Articles Barry Hawk

Antitrust Law & Policies from Brussels to Washington D.C.: The Atlantic Divide

Brussels 16 October 2015

Selection of articles Barry Hawk

Skadden Arps and Fordham University Law School – New York, USA

Separating the Wheat from the Chaff: How the U.S. Courts Analyze Antitrust Evidence

written with James Keyte

European University Institute Robert Schuman Centre for Advanced Studies 2009 EU Competition Law and Policy Workshop/Proceedings

To be published in the following volume: Claus-Dieter Ehlermann and Mel Marquis (eds.), European Competition Law Annual 2009: Evaluation of Evidence and its Judicial Review in Competition Cases, Hart Publishing, Oxford and Portland (forthcoming 2010).

Separating the Wheat from the Chaff: How the U.S. Courts Analyze Antitrust Evidence

Barry Hawk and James Keyte*

Unlike in many countries, the U.S. court system provides essentially open access to potential antitrust plaintiffs: a plaintiff who loses a case does not have to pay defendant's attorney fees and costs, and winning plaintiffs can get their fees and costs reimbursed by defendants. While courts have the power to sanction a plaintiff for bringing a frivolous case, those cases are relatively rare.1 As a result, U.S. courts have the potential to be clogged with an enormous number and variety of cases – many perfectly legitimate, but many not. It is left to the courts, then, to apply U.S. procedure and law to determine which cases may proceed to trial. This paper addresses how U.S. courts analyze evidence in antitrust cases with a particular focus on how U.S. court procedure and application of law have evolved to limit cases to those aimed at redressing legitimate antitrust violations. No doubt it is a broad topic; one could write dozens of articles that touch upon the evaluation of evidence by U.S. courts in antitrust cases. The objective here, however, is both less ambitious and hopefully more useful for those who do not practice antitrust in the U.S., yet wish to understand more generally how U.S. courts assess the merits of antitrust cases. With this objective in mind, a review of recent U.S. antitrust cases suggests that in order to understand the evidentiary battlegrounds in U.S. antitrust cases, one must understand (i) when courts have an opportunity to assess evidence (both alleged and offered), (ii) what types of antitrust cases are being litigated and which particular elements of a claim raise the most evidentiary challenges for plaintiffs, and (iii) how courts are assessing the quality and quantum of evidence necessary to permit an antitrust case to move forward to final resolution on the merits. The question of when a U.S. court has the opportunity to assess antitrust evidence is an important one that has significant practical implications in many U.S. antitrust cases. Even at the pleading stage (the filing of an initial complaint), antitrust courts often are asked to assess whether the complaint alleges enough "evidence" (facts that must be taken as true at this early stage) to support each required element of the antitrust claim at issue. For example, in conspiracy claims under Section 1 of the Sherman Act,2 the initial and often decisive battle may center on whether the complaint itself sets forth sufficient factual allegations to establish an "agreement" that unreasonably restrains trade. Indeed, under the U.S. Supreme Court's recent opinion in Bell Atlantic Corp. v. Twombly,3 antitrust cases must be dismissed at an

* Barry Hawk and James Keyte are partners in the Antitrust Group of Skadden, Arps, Slate, Meagher & Flom LLP & Affiliates. The authors would like to thank Adam Hosmer-Henner, Anna Peltier, and Michael Roth for their invaluable assistance in the preparation of this paper. 1 See FED. R. CIV. P. 11. 2 15 U.S.C. § 1 ("Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal."). 3 550 U.S. 544 (2007).

Cite as: Barry Hawk and James Keyte, “Separating the Wheat from the Chaff: How the U.S. Courts Analyze Antitrust Evidence”, in Claus-Dieter Ehlermann and Mel Marquis, eds., European Competition Law Annual 2009: Evaluation of Evidence and its Judicial Review in Competition Cases, forthcoming, Hart Publishing

2 EUI-RSCAS / Competition 2009 / Proceedings

early stage if an antitrust plaintiff cannot allege sufficient facts to make a violation likely. It is important, therefore, to understand at what stages of a case U.S. antitrust courts have the opportunity to assess antitrust "evidence" and in which procedural contexts courts are summarily rejecting cases. A review of recent U.S. antitrust cases also confirms that the nature of the evidentiary battle depends significantly on what type of claim is being asserted, as well as the legal elements of a claim that a plaintiff must allege and eventually prove. For example, in many conspiracy cases, the primary evidentiary focus will be on whether there is enough evidence of an illegal agreement among otherwise independent horizontal competitors. By contrast, in cases involving facially legitimate horizontal collaborations (e.g., joint ventures) or "vertical" agreements (e.g., between a manufacturer and a distributor), the existence of an agreement often is not in serious dispute and the evidentiary fight is over whether the agreement unreasonably restrains trade and harms competition in a relevant antitrust market. Similarly, in Section 2 cases, 4 the initial evidentiary fight usually centers around whether the defendant is in fact a "monopolist" in a well-defined market; very few cases make it to the stage where the dispositive issue is whether the defendant monopolist engaged in exclusionary conduct. And, in Section 7 merger cases, 5 the battle rarely focuses on the nature of the parties' agreement and more on market structure (market definition, concentration and entry conditions) and likely competitive effects of the proposed transaction. Finally, in the context of these various claims, it is useful to understand how U.S. courts determine what quality and quantum of evidence are required or most useful for their analysis. For example, there is an established principle in U.S. courts to credit so-called "direct" evidence on an issue much more than circumstantial evidence. In recent years, the over reliance by plaintiffs in Sherman Act cases on what may be characterized as "ambiguous evidence" has tended to lead to the rejection of claims at an early or intermediate procedural stage. Similarly, U.S. antitrust courts have devised common law evidentiary "screens" (e.g., for evidence of a defendant's "market power") in an attempt to ensure that plaintiff's claim redress harm to overall competition rather than just to itself. And there is now a fairly clear body of decisional law defining when a plaintiff must offer expert economic evidence on a particular element (e.g., market definition) and what that evidence must entail. 6 An understanding of these principles and trends not only helps practitioners assess antitrust risk for clients and devise the most effective litigation strategies (for both plaintiffs and defendants), it also enables all of us to step back and assess whether U.S. courts are in fact separating the wheat from the chaff in reviewing the thousands of antitrust cases filed each year.

4 15 U.S.C. § 2 ("Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty of a felony . . . ."). 5 15 U.S.C. § 18 ("No person . . . shall acquire, directly or indirectly, the whole or any part of the stock or other share capital and no person . . . shall acquire the whole or any part of the assets of another person . . . , where in any line of commerce or in any activity affecting commerce in any section of the country, the effect of such acquisition may be substantially to lessen competition, or to tend to create a monopoly."). 6 See FED. R. EVID. 702; see also Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 589 n.7 (1993).

EUI-RSCAS / Competition 2009 / Proceedings 3

It is a lot to sort out, but the results are revealing. For example, we can observe that when the issue is the existence of an illegal agreement among horizontal competitors (a large portion of U.S. antitrust cases), U.S. courts are most active in putting antitrust plaintiffs to their proof early on and with little flexibility. These types of conspiracy cases are intensely practical factual battles with little room or necessity for sophisticated economic analysis. At the other end of the spectrum, in Section 7 merger cases the issue of market definition and the competitive effects resulting from a proposed merger have become intensely "economic" both as a matter of theory and quantitative proof – indeed, here, there is a recent trend to de- emphasize qualitative evidence such as the opinions of customers and other industry participants. Similar discernible evidentiary patterns and trends have emerged in Section 1 cases involving competitors or vertically-related firms as well as in Section 2 monopolization cases. In an attempt to assess these topics in an orderly fashion, this paper is divided into four parts. First, we offer a brief primer on U.S. court procedure and substantive antitrust standards. Second, we assess Section 1 cases, separately analyzing those in which there is not supposed to be an agreement at all (e.g., price fixing among independent competitors) and those in which the existence of an agreement is not in dispute and the issue focuses on alleged anticompetitive effects. Third, we review recent Section 2 cases to evaluate the types of issues and evidence – often economic – that come into play to a greater degree when courts assess monopolization claims, particularly on issues of monopoly power and alleged "exclusionary conduct." Fourth, we briefly review some of the evidentiary issues relating to antitrust injury and causation – the bane of plaintiffs who in fact may have been injured as a result of defendant's conduct but must still prove that it flows from harm to the market. Finally, we review several recent Section 7 merger cases, again focusing primarily on the type of factual and economic evidence – in the context of applicable preliminary injunction standards – that courts look for in determining whether to block a proposed merger between competitors.

I. SETTING THE STAGE FOR UNDERSTANDING U.S. ANTITRUST LITIGATION AND EVIDENCE

A. The Stages of a Typical U.S. Antitrust Litigation

With the exception of merger challenges – which almost invariably proceed on a preliminary injunction track7 – most civil antitrust cases involve three stages where the court assesses the sufficiency of evidence. At the initial stage, a trial judge may dismiss a case if the complaint itself fails to state "evidence" that can support a claim.8 Although the court will presume the

7 See infra Section V.A. 8 See FED. R. CIV. P. 12(b)(6). A complaint may be dismissed if the plaintiff "can prove no set of facts, consistent with its complaint, that would entitle it to relief." Elecs. Commc'ns Corp. v. Toshiba Am. Consumer (cont'd)

4 EUI-RSCAS / Competition 2009 / Proceedings

truth of all factual allegations in the complaint and make all reasonable inferences in favor of the non-moving party, to survive a motion to dismiss the complaint must still present "enough facts to state a claim to relief that is plausible on its face."9 Motions to dismiss are critical points in litigation as a victory for defendants allows them to escape costly and burdensome discovery altogether. If a plaintiff survives a motion to dismiss the complaint, the next stage at which a court seriously assesses the sufficiency of evidence is on a defendant's motion for summary judgment. Under U.S. procedural rules, a court may grant summary judgment only if the pleadings, depositions, answers to interrogatories, and admissions on file, together with affidavits, if any, "show that there is no genuine issue as to any material fact and that the [moving party] is entitled to judgment as a matter of law."10 Again, the moving party bears the burden of proving that no genuine issue of material fact exists, and the court will view the underlying facts and construe all reasonable inferences in the light most favorable to the party opposing the motion.11 The mere existence of some fact in support of the nonmoving party is insufficient; the evidence must enable a reasonable jury to find for the nonmoving party on that issue.12 A material fact is one which could alter the outcome of an issue and a genuine dispute is one where a rational person could support the viewpoint of the holder of the burden of proof on the disputed issue.13 If neither plaintiffs nor defendants are granted summary judgment, then the case will usually proceed to trial. Finally, even after trial, a judge has the ability to set aside the jury's ruling on a motion for judgment notwithstanding the verdict. In deciding these motions, the court considers all evidence and makes all reasonable inferences in favor of the party opposed to the motion.14 The court may then set aside the verdict only if there are insufficient facts to support the jury's verdict.15 Put another way, if the facts and inferences overwhelmingly favor one party and the court believes that reasonable people could not arrive at a contrary verdict, a motion for judgment notwithstanding the verdict is proper.16

______(cont'd from previous page) Prods. Inc., 129 F.3d 240, 242-43 (2d Cir. 1997). The moving party has the burden of proving that no claim exists. See Broadcom v. Qualcomm Inc., 501 F.3d 297, 307 (3d Cir. 2007); see also FED. R. CIV. P. 8. 9 Bell Atl. Corp. v. Twombly, 550 U.S. 544, 570 (2007). 10 FED. R. CIV. P. 56(c). 11 See Celotex Corp. v. Catrett, 477 U.S. 317, 322-23 (1986); see also Matsushita Elec. Indus. Co. v. Zenith Radio Co., 475 U.S. 574, 585 n.10 (1986); Toledo Mack Sales & Serv., Inc. v. Mack Trucks Inc., 530 F.3d 204, 219 (3d Cir. 2008). 12 See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 249-50 (1986). 13 See Horowitz v. Fed. Kemper Life Assurance Co., 57 F.3d 300, 302 n.1 (3d Cir. 1995). 14 See Key Enters. of Del. Inc. v. Venice Hosp., 919 F.2d 1550, 1556 (11th Cir. 1990), opinion vacated, 979 F.2d 806 (11th Cir. 1992), order vacated, 9 F.3d 893 (11th Cir. 1993). 15 Ibid. 16 Ibid.

EUI-RSCAS / Competition 2009 / Proceedings 5

B. The Potential Substantive Standards and Burdens of Proof

In addition to the procedural context, the substantive standard of review has a significant impact on the type and quality of evidence that the court may require before an antitrust case can proceed to trial. In the U.S., the three substantive standards of review are per se, "quick look," and rule of reason. These standards will be reviewed from the least burdensome on an antitrust plaintiff – per se – to the most burdensome – full rule of reason analysis.

1. Per se Categories

It is well established in the U.S. courts that "there are certain agreements or practices which because of their pernicious effect on competition and lack of any redeeming virtue are conclusively presumed to be unreasonable and therefore illegal without elaborate inquiry as to the precise harm they have caused or the business excuse for their use."17 This principle of per se unreasonableness avoids complicated and prolonged economic investigation into the market when the restraint is obviously harmful.18 The Supreme Court has limited per se analysis to categories of restraints for which no elaborate study is needed to establish that their nature and effect are anticompetitive. The classic examples of Section 1 per se violations are horizontal price fixing agreements,19 certain types of group boycotts, 20 bid-rigging, 21 market allocations between horizontal competitors22 and certain tying arrangements.23 Recently, the Supreme Court decided that "vertical" price fixing (resale price maintenance) should no longer be considered per se illegal and is now subject to the full rule of reason.24 For the restraint to be condemned as naked price fixing, bid-ridding or market allocations, a plaintiff need prove only that an agreement among truly independent horizontal competitors existed, the notion being that rarely is there a plausible pro-competitive justification for the restraint.25 By contrast, group boycotts and tying arrangements may have business justifications as well as pro-competitive effects. Group boycotts usually are condemned under the per se rule only when there is evidence of an agreement among direct competitors and the combined companies have a significant degree of market power.26 For tying arrangements, courts apply the per se rule only if the supplier has substantial market

17 N. Pac. Ry. Co. v. United States, 356 U.S. 1, 5 (1958). 18 See ibid. 19 Arizona v. Maricopa County Med. Soc'y, 457 U.S. 332, 348 (1982). 20 FTC v. Superior Court Trial Lawyers Ass'n, 493 U.S. 411, 452 n.9 (1990). 21 See Nat'l Soc'y of Prof'l Eng'rs v. United States, 435 U.S. 679, 695 (1978). 22 See United States v. Topco Assocs., Inc., 405 U.S. 596, 608-09 (1972). 23 See Int'l Salt Co. v. United States, 332 U.S. 392, 399 (1947). But see Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 14 (1984) (requiring market power to find tying). 24 See Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877 (2007). 25 See Arizona v. Maricopa County Med. Soc'y, 457 U.S. 332, 357 (1982); Nat'l Soc'y of Prof'l Eng'rs, 435 U.S. at 695; Behrend v. Comcast Corp., 532 F. Supp. 2d 735 (E.D. Pa. 2007). 26 See Nw. Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284 (1985).

6 EUI-RSCAS / Competition 2009 / Proceedings

power over the tying product and there is a substantial potential for impact on competition in the tied market – i.e., the notion that tying is a per se violation is a mischaracterization.

2. The "Quick Look"

When a challenged practice does not fall within one of the traditional per se categories, but still appears facially anticompetitive, a court may assess it under a "quick look" analysis rather than going through the full rule of reason framework.27 Quick look analysis essentially is reserved for horizontal arrangements that appear similar to traditional per se violations – when the great likelihood of anticompetitive effects can easily be ascertained – i.e., when an observer with even a "rudimentary understanding of economics could conclude that the arrangements in question would have an anticompetitive effect on customers and markets."28 In undertaking a quick look review, courts initially determine whether the challenged practice in theory and in all practical likelihood has anticompetitive effects,29 and, if so, the burden shifts to the defendant to come forward with plausible justifications for its conduct.30 Although the defendant need not produce detailed evidence at this stage, it must articulate the specific link between the challenged restraint and the purported justification to merit a more searching inquiry into whether the restraint may advance pro-competitive goals, even though it may facially appear to suppress competition. If that burden of coming forward is met, the case then flips into a rule of reason analysis and the significant evidentiary hurdles that go with it.

3. The "Full" Rule of Reason

In contrast to the truncated approach of the quick look standard, the full rule of reason places a heavy burden on antitrust plaintiffs. As the Sherman Act condemns only unreasonable restrictions on competition, "most antitrust claims are analyzed under [the full rule of reason], according to which the finder of fact must decide whether the questioned practice imposes an unreasonable restraint on competition, taking into account a variety of factors, including specific information about the relevant business, its condition before and after the restraint was imposed, and the restraint's history, nature and effect."31 The inquiry mandated by the rule of reason is whether the challenged agreement is one that promotes competition or one that suppresses competition: "[t]he true test of legality is whether the restraint imposed is such as merely regulates and perhaps thereby promotes competition or whether it is such as may suppress or even destroy competition."32

27 See Cal. Dental Ass'n v. FTC, 526 U.S. 756, 770 (1999). 28 Ibid. 29 See N. Tex. Specialty Physicians v. FTC, 528 F.3d 346, 361-62 (5th Cir. 2008) (a justification is plausible if it cannot be rejected without extensive factual inquiry). 30 See ibid. 31 State Oil Co. v. Kahn, 522 U.S. 3, 10 (1997). 32 Chi. Bd. of Trade v. United States, 246 U.S. 231, 238 (1918).

EUI-RSCAS / Competition 2009 / Proceedings 7

There is no precise formulation that all the U.S. courts use to apply the rule of reason, but the approaches are generally similar. Under the rule of reason, the plaintiffs have the initial burden to prove that an agreement has had or is likely to have an anticompetitive effect. After the plaintiffs have met this burden, the burden shifts to the defendant to produce evidence of pro-competitive effects of the agreement. If the defendants offer evidence of pro- competitive justifications or effects, plaintiffs then may attempt to demonstrate that the challenged conduct is not reasonably necessary to achieve the objective or that the anticompetitive effects outweigh the pro-competitive effects. Ultimately, "the factfinder weighs all of the circumstances of a case in deciding whether a restrictive practice should be prohibited as imposing an unreasonable restraint on competition."33 Finally, in the absence of direct evidence of anticompetitive effects (which is rare), most U.S. courts apply a "market power" threshold in assessing evidence under the rule of reason. The notion is that if the defendant (or defendants) lacks market power, the restraint at issue – even if found – is not likely to harm competition.

C. The Role of Economic Analysis in Antitrust Litigation

It has become commonplace for antitrust litigants to use economic evidence in support of one or more of their arguments, especially on matters relating to market definition, market power injury to competition and causation. With economic testimony playing such a prominent role in antitrust litigation, the importance of evaluating an expert's qualifications in providing evidence has increased dramatically. Judges evaluate the credibility of expert testimony according to Federal Rule of Evidence 702, which provides:

“If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.”34

Federal Rule of Evidence 702 is based primarily on the Supreme Court's landmark decision in Daubert v. Merrell Dow Pharmaceuticals, Inc.35 There, the Supreme Court increased judicial oversight over expert testimony by demanding that courts act as "gatekeepers" for scientific evidence.36 Specifically, the Daubert Court provided a list of factors for courts to consider when evaluating the legitimacy of expert testimony: (a) "whether the theory or technique can be (and has been) tested," (b) "whether the theory has

33 Cont'l T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 49 (1977). 34 FED. R. EVID. 702. 35 509 U.S. 579 (1993). 36 See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147-48 (1999). Kumho also ruled that Daubert extended to testimony that was not necessarily scientific. See ibid. at 147-49.

8 EUI-RSCAS / Competition 2009 / Proceedings

been subjected to peer review and publication," (c) "the known or potential rate of error, and the existence and maintenance of standards controlling the technique's operation," and (d) whether "technique is 'generally accepted' as reliable in the relevant [expert] community."37 As we will see later in case discussions, some clear patterns emerge when courts assess how expert testimony related to economic analysis should influence the course of antitrust litigation. For example, economic analysis is rarely debated in the Daubert context at the pleading stage. In the motion to dismiss context, the parties generally contest the validity of factual information alleged, with little economic analysis. To be sure, the parties may argue over whether the alleged conduct is "economically" rational, but this discussion is largely theoretical. Moreover, it is relatively rare to find expert economic testimony offered in the context of proving an agreement or conspiracy among parties; that element of an antitrust claim typically involves evidence showing specific communication or collusion among the defendant parties rather than a broader inquiry into market definition and dynamics. Far more common, particularly at the summary judgment stage, is a discussion of economic analysis when dealing with the effects of an agreement, alleged monopoly, or proposed merger. In the Section 1 context, expert testimony on the relevant product and geographic market often ignites Daubert challenges. For example, in Nilavar v. Mercy Health System-Western Ohio38 (discussed in Part II below) the defendants successfully defeated plaintiff's attack on defendant's exclusive dealing arrangements under Section 1, finding that plaintiff's expert failed to adequately frame an appropriate market definition. Ironically in that case, the court found that the expert's opinion satisfied Daubert, but nonetheless had failed to satisfy the basic Federal Rule of Evidence 702 requirements because the testimony's unreliable and arbitrary methodology produced a questionable market definition. Several other cases have rejected expert testimony related to market definition for similar reasons.39 In addition, expert economic testimony related to the anticompetitive impact of the challenged agreements have raised Daubert-related challenges. For instance, in In re Linerboard Antitrust Litigation,40 the defendants challenged the reliability of an expert's prediction model which demonstrated how the defendants' agreement to restrict product output would lead to higher prices for consumers. Essentially, the expert provided a model which explained that, "but for" the defendants' alleged agreement, prices would remain at competitive levels.41 The court in that case ultimately allowed the expert testimony over the defendants' objection that the evidence did not properly "fit" the specific facts of the case.42

37 Daubert, 509 U.S. at 593-95 (citation omitted). 38 244 F. App'x 690 (6th Cir. 2007). 39 See, e.g., Champagne Metals v. Ken-Mac Metals, Inc., 458 F.3d 1073 (10th Cir. 2006); Lantec, Inc. v. Novell, Inc., 306 F.3d 1003 (10th Cir. 2002). 40 497 F. Supp. 2d. 666 (E.D. Pa. 2007). 41 See ibid. at 671. 42 Ibid. at 684.

EUI-RSCAS / Competition 2009 / Proceedings 9

One potential pitfall for expert witnesses is to undertake analysis that makes assumptions about causation. Experts providing economic analysis have fallen victim to this defect in several cases, including in Craftsmen Limousine, Inc. v. Ford Motor Co.43 (also discussed below). There, plaintiff's expert provided economic analysis on defendant's alleged misconduct without considering whether separate economic factors might have caused plaintiff's harm. In light of its failure to provide a complete economic analysis, the court excluded the expert's testimony on damages despite the fact that the expert was otherwise well qualified.44 In another case, Concord Boat Corp. v. Brunswick Corp.,45 the Eighth Circuit reversed a $133 million verdict for the plaintiffs because it found that the plaintiff expert's economic model did not properly "incorporate all aspects of the economic reality" when concluding that the defendant's conduct caused the plaintiff's harm.46 As these cases demonstrate, expert assumptions regarding causation evidence can be a determinative factor in the outcome of antitrust litigation. Among all antitrust actions, however, the value of expert economic analysis is most prominent in the context of merger challenges under Section 7 of the Clayton Act. Indeed, expert testimony has become the basic focal point of merger litigation. The reason for this is self-evident: under the merger guidelines, defining the appropriate market for assessing the plaintiff's prima facie case has become the lynchpin of Section 7 actions. Likewise, in assessing competitive effects, the parties routinely hire economic experts to assess the likely increase in the ability of the remaining firms in the market to engage in anticompetitive coordination (coordinated effects theory) or likely increases in the market power of the merged firm ("unilateral" effect theory). For instance, in FTC v. Whole Foods Market, Inc.47 (discussed later in detail), the D.C. Circuit was forced to decide which expert model (the plaintiff's or defendant's) it would follow when defining the appropriate market. Each side proposed a different model: the FTC used a theory that emphasized average consumer reactions to price increase – which would create a smaller market, and thus higher market concentration, while the defendant's theory focused on marginal consumer reactions to the same price increase – which would increase the market size and thus reduce market concentration.48 The court ultimately sided with the FTC's model, which, by no coincidence, ultimately led to a verdict against the defendant.49 Thus, it is clear from merger litigation under Section 7 that the role of experts is vitally important to a case's outcome. In fact, over the past several decades, antitrust litigation in the U.S. has spawned a cottage industry of experts whose primary function is to provide economic analysis in a variety of antitrust litigations. Based on the success of that lofty industry, as well the legal

43 363 F.3d 761 (8th Cir. 2004). 44 See ibid. at 777. 45 207 F.3d 1039 (8th Cir. 2000). 46 Ibid. at 1057. 47 548 F.3d 1028 (D.C. Cir. 2008). 48 See ibid. at 1038. 49 See ibid. at 1040-41.

10 EUI-RSCAS / Competition 2009 / Proceedings

community's continued reliance on economic analysis in antitrust matters, one should expect this relatively new "industry" to continue to grow.

