As digital platforms have become ubiquitous, introducing novel services, revolutionizing business methods and triggering enormous changes throughout the economy and society, calls for limits on (and even breaking up) the largest digital platforms have emerged. Complaints range over a broad spectrum—privacy and data security, working conditions and freedom of expression, just to name a few. Concerns sounding in antitrust have also appeared on the long list of asserted grievances: Do platforms unlawfully squelch competitors, or tilt online markets toward their own products at the expense of independent suppliers? A wide variety of antitrust investigations, cases, and legislative proposals attempt to address these complaints. Some rely on well-recognized antitrust approaches; others seek to stretch or supplement current law, or even to create entirely new regulatory regimes.

Within the legal cauldron stirred up by the pervasive impact of digital platforms, there exists a class of complaints seeking to mandate access to the resources of the platforms. Legislative and regulatory proposals in a variety of jurisdictions (including the US) would substantially expand obligations for digital platforms to provide competitors or other third parties with access to various elements of the platforms’ businesses—their networks, services, accumulated data and the algorithms and other mechanisms they employ to manage the flow of information and transactions.

Among the legal tools potentially available under U.S. antitrust law to those seeking such access is the essential facilities doctrine (EFD). First identified by name in 1977, a classic formulation of EFD holds that a competitor establishes illegal exclusionary conduct by a monopolist under Sherman Act Section 2 by showing:

  1. the monopolist controls access to an essential facility;
  2. the facility cannot be duplicated practicably by the competitor;
  3. the monopolist has denied access to the competitor, and
  4. it was feasible for the monopolist to grant such access.[1]

This chapter describes the emergence of the EFD and reviews its development against the background of major trends in the continuing evolution of federal antitrust law. It then analyzes recent Supreme Court cases that have questioned and thereby undermined EFD, even though lower courts claim to derive the doctrine from the Court’s earlier precedents. The main focus is on the emerging recognition by the Court that ongoing economic regulation of a monopoly business under the guise of antitrust is neither consistent with the fundamentals of the federal antitrust statutes, nor with basic institutional capacities of courts and antitrust enforcement agencies, as distinct from legislatures and the purpose-built agencies that engage in economic regulation pursuant to statute. The conclusion suggested by this analysis is that EFD is no more useful as a response to concerns about access to digital platforms than to the other situations that have led to the Court’s profound doubts regarding EFD. The asserted competitive problems targeted by EFD claims may change as technology and business practices evolve, but the fatal weaknesses of EFD persist. If and to the extent digital platforms can be objectively shown to demonstrate any need for mandatory access by customers, competitors or others, mandatory access remedies should be addressed—if justified by responsible and conscientious policy analysis—through rules and institutions distinct from those of antitrust.

I. Origins of EFD

EFD initially emerged in the US as a lower-court gloss on three Supreme Court antitrust decisions: United States v. Terminal Railroad Association,[2] Associated Press v. United States,[3] and Otter Tail Power Co. v. United States.[4] The first known judicial use of the phrase “essential facilities doctrine” did not occur until 1977, in Hecht v. Pro-Football, Inc.[5] The emergence of EFD is not a simple story, as one might guess from the basic fact that it took 65 years after Terminal Railroad before lower courts identified EFD as a distinct theory of Sherman Act liability. Over that same period both the antitrust statutes (notably, the enactment and strengthening of the Clayton Act) and the Court’s basic approach to antitrust-law interpretation underwent several important refinements. Moreover, several of these key advances in antitrust doctrine occurred just following the last of the three fundamental decisions (Otter Tail) in 1973. Each of the three Supreme Court cases involved both factually and legally complex disputes and all are widely separated by time, circumstances, and key characteristics. As a result, it takes a bit of scholarly gasconade to claim that they are united by any single doctrine. Even if such a unifying theme could be elicited from these early cases, it is far from clear that the central meaning of the trilogy survives the post-1973 developments outlined below.

Although the doctrine is said to be based on early Supreme Court precedents, the Supreme Court has never validated or endorsed the doctrine, despite at least two clear opportunities to do so. The first opportunity was in Aspen Skiing v. Aspen Highlands Skiing,[6] in which EFD was one of two lower-court monopolization theories that were brushed aside in a cursory footnote.[7] Subsequently EFD was the target of a skeptical dictum by Justice Breyer in a Communications Act case, AT&T Corp. v. Iowa Utilities Board.[8] These brief mentions of EFD provided a limited but nevertheless tangible indication that the Court would distance itself from the doctrine, especially given that Justice Stevens (who wrote for a unanimous Court—Justice White not participating—in Aspen) had, and Justice Breyer has, a strong background in antitrust. Moreover, both Justices were (and Justice Breyer remains) free of any reasonable suspicion of habitual favoritism to antitrust defendants, or more specifically, that they subscribe to any prodefense dogma regarding antitrust law and/or economics. Finally, in Verizon Communications v. Trinko,[9] the Court cast significant doubt on the viability of EFD, based inter alia on the considerations identified by Justice Breyer in Iowa Utilities Board. As will be explained below, the early cases usually cited as the source of EFD would probably reach very different results if assessed using present-day antitrust standards. Thus, viewing the evolution of EFD in its overall context strongly reinforces the view—long advocated by leading scholars—that EFD should not be recognized as a theory of federal antitrust liability.

A.  Terminal Railroad—and Terminal Railroad II, III and IV

Terminal Railroad arose from a comprehensive consolidation of entities that owned and operated facilities used to provide railroad transportation services in and around St. Louis, Missouri and East St. Louis, Illinois—cities that face each other from opposite banks of the Mississippi River.[10] In the late 19th and early 20th centuries—the heyday of American railroading—two dozen railroads converged on this geographic area, making it a significant hub for both passenger and cargo transportation by rail. The main relevant features of the St. Louis/East St. Louis railroad transportation complex were as follows: numerous independent railroad lines terminated on the Missouri (West) side of the river, and numerous others terminated on the Illinois (East) side.[11] Access by the railroads to the extensive track and terminal facilities in the area and to its many commercial and industrial sites (warehouses, factories, etc.) had generally been provided by independent terminal lines operating over the intricate local rail network. Finally, rail transport across the Mississippi River was available through three functionally distinct facilities: two railroad bridges and a railroad ferry.[12] (The closest alternative river crossings were more than 200 miles distant up- and down-river.) The Terminal Railroad Association, a consortium of railroads (critically, less than all of them) led by notorious Gilded-Age financier Jay Gould, managed to acquire control of all of these facilities—tracks, terminals and other assets used to provide services throughout the St. Louis/East St. Louis complex, including all three Mississippi River crossings.[13]

The conduct of the Association in disadvantaging non-member railroads by manipulating rates and imposing special charges led the Attorney General to file federal antitrust claims under Sections 1 and 2 (conspiracy to monopolize) of the Sherman Act. The complaint was initially dismissed by an equally divided four-judge Circuit Court (the court having original jurisdiction of the case, given the government’s designation of the suit under the now-repealed Expediting Act[14]). On direct appeal under the Expediting Act, the Supreme Court upheld both government claims, holding that to combine all of the terminal facilities of the area under the ownership of less-than-all of the railroads dependent on them was illegal given the imposition of discriminatory terms on non-member railroads.

The Court did not accept the government’s proposal for a dissolution of the Association; rather, considering the “public advantages of a unified system,”[15] the Court remanded for formulation of a decree that would “permit the proper and equal use [of the system] by nonproprietary companies, and abolishing the obnoxious practices in regard to transportation of merchandise.”[16] The Court itemized several requirements for an acceptable decree, and specifically warned that it would order dissolution of the Association should the parties fail to arrive at a decree that would “reorganize the contracts unifying the terminal facilities . . ..”[17] As matters evolved, however, it appears that no mutually agreeable decree was ever fully implemented. Moreover, despite three subsequent appeals following remand, the Court declined to dissolve the Association or to order any divestitures.

