U.S. Department
of Justice
Ranking Exclusionary Conduct
RANKING EXCLUSIONARY CONDUCT
By Susan Creighton
Director, Bureau of Competition
Federal Trade Commission
Fall Forum, November 15, 2005
As in the past, it is a pleasure to speak at the Fall Forum. I am particularly pleased to appear at this session because the topic is one in which I have considerable interest and one about which I have spent some time thinking since arriving at the Federal Trade Commission, and especially since becoming the Director of the Bureau of Competition. My remarks today are, as always, strictly my own, and not those of the Commission or any Commissioner.
As Kevin mentioned, I am here to address the question of whether the Sherman Act allows dominant firms to compete aggressively on price while requiring more caution when firms are engaged in non-price exclusionary conduct? Let me begin by providing the formal, legalistic answer to that question, which of course is "no." So long as all of the other elements of a Sherman 2 case are satisfied (monopoly power or a dangerous probability of attaining such power and specific intent to monopolize in the case of attempted monopolization), conduct that is "exclusionary" within the antitrust meaning of that term is proscribed by the Sherman Act, irrespective of whether the conduct is price or non-price conduct.
As the rest of my remarks will reflect, however, I believe that, at another level, the answer has been, and should be, yes.
As antitrust history has shown, and as all of you know, a central difficulty for Sherman 2 cases has been arriving at the proper boundaries for the meaning of "exclusionary" conduct. To state the obvious, that is a conceptual problem that has not yet been fully solved -- at least to the extent that there is no consensus on a universal test for determining exclusionary conduct. In addition, and inextricably related to the conceptual difficulty, is the practical problem of defining exclusionary conduct in a manner that permits courts to recognize it when they see it. Short of being able to accomplish that kind of clarity in all instances, we at least want an understanding of exclusionary conduct that - to use terminology out of statistics - minimizes the sum of consumer harm due to type 1 error and consumer harm due to type 2 error - that is, minimizes the harm done to consumers from failures to prohibit anticompetitive acts and from inadvertently proscribing procompetitive acts. At the same time, we also want to minimize the transactions costs imposed on companies in the way of counseling and other costs created by the uncertainty regarding legal constraints with respect to particular business practices.
As we all know, minimizing such harm, and reducing such costs, is tough to do in the real world of antitrust enforcement policy. Indeed, perfection is no more achievable in the enforcement decisions made by government agencies, or in the antitrust law, than it is in any other sphere of human affairs. Yet, even granting all of this uncertainty and room for error, I think that we can begin to make progress in this area by developing a consensus regarding what can be viewed as "extreme" forms of conduct on either end of the spectrum of exclusionary conduct, and then begin to whittle away at the broad range of conduct in the middle. In so doing, I believe that we can "rank" (if that is the proper term) forms of exclusionary conduct in ways that allow us to set priorities in our enforcement efforts and, in so doing, direct the efforts to those areas in which we can be more confident that we are promoting consumer welfare and not proscribing efficient, procompetitive behavior.
"Competition on the Merits"
Let me begin with the language that we often use and the question that we often ask courts to answer in Sherman 2 actions, which is whether or not the defendant's conduct represents "competition on the merits." Care must be taken in using the phrase "competition on the merits" because it is often used as though it brings specific content to Sherman 2 analysis. In fact, defining exclusionary conduct by reference to "competition on the merits" tends to be circular because there is no clear agreement on what, if anything, "competition on the merits" actually means. We only know that a monopolist cannot be found to have committed an illegal act if all the firm did was to engage in "competition on the merits." Conversely, if a monopolist is confronted with rivals or the prospect of their entry, it cannot lawfully counter the competitive challenge with conduct that is not "competition on the merits."
Notwithstanding this circularity, I think that there are a number of interrelated policy considerations surrounding the concept of "competition on the merits" for which, I believe, there is universal agreement and thus provide some common ground. First, as we all know, monopolies, other things equal, are not unlawful. Nor is monopolization itself, without more, illegal. A monopoly that is obtained by "superior skill, foresight, and industry" does not violate competition law.1 It is important to recall that, in setting precedent for the kinds of behavior that competition agencies are trying to deter, striving for monopoly is "an important element of the free market system," because "it induces risk taking that produces innovation and economic growth."2 Accordingly, "the successful competitor, having been urged to compete, must not be turned upon when he wins,"3 even where that success might have a short-term adverse effect on the welfare of consumers.4 The consumer-oriented goal of antitrust policy dictates an overarching need to protect the freedom of even dominant firms to compete.
