Ten years ago the U.S. antitrust agencies began using conduct remedies more frequently and more expansively than ever before. Research and experience, however, highlighted the limitations of such remedies, and so more recently the agencies have restated their determination to avoid the use of conduct remedies wherever possible and to strengthen remedies policy generally. Despite those statements, the agencies have in fact continued to rely on conduct remedies and indeed further expanded their use in ever more problematic ways. Three examples discussed in this article are the recent amendment to the original Ticketmaster-Live Nation settlement, the merger of Staples and Essendant, and the Sprint/T-Mobile merger. The conclusion of this review is that little appears to have been learned about the weaknesses of conduct remedies over the past decade.
By John Kwoka1
A decade ago, U.S. antitrust policy embarked on an experiment in expansive use of conduct remedies for mergers. Several major cases were settled with agreements that the merged firms—as a condition for approval of their mergers—would not engage in specific anticompetitive actions. In 2011 the Antitrust Division of the Justice Department issued a revised Remedies Guide that indicated greater receptivity to conduct remedies than in an earlier version of the Guide.
But a growing body of experience and research was finding that conduct remedies were hard to write, even more difficult to enforce, and often simply ineffective. More recently, these cautionary lessons seem to have had an impact. Three years ago, the new Assistant Attorney General for Antitrust described conduct remedies as “fundamentally regulatory, imposing on-going government oversight” on markets and promised “to return to the preferred focus on structural relief.”2 He pointedly withdrew the revised Remedies Guide in favor of the earlier version that emphasized the very limited role of conduct remedies. Similarly, the incoming chair of the Federal Trade Commission stated his determination to improve on all past remedies policy, not just conduct remedies. He cited as unacceptable a 30 percent failure rate on certain divestiture remedies and asserted that the rate “need[ed] to be “lowered substantially or, ideally, zeroed out altogether.”3
Since that time, however, the agencies have not only failed to limit reliance on conduct remedies: they have continued to use them and even extended their use in more problematic directions. This essay begins with a reminder of the flaws inherent in conduct remedies and then describes three recent cases that raise a question of whether anything has been learned from recent experience with such remedies.
II. CONDUCT REMEDIES: THE BAD AND THE TRULY UGLY
Conduct remedies represent an effort to allow a merger to proceed while preventing the competitive harms that it represents through a series of prohibitions or prescriptions. The prohibitions might be on foreclosure or leveraging, the exchange of competitively sensitive information, or retaliation against independent rivals, while the prescriptions might entail mandatory access or must-supply conditions. But conduct remedies suffer from several inherent defects. Most fundamentally, they require a firm to act against its own interests, that is, in ways that diminish its profit. As a result, they inevitably create incentives for the firm to evade or avoid them. Evasion is facilitated by the fact that conduct remedies are difficult to write without ambiguities or omissions that firms can use to circumvent their intent. And under any circumstances, they are costly to enforce, since offending actions can be hard to observe, and the agencies are not structured as on-going regulators.
Individual case experience includes numerous examples of flawed conduct remedies and the failure to preserve competition. Broader economic evidence of the actual outcomes of conduct remedies corroborated these concerns.4 All of this led the Antitrust Division and the FTC most recently to declare their strong preference for structural remedies—divestitures—in all but the most unusual of circumstances. As the following examples show, however, those statements and the actual practice of the agencies have diverged.
III. TICKETMASTER-LIVE NATION: THIS TIME WE REALLY MEAN IT
At the time of their 2010 merger, Ticketmaster was the dominant ticketing services company and Live Nation a very large concert promoter. The merger raised horizontal concerns since Live Nation had begun to deploy a ticketing services operation in competition with Ticketmaster. The merger also raised vertical concerns, most especially that the merged company might condition the provision of either Ticketmaster’s ticketing services (or Live Nation’s concerts) on a venue also taking the other business of the merged company. The Justice Department approved the merger subject to an order that sought to prohibit these practices.5
The remedy contained two major relevant provisions. One prohibited retaliation, which was defined as “refusing to provide live entertainment events to a venue owner, or providing live entertainment events to a venue owner on less favorable terms, for the purpose of punishing or disciplining a venue owner because the venue owner has contracted or is contemplating contracting with a company other than defendants for primary ticketing services.”
