Roughly a year into control of the federal antitrust agencies, President Biden’s antitrust team is turning its attention to policies and enforcement practices. They seem poised to start, as antitrust so often does, with market definition. This is an appropriate target for review but also perilous territory for the administration. Even slight missteps in market definition could spell disaster for broader enforcement objectives. To help policy work start from a solid foundation, this essay identifies seven common myths of market definition and explains how to avoid them.
By Sean P. Sullivan
Roughly a year into control of the federal antitrust agencies, President Biden’s antitrust team is turning from lofty goals and aspirations to the honest work of converting those objectives into practicable policies and enforcement practices. They seem poised to start, as antitrust so often does, with market definition.
The DOJ’s and FTC’s recent Request for Information on Merger Enforcement asks sixteen questions about whether and how the Agencies should revise market definition. It implies interest in revising market definition under other headings as well. Individual remarks by Assistant Attorney General Kanter highlight market definition as a special target of reform efforts. And market definition and market structure analysis are among the subjects addressed in a recent wave of proposed antitrust legislation in Congress.
Market definition is an appropriate focus of review. Improvements to how the Agencies and the courts approach market definition could enhance not just the clarity of antitrust analysis but also the impact and efficiency of enforcement efforts. By the same token, however, changes to market definition should not be undertaken lightly. The negative consequences of even slight missteps could quickly overwhelm every other benefit the administration seeks to unlock. Market definition is that important.
It is thus alarming to see some old points of confusion swirling about in the current discussion. Several examples of confused thinking are apparent in the Agencies’ Request for Information. Others will undoubtedly be introduced by responsive commentary. Still others lurk in the background context from which the current push for policy reform emerges. To help policy work start from a solid foundation, the following identifies seven common myths of market definition and explains how to avoid them.
I. Myth #1 – Market Definition Is a Necessary Step in Rule of Reason Analysis
Decades of briefs and court opinions have peddled the shopworn claim that “[d]etermination of the relevant market is a necessary predicate to a finding of a violation of the Clayton Act.” Not one has ever offered a plausible explanation why legality is unknowable outside the context of a relevant market. It is sometimes suggested that the language of Section 7 justifies this supposed requirement. The statute does not support this reading. It is also sometimes suggested that a market context is needed to establish substantiality of injury. The illogic of that suggestion becomes clear, however, upon even a moment’s reflection. A pound of flour is a pound of flour, whether placed in a small bowl or a large one. So, too, substantiality of injury is the same however relevant markets are defined.
As merger law is understood and enforced today, market definition is not a necessary step in analysis. This is nearly a truism given the primacy of proof of antitrust injury in proving illegality. Yet, the lesson still eludes some courts and advocates, resulting in flawed reasoning like the rejection of claims for no reason other than failure to define a relevant market. Interest in establishing the non-necessity of market definition is thus encouraging. Agency guidance or legislation that clarified that market definition is not a necessary step in rule of reason analysis could improve and focus antitrust analysis.
It would be a grave mistake, however, to go too far and imply that market definition was not a helpful step in analysis. In some cases, market definition can become necessary, as when market shares and other structural considerations are used to draw competitive inferences. And even when it is not necessary, market definition can still help to clarify and direct analysis. Market definition can focus attention; it can identify important actors for understanding challenged conduct; it can contextualize inquiries about future conduct (like the future entry of potential competitors); and it can provide other benefits besides these. In short, while market definition is not always necessary, it is often useful and important in analysis, and it should be omitted only when there are good reasons for doing so.
II. Myth #2 – Relevant Markets Are Freestanding Entities
How many times have proposed relevant markets been criticized as unrealistic? How many times have markets been described as gerrymandered — not for failing to fit a recognized test but for being selected, from among valid options, as the choice that best matches the plaintiff’s theory of the case? How many times have plaintiffs marched into court with nothing but casual intuition to justify their definition of a relevant market? How many times have markets been defined in terms of industry classifications or abstract inquiries divorced from a particular theory of harm?
All of these arguments fail for the same reason: relevant markets are not things that exist in the world, but merely analytical lenses through which we think about particular theories of harm. The market that is relevant for considering a merger’s ability to encourage anticompetitive coordination is not typically the market that is relevant to considering the same merger’s tendency to stifle head-to-head competition between the merging parties. Neither of these markets is the same as the “industry” in which casual speakers might place the firms or perceive anything about trends in consolidation. The emphatic word in “relevant market” is not market but relevant. The point of market definition is to identify a scope of trade that is helpful for assessing a particular theory of harm.