Competition policy, therefore, requires distinguishing permissible from impermissible means to obtain or maintain a monopoly. The need to draw this distinction highlights a second consideration that policy-makers and courts must keep in mind: conduct that will defeat competitors is what monopolists engage in to obtain and maintain monopoly power, but it is also what we expect competitors to do in open, freely competitive markets. As one judge put it: "Competition is a ruthless process. A firm that reduces cost and expands sales injures rivals - sometimes fatally. . . . These injuries to rivals are by-products of vigorous competition, and the antitrust laws are not balm for rivals' wounds."5 In other words, competitive conduct frequently looks like exclusionary conduct, because aggressive competition may harm less-efficient firms - even though it is "precisely the sort of competition that promotes the consumer interests that the Sherman Act aims to foster."6
The importance and the difficulty of the need to distinguish between exclusionary and competitive conduct illuminates yet a third consideration common to all legal regulation of marketplace behavior. Specifically, in evaluating market behavior, we want to get correct results but, at the same time, we also want to give clear and specific guidance to those who have not yet acted. So, simply to say that antitrust enforcers will just weigh all the factors carefully after the fact may generate very defensible results after the fact, but such an enforcement policy will justifiably be criticized for failing to give sufficient guidance to businesspeople who must make choices before all the consequences of those choices can be known.7
Collectively these three points, about which I think that we can all agree, make it clear that it is essential that antitrust policy directed toward unilateral business conduct be carefully delineated. At a minimum, antitrust enforcement must account for the fact that legal actions vary with respect to their potential to chill desirable competition, an unintended consequence to be sure, but a likely fact of life. In addition, in any given action, enforcement policy must consider both the ease with which exclusionary conduct can be distinguished from aggressive competition and the feasibility and importance of creating "safe harbors" for firms in the marketplace.
Application of These Policy Considerations to Potential Monopolization Actions
For these, as well as possibly other, reasons, U.S. antirust law has evolved a textured approach to the question of what constitutes illegal monopolizing conduct or illegal monopoly maintaining conduct. For example, the Supreme Court has been emphatic in telling us that pricing practices, especially aggressive price cutting, entail the greatest danger that restricting single-firm conduct ultimately will harm consumers by chilling the competitive process.8 U.S. courts thus treat with particular caution claims that low pricing has led to a monopoly or has maintained a monopoly.
There are several reasons for this caution. First, although aggressive price cutting may be the mechanism through which competition is excluded, it "is the same mechanism by which a firm stimulates competition."9 The exclusionary and competitive acts thus look precisely alike. Second, "mistaken inferences" of predatory pricing are "especially costly, because they chill the very conduct the antitrust laws are designed to protect."10 Third, given the heavy costs of predation to the would-be predator, and the usual lack of entry barriers to potential rivals, the strategy is unlikely to succeed. Thus, the risk of consumer harm is relatively low.11
At the other end of the spectrum, I would submit, is conduct that nearly always leads to consumer harm. In my view, the improper manipulation of governmental processes to create or maintain monopoly power plainly falls within this category. Improperly using the rules of government against competitors can be one of the most effective ways for a firm to acquire or maintain market power. Indeed, "[m]isuse of courts and governmental agencies is a particularly effective means of delaying or stifling competition."12 Obtaining a patent by perpetrating fraud on the patent office, for example, is a well-known example of such abuse that may be part of a scheme of unlawful monopolization.13 Significantly, such conduct, I believe, lacks any cognizable justification in the form of socially beneficial efficiency gains. For this reason, in other fora, I have labeled such conduct "naked" exclusion.14 By that term I mean that it is conduct that is likely (assuming all other elements of an antitrust violation are made out) to produce only anticompetitive effects.
Significantly, naked exclusion need not be confined only to abuses of governmental processes. In can also occur in the purely private sphere. Opportunistic or deceptive behavior in standard-setting organizations and the perpetration of intentional business torts, for example, can be effective forms of excluding competitors that lack any cognizable social efficiency benefits. Of course, not all business torts rise to the level of an antitrust violation - one must prove harm to competition - but when the other elements of a Sherman 2 case are made out, I submit that business torts and deceptive practices not only are not in any way "competition on the merits," but may properly be considered to constitute actionable exclusionary conduct under Sherman 2.