The Supreme Court’s remedy involved two basic requirements for reformation of the Association’s organic contractual arrangements, plus three prohibitions on specific competitive practices. The reformations required the Association (1) to offer any non-owner railroad the option to become an owner in the Association on terms equal to those of the present owners, and, (2) for railroads electing not to become owners, to allow use of the Association’s facilities on “just and reasonable” terms so as to create “as nearly an equal plane as may be” for owner and non-owner railroads.[18]

In addition to reformation of the Association’s basic arrangements for ownership and use, the Court prohibited three specific competitive practices by the Association: (1) exclusive dealing provisions governing the proprietary railroads, (2) “billing to East St. Louis, or other junction points, and then rebilling traffic destined to St. Louis or to points beyond,” and (3) imposing any “charge for the use of the terminal facilities in respect of traffic originating within the so-called 100-mile area that is not equally and in like manner applied in respect of all other traffic of a like character originating outside of that area.”[19]

The Supreme Court also included administrative and jurisdictional provisions. Any disputes regarding admission to ownership in or access to facilities of the Association would be subject to resolution by the district court, with any appeal to be treated as part of the same case (and therefore presumably subject to Supreme Court review).[20] There was also a clause preserving the exclusive jurisdiction of the Interstate Commerce Commission to regulate railroad rates and practices under the Interstate Commerce Act and related legislation.[21]

Finally, the Supreme Court specified the main elements of the divestiture that would be imposed if the parties failed to agree on decree terms meeting the Court’s other requirements. Specifically, in that event the Court would require “complete disjoinder of the three systems [i.e., the two bridges and the ferry], and their future operation as independent systems . . . .”[22]

This mandate proved easier to state than to implement. The case returned to the Supreme Court on three separate occasions over the next twelve years, with the last occurring in 1924.[23] Although the behavioral remedies prescribed by the Court were never fully implemented, no divestiture was ever ordered. A brief summary of this history follows: The second Supreme Court decision in the case resulted from some procedural quirks unique to that time, as summarized in the footnote.[24] Once the case was back on track before a lower-court panel appointed under the Expediting Act, the government and the defendants disagreed over several key issues regarding implementation of the mandate. As relevant to whether Terminal Railroad serves as a precedent for EFD, the government argued for an explicit prohibition on the Association’s ability to levy a transportation charge known as an “arbitrary” (referred to in item five of the Court’s decree requirements), imposed on traffic originating within 100 miles of the Association’s facilities but destined for delivery outside that area.[25] The final decree entered by the lower-court panel rejected this government request.[26]

This led to the third direct appeal to the Court, Terminal Railroad III.[27] The government tried to leverage the disagreement with defendants over implementation of the mandate, insisting that the dispute over the prohibition of “arbitraries” (among other issues) triggered the Supreme Courts’ dissolution option. The Court, however, dismissed this as a hyper technical reading of the mandate and thus did not consider any divestiture.[28] As to the requested prohibition on “arbitraries,” the Court showed no hesitation in concluding that the government’s proposal was the type of interference with Interstate Commerce Commission regulation that had been anticipated and prohibited in the original mandate (in item seven).[29]

The Court’s unhesitating dismissal of a proposed decree provision binding the Association to a particular course of action with regard to transportation rates (the so-called arbitraries) seems inconsistent with item five of the mandate. The government’s proposal for limitations on the Association’s rates was rejected as “plainly repugnant to the provisions of the [Interstate Commerce] Act, and contrary to the exercise by the state authorities of their power over charges of the terminal company . . . .”[30] This ruling has had its own durable influence on antitrust jurisprudence, specifically with respect to the standard for implied antitrust immunity for the conduct of regulated firms. Henceforward, the test for implied immunity has been whether application of an antitrust standard would be “plainly repugnant” to the operation of a regulatory scheme enacted by Congress—a standard that still governs.[31] Thus, while the Court has refused any recognition of EFD (as explained later), the “plain repugnancy” test for implied antitrust immunity endures.

The case reached the Supreme Court for the fourth and final time in a dispute between different groups of “proprietary” lines, with one group alleging that the other lines, “through the domination and control of the board of directors of defendant the Terminal Railroad Association of St. Louis and its subsidiaries” had been forcing the complainants to pay certain charges that the latter did not themselves pay for the same services.[32] The complaining lines charged that this was in contempt of the decree entered following remand from Terminal Railroad III.[33] The lower court appointed an examiner to review the impugned conduct, and that review essentially confirmed the allegation.[34] On the fourth direct appeal to the Court, however, the order embodying the examiner’s finding and requiring the offending railroads to make good on the missed payments was rejected.[35] The main point of interest in the Court’s reasoning is its repeated emphasis on the sanctity of the ICC’s exclusive regulatory authority: “In the exercise of its powers under existing law, the [Interstate Commerce] Commission is untrammeled by the decree, and may make and regulate rates on through freight and the divisions thereof.”[36] Again, and for the final time (so far as research discloses), the Court was clear in its determination to prevent any aspect of the decree from interfering with the ICC’s authority to regulate railroad rates and other terms of service.[37] This was to remain a matter for the ICC (and state regulators), not the courts.

B.  Associated Press

The second case cited in the EFD lineage is Associated Press v. United States.[38] Associated Press (AP), with a membership of more than 1,200 newspaper publishers at the time of the case, had been formed (beginning in 1846 with a cost-sharing arrangement among five New York City newspapers) to provide a network for the collection and dissemination of news throughout the US and worldwide.[39] Access to this network enabled member newspapers to provide news about events in other locations. (It is important to recall that prior to the advent of inexpensive long-distance data communication post-World War II, the newspaper business was a local business, given the technologies available for the production and transportation of newspapers in mass quantities. It would have been prohibitively expensive for a newspaper in one locality to provide regular coverage of news arising in many distant places.) AP prohibited members from supplying news reports to non-members, and allowed each member a de facto veto of any application for membership by any competing newspaper (i.e., a newspaper serving the same locality).[40] When the government challenged these practices under Sections 1 and 2 (conspiracy to monopolize), a panel of judges appointed pursuant to the Expediting Act granted it summary judgment in a split decision.[41] On direct appeal the Supreme Court affirmed 5-3. The various opinions in Associated Press—both the lower court and the Supreme Court — offer meandering and to some extent conflicting assessments of the issues and rationale.

An especially consequential point is that the panel granted a summary judgment for the government—finding liability without trial.[42] The competitive position of AP was a disputed issue, so the panel could not assume that AP possessed monopoly power. Moreover, given the procedural posture, the court lacked any basis to conclude that AP members possessed a competitive advantage over local rivals who did not have access to AP news stories due to the exclusivity requirement and membership-veto provision. The lower court nevertheless seemed to assume that by virtue of AP’s greater size and scope relative to other similar news services (of which there were many—the larger of which included United Press and International News Service) and the tendency of many successful newspapers to be AP members, the effect of the AP’s arrangements was to create at least some competitive advantage for its members in comparison to non-members.[43] Seizing on the First Amendment role of the newspaper industry, Learned Hand’s majority opinion reasoned that it was especially important to apply a strict Sherman Act standard to such organizations. The lower court therefore decreed that henceforward AP was forbidden to base admission decisions on whether the applicant competed with any existing member.[44] By holding that an agreement creating a competitive advantage for its participants relative to non-participants would automatically be unlawful under the Sherman Act—absent any showing of monopoly power, regardless of any business rationale, and without any demonstration that AP membership was necessary for successful competition by individual publishers or by groups of publishers forming their own news-agency arrangements—the district court strongly suggested that it was imposing a kind of public utility obligation on AP.

Judge Swan dissented, based on considerations that would be in reasonable accord with present-day antitrust analysis of competitor collaborations.[45] He pointed out that there is nothing suspect in an agreement between newspapers to exchange news reports on an exclusive basis, and that any purported effect of the AP restrictions was disproven (or at least remained subject to dispute at the summary judgment stage) as to news-gathering services by the durable and continuing prosperity of competing press syndicates United Press, International News Service, and a variety of other similar organizations, and as to newspapers by the fact that no newspaper was ever known to have failed due to denial of AP membership.[46] Judge Swan also emphasized that on the summary judgment record there could be no reliance on the notion that AP (or its members) held monopoly power or any particular advantage over competitors.[47] Finally, in Judge Swan’s view the panel majority had relied improperly on the notion that AP should be regarded as a public utility—a determination that should in any event be left to legislatures rather than courts.[48] Reviewing analogous public utility cases involving railroads, stock exchanges, cotton warehouses, and stockyards, Judge Swan found support in a dissent by Justice Brandeis in an earlier case involving AP itself, warning of the dangers of judicial encroachment into the field of public utility regulation:

Courts are ill-equipped to make the investigations which should precede a determination of the limitations which should be set upon any property right in news or of the circumstances under which news gathered by a private agency should be deemed affected with a public interest. Courts would be powerless to prescribe the detailed regulations essential to full enjoyment of the rights conferred or to introduce the machinery required for enforcement of such regulations.[49]

Thus, Judge Swan, following Justice Brandeis, relied on key distinctions between the institutional capacities of legislatures and regulatory agencies on the one hand versus courts and antitrust law on the other. As discussed further infra, these distinctions were critical to the Supreme Court’s inclination to undercut EFD when it finally considered the matter more than a half-century later.