The order then assured the merging company that “‘retaliate’ does not mean pursuing a more advantageous deal with a competing venue owner” and that the merged companies were not prohibited from “bundling their services and product in any combination or from exercising their business judgment in whether and how to pursue, develop, expand, or compete for any business, but subject to the terms of the order. Clearly, the effectiveness of this anti-retaliation provision was dependent on whether terms such as “less favorable,” “for the purpose of,” “bundling,” “business judgment,” and “whether … and how to … compete” could operationally distinguish normal from anticompetitive behavior. Moreover, contracts between Ticketmaster and venues involved numerous terms and provisions, typically over several years, making identification of retaliation nearly impossible.
But it was the other provision, as interpreted by Live Nation, that gutted the remedy. The key language stated that the merged firm must not “condition or threaten to condition the provision of Live Nation Entertainment Events to a venue owner” depending on whether the latter had contracted with another company for ticketing services. But as the Justice Department acknowledged in 2020, beginning “shortly after the decree was entered in 2010,”6 the merged companies “repeatedly conditioned and threatened to condition Live Nation’s provision of live concerts on a venue’s purchase of Ticketmaster’s ticketing services, and they have retaliated against venues that opted to use competing ticketing services, all in violation of the plain language of the decree .”7 The Justice Department recounted six specific episodes, dating back to 2012, in which Ticketmaster threatened to withhold, and in many cases in fact did withhold, Live Nation concerts unless a venue also used Ticketmaster’s services.
How did Ticketmaster-Live Nation justify these actions? The company’s position was that language prohibiting it from withholding “Live Nation Entertainment Events” meant all events, so that it was altogether free to condition or withhold some events from venue owners.8 Whatever the plausibility of this interpretation, the Justice Department acknowledged that it had resulted in a decade of harm to consumers and competition of the very sort that was originally feared. Its response was to submit an amended order “clarifying” the original language so as “to avoid any doubt” about its intent, effectively restating in 2020 what it obviously intended in 2011. It did not charge Ticketmaster with violating the “plain language” of the earlier decree. It imposed no penalty or fine for a decade in which Ticketmaster conditioned, retaliated, and profited. It sought no disgorgement of profits earned based on this strategic misinterpretation.9
In the starkest terms, this experience demonstrates the many inherent defects of a conduct remedy. As noted at the outset, these remedies are difficult to write, and companies have the time, resources, and incentives to find ways of avoiding their intent. They are difficult to enforce, even in the face of years of violations of the “plain language” of an order. And at the end of the day, it appears to be impossible to penalize them for any behavior for which there is a pretextual excuse. Ticketmaster has simply gotten away with it.
IV. STAPLES-ESSENDANT: GOING VERTICAL
Over many years, indeed decades, the antitrust agencies have accepted the specious argument that few if any vertical mergers raise competitive concerns but instead, they confer cost savings from the elimination of double marginalization and other efficiencies.10 As a result, there have been few challenges to vertical mergers, and in those few cases where competitive concerns have been identified, conduct remedies have been adopted as solutions. Such is the case with the FTC’s decision with respect to Staples’s acquisition of controlling interest in Essendant.11 Staples is one of two large two office superstores that sell office supplies to businesses and retail customers. Essendant is the larger of two national distributors of office supplies, selling both to Staples and to independent dealers in competition with Staples. In September 2018, Staples announced its intention to acquire Essendant. An obvious competitive concern with this vertical merger was that the merged firm would become the crucial supplier to its downstream rivals, so that it could raise price to those rivals, relax the competitive constraint that they posed, and thereby raise its own price and profit.
The FTC majority nonetheless approved the acquisition in January 2019. It dismissed concern over foreclosure by alluding to nonpublic evidence that, in the face of a price increase by the merged company, “many” independent dealers could simply switch to the one other large distributor and that any customers lost by the independents would “not likely” switch to Staples anyway. The minority FTC commissioners disputed both the evidence and the conclusions of the majority with respect to concerns over foreclosure.