Several reform opportunities flow from this observation. First, since relevant markets cannot be defined in the abstract, market definition should not be described or attempted outside the context of a specific theory of harm. In a merger context, market definition should not be performed before deciding what concerns need to be evaluated. Market definition should be both described and conducted within analysis of a specific theory of harm: one market definition process for coordinated effects analysis, another market definition process for unilateral effects analysis. Second, because confusion about this subject is prevalent, it would be helpful to clearly and unambiguously refute the idea that all relevant markets should align with casual ideas about industry boundaries. Third, because there is not some freestanding “correct” market waiting to be discovered, courts should not approach market definition as presenting a choice between the plaintiff’s and the defendant’s conflicting proposals of market boundaries. So long as plaintiffs’ proposed markets are consistent with appropriate tests, those are the relevant markets that should guide analysis.
III. Myth #3 – There Is One Relevant Market Per Case
One question stands out in the Request for Information: “Are the guidelines sufficiently clear that the same product or service may be in multiple relevant antitrust markets depending on the competitive effects being evaluated?” The answer is, of course, “No.” Despite their many strengths, the 2010 Horizontal Merger Guidelines do not adequately discredit the common misperception that market definition is a discrete step in analysis that produces one relevant market per case.
A simple example illustrates the problem with that thinking. Suppose that a horizontal merger is being evaluated for two separate competitive concerns. First, the elimination of a previously fierce competitor could embolden coordination among a group of five remaining firms. Second, the elimination of head-to-head competition with the acquired company could result in unilateral harm for a subset of consumers. Suppose, also, that the Hypothetical Monopolist Test would validate as relevant markets either the broad group of five firms or the narrow group of just the merging firms (an implication of a substantial unilateral effect, at least in a differentiated-product space). Which of these is the “right” choice of relevant market? The answer is “both” and “neither.” The broad market is the appropriate choice for assessing coordination concerns; the narrow market is the appropriate choice for assessing unilateral effect concerns. Neither is appropriate for assessing the other concern. Unless one or the other of these two theories of harm is dropped, careful analysis of the merger requires two separate instances of market definition.
Agency guidance could helpfully clarify that it is appropriate to define different relevant markets for different theories of harm. The Agencies would have a solid foundation for this position not only in logic but in law. The Supreme Court adopted just such an approach in Brown Shoe.
IV. Myth #4 – There Is Only One Test for Defining Relevant Markets
Sticking with the same Request for Information question addressed in Myth #3, two related misconceptions about tests for defining markets are also implicated by this question. The first is the false belief that there is just one test for defining markets. In some respects, the 2010 Horizontal Merger Guidelines appear to adopt this position. AAG Kanter’s remarks also seem to adopt the error in the claim that “the static formalism of market definition may not always be the most reliable tool for assessing the potential harms of mergers.” The second is the false belief that every test for defining markets is the same. Modest examples of this error include anachronistic reinterpretation of old tests, like the Request for Information’s suggestion that Brown Shoe’s practical indicia test is focused on “qualitative evidence about substitution.” More vivid examples emerge when language about practical indicia is mashed together with soundbites about cross-elasticity of demand and extracted portions of the HMT in constructing a sort of Frankenstein’s monster of a test for defining relevant markets.
Inherent in the idea that the same product or service may be in multiple relevant markets depending on the competitive effects being evaluated is the idea that different tests are needed to define helpful relevant markets for different theories of harm. By its terms, the HMT is a test for defining markets relevant to coordination concerns. It also defines appropriate relevant markets for assessing output suppression theories of unilateral effects. The HMT does not typically define relevant markets that are helpful for evaluating differentiated-product unilateral effects theories. This is not because market definition is unhelpful in evaluating these theories. It simply reflects the need for a different test of market definition in that context. Perhaps something similar applies to AAG Kanter’s concern that “potential harms of mergers” are going unaddressed by current market definition practices. If so, the implication is not that market definition is unhelpful, but that new tests of market definition are needed to address these potential harms.