The key points of tension in the panel split were reflected in the Supreme Court decision to affirm, by a 5-3 vote (Justice Jackson not participating).[50] For the Court, Justice Black seemed to endorse the lower-court panel majority’s approach, holding that AP must be required to admit any applicant without regard to its competitor status—essentially the “public utility” approach.[51] Justice Douglas, in a separate concurrence, denied that any form of utility regulation was involved,[52] but an additional concurrence by Justice Frankfurter seemed to state precisely the opposite, focusing on the protected First Amendment status of newspaper publishers as a particular justification for obligating AP to treat all applicants without regard to their status as competitors of other AP members.[53] Dissents by Justice Roberts (joined by Chief Justice Stone) and Justice Murphy generally followed the themes of the Swan dissent to the lower-court panel decision.[54]

C.  Otter Tail

The last of the Supreme Court cases in the EFD lineage is Otter Tail Power Co. v. United States.[55] Otter Tail marks a distinct departure from both Terminal Railroad and Associated Press. The opinions do not cite Terminal Railroad, and Justice Douglas’ majority opinion (a 4-3 vote, Justices Blackmun and Powell not participating) cites Associated Press only in support of the innocuous contention that efforts to protect or extend a monopoly may be illegal even when the defendant “had not yet achieved a complete monopoly.”[56] In the earlier cases defendants were associations controlling a resource of competitive significance (asserted but indeterminate competitive significance, in Associated Press) to their members and their members’ competitors—in one case (Terminal Railroad) rail facilities essential to competition in identifiable markets (railroad transportation in and through the St. Louis/East St. Louis complex) and in the other (Associated Press) a news exchange among newspapers published in distinct localities throughout the US and around the world. By contrast, Otter Tail involved only unilateral conduct by a single defendant, Otter Tail Power Co. (OTP), a public utility engaged in generating and transmitting wholesale electrical energy within a substantial contiguous territory in the upper Midwest.[57] OTP also operated a number of retail power distribution systems pursuant to franchises granted by numerous individual localities within OTP’s wholesale service area.[58]

The impugned conduct mainly involved OTP’s efforts to prevent four such municipalities from distributing their own local retail power—an opportunity that arose as OTP’s local franchise expired for each town. Two of these had access to power from other sources, and were faced with aggressive litigation efforts by OTP to prevent initiation of their own retail distribution service.[59] The other two were more geographically isolated within OTP’s wholesale distribution network, and had to rely either upon sales of power to them by OTP, or on OTP’s willingness to “wheel” power from other available sources over OTP’s distribution lines to a point of connection to their local systems.[60] OTP refused all such requests—either to sell power at wholesale, or to wheel power from other sources to a point of connection to the towns. Aside from OTP’s litigation tactics, the case therefore presented as a refusal to deal monopolization claim. Relying on other cases in that line, the Court affirmed (with one exception involving the Noerr-Pennington doctrine) the district court judgment finding OTP liable for monopolization, and ordering it to provide or to wheel power to municipalities subject to rates, terms, and conditions to be approved by the Federal Power Commission (the federal agency with jurisdiction over the electric power industry, including OTP, at the time).[61] As things worked out OTP eventually recaptured the local franchises in three of the four towns in dispute, while one managed to construct its own generating facility and to obtain orders from the FPC for a permanent connection to the OTP grid.

The Court’s mandate for a compelled course of dealing to be regulated by the FPC elicited a dissent from Justice Stewart, joined by Chief Justice Burger and Justice Rehnquist, pointing out that in fashioning the Federal Power Act Congress had specifically rejected proposals to authorize the FPC to require electric utilities to act as common carriers obligated to supply power on request, or to order the wheeling of electric power as the Court’s mandate required.[62] The policy concern underlying that legislative choice was that the imposition of such service and interconnection mandates on regulated utilities could jeopardize their financial soundness—a critical consideration in utility regulation in light of the substantial long-term financial obligations required to build and operate facilities that generate and distribute power, as well as the critical importance of affordable power to so many activities in our society.[63] Thus, it seemed that the majority had ordered precisely what Congress had intentionally refused to authorize—creating an apparent direct conflict with the relevant federal regulatory legislation. The dissent also emphasized the fact—obvious in the circumstances of the case—that the retail supply of electric power would be a monopoly regardless whether OTP or the affected municipality would be the supplier. It was therefore unclear what antitrust interest was ultimately at stake in the dispute.

II. EFD’s Foundational Cases as Seen Through a Post-Sylvania Lens

Taking these three cases together, it is difficult to perceive any consistent theme. Otter Tail involved single-firm conduct and two vertically related markets, each destined to be served by a regulated monopolist; viewed in terms of its ultimate competitive impact, the case at most involved a squabble regarding the identity of the monopolist that would occupy the downstream retail power distribution level (i.e., either OTP or the local municipality).[64] By contrast, central to Terminal Railroad is the key (but unanswered) question whether the numerous and diverse facilities acquired and controlled by the Association should have been reestablished as a competitive sector or whether operation of such facilities by a single consolidated entity (subject to some form of shared ownership and access regulation as contemplated by the Court) was likely to be more productive—given that there already existed a comprehensive regulatory scheme administered by a powerful agency, governing the rates, terms and conditions on which the Association’s various services were offered.[65] And Associated Press—a limited cooperative arrangement (with obvious significant procompetitive benefits) among publications serving distinct, geographically separated markets—very likely did not involve any monopoly at all.[66]

Aside from the broad diversity in the nature of the business sectors and the conduct involved in these cases, another key difficulty in relying on them for the derivation of a coherent EFD lies in the fact that all three occurred before—and therefore do not reflect—numerous important and valuable refinements of federal antitrust law. Most importantly, the Supreme Court’s approach to antitrust shifted materially beginning in the mid-1970’s. As explained briefly below, soon after Otter Tail the Court shifted away from antitrust rules based on legal formalism and toward rules informed by sound economic analysis.[67] Otter Tail came at the peak of a quarter-century-long trend expanding the use of per se rules in federal antitrust interpretation and enforcement. Per se rules deprive antitrust defendants of the ability to present any substantive arguments in defense of their conduct (lack of market power, lack of anticompetitive effect, procompetitive business justifications and effects, etc.) or any opportunity to explain their conduct in terms of economic analysis. Associated Press was arguably one of the first major antitrust decisions in this per se trend. As explained below, Terminal Railroad was decided at a time when the Supreme Court was just settling the most basic elements of antitrust interpretation, while Associated Press and Otter Tail stand at the endpoints of a relatively brief but intense later period in which antitrust interpretation became increasingly intolerant of economic reasoning. To boil down the meaning of these cases, it is helpful to review each against the background of these broader developments, and to ask how each might have been resolved applying more sophisticated and economically attuned methods of analysis that have since been championed by the Court, the antitrust enforcement agencies, and scholars over the forty-six years since the end of the per se craze.

As briefly mentioned above, only the most basic pattern for assessment of competitive practices under the Sherman Act had been set by several early Supreme Court cases. The earliest form of the per se rule as applied in United States v. Trans-Missouri Freight Ass’n,[68] and United States v. Joint Traffic Association,[69] had been limited to naked cartel conduct by Addyston Pipe & Steel Co. v. United States.[70] Although per se treatment was also extended to vertical price agreements in Dr. Miles Medical Co. v. John D. Park & Sons Co.,[71] the per se approach made little additional headway until the 1940’s. Standard Oil Co. v. United States,[72] and United States v. American Tobacco Co.,[73] established that other claims would be assessed under the rule of reason, allowing defendants to defend or justify their conduct according to the particular facts and circumstances of the case. The classic statement of the rule of reason is usually taken from Chicago Board of Trade v. United States.[74] Later cases discussed below have significantly refined the rule and explained how it is best applied to specific categories of conduct, including those at issue in Terminal Railroad and Associated Press. Notably, while the per se rule is still applied to practices found by repeated judicial assessment to be anticompetitive except in rare instances, more recent cases have used rebuttable presumptions and burden-shifting to streamline the search for sound evaluation of the ultimate competitive effects of conduct (other than conduct subject to per se condemnation).

