"VERTICAL RESTRAINTS WITH HORIZONTAL CONSEQUENCES: COMPETITIVE EFFECTS OF "MOST-FAVORED-CUSTOMER" CLAUSES"
Jonathan B. Baker
Director, Bureau of Economics
Federal Trade Commission
Business Development Associates, Inc.
Antitrust 1996 Conference
Loews L'enfant Plaza
September 28, 1995
Note: This speech has been revised and is published in the Antitrust Law Journal, vol. 64, Spring 1996, pp. 517-34.
Note: This speech has been revised and is published in the Antitrust Law Journal, vol. 64, Spring 1996, pp. 517-34.
George Orwell's novel Animal Farm recounts the course of a revolution against the discredited old order on farmer Jones's Manor Farm. The victorious animals developed a carefully crafted and sensible set of rules to govern their own new order " the Seven Commandments. But the Seven Commandments were hard for most of the animals to remember. After much thought, one of the farm's leading intellectuals, a pig named Snowball, declared that the Seven Commandments could be reduced to a single maxim: "Four legs good, two legs bad." This seductive simplification was the beginning of the end for the animal revolution. In the name of this maxim, the pigs perverted each of the seven commandments, one after the other, to reestablish the old prerevolutionary order with themselves, rather than the farmer, on top. Antitrust revolutions also can founder on oversimplification. The Chicago school revolution of the last twenty years uprooted and discarded old, oversimplified per se rules and rubrics, because they discouraged efficient conduct. One theme of the Chicago revolutionaries was to focus antitrust attention on how collusion, that is, horizontal arrangements, can harm competition and consumers. Thus they advocated reorienting enforcement to emphasize prosecuting price-fixing and enjoining horizontal mergers that threaten to create monopolies or near monopolies. Yet the Chicago revolution also taught how vertical arrangements between sellers and buyers can achieve efficiencies. One of the most far reaching doctrinal shifts attributable to the Chicago revolution was the Supreme Court's rejection of the per se prohibition against vertically imposed territorial restraints in GTE Sylvania. By treating these restraints under the rule of reason, the Court permitted manufacturers to write contracts that induce dealer sales effort by discouraging dealer free riding on the promotional efforts of neighboring retailers. Other decisions have effectively relaxed old per se rules against other vertical arrangements, principally tying. Some enthusiastic commentators have advocated reversing all the old rules and treating vertical arrangements, including agreements setting resale prices, as per se legal. These developments could reduce to a deceptively simple maxim: "Vertical good, horizontal bad." If antitrust were to follow that maxim, it would give a free pass to all kinds of agreements between firms and their distributors or suppliers, while closely scrutinizing agreements among competitors.
Each side of the "vertical good, horizontal bad" maxim is an oversimplification. The "horizontal bad" half underplays the many important ways in which collaboration among competitors can achieve efficiencies. To be sure, another legacy of the Chicago revolution was a more nuanced treatment of horizontal restraints as well. Decisions like BMI and Rothery move away from per se treatment of some horizontal restraints expressly to permit efficiencies.
My focus today is the "vertical good" half of the maxim. This slogan understates the competitive problems that can result from vertical restraints. Even believers in "vertical good, horizontal bad" should still pay attention to vertical restraints. Not because "vertical" is by itself "bad", but because vertical restraints can impair horizontal competition. I will describe three ways in which vertical restraints can harm competition by having horizontal effects, which I will term "facilitating practices," "raising rivals" costs," and "dampening competition." In pointing out the anticompetitive potential of vertical practices, I do not mean to minimize their efficiency benefits; I merely intend to highlight the part of the story that has not been emphasized in recent years. To focus my discussion, I will illustrate the three kinds of anticompetitive effects with examples, including some antitrust cases, involving one particular kind of vertical restraint, namely most-favored- customer clauses.
Analyzing vertical restraints is complicated by the variety of forms they take. The most-favored-customer clause is a good example because it is familiar and common and it has frequently been subject to antitrust attention.
A most-favored-customer clause, also called an "antidiscrimination" or a "most favored nations" clause, is a promise by one party, for example a supplier, to treat a buyer as well as the supplier treats its best, "most favored" customer. If the supplier lowers price to someone else, then the buyer's price will be lowered to match. The immediate effect of a most- favored-customer clause is uniformity in how one supplier treats different customers.
In applying rule of reason antitrust analysis to vertical restraints such as a most-favored-customer clause, the anticompetitive effects must be identified and compared with efficiencies. I will describe three anticompetitive mechanisms by which a most-favored-customer clause in vertical contracts could harm competition, one corresponding to each of the general vertical theories I have noted: practices facilitating coordination, raising rivals' costs, or dampening competition. My broader purpose " of highlighting the many ways vertical practices can harm competition " is, in one respect, not well served by focusing upon most-favored-customer provisions: the distinction between the facilitating practices and dampening competition theories could be made more apparent by choosing some other vertical practice for study. On the other hand, my goal of encouraging careful analysis of individual practices rather than careless application of slogans is well served: I will explain why some, though not all, of the commonly-supposed efficiency benefits of most-favored-customer provisions are illusory.