To correct deep-rooted confusion on this point, Agency guidance should acknowledge that different tests of market definition serve different purposes. Agency guidance should clarify that it is appropriate to define different relevant markets for different theories of harm and that it is appropriate to use different tests to define relevant markets for different theories of harm. Objections to the complexity of this approach do not withstand scrutiny. Nobody criticizes a doctor for using different medical tests to look for signs of different illnesses. Nobody should criticize the Agencies for using different market definition tests to look for signs of different competitive concerns.
V. Myth #5 – Market Definition Is Needlessly Overcomplicated
Another myth is the idea that complexity in market definition could be mitigated by simply changing the rules of market definition. This idea might motivate the Request for Information question that asks: “Do the guidelines imply that precision is necessary or possible in defining relevant markets?” Qualitative reasoning is acceptable and usually necessary in market definition. There is also room to streamline the process — for example, by addressing Myths #3 and #4. But much of market definition’s complexity is derivative of complexity in the underlying theory of harm being evaluated. To simplify market definition, while leaving the underlying theory of harm unchanged, would be to mutilate the connection between relevant markets and the purposes they are meant to serve.
As a concrete example, suppose a merger is being investigated for its potential to lead to anticompetitive coordination. Relevant markets defined by the HMT are useful for this investigation. The test articulated by the HMT could be restated (not simplified) as this: a relevant market is a group of competitors with the joint market power to successfully undertake anticompetitive actions if they were able to perfectly coordinate their conduct. True, this is a complicated inquiry, but the payoff is a market with special relevance to the coordination concern. By construction, participants in the HMT market are a group of competitors with the potential to engage in anticompetitive coordination. By construction, market structure in an HMT market provides relevant information for assessing coordination prospects and incentives. By construction, historic information about the behavior of participants in this market is relevant to assessing coordination concerns.
Now suppose that, instead of using the HMT, we had defined the relevant market using some convenient test like intuition, industry recognition, or a similarly simplistic rule of decision. Effort would be saved, but the resulting market would no longer serve the functions that the HMT market had served. Nothing would guarantee that participants in this new market would be potential colluders. Nothing would relate market structure in this new market to the risk of coordinated conduct. And nothing in the history of this new market would be of obvious relevance to coordination concerns. The market would be just as likely to miss coordination by being too broad (gathering too much competition to reveal important patterns) as too narrow (failing to capture enough competition to see important patterns).
In short, when appropriately conducted, the complexity of market definition reflects the complexity of the underlying theory of harm. Attempting to suppress the complexity of market definition by defining casually intuitive or arbitrary markets with no connection to the underlying theory of harm will only confuse investigations and frustrate enforcement efforts. To state the obvious, the Agencies should not pursue this option.
VI. Myth #6 – Direct Evidence of Market Power Obviates Market Definition
Imprecise thinking and language can lead discussions of direct evidence down dangerous paths. The Request for Information toes this line in asking whether the guidelines should “make it clearer that the tests for an antitrust market can often be satisfied using direct evidence of likely effects.” Individual remarks by Chair Khan ask an apparently similar, but importantly different, question: “Are there certain markets where the guidelines should provide a framework to assess direct evidence of market power?”
Direct evidence can be very valuable in antitrust analysis, but there is a tendency to overstate the frequency with which this happens. In merger cases, direct evidence of prospective mergers is unavailable. Merger simulations and other economic models offer qualitative predictions, not “direct” evidence of competitive effects. Consummated mergers may allow for empirical proof of a merger’s effects, though establishing causality can be challenging. Direct evidence of market power is generally not the same as direct evidence of likely effects. Where the worry is that a merger will increase market power, for example, proof of current market power does little to advance analysis. The error in thinking otherwise could be seen as a reverse Cellophane fallacy: proof of existing market power does not prove ability to gain market power.
In short, direct evidence is often informative for market definition but not often dispositive of it. Proof of past competition can be an important factor in market definition, but mere proof of past competition is rarely enough.
VII. Myth #7 – Market Concentration Has Meaning Independent of How the Market Is Defined
Concerns about concentration motivate the current push for policy reform. The executive order that encouraged the Agencies to review the merger guidelines suggested that they do so “[t]o address the consolidation of industry in many markets across the economy.” Referencing the executive order, but changing its language, AAG Kanter asserts a connection between “concentrated market structures” and harm to trading partners, possibly related to coordination or exclusion. The various ideas reflected in even these two statements reflect a background tendency to treat concentration as though it has significance independent of how the market is defined.