The FTC majority did, however, acknowledge one mechanism through which the merged company might gain an anticompetitive advantage. Since Essendant compiled data from resellers about end customers, the post-merger Staples would have access to this competitively sensitive information. Together with reseller cost data, this would permit Staples to determine how much it could raise its prices without losing customers to rivals. To address this one concern, the majority adopted a conduct remedy in the form of a firewall intended to prevent the exchange of competitively sensitive information between parts of the merged company’s operations. It is this one acknowledged competitive concern and the associated conduct remedy that will be the sole focus here.
The firewall in this case provided that employees who are supposed to perform Essendant’s wholesaling functions with respect to its resellers are prohibited from disclosing competitively sensitive information to any other employees. But like most firewalls, this order contains exceptions to allow for what are deemed to be normal and necessary information transfer, as well as provisions concerning the movement of employees across the relevant divisions. In this case the order provides for three exceptions.
First, certain Staples employees “performing wholesale, legal and regulatory, or shared services functions or members of a prescribed management oversight group will have access to protected” information. This exception was intended to permit a management group — newly formed to oversee joint operations — but not other employees to access the data for normal business decisions. Secondly, the rule allowed access to information about Essendant resellers if presented in an “aggregated … and anonymized form.” It is not difficult to envision how these channels can result in the transfer of at least some useful competitively sensitive information.
Furthermore, like most firewalls, this contains a provision for the routine transfer within the merged company not of the competitively sensitive information itself but of employees that know or have access to such information. Unrestrained, of course, the transfer of knowledgeable employees can be a near-perfect substitute for direct information exchange, so conduct orders involving firewalls strive to prevent misuse of such transfers, again without impeding normal internal reassignments.
The difficulties associated with these conflicting objectives have in the past resulted in some strange conduct remedies. In approving the merger of Google and ITA in 2011, for example, the Justice Department decree expressly permitted the merged company to assign an employee to any job, even if that person had access to competitively sensitive information, unless there was “evidence of intentional reliance on information other than information that was retained in the unaided memory of such employee (provided that memory is “unaided” if the employee has not intentionally memorized the information for the purpose of retaining and subsequently using or disclosing it).”12
That is, transfers of knowledgeable people were permitted unless the person deliberately memorized the competitively sensitive information in order to port it across the firewall. It is difficult to believe anyone would view this provision as enforceable and effective — or as antitrust policy.13
The Staples-Essendant order does not include any language similar to that in Google-ITA, but it does explicitly permit otherwise firewalled persons to be transferred so long as that transfer is in accordance with “usual and customary business practices,” subject to two further conditions. The first is a requirement for advance notice to an outside monitor of such an intended transfer, and secondly, a minimum waiting period of 6 months prior to the transfer during which the person would no longer have access to such information. While these provisions avoid the obviously vacuous nature of the Google-ITA remedy, they ultimately permit transfers of competitively sensitive information, the potential for which undoubtedly fosters discomfort of independent rivals and the erosion of competition.
Apart from these specifically discussed and disputed channels, information conveys in many ways, not all formal, and not all intentional. As a result, there is little direct evidence about the effectiveness of firewalls. Evidence from an arguably similar context, namely, firewalls in financial settings, is not encouraging. One summary of that literature reports that “informational firewalls are extremely porous and that traders and executives in diversified financial firms are able to access and misuse information obtained across internal firewalls.”14
Given the often facile reliance on information firewalls, there would seem to be an urgent need to study their effectiveness, in particular, determining what type of firewall and associated conditions, and what firm and market circumstances are more vs. less likely to result in their successful application. The present practice of simply writing and relying on firewalls of unproven, and sometimes dubious, effectiveness needs to be stayed while the agencies conduct retrospective analyses of past firewalls. Retrospectives have been shown extremely valuable in better understanding the outcomes of mergers and of remedies in general. It is now time to apply that methodology to firewalls rather than assuming their effectiveness to justify approval of anticompetitive mergers.