Whatever one could say about AP, it could not by any stretch be regarded as a naked cartel, which would have qualified it for per se treatment. The arrangement was not limited to the fixing of prices or other terms of trade, to division of markets, or similar conduct eliminating competition without any plausible procompetitive contribution; rather, AP established an exclusive network for the exchange of news reports among publications serving geographically distinct areas. The arrangement allowed expanded national and global news coverage through cooperation of newspapers serving widely separated areas and therefore incapable of providing comprehensive news coverage on their own.[75] Thus, it had powerful indications of competitive merit. Moreover, the concessions made in the district court opinion included; (1) that there were many collective news services similar to the AP, (2) that several of these other services (United Press, International News Service) had significant scope and membership, (3) that many newspapers participated in several such services simultaneously, and (4) that a number of significant and successful newspapers were not members of AP.[76] Nevertheless, the district court awarded summary judgment to the government.

From our contemporary viewpoint it seems astonishing that the restrictions adopted by AP would have been condemned without trial. As explained below, in a series of decisions beginning with Broadcast Music, Inc. v. CBS, Inc.,[77] application of the per se rule to collective conduct undertaken by associations like AP has been reserved for naked cartel restraints. This illustrates the profoundly different perspective that seemed to take hold in the antitrust enforcement community and the federal courts at about the time Associated Press was decided. Premonitions of a new and more aggressive approach could be found in two major antitrust cases brought just prior to Associated Press: United States v. Aluminum Co. of America,[78] (ultimately decided by the same panel as the lower court decision in Associated Press, also in 1945) and United States v. Socony-Vacuum Oil Co.[79] Both cases represented a surge in antitrust enforcement bravado championed by the FDR Administration, largely at the urging of then AAG for Antitrust Robert Jackson (later successively appointed Solicitor General, Attorney General, and then Associate Justice of the Supreme Court by FDR) following the Supreme Court’s rejection of FDR’s government-sponsored cartelization policy embodied in the National Industrial Recovery Act in Schechter Poultry Corp. v. United States.[80] Both antitrust cases were initiated by Jackson during his brief tenure as AAG for Antitrust and pursued to conclusion by his enthusiastic successor in the Antitrust Division, Thurman Arnold, a former antitrust skeptic who displayed zeal characteristic of the recent convert when he became the Administration’s chief antitrust prosecutor in 1938.

Alcoa established a presumption that a monopolist is guilty of monopolization, shifting the burden of proof to the defendant to show that its monopoly power had been “thrust upon” it.[81] Alcoa paid lip service to the notion that “the successful competitor, having been urged to compete, must not be turned upon when he wins.”[82] Contrary to the spirit of that famous epithet, however, Alcoa imposed liability on the defendant—indeed, mandated that the firm be broken up, as in the landmark 1911 Standard Oil case—based inter alia on Alcoa’s habit of expanding capacity to meet demand. Similarly, Socony-Vacuum, while outwardly a conventional application of the per se rule to a straightforward example of horizontal output allocation and price fixing, provided a very expansive statement of the price-fixing rule. A famous footnote in the Court’s opinion (footnote 59) suggested that any agreement that might “tamper” with prices must be condemned out of hand even absent monopoly power or the capacity to influence market prices.[83] In so stating, the Court suggested that a joint venture must avoid any arrangement that could affect prices, regardless of market position, competitive effect, or procompetitive business rationale. As matters evolved, the Court actually went farther than that—holding first in United States v. Sealy, Inc.,[84] that any price restriction adopted by a joint venture was per se illegal, and then in United States v. Topco Assocs., Inc.,[85] that even non-price (territorial) restrictions adopted by a joint venture are per se illegal.

Beginning about the time of Associated Press, various extensions of the per se rule followed regularly for the next quarter century, resulting in an almost-total displacement of the rule of reason from antitrust analysis. In 1947 patent tie-ins were moved to the per se category,[86] and a long list of other patent licensing practices followed during the next quarter-century, encouraged by the federal enforcement agencies. Although the Court once hesitated to subject non-price vertical restraints to the per se rule,[87] it soon lost its inhibitions in that regard, condemning all vertical restraints per se based on the early vertical price-fixing precedent Dr. Miles Medical Co. v. John D. Park & Sons Co.,[88] and the ancient rule prohibiting restraints on alienation of chattels. Defendants were thus precluded from appealing to economic reasoning or relying on any substantive defense with regard to vertical restraints.[89]

In the same vein, a rigid structural presumption against horizontal mergers was adopted in United States v. Philadelphia National Bank,[90] and applied even to mergers involving entities of no competitive significance in United States v. Von’s Grocery Co.,[91] and later casesestablishing a virtual per se rule against mergers. Alcoa became the rule where monopolization was concerned, placing the burden of proof upon monopolists to show that their success had been “thrust upon”[92] them, even if their conduct had been “honestly industrial”[93] and not otherwise wrongful. Arguably the high-water mark of the per se trend was reached in United States v. Topco Assocs., Inc.,[94] a then-recent case cited by the majority in Otter Tail.[95] In Topco the Court not only disregarded economic justifications for the particular arrangement (a joint venture among independent grocery stores to establish their own private-label brand to respond competitively to those offered by the large national chain stores such as A&P), but the opinion openly mocked the idea of relying on economic analysis in assessing the legality of even non-price restrictions (territorial restrictions) in a joint venture agreement.[96]

Fortunately, the Court began to turn away from this brutalist approach to competition analysis almost immediately after Otter Tail. The intensifying reliance on per se rules and heavy presumptions of liability from about 1945 until the 1972 decision in Topco attracted broad criticism from respected scholars of antitrust law and economics. Worsening economic conditions in the U.S. during the early 1970’s—a period of high inflation, low growth, and loss of U.S. competitiveness in key economic sectors such as automobiles and consumer electronic products—created an environment in which a broad range of U.S. economic and regulatory policies (including antitrust) were subject to focused reexamination. In United States v. General Dynamics Corp.,[97] decided in the term following Otter Tail, the Court began to display sensitivity to the need for antitrust rules to reflect sound economic analysis. The Court accepted economic arguments to overcome the rigid structural presumptions established by its merger cases from the 1960’s. The government had challenged a merger in the coal industry, relying on recent production data in calculating market shares for purposes of measuring concentration and claiming the presumption of illegality.[98] The Court noted, however, that in the particular sector involved—supply of coal to electricity generation facilities—available uncommitted reserves rather than recent production were far more meaningful in weighing the competitive significance of particular firms.[99] (Since coal purchasers were electric-power utilities requiring very long-term commitments from their fuel suppliers, a coal supplier lacking uncommitted reserves was incapable of competing, regardless of past and present rates of output.) Then the situation was clarified and a powerful new approach established with Continental T.V., Inc. v. GTE Sylvania, Inc.,[100] a case involving run-of-mill vertical territorial restrictions on the distribution of consumer electronic products. Sylvania overruled Schwinn, removing non-price vertical restraints from the per se category.[101] Eventually, virtually all vertical restraints (including price limitations) were restored to rule-of-reason status.[102] More fundamentally, the case turned Topco around 180 degrees on the question of using economic analysis to assess competitive effects. Whereas Topco had exalted the need for legal certainty and mocked the use of economics in antitrust analysis, Sylvania observed that “an antitrust policy divorced from market considerations would lack any objective benchmarks.”[103] Sylvania decisively rejected strict legal formalism as a basis for assessment of competitive effect.