II. SECTION 1 SHERMAN ACT CASES

One way to assess evidence in Section 1 conspiracy cases is to consider whether there are any categorical distinctions that make it easier to determine the type and quality of evidence needed to take a case all the way to trial on the merits. A review of Section 1 decisions over the past several years suggests a simple divide: those cases in which horizontal competitors are not supposed to have agreements at all (which certainly leads to an evidentiary focus on the existence of such an agreement), and those Section 1 cases in which the existence of an agreement is not in dispute (e.g., joint ventures, certain trade association activity, vertical distribution arrangements, etc.) and the issue is whether there is sufficient proof that the subject "restraint" is unreasonable and harmful to consumers. As we will explain below, this simple dichotomy provides an easy way to understand both when and how courts go about separating plausible Section 1 cases from the frivolous.

A. Cases in Which the Primary Issue is Whether There is an "Agreement" at All

Historically, an overwhelming number of U.S. antitrust cases – spurred on by liberal class action rules50 and treble damage awards51 – have focused on plaintiffs' attempts to prove that an industry price increase or other type of "parallel" competitive activity (e.g., pricing, territorial restrictions, refusals to deal) was the result of an illegal agreement rather than a common response to other market forces. Moreover, because of the U.S. courts' "notice pleading" standards,52 plaintiffs historically were usually able to take these cases through discovery to summary judgment before a court seriously assessed the evidence – by which time many of these cases, however questionable, settled for large sums. This all changed with the Supreme Court's 2007 decision in Bell Atlantic Corp. v. Twombly,53 in which the Supreme Court addressed whether a Section 1 price fixing complaint could survive dismissal without specific allegations that, if true, would demonstrate the existence of a conspiracy. The Court held that even an allegation of deliberate parallel conduct with a bald assertion that the defendants were participants in a "conspiracy," without any allegations that, if later proved true, would establish the existence of a conspiracy was

50 See FED. R. CIV. P. 23. 51 See 15 U.S.C. § 15(a) ("[A]ny person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue therefor . . . and shall recover threefold the damages by him sustained, and the cost of suit, including a reasonable attorney's fee."). 52 FED. R. CIV. P. 8. 53 550 U.S. 544 (2007).

EUI-RSCAS / Competition 2009 / Proceedings 11

insufficient to state a claim under the Sherman Act.54 The Court did not specify exactly what additional facts must be pled, but indicated that an assertion of a conspiracy, even a plausible one, was not enough without additional detail in the form of "plus factors." These "plus factors" made the existence of an agreement more likely and include things such as simultaneous changes in price by multiple competitors made without an ordinary economic justification.55 In practical effect, Twombly took a long-standing summary judgment concept for dealing with ambiguous evidence and imposed it on antitrust plaintiffs at the pleading stage.56 The immediate and likely long-term impact of this new pleading standard in the U.S. cannot be overstated. The effect has been the dismissal of a very large number of cases at the 12(b)(6) motion to dismiss stage that otherwise would have gone forward at least through discovery and summary judgment. This is true, especially when there is no "direct" evidence of an illegal agreement and any circumstantial evidence is just as indicative of unilateral conduct as of an illegal agreement. The best way to understand these cases – and the critical evidence on the agreement issue – is to break down the cases into those that turn on direct evidence (both at the complaint and summary judgment stages) and those that require courts to assess circumstantial evidence of an illegal agreement. Given the plethora of cases in the U.S., we necessarily provide here only an overview of the case law and evidentiary issues with a few recent examples.

1. Direct Evidence Conspiracy Cases Will Proceed but Are Few and Far Between

Section 1 cases can certainly survive at the pleading stage, even in light of Twombly, where a court finds that there is direct evidence of an unlawful agreement. The problem for antitrust plaintiffs, however, is that as the U.S. courts generally take a narrow view of direct evidence and, absent criminal conduct, relatively few horizontal competitors are that ill-advised or foolish to engage in such conduct.

a. Direct Evidence of Agreement at the Pleading Stage.

Direct evidence is the simplest way to demonstrate an agreement,57 which is why plaintiffs often try to cloak inferential evidence as direct evidence.58 Courts, however, have held that "[d]irect evidence in a Section 1 conspiracy must be evidence that is explicit and requires no

54 Ibid. at 557-58. 55 Ibid. at 553. 56 See Matsushita Elec. Indus. Co. v. Zenith Radio Co., 475 U.S. 574, 588 (1986); supra notes 10-13. 57 See, e.g., Addyston Pipe & Steel Co. v. United States, 175 U.S. 211, 243 (1899) (holding that an express agreement can show concerted action). 58 See, e.g., InterVest, Inc. v. Bloomberg, L.P., 340 F.3d 144, 162-63 (3d Cir. 2003) (finding that the proffered direct evidence was unpersuasive and insufficiently clear, especially when the underlying activity is not facially anticompetitive).

12 EUI-RSCAS / Competition 2009 / Proceedings

inferences to establish the proposition or conclusion being asserted."59 Direct evidence such as a written contract60 or videotaped surveillance61 has been used to demonstrate concerted action. Alleging the existence of direct evidence in a complaint is generally sufficient to survive a motion to dismiss because it pushes the complaint past being merely a "bald assertion" of conspiracy. For example, Twombly clearly can be satisfied when plaintiffs provide "robust factual details" of a "predicate" illegal agreement such as naming the individuals who initiated discussions on the agreement, the object of the agreement, details of the agreement itself, and where and when the agreement was reached.62 As a general matter, however – with a few notorious exceptions – U.S. businesses are long past the time when principals sit down and hammer out price fixing arrangements.

b. Direct Evidence of Agreement at the Summary Judgment Stage.

Direct "evidence" at the pleading stage is one thing, showing enough actual evidence to survive summary judgment is quite another. The central evidentiary issue in direct evidence situations is whether the direct evidence is sufficient. Much like the pleading stage, at summary judgment direct evidence in a Section 1 conspiracy "'must be evidence that is explicit and requires no inferences to establish the proposition or conclusion being asserted. . . . [W]ith direct evidence the fact finder is not required to make inferences to establish facts.'"63 When direct evidence is shaky, courts look to the overall context of the claim to determine the sufficiency of the evidence.64 Moreover, additional circumstantial evidence is necessary to avoid summary judgment when the direct evidence is weak.65 Only where the alleged conspiracy is economically plausible, the actors are alleged to behave rationally, and the direct evidence would constitute concerted action, will the complaint likely survive summary judgment.66 Given this standard, courts recognize that only "rarely will there be direct evidence of an express agreement" in conspiracy cases.67 And canvassing cases over the last several years indeed confirms that direct evidence in civil conspiracy cases is relatively rare; this is the kind of evidence that typically results in criminal prosecutions followed by the predictable slew of class action follow-on cases.

59 In re Baby Food Antitrust Litig., 166 F.3d 112, 118 (3d Cir. 1999). 60 See, e.g., New York v. Anheuser-Busch, Inc., 811 F. Supp. 848, 869 (E.D.N.Y. 1993) (vertical agreement to allocate territories established by contract between parties). 61 See, e.g., United States v. Andreas, 216 F.3d 645 (7th Cir. 2000) (horizontal agreement to fix prices and reduce output established by videotaped surveillance of conspirators' meetings). 62 See In re Rail Freight Fuel Surcharge Antitrust Litig., 587 F. Supp. 2d 27, 34 (D.D.C. 2008). 63 Champagne Metals v. Ken-Mac Metals, Inc., 458 F.3d 1073, 1083 (10th Cir. 2006) (alterations in original) (citation omitted). 64 See ibid. at 1082-85. 65 See ibid. 66 See, e.g., In re Mercedez-Benz Antitrust Litig., No. 99-4311, 2006 WL 2129100, at *11-12 (D.N.J. July 26, 2006). 67 Local Union No. 189, Amalgamated Meat Cutters v. Jewel Tea Co., 381 U.S. 676, 720 (1965).

EUI-RSCAS / Competition 2009 / Proceedings 13

2. Circumstantial Evidence Cases of Horizontal Conspiracy

There are innumerable cases in which antitrust plaintiffs attempt to prove an illegal conspiracy based on circumstantial evidence, in part the result of the pre-Twombly notion that dismissal could be avoided by an allegation of facts that suggested an environment ripe for illegal agreements. On the heels of Twombly, however, a large number of complaints have been dismissed based on evidentiary allegations that are insufficient to suggest a conspiracy. And, of course, there continue to be a large number of cases that assess circumstantial evidence of horizontal agreements at the summary judgment stage.

a. Circumstantial Evidence of Agreement at the Pleading Stage.

There is now a growing stack of Section 1 court decisions dismissing conspiracy cases in which plaintiff can allege only parallel conduct (e.g., pricing, refusal to deal, etc.) and "plus factors" that no longer can provide the basis to survive a motion to dismiss.68 For example, it is now fairly settled, under Twombly, that any conduct that might be viewed as ordinary competitive activity – e.g., attending trade association meetings, mere "opportunities"69 to conspire arising from normal business practices, and even very clear "oligopoly" pricing70 – will not push the case across the line into likely conspiratorial behavior. This is not to say that in all U.S. courts, Twombly has been so readily embraced, and there is no doubt that circuit splits will develop and Twombly itself will need to be clarified. Indeed, several courts continue to read Twombly itself narrowly and continue to invoke the concept that, looked at as a whole,71 acts that are not themselves indicative of agreement can, in context, collectively imply a conspiracy. One notable recent example involved plaintiffs' bare allegations that certain private equity firms conspired to fix prices and bids as evidenced by their common involvement in the same transactions.72 This case represents the extreme side of the spectrum where the plaintiffs' burden is the easiest. Other courts avoid Twombly, but with little confidence. For example, in In re Chocolate Confectionary Antitrust Litigation,73 chocolate makers allegedly implemented three coordinated prices increases from 2002 to 2007, raising prices in nearly identical proportion to one another on each occasion. During this period, plaintiffs alleged that barriers to entry protected the chocolate market, defendants' raw material costs remained stable, and

68 Bell Atl. Corp. v. Twombly, 550 U.S. 544, 553 (2007). 69 In re Late Fee & Over-Limit Fee Litig., 528 F. Supp. 2d 953, 963 (N.D. Cal. 2007). 70 See, e.g., In re Graphics Processing Units Antitrust Litig., 527 F. Supp. 2d 1011 (N.D. Cal. 2007); In re Travel Agent Comm'n Antitrust Litig., No. 1561, 2007 WL 3171675 (N.D. Ohio Oct. 29, 2007); Shafer v. State Farm, Fire & Cas. Co., 507 F. Supp. 2d 587, 594-97 (E.D. La. 2007); In re Ins. Brokerage Antitrust Litig., No. 1663, 2007 WL 1100449, at *15-16 (D.N.J. Apr. 5, 2007). 71 See Cont'l Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690 (1962). 72 See Dahl v. Bain Capital Partners, LLC, 589 F. Supp. 2d 112 (D. Mass. 2008). 73 No. 1935, 2009 WL 931530 (M.D. Pa. Apr. 8, 2009).

14 EUI-RSCAS / Competition 2009 / Proceedings

consumer demand waned, providing the defendants with market power. Plaintiffs also alleged that defendants had integrated their American and Canadian operations through coordinated manufacturing and distribution systems, cross-border licensing agreements, and fusion of corporate oversight. The plaintiffs asserted that this market integration lent plausibility to the alleged pricing conspiracy in the U.S.74 The court denied defendants' motions under Twombly, but certified the case for interlocutory appeal, stating that the issue "is not whether either characteristics of a mature market or anticompetitive foreign conduct satisfy Twombly when placed alongside allegations of parallel conduct," but "whether the totality of the amended complaints raise a plausible inference of price fixing."75 Suffice it to say that the court's sentiment is typical of some judges still unwilling to pull the Twombly trigger.

b. Circumstantial Evidence of Agreement at the Summary Judgment Stage

It often is equally difficult for plaintiff to survive summary judgment on the conspiracy issue after discovery. A good example is the Seventh Circuit's decision in R.J. Reynolds Tobacco Co. v. Cigarettes Cheaper!.76 There, the court upheld a summary judgment ruling for Reynolds on a Sherman Act counterclaim alleging that it organized a retail price fixing conspiracy to drive the discounter, Cigarettes Cheaper!, out of business. Cigarettes Cheaper! operates a chain of retail outlets selling discount cigarettes. For years, Cigarettes Cheaper! sold Reynolds' brands that it had procured outside the U.S. and could sell more cheaply than other retailers. Cigarettes Cheaper! accused Reynolds of violating the Sherman Act by conspiring with retail dealers to drive Cigarettes Cheaper! out of business. The court explained that since Cigarettes Cheaper!'s best evidence would not permit a reasonable jury to find a retail conspiracy, it was not enough for plaintiff "to regale a jury with evidence that its retail rivals (and perhaps Reynolds too) intended to deal it a fatal blow."77 For example, it secured "tidbits" in discovery to support its theory.78 "Cigarettes Cheaper! sold Marlboros for less than many outlets sold Camels and Winstons; neither Reynolds nor its retail outlets were willing to take this lying down. They responded with discounts of their own and generated reams of paper expressing unhappiness about the need to do so."79 However, Cigarettes Cheaper! had not cited anything in the voluminous record that showed a horizontal retail conspiracy. Similarly, in Abraham v. Intermountain Health Care Inc.,80 summary judgment was awarded to Intermountain ("IHC"), Utah's largest managed care company, on Sherman Act Sections 1 and 2 claims stemming from its decision to use ophthalmologists to provide all eye

74 See ibid. at *1. 75 Ibid. at *2. 76 462 F.3d 690 (7th Cir. 2006). 77 Ibid. at 696. 78 Ibid. 79 Ibid. 80 461 F.3d 1249 (10th Cir. 2006).

EUI-RSCAS / Competition 2009 / Proceedings 15

care services to certain members and, thereby, to exclude independent optometrists. The plaintiffs were forty-nine optometrists and their affiliated professional organizations; and an eye clinic that employs optometrists. IHC served an estimated 60% of managed care enrollees in this area. Although IHC employed some physicians directly, it provided most health care only through its managed care subsidiaries. In 1995, there were indications that IHC would add optometrists to the panel because they typically charged about 20% less for NSEC than do ophthalmologists. According to IHC's director of provider relations, each time IHC considered adding optometrists, "the ophthalmologists 'write all kinds of letters and [make] phone calls and raise such a stink' that IHC decide[d] not to do it."81 The court held that there was insufficient evidence of an illegal agreement between IHC and its panel ophthalmologists to survive summary judgment. The applicable standard, the court explained, requires evidence tending to "exclude the possibility that the alleged conspirators acted independently."82 The court found that while there was evidence that IHC acted in response to ophthalmologists' complaints, it is well-established that such evidence is insufficient as a matter of law to establish concerted action.83 The plaintiffs' collusion evidence included the fact that they are a lower-cost alternative to ophthalmologists and that IHC allegedly had a motive to conspire in that it hoped to benefit from increased utilization of its hospitals and surgical facilities. Mindful that "'antitrust law limits the range of permissible inferences from ambiguous evidence in a § 1 case,'"84 the court held that "the Plaintiffs have not presented enough evidence from which a jury may infer an antitrust conspiracy."85 The court observed that:

IHC prefers panel health care providers who have staff privileges at IHC hospitals. Ophthalmologists, by law, can perform more procedures than optometrists. Therefore, even if IHC were to include optometrists on its provider panels, it would still need to panel ophthalmologists. This fact, coupled with IHC's pro-competitive justification for limiting the number of paneled health care providers – that is, limiting the number of providers performing a given service increases the volume of patients each provider sees, which, in turn, enables IHC to negotiate lower reimbursement rates to the panel providers – suggests that IHC may have acted independently in deciding not to panel optometrists . . . .86

81 Ibid. at 1255 (first alteration in original). 82 Ibid. at 1257 (quoting Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 588 (1986)) (internal quotation marks omitted) (citations omitted). 83 See ibid. at 1259 84 Ibid. at 1257 (citation omitted). 85 Ibid. at 1261 86 Ibid. at 1263.

16 EUI-RSCAS / Competition 2009 / Proceedings

B. Cases in Which the Primary Issue Is the Competitive Effect of the Agreement at Issue

The other major type of Section 1 case is one in which there is little dispute about the existence of an agreement, but instead concerns a claim that the agreement unreasonably restrains trade, harms competition and, as a result, injures the plaintiff. The evidentiary battles in these cases are best understood by dividing the cases into horizontal and vertical agreements. And, again, U.S. courts at least purport to make distinctions between direct and circumstantial evidence of effects – although direct effects evidence appears to be more of a myth than reality. Instead, in all of these rule of reason (or, at best, "quick look") cases, market definition and market power tend to be the key threshold issues.

1. Horizontal Collaborations

a. The "Reasonableness" Quagmire for Assessing the Activities of Legitimate "Horizontal" Ventures and Associations

Where Section 1 is presumed or found to apply to the alleged restraints of facially legitimate ventures or other "competition" collaborations, 87 the evidentiary battles are, at best, understood to be in a serious state of flux. An initial question in these cases – with serious evidentiary implications – is whether the "quick look" will or will not apply. As described above, if quick look is the appropriate framework, plaintiff's burden of proof is much reduced, focusing on whether the anticompetitive effects are fairly obvious and the potential pro- competition justifications equally implausible. The per se and quick look cases, however, have not been kind to antitrust plaintiffs in the venture or collaboration context. For example, in Stop & Shop Supermarket Co. v. Blue Cross & Blue Shield of Rhode Island,88 even after a full trial, a plaintiff was still unable to avoid rule of reason analysis by cabining its claims into one of the standard per se categories. Affirming a directed verdict for defendants, the First Circuit held that if the plaintiffs fail to "explain in detail . . . just what the arrangements were and why they plausibly constituted antitrust violations," then their intimations of illegal horizontal agreements will still not receive per se analysis.89 Attempts to apply per se a quick look analysis to integrated joint ventures, such as sports leagues, are also now routinely rejected.90

87 This paper does not address the interesting debate concerning the circumstances under which distinct companies may be viewed as a single economic enterprise and therefore incapable of conspiring under Section 1. See Copperweld v. Independence Tube Corp., 467 U.S. 752 (1984); Am. Needle Inc. v. NFL, 538 F.3d 736 (7th Cir.), petition for cert. filed, 77 U.S.L.W. 3326 (U.S. Nov. 17, 2008) (No. 08-661); City of Mt. Pleasant v. Associated Elec. Coop., Inc., 838 F.2d 268, 278 (8th Cir. 1988). 88 373 F.3d 57 (1st Cir. 2004). 89 Ibid. at 65. 90 See, e.g., Major League Baseball Props., Inc. v. Salvino, Inc., 542 F.3d 290 (2d Cir. 2008). See generally Texaco Inc. v. Dagher, 547 U.S. 1 (2006).

EUI-RSCAS / Competition 2009 / Proceedings 17

Indeed, in some limited circumstances, courts have found that a purported "restraint" can be rejected in defendants' favor on a "quick look." For example, in Wallace v. IBM,91 plaintiff was a competitor of the Linux operating system, which was under the GNP General Public License (GPL), and thus, like any derivatives or improvements to the Linux system had to be available free of charge. Wallace alleged that the defendants conspired among themselves and with others, including the Free Software Foundation, to eliminate competition in the operating system market by making Linux available at the unbeatable price of zero, thus undermining his ability to compete with his own operating system or a derivative work. The court found that the GPL did not even restrain trade; it is "a cooperative agreement that facilitates production of new derivative works, and agreements that yield new products that would not arise through unilateral action are lawful."92 The court explained that GPL cannot be understood as price fixing because the GPL sets a maximum price of zero; similarly, plaintiff "does not contend that Linux has such a large market share, or poses such a threat to consumers' welfare in the long run, that evaluation under the rule of reason could lead to condemnation."93 Accordingly, the court found that the claim was rejected after a "quick look."94 Where the full rule of reason standard has been applied to ventures and collaborations, the evidentiary hurdles for most plaintiffs becomes quite difficult. While Twombly motions are relatively rare on issues such as market definition, market power or antitrust injury, they certainly are not unheard of, and they are growing in number.95 For example, in Eastern Food Services v. Pontifical Catholic University Services Ass'n, Inc.,96 the court inferred from the face of the complaint that the defendant did not have market power because the alleged market was the supply of vending machines at one college campus in a large city. More typically, many of these cases cannot make it past the summary judgment stage on these standard Section 1 elements – which, of course, is why plaintiffs usually try the "quick look" approach in the first place. For example, in Craftsman Limousine, Inc. v. Ford Motor Co.97 Craftsman claimed that it was the victim of a conspiracy to influence the advertising and trade show policies of the major publications serving the limousine industry. Craftsman's vehicles did not conform to industry standards set by Ford and General Motors in that they were longer than the 120 inches allowed by Ford and the 130 inches permitted by GM. The conspiracy allegedly began in the early 1990s and prevented Craftsman from promoting its products through the usual sources that restricted promotion to conforming products, harming Craftsman's business.

91 467 F.3d 1104 (7th Cir. 2006). 92 Ibid. at 1107; see also Broad. Music, Inc. v. CBS, Inc., 441 U.S. 1 (1979); Polk Bros., Inc. v. Forest City Enter., Inc., 776 F.2d 185 (7th Cir. 1985); cf. Dagher, 547 U.S. 1. 93 Wallace, 467 F.3d at 1108. 94 Ibid. 95 See James Keyte, “Twombly: How Courts Are Interpreting and Extending Its Principles”, Antitrust (Fall 2008), at pp. 65-71. 96 357 F.3d 1, 7 (1st Cir. 2004). 97 491 F.3d 380 (8th Cir.), cert. denied, 128 S. Ct. 654 (2007).

18 EUI-RSCAS / Competition 2009 / Proceedings

The Eighth Circuit held that the trial court erroneously applied per se scrutiny since the defendants' motive for their actions may have been safety. It also upheld the exclusion of Craftsman's economic expert who analyzed the effect of the restraint on Craftsman, but not on competition in the industry generally as is required under the rule of reason. The court then found that Craftsman did not meet its burden under the rule of reason of proving a relevant market that it attempted to define as specialty limousines. This, the court explained, would require a finding that a Ford Town Car stretched by 120 inches is in a different market than a Ford Town Car stretched by 121 inches, and no reasonable juror could so find. In addition, the court found that Craftsman did not meet its burden of showing the unreasonableness of the alleged restraint. Craftsman did not either offer direct proof that the challenged conduct had actual adverse effects on competition or meet its burden under the alternative method, which allows proof of adverse effects indirectly through an inquiry into market power and market structure. The court concluded that the defendants did not have the market power to affect the limousine market.98 Further, as the court observed in Eastern Food Services, in a full rule of reason analysis, courts typically undertake a threshold inquiry into market power to determine whether the restraint will likely have a substantial and adverse effect on market competition. In fact, since plaintiffs face such a heavy evidentiary burden when trying to prove direct anticompetitive harm, they almost always have to overcome the market power hurdle in order to demonstrate that their particular claim poses a serious threat to competition.99 Generally, this inquiry mirrors that used for Section 2. The court will assess whether plaintiff has established a proper relevant market, and then assess whether plaintiff's evidence demonstrates that defendants' market share will provide them "the ability to raise prices above those that would be charged in a competitive market."100 If a plaintiff can demonstrate that the defendants' agreement will create or enhance market power, he must also provide, at a minimum, evidence of a proximate connection between that market power, the alleged restraint101 and harm to the overall market; that is, the market power and restraint analysis must provide evidence of harm to competition, not simply to competitors.102

2. Vertical Agreements

Vertical agreements – both price and non-price related – have a long history in U.S. law with fairly established evidentiary patterns.103 Again, absent inexperienced counsel or woefully bad facts, these cases tend to survive Twombly-based challenges with a few exceptions. The

98 Ibid. at 393. 99 See, e.g., Grappone, Inc. v. Subaru of New England, Inc., 858 F.2d 792, 797 (1st Cir. 1988) (Breyer, J.). 100 NCAA v. Bd. of Regents of the Univ. of Okla., 468 U.S. 85, 109 n.38 (1984). 101 See, e.g., Gen. Leaseways, Inc. v. Nat'l Truck Leasing Ass'n, 744 F.2d 588, 596 (7th Cir. 1984). 102 See, e.g., K.M.B Warehouse Distribs., Inc. v. Walker Mfg. Co., 61 F.3d 123, 129-30 (2d Cir. 1995); see also, e.g., Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 458 (1993); Atl. Richfield Co. v. USA Petrolium Co., 495 U.S. 328, 338 (1990). 103 Leegin Creative Leather Prods., Inc. v. PSKS Inc., 551 U.S. 877 (2007); Cont'l T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977).

EUI-RSCAS / Competition 2009 / Proceedings 19

evidentiary question typically becomes whether these vertical arrangements truly harm competition, which requires plaintiffs to prove a relevant antitrust market and that the vertical restraint at issue in fact harms competition in that market. As described below, for many years now this has been quite difficult for antitrust plaintiffs to establish.

a. Vertical Price Fixing Is Now a Rule of Reason Inquiry in the U.S.

As an initial matter, it needs to be highlighted that the U.S. Supreme Court recently held that vertical price fixing would no longer be subject to per se condemnation and would now be scrutinized by courts under the full rule of reason. The Court's decision in Leegin has generated considerable reaction in Congress – and the FTC – while the antitrust plaintiffs' bar lost a major weapon. These cases are now being subjected to the same evidentiary standards of non-price vertical restraints, and few cases can withstand that scrutiny whether for lack of a clearly defined market or proof of harm to competition. This is an example where the current make- up of the Supreme Court has significantly redefined what is "wheat" and what is "chaff" in U.S. antitrust jurisprudence.

b. The Continued Evidentiary Battles Involving Non-Price Vertical Restraints.