V. SPRINT/T-MOBILE: MASQUERADING AS MERGER CONTROL
Classic remedy policy toward horizontal mergers involves divestiture of overlapping assets, perhaps accompanied by some temporary transition services to ensure the business prospects of the divested assets. Over time, this model has been extended to involve less than complete divestitures. But divestitures that do not include a full complement of closely associated functions, supply arrangements, etc., have often failed,15 and an even worse fate might be expected for remedies involving ever fewer asset divestitures and requiring more third-party assistance subject to a conduct order.
The new frontier in this model of “competitor creation” would seem to be DOJ’s remedy in the case of the Sprint/T-Mobile merger. In early 2018, Sprint and T-Mobile, two of the four national wireless carriers, announced their intent to merge. The companies promised benefits in the form of faster deployment of new technologies and, later, consumer price freezes, but most observers doubted that a four-to-three merger would in fact improve market performance–or that the Justice Department would approve such a major consolidation of the industry. Indeed, DOJ itself later stated:
The merger would eliminate Sprint as an independent competitor, reducing the number of national facilities-based mobile carriers from four to three. The merger would cause the merged T-Mobile and Sprint (“New T-Mobile” to compete less aggressively. Additionally, the merger would likely make it easier for the three remaining national facilities-based mobile wireless carriers to coordinate their pricing, promotions, and service offerings.
Having conceded the need for four competitors, it was somewhat unexpected in July 2019 that DOJ indicated it had settled with the parties on terms that would permit the merger to proceed. The settlement imposed an unusual set of conditions on Sprint and T-Mobile, designed to create a new wireless competitor to replace the one whose elimination it was approving. That new competitor would be Dish, a national satellite TV distributor with no wireless operation or experience. Dish is supposed to evolve into that new competitor based on some modest asset and operations divestitures, together with crucial conduct-based services and assistance from the merged Sprint/T-Mobile.
The settlement set out a long list of actions to be taken by the merging parties and by Dish. First, in the short term, the merged T-Mobile would transfer to Dish its prepaid wireless operation — a minor and not very valuable business. Since Dish has no wireless infrastructure, it would simply be reselling T-Mobile services, rebranded as its own but completely dependent on T-Mobile as its supplier. Research as well as common sense shows that resale does not constitute full scale competition and, indeed, when dependent on a direct rival, may not constitute competition at all.16
Second, within a year, Dish must begin providing national postpaid services and continue to expand that service over seven years, at every step supposedly assisted by the merged T-Mobile. T-Mobile is supposed to start transferring thousands of its cell sites and hundreds of its retail stores as it decommissions those that it no longer needs. Since these involve duplication, T-Mobile would presumably shed its inferior assets, cementing Dish’s market disadvantage.
Third, for up to three years, T-Mobile would have to provide necessary “transition services” to Dish for any prepaid as well as postpaid services. These transition services include billing, customer care, SIM card procurement, device provisioning, and “all other services used by Prepaid Assets.” This arrangement would make Dish’s nascent service offerings fully dependent on T-Mobile’s willingness to assist its new competitor in providing high quality service, assistance that is obviously contrary to T-Mobile’s own interests.
And fourth, since Dish will only slowly be able to provide its own services, T-Mobile would have to provide Dish with sound and seamless — “robust” — access to its cell sites for up to seven years, in order for Dish to offer postpaid services. This provision makes Dish fully dependent on its direct competitor to launch and expand its services. Moreover, even if successful, the remedy does not envision Dish as a new facilities-based competitor capable of replacing the one eliminated by the merger until at least 2026.
This remedy fails the test of plausibility. More than perhaps any other remedy, it involves a linkage between the putative replacement wireless carrier and the merged company that puts the fate of the new competitor fully in the hands of its incumbent rival that has both the means and the incentive to prevent its emergence as a strong new firm. As written, Dish will be dependent on the merged T-Mobile providing good support for the divested prepaid business, support that does not compromise Dish’s start up so that it is branded as a poor-quality rival that loses customers and perhaps employees. Then it will be dependent on T-Mobile transferring thousands of cell sites and hundreds of retail locations on a schedule and of a quality that allows Dish to start its own competitive facilities-based service. Dish will also be dependent on T-Mobile providing complex and crucial transition services to assist Dish. And as Dish ramps up over seven years or more, it must hope that T-Mobile provides “robust” access to its cell sites so that Dish can piggyback its own nascent services on that of its direct competition.