Assuming, for sake of argument, that consolidation means concentration, and that industry means market, it is not true that concentration has the same meaning in every market. Forty years ago, Landes and Posner illustrated this point by showing how the same market share could have different competitive implications as the size of the market and elasticity of demand varied. The economic significance of market concentration also depends on the way that a relevant market is defined. HMT markets are defined in such a way that greater concentration says something about the prospects of coordination. The process of defining the HMT market is what gives concentration this meaning. In a market defined by NAICS code, for example, concentration is not an economically defensible predictor of coordinated effects.
The Agencies have much to gain by leaning into the insight that market concentration gets its meaning from market definition. One implication of this insight is that market concentration thresholds should be applied contextually. High concentration in an HMT market supports an inference that a merger could have coordinated effects. Low concentration suggests it could not. But what low concentration in that HMT market does not do is suggest that the merger is unlikely to have other effects: it could well risk unilateral effects or other harms. The Agencies would both strengthen and clarify merger enforcement by building this reasoning into their use of market concentration thresholds. At a minimum, the triggering of a safe harbor provision in one relevant market should not preclude continued investigation of competitive harms that would arise in other relevant markets.
1 Associate Professor of Law, The University of Iowa College of Law. Contact the author at email@example.com. The author thanks Kassandra DiPietro for research assistance.
2 U.S. Dept. of Justice & Federal Trade Comm’n, Request for Information on Merger Enforcement 5–6 (Jan. 18, 2022) [hereinafter Request for Information].
3 E.g. id. at 3 (asking “should the guidelines make it clearer that the tests for an antitrust market can often be satisfied using direct evidence of likely effects . . . or qualitative evidence about substitution?”).
4 Jonathan Kanter, Modern Competition Challenges Require Modern Merger Guidelines 4–5 (2022).
5 See e.g. Competition and Antitrust Law Enforcement Reform Act, S. 225 117th Cong. §§ 4(b), 9(a), 13 (2021).
6 United States v. E. I. du Pont de Nemours & Co., 353 U.S. 586, 593 (1957).
7 E.g. Brown Shoe Co. v. United States, 370 U.S. 294, 324 (1962) (“The ‘area of effective competition’ must be determined by reference to a product market (the ‘line of commerce’) and a geographic market (the ‘section of the country’).”).
8 See Herbert Hovenkamp, Markets in Merger Analysis, 50 Antitrust Bull. 887, 890–93 (2012) (providing a better reading of the statutory language); Louis Kaplow, Replacing the Structural Presumption § III.B.1 (working paper, Dec 14, 2021) (similar).
9 E.g. Brown Shoe Co., 370 U.S. at 324 (“Substantiality can be determined only in terms of the market affected”).
10 See David Glasner & Sean P. Sullivan, Logic of Market Definition, 83 Antitrust L.J. 293, 339–40 (2020).
11 See e.g. United States v. Pabst Brewing Co., 384 U.S. 546, 549–50 (1966) (“Proof of the section of the country where the anticompetitive effect exists is entirely subsidiary to the crucial question . . . whether a merger may substantially lessen competition anywhere in the United States.”).
12 E.g. City of New York v. Grp. Health Inc., 649 F.3d 151, 155 (2d Cir. 2011) (“To state a claim under § 7 of the Clayton Act, §§ 1 or 2 of the Sherman Act . . . a plaintiff must allege a plausible relevant market in which competition will be impaired.”).
13 E.g. Request for Information, supra note 2, at 5 (“Is it necessary to precisely define the market in every case?”); Competition and Antitrust Law Enforcement Reform Act, S. 225 117th Cong. § 13 (2021) (“Establishing liability under the antitrust laws does not require the definition of a relevant market . . . .”).
14 See Sean P. Sullivan, Modular Market Definition, 55 U.C. Davis L. Rev. 1091, 1117–29 (2021) (discussing various purposes of market definition); Gregory Werden, Why (Ever) Define Markets? An Answer to Professor Kaplow, 78 Antitrust L.J. 729 (2013) (similar).