Occupying the wide and expansive middle ground is business conduct that falls within the general ambit of selling and distribution practices. These practices would include, for example, such things as exclusive dealing arrangements, product bundling, and requirements contracts. Determining when and even whether practices such as these are anticompetitively exclusionary in any given instance is by far the most difficult problem facing enforcers and courts in Sherman 2 actions. For one thing, these kinds of practices are very common in highly competitive markets, reflecting the fact that such distribution methods can reduce costs and improve firm efficiency. When this occurs, and if we can accurately recognize it, we can all agree that the practice is "competition on the merits." It is also the case, however, that in some circumstances denying rivals access to key distribution channels may be an effective strategy for acquiring or maintaining market power. Asking courts consistently to sort out the good from the bad may be simply asking too much given current empirical understanding. Moreover, because distribution restraints are a frequent and effective form of competition, claims of exclusion based on such practices have the potential to chill conduct that benefits consumers. It is therefore important not only for antitrust enforcement to exercise considerable caution in challenging selling and distribution practices, but also to provide clear guidance to firms to minimize the potential for chilling efficient conduct.
Implications for Sherman 2 Enforcement Policy
Where does all of this leave us? I believe that viewing the range and variety of business conduct that might be subject to claims of "exclusion" in light of the issues that I have just discussed has important implications for the setting of antitrust enforcement priorities. The impact of antitrust enforcement is felt not only when an enforcement action is initiated or a judicial decision is rendered, but also when businesses take steps (or avoid taking steps) in the context of prevailing legal standards. Accordingly, in deciding enforcement priorities respecting actions taken against single-firm conduct, I believe that enforcement agencies should take into consideration not only the risks of "false positives" and "false negatives" in the particular case, but also the effect of the articulated standard generally on business conduct in the marketplace. That is, enforcers should give careful thought to the relative balance between identifying conduct that may be exclusionary, and the risk of deterring a wide range of conduct that might be highly beneficial to consumers.
Thus, in the allocation of always-scarce enforcement resources, a sound and sensible enforcement program might focus first and foremost on forms of exclusionary conduct that do not even arguably raise cognizable efficiency justifications. In this regard, abuse of governmental processes presents a very different trade-off of risks and benefits than aggressive price cutting. Unlike predatory pricing, abuse of governmental processes frequently is likely to succeed from the firm's perspective because the exclusionary effect often operates by force of law. Moreover, as I and several of my FTC colleagues have argued elsewhere, such exclusionary conduct is often going to be "cheap."15 By that I mean that, by comparison with other forms of unilateral conduct, it often costs little to attempt. Therefore, as economists would tell us, because the exclusion is "cheap" and yet can produce substantial returns to the purveyor in the form of supracompetitive profits, it is likely to be attempted frequently. If that claim is correct, then our enforcement program should be especially alert to discovering instances of such exclusion. Most fundamentally, because the conduct does not in any way resemble "competition on the merits," the risk of chilling procompetitive conduct is nil and, thus, the payoff to consumers is high. False statements to government agencies are not susceptible to any justification. They cannot be explained in terms of the defendant's effort to increase output or improve product quality, innovation, or service. They therefore do not raise the same concerns that challenging other kinds of conduct entails.
Looking for anticompetitive opportunism in standard-setting organizations and for intentional torts that exclude competition similarly can yield significant bang for the enforcement buck. Like the ease with which regulatory structures can be "gamed," and the relatively low cost of trying, such business conduct within the private sphere not only lacks in any benefit to consumers, but is likely to be commonplace relative to other forms of exclusionary conduct. These acts - plainly outside of the area of "competition on the merits" -- should have an important place on an enforcement agenda that challenges exclusionary conduct.
And indeed, at the FTC, both abuse of governmental process and anticompetitive strategic opportunism have been and will continue to be an important focus of our enforcement efforts. The FTC's administrative actions against Unocal,16 South Carolina Dentists,17 and Rambus,18 as well as the Orange Book cases,19 were all instances in which we alleged that the defendant's conduct cannot be explained in terms of the defendant's effort to increase output or improve product quality, innovation, or service. Thus, we believed that each was a case of naked exclusion. In each case, moreover, we believed that the alleged exclusionary conduct was, by comparison to alternatives, a relatively cheap form of exclusion.