By contrast, the evidentiary fights involving non-price vertical restraints are well defined and clearly biased toward defendants. Indeed, much like with the Twombly and Matsushita horizontal cases, this is an area where "ambiguous" evidence of anticompetitive effect is common, pro-competitive restrictions are plausible and fairly easy to document, and U.S. courts tend to dispose of these cases fairly early and routinely. The primary evidentiary focus of these cases is whether the restraint harms competition in a relevant market. To be sure, there are a small number of cases that can survive summary judgment under this lens, but it takes facts that are fairly unusual because there is a serious level of "market power" in the hands of the firm or firms allegedly imposing the restraint. The overwhelming majority of Section 1 vertical cases, however, just have not met these evidentiary burdens and typically have been dismissed at the summary judgment stage. A typical example is PepsiCo, Inc. v. Coca-Cola Co.,104 in which the court addressed one of the many battles between Coke and Pepsi in the "vertical" context. Coca-Cola and PepsiCo compete to sell fountain syrup to customers with soda fountains on their premises. Coca-Cola was first to use independent food service distributors ("IFDs") to distribute fountain syrup. When PepsiCo attempted to use IFDs as well, Coca-Cola enforced loyalty provisions in its agreements with IFDs to exclude PepsiCo.105 PepsiCo alleged that the

104 315 F.3d 101 (2d Cir. 2002). 105 Ibid. at 104.

20 EUI-RSCAS / Competition 2009 / Proceedings

inclusion and enforcement of the loyalty provisions constituted monopolization and attempted monopolization in violation of Section 2 of the Sherman Act and concerted action in restraint of trade in violation of Section 1.106 After Coca-Cola's motion to dismiss was denied, the company was subjected to eighteen months of discovery. PepsiCo's complaint defined the relevant market as the "market for fountain-dispensed soft drinks distributed through [IFDs] throughout the United States" and then sought to narrow this market definition on summary judgment by confining it to customers with certain characteristics, specifically "large restaurant chain accounts that are not 'heavily franchised' with low fountain 'volume per outlet.'"107 The Second Circuit affirmed a grant of summary judgment for Coca-Cola on the basis that PepsiCo's market definition was not substantiated by the evidence, nor was it legally sufficient.108 Although many customers had a preference for using IFDs, this preference alone failed to create a distinct group but should be included in the whole fountain syrup purchasing market.109 In fact, PepsiCo's own definition excluded franchisees who purchased over half of the fountain syrup Coca-Cola sold through IFDs.110 Consequently, the market definition was improper because it arbitrarily selected one non-determinative factor that was used in purchasing decisions and lacked evidentiary support for the narrower definition.111 PepsiCo is typical of numerous cases in which a plaintiff can claim an antitrust harm only by arguing for artificially narrow markets while in fact the defendant has no real market power at all. Plaintiffs have similarly failed in the vertical agreement context based on what courts have construed as suspect geographic market definitions for the alleged restraint of trade. For instance, in Nilavar v. Mercy Health System-Western Ohio,112 the Sixth Circuit dismissed the plaintiff's expert testimony regarding potential anticompetitive effects of an exclusive medical services agreement with radiologists and hospitals, in large part, on the expert's attempt to artificially narrow the geographic region.113 When considering the proposed market definition, the court of appeals determined that the expert artificially excluded various geographic regions based on inexact and arbitrary methods (such as relying on ZIP codes rather than county borders).114 In doing so, the court deemed the expert's model unreliable when attempting to prove considerable damage to competition from the defendant's employment of exclusive contracts. Additionally, the court found that the expert's proposed geographic region in some cases conflicted with its own product definition, a dilemma which plaintiffs often encounter when attempting to carve narrow market definitions in both product

106 Ibid. 107 Ibid. at 105 (internal quotation marks omitted) (citation omitted). 108 Ibid. at 106-7. 109 Ibid. 110 Ibid. 111 Ibid. 112 244 F. App'x 690 (6th Cir. 2007). 113 Ibid. at 697-99 114 Ibid. at 698.

EUI-RSCAS / Competition 2009 / Proceedings 21

and geographic categories.115 Through this lens, Nilavar provides an appropriate example of how difficult the hurdles may be for plaintiffs when asserting non-pricing vertical agreement actions under Section 1.

III. SECTION 2 MONOPOLIZATION CASES

Surveying the evidentiary battleground for Section 2 monopolization (or attempt) claims, we see a great deal of overlap with Section 1 vertical cases in terms of fights over market definition, harm to competition and the like. But there is more to be shown, of course, under Section 2: an antitrust plaintiff must allege and prove that the defendant has sustainable monopoly power116 and that it achieved or sustained that power through means that may be deemed improper under Section 2. Recent Section 2 cases delve into unique and complex issues of "evidence."

A. Proof of Monopoly Power

It is easy to allege that a defendant has a "monopoly"; it is another thing altogether to prove it. Indeed, in the dynamic, rapidly changing marketplaces of today, the notion of a firm having sustained monopoly power may be fairly rare with some notable, well-publicized exceptions. Looking at the cases in the Section 2 area, therefore, we see a great number of decisions in which plaintiff simply cannot provide the evidence – again much of it economic – that the defendant possesses monopoly power. Very few cases survive the summary judgment stage on this issue. As a threshold matter, courts often presume that significant market share is dispositive evidence that market or monopoly power exists.117 Indeed, market share percentage can be a prime indicator of whether a court will conclude that a defendant possesses monopoly power under Section 2.118 In addition, courts will also look to evidence related to the particular market's barriers to entry.119 As a general matter, plaintiffs can establish a prima facie case of monopoly power if they can show that a firm possesses market share in excess of seventy

115 Ibid. 116 See United States v. E.I. duPont de Nemours & Co., 351 U.S. 377, 391 (1956) (defining monopoly power as "the power to control prices or exclude competition"). 117 See United States v. Aluminum Co. of Am., 148 F.2d 416, 424 (2d Cir. 1945) (Hand, J.) (discussing market share percentage as a vital barometer in determining a firm's monopoly status). 118 See, e.g., Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451, 481 (1992) (noting that one can infer monopoly power at eighty percent monopoly share); United States v. Grinnell Corp., 384 U.S. 563, 571 (1966) (ruling that the "existence of such [monopoly] power ordinarily may be inferred from the predominant [eighty-seven percent] share of the market"); E.I. duPont de Nemours & Co., 351 U.S. at 379, 391 (observing that a seventy-five percent market share would constitute monopoly power). 119 See, e.g., United States v. Microsoft Corp., 253 F.3d 34, 51 (D.C. Cir. 2001) (stating that "monopoly power may be inferred from a firm's possession of a dominant share of a relevant market that is protected by entry barriers").

22 EUI-RSCAS / Competition 2009 / Proceedings

percent in a market characterized by substantial entry barriers.120 In contrast, courts generally will not find monopoly power when a firm's market share is less than fifty percent.121 A good example of a heated dispute that turned on a lack of monopoly power (here, monopsony) power is Campfield v. State Farm Mutual Automobile Insurance Co.,122 in which an auto glass repair shop owner asserted that State Farm Mutual Automobile Insurance monopolized the auto glass repair market. According to plaintiffs, State Farm had stopped handling glass-only claims by insureds through local personnel and instead developed a nationwide program for handling such claims. One aspect of this change was the State Farm Offer and Acceptance Program (O&A Program), through which State Farm contracted with glass shops to provide services in accordance with its policy and pursuant to a fixed reimbursement schedule. State Farm outsourced management of the O&A Program to Lynx Services, a company that specialized in insurance claim processing. Lynx followed a script developed by State Farm to determine the type of windshield damage and process the claim. The plaintiff owned an automobile glass repair shop and was the inventor of the patented Ultra Bond repair technique which can reliably repair cracks up to 18 inches long, making the crack nearly invisible. He gave a demonstration of the Ultra Bond Process but State Farm was unimpressed. Campfield and Ultra Bond filed a complaint against State Farm and Lynx, alleging both Section 1 and 2 violations. The court explained that Campfield's allegation described a monopsony, not a monopoly. When considering market power in a monopsony situation, the court observed, "'the market is not the market of competing sellers but of competing buyers. This market is comprised of buyers who are seen by sellers as being reasonably good substitutes.'"123 Thus, according to the court, if the market described in the complaint failed to include "reasonably good substitutes," it would follow that the plaintiff had not adequately alleged a relevant market. Such was the situation in this case. Campfield asserted that the market consists of auto-glass shops that compete to provide windshield repair or replacement services to individuals who are insured by State Farm within the United States. The court considered this market definition to be under-inclusive. Campfield's market description "failed to include competitive substitutes; there are other consumers to which auto glass repair shops may cater their services."124 Again, this lack of evidence of monopoly power is repeated time and time again in cases granting summary judgment on Section 2 claims.125

120 See ibid.; see also Tops Mkts., Inc. v. Quality Mkts., Inc., 142 F.3d 90, 98-99 (2d Cir. 1998). 121 See, e.g., Bailey v. Allgas, Inc., 284 F.3d 1237, 1250 (11th Cir. 2002); AD/SAT v. Associated Press, 181 F.3d 216, 229 (2d Cir. 1999). 122 532 F.3d 1111 (10th Cir. 2008). 123 Ibid. at 1118 (quoting Todd v. Exxon Corp., 275 F.3d 191, 202 (2d Cir. 2001)). 124 Ibid. (citing Tampa Elec. Co. v. Nashville Coal Co., 365 U.S. 320, 331-32 (1961)). 125 See, e.g., United States v. Grinnell Corp., 384 U.S. 563 (1966); United States v. E.I. duPont de Nemours & Co., 351 U.S. 377 (1956).

EUI-RSCAS / Competition 2009 / Proceedings 23

B. Proof of the Requisite Misconduct

Even when an antitrust plaintiff – including the government – can surmount the evidentiary hurdles of proving a monopoly, it still must prove that the monopolist has engaged in conduct that violates the Sherman Act – often referred to as "exclusionary" conduct126 or conduct that is not "competition on the merits."127 But before reviewing recent cases assessing evidence on the misconduct element, we must note a trend in the case law – for the most part opposed by the current administration's antitrust enforcers128 – that is making it harder and harder to find evidence that violates Section 2. Much like the U.S. Supreme Court's recent rejection of the per se treatment of resale price maintenance, the Court in recent years has completely eradicated or seriously cut back on several traditional Section 2 theories. For example, over the past decade, the Court has essentially rejected Section 2 theories of leveraging,129 "essential facilities," 130 "predatory bidding"131 and, most recently price squeezes (except in rare circumstances).132 In short, the most interesting developments in Section 2 law have nothing to do with the treatment of evidence and everything to do with the scaling back of the substantive reach of Section 2 itself. Indeed, even where the scope of Section 2 is not itself challenged, courts are increasingly finding ways to neutralize its application. The recent decision in Rambus Inc. v. FTC133 is a good example. There, the D.C. Circuit investigated the critical question of whether Rambus engaged in exclusionary conduct by secretly securing patents on technology that it knew would become the industry standard due to its participation in a standard-setting organization. In rejecting the FTC's position, the court relied on two principles: first, that the exclusionary conduct must have an anticompetitive effect and second that the antitrust plaintiff bears the burden of proving the anticompetitive effect.134 In making its determination the court noted that deception alone is not necessarily anticompetitive, even if its effect is to raise price.135 In order for deception to be actionable, it must impair rivals' ability in a manner tending to bring about or protect a defendant's monopoly power.136 The court noted that, at worst, the plaintiff only lost out on an opportunity to profit due to Rambus' actions, but this loss was not a harm to competition because the only harm was to one competitor while the

126 Verizon Commc'ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398 (2004); Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985). 127 Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 222-23 (1993). 128 Up until 2009, the Department of Justice had abided by a report posted on its website that interpreted Section 2 actions favorably for defendants. The current administration promptly withdrew the Section 2 report when it took office. See Press Release, U.S. Dep't of Justice, Justice Department Withdraws Report on Antitrust Monopoly (May 11, 2009), http://www.udsoj.gov/atr/public/press_releases/2009/245710.htm. 129 Trinko, 540 U.S. 398. 130 Ibid. 131 Weyerhaeuser Co. v. Ross-Simmons Hardwood Lumber Co., 549 U.S. 312 (2007). 132 Pac. Bell Tel. Co. v. Linkline Commc'ns, Inc. 129 S. Ct. 1109 (2009). 133 522 F.3d 456 (D.C. Cir. 2008), cert. denied, 129 S. Ct. 1318 (2009). 134 Ibid. at 462-64. 135 Ibid. at 464-65. 136 Ibid.

24 EUI-RSCAS / Competition 2009 / Proceedings

marketplace could benefit overall from the creation of alternative technologies. 137 Additionally, the FTC had contended that, because "the ability to profitably restrict output and set supracompetitive prices is the sine qua non of monopoly power, any conduct that permits a monopolist to avoid constraints on the exercise of that power must be anticompetitive."138 The court held, however, that a lawful monopolist's circumvention of price constraints, even if they are deceptive, does not necessarily present harm to competition.139 Accordingly, the court vacated the FTC's finding of anticompetitive and monopolistic behavior and remanded for further proceedings.140 Courts, however, have not completely shut the door on actions similar to those presented in Rambus. In Broadcom Corp. v. Qualcomm, Inc.,141 the Third Circuit reversed a district court's dismissal of a Section 2 claim, under facts almost identical to Rambus, in which the plaintiff essentially alleged that the "patent holder's intentionally false promise to license essential proprietary technology" to a standard-setting organization constituted anticompetitive conduct under Section 2.142 In its opinion, the court noted that the defendant's conduct "increase[ed] the likelihood that patent rights will confer monopoly power to the patent holder" and thus satisfied Section 2.143 Although Qualcomm does provide a glimmer of hope for plaintiffs, its potential impact should not be overstated. For one, Qualcomm placed heavy emphasis on the FTC's initial finding in the Rambus case; one which the D.C. Circuit promptly reversed one year after the Third Circuit decided Qualcomm.144 Moreover, in an attempt to curtail Qualcomm's breadth, the D.C. Circuit reasoned in Rambus that Qualcomm should to be limited to circumstances in which the plaintiff can show a standard-setting organization's actual reliance on the defendant's deceptive language.145 To be sure, Rambus carries no substantive weight in its attempt to narrow its sister circuit court's decision in Qualcomm. Yet, in light of Qualcomm's own reliance on the initial FTC action in Rambus, one might suspect that a plaintiff's window of opportunity remains exceedingly narrow.

IV. ANTITRUST INJURY AND CAUSATION REMAIN A MAJOR HURDLE

Whether based on Section 1 or Section 2 of the Sherman Act, all private antitrust plaintiffs must be able to allege and prove that its injury is of the "type" that the antitrust laws were intended to prevent and that the injury (and damages) flow from that which makes

137 Ibid. at 465. 138 Ibid. at 466. 139 Ibid. 140 Ibid. at 459. 141 501 F.3d 297 (3d Cir. 2007). 142 Ibid. at 314. 143 Ibid. 144 Rambus v. FTC, 522 F.3d 456, 466 (D.C. Cir. 2008), cert. denied, 129 S. Ct. 1318 (2009). 145 Ibid.

EUI-RSCAS / Competition 2009 / Proceedings 25

defendant's conduct unlawful under the antitrust laws.146 These are significant limiting principles in U.S. antitrust law intended to ensure that private antitrust actions – especially in the context of treble damages – are in fact redressing harms (or prospective injuries in the context of injunction actions) to overall competition in a market and not just harm to individual competitors or injuries caused by other factors.

A. Injury to Competition

Harm to competition – overall competition, that is, and not just harm to competitors – must be identified in every antitrust suit. The alleged harm to competition must be not only possible but likely and significant, and this is ordinarily ascertained by reference to the market circumstances. Generally, harm to competition takes the form of higher prices, reduced output, or a deterioration in quality,147 but it can also be found if market participants are denied access to services.148 While higher prices are relatively easy to identify, there are substantial debates over output measurements. First, the proper definition of output is often disputed. The Supreme Court split over the definition of output when a professional association had conspired to suppress price and quality advertising in relation to dental services. 149 The majority categorized output as dental services whereas the minority categorized output as the sum total of the goods, services, and information that a dentist provides.150 The disagreement over categorization heavily influenced the outcome of the case, as the majority held that, within their limited definition of output, the plaintiff's practices were neutral towards overall competition.151 Second, even when the Court agrees upon a definition of output, it still may have difficulty ascertaining whether output has increased or decreased. Because output is relatively difficult to measure, the test is whether the challenged practice would "tend to restrict competition and decrease output."152 Market power is also helpful as a shortcut to determine whether the challenged practice could plausibly cause an output reduction. The reverse is not necessarily true, as the Court has held that market analysis is not required where there is actual proof of detrimental effects such as output reduction.153

146 See generally Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477 (1977); see also 15 U.S.C. § 15(a). 147 Babyage.com, Inc. v. Toys "R" Us, Inc., 558 F. Supp. 2d 575, 583 (E.D. Pa. 2008). 148 FTC v. Ind. Fed'n of Dentists, 476 U.S. 447, 459 (1986) ("A refusal to compete with respect to the package of services offered to customers, no less than a refusal to compete with respect to the price term of an agreement, impairs the ability of the market to advance social welfare by ensuring the provision of desired goods and services to consumers at a price approximating the marginal cost of providing them . . . . [S]uch an agreement limit[s] consumer choice by impeding the 'ordinary give and take of the market place.'" (citation omitted)). 149 Cal. Dental Ass'n v. FTC, 526 U.S. 756, 776-77 (1999). 150 Ibid. at 773-74. 151 Ibid. at 776-77. 152 Broad. Music, Inc. v. CBS, Inc., 441 U.S. 1, 20 (1979). 153 Ind. Fed'n of Dentists, 476 U.S. at 457.

26 EUI-RSCAS / Competition 2009 / Proceedings

There is a constant flow of cases that turn on the antitrust injury element, some that survive it,154 many that do not.155 This is one subject area that will continue to be attractive for courts to cite when ridding themselves of dubious antitrust cases.

B. "By Reason Of" the Violation and That Which Makes the Conduct Unlawful

Even where antitrust plaintiffs are able to provide an evidentiary basis for proving an injury to competition, they still must be able to prove that their particular injury flows from the misconduct itself (injury in fact) and from the reduction in competition that violates the antitrust laws (sometimes referred to as "causal antitrust injury"). Moreover, in contrast to the calculation of damages – where U.S. courts allow for some degree of flexibility – causation is viewed as an element of liability, and courts therefore require plaintiffs to prove the causal connection with "reasonable certainty."156 Again, this causation requirement has proven to be a difficult hurdle for antitrust plaintiffs in many different contexts because a lack of "causation in fact is fatal to the merits of any antitrust claim."157 At the motion to dismiss stage an antitrust plaintiff generally need allege only plausible causation whereas in later stages an antitrust plaintiff must prove with reasonable certainty that the defendant's antitrust violation caused the plaintiff's harm.158 For example, "[a]t summary judgment, plaintiff only needs to set forth sufficient facts to allow a reasonable fact finder the inference that defendants were a substantial factor in causing injury."159 In Rome Ambulatory Surgical Center, the defendants presented a veritable laundry list of factors that could have caused plaintiff's antitrust injury. 160 The court held, however, that "[d]espite the evidence of alternative factors and alternative causes for those factors, plaintiff has set forth facts sufficient to support a reasonable inference that defendants were a material cause of injury."161 Thus, it appears clear that where plaintiffs can present a plausible argument that defendant's actions were in some part responsible for the injury, they are not required to eliminate every alternative explanation.162

154 Lucas v. Citizens Commc'ns Co., 409 F. Supp. 2d 1206 (D. Haw. 2005) (finding for defendants upon summary judgment because plaintiff's description of the harm to competition was implausible - the challenged practice likely benefited competition), aff'd, 244 F. App'x 774 (9th Cir. 2007); Rome Ambulatory Surgical Ctr., LLC v. Rome Mem'l Hosp., 349 F. Supp. 2d 389, 405 (N.D.N.Y. 2004) (denying defendant's motion for summary judgment because plaintiff properly alleged harm to competition by stating that defendant foreclosed competition through improper exclusive dealing). 155 Babyage.com, Inc. v. Toys "R" Us, Inc., 558 F. Supp. 2d 575, 580 (E.D. Pa. 2008) (dismissing complaint in part because there was no harm to competition). 156 See O.K. Sand & Gravel v. Martin Marietta Techs., Inc., 36 F.3d 565, 573 (7th Cir. 1994); Nat'l Indep. Theatre Exhibitors, Inc. v. Buena Vista Distrib. Co., 748 F.2d 602, 607 (11th Cir. 1984). 157 Argus Inc. v. Eastman Kodak Co., 801 F.2d 38, 41 (2d Cir. 1986). 158 Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 562 (1931). 159 Rome Ambulatory Surgical Ctr., 349 F. Supp. 2d at 403. 160 Ibid. at 397. 161 Ibid. at 404. 162 Ibid. at 403 (citing Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 114 (1969)); see, e.g., Churchill Downs Inc. v. Thoroughbred Horsemen's Group, 605 F. Supp. 2d 870, 880 (W.D. Ky. 2009) (holding (cont'd)

EUI-RSCAS / Competition 2009 / Proceedings 27

V. SECTION 7 MERGER CHALLENGES

A. Why Section 7 Cases Are Different

In contrast to Sherman Act cases, in which ambiguous evidence favors early dismissal of cases, in Section 7 merger cases the opposite (although certainly not invariably) tends to be true. Part of this reality is due to the statute itself: Section 7 prohibits transactions that "may" tend to substantially reduce competition and tend to create a monopoly. While early cases (e.g., during the 1970s) read this language quite broadly,163 more recent cases make clear that "may" at least requires reasonable probability164 – again, not always easy to meet but far less rigorous than the actual effects required to be demonstrated in all Sherman Act cases that fall outside of per se condemnation. The more practical reason Section 7 cases tend to favor plaintiffs when assessing the type and sufficiency of evidence is that the overwhelming number of cases arise in the context of the government seeking preliminary injunction relief. Unlike a private plaintiff, the government does not have to prove that it has "standing" or is otherwise a proper plaintiff to pursue a Section 7 case – requirements that have, on occasion, blocked some of the few private Section 7 cases brought over the years.165 Further, in the preliminary injunction context, the government, as a general procedural matter, need establish only a likelihood of success on the merits of its Section 7 claim. Indeed, very recently, a few courts have found that the FTC – in contrast to the DOJ – enjoys a statutory mandate under Section 13(b) of the FTC Act that allows for injunctive relief even if the FTC establishes only "serious questions" going to the merits of its claims. 166 Whether that standard survives further appellate consideration, or the slow hand of Congressional reform of Section 13(b), is yet to be seen. But what is certain is that the government, and perhaps especially the FTC for the time being, have a built-in advantage in seeking to enjoin mergers and other transactions under Section 7.

______(cont'd from previous page) that "causation can be found where Plaintiffs allege the violation caused the injury and no evidence of an independent, non-violative cause exists"). 163 United States v. Vons Grocery, 384 U.S. 270 (1966). 164 FTC v. Arch Coal, Inc., 329 F. Supp. 2d 109,115-16 (D.D.C. 2004); United States v. Oracle Corp., 331 F. Supp. 2d 1098, 1110 (N.D. Cal. 2004). 165 See generally Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104 (1986); Brunswick Corp. v. Pueblo Bowl- O-Mat, Inc., 429 U.S. 477 (1977); Alta. Gas Chems. Ltd. v. E.I. Du Pont De Nemours & Co., 826 F.2d 1235 (3d Cir. 1987). 166 FTC v. Whole Foods Market, Inc., 548 F.3d 1028, 1042 (D.C. Cir. 2008); FTC v. CCC Holdings, Inc., 605 F. Supp. 2d 26, 36 (D.D.C. 2009).

28 EUI-RSCAS / Competition 2009 / Proceedings

B. Economic Theory and Evidence Plays a Critical Role in Section 7 Cases

As for the type of evidentiary battles that play out in these Section 7 cases, a few general observations help inform the review of recent cases. Again, in sharp contrast to Section 1 conspiracy cases, the issue is primarily about likely effects. This, in turn, focuses the evidentiary inquiry primarily on two issues. The first is whether the government has met its burden on proving a sufficient "structural" case to enjoy a rebuttable presumption that the proposed transaction may tend substantially to harm competition.167 This inquiry focuses on market definition (product and geographic), changes in market concentration (measured by HHI168), and the presence of entry barriers so that any implied anticompetitive effect is not merely ephemeral.169 Section 7 defendants typically, but not always, offer enough "rebuttal" evidence to focus the analysis on likely competitive effects, with the government carrying the ultimate burden of proof.170 Second, the evidentiary contest centers on proof of anticompetitive effects which, in recent years, has focused on two economic theories – "coordinated" effects and "unilateral" effects. Standard coordinated effects analysis asks whether the proposed transaction will make anticompetitive coordination (whether lawful in its own right or not) more likely and sustainable. Unilateral effects analysis assesses whether, among other things, a firm, post- merger, has eliminated a uniquely close rival such that the proposed merger or transaction enables the merged firm to raise prices or reduce competitiveness "unilaterally" – i.e., irrespective of the likely responses of the primary firms in the market. As even the general principles suggest, Section 7 evidence – both on market structure and effect – tends to be heavily "economic" and involves experts, economic modeling or simulations and econometrics that often seem inaccessible to mere mortal antitrust practitioners and courts. Moreover, much of the complexity in Section 7 economic analysis is driven by an inherent ambiguity in nearly all Section 7 cases: no one really knows in advance whether a merger will reduce competition in a sustained manner; indeed, most of the time, there is not even enough data to predict what would happen to purchasing patterns or prices if, subsequent to the transaction, the merged firm were to raise prices – the elusive cross-elasticity of demand inquiry that is so critical to the "what if" questions underlying both market definition and effects analysis.

