The following article summarizes the discussion at the event.
1. Focus on concentration, not competition
Several panelists suggested that the draft guidelines appear to incorporate propositions at odds with foundational antitrust principles. Leemore Dafny (Harvard Business School) and Kristen Limarzi (Gibson Dunn) highlighted that the guidelines tend to suggest that their aim is to protect competitors and achieve deconcentration, without clearly explaining how that would benefit competition. Likewise, Daniel Francis (NYU Law School) said that Guideline 7 (mergers should not entrench or extend a dominant position) seemed inconsistent with Supreme Court precedent by suggesting that even mergers that lowered prices for consumers with no evidence of anticompetitive harms could be condemned due to increased market share.
2. Competing visions of guidelines
The panelists and enforcers articulated very different visions of how the guidelines are to be interpreted. It remains to be seen how the courts will ultimately treat the forthcoming guidelines.
Panelist vision focuses on unpredictability for merging parties: Multiple panelists were concerned that the draft guidelines do not appear to constrain agency discretion. For example, Guideline 7 on dominant firms gives the agencies a blank check to define any firm with a 30% market share in a particular product as dominant, regardless of the overall level of market concentration. Likewise, by Aviv Nevo’s (Director, FTC Bureau of Economics) admission, agencies would be willing to challenge a merger on any one of the three theories in Guidelines one through three (HHIs, unilateral effects, and coordinated effects), which could result in a lack of predictability for practitioners. Accordingly, one panelist suggested that there be a quick look in the merger guidelines with a safe harbor for defendants.
Agency vision focuses on flexibility for agencies and courts: Susan Athey (Chief Economist, DOJ Antitrust Division) stated that the guidelines merely illustrate theories of harm and that they are subject to rebuttal. Athey also claimed that there are portions of the guidelines that merging parties could cite to attempt to defend their transactions in court.
3. Panelist consensus that the vertical structural presumption in Guideline 6 is poorly supported.
50% threshold is unsupported: Almost every panelist thought the 50% structural threshold number in Guideline 6 was poorly supported and untethered to a theory of harm. Beyond that, most panelists found vertical mergers to be a complex field, for which a structural presumption was ill-suited. One panelist stated that while the guidelines are correct to reject a presumption that vertical mergers are efficient, the 50% number did not relate to how a particular merger might dampen competition. Another panelist agreed with this assessment and offered it as a reason that courts will not follow an unsupported presumption. Indeed, the only panelist to defend the presumption did so for expediency because the incentive and ability to foreclose test in Guideline 5 is very difficult for an enforcer to meet.
4. Panelists: Market definition analysis needs work
Courts will look for the best market: Nevo explained that the draft guidelines rely more heavily on the idea that there really is no “best” market, and the requirement for the agency should be to show harm in any market as opposed to the current guidelines which state that the correct market definition is usually the smallest relevant market that satisfies the Hypothetical Monopolist Test.
Panelists were skeptical of the revision. One stated that the guidelines should constrain the government on market definition because courts will not accept guidelines under which the government can always win using some option on the menu. Another indicated that the smallest relevant market language in the current guidelines is a useful constraint on the agencies.
The panelists also thought that there needs to be some explanation of how the three factors listed in Section 3 of the Guidelines as tools for market definition (direct evidence, Brown Shoe factors, and HMT) interact. One panelist stated that her biggest issue with the current draft guidelines on market definition is that they appear to put the HMT on the same footing as practical Brown Shoe indica, when the former is higher quality evidence of a relevant market.
5. Panelists were critical of Guideline 10 (on mergers involving a multi-sided platform, the agencies examine competition between platforms, on a platform, or to displace a platform).
Susan Athey stated the guidelines are meant to address increasingly common fact patterns like those associated with platforms, which touch on multiple theories of harm at the same time. Athey highlighted a few problems in the platform space that inform agency thinking about why there needs to be a guideline on platforms. Athey explained that dominant platforms can steer customers to their platform for products because of the benefits realized from economies of scale and then withhold those products from other platforms (i.e., a refusal to deal), never allowing those platforms to develop. This sequence is allegedly destructive to competition.
