Abstract: Market power can be assessed in antitrust cases by two methods often described as “alternatives”. For unilateral conduct, however, they are inconsistent. Courts insist on market shares well above 50%, which entails that only one firm can be a monopolist. Direct measures by examining price-cost margins, however, can sometimes produce two or more “monopolists” in the same market. Such outcomes are inconsistent with the historic and economic meaning of “monopoly”.
***
Prior to the Sherman Act, Anglo-American law used the term “monopoly” to describe a firm with an exclusive right from the government, either by patent or a grant in a corporate charter.[1] The drafters of the Sherman Act intended to reach further. By 1905, the Supreme Court acknowledged that the term “monopoly” was no longer “confined to a grant of privileges,” but also included “a condition produced by the acts of mere individuals.”[2]
Monopolies created by a patent or government grant are easy to identify by the language that creates them, but what about monopolies that result from private conduct? For those, one must look at the relationship between the firm and the market. Exclusive grants gave firms 100% of the territory covered by the grant, and even the legal authority to remove rivals from their exclusive space.[3] From the beginning, however, antitrust policy confronted dominant firms that had competitors. The Supreme Court has never given a decisive answer to the question of how much of the market a firm must control in order to be a Sherman Act monopolist. In its 1927 International Harvester decision, which involved the interpretation of an antitrust consent decree, the Court suggested that a firm no longer had monopoly power when its market share was 64% and declining.[4] Building on that, Judge Hand famously concluded in the Aluminum case (1945) that a 33% market share was clearly not enough, a 64% share was “doubtful,” but that a 90% market share was sufficient for unlawful monopolization.[5]
The Supreme Court’s American Tobacco decision, a year later, endorsed Judge Hand’s conclusions but also made an important distinction between the power to raise prices and the power to exclude.[6] That decision involved three large tobacco companies (American, Liggett, and Reynolds). They had been convicted of cartel violations of §1 of the Sherman Act as well as conspiracy to monopolize under §2. Without stating a precise market share, the court held that a conspiracy to monopolize involved “the power to control and dominate interstate trade,” and this required more than the power to fix prices.[7] The Court aggregated the market shares of the three defendants. Although the three firms’ market shares were falling, the Court found it sufficient that they consistently accounted for more than 68% of the market, and sometimes more than 75%.
No Supreme Court decision has ever reached that low for purely unilateral conduct, as opposed to a conspiracy.[8] The Boxing Club decision in 1959 concluded that 81% was sufficient.[9] A few years later, the Court defined monopoly power as the “power to exclude competitors” and concluded that an 87% market share sufficed.[10] The minimum share requirement experienced some downward creep in some lower court decisions. The Second Circuit’s Broadway Delivery decision held that a 50% share “rarely’ showed monopoly power, a share between 50% and 70% could “occasionally” show it, and a share above 70% is unusually strong evidence of monopoly power.[11]
Other federal courts have tracked the Supreme Court more closely. The Tenth Circuit concluded in 1989 that courts “generally require a minimum market share of between 70% and 80%.[12] The Fifth Circuit observed that monopoly power was “rare” with a market share of less than 70%.[13] The Ninth Circuit suggested a 65% minimum.[14] The Fourth Circuit indicated that 70% was enough.[15] The Third Circuit found 75% to 80% sufficient in one case,[16] and held in another that 55% would not be enough unless other factors indicated power “to restrict entry, supply, or price.”[17] In Microsoft, the D.C. Circuit found a market share exceeding 90% in an operating system market that excluded Apple’s Mac OS was sufficient, and even if MAC OS were included, it exceeded 80%.[18]
Most recently, Judge Boasberg observed in the 2025 Meta (Facebook) case that the Supreme Court has “never found a party with less than 75% market share to have monopoly power.”[19] He concluded that Meta’s market share was “almost certainly under 54%” and falling, and that was in a market that did not even include YouTube, which the court believed should be included. He also noted that while “almost every business enjoys some degree of market power,” this is “a far lower bar than monopoly power.”[20]
Market share is a sensible metric for measuring monopoly when it enables a court to assess a firm’s unilateral ability to control the market – that is, to raise the market price by reducing its own output. If competitors can simply increase their own production to offset the larger firm’s output reduction, however, market prices will not rise. While a firm with a larger market share is in a better position to pull this off, two other factors are essential.[21] First, the market elasticity of demand must be low, which requires that substitution between goods inside and goods outside of the defined market be minimal. If a large firm’s output reduction triggers a flood of goods entering from outside, the market must be defined more broadly.
Second, the supply elasticity of competing firms, or their ability to increase sales in the market, must also be low. This number is particularly critical as a defendant’s market share falls further below 100%. At one extreme, if rivals have no ability whatsoever to increase their own output, then even a firm with a fairly low market share can raise the market price by reducing its own output. Rivals will not be able to make up the shortfall. At the other extreme, if rivals can increase output instantaneously and at no cost, then even a firm with 90% of the market will have no power to raise the market price. As soon as it reduces output, rivals will immediately compensate with an offsetting output increase, assuming they are behaving competitively.