167 United States v. Phila. Nat'l Bank, 374 U.S. 321 (1963). 168 U.S. Dep't of Justice & Fed. Trade Comm'n, Horizontal Merger Guidelines § 1.51 (1992) (rev. 1997); see also FTC v. Univ. Health, Inc., 938 F.2d 1206, 1211 & n.12 (11th Cir. 1991); United States v. Oracle Corp., 331 F. Supp. 2d 1098, 1102-03 (N.D. Cal. 2004). 169 United States v. Syufy Enters., 903 F.2d 659 (9th Cir. 1990). 170 United States v. Baker Hughes, Inc., 908 F.2d 981 (D.C. Cir. 1990).

EUI-RSCAS / Competition 2009 / Proceedings 29

C. Recent Section 7 Cases the Government Has Lost and the "Evidence"

A District of Columbia district court's weighing of qualitative and quantitative evidence figured heavily in FTC v. Arch Coal, Inc.,171 in which the FTC brought suit against a proposed merger in the coal industry. The FTC sought a preliminary injunction against Arch Coal, Inc. following its proposed merger with New Vulcan Coal Holdings, LLC.172 The FTC argued that the proposed merger would lead to "future 'tacit coordination' among competitors to restrict production."173 The parties' dispute regarding product market definition centered principally on whether 8800 Btu SPRB coal could itself constitute the relevant product market or whether there existed sufficient cross-elasticity of demand to include additional grades of coal, specifically 8400 Btu SPRB.174 In reaching its conclusion that a broader product market should be utilized, the court focused on how reluctant the government expert was to definitively assert that the 8800 Btu SPRB coal could suffice as a complete market.175 The court did not agree that a distinguishable market could be discerned based largely on qualitative evidence that consumers preferred 8800 Btu over 8400 Btu.176 Since those consumers stated that they were willing to purchase 8400 Btu, their preference of 8800 Btu could not alone justify a separate, narrow product market.177 Finally, the FTC provided little quantitative evidence regarding the coal industry's previous competitive marketing or profit structure that could justify an 8800 Btu market on its own. No evidence provided by the government could lead the court to definitively conclude that the two species of coal were not interchangeable. Since "virtually all the utilities acknowledged that they can and do purchase and consume both 8800 and 8400 Btu coal," the court ruled that the government had not met its burden.178 Despite defining a broader market, the court nonetheless determined that the FTC had met its burden in showing a prima facie case of significant market concentration, albeit not a particularly strong prima facie case.179 The court next considered potential market trends that, as the FTC asserted, would lead to tacit coordination among competitors in a post-merger market. To show this likelihood, the FTC presented evidence regarding past output restriction conduct among SPRB coal competitors that led to an increase of price.180 However, based on the speculative nature of the proposed future collaboration, the district court refused to accept the FTC's argument that past practices might lead to future tacit coordination. Absent statistical

171 329 F. Supp. 2d 109 (D.D.C. 2004). 172 Ibid. at 114. 173 Ibid. at 115. 174 Ibid. at 121. 175 Ibid. ("Nonetheless, the clear impression left by Dr. Morris is that he is not entirely sure that 8800 Btu SPRB coal is a relevant product market, and that he would prefer to undertake further analysis before so concluding."). 176 Ibid. at 122. 177 Ibid. at 122-23. This is somewhat similar to Oracle's analysis of consumer preferences and the lack of hard, statistical analysis. See infra pp. 48-50. 178 Arch Coal, 329 F. Supp. 2d at 121. 179 Ibid. at 130. 180 Ibid. at 131.

30 EUI-RSCAS / Competition 2009 / Proceedings

certainty or quantitative evidence, the FTC failed to demonstrate that such actions would result. Specifically, the FTC submitted evidence regarding a 2001 price spike, but the court did not consider this to be probative of anticompetitive activity or collusive coordination.181 The court also rejected evidence of collusion based on a period of reduced output in 1999- 2000 because it determined that each competitor possessed independent legitimate business justifications for its conduct.182 Meanwhile, the court found that the FTC's qualitative evidence suggesting potential interest in price fixing among competitors, based on previous statements made by executives, was too speculative and likely involved legitimate business strategy discussions.183 Similarly, in United States v. Oracle Corp.,184 a California district court rejected the government's argument that the relevant product market should be construed so narrowly as to include only "high function software" – a term conceived by the government itself, and not commonly associated with the defendants' industry.185 The merger between Oracle and PeopleSoft prompted the DOJ to initiate litigation to block the transaction under Section 7 of the Clayton Act based on the alleged anticompetitive effect the merger would have on a particular segment of the computer software industry. Specifically, the government alleged that the merger would lead to significant market concentration between two office software products known as human relations management ("HRM") and financial management systems ("FMS"), since their value among large complex enterprises ("LCE") was significantly higher than other, mid-range customers who would be interested in the product.186 The court's refusal to accept the "high function software" market definition stemmed largely from the fact that the government could not provide quantitative evidence (i.e., market statistics or financials) to support its proposed market. The government's evidence included an extensive list of customer witnesses who testified that, as consumers in the alleged "high function software" market, they understood Oracle and PeopleSoft to have no practical competition in the industry. Though the court respected this form of qualitative evidence, it questioned whether "[c]ustomer preferences towards one product over another [could] negate interchangeability" among different software products.187 The court, in essence, refused to accept subjective business decisions as conclusive evidence of a definitive product market. While such forms of qualitative evidence could have value, the court held that "customer apprehensions do not substitute for hard evidence."188 Aside from customer-based qualitative evidence (and similar testimony by industry- based witnesses who provided similar testimony), the government relied principally on the testimony of its primary expert, Professor Kenneth Elzinga, who proffered quantitative

181 Ibid. at 133-34. 182 Ibid. at 135. 183 Ibid. at 138-39. 184 331 F. Supp. 2d 1098 (N.D. Cal. 2004). 185 Ibid. at 1158. 186 Ibid. at 1102-03. 187 Ibid. at 1131. 188 Ibid.

EUI-RSCAS / Competition 2009 / Proceedings 31

evidence in support of the product market definition sought by the government. For instance, Elzinga provided statistics that Oracle salespersons would offer larger-than-normal customer discounts most often on its HRM and FMS products when it knew it was in direct competition with either PeopleSoft or SAP AG ("SAP") (the third competitor).189 Despite these facts, the court assailed the quantitative findings as painting still too "equivocal and vague" a picture of the narrow submarket.190 The court emphasized the potential uncertainty that might surround the definition of a product market, especially when the government's witness himself admitted that "no 'quantitative metrics' . . . could be used to distinguish a vendor of high function [software] from a vendor of mid-market software."191 The court appeared to hone in on the government expert's own uncertainty, suggesting that experts must be confident and sure of their estimations before a court should accept their findings and recommendations.192 To be sure, the court's conclusion was based on a concept of degree: while the government had provided some evidence to suggest the presence of a submarket, it had simply not shown enough. Thus, without the narrow market advocated by the DOJ, no unilateral threat could be asserted as a result of the merger.

D. Recent Section 7 Cases the Government Has Won and Why

More recently, the government – primarily the FTC – has made a significant comeback, albeit in large part due to the more generous Section 13(b) standard. In FTC v. Whole Foods Market, Inc.,193 the Court of Appeals for the District of Columbia reversed a district court holding and ruled that the FTC's proposed product market, "premium, natural, and organic supermarkets" ("PNOS"), for the merger between Whole Foods and Wild Oats was appropriate given the contours of the food market and "core" consumer demand. The court rejected a critical loss analysis, where the focus would center on the response of marginal customers to a SSNIP ("small but significant non-transitory increase in price," a technique developed by the FTC and the DOJ), in favor of a critical diversion analysis, whereby the court concentrated on how many consumers would be diverted to Whole Foods compared to conventional supermarkets if a local Wild Oats were shut down.194 This decision allowed the court to consider a broad spectrum of market evidence which pointed to a distinct PNOS market, whereas a critical loss analysis focused solely on the market reaction of marginal (not core, or average) consumers. In ruling against Whole Foods, the Court of Appeals focused on quantitative evidence produced by the FTC which, in its view, showed that Whole Foods' profits and ability to price discriminate suffered considerably when it competed in close proximity to Wild Oats stores,

189 Ibid. at 1145-46. 190 Ibid. at 1158 191 Ibid. at 1154. 192 Ibid. at 1159. 193 548 F.3d 1028 (D.C. Cir. 2008). 194 Ibid. at 1038.

32 EUI-RSCAS / Competition 2009 / Proceedings

whereas Whole Foods fared far better when competing with basic grocery stores.195 Further, when Whole Foods sought to price discriminate in competition with conventional supermarkets, it would focus on its smaller collection of "dry grocery" products, whereas 70% of Whole Foods business existed in the perishable (non-dry) market.196 The government also provided statistical evidence suggesting "that PNOS competition had a greater effect than conventional supermarkets on PNOS prices."197 Thus, the court reasoned that PNOS prices were often insulated from developments or changes in the general grocery industry, suggesting a distinct submarket in itself. From a qualitative standpoint, the court also emphasized that Whole Foods' corporate structure catered to a specific group of food consumers who valued the advantages of an organic nutrition lifestyle.198 It was clear, according to the court, that Whole Foods and Wild Oats provided unique services and a food-shopping environment which focused on particular values espoused by PNOS customers. The court also highlighted that a majority of Whole Foods customers identified with the PNOS' goals and values.199 Finally, the court relied on various comments made by executives and separate industry-related reports indicating that Whole Foods indeed identified the PNOS market as its primary realm of competition.200 To be sure, the more lenient FTC standard did provide some advantage to the government, as it did not have to show, in the context of a preliminary injunction, irreparable harm. However, the court did not hesitate to define a distinct "submarket" based on evidence of isolated market competition between Whole Foods and Wild Oats, along with general marketing and consumer identification testimony. More recently, in FTC v. CCC Holdings, Inc.,201 a District of Columbia district court granted the FTC's request of a preliminary injunction for a proposed merger that would create a "duopoly" in the relevant product market based on a coordinated effects theory. The proposed merger involved computer software enterprises who sold specialized software that would assist insurance companies in computing costs of repair or replacement costs.202 The two merging companies, CCC and Mitchell International, constituted two-thirds of what was commonly known as the "Big Three" in total loss software systems (TLV).203 Before considering potential collusive effects post-merger, the court focused on whether the product market should be defined narrowly to include only the TLV produced by the Big Three or whether the market would also include various books used in-house by insurance companies. The court placed heavy emphasis on both qualitative and quantitative evidence produced by the FTC to conclude that the two products were, in fact,

195 Ibid. at 1039-40. 196 Ibid. at 1040. 197 Ibid. 198 Ibid. at 1039. 199 Ibid. at 1039-40. Under a critical loss analysis, courts would look to a canvassing of an entire consumer base to view general characteristics, since their focus would be solely on the actions of marginal customers. 200 Ibid. at 1044-48 (Tatel, J., concurring). 201 605 F. Supp. 2d 26 (D.D.C. 2009). 202 Ibid. at 30. 203 Ibid. at 32. The third company in the market was Audatex.

EUI-RSCAS / Competition 2009 / Proceedings 33

distinguishable and so the TLV could stand as its own defined product market. The defendant initially sought to dispel the FTC's claim, providing isolated examples, where, for example, one insurance company calculated roughly 500,000 of its 700,000 total loss claims in- house.204 Another company, Progressive Insurance, calculated 75% of its claims in-house.205 However, the court rejected these findings, citing the fact that, regardless of these isolated figures, "over 90% of all total loss claims are calculated using TLV sold by one of the three major competitors."206 The court also relied on the FTC's expert's SSNIP testimony involving a critical loss analysis. The court alluded to qualitative statements made by Insurance executives who said they would not switch to in-house service even at a 10% increase of TLV prices.207 Once gaining the benefit of a narrow product market definition, the FTC, relying on FTC v. H.J. Heinz Co.,208 argued that the merger of two competitors where the post-merger market is duopolistic, provides sufficient evidence that collusive behavior will result.209 The district court, however, rejected this per se-type approach and instead required a more in- depth analysis of the particular industry in question. The court nonetheless found high barriers to entry in the TLV market. The court dismissed defendant's arguments that since previous market entry had been relatively low, the post-merger barrier would likewise remain low. The court alluded to the relative lack of success among previous market entrants as demonstrating that future, potential market entrants in a post-merger duopoly market would be high.210 Based on the high barriers of entry, the court found that coordinated activity between the remaining competitors (CCC and Audatex) was likely based on the defendant's own admission of "minimum differentiation in offerings."211 Indeed, the defendant's own view of the market was crucial evidence. The court also pointed out the transparency of customary TLV business practices, specifically that the TLV would be sold on a per valuation basis and thus concluded that there would only be one price to track.212 Overall, the general homogeneity of the product's pricing and offering scheme was the critical factor in determining that anticompetitive coordination following the merger was probable.213 Finally, the court specifically rejected the unilateral effects argument made by the FTC. The court refused to accept the models and predictions of the FTC’s expert because they did not align with the specific evidence associated with the TLV market.214 Specifically,

204 Ibid. at 39. 205 Ibid. 206 Ibid. at 40. 207 Ibid. at 41. 208 246 F.3d 708 (D.C. Cir. 2001). 209 In Heinz, the court of appeals noted that the relatively high barriers of entry in the baby food market (given the trusted brand labels) and complete transparency in pricing in that market demonstrated that a likelihood of anticompetitive collusion post-merger was high. See FTC v. CCC Holdings, Inc., 605 F. Supp. 2d 26, 46 (D.D.C. 2009). 210 Ibid. at 47-48. 211 Ibid. at 62 (citation omitted). 212 Ibid. at 63. 213 Ibid. at 66 & n.42. 214 Ibid. at 69-70.

34 EUI-RSCAS / Competition 2009 / Proceedings

the model presented by the FTC had assumed that consumers preferred the merging companies as either their first and second choice, whereas Audatex would allegedly finish a distant third.215 On this basis, the FTC had argued that, post-merger, the firms would act independently given the disparity in consumer preference between the two remaining duopoly firms. However, the FTC failed to produce evidence suggesting similar consumer preferences. Thus, the court found no evidence to support the conclusion that the post-merger firms would have their own unilateral anticompetitive effect. The general consensus coming out of CCC Holdings, and the Section 13(b) standard adopted by the court, is that a "tie" goes to the government (or at least the FTC). It has yet to be seen whether this evidentiary standard will hold and whether the DOJ will step up its merger enforcement activity to get back in the game.

E. Section 7: Concluding Thoughts

In assessing the quality and quantity of evidence necessary for the government to prevail in a Section 7 case, we are struck by two fairly unpredictable factors: the particular judge that is assigned to the case (randomly); and the extent of the impact of the relatively low Section 13(b) standard in FTC matters. Nevertheless, as the cases discussed in this section make clear, the key evidentiary battles will remain well defined. The contest to define the relevant antitrust market (product and geographic) will continue to be the primary battleground in these types of actions; this is because the government's entire structural case flows from the results of that contest. In this respect, courts will remain hungry for both qualitative and quantitative evidence demonstrating, if possible, the extent of historical competition between and among potential substitute products and, also if possible, a prediction of the likely effect if market power (unilateral or in coordination) were to be exercised following the transaction – the infamous SSNIP test. In addition, evidence showing that a firm's profits, pre-merger, were either affected or insulated from a competitor's proximity will remain crucial, as it was in Whole Foods. And, of course, courts will be hesitant to embrace a market definition when the principal expert himself expresses doubt about whether his advocated market can be definitively construed. Moreover, while testimony by industry experts or consumers will remain relevant (especially if they favor the government), courts generally will not rely on consumer preference as a substitute for concrete, quantitative evidence. Finally, a court's determination on whether to adopt a critical loss analysis (which emphasizes marginal investors) or a diversion outlook (which focuses more on average or core consumers) may influence the outcome of these market definition inquiries, especially in differentiated product markets. Ultimately, however, the dispositive evidence will relate to competitive effects. Here the evidence will center around: (i) how many competitors remain in the market and whether market conditions substantially increase the likelihood of coordination; and (ii) whether the merger is between two uniquely situated rivals which are already insulated from competition

215 Ibid. at 71.

EUI-RSCAS / Competition 2009 / Proceedings 35

and which may be free, following the transaction, to exercise an incremental amount of market power. While evidence on these issues will continue to be heavily "economic," practitioners and their clients should always remember that what is said or written about the likely effects of a merger can greatly influence how U.S. courts view the remaining evidence.

Antitrust Trade and Practice EU and Google: Study in Divergence for Antitrust Enforcement

Shepard Goldfein and James Keyte, New York Law Journal

May 12, 2015

On April 15, 2015, the European Commission levied formal charges against Google, the culmination of a long-simmering and politically charged investigation into the Internet giant's search practices. Despite various inquiries in recent years, the announcement marks the first time a government regulator has gone so far as to charge Google with an antitrust violation. The charges—which assert that Google abused its dominant position in the European search engine market to favor its own "vertical" services—are coupled with the launch of a formal investigation into allegations that Google currently bundles its Android mobile operating system with Google applications. The news has been met with a defiant response by Google, as well as both praise and criticism from state governments and antitrust commentators.

The recent EC decision is not the first time a prominent government agency has examined Google's potential anti-competitive behavior. In 2013, the Federal Trade Commission concluded its own investigation into Google's search practices, recognizing a pro-competitive basis for Google's prioritization of certain content. Complainants had alleged that Google utilized an algorithm specifically tailored to favor the return of Google's own content above that of competitors during an Internet search. This practice, known as "search bias," resulted in the supposed favoring of "vertical" Google content—i.e., Google-sponsored shopping and travel searches—at Google's competitors' expense.

The FTC ultimately concluded that these design changes were aimed at improving the user experience (by offering more responsive content) and that any harm to competitors was purely incidental. Indeed, the FTC found that Google frequently conducted testing to measure the effects of these changes on consumers, as well as determining that other general search engines had embraced similar tweaks. In declining to file charges against Google, the FTC stated that to "second-guess a firm's product design decisions where plausible procompetitive justifications have been offered, and where those justifications are supported by ample evidence" would be inappropriate from an antitrust perspective.1

The FTC was, however, able to secure a number of concessions from Google, including the company's promise to provide competitors with access to certain patents, allow advertisers increased flexibility to manage ads on Google's AdWords platform and grant certain websites an "opt out" option from Google searches. Despite these remedies, the decision was largely hailed as a victory for Google that allowed the company to bypass the financial and reputational drain of a prolonged antitrust battle similar to the one Microsoft endured in the 1990s.2

The FTC's 2013 decision was largely met with positive reactions from antitrust commentators, many of whom saw the investigation as (in the words of former FTC chairman James Miller) a "shameless attempt at rent-seeking" by Google's rivals.3 In early 2015, The Wall Street Journal obtained a copy of an FTC staff report, which referred to the FTC's 2013 decision not to file charges as a "close call." The Wall Street Journal article suggested that the decision may have been influenced by a political agenda, a claim that the FTC vehemently denied in a press release shortly after the article's publication.

Action in Europe

While the FTC was conducting its Google investigation, the European Commission was proceeding with its own inquiry into Google's search practices. The EC specifically looked at three primary areas: (1) search bias, (2) copying (or "scraping") of content from other search engines and (3) restrictions on the use of certain Google advertising features. The EC eventually reached a tentative settlement with Google on the first issue (which had become the focus of the investigation)4 that would have allowed competitors to purchase space near the top of Google search results pages. The proposed remedy was met with a flurry of harsh criticism from EU government officials and commissioners, as well as Google's European competitors, arguing that the EC had failed to extract an adequate settlement to quell search bias fears.

Despite Google's public appeal that its search algorithm was pro-competitive and benefited consumers, it appears that intense political pressure culminated in EC antitrust head Joaquin Almunia's decision in September 2014 to re-open the Google investigation. Additionally, Almunia initiated a separate investigation track focused on Google's supposed bundling of its mobile phone Android operating system with Google applications.

Less than a year later, new EC antitrust chief, Margarethe Vestager, announced the EC's decision to file charges on the search bias issue, while simultaneously formalizing the EC's investigation into Google's use of Android. Vestager noted that the EC's decision to file charges was based on the "preliminary conclusion that Google had abused its dominant position to systematically favor its own comparison shopping service, Google Shopping, over rival services on its general search page."5 Meanwhile, the Android investigation would focus on whether Google improperly utilized its market leading position in mobile operating systems to hinder the development of competitors' products by, for example, requiring the bundling of Android with Google applications.

Google responded to the charges with a mix of public statements and an internal memo that focused on the pro-competitive justifications for Google's alleged search bias and the high degree of competition amongst search engines and vertical services. Specifically, Google highlighted competition from general search engines (like Bing and Yahoo), specialized websites (like Amazon and eBay) and social media as vying for "vertical" sale and marketing opportunities. On the Android issue, Google countered that Android is an open-source operating system that can be freely accessed and that the pre-loading of certain applications on Android enhances the user experience.

Much of the commentary that followed the announcement was muted in its support of the EC's decision. In fact, a number of commentators viewed the EC's decision as typical of the European Union's focus on European protectionism, a practice that critics argue has stifled the rise of major European tech companies capable of competing with U.S. giants such as Google and Facebook.6 Other commentators7 were quick to draw a comparison to the EU's decision to challenge Microsoft's bundling practices, an investigation that led to a €2.2 billion fine for Microsoft.8 In fact, Microsoft has found itself connected to the recent EC charges in more ways than one—a recent New York Times article suggests that Microsoft has deep ties to a number of the entities that lobbied the EC to bring its antitrust campaign, perhaps in an attempt to stifle the growth of a key rival.9

Issues and Impact

There are no indications that the announcement of charges will affect the United States' position on the issue. The decision is, however, indicative of key differences in U.S. and EU approaches to antitrust enforcement. The 2013 FTC decision was largely based on the fact that, despite potential incidental harm to competitors, consumers were benefited by Google's practices. U.S. antitrust law seeks to protect consumers first and foremost, whereas EU law requires an additional focus on competitors' welfare.

The EC decision is also fraught with potential political implications. President Barack Obama recently cautioned the EU against making "commercially driven" decisions against major U.S. tech companies. Indeed, shortly before the EC announcement, Daniel Sepulveda, deputy assistant secretary in the U.S. State Department Bureau of Economic and Business Affairs, warned the EU against basing its decision on a protectionist agenda, publicly stating that it is "important that the process of identifying competitive markets and remedies be based on impartial findings and not be politicized."10

So where do the parties go from here? Vestager has stated that the EC is open to settlement options, but noted that any acceptable settlement will need to differ significantly from the previous proposal that was based too heavily on "a particular look of the screen" rather than a change in corporate principles.11 Google now has approximately 10 weeks to respond to the charges, a process that ultimately could entail a hearing request before the commission and, in the event fines or an injunction are imposed, an appeal to the EU appeals court in Luxembourg.

An EC victory would almost certainly have a widespread impact on other U.S. tech companies currently facing antitrust scrutiny in Europe.12 Additionally, many will continue to closely watch the Android investigation, as some of Google's largest competitors (Microsoft and Apple, for example) also routinely provide their operating system in conjunction with various applications.

Of the many issues raised by the EC's Google investigation, one of the most interesting centers on whether antitrust investigations in the tech industry are inevitably antiquated. Since tech companies are rapidly innovating, an antitrust claim centered on perceived anti-competitive use of these technologies is often obsolete by the time a decision is rendered. Nonetheless, while the outcome of the instant case remains uncertain, it is clear that the EC does not fear plotting its own course of antitrust enforcement, the next step in the continuing divergence of U.S. and EU competition regulation. It is also starkly evident that this path, at least in the tech context, is riddled with dangers posed by the political motives of state actors and, perhaps, the self-interested agendas of competitors as well.

Endnotes:

1. Press Release, FED. TRADE COMM'N, Statement of the Federal Trade Commission Regarding Google's Search Practices, In the Matter of Google, at 3 (Jan. 3, 2013), FTC File No. 111-0163, https://www.ftc.gov/system/files/documents/public_statements/295971 /130103googlesearchstmtofcomm.pdf.

2. In the late 1990s, the U.S. Department of Justice, along with 20 individual states and the District of Columbia, filed suit alleging that Microsoft had utilized monopoly power to bundle products and commit a host of other antitrust violations. After a multi-year process that involved numerous appeals and remands, Microsoft settled with the Justice Department, but was later subject to suits by private parties that resulted in significant settlements and an extended European Commission investigation that ended with a massive fine. A. Douglas Melamed and Daniel L. Rubinfeld, U.S. v. Microsoft: "Lessons Learned and Issues Raised," in ANTITRUST STORIES (Eleanor M. Fox & Daniel A. Crane eds., 2007).

3. James C. Miller III, "The Case Against Google Was Always Weak," WALL ST. J., Jan. 3, 2013, http://www.wsj.com/articles/SB10001424127887323874204578220050460529918.