Guideline 10 does not properly account for network effects: Three panelists criticized Guideline 10’s treatment of network effects. One said the guideline ignores the potential beneficial network effects arising from platform growth via merger, and that if two smaller platforms merge, they can better constrain the dominant platform. Regarding a nascent competitive threat, the panelist stated that the guideline does not properly acknowledge the benefits of network effects and how to weigh them against potential harm. Another panelist criticized the guideline’s suggestion that network effects favor the incumbent monopolist, which is not always the case, as differentiation and congestion effects at certain levels of scale show. A third panelist offered a more muted criticism: he stated that there might need to be a note in the guidelines that a small, but significant and non-transitory increase in price needs to be adjusted for platform markets to recognize the efficiencies provided by network effects.
Guideline 10 suggests that there is a theory of harm unique to platforms: One panelist cautioned that Guideline 10 also suggests that there should be special presumptions or theories of harm in a platform context. In response, Athey stated that she envisions Guideline 10 as just an application of the other guidelines, like the labor market guidelines.
Conflict of interest language misguided: Athey asked about the conflict of interest language in Guideline 10E. One panelist stated that it is not an antitrust concern and expressed that a presumptive equal treatment obligation suggested by the conflict of interest language would be destructive to the core benefits of vertical integration because vertical integration benefits can arise from special treatment to an integrated division. Another panelist agreed that conflict of interest does not map onto antitrust law and further noted that Guideline 10 does not connect conflict of interest to market power analysis. The panelist also noted the unanswered questions resulting from reading Guideline 10 broadly: can streaming services prefer their own products, or grocery stores prefer their own brands?
The pros of Guideline 10: One panelist stated that Guideline 10’s taxonomy of competition between platforms, among platforms, and displacement was helpful and consistent with economics. Another noted that the platform guideline is well structured with its inclusion of scenarios A through D.
How to improve Guideline 10: One panelist noted that the displacement language in the guideline is too strong because displacement is not a winner-take-all issue. Therefore, he cautioned that the guideline should clarify that displacement does not mean one competitor disappears. In contrast, the panelist suggested that the language on nascent competitive threats needs to be stronger because it currently includes threats from systematic acquisition of rival networks, whereas there need not be a series of such acquisitions to cause competitive harm. Another panelist agreed that the “series of acquisitions” language is underinclusive in capturing the harms. A third panelist advised that if there is going to be a guideline on platforms (which he disagrees with), he would state in the guidelines that a multi-sided business can compete with a single-sided business, to avoid the outcome in Sabre/Farelogix.
6. Panelists were mostly critical of Guideline 7 (that a merger may be used to entrench a dominant firm).
Athey stated that the 30% number in the guideline has captured a lot of attention and it is a number the agencies are thinking about. She asked the panelists for views on Guideline 7 other than on the 30% threshold.
Guideline 7 punishes efficiencies: One panelist stated that the biggest problem with Guideline 7 is that it pays lip service to not punishing efficiencies but does not delineate when they should not be punished. In contrast, he stated that the content of the guideline “winks at an efficiency offense” (e.g., in its discussion of scale economies and network effects). He mentioned that the guideline seems to imply that an enforcer can punish a dominant firm for a merger that has efficiencies that result in lower consumer costs but further increases the merged firm’s market share. The panelist cautioned that would violate the core of Section 2, which for 130 years has established a distinction between good ways and bad ways to acquire monopoly power and be contrary to the Supreme Court precedent. Likewise, another panelist criticized Guideline 7’s definition of a dominant firm as one with a market share of 30% in particular products, because that threshold is too low.
Enforcer view that Guideline 7 is the only way to prevent refusal to deal: One panelist lamented that the “refusal to deal” label has been used too broadly to immunize conduct in a way that could entrench or create or maintain a monopoly. She noted that Guideline 7 (the entrenchment guideline) attempts to deal with this problem by focusing on incipiency if an enforcer suspects the merging parties will refuse to deal post-merger. Athey stated she agreed, and that merger review may be a way to preclude certain refusals to deal.
7. Panelists and moderators agreed that economic evidence need not be econometric, and that the guidelines should add analysis on economic harms.
The panelists indicated a view that economic evidence can be in the form of ordinary course documents and need not be based on perfect economic models. Two of the panelists thought the guidelines should express that what constitutes productive economic analysis differs from case to case. One panelist expressed that defendants frequently spend too much time discrediting the economic models instead of engaging with the theory of harm.