This fact serves as a special warning in cases involving purely digital products, such as Facebook. For example, a digital rival like TikTok would likely be able to sign up new members very quickly and at very low additional cost. If a large firm’s output reduction of, say, 1000 units is immediately offset by a 1000 unit increase by rivals, then competition has not been injured at all. The only thing that will happen is that the larger firm will lose market share. For purely digital output, the cost of adding further users is typically close to zero. The extent to which rivals can offset the dominant firm’s output decrease is, of course, a question of fact. Some firms may be limited by capacity constraints that make them unable to expand, but that is uncommon enough in digital markets that it would have to be proven.[22] The more serious takeaway is that determining a firm’s market share is only one part of the assessment of market power. A court must also ensure that the market is well defined and that rivals are not in a position to offset the larger firm’s reduction in output.
Determining the share of a relevant market has been a nearly universal method for assessing power in monopolization cases. However, “direct” alternatives are available, and some courts have discussed them. What is not clear is whether these methodologies can be applied to claims of single firm monopoly, or must be limited to “collusive” practices such as cartels or mergers. Courts also dispute whether a relevant market must be defined in a monopolization case when direct method alternatives are available.[23] Here, however, the Supreme Court itself has insisted that a relevant market must be defined in an attempt to monopolize case.[24] Further, that decision made clear that it was “beyond doubt” that a relevant market muse be established in monopolization cases as well.[25] So as a matter of Supreme Court law, a monopolization or attempt case requires a market definition.
The more technical direct measurements assess power by examining price-cost margins or, more controversially, profit levels. They query whether a firm can prices above its costs, and do not require a market definition. Some of these metrics are simply not useful for assessing monopoly power. For example, in Meta the court noted that Facebook engaged in price discrimination, which implies that a firm has power in at least that portion of the market being charged the higher price. As Judge Boasberg noted, the problem with using price discrimination is that while it can identify the presence of some power, it does not provide a good basis for identifying monopoly power. In fact, price discrimination is fairly common in nearly any product differentiated market, and is often used by firms that are clearly nondominant.[26]
Another problem with estimating power by reference to price/cost margins is that prices above marginal cost are almost universal in product-differentiated markets for manufactured products, particularly if fixed costs are substantial. In one prominent study economist Robert Hall concluded that markups across manufacturing averaged around 40-50% above marginal cost.[27] But clearly not everyone is a monopolist. It also seems to be the case that, while price-cost margins have lately been rising the cause is more often falling costs rather than rising prices.[28] This creates false positives to the extent that antitrust policy does not wish to be in the business of condemning lower costs instead of higher prices.
An additional problem with direct measurement is that it can be very difficult or even impossible to distinguish monopoly power from high markups that result from the cooperation of rivals – that is, oligopoly or even collusion. For example the Dimmitt decision suggested that when market shares were low the most likely assumption is that the industry is an oligopoly and that the exercise of the defendant’s market power was “dependent upon joint action by at least some of its rivals.”[29] Both a monopolist and fellow oligopolists or cartel members could have high price-cost margins.
A final serious conceptual problem is that if our only criterion for identifying monopoly power is high margins or profits, then it becomes possible for two or more firms to be “monopolists” in the same market. That idea was given credence in the Geneva Pharmaceuticals decision, which held that monopoly power could be proven “directly through evidence of control over prices or the exclusion of competition.”[30] Following that, in the US Airways case a court permitted a plaintiff’s claim where the defendant’s market share ranged from 49 to 52 percent. In addition, the plaintiff showed prices well above the competitive level; excessive profits; a flow of rent payments that would not have existed in a competitive market; the ability to price discriminate; maintenance of a stable market share even with obsolete technology; and high entry barriers, with no new entrants in some 30 years. The court rejected the defendant’s protest that under this methodology three firms had monopoly power in the market at issue. It could not find “any controlling case law to suggest that an entity cannot be liable under §2 for exercising the power to control prices or exclude competition in a market because some other entity also exercises the same power in the market.”[31]
That outcome seems flatly inconsistent with the Supreme Court’s twin propositions noted above, that a relevant market must be defined in a monopolization case and that market shares must be substantially greater than 50%. In that case indirect and direct measures are not merely “alternative” ways of measuring power, they are also contradictory. The cases that measure monopoly power by reference to market share insist that a market can have only one monopolist. And indeed, that is both the historical and intuitive meaning of “monopoly”. So jumping to direct measure is not merely seeking out an alternative set of tools, but it also introduces a fundamentally different conception of monopoly.
What are we looking for when we speak about monopoly power? Do we mean simple power to charge a price substantially above cost; or do we mean market dominance, which is the power to control market prices? Concluding that two or more firms in a market are monopolists is inconsistent with the definition in early cases such as Grinnell that monopoly power is the power to exclude competition. Neither of two “monopolists” has excluded the other one.
While econometric tools are useful for identifying market power, it seems clear that the evidence from whatever source must also establish dominance. As a result, market shares of less than 50% should automatically be dismissed. Even when market power is established more directly from econometric evidence, that evidence must also indicate that only one firm in this particular market has the requisite power, and that it includes the unilateral power to raise the market price. That is the only rule that does justice to the meaning of “monopoly”.