4. The advertising issues were settled separately.

5. Tom Fairless, EU Charges Google on Its Searches, WALL ST. J. EUR., Apr. 16, 2015, http://www.pressreader.com/belgium/the-wall-street-journal-europe/20150416/281865822000815 /TextView.

6. See James Kanter and Mark Scott, "Europe Challenges Google, Seeing Violations of Its Antitrust Law," N.Y. Times, April 15, 2015, http://www.nytimes.com/2015/04/16/business /international/european-union-google-antitrust-case.html.

7. Tom Fairless, "EU Charges Google on Its Searches," Wall St. J. Eur., April 16, 2015, http://www.pressreader.com/belgium/the-wall-street-journal-europe/20150416/281865822000815 /TextView.

8. If the EC prevails in the instant case, it could seek a fine of over €6billion,approximately10 percent of Google's annual revenue.

9. Danny Hakim, "Microsoft, Once an Antitrust Target, Is Now Google's Regulatory Scold," N.Y. TIMES, April 15, 2015, http://www.nytimes.com/2015/04/16/technology/microsoft-once-an-antitrust- target-is-now-googles-regulatory-scold.html?_r=0.

10. James Kanter and Mark Scott, "Europe Challenges Google, Seeing Violations of Its Antitrust Law," N.Y. TIMES, Apr. 15, 2015, http://www.nytimes.com/2015/04/16/business/international /european-union-google-antitrust-case.html.

11. Fairless, supra note 5.

12. The EC is also investigating certain tax concessions granted to Apple and Amazon, as well as Facebook's privacy protections.

Shepard Goldfein and James Keyte are partners at Skadden, Arps, Slate, Meagher & Flom. Michael Singer, an associate at the firm, assisted in the preparation of this column. Chapter

Private Antitrust Litigation in the EU: a James Keyte New Age of Advocacy

Skadden, Arps, Slate, Meagher & Flom LLP Paul Eckles

1. Introduction them, has also gotten involved in the new age of antitrust advocacy in the EU. CDC acquired claims of paper companies As the European Union ("EU") and its Member States prepare to and brought a follow-on damages action against chemical makers implement the EU Directive on antitrust damages actions Kemira, AkzoNobel and Eka.iv It also brought a case against (“Directive”),i including collective redress, it strikes those of us cement manufacturers in Germany but ran into a road-block when across the pond that antitrust litigators, competition authorities and a national court determined the cement buyers’ assignments of national courts throughout the EU are on the verge of a new era of their claims to CDC were invalid under German law.v These advocacy. No longer constrained by many of the procedural and issues will likely become more significant as other Member States substantive strictures that characterise administrative procedures adopt private enforcement procedures in line with the Directive. (both in the EU and Member States), parties (including collective That said, the US system may be of little guidance here. In redress companies such as CDC) on both sides of cases are poised general, the US and the EU diverge on how claimant parties to litigate substantive, procedural and economic issues in ways and should be formed in collective redress actions. In the US, depths not seen before within the EU. While such vigorous collective action (called "class action") rules specify that all class litigation has been standard in the US courts for many decades, in members are bound by any judgment, whether favourable or the EU we have only seen hints of what is to come by observing unfavourable, unless they affirmatively request exclusion from – what is taking place in those few jurisdictions (e.g., the UK, i.e., “opt-out” of – the certified class.vi The EU favours the Netherlands, Germany and Italy) that have gotten a jump-start on opposite approach. Although the Directive does not speak to the new private enforcement era. Indeed, the fact that Italy and collective redress, the EC Recommendation on common France recently announced a programme to educate judges on principles for injunctive and compensatory collective redress antitrust legal and economic principles aptly highlights what mechanisms (which is not binding on Member States) advises that clients, antitrust practitioners and courts are in store for in the the claimant party should be formed based on the express consent coming years and decades. of the persons claiming to be harmed. In other words, members This article briefly outlines some of the key areas that are likely to must affirmatively “opt-in” to collective actions.vii Courts in become the central battlegrounds of private antitrust enforcement some Member States (e.g., the Netherlands, UK, Spain and in EU Member States, highlights the primary US issues in those Portugal) have rejected this approach and adopted “opt-out” areas and notes where we are already seeing antitrust advocacy in systems. Whatever approach ultimately prevails among Member EU private litigation. States, policy considerations that underlie the US system – facilitating redress, but ensuring that collective claims are typical 2. Who May Bring Actions? of any asserted "class" – may find their way into Member State procedures and, in turn, decisions. One area that has been, and will continue to be, the subject of private litigation is who, or what types of entities, can bring 3. How Will National Courts Weigh Decisions of private cases, including for "collective redress". This is a National Competition Authorities? relatively nascent issue today in Member States. The Directive does not address the issue, but Member States continue to explore While EC final infringement decisions remain binding on national different methods for bringing claims. France has adopted a class courts, under the Directive final infringement decisions of other action system for competition cases, but only allows a limited national competition authorities only constitute prima facie group of government-approved consumer associations to evidence that a party has violated the competition laws.viii The represent consumers.ii Other countries have experimented with final Directive differed from the proposed Directive in this regard, assigning claims to third parties and claims vehicles. The German as the proposed Directive sought to also make the final government signed over its claims against a cartel of pre-stressing infringement decisions of national competition authorities steel producers to rail operator Deutsche Bahn, which then binding.ix As a result, when parties bring a follow-on action initiated legal proceedings on its assigned claims in the arising from a decision of another Member State’s competition Netherlands where courts are more permissive towards claims authority, national courts will need to determine the weight that assignment.iii Cartel Damage Claims (“CDC”), a private decision deserves. company that purchases antitrust damage claims and litigates

ICLG TO: CARTELS & LENIENCY 2015 WWW.ICLG.CO.UK Skadden, Arps, Slate, Meagher & Flom LLP Private Antitrust Litigation in the EU

The US may provide a useful model on this issue. Judgments of US overcharge was passed on down the supply chain.xiv Further, in antitrust enforcement agencies are not binding evidence in follow- most circumstances, indirect purchasers lack standing to bring on private antitrust actions. Instead, agency judgments are claims for alleged overcharges.xv Through this approach, the US admissible as prima facie evidence of matters actually and system promotes simplicity by not asking courts to apportion necessarily decided against a defendant in a government action. damages among purchasers along the distribution chain and helps Further, the plaintiff is only entitled to the prima facie effect where avoid the possibility of multiple damages. The goal of this system the government judgment (i) was final, (ii) was entered in a civil or is to avoid adding additional complexity to already complex criminal proceeding brought on behalf of the United States, (iii) antitrust litigation. resulted from an action brought under the antitrust laws, and (iv) Courts in the UK have already encountered the complexity of pass- was not a consent judgment entered before any testimony was on defence cases. In the Dow Chemicals case, the defendant, Dow, x taken. While national courts in the EU may adopt less demanding raised the pass-on defence to counter claims that tyre makers standards on the evidentiary value of national competition suffered damages from a rubber price-fixing cartel that included authorities’ decisions, we will need to wait and see how the courts Dow. Lawyers for the plaintiffs argued against the pass-on defence, make this determination. saying that “public policy considerations … strongly favor the drawing of a line to crystallize the loss at the point of overcharge.” 4. What is a Sufficient Claim on its Face? They continued, “[o]nce you enter into this question of whether, ultimately, the claimant has gained or lost, and to what extent the We are also likely to see other types of antitrust cases beyond the business has suffered from the overcharge, you … enter an infinite now common context of follow-on private litigation from cartel regression.” “It is possible to trace the economic consequences of liability findings of the EC or a national competition authority. the defendants’ overcharge almost indefinitely through the market.” Indeed, if the US is any guide, parties are likely to begin to bring Dow, for its part, argued that it should be able to invoke the pass-on actions for cartel-related conduct not covered by any decision or for defence to avoid paying out twice for the same damages.xvi plenary actions covering any number of potential antitrust Although the Dow case settled before the court resolved these violations (dominance, object or effect). In these cases, without the issues,xvii the back-and-forth between the parties illustrates the guidance of the EC or a national competition authority, Member challenging task for national courts. One possible outcome is States’ national courts may need to address whether certain claims, increased consolidation of claims initiated by direct and indirect on their face, actually raise antitrust issues. purchasers, but we anticipate significant litigation over these points This is an essential aspect of the US system, where class action until case law is more settled. cases can be rejected at the onset if what they allege factually – even if presumed true – would not amount to an antitrust violation. 6. What Battles can we Expect Over Disclosure? While many Member States have not needed to develop procedural mechanisms for such a review, we see no reason why national The EU placed disclosure issues front-and-centre in the Directive. courts would not eventually explore such threshold requirements. The Directive’s disclosure provisions reflect an effort to walk a As they do so, we can expect vigorous fights – well known in the fine line between several important policy goals, several of which US – over what may constitute antitrust misconduct. often conflict. The Directive balances disclosure rights of private claimants with protection of the EC’s leniency programme. At the 5. How Will Courts Deal With Indirect Purchaser same time, the Directive strives to correct the “information asymmetry” between claimants and alleged infringers and third Issues? parties. It attempts to provide private claimants with enough The Directive codifies the position of several Member States on the access to relevant evidence to make their case but also to avoid the standing of indirect purchasers and the availability of the “pass-on disclosure of confidential business information and to resist defence”. Indirect purchasers may bring claims for damages transformation to a full-blown US discovery system. While this depending on whether, or to what degree, an overcharge was passed balance protects important policies in the EU, it may also lead to to them, but the indirect purchaser also has the burden of providing protracted disclosure battles. In addition, parties have already the existence and scope of its damages.xi The Directive also litigated the limits of disclosure in a few follow-on cartel cases, permits defendants to invoke the pass-on defence – an affirmative and we can expect to see more of these disputes going forward. defence that a claimant passed on the whole or part of the The Directive itself confirms that disputes over access to files overcharge resulting from the infringement of competition law and possessed by competition authorities will continue to be a focal as a result, the claimant is either not entitled to damages or is point. Unlike in the US, where participation in leniency entitled to less damages than it is claiming.xii And of course, the programmes affords little protection against discovery of non- Directive also notes that full compensation for claimants should not privileged materials, the Directive prevents national courts from lead to overcompensation through multiple damages.xiii In its ordering disclosure of leniency statements and settlement attempt to provide compensation to all harmed parties, the Directive submissions.xviii National courts can, however, order disclosure has also left national courts with a complicated task. They must of all other files in the possession of a competition authority, simultaneously consider whether a direct purchaser was harmed, provided that they evaluate whether a claimant’s request is whether that direct purchaser passed on any of the overcharge to specifically formulated, whether the request is made in relation to consumers further down the supply chain and determine the amount an action for damages before a national court, and arguably most of harm, if any, at each level. importantly, whether disclosure is consistent with the need to In the US, the approach is much more streamlined. In the vast safeguard the effectiveness of the public enforcement of xix majority of cases, federal law precludes the pass-on defence. The competition law. The Directive envisions a dialogue between victim of an overcharge is found to be harmed whether or not the competition authorities and national courts on when the need to

WWW.ICLG.CO.UK ICLG TO: CARTELS & LENIENCY 2015 Skadden, Arps, Slate, Meagher & Flom LLP Private Antitrust Litigation in the EU

safeguard public enforcement of competition law outweighs 7. How Will Causation and Antitrust Injury disclosure, as it permits competition authorities to submit Inquiries Affect Private Litigation? observations about the proportionality of a disclosure request for files in the possession of the authority.xx As national courts Based on what has developed over the last several decades in the continue to balance the needs of claimants against public US – as well as a few early illustrations in some Member States enforcement needs, we can expect significant interaction between – one of the more active areas for private litigation in the EU the parties, national courts and competition authorities. could be on the subject of causation and so-called "causal National courts will also continue to balance providing private antitrust injury" – i.e., ensuring that private plaintiffs do not claimants with enough information to effectively assert their obtain remedies unless their harm actually flows in fact from claims and protecting confidential business information, all while antitrust misconduct, and (as developed further in the US) only xxv avoiding full-blown discovery. In a defending a series of cases, from "that which makes the conduct unlawful". It is a MasterCard resisted disclosure on confidentiality and relevance mouthful, but has been critical in US cases for separating grounds. Further complicating the matter, the EC had possession "antitrust injury" from harm that flows from other causes or harm of the documents in question, which it had gathered in an that may flow from antitrust misconduct but not from that which antitrust probe. To help resolve the complicated disclosure made the misconduct unlawful under the antitrust laws (e.g., a issues, the UK court sought guidance from the EC. The EC, in merger may be illegal under antitrust principles, but plaintiffs turn, advised the court that the requested documents could be still must prove that their harm flows from the reduction of disclosed, provided that adequate protection was given to competition that made the merger illegal rather than from the business secrets. The EC suggested a confidentiality ring – a mere change in management or strategies resulting from the xxvi device that limits disclosure among a limited group of lawyers – combination). Granted, this latter principle is complex even as the means to provide that protection.xxi Courts can expect to within US case law and may take some time to surface in EU continually confront these issues as the Directive requires private litigation. But, given the new age of advocacy and the Member States to ensure their courts have adequate measures in moneys involved, it is likely to gain traction sooner than we may place to protect confidential information.xxii We also suspect that expect. courts, much like in the US, are going to become increasingly flexible in granting targeted discovery requests. Although the EU 8. What Role Will Economic Experts Play in strongly opposes the US system of discovery, the Directive’s Private Antitrust Litigation? focus on informational asymmetry suggests at least a slight move away from the limited disclosure that has historically prevailed in Finally, there is no doubt in our minds that one of the most hotly many Member States. contested areas for private litigation going forward will be Finally, there has been a great deal of litigation (and ambiguity) surrounding economic experts – a virtual cottage industry in the surrounding disclosure in follow-on cartel cases. Here again, the US in large part because of private antitrust litigation. As in the UK may have provided a preview of what is to come. In the US, these issues cover the gamut, ranging from the traditional National Grid case, the claimant, National Grid, requested merits (defining markets, market power, coordination versus information from each of the defendants, all of whom were unilateral conduct) to causation and "antitrust injury" (as noted, already fined by the EC for their role in a cartel. The claimant tracking whether the alleged harm "flows" from the alleged wanted information on how the cartel worked in the UK.xxiii antitrust misconduct) to damages (both as to causation and Defendants Siemens and Alstom both refused to reply to the calculation). requests and argued that National Grid’s allegations covered Economic experts will also confront several issues unique to the matters outside the EC decision fining the defendants for their role EU. As discussed above, the pass-on defence and indirect in the cartel and therefore could not be disclosed. Judge Roth, the purchaser issues raise complicated issues of damage assessment and presiding judge, ordered disclosure but provided ambiguous apportionment among parties. In addition, in Kone AG and Others, guidance for future litigants. Roth acknowledged that a party’s the European Court of Justice (“ECJ”) held Member States cannot claim must fall within the scope of the EC decision to order exclude civil liability for umbrella damages in cartel cases.xxvii The disclosure of the requested materials. He stated that disclosure umbrella damages doctrine allows a court to hold a cartel’s required, “looking at the decision [and] whether it encompasses members liable for damages caused by price increases of non-cartel the UK and in what way”. Roth also noted that the EC decision members that were able to free-ride on the cartel’s price effects. In set out a breach of the antitrust laws but did not provide the addition to exposing cartel members to even greater damages, xxiv “whole detail of how the cartel actually operated”. Future umbrella damages require complex proof that will call for expert courts will need to further determine whether follow-on cartel economic testimony. allegations fall within the scope of an EC decision. While these Member states are beginning to prepare their courts for these issues may be clarified as Member States implement the complicated economic issues. Last spring Italy and France Directive, there likely will be ongoing advocacy over the announced they will begin educating their generalist judges about propriety and scope of disclosure, especially with the settled competition economics and law.xxviii These are issues already policy favoring leniency programmes. being litigated in a few of the more active Member States, and there is likely to be an explosion in this area in the years to come.

ICLG TO: CARTELS & LENIENCY 2015 WWW.ICLG.CO.UK Skadden, Arps, Slate, Meagher & Flom LLP Private Antitrust Litigation in the EU

Endnotes xvi. Lewis Crofts, Michelin, Pirelli, others seek to scotch “passing on” claims as Dow trial opens, MLEX, 12 May i. Directive of the European Parliament and of the Council on 2014. certain rules governing actions for damages under national xvii. Lewis Crofts, Dow settles UK lawsuit with tiremakers over law for infringements of the competition law provisions of rubber-cartel damages (update), MLEX, 28 May 2014. the Member States and of the European Union. Document A7-0089/2014, adopted 17 April 2014, available at xviii. Article 6(6) of the Directive. http://www.europarl.europa.eu/sides/getDoc.do?pubRef=- xix. Article 6(4). Note that although national courts can order //EP//NONSGML+AMD+A7-2014-0089+002- disclosure of any information besides leniency statements 002+DOC+PDF+V0//EN. and settlement submissions, the Directive limits the time ii. Melissa Lipmann, New French Law OKs Antitrust, when national courts can order disclosure of certain Consumer Class Actions, Law 360, March 20, 2014, materials. Only after the competition authority has closed its available at http://www.law360.com/articles/520536/new- proceedings can national courts order disclosure of french-law-oks-antitrust-consumer-class-actions. information prepared for competition authority proceedings, information the competition authority has drawn up and sent iii. Lewis Crofts, Comment: Deutsche Bahn cartel claim may to the parties in the course of proceedings and withdrawn break new ground for Dutch Courts, MLEX, 8 May 2014. settlement submissions after. See Article 6(5). iv. Christiaan Nelisse, Kemira, AkzoNobel lose fight against xx. Article 6(11) of the Directive. damage claim in paper bleach cartel, MLEX, 4 June 2014. xxi. Lewis Crofts, EU tells UK court it can disclose some v. Melissa Lipman, German Court Nixes €176M Cement Cartel MasterCard antitrust information, MLEX, 27 June 2014. Suit, Law 360, December 17, 2013, available at http://www.law360.com/articles/496512/german-court- xxii. Article 5(5) of the Directive. nixes-176m-cement-cartel-suit. xxiii. National Grid Electricity Transmission Plc v ABB Ltd & Ors vi. In the US, federal class actions are governed by Federal Rule [2013] EWHC 822 (Ch) (11 April 2013). of Civil Procedure Rule 23 and the vast majority of class xxiv. Sille Ruubel, National Grid, Siemens fight over UK cartel actions are governed by Rule 23(b)(3). Rule 23(c)(2-3) data in damages claim, MLEX, 1 May 2014. provide rules on opting out of a certified class. xxv. Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 US 477 vii. Commission Recommendation of 11 June 2013 on common (1977). principles for injunctive and compensatory collective redress xxvi. See Alberta Gas Chems. V. E.I. DuPont de Nemours & Co., mechanisms in the Member States concerning violations of 826 F.2d 1235, 1240 (3d. Cir 1987). rights granted under Union Law, Article V, ¶ 21, available at xxvii. See Case C-557/12, Kone AG and Others, judgment of June http://eur-lex.europa.eu/legalcontent/EN/TXT/ 5, 2014. See also Irene Fraile, EU High Court Opens The HTML/?uri=CELEX:32013H0396&from=EN. Door To Umbrella Liability, Law 360, June 10, 2014, viii. Article 9(2) of the Directive. available at http://www.law360.com/articles/546044/eu- ix. See Henry Vane, Finish line in sight for EU cartel damages high-court-opens-the-door-to-umbrella-liability. directive, GCR, 19 March 2014 available at xxviii.Ron Knox, Italian training project pushes for understanding http://globalcompetitionreview.com/news/article/35534/finis in court, Global Competition Review, 25 April 2014. h-line-sight-eu-cartel-damages-directive/. x. 15 USC §16(a). Acknowledgment xi. Article 14 of the Directive. See also Article 12 of the Directive. James Keyte and Paul Eckles are partners in the antitrust group of xii. Article 13 of the Directive. Skadden, Arps, Slate, Meagher & Flom LLP. They would like to xiii. Preamble ¶12 of the Directive. thank Mike Folger, an associate in the group, for his invaluable xiv. See Hanover Shoe, Inc. v. United Machinery Corp., 392 US assistance in preparing the article. 481, 491-94 (1968). xv. See Illinois Brick Co. v. Illinois, 431 US 720, 730-32 (1977).

WWW.ICLG.CO.UK ICLG TO: CARTELS & LENIENCY 2015 Skadden, Arps, Slate, Meagher & Flom LLP Private Antitrust Litigation in the EU

James Keyte Paul Eckles

Skadden, Arps, Slate, Meagher & Flom LLP Skadden, Arps, Slate, Meagher & Flom LLP Four Times Square Four Times Square New York, NY 10036 New York, NY 10036 USA USA

Tel: +1 212 735 2583 Tel: +1 212 735 2578 Fax: +1 917 777 2583 Fax: +1 917 777 2578 Email: [email protected] Email: [email protected] URL: www.skadden.com URL: www.skadden.com

James A. Keyte handles a wide variety of antitrust litigation, Paul M. Eckles represents clients across numerous industries in transactional and advisory matters across numerous industries. antitrust and other complex litigation matters. In the litigation area, Mr. Keyte has handled numerous cases Mr. Eckles has extensive experience in defending class actions. involving alleged price-fixing, monopolisation, litigated mergers, He is currently defending the National Hockey League in class other restraints of trade and class actions. Matters include cases action litigation relating to the league’s exclusive broadcast for SanDisk, Anheuser-Busch InBev, Pfizer, USIM, New York- territories. He is also defending Actavis in reverse payment class Presbyterian Hospital, IASIS Healthcare as well as several actions. foreign clients. In addition, he has handled or played significant Other notable representations include, among others, roles in a number of sports-related litigations and trials, including successfully defending: CEMEX in putative class actions alleging high-profile matters for the NHL, NFL and NBA. price-fixing of cement and concrete; HarperCollins Publishers in In the transactional arena, Mr. Keyte has represented numerous class actions and government investigations relating to e-books; clients before the DOJ and FTC and in litigated mergers. He also DeBeers in antitrust actions alleging monopolisation and other advises clients regarding antitrust issues, such as compliance anticompetitive conduct; Morgan Stanley in class actions alleging with basic antitrust statutes, including issues relating to bid-rigging relating to municipal derivatives; Arclin in a Robinson- competitor collaborations, unilateral conduct distribution and Patman case alleging price discrimination; Activision Publishing, intellectual property implications. Inc. in an antitrust action challenging the purported tying of Mr. Keyte is an adjunct professor at Fordham Law School, products; NewYork-Presbyterian Hospital in a putative antitrust teaching comparative antitrust law. class action brought by resident physicians; and International For a more detailed biography visit: www.skadden.com/ Paper Company in an antitrust class action alleging price-fixing. professionals/james-keyte. For a more detailed biography visit: www.skadden.com/ professionals/paul-m-eckles.

With approximately 1,600 attorneys in 23 offices on five continents, Skadden serves clients in every major financial centre. For more than 60 years, Skadden has provided legal services to the corporate, industrial, financial and governmental communities around the world in a wide range of high-profile transactions, regulatory matters, and litigation and controversy issues. Our clients range from small, entrepreneurial companies to the largest global corporations. Skadden’s Antitrust and Competition Group is a global leader in its field. Chambers Global: The World’s Leading Lawyers for Business recognises Skadden as one of the top-tier firms in the area of antitrust and competition. Skadden’s European competition law practice advises and represents clients on a wide variety of cutting-edge EU competition law issues, including conduct cases (abuse of dominance proceedings under Article 102 TFEU and cartel proceedings under Article 101 TFEU) as well as mergers and acquisitions. Our attorneys work closely with in-house counsel to advise on compliance and defend against enforcement actions brought by the European Commission or Member State authorities and, where necessary, represent clients in appeals before the European courts.