Otter Tail missed the transition that ended this antitrust prosecutor’s/plaintiffs’ utopia by a whisker: It may be the last Supreme Court case to base antitrust liability on a monopolist’s use of sharp business tactics without any persuasive demonstration that such conduct would or could result in a material anticompetitive effect. The author of the four-Justice majority in Otter Tail, Justice Douglas, was a committed New Deal Democrat, well-connected at the highest levels of FDR’s Administration. He had been heavily promoted by FDR “brain trust” member Justice Louis Brandeis as his successor on the Court when Brandeis retired in 1939. As a Justice, Douglas was generally an enthusiastic supporter of the per se trend. He authored the majority opinion in Socony-Vacuum. The same year he decided Otter Tail, he wrote (in a concurrence) one of the most expansive interpretations of antitrust policy and the nature of merger control in the history of antitrust, and aligned himself with the Brandeis view that the antitrust laws should be construed to protect local control of business even if that approach requires economic sacrifice.[104] However, in a notable albeit limited departure he authored the Court’s White Motor opinion,[105] declining an opportunity to place all non-price vertical restraints into the per se category—a step the Court took, with Douglas’ support, a few years later in Schwinn.[106]

It is therefore unsurprising that Justice Douglas’ majority opinion in Otter Tail shows little regard for careful analysis of tangible competitive effects, based on balanced evaluation of diverse economic understandings offered by the parties. It had been characteristic for some time for the relevant markets involved in Otter Tail—wholesale and retail distribution of electricity—to be subject to public utility-style regulation at both the federal (wholesale) and state (retail) level. Under no foreseeable conditions could competition occur between OTP and the local municipal utilities involved. Power was provided locally by a single franchisee in each case—only the identity of the monopolist was at stake.

Following identification of EFD in Hecht, private litigants showed exceptional creativity in finding “essential facilities” to which they could claim access through litigation. Numerous cases considered and sometimes applied the EFD in a broad variety of circumstances following Hecht. Despite a great deal of lower-court activity involving EFD, specific claims involving the doctrine managed to avoid Supreme Court review until an unusual matter arose due to a competitive stand-off between the two leading ski resorts in Aspen, Colorado. There were four popular skiing mountains in the Aspen area, three controlled by Aspen Skiing Co. and one by Aspen Highlands Skiing Co.[107] For some period of years, the two companies cooperated in providing an all-Aspen pass, allowing skiers the convenience of obtaining access to the full range of ski slopes in Aspen through a single transaction.[108] Revenues from sales of the all-Aspen pass were allocated based on random-sample surveys of the number of skiers using each mountain.[109] For the 1977-1978 ski season, however, Aspen Skiing conditioned its participation in the all-Aspen pass on Aspen Highlands’ agreement to accept a fixed percentage of revenues—a percentage falling below Aspen Highlands’ average historical share of revenue.[110] The following season Aspen Highlands refused to accept a fixed and below-average percentage revenue allocation, and Aspen Skiing discontinued sale and acceptance of the all-Aspen pass.[111] Aspen Skiing shifted to a three-mountain pass honored at its own facilities only.[112]

Aspen Skiing went farther than mere termination of the previous joint all-Aspen pass arrangement—notably, it refused to sell its three-mountain pass to Aspen Highlands even when the latter offered to pay full retail price in cash. This prevented Aspen Highlands from offering its own customers the opportunity to obtain an all-Aspen experience without the inconvenience of entering into separate transactions with each company. Aspen Highlands challenged Aspen Skiing’s conduct as unlawful monopolization and obtained a favorable verdict.[113] The jury had been instructed on a number of monopolization theories, including EFD. The 10th Circuit affirmed.[114] On cert. Justice Stevens, writing for a unanimous Court (Justice White not participating), upheld the verdict.[115]

An initial problem in understanding Aspen is the petitioner’s tactical concession on appeal that it possessed monopoly power in the relevant market. Given the numerous other ski facilities in Colorado, other Rocky Mountain states, and other destinations around North America, this provides a somewhat unrealistic framework for analysis of the competitive situation. One must ignore serious doubts about the idea that any single local ski area could be a monopoly in order to look at whether Aspen Skiing was in possession of an “essential facility,” or whether its conduct was unreasonably exclusionary.

A classic antitrust analysis of a monopolist’s refusal to trade with customers that patronize a competitor is provided by Lorain Journal Co. v. United States,[116] involving a dominant local daily newspaper’s refusal to sell advertising to customers if they also placed advertising with a local media upstart, a fledgling radio station. There is no plausible economic or technical rationale for such a refusal to deal, other than squelching the obvious competitive risk to the incumbent’s business. Since that is an anticompetitive rationale, the Court had little trouble finding liability under Section 2. The key conduct in Aspen—especially Aspen Skiing’s refusal to sell its own tickets to Aspen Highlands at their full retail price in exchange for cash—reads directly on Lorain Journal. Given these questions about the vitality of the decision, the Court subsequently described Aspen—to the extent it is construed as requiring a monopolist to engage in a course of dealings with a competitor—as lying “at or near the outer boundary of § 2 liability.”[117] But the Court in Aspen engaged in no extended analysis of EFD. Instead it lightly dismissed the 10th Circuit’s effort to weave its decision from the doctrinal material provided by other lower courts, including EFD (as well as the other then-prevailing lower-court method of analyzing a unilateral refusal to deal, the so-called “intent test”). In Aspen’s concluding footnote, the Court simply observed:

Given our conclusion that the evidence amply supports the verdict under the instructions as given by the trial court, we find it unnecessary to consider the possible relevance of the “essential facilities” doctrine . . . .[118]

It required another fourteen years for the Court to further reflect on EFD, and in the event it was through a short but trenchant dictum on a related regulatory issue. That dictum has been especially influential because it came from Justice Breyer, who has had unique influence on the application of antitrust law to regulated sectors. Justice Breyer spent early career time as a senior staff member in the Antitrust Division and then as counsel to the Senate Judiciary Committee, which has legislative oversight of antitrust enforcement. He also became a well-regarded scholar on the related subjects of economic regulation and administrative law. His tenure as Senate Judiciary Committee counsel coincided with significant legislative efforts to deregulate several major US industry sectors. He was instrumental in the specific effort to liberalize economic regulation of commercial aviation by the Civil Aeronautics Board. The CAB had been heavily criticized for blockading entry and eliminating price competition in the aviation sector for decades.

Justice Breyer joined the U.S. Court of Appeals for the First Circuit in 1980, and began to have direct influence on how the borderline between regulation and antitrust is defined for purposes of antitrust litigation. In Town of Concord v. Boston Edison,[119] a ruling that foreshadowed a subsequent line of Supreme Court cases addressing the issue, then-Judge Breyer was faced with a claim that a “price-squeeze” by an electric power utility constituted unlawful monopolization. The basic claim was that Boston Edison, a power supplier operating at both wholesale and retail levels (like OTP in Otter Tail), kept its wholesale rates high and its retail rates low, making it uneconomical for municipalities within its service area to operate at the retail level. Just as in Otter Tail, wholesale rates were controlled by federal regulation, while retail rates were controlled by state utility regulation.[120] Emphasizing the complexity and likely futility of applying a layer of antitrust intervention to pricing decisions already subject to comprehensive administrative agency supervision at both state and federal levels, Breyer wrote for a panel that refused to recognize such a claim, reversing a lower-court verdict in plaintiff’s favor.[121] Although the opinion defies quick summarization, Breyer’s extensive and careful reasoning traversed questions involving the likely effects (both pro- and anticompetitive) of the conduct in question, as well as the dangers of leaving antitrust courts free to ramble through the labyrinth of public utility regulation, as would have been required to administer any remedy in a case of this character.

This brings us to Justice Breyer’s first explicit contribution to the EFD debate from his later position on the Supreme Court. Fourteen years after Aspen Skiing deftly sidestepped EFD in a footnote, AT&T Corp. v. Iowa Utilities Bd.,[122] reached the Court. The case involved appellate review of rules adopted by the Federal Communications Commission (FCC) to implement the Telecommunications Act of 1996.[123] This legislation sought to resolve a number of significant legal and regulatory questions plaguing the communications industry for more than a decade following dissolution of the former Bell System, which had held a monopoly (subject to FCC and state regulation) on telecommunications service in much of the U.S. for most of the 20th Century. The challenged regulations had in part implemented a statutory mandate requiring the FCC to compel regulated local monopoly telephone companies (the “Baby Bells” spun off from the former Bell System) to provide various forms of network access to would-be competitors. Among numerous other claims, petitioners challenging the FCC rules asked the Court to treat these mandatory access provisions of the statute as analogous to EFD.

