The essential facilities doctrine has received a resurgence of interest recently, especially with regard to platform markets. Many references to the doctrine exaggerate its centrality to antitrust law; it is a “doctrine” in name only. The question is why the essential facilities doctrine has been so popular among recent proponents while it remains unpopular with courts. The answer lies in the doctrine’s ability to simplify inquiries into anticompetitive conduct, which can be complicated in platform markets. The next question, then, is whether that is a good thing. I conclude that it is not. The essential facilities doctrine, appealing as it is for solving a host of competitive problems posed by large Internet platforms, is likely to lead antitrust commentators and courts to ignore more significant – but much harder – questions of what constitutes anticompetitive conduct in the context of those markets. As a theory of liability, the essential facilities doctrine shifts attention away from the question of anticompetitive conduct, which is a question that deserves more attention, not less. Instead, the concept of essential facilities is better thought of as a remedy rather than as a distinct theory of antitrust liability.

By Thomas B. Nachbar[1]

The essential facilities doctrine has received a resurgence of interest recently. Commentators have been advancing it[2] for its potential to bring meaningful antitrust regulation to platforms,[3] and regulators have similarly suggested either the doctrine itself[4] or remedies sounding in essential facilities, such as the requirement for monopolists to provide nondiscriminatory access to their platforms,[5] as ways to tame the commercial power of those platforms. I’ve described elsewhere why I think the essential facilities doctrine is a poor match for platform regulation.[6] My larger point here is a more general one about the essential facilities doctrine: that the thing that makes it popular among commentators and regulators is the same thing that suggest that courts and commentators should eschew it. The essential facilities doctrine is too useful to be useful.

Although frequently presented (especially in popular media) in terms such as “a long-standing element of antitrust law,”[7] the essential facilities doctrine is anything but.[8] Indeed, given how infrequently it is applied, one hesitates to refer to it as a “doctrine” so much as a “theory” or, because I doubt that it actually serves as a distinct theory of liability at all, maybe just a “concept.”

Essential facilities might be unique among antitrust theories for having its most recognizable references be in the form of criticisms. Phil Areeda’s 1989 article on essential facilities, Essential Facilities: An Epithet in Need of Limiting Principles,[9] arguably remains the definitive work on the doctrine,[10] and the only time the Supreme Court addressed the doctrine by name, it refused to recognize its existence.[11] Given the lack of favorable Supreme Court precedent, citations to the doctrine usually point to U.S. Court of Appeals cases like MCI v. AT&T, in which the Seventh Circuit laid out the elements of the claim:

(1) control of the essential facility by a monopolist

(2) a competitor’s inability practically or reasonably to duplicate the essential facility

(3) the denial of the use of the facility to a competitor, and

(4) the feasibility of providing the facility.[12]

The doctrine is a creature of lower court decisions, although even there, successful claims are uncommon.[13]

It is, nevertheless, extremely popular among academics. Even though there are arguably zero Supreme Court cases applying the doctrine,[14] there are many, many law review articles offering to “revitalize,”[15] “revive,”[16] renew[],”[17]or otherwise resurrect the essential facilities doctrine to solve any range of competitive ills, from self-dealing by Internet shopping platforms to social media platform refusals to allow application interfaces. The question is: Given that the doctrine is so rarely applied by courts, why is it so popular among commentators and regulators? And why is a doctrine so popular with the aforementioned commentators and regulators so disfavored by courts?

The answer, I think, lies not in Areeda’s cogent criticism of the essential facilities as a doctrine in need of limiting principles. Areeda himself offered six,[18] which haven’t been as helpful in application as Areeda’s criticism of the doctrine itself. For instance, the first of the six argues that the doctrine should only be applied in “very exceptional” cases while the second concedes the doctrine should be applied when a competitor cannot effectively compete without the facility and cannot practically replicate the facility.[19] Both limitations are hopelessly vague, and the second essentially repeats the requirements of the doctrine as nominally applied, negating the “very exceptional” cautioning of the first. None of the others, I think, have fared much better.[20]