ICLG TO: CARTELS & LENIENCY 2015 WWW.ICLG.CO.UK Disclosure and Protection of Evidence in Light of the Ingrid Vandenborre Damages Directive and Recent EU Case Law

Skadden, Arps, Slate, Meagher & Flom LLP Thorsten C. Goetz

1. Introduction items of evidence” (Preamble, recital 15). However, this statement of principle is circumscribed by a number of conditions and exceptions. On 26 November 2014, the European Parliament and the Council adopted Directive 2014/104/EU on certain rules governing actions First, the claimant must present a reasoned justifcation containing for damages under national law for infringements of the competition reasonably available facts and evidence suffcient to support the law provisions of the Member States and of the European Union plausibility of its claim for damages (Article 5.1). Disclosure can (the “Damages Directive”). The Damages Directive entered into thus only be ordered if the claimant has made a plausible assertion, force on 26 December 2014 and Member States need to implement on the basis of facts reasonably available to the claimant, that it has it in their legal systems by 27 December 2016. suffered harm that was caused by the defendant. While the claimant does not have to specify individual items of evidence, the claimant is The Damages Directive contains detailed provisions relating to under a duty to specify items or categories of evidence “as precisely the disclosure of evidence in actions for damages before national and as narrowly as possible” (Article 5.2). The Damages Directive courts that seek to strike a balance between a claimant’s right to is unambiguous that “fshing expeditions”, i.e. non-specifc or overly access evidence in support of its private damages claim and the broad disclosure requests, should be prevented. Requests for the protection of leniency programmes through the distinct treatment general disclosure of documents in a competition authority’s fle of material provided to competition authorities in the course of the or of documents submitted by the defendant in the administrative administrative investigation pursuant to an immunity or leniency procedure would therefore not be permissible under Article 5.2 of application. the Damages Directive. The Damages Directive is without prejudice to rules and practices Second, the claimant’s obligation to precisely and narrowly on public access to documents under Regulation (EC) No 1049/2001 circumscribe the evidence in its disclosure request is mirrored by an (the “Transparency Regulation”). However, the EU courts obligation on the national court to order the disclosure of evidence have consistently confrmed that a different treatment of requests only to the extent that disclosure would be proportionate, taking into for access to the fle under the Transparency Regulation would account: (i) the extent to which the claim or defence is supported undermine the protection of commercial interests or the purpose by available facts and evidence justifying the request to disclose of the European Commission’s (the “Commission”) investigations, evidence; (ii) the scope and cost of disclosure, especially for any thereby signifcantly restricting the use of the Transparency third parties concerned, including preventing non-specifc searches Regulation as an alternative path for access. The Damages for information which is unlikely to be of relevance for the parties in Directive, and national legislation implementing its provisions, will the procedure; and (iii) whether the evidence the disclosure of which therefore likely constitute the most effcient and reliable basis for is sought contains confdential information, especially concerning private damage claimants to obtain access to the Commission’s fle. any third parties, and what arrangements are in place for protecting At the same time, the General Court has affrmed the Commission’s such confdential information (Article 5.3 (a)-(c)). policy of disclosing more detail in its public decisions, thereby Third, as further discussed below, where national courts order supporting an alternative or complementary source of evidence in evidence that is included in the fle of a competition authority, actions for damages. i.e. the Commission’s or a national competition authority’s administrative fle, specifc rules apply that seek to protect certain 2. The Damages Directive’s Provisions categories of information from disclosure. Importantly, those rules on the Disclosure and Protection of are not restricted to orders directed against the competition authority Evidence itself, but also to disclosure orders against private parties, e.g. the defendant in a private damages action, in relation to copies of those same fle documents that are in the possession of the private party. The right to proportional disclosure The protection of specifc categories of evidence As a matter of principle, and based on the observation that “competition law litigation is characterized by an information As a general rule, complementing the proportionality requirement asymmetry”, the Damages Directive acknowledges a general right in Article 5.3, Article 6.4 of the Damages Directive stipulates that for a claimant “to obtain the disclosure of evidence relevant to their when assessing the proportionality of an order for disclosure of claim, without it being necessary for them to specify individual iclg to: cARtElS & lENiENcY 2016 www.iclg.co.uk 1 Skadden, Arps, Slate, Meagher & Flom LLP Disclosure & Protection of Evidence

evidence that is in the fle of a competition authority, the national also pending an appeal. Given that a competition authority may court shall consider: (i) whether the request has been formulated have to reopen an investigation after a successful appeal, there are specifcally with regard to the nature, subject-matter or contents good arguments that a temporary exemption from disclosure should of documents submitted to a competition authority or held in the be considered to apply throughout the entire appeal procedures until fle thereof, rather than by a non-specifc application concerning the competition authority’s decision has become fnal. documents submitted to a competition authority; (ii) whether the Importantly, even after the competition authority has closed its request relates to an action for damages before a national court; and proceedings, the disclosure of information that qualifed for the (iii) the need to safeguard the effectiveness of the public enforcement temporary exemption from disclosure would still be subject to the of competition law (Article 6.4 (a)-(c)). The latter consideration general requirement of proportionality as embodied in Articles 5.3 would allow national courts to take into account general public and 6.4 of the Damages Directive. This applies in particular for enforcement interests (e.g., the interest of encouraging undertakings the national court’s obligation to assess the “need to safeguard the to cooperate with the investigation and volunteer incriminating effectiveness of the public enforcement of competition law” (Article evidence pursuant to a leniency programme). 6.4 (c)). To that end, the balancing exercise introduced in the More specifcally, the Damages Directive provides for an increased Pfeiderer and Donau Chemie judgments would apply to information level of protection for certain categories of documents. that would no longer qualify for the temporary exemption from disclosure. Indeed, in its Opinion dated 5 May 2014 submitted to the English High Court in the context of the MasterCard litigation Absolute exemption from disclosure (C(2014) 3066 fnal – Opinion of the European Commission – Interchange fee litigation before the judiciary of England & Leniency statements and settlement submissions beneft from an Wales: Wm. Morrison Supermarkets plc and Others v. MasterCard absolute exemption from disclosure, i.e. national courts cannot Incorporated and Others), the Commission proposed that as regards order disclosure at any point in time (Article 6.6). Leniency materials voluntarily provided to the Commission, such as replies to statements only comprise the corporate statement, i.e. the oral Statements of Objections, it is for the national courts to assess on a or written information provided voluntarily by, or on behalf of, case-by-case basis whether there are overriding reasons for refusing an undertaking or natural person to the competition authority the disclosure of such documents. In the Commission’s view, the specifcally for the purpose of obtaining immunity or a reduction of disclosure of replies to a Statement of Objections may not be liable fnes under the Commission’s or a national competition authority’s to deter the undertakings under investigation from cooperating with leniency programme. Pre-existing information, i.e. information that competition authorities, although there may be a need to protect the exists irrespective of the proceedings of a competition authority, is confdentiality of commercially sensitive third-party information not protected by the absolute ban of disclosure (Preamble, recital in the context of disclosure. The MasterCard case fle included 16). The exemption from disclosure does, however, extend to literal a substantial volume of third-party business secrets that required quotations of a leniency statement or a settlement submission in protection. The Commission therefore considered that the national other documents (Preamble, recital 26). court is obligated to take such measures, as appropriate, to protect Importantly, by categorically exempting leniency statements and third-party confdential information, e.g., through a confdentiality settlement submissions from disclosure, the Damages Directive ring or further redactions. inherently takes the position that the CJEU’s rulings in Pfeiderer (Case C-360/09) and Donau Chemie (Case C-536/11), that required national No exemption from disclosure courts to balance interests on a case-by-case basis when assessing the scope of disclosure, do not apply to those two categories of documents. The Pfeiderer and Donau Chemie judgments expressly noted that the The evidence that does not fall in any of the categories qualifying national court’s competence to conduct a balancing exercise derives for absolute or temporary exemption from disclosure may be the from the “absence of EU rules governing the disclosure of documents subject of an order for disclosure by national courts at any time for the purpose of antitrust damages actions” (see, e.g., Donau Chemie, (Article 6.9). This category of information includes “pre-existing at paragraph 25). Following the entry into force of the Damages information”, such as emails and minutes of meetings, even if Directive, including “EU rules governing the disclosure of documents submitted in the context of an immunity or leniency application. for the purpose of antitrust damages actions” that were absent at the However, even for that category of information, the general rules time of the Pfeiderer and Donau Chemie judgments, the categorical on proportionality of the disclosure and the specifcity of the exemption of leniency statements and settlement submissions now is disclosure request apply. Arguably, the requirement introduced by provided for in EU rules. the Pfeiderer and Donau Chemie rulings that the interests for and against disclosure are balanced also fnd application in relation to these documents. Temporary exemption from disclosure

Use restrictions Information that was prepared by a party specifcally for the proceedings of a competition authority (e.g., a party’s responses to data requests or replies to a Statement of Objections or Letter of The Damages Directive provides for additional safeguards in the Facts), information that the competition authority has drawn up and form of use restrictions imposed on parties that obtained evidence sent to the parties in the course of the proceedings (e.g., the Statement through access to a competition authority’s fle. For example, of Objections or Letter of Facts), and settlement submissions that evidence falling within the scope of the absolute or temporary have been withdrawn beneft from a temporary exemption from exemption from disclosure, and which is obtained solely through disclosure, i.e. these categories of evidence can be ordered to be access to the fle of a competition authority, is deemed inadmissible disclosed by a national court only after the competition authority, or otherwise protected under applicable national rules (Articles 7.1 by adopting a decision or otherwise, has “closed its proceedings” and 2.2; in the case of documents that are subject to a temporary (Article 6.5). It is not entirely clear from the text of the Damages exemption from disclosure, the additional safeguard applies also Directive whether the temporary exemption from disclosure applies only until the competition authority has closed its proceedings).

2 www.iclg.co.uk iclg to: cARtElS & lENiENcY 2016 Skadden, Arps, Slate, Meagher & Flom LLP Disclosure & Protection of Evidence

of the documents in a fle. In its reasoning, the CJEU expressly Penalty provisions referred to the restrictive fle access rules under the specifc antitrust legislation, i.e. Regulations 1/2003 and 773/2004, which would need The Damages Directive’s rules on disclosure are reinforced by the to be respected in the interpretation of the general fle access rules penalty provision of Article 8, which requires Member States to set forth in Regulation 1049/2001. The CJEU’s ruling confrms ensure that national courts are able to effectively impose penalties on that Regulation 1049/2001 does not take primacy over Regulation “parties, third parties and their legal representatives” in the event 1/2003 or 773/2004 and that each of the regulations must be applied of: (i) their failure or refusal to comply with the disclosure order of in a manner compatible with the other and which enables them to be any national court; (ii) their destruction of relevant evidence; (iii) applied consistently. The same reasoning arguably applies also in the their failure or refusal to comply with the obligations imposed by context of the disclosure rules of the Damages Directive. a national court order protecting confdential information; and (iv) However, the general presumption against disclosure under the their breach of the limits on the use of evidence provided for in the Transparency Regulation does not rule out the possibility of showing Damages Directive. that disclosure of a specifc document is necessary and therefore not With respect to the “destruction of evidence”, the Preamble of the covered by the general presumption against disclosure, or that there Damages Directive suggests that the relevant point in time when is an overriding public interest in disclosure of the document. evidence should be preserved is when “a claim for damages is In its recent Axa judgment of 7 July 2015 (Case T-677/13), the General initiated or … an investigation by a competition authority is started” Court held that the Commission’s systematic deletion of references to (Preamble, recital 33). leniency documents, namely submission dates and document names, Although Member States are free in determining what penalty to could not be justifed by a general presumption that all information apply in relation to breaches of the disclosure requirements, the relating to leniency applications must be protected from disclosure in Damages Directive prescribes that the penalties shall include the order to safeguard the effcacy of its leniency programme. Referring possibility “to draw adverse inferences”, such as presuming an issue to the CJEU’s judgment in Donau Chemie, the General Court held to be proven or dismissing claims or defences in whole or in part, that the Commission has to assess, on a case-by-case basis, what and the possibility to order the payment of costs (Article 8.2). information in its fle index must be disclosed, in particular where – as in the case at issue – the claimant already commenced an action for Amendments to Regulation 773/2004 and Commission damages before a national court for damages allegedly incurred by the Notices cartel that was investigated by the Commission. While disclosure of the Commission’s fle index, including On 3 August 2015, the Commission adopted amendments to references to submitted leniency documents, falls far short of Regulation 773/2004 and four related Notices (Access to the File, obtaining access to the actual documents on the fle, a claimant may Leniency, Settlements, Cooperation with National Courts) aimed be able to use the Commission’s fle index to substantiate disclosure at aligning the Notices with the Damages Directive. In addition, requests in damages actions before national courts, as required also with respect to the File Access Notice, a provision (§ 9) was by the Damages Directive. added pursuant to which evidence that proved unrelated to the subject matter of the Commission’s investigation may be returned to the undertaking from which they have been obtained. Upon 4. Publication of More Detailed Cartel return, these documents will no longer constitute part of the fle, Decisions thus removing them from the scope of material to be subject to a request for access. This helps ensure that leniency applicants face Another important aspect relating to the disclosure of evidence was no impediments to disclosing information to the Commission, even addressed in two judgments of the General Court of 28 January 2015 when it is not certain whether this information will ultimately be relating to the hydrogen peroxide cartel (Case T-345/13 Akzo Nobel relevant to the infringement identifed. and Case T-341/12 Evonik Degussa). In those judgments, the General Court endorsed the Commission’s new policy of publishing, years after the initial decision, a more detailed decision which included 3. Restricted Access to Evidence Under the information that the Commission, due to the passing of time, considered Transparency Regulation 1049/2001 no longer confdential. Importantly, it also included information based on leniency applications that had been redacted in the public Private damages claimants have frequently tried to invoke the version of the decision so that the information could not be traced back Transparency Regulation as a legal basis for requesting access to to a particular leniency applicant. In relation to the latter point, the documents in the Commission’s administrative fle. Application of General Court confrmed the Commission’s position that it is for the the Transparency Regulation is not specifc to competition law or Commission to balance the effectiveness of its leniency programme competition proceedings but is available to all parties, regardless of against the interest of parties in the disclosure of information contained whether they are involved in the Commission’s investigation. in the Commission’s cartel decisions. The General Court expressly However, application of the Transparency Regulation in relation to held that the specifc protection accorded to leniency statements, fles of competition law proceedings has been consistently rejected by including under the Commission’s Leniency Notice, “relate only to the Commission on the basis of a general presumption that disclosure the disclosure of documents submitted to it voluntarily by undertakings of documents in competition proceedings will undermine the wishing to beneft from the leniency programme and to the disclosure protection of the commercial interests of the companies involved in of statements made by those undertakings in that connection.” In those proceedings and the protection of the purpose of the inspections. other words, the protection may not extend to the content of the In its judgment in EnBW Energie Baden-Württemberg of 27 February submitted information provided that, as the Commission has done, “all 2014 (Case C-365/12P), the CJEU sanctioned the Commission’s information that might permit, directly or indirectly, identifcation of practice and expressly confrmed that the Commission is not under the source of the information communicated to it by the applicant with an obligation to carry out a specifc, individual examination of each a view to benefting from the leniency programme” has been removed in the non-confdential version of the decision. iclg to: cARtElS & lENiENcY 2016 www.iclg.co.uk 3 Skadden, Arps, Slate, Meagher & Flom LLP Disclosure & Protection of Evidence

5. Conclusion Commission’s consistent practice is to invoke a general presumption against a broad or general disclosure of antitrust fle documents to ensure that the specifc provisions of the Damages Directive are The Damages Directive contains carefully balanced rules relating not undermined. However, a damages claimant may be able to to the disclosure of information in the fles of the EU Commission obtain access to documents that can be specifcally identifed, such or national competition authorities. These provide inter alia, as the Commission’s fle index. Moreover, the Commission may that leniency statements are permanently exempt from disclosure increasingly seek to make more detailed information available in whereas pre-existing documents can be disclosed subject to a the public version of its decisions, either immediately or through proportionality and balancing test. While documents can also re-issuance of the decision after a period of time. be made accessible pursuant to the Transparency Regulation, the

Ingrid Vandenborre Thorsten C. Goetz Skadden, Arps, Slate, Meagher & Flom LLP Skadden, Arps, Slate, Meagher & Flom LLP 523 Avenue Louise, Box 30 An der Welle 3 1050 Brussels 60322 Frankfurt Belgium Germany

Tel: +32 2 639 0336 Tel: +49 69 74 220 167 Fax: +32 2 641 4036 Fax: +49 69 133 839 167 Email: [email protected] Email: [email protected] URL: www.skadden.com URL: www.skadden.com

Ingrid Vandenborre is a partner in Skadden’s Brussels offce, Thorsten Goetz is a European Counsel in Skadden’s Frankfurt and focusing on EU and international merger control and competition Brussels offces. Prior to joining Skadden, Mr. Goetz worked in the law enforcement. Ms. Vandenborre repeatedly has been selected London offce of a leading U.K. frm. He has wide-ranging experience for inclusion in Chambers Global: The World’s Leading Lawyers for in European Union and international merger control cases, as well as Business, and she is described as “a technically excellent lawyer who cartel enforcement and abuse of dominance matters. His clients are holds her own at oral hearings, and has excellent contacts”. She from a broad range of industries, including pharmaceutical/life sciences, consistently has been named as a leading practitioner in her feld by chemicals, fnancial services, energy, travel and telecommunications, Chambers Europe and repeatedly has been selected for inclusion in among others. Mr. Goetz advises clients on antitrust aspects of Who’s Who Legal guides in both competition and life sciences. She complex cross-border M&A and joint ventures. He has worked on was chosen as a “Rising Legal Star” for antitrust by Law360 and numerous transactions requiring international antitrust merger control recognised by Global Competition Review on various occasions, approvals both in Europe and worldwide. Mr. Goetz also advises including being profled as a leading antitrust attorney in its 2013 clients in cartel cases, as well as competition law issues relating to “Women in Antitrust” issue, selected for its 2012 “40 Under 40” list vertical agreements and dominance. He has represented clients in and shortlisted in the “Lawyer of the Year – 40 and Under” at the GCR Article 101 investigations in relation to cartels, strategic alliances, Awards 2013. distribution arrangements and other vertical agreements, as well as in Article 102 investigations, both before the European Commission, the European Courts and national competition authorities.

With approximately 1,700 attorneys in 23 offces on fve continents, Skadden serves clients in every major fnancial centre. For more than 60 years, Skadden has provided legal services to the corporate, industrial, fnancial and governmental communities around the world in a wide range of high- profle transactions, regulatory matters, and litigation and controversy issues. Our clients range from small, entrepreneurial companies to the largest global corporations. Skadden’s Antitrust and Competition Group is a global leader in its feld. Chambers Global: The World’s Leading Lawyers for Business recognises Skadden as one of the top-tier frms in the area of antitrust and competition. Skadden’s European competition law practice advises and represents clients on a wide variety of cutting-edge EU competition law issues, including both conduct cases (abuse of dominance proceedings under Article 102 TFEU and cartel proceedings under Article 101 TFEU) as well as mergers and acquisitions. Our attorneys work closely with in-house counsel to advise on compliance and defend against enforcement actions brought by the European Commission or Member State authorities and, where necessary, represent clients in appeals before the European courts.

4 www.iclg.co.uk iclg to: cARtElS & lENiENcY 2016 FROM HYDROGEN PEROXIDE TO COMCAST: THE NEW RIGOR IN ANTITRUST CLASS ACTIONS$

James Keyte, Paul Eckles and Karen Lent

ABSTRACT

In 2009, the Third Circuit decided Hydrogen Peroxide, which announced a more rigorous standard under Federal Rule of Civil Procedure 23(b)(3) for assessing whether a putative class could establish antitrust injury. Earlier this year, the Supreme Court decided Comcast v. Behrend,a case that carries potentially broad implications for both antitrust cases and Rule 23(b)(3) class actions generally. A review of the case law starting with Hydrogen Peroxide and continuing through Comcast and its progeny reveals the new rigor in antitrust class action decisions and suggests what the future may hold, including the type of arguments that may provide defendants the most likely chance of defeating class certiﬁ- cation. After Comcast, rigor under 23(b)(3) can no longer be avoided Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) in assessing all class actions questions, and courts should now apply Daubert fully in the class setting concerning both impact and damages.

$James Keyte, Paul Eckles and Karen Lent are partners in the antitrust group of Skadden, Arps, Slate, Meagher & Flom LLP.

The Law and Economics of Class Actions Research in Law and Economics, Volume 26, 11 63 À Copyright r 2014 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0193-5895/doi:10.1108/S0193-589520140000026002 11 12 JAMES KEYTE ET AL.

Courts should also closely evaluate plaintiffs’ proposed methodologies for proving impact to determine if they apply to each class member. Finally, courts will inevitably have to determine how rigorously to scrutinize experts’ damages methodologies and whether Comcast requires or suggests more scrutiny in assessing common evidence for measuring damages. Keywords: Class certiﬁcation; Daubert; common impact; Comcast JEL classiﬁcations: K41; K21

TABLE OF CONTENTS

Introduction...... 13 Rigor Under 23(b)(3) Can No Longer be Avoided...... 15 Step 1: Apply Daubert Fully in the Class Setting on the Questions of Both Impact and Damages...... 16 Where the Class Action Law on Daubert Has Been and Is Headed . 16 Types of Rule 23(b)(3) Expert Analysis That Should Not Make It Past a Daubert Screen...... 18 Where the Theory of Impact Does Not Match the Theory of Antitrust Injury ...... 18 Where the Methodology Cannot Pass the Most Basic Daubert Academic Requirements...... 20

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Where the Methodology “Presumes” or “Assumes” Impact or a Basis for Calculating Damages...... 21 Other Analyses that “Mask” Potential Individual Questions of Proof ...... 21 “Pricing Structure” or “Correlation Analysis” ...... 22 Methodologies Based on “Average” Impact ...... 23 Methodologies that Ignore Critical Variables and Issues with the But-For World ...... 25 Step 2: Evaluate if the Proposed Methodology for Proving Impact Could be Used by Every Class Member if Applied in Individual Cases ...... 27 Courts Must Perform a “Rigorous Analysis” of an Expert’s Opinion and Resolve the “Battle of the Experts,” Even if an Expert’s Report Survives Daubert ...... 27 From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 13

Defendants Should Attempt to Develop a Factual Record Demonstrating that Plaintiffs’ Expert’s Assumptions about the Actual World Are Incorrect ...... 28 Price Lists and Price Increase Announcements...... 29 Market Characteristics...... 31 Defendants Challenging an Expert’s Opinion Under Rule 23 Should Also Attack Any Unreasonable Assumptions Made about the “But-For” World ...... 32 If Common Questions Predominate for Only Some of the Plaintiffs, the Court Can Sometimes Narrow the Scope of the Proposed Class ...... 34 Step 3: Can Damages be Proven on a Class Wide Basis? ...... 34 Preliminaries A Whole New World after Comcast? ...... 35 Comcast Majority:À The Need for Rigor Regarding Damages. . . 35 Comcast Dissent: Attempting to Fence in the Majority ...... 36 How Comcast Is Being Applied to Class Damages Analyses ...... 37 Greater Rigor Required ...... 37 Cases Vacated and Remanded (GVR) by the Supreme Court 37 Cases Applying Comcast’s Rigor Requirement to Proof of Damages ...... 41 Cases Purporting to Follow the Comcast Dissent...... 43 Ways to Attack Damages Evidence under the New Rigor ...... 45 Does the Plaintiff Provide a Formula to Calculate Damages? . . 45 Are Damages Susceptible to “Mathematical” or “Formulaic” Calculation? ...... 46 Does the Product at Issue Involve Individualized

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Negotiations? ...... 46 Do Individual Plaintiffs Differ in Ways that Make Proof of Damages Highly Individualized? ...... 48 Do Plaintiffs Propose a Damages Formula Clearly Inadequate to Calculate Individual Damages? ...... 49 Conclusion...... 50 Acknowledgments ...... 62 References ...... 62

INTRODUCTION

In the wake of the Third Circuit’s 2009 decision in Hydrogen Peroxide,1 numerous circuit and lower courts adopted a more rigorous standard under 14 JAMES KEYTE ET AL.

Federal Rule of Civil Procedure 23(b)(3) for assessing whether a putative class could establish antitrust injury or “impact” through common proof. And, no doubt, that newly found rigor resulted in a number of decisions rejecting class certification in cases where, on the proffered expert evidence, individual issues of proof of impact would likely predominate.2 Certainly, however, the cases denying class certification on this basis vary quite a bit in identifying the particular failure of evidence under Rule 23(b)(3), ranging from cases that exclude expert testimony at the class stage under Daubert3 to those that, in effect, make findings (for class purposes only) that plaintiffs’ proposed common proof of impact did not overcome the individualized issues identified by defendants’ expert. By contrast, on the subject of common proof of the amount of damages, relatively fewer courts both pre- and post-Hydrogen Peroxide rejected class certification on thisÀ basis alone, although some did whereÀ the proffered methodology clearly could not identify and measure a class member’s damage.4 The pervasive theme of many 23(b)(3) damages opinions is that, in contrast to the essential element of antitrust impact, plaintiffs and their experts had more flexibility in showing how damages would be assessed through common proof, just as in an individual case where proof of the “amount” of damages historically is given more leeway.5 Then, this past year, the Supreme Court decided Comcast Corp. v. Behrend, a Rule 23(b)(3) damages case.6 Read narrowly, Comcast merely stands for the proposition that any proposed common proof of damages under Rule 23(b)(3) must flow solely from the basis of antitrust liability accepted by the court considering class certification. But the Court’s majority opinion has much broader implications both in antitrust cases

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) and in Rule 23(b)(3) actions generally. The opinion itself highlights that the same “rigor” the Supreme Court had earlier found applied to Rule 23 (a) also applied to Rule 23(b)(3), including presumably for assessing theories and evidence of damages. And while the dissent in Comcast did what it could to fence in the implications of the majority opinion attempting to limit it to its holding on the causal link between the basisÀ for liability and damages the Supreme Court itself has remanded cases addressing broader impactÀ and damages issues in light of Comcast.7 Similarly, several cases and commentators have read Comcast to require the same rigor now being applied on impact to be applied in the damages setting, whether on the subject of disaggregation, damages methodologies or simply whether plaintiffs met their burden of proving that common issues of proof concerning damages would predominate over any individual issues of proof.8 From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 15

In contrast to most disclaimers for articles such as these, the objective here is to be reasonably comprehensive in collecting, assessing and categor- izing successful challenges to putative Rule 23(b)(3) direct purchaser antitrust classes since Hydrogen Peroxide and in the wake of Comcast. Moreover, the goal is not only to understand what “rigor” has come to mean in antitrust cases both for experts and plaintiffs generally but also to suggest what the futureÀ may hold given the principles and trajectoryÀ we see. For example, Comcast and the cases in its wake (including those caught midstream) demonstrate that as a practical matter, Daubert could play an increasingly important role at the class certification stage e.g., practitioners are going to have to make more Daubert (or later, inÀ limine) motions to avoid waiver issues. In turn, and given the rigor now clearly mandated by the Supreme Court under Rule 23(b)(3), a body of Daubert class action law will continue to expand as it has already since Hydrogen Peroxide. And, of course, courts will continue to hone analytical frame- works and principles for assessing Rule 23(b)(3) issues including those related to damages to the benefit of both experts andÀ practitioners who must continually adaptÀ to the new rigor. In this evolving legal environment, the body of Rule 23(b)(3) antitrust decisions themselves suggests some clear demarcations of where decisions have come out in the past and where they are likely to come out in the future. The rigor we see from Hydrogen Peroxide through Comcast and beyond suggests a more stepwise analytical framework for assessing class certification under Rule 23(b)(3) beginning with Daubert issues, proceeding to questions surrounding issues of common proof of impact, and finally, if necessary, concluding with a rigorous analysis of proffered classwide

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) proof concerning damages. In the pages that follow, we seek both to identify the types of challenges to class certiﬁcation that defendants should consider making going forward and to suggest where the future analytical framework for assessing class certiﬁcation may eventually settle.