15 See Glasner & Sullivan, supra note 10, at 299 n.29 (collecting examples of this argument).
16 E.g. Subcomm. on Antitrust, Commc’n. & Admin. L. of the Comm. on the Judiciary, Investigation of Competition in Digital Markets: Majority Staff Report and Recommendations 12, 15, 17 (2020) (declaring properties of “the market for social networking,” “the U.S. online retail market,” and “the mobile operating system market”).
17 Glasner & Sullivan, supra note 10, at 308–09; Magali Eben, The Antitrust Market Does Not Exist: Pursuit of Objectivity in a Purposive Process, 17 J. Comp. L. & Econ. 586 (2021).
18 See Sullivan, supra note 14, at 1114–17 (observing the unhelpfulness of HMT markets in unilateral effects analysis).
19 See id. at 1108–09 (observing the unhelpfulness of industry concepts in coordinated effects analysis); see also id. at 1124, 1126–27 (discussing how industry concepts did respond to Warren-Court era concerns about industrial concentration as it was popularly perceived).
20 Request for Information, supra note 2, at 5.
21 See Glasner & Sullivan, supra note 10, at 331 (“Despite drawing a sharp distinction between the analysis of unilateral and coordinated effects, the 2010 Horizontal Merger Guidelines can be read to imply that the same relevant market should be used to assess both of these different theories of harm.”); see also id. at 326–30 (describing the “single market fallacy”).
22 Sullivan, supra note 14, at 1116.
23 See Glasner & Sullivan, supra note 10, at 333 (working through a similar thought experiment).
24 Brown Shoe Co. v. United States, 370 U.S. 294, 325–28 (1962) (defining the relevant market for analysis of the vertical aspects of the merger); id. at 336–39 (separately defining the relevant market for analysis of the horizontal aspects of the merger).
25 Compare U.S. Dep’t of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines § 4 (Aug. 19, 2010) (presenting the Agencies’ approach to market definition in merger cases), with id. § 6.1 paras. 6–7 (describing unilateral effects analysis as not requiring market definition).
26 Kanter, supra note 4, at 4.
27 See Sullivan, supra note 14, at 1096–97 (labeling this thinking the “hammer analogy” of market definition).
28 Compare Request for Information, supra note 2, at 3; with Glasner & Sullivan, supra note 10, at 302 (noting that Brown Shoe’s discussion and test of market definition was “never meant to fit the focus of modern antitrust analysis”); Sullivan, supra note 14, at 1105 (“Only a few of the practical indicia have any plausible connection to antitrust’s current focus on market power and constraints on that power.”).
29 E.g. United States v. Aetna Inc., 240 F. Supp. 3d 1, 20–21 (D.D.C. 2017) (presenting multiple different tests for defining markets as though they were interchangeable).
30 See Sullivan, supra note 14, at 1107–11, 1130–33.
31 Id. at 1136–37.
32 Id. at 1111–17.
34 See supra note 26 and accompanying text.
35 Request for Information, supra note 2, at 5.
36 Sullivan, supra note 14, at 1110 (“At base, the HMT defined a market as a group of competitors who could, at least under ideal circumstances, collude to jointly raise their prices.”).
37 Id. at 1111.
38 Id. at 1123–24.
39 Id. at 1125–26.
40 Request for Information, supra note 2, at 3.
41 Fed. Trade Comm’n, Remarks of Chair Lina M. Khan Regarding the Request for Information on Merger Enforcement 3 (Jan. 18, 2022); see also id. (“What types of indicia of market power should the guidelines consider?”).
42 See Sullivan, supra note 14, at 1116–17, 1127–28.
43 See 5 Philip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 1205 at 310–11 (4th ed. 2014) (noting difficulty of proving causality of post-merger price increases).
44 Cf. Luke M. Froeb & Gregory J. Werden, The Reverse Cellophane Fallacy in Market Delineation, 7 Rev. Indus. Org. 241 (1992) (describing a different but related reverse Cellophane fallacy).
45 Exec. Order No. 14,036, 86 Fed. Reg. 36,987 § 5(c) (July 14, 2021).
46 Kanter, supra note 4, at 2.
47 See William M. Landes & Richard A. Posner, Market Power in Antitrust Cases, 94 Harv. L. Rev. 937, 954–55 (1981).
48 Sullivan, supra note 14, at 1123 (“[M]arket definition really is where this translation [of market features into economic significance] takes place.”).
49 Id. at 1145–47 (making this point in greater detail).