Many scholars have effectively offered their own limits on the doctrine by focusing on particular subject matter. The doctrine is most frequently invoked with regard to specific industries, or at least industries displaying particular characteristics. It is a staple of commentary on regulation of the Internet,[21] a field of research closely related to communications law. Perhaps the leading modern proponent of the doctrine, Spencer Weber Waller, has argued (in conjunction with Brett Frischman) that it should be applied more aggressively to “infrastructure,” a set of industries (or resources) defined by economic characteristics that make sharing both easier and particularly beneficial,[22] such as roads, communications networks, and electricity grids, but also courts, schools, and basic research.[23] Some of these items (like communication networks and electricity grids) are subject to other regulatory regimes, some (like roads, courts, and public schools) are provided by government and unlikely objects of antitrust regulation, but others (like basic research) open the aperture of antitrust considerably. This is the move underlying Lina Khan’s argument that the essential facilities doctrine should be applied to Amazon. Khan, the current chair of the FTC, relies heavily on arguments like Frischmann and Waller’s to analogize Amazon to a public utility.[24] One, of course, could make the same argument for any major Internet platform, such as Facebook,[25] or, indeed, for access to any product that exhibits network effects.[26]

Those subject-matter limits, as much as they might increase one’s comfort with application of the doctrine, come not from the doctrine itself but rather from its proponents; there’s nothing inherent in the essential facilities doctrine that limits its application to firms having those characteristics. Many lower court cases applying the doctrine concern firms that resemble public utilities, such as the MCI case cited above. The Supreme Court actually suggested in Trinko that the availability of other forms of regulation (including utility regulation) make those cases poor candidates for the essential facilities doctrine,[27] but even more significantly, two of the four Supreme Court cases usually cited for the doctrine, Aspen Skiing and Associated Press, involved businesses as about as far from utilities (a ski resort and a newspaper wire service) as imaginable.[28]

But if the doctrine is poorly specified seemingly unlimited as to its industrial subject matter, it is even less so with regard to the conduct it seeks to police. The trigger for scrutiny under the essential facilities doctrine is a refusal to deal. But a refusal to deal itself tells one very little about the circumstances underlying that refusal. The elements of a Section 2 violation are market power and anticompetitive conduct,[29] and consequently, in order for a refusal to deal to be a Section 2 violation, it must be an expression anticompetitive conduct.[30] The essential facilities doctrine lumps together refusals to deal that might have very little to do with each other. Both Apple[31] and Facebook[32] have recently been accused of refusing to deal in violation of the antitrust laws, but their business models are completely different and so are the likely reasons (good or bad) for their relative refusals. Any such nuance, though, is lost by labeling either one of them “essential.”

As theories of antitrust become increasingly sophisticated, antitrust needs to pay more attention to anticompetitive conduct, not less, and the essential facilities doctrine, by focusing on whether the facility is essential, shifts the focus of the inquiry away from the reasons for the refusal and toward the refusal itself (or whether the facility is truly essential). There is not, under typical formulations of the doctrine, any distinct requirement that the refusal be anticompetitive.[33] The doctrine is not a method for analyzing whether a specific refusal is anticompetitive, it is a conclusion that such a refusal is anticompetitive. Areeda correctly noticed this problem in his treatment of essential facilities, suggesting as one of his limits a rule-of-reason inquiry into the refusal, but he provided little guidance on how that would happen,[34] for good reason. If one were to conduct a normal Section 2 analysis on the underlying refusal, the essential facilities doctrine would serve no function. The market power inherent in finding that the input is essential combined with an anticompetitive refusal would be sufficient for a Section 2 violation,[35] obviating the need for the essential facilities doctrine as a distinct theory of liability. That’s why it is so hard to characterize cases like Aspen Skiing as essential facilities cases – because the evidence in that case showed conduct (a refusal to sell even at retail and the deviation from a past course of dealing) that could otherwise be characterized as anticompetitive.[36] When combined with market power, that refusal would be sufficient to support a claim under Section 2. The essential facilities doctrine only does any work when the plaintiff cannot explain how the refusal is otherwise anticompetitive,[37] and that itself should be reason for caution.