Apart from enforcement priorities, there may be other important implications of thinking about claims of exclusionary conduct on the basis of the principles that I have set out. To begin with, because the risk trade offs are fairly one-sided at both ends of the spectrum - that is, aggressive pricing is seldom likely to be anticompetitive, while chilling such conduct carries great costs; even as instances of naked exclusion are both more likely to be successful, and less likely to be efficiency-enhancing - such trade-offs might be incorporated into the standards courts apply when analyzing such conduct. By this, I mean that judges (and with apologies to statisticians if I misuse the term!) might engage in "Bayesian" thinking (that is, having prior notions about probabilities in mind) with regard to the standards they apply to different forms of exclusionary behavior. I believe that this is precisely what the Supreme Court did in Brooke Group when it instructed us to apply a very strict standard to assess whether aggressive pricing is or is not "competition on the merits." By contrast, for naked exclusion claims, no such comparable screen appears necessary: an intentional tort, assuming the requisite market power and causation elements are satisfied, should not separately need to satisfy some further definitional hurdle to be recognized as "exclusionary."
What about the broad and perplexing set of claims in the middle, typically involving different types of distributional restraints? Are there ways to push particular types of conduct more towards the aggressive pricing end of the spectrum, or towards the end of the spectrum occupied by naked exclusion? It seems to me that two approaches - one empirical, the other evidentiary - have been, and should continue to be, used to chip away at the uncertain middle. First, one implication of the lack of agreement regarding various distributional practices is that further empirical research is needed better to understand the competitive consequences of ubiquitous selling and distribution practices under alternative factual scenarios. For example, with respect to concerns regarding false negatives and false positives, it would be good to know how likely are various practices, frequently attacked as creating or maintaining monopoly power, in fact useful as methods to reduce costs or lower output. It would also be important to know under what, if any, circumstances behavior that seems reasonably likely to be efficient might also threaten to generate market power of sufficient magnitude and probability that net social welfare will be greatly reduced.
Such empirical research, it seems to me, might assist in developing factual screens or other tests that could profitably be employed better to select actions that challenge conduct that falls within this middle area of the spectrum. For example (and I am only offering this as a hypothetical), empirical research showing that bundled discounts are not likely to be exclusionary when the potential defendant's share of industry sales in one or more components of the bundle is below a certain level would significantly aid both enforcement agencies and courts. To the extent that such a screen creates a "safe harbor," it would also provide important guidance to the business community.20
Evidentiary screens also have been used to distinguish where conduct lies along the spectrum between aggressive pricing and naked exclusion. For example, whether or not one agrees with the outcome in that case, I believe that the Court's emphasis in Aspen Skiing on the fact that the defendant did not adopt a particular business practice until it had acquired market power (a factor emphasized by the Court again in Trinko) is an example of an effort to identify such an evidentiary screen. The rationale for such a screen appears to be that if a defendant did not find a particular business practice profitable under competitive conditions, but does so once it has market power, the business practice -- in the absence of any other changed circumstance that can explain the change -- fairly may be identified as resulting from the acquisition of market power. Although that is not the same thing as concluding that the practice is exclusionary, the Supreme Court may view this evidentiary screen as significantly changing the risk trade-off that should apply to such cases, so that if there is also then evidence that the practice had the effect of dampening competition from rivals, the conduct meaningfully may be distinguished from ordinary "competition on the merits."
The Department of Justice's "no economic sense" test, applied for example in the Dentsply case, I believe also may be understood as an evidentiary screen. If an exclusive dealing arrangement would not make sense but for its exclusionary effect, as the Department of Justice found in Dentsply, the arrangement may readily be distinguished from the many exclusive dealing arrangements that deter free-riding and improve efficiency.
As a practical matter, I believe it makes sense for us to continue to work towards the identification of particular evidentiary and empirical rules as a way forward, both from an enforcement perspective and also as a way of working towards clearer and more administrable judicial standards. As enforcement officials, our enforcement priorities should be towards those cases where we can with some confidence distinguish the conduct at issue from conduct that might be beneficial to consumers. The same concerns arguably should also inform the courts' approach to the many private damages actions brought under Sherman 2. If indeed I am correct in thinking about unilateral business conduct in the context of a spectrum that, with varying degrees of confidence, can be delineated by forms of conduct with disparate likelihoods of falling within or without the ambit of "competition on the merits," then, it seems to me, it would be well for courts also to account for this condition
Conclusion
I will close by coming full circle back to the original question: whether the Sherman Act allows dominant firms to compete aggressively on price while requiring more caution when firms are engaged in non-price exclusionary conduct? My first answer was "no" but, in light of my remarks today, I would add the important qualification that it is possible - and indeed, desirable - to rank antitrust attacks on forms of unilateral business conduct by their likelihood to be consistent with important policy goals. Although, in principle, many varieties of conduct, including aggressive pricing, can in a given factual context, be exclusionary - i.e., not "competition on the merits" - some claims of exclusionary conduct are far more likely to be meritorious than others. Both enforcement priorities and legal standards should account for this fact.