RIGOR UNDER 23(B)(3) CAN NO LONGER BE AVOIDED

The underlying theme of Comcast is that the same (if not more) rigor applied to Rule 23(a) must now apply to Rule 23(b)(3) analysis. Hence, the Court conﬁrmed that at the class certiﬁcation stage, lower courts must “probe behind the pleadings” even if arguments “would be pertinent to the merits determination,” including with respect to expert testimony.9 Going 16 JAMES KEYTE ET AL.

forward, this deﬁnitively precludes any courts from essentially “kicking the can down the road” to avoid addressing “merits” issues at the class stage. Moreover, it in effect conﬁrms that some of the “shortcuts” or presump- tions utilized in past Rule 23(b)(3) cases e.g., based on Bogosian10 will no longer have traction with courts to theÀ extent they were not alreadyÀ limited by Hydrogen Peroxide and other cases.11 These or similar truncated analyses cannot meet the Court’s directive that lower courts have a “duty to take a ‘close look’ at whether common questions predominate over individual ones.”12 Certainly, this is now the state of the law in assessing “predominance” with respect to the subject of “impact.” But, as we describe below, there remains some ambiguity on the scope of rigor that now is to be applied to proof of damages, which courts already are addressing in the post-Comcast environment. While the authority is split, we believe the better view is that the rigor required under Rule 23(b)(3) on the essential element of impact does not simply dissolve when assessing plaintiffs’ proffered expert evidence for calculating damages. Presumably a settled view will emerge, but one cannot rule out another Rule 23(b)(3) case making its way to the Court and perhaps another opportunity for a divided court to limit the scope of class actions damages litigation.

STEP 1: APPLY DAUBERT FULLY IN THE CLASS SETTING ON THE QUESTIONS OF BOTH IMPACT AND DAMAGES

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Where the Class Action Law on Daubert Has Been and Is Headed

In the wake of Comcast, the subject of Daubert’s application at the class certification stage will be a primary focus, for two different reasons. First, as a purely practical matter, the waiver risks of not making a Daubert motion at the class stage are now sufficiently high that we are likely to see a steady flow of such motions in all but the most mundane price-fixing cases. Second, now that both Comcast (implicitly) and Dukes (more directly) have endorsed the use of Daubert scrutiny at the class certification stage,13 it is reasonable to conclude that those decisions requiring or endor- sing a “full” Daubert screen prior to ruling on class certification are likely to emerge as the prevailing standard. When testimony challenged under Daubert relates to Rule 23(b)(3) requirements, to defer a Daubert screen From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 17

until after class certification would effectively remove, or at least severely diminish, the impact of the rigor requirement. This would ensure exactly the kind of provisional, wait-and-see approach that the rigorous approach to class certification rejects. Indeed, it makes little sense to claim rigor is required at the class certification stage but then push off tough (or even easy) Daubert calls for some later procedure in the case. While Dukes suggests that Daubert scrutiny applies at the class certification stage, other courts have held so more explicitly. This was precisely the reasoning employed by the Seventh Circuit in American Honda Motor Co. v. Allen, a case decided before Dukes and Comcast.14 There, purchasers of Honda motorcycles brought suit, alleging a design defect that prevented the front steering assembly from wobbling while riding.15 In support of its class certification motion, plaintiff’s expert presented a report that relied on a standard created by the expert himself.16 Although the District Court acknowledged it had reservations about the reliability of the expert’s standard, it refused to exclude the report and instead certified two classes of motorcycle purchasers.17 The Seventh Circuit reversed, holding that the district court should have performed a full Daubert analysis before certifying the two classes.18 The Seventh Circuit reasoned that if an expert’s information is relevant to proving any of the Rule 23 requirements for class certification, then the district court must resolve any challenge to the reliability of that information.19 A year later, but also before Dukes and Comcast, the Eleventh Circuit followed suit in Sher v. Raytheon Co.20 Relying on American Honda, the court overturned a district court’s certification of a class, holding that the district court’s failure to apply Daubert during its Rule 23 analysis was 21

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) clear error. And, of course, the Supreme Court originally granted certiorari in Comcast for purposes of addressing the role of Daubert at the class certification stage before concluding that defendants had waived their rights by failing to file a motion objecting to the plaintiffs’ expert report.22 Accordingly, we are likely to see the maturing of a clear body of class certification law that is centered around traditional Daubert factors, but plays out in the landscape and language of class certification. Moreover, even past decisions that chose not to apply Daubert fully (or at all), yet clearly highlighted standard Daubert-like flaws in the expert’s qualifications or analyses, are a good resource for assessing the type of class-related Daubert exclusions we are likely to see going forward. These and other can- didates are described below. 18 JAMES KEYTE ET AL.

Types of Rule 23(b)(3) Expert Analysis That Should Not Make It Past a Daubert Screen

With rare exception, class action experts in the antitrust field are “qualified” in terms of their academic backgrounds and professional experience. In fact, there is a fairly robust cadre of antitrust economists that tend to work on the plaintiffs’ side of the bar and, in turn, have their work scrutinized in a number of court opinions. Instead, the problem for many of these economists is that their professional experience in the class action setting often is based largely on the pre-rigor legal framework. In the earlier framework, the validity of allegations was assumed, and courts did not address issues that went to the “merits” until after class certification. Under this framework, at times, the mere assurance that a methodology based on common proof existed or could be developed at some later point in the litigation was enough to support class certification.23 Those days are now offi- cially behind us. Indeed, class experts must now be prepared to defend their work from any number of Daubert perspectives, all of which should be pressed by defendants.24

Where the Theory of Impact Does Not Match the Theory of Antitrust Injury One predictable Daubert motion in light of Comcast is one arguing that the expert has proffered a theory or analysis of impact or damages that does not clearly limit the harm to the exact basis of antitrust liability. In Comcast, cable television customers brought an antitrust action against Comcast and its subsidiaries, alleging that Comcast’s practice of “clustering” resulted in Comcast monopolizing the Philadelphia market. The District

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Court and Court of Appeals both approved certiﬁcation of the class under Rule 23(b)(3). The Supreme Court reversed, holding that class certiﬁcation was improper because the damages model developed by plaintiffs’ expert did not prove that damages could be measured on a classwide basis. Consequently, because individual questions would overwhelm the Court’s damages analysis, it held that plaintiffs did not satisfy Rule 23(b)(3)’s predominance requirement. The case arose out of a Comcast practice called “clustering,” a business strategy intended to concentrate operations within a particular region. In order to implement its strategy, Comcast purchased competing cable providers in the Philadelphia cluster25 and then replaced those providers with Comcast services. As a result of the practice, Comcast’s share of subscribers in the region allegedly jumped from 23.9% in 1998 to 69.5% in 2007. From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 19

Plaintiffs presented four theories of antitrust injury that allegedly affected the entire class.26 The district court rejected all but the “overbuilder deterrence theory,” which posited that Comcast’s strategy reduced competition by deterring “overbuilders” companies that chose to build competing networks in areas where an incumbentÀ cable company already operates. But plaintiffs’ damages model was prepared before the court’s ruling regarding which liability theories were viable and hence relied on all four theories.27 The Supreme Court thus had to consider whether, when calculating damages, the plaintiff could use a model that measured harm based on the original four theories of injury or instead had to use a model based only on the remaining theory of overbuilder deterrence. The Court held that plaintiffs’ model could not meet the predominance requirement of Rule 23(b)(3). Plaintiffs were required to put forth a damages model that was speciﬁc to the overbuilder theory in order to sus- tain the predominance standard. Instead, plaintiffs’ model calculated damages based on the harm caused by “the alleged anticompetitive conduct as a whole” rather than the harm caused solely by overbuilder deterrence. This model could not differentiate between higher prices in general and higher prices attributable to overbuilder deterrence. Based on this ﬂawed methodology, customers who were harmed by anything other than overbuilder deterrence still would be awarded damages. As a result, the trial court would have to investigate whether overbuilder deterrence or entirely different effects were the true cause of individual class members’ harm. Recognizing this, the Court found that individual questions necessarily would predominate over common questions during the damages inquiry, requiring reversal. As anticipated, since Comcast, a number of courts have

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) applied this fundamental principle in either rejecting class certification or having to reassess class certification in light of Comcast.28 The practical implications of Comcast just on the subject of disaggregation are far reaching and potentiallyÀ very complicated in cases where there isÀ a variety of alleged misconduct. Experts are likely to find themselves offering several (or a myriad of) alternative impact and damages scenarios depending on the nature of the allegations. Moreover, the fact that, under Rule 23, class certification is supposed to be addressed early in a case29 puts plaintiffs (and courts having to manage the case) in somewhat of an awkward position: class experts have to account for alternative or completely different bases for antitrust liability often before discovery is over and plaintiffs themselves have settled on their preferred theory of liability. And, on top of this, all of these positions must be taken under a much more rigorous certification standard in light of Comcast. 20 JAMES KEYTE ET AL.

Finally, there also are disaggregation and causation issues that likely will come into play that go beyond the alternative bases for liability. Courts have long struggled with how to address the subjects of causation, impact and damages where plaintiffs’ harm ﬂows in part from lawful or procompetitive conduct.30 Accordingly, not only will class experts have to deal with alternative analysis of alleged misconduct, but courts (and defendants) too are likely to consider whether economic experts can offer common proof of impact and damages that properly disaggregate other causes of plaintiffs’ harm, irrespective of whether it is other alleged misconduct. Moreover, the dissent in Comcast was quite upset that the majority opinion strongly suggests that plaintiffs and their experts must prove precisely “how” the alleged antitrust misconduct actually resulted in higher prices. The dissent argued that it should be sufﬁcient simply to show “that” prices were in fact higher in the geographic market where Comcast had accumulated a high share.31 While these issues are beyond the scope of this article, they surely will be explored further in the lower courts.

Where the Methodology Cannot Pass the Most Basic Daubert Academic Requirements Perhaps the easiest application of Daubert in the class setting is where plaintiff’s expert essentially makes up his or her own analysis for the case i.e., where the expert is not relying on an academically established methodol-À ogy. It is not uncommon for experts to rely on theories uniquely developed within a case and not based on an established methodology, and these situa-

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) tions often warrant a Daubert challenge. For example, in American Honda, plaintiffs’ expert created a report attempting to analyze the oscillations of Honda motorcycles’ steering col- umns.32 The report’s analysis rested on the application of a “wobble decay” standard, which was created by plaintiffs’ expert himself.33 The Seventh Circuit excluded the report under Daubert for three reasons. First, there was no indication that the standard used in the report had been generally accepted by anyone other than plaintiffs’ expert.34 Second, plaintiffs’ expert had no baseline to which he could compare his ﬁnd- ings.35 Third, the expert’s methodology was ﬂawed because it relied on an impermissibly small sample size.36 All of these reasons, the court found, indicated that the report did not pass the most basic Daubert testing requirements.37 From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 21

Where the Methodology “Presumes” or “Assumes” Impact or a Basis for Calculating Damages Another common flaw though experts now are acutely aware of the pitfall is to offer a methodologyÀ that simply ignores the key question: is thereÀ common proof that could be used at trial to establish that the alleged misconduct in fact caused injury to each class member. Proposed expert testimony that avoids this central issue will not survive. For example, in Blades v. Monsanto Co.,38 purchasers of Monsanto’s genetically modified seeds alleged that Monsanto conspired with competitors to inflate the price of their seeds. Plaintiffs moved to certify two classes under Rule 23, but the district court rejected certification of both classes. The Eighth Circuit affirmed the district court’s decision and held that any attempt to analyze the harm caused by the alleged conspiracy would require consideration of individual rather than common questions.39 After describing the nationwide prices of genetically modified seeds, the court concluded that plaintiffs’ expert mistakenly assumed a classwide injury. Rather than rely on this assumption at the class certification stage, the court required plaintiffs’ expert to establish defendants’ alleged antitrust violations on a classwide basis through common proof. However, because the supply-and-demand conditions for the seeds varied so greatly, seed prices fluctuated within individual geographic markets across the country. Moreover, sellers throughout the country utilized heavy and variable discount approaches, which resulted in sellers charging a variety of prices for the seeds. In light of such differences in seed pricing, the court found that plaintiffs could not, through common proof, prove that each class member suffered injury by paying an inflated price for the

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) seeds. In other words, when assessing whether each class member was harmed, the court would have to answer too many individual questions about which farmers were injured due to inflated prices and how significantly they were injured. As such, both groups of plaintiff seed purchasers failed to satisfy Rule 23’s predominance requirement and could not be certified as a class.40

Other Analyses that “Mask” Potential Individual Questions of Proof Separate and apart from these rather obvious defects, a rigorous approach to Rule 23(b)(3) will increasingly compel courts to reject expert methodologies and analyses that tend to avoid or mask potential individualized issues of impact or damages. Below are a few of the primary examples. 22 JAMES KEYTE ET AL.

“Pricing Structure” or “Correlation Analysis.” In vogue for some time was the so-called “pricing structure” analysis i.e., the notion that if prices of differentiated products move together inÀ response to marketplace factors, then a price-ﬁxing agreement also would affect all class members in a similar fashion.41 Such a structural approach, however, is ripe for masking individual issues of proof.42 For example, in In re Plastics Additives Antitrust Litigation,43 all direct purchasers of organotin heat stabilizers (“tins”) and epoxidized soybean oil (“ESBO”) alleged that defendant producers conspired to ﬁx prices. Plaintiffs moved to certify two classes: one comprised of tin purchasers and the other comprised of ESBO purchasers. The plaintiffs’ expert opined that pricing for the relevant products displayed a “structure” in which prices “moved similarly” over time, such that if a conspiracy existed, it would have impacted every purchaser.44 After plotting transaction data on a graph, the expert relied on “visual observation” as the sole proof that prices were moving similarly.45 The court rejected the pricing structure methodology for three reasons. First, as a preliminary matter, some of the graphs presented did not, on their face, show that prices moved similarly throughout the class period.46 Second, the graphs relied on did not include transactional data for all customers and all products. Instead, the graphs considered a very limited sample size: 18 of the 256 tin products at issue, less than two dozen of the 508 tin class members, and less than two dozen of the 503 ESBO class members.47 Third, evidence at trial showed that, at certain times during the class period, prices moved in opposite directions.48 Thus, the court found that the pricing structure model was inadequate to measure antitrust injury on a classwide basis.

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Similarly, in In re Florida Cement & Concrete Antitrust Litigation, the court rejected the plaintiffs’ expert’s correlation analysis as an insufficient basis for showing common proof of impact.49 The court focused on the lack of connection between general pricing structure in the market and the impact of a price-fixing conspiracy, noting that “[p]laintiffs do not explain how the mere fact that average prices move similarly over time in response to general changes in market conditions necessarily implies those prices will also respond similarly to a price-fixing scheme such as the one alleged here.”50 As the court noted, “[t]he implementation of a price-fixing scheme may involve different considerations besides changes in market conditions.”51 Economists similarly have rejected a correlation analysis as an adequate way to determine whether antitrust impact exists.52 As two commentators note, “[t]he conclusion that the ‘prices move together’ is entirely in the eye From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 23

of the beholder, and this subjectivity is exactly what makes this ‘analysis’ non-scientiﬁc.”53

Methodologies Based on “Average” Impact. While regression analysis can be a valuable tool in determining whether common proof of impact exists, a common fatal ﬂaw in plaintiffs’ experts’ regression models is a reliance on averages or average pricing. As the Northern District of Illinois recently explained, quoting an ABA publication:

Sometimes the prices used by economists are averages of a number of different prices charged to different customers or for somewhat different products. Using such averages can lead to serious analytical problems. For example, averages can hide substantial variation across individual cases, which may be key to determining whether there is a common impact. In addition, average prices may combine the prices of different package sizes of the same product or of somewhat different products. When this happens, the average price paid by a customer can change when the mix of products that the customer buys changes even if the price of [no] single product changed.54 À Hence, the problem is that an “average” impact, by deﬁnition, fails to demonstrate that all class members have been impacted whether in a Section 1 or a Section 2 Sherman Act case. À In Bell Atlantic Corp. v. AT&T Corp., for example, plaintiffs alleged that AT&T attempted to monopolize the market for caller-ID services by blocking the free transmission of caller-ID signals over its long-distance network.55 Plaintiffs moved to certify two classes, one comprised of businesses and organizations that purchased AT&T’s long-distance service and a second comprised of businesses and organizations that were actual or poten- 56

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) tial purchasers of caller-ID services for long-distance calls. Because of AT&T’s actions, plaintiffs alleged they were unable to enjoy the substantial efficiency gains and cost savings that come with caller ID.57 Plaintiffs’ proposed damages formula utilized two national averages: the average cost of labor and the average amount of time that class members would have saved per call had caller ID been available.58 Using these averages as a baseline, plaintiffs argued, the court could assess the class members’ harm by taking the difference between plaintiffs’ averages and the national averages.59 The Fifth Circuit denied class certification, finding that plaintiffs did not satisfy Rule 23(b)(3)’s predominance requirement because their method to prove harm could not reasonably estimate the harm suffered by every class member.60 The court found that in order to adequately gauge harm, it had to consider important individualized questions like the varied nature of the businesses that made up the classes and, depending on those businesses, the 24 JAMES KEYTE ET AL.

range of uses for which caller ID could be employed.61 Because plaintiffs’ averaging formula failed to account for such critical and individualized factors, plaintiffs’ measure of harm and eventual damages projections could not be applied universallyÀ to all the businesses in the proposedÀ classes.62 Consequently, the court held that individual questions rather than common questions dominated both the antitrust injury and damages inquiry.63 Similarly, in Reed v. Advocate Health Care, plaintiffs alleged that entities controlling several hospitals in the Chicago area conspired to suppress the wages of their registered nurse (“RN”) employees.64 To prove the conspiracy impacted all class members, plaintiff’s expert proffered econometric models referred to as “wedge” analyses, which compare the actual wages paid by defendants to the wages defendants would have been willing to pay given the actual supply of nursing hours.65 Calling plaintiffs’ expert’s statements “vague and inscrutable,” the court rejected the wedge analysis.66 The court found many problems with the wedge analysis, one of which was averaging.67 While the average wage may have been reduced from the alleged conspiracy, it does not follow that each class member’s wage was actually reduced as a result of the conspiracy.68 Conversely, because the evidence showed a substantial variation in the compensation of individual RNs during the class period, the court found that some members of the class were treated differently than others and thus not affected by the alleged conspiracy.69 In addition, the wedge analysis made no attempt to distinguish between registry nurses and nonregistry nurses despite the fact that registry nurses made up 20% of the class and that registry nurses’ base wages are calculated differently.70 The failure to account for such a dispar-

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) ity, the court found, rendered the model utterly unable to show common impact on a classwide basis. While the court did not exclude the evidence on the grounds that it failed under Daubert, it held that because the method failed to provide a reliable basis for plaintiffs to show common impact, it need not decide whether the wedge analysis passed muster under Daubert.71 In light of Comcast, future courts are much more likely to grant the Daubert motion in these circumstances and reserve on Daubert only where it is a much closer call. These principles apply with equal force outside of the price-ﬁxing context. For example, in In re Wholesale Grocery Products Antitrust Litigation, the plaintiffs alleged that two large grocery wholesalers conspired to allocate territories and customers using an Asset Exchange Agreement (“AEA”).72 The plaintiffs presented several theories of common proof that From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 25

could be used to show impact, including price lists and two methodologies used by their expert: contrary hypothesis theory and a variance test. The court rejected these methodologies as insufficient to show damages on a classwide basis because they improperly relied on averages and oversimpli- fied the individual plaintiff’s price negotiations. Specifically, the court rejected the contrary hypothesis theory because it could not confirm that each class member was impacted:

Here, the contrary hypothesis test does not address the price levels after the AEA and whether each member of the New England Class was in fact charged a supracompetitive price. If accepted as valid, the contrary hypothesis test proves that SuperValu’s presence in New England influenced C & S’s prices throughout that region. However, even accepting that premise, the contrary hypothesis shows nothing about prices for C & S customers after the AEA (and in fact does not analyze upcharges after the AEA at all).73 The court similarly rejected the variance test, which “is a statistical measure of the spread of data calculated by comparing each data point’s value with the average value of the data set.”74 The court found that this methodology “only compares averages.”75 However, “[t]hat profits may have increased on average, does not mean that monopolist profits were extracted from each class member.”76 Rather than accept these theories that assumed all class members were impacted in the same way, the court focused on the fact that “prices are all the result of individual negotiations “influenced by, among other factors,” the size of orders, frequency of orders, and transportation costs.”77 Accordingly, the court denied class certification. Again, the court did not conduct a Daubert analysis, but after Comcast, courts will be much more likely to exclude expert testimony under Daubert where it erro-

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) neously relies on averages and oversimpliﬁed information that is not consistent with the anticompetitive theory of the case.

Methodologies that Ignore Critical Variables and Issues with the But-For World. Finally, courts will continue to be presented with methodologies that, in theory, could pass Daubert, but remain flawed in the face of facts highlighted by the defendants that affect the price paid by individual putative class members. For example, regression models that ignore significant potentially individualized variables have been held to be an inadequate form of common proof. While these factors often exist in price-fixing cases, they are just as likely to appear in vertical nonprice restraint cases or monopolization cases where prices are negotiated on an individualized basis. 26 JAMES KEYTE ET AL.

For example, in In re Live Concert Antitrust Litigation, plaintiffs’ expert presented four types of statistical analyses, one of which was called the “yardstick approach.”78 At the first step of the “yardstick approach,” plaintiffs’ expert calculated the average ticket price for rock concerts in each geographic market. He then calculated the average ticket price for concerts promoted by the defendants in those markets and compared the two average prices. Finding that the concerts promoted by defendant had a higher average ticket price, the expert concluded that: (1) the average higher price was a direct result of the alleged conspiracy and (2) all class members were injured because they paid the higher price.79 The court flatly rejected this method, finding that the “yardstick” comparison ignored other critical factors that determine average ticket prices. Because the yardstick method did not account for other possible explana- tions for the difference in average ticket prices, including artist quality, popularity and venue size, it could not properly estimate impact in the but- for world, and the court excluded the evidence generated from the “yardstick” method.80 Similarly, in In re Cox Enterprises, Inc. Set-Top Cable Television Box Antitrust Litigation, plaintiffs alleged that defendant illegally forced its customers to rent a Cox set-top box in order to gain full access to defendant’s premium cable services.81 As a result of defendant’s alleged tying scheme, plaintiffs claimed they paid supracompetitive prices for the tied product, the set-top box. In order to prove that all class members actually were injured by paying the supracompetitive price, plaintiffs’ expert proposed the “GRS test.”82 The GRS test estimated the overcharged price by comparing the actual price paid by Cox customers to the “but-for” price the price Cox cus- À

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) tomers would have paid absent the alleged conduct. The proposed damages would be calculated by multiplying the overcharge price by the number of set- top boxes rented. Although the court refused to decide whether the GRS test satisfied Daubert,83 it considered the Daubert requirements when evaluating whether the GRS test could be applied to the entire plaintiff class.84 The court particularly found flawed the test’s factor that weighed demand elasticities, faulting the expert for relying on academic estimates that were not Cox-specific to calculate demand elasticities. The court found that in order to estimate demand elasticities, an expert must consider the extent of market competition. And because market competition varies in different parts of the country, the court held that it was improper for plaintiffs’ expert to rely on common market data contained in the academic estimates. Thus, because the GRS test could not estimate demand elasticities unless it considered individualized market data, which it did not do in this case, the court denied certification.85 From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 27

Going forward, courts are likely to apply the rigor required under Rule 23(b)(3) to regression models (or other quantitative analyses) that ignore obvious market facts or variables affecting price that, if addressed, may show a lack of impact to some number of class members. And where such potential is evident on the record, courts are very likely to exclude the expert’s opinion based on a threshold Daubert review without ever reaching the full review of the record that is required where an expert’s testimony is not facially defective.