The potential reach of the essential facilities theory can be demonstrated by reconsidering any number of antitrust cases as essential facilities cases. In Eastman Kodak v. Image Technical Services v. Eastman Kodak,[38] Kodak had refused to sell (and entered into agreements with suppliers to refuse to sell) Kodak copier parts to independent service organizations (ISOs).[39] The parts were essential to the ISOs in order to compete with Kodak in providing service to Kodak copiers, there was no other source for the parts, and Kodak could easily have provided the parts, so that should be enough alone to establish the essential facilities case. But what made the case against Kodak compelling was not the refusal to deal itself but that it was combined with a pattern that otherwise showed that Kodak’s refusal was intended to be, and was, anticompetitive.[40] Kodak could have been an essential facilities case, but if it had been, the underlying conduct the Court found so critical to deciding the case would have been irrelevant.[41]

One way to square the essential facilities theory with antitrust doctrine would be to acknowledge the addition of a fifth element in cases like Aspen Skiing (or my re-imagined Kodak): that the refusal to deal be anticompetitive. That makes sense of the theory, but because a Section 2 violation requires only market power and anticompetitive conduct, the addition of the anticompetitive conduct element essentially transforms the essential nature of the facility into an inquiry into the defendant’s market power, as described above. [42] I actually think the doctrine might be useful for those purposes. The current inquiry into market power, dominated as it is by what Louis Kaplow calls the market share/market power paradigm,[43] is both irrational and has led to an anemic understanding of market power.[44] Using the essentiality of the input as part of the inquiry into market power might be one way the essential facilities concept can contribute to Section 2 analysis.

By shifting the focus away from conduct, the essential facilities doctrine could potentially simplify antitrust cases, except that its other requirements negate any such simplification. Mandating access, and placing a court in charge of monitoring it, places courts in an inherently fraught role.[45] Managing access is a complicated task on its own. In the late twentieth century, when telephone networks started becoming useful for things other than telephony, the FCC mandated that AT&T provide access for providers of other services to be carried on those networks. But even that comparably simple access mandate gave rise to a years-running set of regulatory proceedings to define and manage those network interfaces.[46] Access mandates only work if the need for access and the definition of access itself are relatively stable.[47] In modern Internet platform markets, neither of those things are true. Nor is the business context surrounding access subject to ready and stable definition. The doctrine only applies to refusals to deal with competitors,[48] but in the modern markets in which proponents would apply the essential facilities doctrine, it is sometimes difficult to identify exactly what the competing products are, much less whether two firms are in competition.

Instead of being a tailored solution to a specific form of anticompetitive conduct, the essential facilities doctrine is a solution in search of a problem – it is a generic, but lightly theorized, form of traditional utility regulation. Like modern information platforms, industrial age monopolies presented a host of problems, from the potential for overcharging to limiting access to both potential customers and competitors. Comprehensive utility regulation could solve all of those problems by controlling many aspects of how those firms did business without the need to identify a particular practice. That might be why Lina Khan, after suggesting the essential facilities doctrine as a response to Amazon, later moved on to suggest a different remedy more closely tied to utility regulation – structural separation – to cure Amazon’s ills.[49]

And that might be the best lesson for the essential facilities doctrine: that its power is remedial. Mandatory access has been used in previous antitrust cases involving platforms, like the one against Microsoft for its anticompetitive conduct related to Windows and Internet browsers.[50] Cases like Otter Tail and Aspen Skiing, and the essential facilities theory itself, might best be considered a category of remedy – access – rather than a distinct theory of antitrust liability.[51] That modification to the essential facilities literature would do much to bring it more into line with the Supreme Court’s antitrust doctrine.

We should be wary of any doctrine with the selectivity of a hammer because it is likely to turn any number of controversial business practices into nails.[52] The essential facilities doctrine, which has been offered as the solution to so many problems, presents exactly that risk. The doctrine is little more than a direct assault on the notion of anticompetitive conduct, which is an area of antitrust law that requires more attention, not less.

The question is not whether the essential facilities theory should be its own theory of liability – it clearly should not. Given its moribund status in the courts, the essential facilities concept serves primarily as fodder for conversations among academics, not as a rule of decision in antitrust cases. Because it is being applied so infrequently and unevenly, the real question is whether the concept of essential facilities is helpful for understanding antitrust law. Its lack of precision and amenability to being molded to fit virtually any refusal by a dominant firm to deal with its competitors suggests that it is not.