FOOTNOTES
1. The phrase comes from Judge Learned Hand's opinion in United States v. Aluminum Co. of Am., 148 F.2d 416,430 (2d Cir. 1945).
2. Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 407 (2004).
3. Judge Hand also coined this phrase in the Alcoa case. 148 F.2d at 430.
4. Harm to consumers from an efficient monopolist may come about in some cases where the monopolist would restrict market output further (and consequently raise market price higher) than would have been the case in a market with, for example, three rivalrous but slightly less efficient firms. The monopolist is not condemned, nevertheless, because it merely engaged in "competition on the merits." An example of how similar harm to consumers may occur from a merger that generates both market power and merger-specific efficiencies is neatly explained in Ken Heyer, A World of Uncertainty: Economics and the Globalization of Antitrust, 72 Antitrust L.J. 375, 404-06 (2005).
5. Ball Memorial Hospital, Inc. v. Mutual Hospital Insurance, Inc., 784 F.2d 1325, 1338 (7th Cir. 1986) (Easterbrook, J.).
6. Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 767 (1984).
7. Still, a "case-by-case approach, one that bases decisions more explicitly on their likely impact on welfare," may be better than "'bright line proxies' and rules of thumb" at reducing business uncertainty and enhancing welfare. Heyer, A World of Uncertainty, supra note 6, at 417 et seq.
8. Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U. S. 574, 594 (1986).
9. Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 226 (1993).
11 Given these concerns, the Supreme Court has articulated a rule that seeks to minimize the risk of falsely proscribing competitive acts. In particular, before low pricing can be condemned, the conduct must be shown to involve the short-run sacrifice of profits in order to attain the probability of long-run market power. At the same time, aggressive, but above-cost, pricing is considered "competition on the merits" and is placed within a safe harbor. Brooke Group, supra.
12. Robert Bork, The Antitrust Paradox 159 (1978, rev. 1993).
13. See Walker Process Equip., Inc. v. Food Mach. & Chem. Corp., 382 U.S. 172 (1965); see also American Cyanamid Co., 72 F.T.C. 623, 684 (1967), aff'd, Charles Pfizer & Co. v. FTC, 401 F.2d 574 (6th Cir. 1968).
14. Susan A. Creighton, "Cheap Exclusion," remarks before Charles River Associates 9th Annual Conference on Current Topics in Antitrust Economics and Competition Policy (Feb. 8, 2005).
15. See Creighton, Hoffman, Krattenmaker & Nagata, Cheap Exclusion, 72 Antitrust L.J. 975 (2005) (describing false government filings, abuse of standard-setting processes, certain intentional tortious conduct, and abusive litigation as examples of "cheap exclusion").
16. In re Union Oil Company of California, Dkt. No. 9305 (March 4, 2003) (Statement of the Commission) available at http://www.ftc.gov/os/adjpro/d9311/index.htm.
17. In re South Carolina State Board of Dentistry, Dkt. No. 9311 (Sept. 12, 2003) (Complaint and Commission Decision), available at http://www.ftc.gov/os/adjpro/d9311/index.htm, appeal docketed, No. 04-2006 (4th Cir.).
18. In re Rambus Inc., Dkt. No. 9302 (F.T.C. Feb. 23, 2004) (Initial Decision dismissing complaint), available at http://www.ftc.gov/os/adjpro/d9302/040223initialdecision.pdf, appeal docketed before Commission. I am recused in Rambus, and so I take no position on the merits of the case.
19. In re Bristol-Myers Squibb Co., Dkt. No. C-4076, 2003 WL 21008622 (F.T.C. Apr. 14, 2003) (consent order), available at http://www.ftc.gov/os/2003/03/bristolmyersconsent.pdf; In re Biovail Corp., Dkt. No. C-4060, 2002 WL 31233020 (F.T.C. Oct. 2, 2002) (consent order), available at http://www.ftc.gov/os/2002/06/biovailelanagreement.pdf.
20. Of course, being outside of any such "safe harbor" does not mean that the practice is anticompetitive.