In construing the FCC’s statutory mandate for facility sharing, Breyer began by mentioning parenthetically that EFD is “an antitrust doctrine that this Court has never adopted.”[124] Having thus distanced the Court from EFD, Breyer questioned whether the statute could have intended that the FCC compel network sharing even where “a new entrant could compete effectively without the facility, or where practical alternatives to that facility are available.”[125] Justice Breyer also cited an article critical of EFD by antitrust legend (and Harvard Law School faculty colleague from Breyer’s time there) Prof. Phillip Areeda,[126] then proceeded to itemize a number of risks posed by any doctrine (including the relevant provision of the Telecommunications Act) suggesting the imposition of mandatory access requirements on firms with competitively valuable resources:

Even the simplest kind of compelled sharing, say, requiring a railroad to share bridges, tunnels, or track, means that someone must oversee the terms and conditions of that sharing. Moreover, a sharing requirement may diminish the original owner’s incentive to keep up or to improve the property by depriving the owner of the fruits of value-creating investment, research, or labor. . . . Nor can one guarantee that firms will undertake the investment necessary to produce complex technological innovations knowing that any competitive advantage deriving from those innovations will be dissipated by the sharing requirement. The more complex the facilities, the more central their relation to the firm’s managerial responsibilities, the more extensive the sharing demanded, the more likely these costs will become serious. And the more serious they become, the more likely they will offset any economic or competitive gain that a sharing requirement might otherwise provide. The greater the administrative burden, for example, the more the need for complex proceedings, the very existence of which means delay, which in turn can impede the entry into long-distance markets that the Act foresees.

Nor are any added costs imposed by more extensive unbundling requirements necessarily offset by the added potential for competition. Increased sharing by itself does not automatically mean increased competition. It is in the unshared, not in the shared, portions of the enterprise that meaningful competition would likely emerge. Rules that force firms to share every resource or element of a business would create not competition, but pervasive regulation, for the regulators, not the marketplace, would set the relevant terms.[127]

Although Iowa Utilities Board did not directly involve any question of antitrust law, Justice Breyer’s contribution identifies generic competitive and institutional issues that affect the wisdom of imposing mandatory access obligations on private resources of competitive value. These issues would be relevant to antitrust as well as to any other field of law in which they arise. Aside from the parties’ explicit invocation of EFD, Breyer’s opening reference to the sharing by a railroad of “bridges, tunnels, or track”—an unmistakable reference to Terminal Railroad—makes a very direct connection to antitrust.[128] As to the specific concerns identified by Justice Breyer, first, there is the cost, complexity and other burdens of regulating the prices, terms, and conditions of access, which neither courts nor antitrust enforcement agencies are equipped to consider. Second, mandatory sharing poses a threat to innovation and investment in resources that provide competitive value to the market and to the regulated entity. Finally, there is the question of whether mandatory sharing has any potential to increase competition: since regulation sets the terms of the mandatory access, the role allowed for competition between the parties is reduced to that extent, and for a fully-shared resource there is essentially no remaining competition as to the affected activity. Given these substantial risks to competition from mandatory sharing as a regulatory approach, Justice Breyer was led to conclude that “it requires a convincing explanation of why facilities should be shared . . . where a new entrant could compete effectively without the facility, or where practical alternatives to that facility are available.”[129]

Five years later the Court made its most direct and recent assessment of the EFD, relying in substantial part on the same policy elements identified by Justice Breyer in Iowa Utilities Board, as well as his key insights on the analysis of monopolization claims in the regulated-firm context in Town of Concord. In Trinko a customer of long-distance carrier AT&T, complained that local exchange carrier Verizon (also a long-distance competitor) had unlawfully discriminated against AT&T in providing elements of local exchange access and thereby deprived plaintiff of the benefit of competition in the long-distance market.[130] But the legislation and rules adopted pursuant to the Telecommunications Act of 1996 supplied a well-elaborated regulatory structure for dealing with such complaints (involving both the FCC and state telecommunications regulators), and indeed remedies had been sought and imposed in the specific instance at issue in Trinko. Thus, the question in Trinko was whether an independent antitrust claim could proceed against Verizon for the precise conduct that had been subject to regulatory scrutiny and relief.

Justice Scalia, writing for a six-Justice majority (three Justices concurring in the judgment solely upon the distinct rationale that plaintiff Trinko lacked standing), followed a mode of analysis very similar to that formulated by then-Judge Breyer in Town of Concord: Although the Telecommunications Act was clear that no antitrust exemption was available (the Act provided no explicit immunity and it contained an antitrust savings clause, precluding implied immunity as well), nevertheless the competitive and institutional circumstances compelled the conclusion that Trinko’s claim could not be recognized. Justice Scalia first itemized the risks inherent in compelled access remedies involving competitively valuable private resources:

Compelling such firms to share the source of their advantage is in some tension with the underlying purpose of antitrust law, since it may lessen the incentive for the monopolist, the rival, or both to invest in those economically beneficial facilities. Enforced sharing also requires antitrust courts to act as central planners, identifying the proper price, quantity, and other terms of dealing—a role for which they are ill-suited. Moreover, compelling negotiation between competitors may facilitate the supreme evil of antitrust: collusion. Thus, as a general matter, the Sherman Act “does not restrict the long recognized right of [a] trader or manufacturer engaged in an entirely private business, freely to exercise his own independent discretion as to parties with whom he will deal.”[131]

This list is very similar to that provided by Justice Breyer in his analysis of the compulsory network sharing regulations in Iowa Utilities Board. The first two items—the disincentives to investment created by enforced sharing and the lack of regulatory capacity in the “antitrust courts”—are more or less cloned from Breyer’s list.[132] Finally, whereas Breyer had noted that competition will be suppressed to the extent there is forced sharing, the Scalia list explicitly refers to the risks of collusion among joint venturers—a key element of antitrust analysis of both contractual relationships between competitors and structural transactions establishing enterprises owned in common by otherwise-independent firms.[133]

III. Conclusion: EFD is Access Regulation Masquerading as Antitrust, and Should be Considered Skeptically, If at All , and Only as a Sector-Specific Regulatory Alternative

Despite the Supreme Court’s tendency to distance itself from and identify policy objections to EFD, the Court has not definitively rejected the doctrine in so many words. (In fact, it explicitly declined to repudiate EFD even in Trinko.) This has invited commentators to suggest a continuing role for the doctrine. In the 1989 article cited by Justice Breyer in Iowa Utilities Board, written a few years after Aspen, Professor Areeda assessed EFD and concluded that given the dangers of broad application and the absence of “limiting principles,” EFD should be applied only in very narrow circumstances.[134] Although the respected Areeda & Hovenkamp treatise endorsed outright abandonment of EFD,[135] some scholars continue to support the legitimacy of EFD and to speculate that it might fulfill some useful role.[136]

Even if limited as Prof. Areeda had proposed in his 1989 article, the EFD deserves no support, and lingering speculation about remaining vitality of EFD in scholarship and lower-court decisions has become a serious liability—first, because it still inspires plaintiffs to continue filing claims based on the doctrine, and second, because it is likely encouraging foreign antitrust regimes to engage in various applications of the doctrine that do not recognize and adequately account for the substantive and institutional weaknesses of EFD.

One voice in favor of EFD is the influential Chief Judge of the Seventh Circuit, Diane P. Wood, who, like Justice Breyer, had considerable academic and governmental antitrust experience prior to becoming a federal judge. In a recent article about EFD she states:

We can see that recognition of an anticompetitive practice sometimes has preceded the formal, rigorous economic explanation of why exactly that practice is capable of harming competition and consumer welfare. My suggestion is that the essential facilities doctrine or idea, may belong in that category.[137]

This suggestion is remarkable for at least two reasons: first, EFD has been under active debate—judicial decisions, extensive legal and economic scholarship, etc.—for more than forty years. Quite a bit of effort has been expended trying to find a “rigorous economic explanation” for why one might adopt EFD. One of the fundamental understandings to emerge from the discussion surrounding EFD is that the prerequisites for application of EFD (by its own terms, as defined by the lower-court cases and by those advocating recognition of EFD at least in some limited form) are met only when the typical conditions calling for possible implementation of economic regulation are present.[138] If a market can be served effectively only by unified control of an indivisible “facility” (accepting the possibility that an intangible asset or a complex collection of assets such as a digital platform might constitute such a facility), the adverse economic effects of that monopoly control cannot be limited by the techniques of antitrust law, which prohibit only conduct that restricts competition. If compelled access is required (as distinct from the prohibition of specific identifiable exclusionary practices by the monopolist, or a divestiture restoring or creating competition among independent firms controlling distinct parts of the “facility”), then necessarily the available public policy options are found within the realm of economic regulation, not competition law. Of course, economic regulation—like antitrust and other policy options—brings its own limitations.[139] Many are severe and often fatal—industry capture, “mission creep”, political cronyism, lack of accountability, etc. Thus, judgments regarding the wisdom of invoking regulatory solutions to essential facilities issues will depend upon a complex and unique set of considerations for each such “facility.”