Facilitating Horizontal Coordination
The first kind of competitive effect to be concerned about is the use of most-favored-customer provisions to facilitate anticompetitive horizontal coordination. Coordination works better if firms have little incentive to cheat to begin with. Most-favored-customer clauses serve just that purpose. A firm that has agreed to offer most-favored-customer treatment in its contracts has reduced its incentive to deviate from a coordinated horizontal arrangement, because it cannot limit its discounts to a single customer.
It may be enough for the firm to offer such protection only to major customers, or even a single major customer " as the court recognized in the old polio vaccine case. The court refused to treat uniform high prices as evidence of criminal conspiracy, where the largest purchaser, the government, insisted on most- favored-customer clauses. Industry witnesses pointed out that the most favored customer clauses set a price floor, and they would be reluctant to cut price to any customer, because they would then have to cut price to the federal government and to states with similar clauses in their contracts.
These facilitating coordination effects were also one object of the FTC's Ethyl litigation, and the Department of Justice's consent order prohibiting anticompetitive practices in the electrical equipment industry.
Raising Rivals' Costs
The second kind of horizontal effect from vertical practices is exclusion, or, more generally, raising rivals' costs. Vertical restraints can harm competition by creating conditions in which downstream firms must participate in, or accede to, what we might call an involuntary or coerced cartel. The restraints may increase the marginal costs of some downstream firms, inducing them to reduce output and raise price. Then the remaining downstream firm or firms can reduce output to raise price without fear that these rivals will undermine the nvoluntary cartel by discounting. The restraints may also increase the costs of potential downstream entrants, discouraging their entry. Then the remaining downstream firms can reduce output to raise price without fear of new competition. For a vertical restraint to implement the involuntary cartel strategy successfully, it must overcome three problems. The benefits of the strategy to the firms undertaking it must exceed their costs; those firms must not cheat on each other; and their target must be unable to avoid the strategy. Yet none of these problems is necessarily insurmountable.
Most-favored-customer clauses could be weapons in a strategy to impose an involuntary cartel. Firms that demand and get most- favored-customer treatment from important input suppliers are assured that new entrants and existing competitors will not be able to obtain lower costs by getting better prices from those suppliers. By reducing the ability of entrants or rivals to lower their costs, firms can achieve or maintain prices above competitive levels.
Antitrust cases have addressed the threat of this kind of harm in the context of large health insurance plans demanding most- favored-customer protection in contracts with hospitals or doctors. Two leading cases point in opposite directions. In the Reazin case, Blue Cross had terminated its provider contract with the largest hospital in Wichita, Kansas after that hospital affiliated with a major, for-profit chain and a sizeable health maintenance organization. Blue Cross's standard provider contract included a most-favored-customer clause. The jury relied on the effects of that clause " they way it discouraged price discounting and thus the entry of new competition " as evidence that Blue Cross had market power, understood in the legal sense of power to control price or prevent entry.
A similar scenario appeared in the Ocean State case, but the result differed. In Ocean State, the Blue Cross demanded most- favored-customer terms from physicians that also affiliated with a new Health Maintenance Organization (HMO). The HMO's payment schedule to physicians contemplated sharing profits, if any. If the HMO was not profitable, its effective payment rate to physicians was lower than what Blue Cross paid. Under the most-favored-customer contracts, the HMO's physicians would then have to accept lower payment from Blue Cross, too. The implication was clear, as was the result: hundreds of doctors terminated their affiliation with the HMO after Blue Cross insisted on these terms. But both the trial judge and the First Circuit were reluctant to label as monopolization, in violation of Section 2 of the Sherman Act, a tactic that, on its face, looked like it was designed to ensure low prices. Federal and state enforcers have been actively scrutinizing the use of most-favored-customer clauses in health care contracts to discourage entry by raising rivals' costs. In the Delta Dental case, for example, the Justice Department and the state of Arizona challenged the enforcement of a most-favored-customer requirement by a health plan that had signed up 85 percent of the dentists in the state.
Vertical restraints can harm horizontal competition through a third route: by dampening competition directly. A party to a vertical restraint may appear to be imposing a burden on itself, limiting its own range of competitive options. That burden can constitute a commitment to take strategic action that will encourage anticompetitive cooperation, or discourage vigorous competition, by its horizontal rivals. In particular, in some industries, a commitment to conduct that appears less aggressive will lead rivals to see that their best interest is allowing industry prices to rise. The practices generating the commitments could have greater power when they are adopted by most or all the firms in an industry.
