders it of vital importance to the public that uniform rules and regulations governing railway traffic should be framed by those who have a practical acquaintance with the subject.
In the 1890s, freight rates had to be computed with pencil and paper by the shipping or receiving agent, and packages often had to travel down many lines. Most of the railroads at the time were chartered by individual states and limited to them geographically. A package shipped from, say, New York to Chicago, might have to travel down four or five independently owned lines. That served to explain and perhaps justify even that portion of the Trans-Missouri agreement that fixed rates. Nevertheless, the Supreme Court reversed, looking only at the freight-rate agreement.
In the American/JetBlue agreement the parties did not fix fares. Further, many of the services covered by the agreement were complementary, covering connecting rather than competing flights. In addition, they rebooked one another’s passengers whose flights had been canceled. However, they also agreed about the allocations of slots and gates and inaugurated a revenue-sharing agreement that blunted the incentives of the two firms to compete for fares. It made the two airlines “indifferent as to which carrier a customer uses.” Overall, the court concluded, the venture led to “decreased capacity, lower frequencies, or reduced consumer choice on multiple routes.” While some offsetting benefits were shown, the court concluded that they could have been attained by less restrictive alternatives—in this case, agreements that coordinated the activities in which the firms operated as complements, but without limiting competition between them.
Just as the Supreme Court has consistently done, the court declined to embrace a “quick look” rule of reason analysis. Rather, it held, the rule of reason creates a “sliding scale” that the district court had properly applied. The court went through the multi-step process explained above, finding that the plaintiff had plainly provided evidence of market power and competitive harm. Further, while the defendant had been able to rebut by showing some benefits, these could have been attained through less restrictive means. It did not have to engage in any “balancing.”
The American Airlines/JetBlue decision illustrates the rule of reason done right. Applying it in an antitrust case will never be cheap, but it can be far less costly in litigation time and effort than the meandering “balancing” queries suggested in decisions like the Chicago Board of Trade case. The key is to focus on power, which concerns the defendants’ ability to limit market output, raise prices, or restrain innovation. With power established, then the court must consider whether the challenged conduct actually did so.
Author Disclosure: the author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