In principle, however, the identification of an essential facility impervious to effective antitrust remedy requires that the key judgments—which business activities and entities to control, which business variables to control (price and other transaction terms, quality, output, capacity, etc.)—be carried out by governmental institutions most suitable for that purpose. These are primarily legislative judgments suitable for implementation (if at all) through specialized (usually administrative) forms of regulation. There is little basis to expect that antitrust courts (or enforcement agencies) have any particular competence either in deciding when the instruments of economic regulation should be applied, or in administrating such regulation over the extended periods required to make any solution effective.

The second point of interest in Chief Judge Wood’s comment about EFD is the notion “that recognition of an anticompetitive practice” should ever “precede[ ] the formal, rigorous economic explanation of why exactly that practice is capable of harming competition.”[140] There has been extensive commentary on and heavy criticism of the tendency of antitrust agencies and courts to treat novel forms of business conduct with suspicion, and tilt toward condemnation in the absence of any persuasive analysis tending to confirm the pro- or anticompetitive characteristics of that conduct. Thus, in this author’s view, Chief Judge Wood has reversed the appropriate order of concern about unexplained competitive conduct. Given the tremendous power of antitrust remedies, as well as the enormous cost and delay inherent in the application of the “great machinery” of antitrust enforcement (per Judge Posner), any judicial suspicion of practices not demonstrated and confirmed to pose genuine competitive risk represents a tangible threat to welfare. Better to advocate the opposite—that no practice should be condemned without persuasive reason, including empirically based economic analysis. I would therefore hesitate long and hard before suggesting that EFD should be preserved in antitrust doctrine because there is suspicion that it might someday be shown to be economically sensible. Given the demonstrable institutional weaknesses of EFD as an antitrust concept—even conceding that there are some industries whose performance may be improved by subjecting them to economic regulation—proper jurisprudential caution suggests continued resistance to any adoption of EFD by the federal courts and antitrust agencies.


[1] MCI Commc’ns Corp. v. American Tel. and Tel. Co., 708 F.2d 1081, 1132-33 (7th Cir. 1983), cert. denied, 464 U.S. 891 (1983).

[2] 224 U.S. 383 (1912).

[3] 326 U.S. 1 (1945).

[4] 410 U.S. 366 (1973).

[5] 570 F.2d 982 (D.C. Cir. 1977), cert. denied, 436 U.S. 956 (1978).

[6] 472 U.S. 585 (1985).

[7] See id. at 600 n.26; see id. at 611 n.44.

[8] 525 U.S. 366, 388, 428 (1999).

[9] 540 U.S. 398, 410-11 (2004).

[10] United States v. Terminal R.R. Ass’n, 224 U.S. 383, 391-92 (1912) (Terminal Railroad).

[11] See id. at 392-93.

[12] See id.

[13] See id. at 391-92.

[14] Prior to its repeal in 1984, the Expediting Act, 15 USCA 28, a procedural antitrust innovation successfully promoted by “trust-busting” President Theodore Roosevelt, allowed the Attorney General to designate any antitrust case brought by the US for hearing by a panel consisting of a minimum of three judges (chosen pursuant to a specified protocol from among judges of the circuit court and the district court having jurisdiction of the matter) rather than before a single district court judge, and to appeal the final judgment directly to the Supreme Court, thus eliminating one step in the typical federal appellate progression.

[15] Terminal Railroad, 224 U.S. at 390.

[16] Id.

[17] Id.

[18] The full text of the remedial elements prescribed by the Supreme Court regarding ownership and use is as follows:

[First, b]y providing for the admission of any existing or future railroad to joint ownership and control of the combined terminal properties, upon such just and reasonable terms as shall place such applying company upon a plane of equality in respect of benefits and burdens with the present proprietary companies.

Second. Such plan of reorganization must also provide definitely for the use of the terminal facilities by any other railroad not electing to become a joint owner, upon such just and reasonable terms and regulations as will, in respect of use, character, and cost of service, place every such company upon as nearly an equal plane as may be with respect to expenses and charges as that occupied by the proprietary companies.

Id. at 411.

[19] The full text is as follows:

Third. By eliminating from the present agreement between the terminal company and the proprietary companies any provision which restricts any such company to the use of the facilities of the terminal company.

Fourth. By providing for the complete abolition of the existing practice of billing to East St. Louis, or other junction points, and then rebilling traffic destined to St. Louis, or to points beyond.

Fifth. By providing for the abolition of any special or so-called arbitrary charge for the use of the terminal facilities in respect of traffic originating within the so-called 100-mile area that is not equally and in like manner applied in respect of all other traffic of a like character originating outside of that area.

Id. at 411-12.

[20] The full text is as follows:

Sixth. By providing that any disagreement between any company applying to become a joint owner or user, as herein provided for, and the terminal or proprietary companies, which shall arise after a final decree in this cause, may be submitted to the district court, upon a petition filed in this cause, subject to review by appeal in the usual manner.

Seventh. To avoid any possible misapprehension, the decree should also contain a provision that nothing therein shall be taken to affect in any wise or at any time the power of the Interstate Commerce Commission over the rates to be charged by the terminal company, or the mode of billing traffic passing over its lines, or the establishing of joint through rates or routes over its lines, or any other power conferred by law upon such Commission.

Id. at 412.

[21] Id.

[22] The full text is as follows:

Upon failure of the parties to come to an agreement which is in substantial accord with this opinion and decree, the court will, after hearing the parties upon a plan for the dissolution of the combination between the terminal company, the Wiggins Ferry Company, the Merchants’ Bridge Company, and the several terminal companies related to the Ferry and Merchants’ Bridge Company, make such order and decree for the complete disjoinder of the three systems, and their future operation as independent systems as may be necessary, enjoining the defendants, singly and collectively, from any exercise of control or dominion over either of the said terminal systems or their related constituent companies through lease, purchase, or stock control, and enjoining the defendants from voting any share in any of said companies or receiving dividends, directly or indirectly, or from any future combination of the said systems in evasion of such decree or any part thereof.

Id. at 412-13.

[23] Ex parte United States, 226 U.S. 420 (1913) (Terminal Railroad II); United States v. Terminal R. Ass’n, 236 U.S. 194 (1915) (Terminal Railroad III); Terminal R. Ass’n v. United States, 266 U.S. 17 (1924) (Terminal Railroad IV).

[24] It happened that the federal Circuit Courts had been abolished while Terminal Railroad was first pending in the Supreme Court. (The Circuit Courts were succeeded by the appellate bodies now known as the U.S. Courts of Appeals for the various geographically-designated circuits, currently numbering twelve.) The case was remanded to the Eastern District of Missouri where it had first been filed, and assigned to a district court judge. But the government insisted that since the case had been designated as a proceeding subject to the Expediting Act (not directly affected by the legislation abolishing the Circuit Courts), it must continue as such, requiring appointment of a (minimum) three-judge panel for the proceedings on remand. The district court did not adopt the government’s view on this issue, but left the government to petition the Supreme Court for a writ of prohibition, requiring appointment of a panel pursuant to the Expediting Act. This writ was granted in Terminal Railroad II, 226 U.S. at 420.

[25] See Terminal Railroad I, 224 U.S. at 411-12.

[26] See Terminal Railroad II, 226 U.S. at 425.

[27] Terminal Railroad III, 236 U.S. at 194.

[28] See id. at 207-09.

[29] Id.

[30] Id. at 207.

[31] Credit Suisse Securities v. Billing, 551 U.S. 264 (2007). For a broader discussion of both Credit Suisse and the plain repugnancy standard, see Bruce H. Kobayashi & Joshua D. Wright, Antitrust and Ex-Ante Sector Regulation, in The GAI Report on the Digital Economy (2020).

[32] Terminal Railroad IV, 266 U.S. at 28.

[33] Id. at 26.

[34] Id.

[35] See id. at 26-27.

[36] Id. at 31.

[37] See id. at 30-31

[38] 326 U.S. 1 (1945).

[39] See id. at 3-4.

[40] See id. at 4.