Most-favored-customer clauses can dampen competition by making firms less aggressive, in settings in which rivals will respond by becoming less aggressive as well. A firm that introduces a most favored customer clause commits to being less aggressive by obligating itself to pay a substantial penalty if it lowers price to any individual customer. If it lowers price to one, it must lower prices to all its customers. If its rivals would respond by becoming less aggressive themselves, some firms will find it profitable to make this commitment. This dynamic may be most likely, and the resulting price increase likely to be the greatest, when the number of firms is small, when higher prices would not lead to new entry, and when exogenous shifts in cost or demand that would tend to lead to lower prices are unlikely.
This effect may have been present in the industry situation of the FTC's Ethyl case, although it was not the theory that was litigated. There were only four firms; the industry was declining, so entry was highly unlikely; and demand was reasonably predictable. Only the top two firms routinely relied on most-favored-customer clauses. The Commission found that unanimous use of these clauses was not necessary to produce an anticompetitive effect. While the Court of Appeals disagreed with the Commission's legal conclusion that unilateral adoption of the challenged practices would support a finding of liability, the court was not asked to consider the dampening competition theory.
The dampening competition theory represents something of a frontier for antitrust enforcement. The theory makes sense, but enforcers seeking to apply it must grapple with judicial skepticism, with limitations on the reach of the antitrust laws, with the challenge of showing that a commitment to less aggressive behavior has led or will likely lead rivals to act likewise, and with the need to untangle efficiencies from harm.
The touchstone of the Chicago School's advice to treat vertical restraints leniently is the potential for creating pro- competitive efficiencies. Two types of efficiencies are frequently cited for most-favored-customer treatment. Yet the most common story of the two appears less convincing upon careful analysis.
It seems obvious to many that a most-favored-customer provision helps a buyer lower its costs by purchasing inputs for less. In a recent decision, Judge Posner describes such contractual clauses as "standard devices by which buyers try to bargain for low prices." The First Circuit found it "hard to disagree" with a district court's view that such clauses are "what competition should be all about." And former FTC Chairman Miller argued in Ethyl that most-favored-customer clauses must be procompetitive because buyers appear to want them. What seems obvious at first glance can look different upon reflection, however. In a market with many buyers, and in which it is costly for buyers to shop for a low price, uninformed buyers may pay more than the informed. An uninformed buyer able to bargain for most-favored-customer status might expect to obtain the benefits of becoming informed without expensive search. If some other informed buyer can convince the seller to lower price to meet the competition, the most-favored-customer can free ride on that buyer's bargaining effort, and so obtain a low price without price-shopping itself. But once one buyer discovers this way of shopping at low cost, other buyers seeking to minimize their own costs are led by competition to seek most- favored-customer status as well. After the resultingproliferation of most-favored-customer provisions, no buyer obtains the product for less. Indeed, prices to informed customers likely rise, as sellers come to see more cost than benefit from discounting.
There is nothing "perverse," to use Judge Posner's colorful phrase, about the idea that competition might reduce buyer benefits. Indeed, GTE Sylvania teaches this. When buyers learn about a product's features from a full service dealer, then make their purchase from a no-frills discounter, such buyer free- riding may drive the full service dealers out of business. Here buyer competition, at the end of the day, can cost buyers as a group. Nor is there anything perverse about the idea that buyers might compete to sign a contractual provision that ends up hurting them. When a monopolist asks its customers to agree not to deal with potential competitors, in order to deter entry by keeping entrants below efficient scale, buyers may compete to sign in order to ensure that they receive the monopolist's product " even when the buyers know that all would be better off if none were to agree. In short, when buyers desire something individually, one cannot assume, as these courts have done, that it is in the buyers' interest collectively to obtain it.
A second efficiency story is also frequently cited. This story applies when firms must write long term contracts with their customers knowing that supply and demand conditions might change unpredictably. For example, natural gas pipelines must contract with well owners without knowing the future demand for gas. A fixed price contract is unattractive because it would not lead production to respond efficiently to changes in demand. The well owners would not be willing to agree to renegotiate the price every year because doing so puts them in a difficult bargaining position: once they drill, they would find themselves at the mercy of single pipeline buyer in future years. One solution is to sign a long term contract, but to constrain the pipeline's ability to exploit individual well owners with a most- favored-customer provision. This efficiency explanation would not apply to every use of a most-favored-customer clause, but when plausible it must be balanced against the harm to competition under the rule of reason.
Unlike what happened in Animal Farm, we will not end up returning to the past. No one proposes making most-favored-customer clauses illegal per se. It is well-established that in evaluating the reasonableness of any vertical practice, both its anticompetitive and efficiency-enhancing potential must be considered. If the "vertical good, horizontal bad" slogan presents a danger, it goes the other way " that antitrust might develop new per se rules of legality that ignore the possibility that vertical restraints can harm competition through their horizontal effects.