This scenario is not a remedy to an antitrust problem. It is an utterly speculative story, contrary to the actual incentives of key parties, about how Dish might eventually grow and become a fourth competitor. That will not happen for the simple reason that that scenario depends at each stage on behavior by T-Mobile that is plainly contrary to its own interest, behavior that would simply help its competitor and lower its own profit. Given the innumerable ways whereby T-Mobile can compromise, handicap, and even sabotage Dish’s emergence as a full-blown competitor, there is little reason to expect this process to end in any other way. Indeed, few remedies embody clearer adverse incentives and more implausible expectations.
Even more worrisome is the fact that, if this plainly anticompetitive merger can be rescued with such an implausible remedy, it is unclear what merger cannot be so resolved.
At the outset, this essay posed the question: what has been learned about conduct remedies over the past decade? The Ticketmaster-Live Nation experience demonstrates again what was always known, namely, that it may be easy for a firm subject to a conduct remedy to evade or avoid its intent and achieve the feared anticompetitive effects. The remedy for the Staples-Essendant merger, apart from dismissing concern with foreclosure, rests entirely on unproven and to some degree doubtful restraint on information exchange. And the resolution of the Sprint/T-Mobile merger illustrates the lengths to which the antitrust agencies now seem to be prepared to use remedies that will not predictably or even plausibly resolve the competitive issues with a merger.
These experiences suggest that the hard lessons about conduct remedies have not in fact been learned.
1 Northeastern University.
2 Makan Delrahim, “Modernizing the Merger Review Process,” Sept. 25, 2018.
3 Joseph Simons, Completed Initial Questionnaire, February 2018.
4 For an elaboration of the arguments and evidence, see John Kwoka, “Merger Remedies An Incentives/Constraints Framework,” Antitrust Bulletin, 2017.
5 For discussion of the horizontal aspects and a fuller description of the vertical issues, see John Kwoka, “Rockonomics: The Ticketmaster-Live Nation Merger,” in The Antitrust Revolution, 7th ed, J. Kwoka & L. White, eds., 2019.
6 Justice Department Memorandum in Support of Motion to Modify Final Judgment and Enter Amended Final Judgment, January 8, 2020, p. 6.
7 Ibid. p. 1.
8 Ibid. p. 11. This confirmed by language in the Amended Final Judgement that Live Nation “waives any argument that this [provision] …only prohibits retaliation or conditioning with respect to all Live Nation content,” p. 19.
9 The amended decree also provides for a monitor and enhanced fines for future violations, neither of which addresses past violations. Moreover, the new decree runs for only five years instead of the original ten, halving the period under which the merged company is subject to any constraint (as well as postponing it by a decade).
10 In reality there is often a risk of foreclosure, and the proposition about double marginalization depends on a number of strong assumptions and limitations. John Kwoka & Margaret Slade, “Second Thoughts on Double Marginalization,” Antitrust, Spring 2020.
11 Federal Trade Commission, “FTC Imposes Conditions on Staples’ Acquisition of Essendant,” Washington DC, January 2019.
12 U.S. v. Google Inc and ITA Software Inc, Competitive Impact Statement, Apr. 8, 2011.
13 Even if strictly followed, this prohibition does not specify any minimum length of stay, so that the relevant person could apparently be transferred as frequently as desired. The entire passage was the cause of some mirth within one of the companies.
14 Christopher Gorman, “Are Chinese Walls the Best Solution to the Problems of Insider Trading and Conflicts of Interest in Broker-Dealers?” Fordham Journal of Corporate and Finance Law, 2004.
15 The 1999 FTC Divestiture Study reported a 40 percent failure rate for divestitures of less than an entire business unit. Despite that, the FTC inexplicably continued to use such remedies, for which their follow-up study in 2017 found an even higher failure rate.
16 This has been the conclusion of the FCC. Federal Communications Commission, 20th Wireless Competition Report, Sept. 2017, n. 99.