Click here for the full article.

[1] F.D.G. Ribble Professor of Law, University of Virginia School of Law.

[2] E.g. Donald I. Baker & William S. Comanor, A U.S. Antitrust Agenda for the Dominant Information Platforms, 35 Antitrust 66, 69-70 (2021); Nikolas Guggenberger, Essential Platforms, 24 Stan. Tech. L. Rev. 237 (2021); Lina Khan, Amazon’s Antitrust Paradox, 126 Yale L.J. 710, 794-802 (2017); Sandeep Vaheesan, Reviving an Epithet: A New Way Forward for the Essential Facilities Doctrine, 2010 Utah L. Rev. 911, 912 (2010).

[3] On platforms, see Thomas B. Nachbar, Platform Effects, 62 Jurimetrics J. 1, 8-10 (2021).

[4] See Majority Staff of the Subcommittee on Antitrust, Commerce & Administrative Law. of the House Committee on the Judiciary, 116th Cong., Investigation Of Competition In Digital Markets 397 (2020); Khan, supra note 1, at 794-802. Although a law student when she first suggested the use of essential facilities doctrine for Amazon, Khan currently serves as Chair of the Federal Trade Commission.

[5] E.g. American Innovation and Choice Online Act, S. 2992, 117th Cong. (2021).

[6] Nachbar, supra note 3, at 38-42. The economic characteristics of platforms are such that they have little incentive to deny access solely to leverage themselves vertically into markets that are likely to be more competitive than the platform market. Id.

[7] Russell Brandom and Adi Robertson, Apple asks court to rule iOS is not an “essential facility,” The Verge (May 19, 2021), https://www.theverge.com/2021/5/19/22443616/apple-requests-court-dismiss-epic-essential-facilities-claim-fortnite-trial.

[8] See Spencer Weber Waller, Areeda, Epithets, and Essential Facilities, 2008 Wisc. L. Rev. 359, 360 (describing the doctrine as “charitably speaking, hanging by a thread”).

[9] Phillip Areeda, Essential Facilities: An Epithet in Need of Limiting Principles, 58 Antitrust L.J. 841 (1989).

[10] Of the 1,156 articles in the Westlaw JLR database that mention the essential facilities doctrine as a feature of antitrust law, over 20 percent of them cite Areeda’s article on essential facilities.

[11] Verizon Communications, Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 409 (2004). The Court did acknowledge the existence of a unilateral “duty-to-deal” claim in antitrust law in Pacific Bell Telephone Co. v. linkLine Communications, Inc., 555 U.S. 438, 455 (2009), although only in the course of rejecting a different (price squeeze) claim. See also IIIB Phillip E. Areeda and Herbert Hovenkamp, Antitrust Law ¶ 770 at 199 (5th ed. 2020). (alternatively referring to it as the “essential facility” doctrine and “unilateral refusals to deal”).

[12] MCI Communications Corp. v. AT&T Co., 708 F.2d 1081, 1133-34 (7th Cir. 1983).

[13] See generally Phillip E Areeda & Herbert Hovenkamp, Antitrust Law ¶ 722 at 97-98 (2022 supp.) (collecting cases).

[14] Areeda, supra note 9, at 847-50; Nachbar, supra note 3, at 38.

[15] Brett Frischmann & Spencer Weber Waller, Revitalizing Essential Facilities, 75 Antitrust L.J. 1 (2008).

[16] See Vaheesan, Reviving an Epithet, supra note 2.

[17] Guggenberger, supra note 2, at 314 (describing “inspiration for a renewed essential facilities doctrine”).

[18] Areeda, supra note 9, at 852-53.

[19] Id. 852.

[20] The sixth, for instance, suggests the doctrine should only be applied when a regulator is available to oversee implementation, id., but the existence of such a regulator, the Court suggested in Trinko, was a reason for not applying the doctrine. Verizon Communications, Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 412 (2004).

[21] See, e.g. Marina Lao, Networks, Access, and “Essential Facilities”: From Terminal Railroad to Microsoft, 62 S.M.U.L. Rev. 557 (2009).