Herbert Hovenkamp
Citation: Herbert Hovenkamp, The Enigmas of Monopoly Power, Network Law Review, Fall 2025
References
[1] Under English common law, see William L. Letwin, The English Common Law Concerning Monopolies, 21 Univ. Chi. L. Rev 355 (1954); in American common law, see Joseph K. Angell & Samuel Ames, A Treatise on the Law of Private corporations Aggregate (1832).
[2] National Cotton Oil Co. v. Texas, 197 U.S. 115, 129 (1905).
[3] State v. Milwaukee Gaslight Co., 29 Wis. 454 (1872) (state granted monopoly of right to produce gas entitled its owner to exclude other producers). Other decisions are discussed in Herbert Hovenkamp, Antitrust Policy Design, Ch. 1 (forthcoming, 2026).
[4] United States v. Int’l Harvester, 274 U.S. 693, 709 (1927).
[5] United States v. Aluminum Co., 148 F.2d 416, 424 (2d Cir. 1945).
[6] United States v. American Tobacco Co., 328 U.S. 781, 813 (1946).
[7] Id. at 788-790, 795, citing United States v. Socony-Vacuum Oil Co., 310 U.S. 150 (1946).
[8] The Antitrust law treatise is more aggressive, recommending 60% of a well-defined market and held for at least five years. Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶801a (5th ed. 2022).
[9] International Boxing Club v. United States, 358 U.S. 242 (1959).
[10] United States v. Grinnell Corp., 384 U.S. 563, 567 (1966).
[11] Broadway Delivery Corp. v. United Parcel Serv. Of Am., Inc., 651 F.2d 122, 129 (2d Cir. 1981). One outlier District Court decision concluded that a market share below 30% would not foreclose the possibility of unlawful monopolization. In re Payment Card Interchange Fee & Merch. Disc. Antitrust Litig., 562 F.Supp.2d 392, 400 (E.D.N.Y. 2008).
[12] Colo. Interstate Gas Co. v. Nat. Gas Pipeline Co. of Am., 885 F.2d 683, 694 n.18 (10th Cir. 1989).
[13] Exxon Corp. v. Berwick Bay Real Estate Partners, 748 F.2d 937, 940 (5th Cir. 1984).
[14] Image Tech. Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195, 1206 (9th Cir. 1997).
[15] E.I. du Pont de Nemours & Co. v. Kolon Indus., Inc. (Kolon I), 637 F.3d 435, 450 (4th Cir. 2011).
[16] United States V. Dentsply, Int’l, 399 F.3d 181, 184-188 (3d Cir. 2005).
[17] Fineman v. Armstrong World Indus., Inc., 980 F.2d 171 (3d Cir. 1992).
[18] United States v. Microsoft Corp., 253 F.3d 34, 54-56 (D.C. Cir. 2001).
[19] FTC v. Meta Platforms, Inc., __ F.Supp.3d __, 2025 WL 3458822 (D.D.C. Dec. 2, 2025), quoting Kolon Indus., Inc. v. E.I. DuPont de Nemours & Co., 748 F.3d 160, 174 (4th Cir. 2014).
[20] Id. at __ citing Richard Schmalensee, Another Look at Market Power, 95 Harv. L. Rev. 1789, 1790 (1982); and Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶501 (5th ed. 2023).
[21] Willam M. Landes & Richard A. Posner, Market Power in Antitrust Cases, 94 Harv. L. Rev. 937 (1981).
[22] See Herbert Hovenkamp, Antitrust and Platform Monopoly, 130 Yale L.J. 1952, 1990 (2021).
[23] Neurontin Antitrust Litig., 2013 WL 4042460, 2013-2 Trade Cas. ¶78,478 (D.N.J. Aug. 8, 2013).
[24] Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447 (1993).
[25] Id. at 457 (section 2 “applies to charges of monopolization as well as to attempts to monopolize, and it is beyond doubt that the former requires proof of market power in a relevant market. United States v. Grinnell Corp., 384 U.S. 563, 570–571 (1966); United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 404 (1956).
[26] 2B Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶517 (5th ed. 2022).
[27] Robert E. Hall, The Relation Between Price and Marginal Cos in U.S. Industry, 96 J. Pol. Econ. 921 (1988).
[28] See, e.g., Christopher Conlon, Nathan H. Miller, Tsolmon Otgon, & Yi Yao, Rising Markups, Rising Prices?, 112 AEA Pap. Proc. 279 (2023).
[29] Dimmitt Agri Indus., Inc. v. CPC Intern., Inc., 679 F.2d 516, 531 (5th Cir. 1982), following Tops Markets, Inc. v. Quality Markets, Inc., 142 F.3d 90, 97-98 (2d Cir. 1998).
[30] Geneva Pharmaceuticals Tech. Corp. v. Barr Laboratories, Inc., 386 F.3d 485, 500 (2d Cir. 2004).
[31] US Airways, Inc v. Sabre Holding Corp., 2022 WL 874945, *10 (S.D.N.Y. March 24, 2022).