[41] The trial-court decision was rendered by a panel consisting of three judges drawn from the Second Circuit Court of Appeals, including Judge Learned Hand, his cousin Augustus Hand, and Thomas Swan. This is the same panel that rendered the historic Section 2 decision in United States v. Aluminum Co. of America, 148 F.2d 416 (2d Cir. 1945) (Alcoa), which served as the final decision due to lack of a quorum in the Supreme Court in that matter. Judge Learned Hand wrote both Associated Press and Alcoa, with Swan dissenting in Associated Press.

[42] United States v. Associated Press, 52 F. Supp. 362, 357 (S.D.N.Y. 1943).

[43] See id. at 337.

[44] Id. at 375.

[45] See id. at 375-77 (Swan, J., dissenting).

[46] Id. at 375.

[47] Id. at 377.

[48] Id.

[49] Int’l News Serv. v. Associated Press, 248 U.S. 215, 267 (1918) (Brandeis, J., dissenting).

[50] Associated Press v. United States, 326 U.S. 1 (1945).

[51] Id. at 19.

[52] Id. at 24-25.

[53] Id. at 28.

[54] See id. at 30 (Roberts, J. dissenting); See id. at 50 (Murphy, J. dissenting).

[55] 410 U.S. 366 (1973).

[56] Id. at 378 (citing Associated Press v. United States, 326 U.S. 1,13 (1945)).

[57] See id. at 368.

[58] See id. at 368-69.

[59] Id.

[60] Id. at 369.

[61] See id. at 381-82.

[62] See id. at 383-85 (Stewart, J., dissenting).

[63] See id. at 392.

[64] Otter Tail Power Co. v. United States, 410 U.S. 366 (1973).

[65] United States v. Terminal R.R. Ass’n, 224 U.S. 383 (1912).

[66] Associated Press v. United States, 326 U.S. 1 (1945).

[67] See, e.g., United States v. General Dynamics Corp., 415 U.S. 486 (1974); Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977).

[68] 166 U.S. 290 (1897).

[69] 171 U.S. 505 (1898).

[70] 175 U.S. 211 (1899).

[71] 220 U.S. 373 (1911).

[72] 221 U.S. 1 (1911).

[73] 221 U.S. 106 (1911)

[74] 246 U.S. 231 (1918).

[75] To appreciate this, it is essential to picture the information sector as it existed in the first half of the 20th Century, when there was no widespread access to means of long-distance communication other than telephone and telegraph. Such access as existed was relatively expensive and/or required costly resources such as specialized equipment and trained personnel.

[76] United States v. Associated Press, 52 F. Supp. 362 (S.D.N.Y. 1943).

[77] 441 U.S. 1 (1979).

[78] 148 F.2d 416 (2d Cir. 1945).

[79] 310 U.S. 150 (1940).

[80] 295 U.S. 495 (1935).

[81] United States v. Aluminum Co. of America, 148 F.2d 416 (2d Cir. 1945).

[82] Id. at 430.

[83] United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 224 n.59 (1940).

[84] 388 U.S. 350 (1967).

[85] 405 U.S. 596 (1972).

[86] International Salt Co., Inc. v. United States, 332 U.S. 392 (1947).

[87] White Motor Co. v. United States, 372 U.S. 253 (1963).

[88] 220 U.S. 373 (1911).

[89] United States v. Arnold, Schwinn & Co., 388 U.S. 365 (1967).

[90] 374 U.S. 321 (1963).

[91] 384 U.S. 270 (1966).

[92] Id. at 429.

[93] Id. at 431.

[94] 405 U.S. 596 (1972).

[95] Otter Tail Power Co. v. United States, 410 U.S. 366, 378 (1973).

[96] See United States v. Topco Assocs., 405 U.S. 596, 609-10 (1972).

[97] 415 U.S. 486 (1974).

[98] Id. at 493-96.

[99] Id. at 502.

[100] 433 U.S. 36 (1977).

[101] Cont’l T.V. v. GTE Sylvania, 433 U.S. 36, 59 (1977) (overruling United States v. Arnold, Schwinn & Co., 388 U.S. 365 (1967)).

[102] See generally, State Oil Co. v. Khan, 522 U.S. 3 (1997)(maximum vertical price agreements); Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877 (2007)(minimum vertical price agreements).

[103] Sylvania, 433 U.S. at 53 n.21.

[104] Excerpt is as follows:

By reason of the antitrust laws, efficiency in terms of the accounting of dollar costs and profits is not the measure of the public interest, nor is growth in size where no substantial competition is curtailed. The antitrust laws look with suspicion on the acquisition of local business units by out-of-state companies. For then local employment is apt to suffer, local payrolls are likely to drop off, and responsible entrepreneurs in counties and States are replaced by clerks.

A case in point is Goldendale in my State of Washington. It was a thriving community — an ideal place to raise a family — until the company that owned the sawmill was bought by an out-of-state giant. In a year or so, auditors in faraway New York City, who never knew the glories of Goldendale, decided to close the local mill and truck all the logs to Yakima. Goldendale became greatly crippled. It is Exhibit A to the Brandeis concern, which became part of the Clayton Act concern, with the effects that the impact of monopoly often has on a community, as contrasted with the beneficent effect of competition.

A nation of clerks is anathema to the American antitrust dream. So is the spawning of federal regulatory agencies to police the mounting economic power. For the path of those who want the concentration of power to develop unhindered leads predictably to socialism that is antagonistic to our system.

United States v. Falstaff Brewing Corp., 410 U.S. 526, 543 (1973).

[105] White Motor Co. v. United States, 372 U.S. 253 (1963).

[106] Schwinn, 388 U.S. at 365.

[107] Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 587 (1985).

[108] Id. at 589-90.

[109] Id. at 592.

[110] See id.

[111] See id. at 594.

[112] Id.

[113] Id. at 596.

[114] Id. at 599.

[115] Id. at 611.

[116] 342 U.S. 143 (1951).

[117] Verizon Commc’ns., Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 409 (2004).

[118] Aspen Skiing, 472 U.S. at 611 n.44.

[119] 915 F.2d 17 (1st Cir. 1990).

[120] Id. at 20.

[121] Id. at 31.

[122] 525 U.S. 366 (1999).

[123] Id. at 371-73.

[124] Id. at 428 (Breyer, J., dissenting).

[125] Id.

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

[127] AT&T Corp. v. Iowa Utils. Bd., 525 U.S. 366, 428-29 (1999).

[128] Id. at 428.

[129] Id. at 428. For more on forced interoperability generally and the interoperability requirements of the Telecommunications Act of 1996, see Gus Hurwitz, Digital Duty to Deal, Data Portability, and Interoperability, in The GAI Report on the Digital Economy (2020).

[130] Verizon Commc’ns., Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 404 (2004).

[131] Id. at 407-08 (citing United States v. Colgate & Co., 250 U. S. 300, 307 (1919)).

[132] Iowa Utils. Bd., 525 U.S. at 428 (Breyer, J., dissenting).

[133] See U.S. Dep’t of Justice & Fed. Trade Comm’n, Antitrust Guidelines for Collaborations Among Competitors (2000); North Texas Specialty Physicians v. FTC, 528 F.3d 346 (5th Cir. 2008), cert. denied, 555 U.S. 1170 (2009); Texaco Inc. v. Dagher, 547 U.S. 1 (2006).

[134] Phillip Areeda, Essential Facilities: An Epithet in Need of Limiting Principles, 58 Antitrust L.J. 841, 852 (1989).

[135] 3A Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶771c (2d ed. 2002) (arguing that “the essential facility doctrine is both harmful and unnecessary and should be abandoned”).

[136] R. Pitofsky, D. Patterson & J. Hooks, The Essential Facilities Doctrine Under United States Antitrust Law, 70 Antitrust L.J. 443-462 (2002) (arguing that, as narrowly defined by various lower federal-court decisions, application of the doctrine is appropriate and may be needful when applied to “essential facilities” that consist of intellectual property).

[137] Diane P. Wood, The Old New (Or is it the New Old) Antitrust: “I’m Not Dead Yet!!”, 51 Loy. U. Chicago L.J. 1, 17 (2019).

[138] J. Gregory Sidak & Abbott B. Lipsky, Jr., Essential Facilities, 51 Stan. L. Rev. 1187, 1222-23 (1999).

[139] Christine S. Wilson & Keith Klovers, The Growing Nostalgia for Past Regulatory Misadventures and the Risk of Repeating These Mistakes with Big Tech, 8 J. Antitrust Enforcement 10 (2020). See also, Kobayashi & Wright, supra note 31.

[140] Wood, supra note 137, at 17.