[22] According to Frischmann and Waller,

Infrastructure resources tend to satisfy the following demand-side criteria:

        1. The resource may be consumed non-rivalrously;
        2. Social demand for the resource is driven primarily by downstream productive activity that requires the resource as an input; and
        3. The resource is used as an input into a wide range of goods and services, including private goods, public goods, and/or non-market goods.

Frischmann and Waller, supra note 15, at 12. See also Waller, supra note 8, at 361 (arguing that infrastructure theory provides a justification for the theory).

[23] Id. at 11, 14.

[24] Khan, supra note 2, at 801-02.

[25] Guggenberger, supra note 2, at 320-21.

[26] See Lao, supra note 21, at 559.

[27] Trinko, 540 U.S. at 412.

[28] Indeed, the application of something resembling utility regulation on a newspaper wire service could potentially raise serious First Amendment concerns. See Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622 (1994).

[29] United States v. Grinnell Corp., 384 U.S. 563, 570­71 (1966).

[30] Trinko, 540 U.S. at 407-09. Areeda and Hovenkamp, supra note 8, ¶ 770 at 199 (5th ed. 2020).

[31] EPIC Games, Inc. v. Apple, 559 F. Supp.3d 989 (N.D. Cal. 2021).

[32] FTC v. Facebook, Inc., 560 F. Supp.3d 1, 21-23 (D.D.C. 2021).

[33] See supra the text accompanying note 12.

[34] Areeda, supra note 9, at 852.

[35] Cf. Marian HealthCare, LLC v. Southern Illinois Hospital Svcs., 41 F.4th 787, 792 (7th Cir. 2022) (Easterbrook, J.) (suggesting that, absent an allegation of market power, it cannot be the case that an input is “essential” for purposes of the doctrine).

[36] Whether the refusal was anticompetitive is subject to considerable doubt, and the Court keyed upon features, such as the termination of a prior course of dealing, that made the case very unusual. Trinko, 540 U.S. at 409. The case highlights how every refusal to deal is unique and cannot be lumped together under an umbrella doctrine like the essential facilities doctrine.

[37] Thus, Areeda and Hovenkamp label such refusals “arbitrary,” an attempt to define actionable refusals to deal negatively. See Areeda and Hovenkamp, supra note 9, ¶ 770 at 198.

[38] Eastman Kodak Co. v. Image Tech. Servs., Inc., 504 U.S. 451, 458-62 (1992).

[39] Id. at 465.

[40] Id. at 483-85.

[41] See Areeda and Hovenkamp, supra note 13, ¶772 at 244.

[42] See supra, the text accompanying note 34.

[43] Louis Kaplow, Why (Ever) Define Markets?, 124 Harv. L. Rev. 437, 446 (2010).

[44] See Thomas B. Nachbar, Qualitative Market Definition, 109 Va. L. Rev. 373 (2023).

[45] See Areeda and Hovenkamp, supra note 13, ¶ 774e at 309-13.

[46] On the years-long evolution of those requirements, see Glen O. Robinson and Thomas B. Nachbar, Communications Regulation 447-60 (2008).

[47] Christopher Yoo, Common Carriage’s Domain, 35 Yale J. Reg. 991 (2018).

[48] See Thomas B. Nachbar, The Public Network, 17 CommLaw Conspectus 85 (2008); Oliver E. Williamson, The Vertical Integration of Production: Market Failure Considerations, 61 Amer. Econ. Rev. 112, 112 (1971).

[49] Lina M. Khan, The Separation of Platforms and Commerce, 119 Colum. L. Rev. 973 (2019).

[50] Daniel L. Rubinfeld, Access Remedies in High-Technology Antitrust Cases, in Merger Remedies in American and European Competition Law 137, 153-56 (François Lévêque and Howard Shelanski, eds. 2003). See also Herbert Hovenkamp, Antitrust Interoperability Remedies, 123 Colum. L. Rev. F. 1, 8 (2023) (comparing the remedy of mandatory interoperability with the application of the essential facilities doctrine in the MCI case).

[51] See Rubinfeld, supra note 50  at 158-59.

[52] See Richard W. Brislin, Cross-Cultural Research Methods: Strategies, Problems, Applications, in Environment and Culture 47, 73 (Irwin Altman et al. eds. 1980) (describing the law of the hammer).