STEP 2: EVALUATE IF THE PROPOSED METHODOLOGY FOR PROVING IMPACT COULD BE USED BY EVERY CLASS MEMBER IF APPLIED IN INDIVIDUAL CASES

Where a plaintiff’s class action expert survives a Daubert challenge, this by no means suggests that class certification is inevitable or easy. Instead, sur- vival of a Daubert motion is a necessary, but not in and of itself sufficient, condition for class certification.86 In these circumstances, plaintiffs must still meet their burden of showing that the record evidence, including the expert’s analysis and testimony, supports a finding that impact for each class member can be proven through evidence that is common to the class. As described below, the general framework for this inquiry is to (i) assess whether there are potential individualized issues of proof with respect to how the market operates, such as how prices are determined, (ii) understand how plaintiffs purport to show that the “but-for” price Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) i.e., that which would exist without the alleged misconduct can beÀ demonstrated with common proof, and (iii) analyze whether plaintiffsÀ met their ultimate burden of demonstrating that, at trial, the essential element of impact can be established through common evidence. The most common pitfalls that plaintiffs face in making this showing are outlined below.

Courts Must Perform a “Rigorous Analysis” of an Expert’s Opinion and Resolve the “Battle of the Experts,” Even if an Expert’s Report Survives Daubert

As a threshold issue, it is the district court’s responsibility to determine the persuasiveness of admissible expert testimony when necessary to resolve class certiﬁcation issues even if any ﬂaws in an expert’s opinion do not rise to the level of excludabilityÀ under Daubert. Plaintiffs who prevail under 28 JAMES KEYTE ET AL.

Daubert may try to sidestep this burden by emphasizing that courts historically have considered price-fixing conspiracies well-suited for class treatment.87 However, “it does not follow [from Amchem] that a court should relax its certification analysis, or presume a requirement for certification is met, merely because a plaintiff’s claims fall within one of those substantive categories.”88 The court still must determine if the expert’s opinion provides sufficient analysis to warrant certification. In Dukes, the Supreme Court clarified once and for all that a court considering class certification under Rule 23 must look beyond the pleadings and conduct a “rigorous analysis” of the evidence to assess whether plaintiffs’ proposed class meets the requirements of Rule 23.89 As the Dukes Court observed, this rigorous analysis often requires a close look at the merits of the evidence put forth by the plaintiffs:

Frequently that “rigorous analysis” will entail some overlap with the merits of the plaintiff’s underlying claim. That cannot be helped. The class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiff’s cause of action.90 Accordingly, it is the court’s job to resolve any inconsistencies between the opinions of the experts. Signiﬁcantly, most of the problems addressed above that may provide the basis for a Daubert motion, also provide defendants a basis to argue that, regardless of whether the plaintiffs’ expert’s testimony is formally excluded, it is still insufﬁcient to satisfy plaintiffs’ burden in proving that common issues predominate. Thus, for example, while reliance on a “price structure” analysis or “averages” may provide fruitful grounds for a Daubert motion, they also provide a basis for defendants to argue that plaintiffs have failed to meet their burden in demonstrating that Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) they could prove each class member’s claim through common proof. Of course, there certainly was a trend in this direction in the case law well before the Supreme Court offered its own guidance in Dukes and Comcast.91 Dukes and Comcast make clear that this trend is now a requirement for lower courts. Thus, courts must perform a rigorous analysis and weigh the persuasiveness of the experts for both sides even where the expert’s testimony survives a Daubert analysis. À

Defendants Should Attempt to Develop a Factual Record Demonstrating that Plaintiffs’ Expert’s Assumptions about the Actual World Are Incorrect

Even if a plaintiff’s expert uses an accepted methodology that is capable of demonstrating common impact under certain circumstances, that does not From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 29

mean those circumstances exist in every case. Expert’s opinions often mask or ignore inconvenient evidence regarding how a market actually operates, which may include individualized issues as to how prices are determined. The new rigor at the class certification stage opens up opportunities for defendants to challenge these experts’ opinions. In cases where the plaintiffs’ expert’s basic methodology can survive a Daubert challenge, the foundation for defeating class certification will often lie in the quality of the factual record that defendants can muster. Critically, even though plaintiffs bear the burden of proof, defendants are wise not to rely on a critique of the plaintiffs’ evidence, but rather to develop their own evidence demonstrating why the plaintiffs’ proffered methodology does not work in that particular case. Specifically, attacking the plaintiffs’ factual assumptions regarding how the market actually operates depends on the defendants developing a robust factual record through declarations, deposition testimony, analysis of the relevant pricing documents and analysis of the available data.

Price Lists and Price Increase Announcements To meet their burden of proof, plaintiffs commonly attempt to rely upon “price lists” or price increase announcements that at least ostensibly relate to all products purchased by putative class members. Prior to the more recent pronouncements regarding the rigor that courts should apply when determining class certiﬁcation, courts were inconsistent in how to handle such evidence. Even in cases in which defendants argued that prices were set through complex, individualized negotiations, courts often found that a

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) uniform price increase raises the base price at which those negotiations begin and hence lends itself to common proof sufﬁcient for the class certiﬁ- cation stage.92 Conversely, other courts declined to certify classes despite the use of price lists or price announcements when defendants were able to proffer evidence that individual variables impacted the setting of prices.93 More recently, courts have been more willing to take a critical look at the relationship between price lists or price increase announcements and actual prices, recognizing that such evidence is only susceptible to common proof if the price lists or announcements are actually used and followed. For example, in Plastics,94 plaintiffs proffered a collection of defendants’ documents that set forth price lists and price increase announcements for tins and ESBO. Because the plans detailed in the documents applied to all tin and ESBO products, plaintiffs claimed, all class members were impacted when they purchased those products. 30 JAMES KEYTE ET AL.

In response, the defendants offered evidence demonstrating that there were wide variations in the prices actually paid by members of the putative class, including that the prices paid by some purchasers remained constant after the price increase announcements were issued and that prices for at least some putative class members declined during the class period.95 Rather than blindly accepting the plaintiffs’ assertion that all putative class members were impacted by the conspiracy in a manner that could be demonstrated through common proof, the Plastics court considered the record evidence in the case and concluded: “[T]he evidence of record shows that the prices paid by customers did not correspond with Defendants’ price increases. Accordingly, the price lists and price increase announcements cannot serve as common evidence of impact.”96 Similarly, in In re Florida Cement and Concrete Antitrust Litigation, the court considered allegations that defendants conspired to raise the price of concrete by $25. Because the defendants all announced a $25 increase during the summer of 2008, plaintiffs asserted that the defendants’ price increase announcements could serve as common proof of impact. The court disagreed, finding evidence “[t]hat some customers avoided paying price increases or artificially-stabilized prices, is inconsistent with Plaintiffs’ theory of common impact.”97 The court highlighted the flaws in the expert’s uncritical reliance on the price lists:

In fact, Plaintiffs present no empirical analysis of the actual effect of the price increase announcements on individual customers to counter the evidence presented by Defendants, and unlike Dr. Ordover, Dr. Mangum did not conduct any signiﬁcant analysis at the individual customer level to determine whether any price changes were consistent across the putative Class. Thus, Plaintiffs have not shown how the price increase

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) announcements even if they were intended to affect all customers across-the- À board constitute common evidence through which impact on the individual class mem- À bers can be proven.98

Finally, in In re Graphics Processing Units Antitrust Litigation, the court considered allegations of a price-ﬁxing conspiracy for the sale of graphics processors, which are “designed based on the speciﬁc application for which they will be used.”99 The court recognized that “[w]ithin these markets, defendants sold chips and cards of varying performance levels” based on the particular needs of the purchaser.100 The court also highlighted that the graphics processor “products at issue were sold to a variety of customers through a number of distribution channels.”101 The court focused on these complex characteristics that distinguished purchasers (all of whom were putative members of the same class) and From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 31

concluded that price lists were not an accurate way of determining how prices were set in the GPU market:

[O]ver 99.5% of defendants’ revenue during the limitations period came from sales with large wholesale purchasers like Microsoft. … [T]he vast majority of sales were primarily executed after customized negotiations between wholesalers and either defendant. These sales were made without any regard to a price list. As such, defendants’ sales contracts varied significantly depending on the wholesale purchaser. … There is no doubt that a myriad of factors played a role in each large transaction. These factors each influenced the final sales price of each transaction.102 Thus, the court held that these price lists could not serve as common proof of impact for the class. As these cases demonstrate, in situations where the plaintiffs focus on price increase letters or price lists to establish common impact, defendants should make every attempt to determine whether individual customers actually paid prices that corresponded with those published prices. Developing evidence demonstrating that the prices in the real-world market do not correspond to those price lists, or that individuals were able to avoid published price increases, should force the plaintiffs’ expert to come up with something more to show common proof of impact.

Market Characteristics Additionally, plaintiffs’ experts sometimes rely upon “market characteristics” to support their conclusion that impact can be shown through common proof. The critical inquiry is generally whether the record evidence demonstrates that the products are homogeneous.103 Even before the trend

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) toward a more rigorous analysis, courts have long found that when “the distinctions among the [products] offered are substantial[,] the process of assessing impact would be anything but mechanical.”104 But this is an area that has seen increased scrutiny since Hydrogen Peroxide and that trend should continue in the post-Comcast environment. For example, in Plastics, the court recognized that, “in theory,” characteristics such as the homogeneity of the products at issue and defendants’ domination of the relevant market could make a market “vulnerable” to a price-ﬁxing conspiracy.105 However, the court then discussed at length the evidence proffered by defendants demonstrating the heterogeneity of the products at issue: the different quality and strength properties in the defendants’ products affected whether such products were suitable for different types of buyers’ end uses, and customers clearly indicated a preference for the products of one defendant over another.106 These qualities made it 32 JAMES KEYTE ET AL.

impossible for the court to find that market characteristics could serve as common proof of antitrust impact. Similarly, in Florida Concrete,107 the court weighed the evidence offered by both sides and found that “Plaintiffs fail to demonstrate that Concrete is indeed a homogenous product such that it can be used interchange- ably.”108 Importantly, the court rejected the notion that a defendant’s expert must “perform enough quantitative analyses to prove that Concrete is not interchangeable and that the market characteristics are not such that the impact of the purported conspiracy would have been unavoidable.”109 Rather, the court highlighted that “it is not Defendants’ burden to prove that the impact of the alleged conspiracy is not susceptible to common proof; to the contrary, it is Plaintiffs’ burden to prove the impact is susceptible to proof by common evidence.”110 Additional factual issues for defendants to consider include whether prices are set by a single individual or by a large group of individuals operating in the field. The same types of individual issues that can undermine an employment class action in which plaintiffs are challenging employment decisions made by scores of individual managers across different offices (i.e., Dukes) can exist in an antitrust case in which purchases or sales are handled by scores of individual employees. Even where prices are ostensibly dictated by a single individual, defendants should scrutinize whether the individuals on the front lines actually followed whatever policies or directives were issued. A related issue is whether the relevant industry is a “relationship” driven business in which sales and prices are determined more by the relationship of the individuals involved than by any directives received from corporate headquarters.111

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Consequently, where there is record evidence of nonprice factors, such as perceived differences in service, quality, timeliness of deliveries, and willingness to accept returns, that may impact plaintiffs’ purchasing decisions, defendants should challenge the notion that a market for homogeneous products is “susceptible” to common proof based on market characteristics alone.

Defendants Challenging an Expert’s Opinion Under Rule 23 Should Also Attack Any Unreasonable Assumptions Made about the “But-For” World

Defendants should attack any unreasonable assumptions built into an expert’s models proffered to show common impact. It is the task of an antitrust plaintiff to “establish a ‘but for’ baseline a ﬁgure that would show À From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 33

what competitive prices would have been if there had been no antitrust violations.”112 Construction of the “but-for” world has never been a perfunc- tory task that plaintiffs (or courts) can take for granted; rather, damages can only be calculated “by comparing to that baseline what the actual prices were during the challenged period.”113 As the trend toward greater rigor in class certification continues, it will be essential that courts evaluate whether the plaintiffs’ “baseline” matches up with what really happened in the relevant market or what would likely happen if the alleged misconduct did not exist. In straightforward instances of an alleged price-fixing conspiracy in a market for a homogeneous product where prices are set by price lists, an expert may be able to construct an economic model that does not contain many or any suspect assumptions about “but-for” pricing and output in theÀ relevantÀ market. In these scenarios, if an expert uses an acceptable methodology and applies it properly to the facts of the case, that expert likely will have offered acceptable opinions that could prove antitrust impact. However, in more complicated cases where pricing is negotiated on an individualized basis or where the alleged misconduct involves nonprice- related conduct expertsÀ are likely to make certain assumptions about the way the marketÀ operates that can skew their analyses. This is especially true in those cases where individual plaintiffs may have complicated specifi- cations, preferences, or other characteristics that make it difficult to extra- polate broad principles for the behavior of the entire class. In such scenarios, experts cannot merely “assume” that all plaintiffs will be impacted in the same way.114 Indeed, such an assumption is grounds for 115

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) the denial of class certiﬁcation. Similarly, as made clear in Comcast, experts can no longer make assumptions that all theories of antitrust liability impact the plaintiffs in the same way.116 Therefore, an expert fails to establish an acceptable “but-for” world if that expert cannot distinguish between supracompetitive prices resulting from antitrust misconduct and “prices whose level above what an expert deems ‘competitive’ [is] caused by factors unrelated to an accepted theory of antitrust harm.”117 Defendants should highlight instances where it is likely that the plaintiffs’ expert is lumping together several theories of harm for example, when plaintiffs include in their complaints marketÀ allocation allegations or allegations related to other nonantitrust violations (such as misrepresentationsÀ or deceptive practices, or even competitive behavior). In such scenarios, if the court ﬁnds that one or more of these theories is not a viable antitrust theory of liability, then the expert’s 34 JAMES KEYTE ET AL.

model must be able to disaggregate those theories of antitrust impact to support certiﬁcation of the class.118

If Common Questions Predominate for Only Some of the Plaintiffs, the Court Can Sometimes Narrow the Scope of the Proposed Class

Even if the court does not deny class certification on the basis of the aforementioned individualized questions of fact, under the rigorous standard, a court may consider narrowing the proposed class based on the predominance requirement. For instance, the In re GPU court denied certification of the plaintiffs’ proposed class but granted certification for a more limited class of direct purchasers.119 The court focused on the fact that some of the proposed class members faced a nonnegotiable price, while other proposed plaintiffs negotiated prices individually with the defendants.120 Accordingly, the court held that “[p]roof that defendants conspired to fix [list] prices would hardly prove that defendants also conspired to fix the non-list prices for the transactions entered into with absent wholesale purchasers.”121 Defendants can ask the court to exclude customers who typically negotiate prices, if defendants find that some plaintiffs did make purchases solely based on list prices. While a court is generally under no obligation to rede- fine the class, it may be tempted to adopt this approach if it is particularly difficult to resolve the predominance issue. Moreover, in situations where defendants cannot make the aforementioned arguments on the basis of all members of the class, narrowing the proposed class to list-price customers

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) can sometimes provide a way reducing the size of the potential damages in a price-ﬁxing case.

STEP 3: CAN DAMAGES BE PROVEN ON A CLASS WIDE BASIS?

The most controversial issue in the wake of Comcast is whether the “rigor” under Rule 23(b)(3) referenced by the majority opinion applies to class action damages methodologies and proof generally or whether, instead, the case can be limited to its facts i.e., the particular ﬂaw in not disaggregat- ing potential causes of harm thatÀ are not the basis of antitrust liability (discussed above). From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 35

We discuss below both the debate in the lower courts (and academic cir- cles) on what Comcast means going forward, including how class action damages analysis is or will be affected in those cases where courts apply the full rigor of Rule 23(b)(3) in assessing whether damages for each class member can be proven at trial through common proof.

Preliminaries A Whole New World After Comcast? À Comcast contains potentially wide-ranging implications for the subject of common proof of the amount of damages. At a minimum, Comcast stands for the proposition that any proposed common proof of damages under Rule 23(b)(3) must ﬂow from the basis of antitrust liability proposed by the plaintiffs and accepted by the court. Thus, an expert’s model calculating damages that failed to disaggregate potential causes of harm from those harms that the lower court accepted as a possible basis for antitrust liability was not consistent with plaintiffs’ liability case.122 In other words, the model did not link antitrust harm to damages, dooming the plaintiffs’ class certiﬁcation motion. But the majority’s opinion suggests a new rigor for scrutinizing class action damages evidence well beyond the narrow proposition that a damages model must link damage to harm. The opinion suggests that the rigor Dukes required courts to apply under Rule 23(a) applies now with equal, if not greater, force under Rule 23(b)(3). This would include, presumably, rigorous analysis in assessing common evidence regarding the measurement of damages.

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) The Comcast dissent claims the majority’s opinion could not possibly be read to require damages be measureable for each class member. But some courts, and commentators, do not see it the dissent’s way at all; sufﬁce it to say that what the future holds is far from certain.

Comcast Majority: The Need for Rigor Regarding Damages In addition to the uncontroversial holding that a damages model must link the theory of harm to impact, the Comcast majority applied additional rigor in evaluating common proof of the amount of damages. The Court started from the observation that the predominance criterion under Rule 23(b)(3) demands even stronger proof than Rule 23(a),123 and noted that Rule 23(b)(3) itself is “an adventuresome innovation designed for situations in which ‘class-action treatment is not as clearly called for’.”124 It held that a close look, including at arguments pertinent to the merits, is required to 36 JAMES KEYTE ET AL.

satisfy predominance.125 Under this rigorous standard, a proposed class cannot show Rule 23(b)(3) predominance in the damages context where “[q]uestions of individual damage calculations will inevitably overwhelm questions common to the class.”126 The Court added,

even if the model had identiﬁed subscribers who paid more solely because of the deterrence of overbuilding, it still would not have established the requisite commonality of damages unless it plausibly showed that the extent of overbuilding (absent deterrence) would have been the same in all counties, or that the extent is irrelevant to effect upon ability to charge supra-competitive prices.127 Applying this rigorous standard to the damages model proffered by plaintiffs’ expert, the Court held that where the model measures “damages” that are not attributable to an accepted theory of liability, “the model cannot possibly establish that damages are susceptible of measurement across the entire class for purposes of Rule 23(b)(3).”128 The Court also identiﬁed three other situations where individualized damages might overwhelm common questions: (1) where plaintiffs’ methodology cannot prove harm for a large portion of the class;129 (2) where the “permutations” of damages are “nearly endless”;130 and (3) where the proposed damages methodology is speculative.131

Comcast Dissent: Attempting to Fence in the Majority The dissent, written by Justice Ginsburg and Justice Breyer, with Justices Sotomayor and Kagan joining, argued that the Court’s holding with respect to damages issues should be limited to the particular facts of the case. These four justices claimed that “[t]he decision should not be read to require, as a prerequisite to certiﬁcation, that damages attributable to a Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) classwide injury be measureable ‘on a class-wide basis.’ ”132 According to the dissent, plaintiffs must demonstrate and courts must ﬁnd that questions of law or fact common to class members predominate over questions affecting only individual members, but predominance does not demand commonality as to all questions.133 The dissent claimed that,

[r]ecognition that individual damages calculations do not preclude class certification under Rule 23(b)(3) is well nigh universal … In the mine rune of cases, it remains the ‘black letter rule’ that a class may obtain certification under Rule 23(b)(3) when liability questions common to the class predominate over damages questions unique to class members.134 Indeed, the dissent argued that if a class action is an efficient mechanism for resolving common liability questions, “the predominance standard is generally satisfied even if damages are not provable in the aggregate.”135 From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 37

The dissent’s analysis with respect to damages is inconsistent with the rigor reﬂected in the majority opinion. Indeed, in the months since the case was decided, the Court has remanded several Circuit level antitrust cases that lacked common proof of damages “in light of Comcast.” Lower courts also have applied the Comcast court’s rigorous analysis to proof of damages in the class setting, although the responses to Comcast are far from uniform.

How Comcast Is Being Applied to Class Damages Analyses

Greater Rigor Required Cases Vacated and Remanded (GVR) by the Supreme Court. Following Comcast, the Supreme Court has granted certiorari on, and vacated and remanded three cases “in light of Comcast.” In two of these cases, In re Whirlpool Corp. Front-Loading Washer Products Liability Litigation136 and Butler v. Sears, Roebuck and Co.,137 the likely cause for remand appeared to be a lack of rigor in analyzing proof of damages.138 In Whirlpool, consumers brought a putative class action against Whirlpool in connection with sales of front-load washing machines. Plaintiffs alleged breach of warranty, negligent design, and negligent failure to warn. According to the plaintiffs, Whirlpool’s front-load washing machines suffered from a design defect that made them prone to develop mold and mildew.139 Plaintiffs alleged a class comprised of current Ohio residents who purchased one of the speciﬁed washing machines for perso- nal, family or household purposes.140

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Whirlpool opposed class certiﬁcation on the following grounds: • The vast majority of washing machine owners had not experienced a mold problem. • Whirlpool made dozens of changes between 2002 and 2009 to increase customer satisfaction and reduce service costs. • Washers owned by class members were built on two different (although substantially identical) platforms, involved 21 different engineering models and spanned 9 model years. • Consumer laundry habits are so diverse that plaintiffs presented individual liability questions.141 The district court held that common questions of liability predominated and certiﬁed the class under Rule 23(b)(3). It noted that “ ‘[n]o matter how individualized the issue of damages may be, these issues may be reserved 38 JAMES KEYTE ET AL.

for individual treatment with the question of liability tried as a class action.’ ”142 In a limited discussion, the Court noted that division of the class into subclasses would be sufficient to overcome any individualized issues with respect to damages.143 In Butler, consumers brought a class action against Sears regarding its sale of allegedly defective Whirlpool washers based on warranty laws in six states.144 The plaintiffs made claims regarding two alleged design defects: mold claims, similar to those made in Whirlpool, and control unit claims.145 The control unit claims involved allegations that a soldering defect caused control units to mistakenly identify a serious error with the washing machine and automatically shut down the machine mid-cycle.146 Plaintiffs sought to certify two classes: a class with “mold” claims and a class with “control unit” claims. For both classes, the court stated that there were common questions on liability. For the mold class, the question was, “were the machines defective in permitting mold to accumulate and generate noxious odors”; for the control unit class, the question was “whether the control unit was indeed defective.”147 The court acknowledged that individual questions existed on the issue of the amount of damages and, at least for the mold class, that most members did not experience a problem. The district court certified the control unit class and denied certification for the mold class. The Seventh Circuit affirmed certification of the control unit class and reversed denial of certification for the mold class,148 with Judge Posner favoring an “efficiency” approach to the predominance inquiry. According to the opinion, efficiency would be served through a single proceeding to resolve a central liability issue whether the washing machines were defective. This liability proceedingÀ “could be

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) followed by individual hearings to determine the damages sustained by each class member.”149 In both Whirlpool and Butler, the lower courts failed to apply rigor to their analysis of proof of damages, instead suggesting that it could be resolved later. At the same time, the lower courts acknowledged that numerous plaintiffs within the certiﬁed classes did not suffer any direct harm, and even those plaintiffs claiming actual harm suffered “nearly endless permutations” of alleged harm. This is precisely the type of harm the Court found inadequate to establish classwide proof of damages in Comcast. Although the Court did not explain why it was remanding these cases in light of Comcast, most commentators reasoned that it must be based on the greater rigor courts must use to assess theories of damages. For example, a Bloomberg Class Action Litigation Report suggested that the From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 39

Supreme Court’s remand of Whirlpool “could prompt additional scrutiny of the lack of injury for the majority of class members” as well as the 21 different types of washing machines purchased by class members. In both cases, any damages evidence must account for intraclass differences under Comcast.150 In another article commenting on the remand of both Whirlpool and Butler, the author noted:

In both Comcast and Wal-Mart[, the] linchpin to certification is assuring that determination of whether defendant’s conduct caused injury to each class member can be made classwide and without resort to individualized assessments of each member’s circumstances … The major infirmity in Butler and Whirlpool is that each overlooked myriad permutations among hundreds of thousands of purchasers of different product models concerning the presence of mold and mildew, their causes, amounts of any resulting damages, customers’ care of washers, whether requests for warranty service were made and timely, and defendants’ responses to warranty claims.151 He suggested, drawing on the language from Comcast, that when “permutations” of plaintiffs’ alleged damages “are nearly endless,” damages may not be capable of classwide measurement and instead may devolve into “labyrinthine individual calculations.” The article also took specific issue with Judge Posner’s “efficiency standard” in Butler. The Butler court admitted that both whether each class member suffered any damages and, if so, the amount of those damages were individualized questions. The author argues the single common issue identified in Butler, “were the machines defective,” cannot satisfy the Rule 23 test.152 Another article concurred: “[T]he Butler court’s treatment of the need for individual damages trials seems flatly inconsistent with the 153

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Comcast court’s statements on the need for proof on a class-wide basis.” Surprisingly, the Supreme Court’s directive to reconsider the cases “in light of Comcast,” and the views of most commentators, largely fell on deaf ears. In Whirlpool, the Sixth Circuit subsequently issued an opinion holding that the front-load washing machine class was properly certiﬁed notwithstanding Comcast. Relying on the Comcast dissent, the Sixth Circuit viewed the Court’s decision of limited relevance to the case before it because the district court “certiﬁed only a liability class and reserved all issues concerning damages for individual determination.”154 In addition to essentially ignoring the GVR order on damages issues, the Sixth Circuit’s ruling over- looks the implications of Comcast for variations of injury within a putative class e.g., the odor problem that manifested in only a small percentage of the washingÀ machines.155 Yet this only highlights that putative class members did not “suffer the same injury.”156 40 JAMES KEYTE ET AL.

Shortly thereafter, in Butler, the Seventh Circuit followed suit and rein- stated the ruling that the classes could be certified, even in light of Comcast. The Butler opinion continued to rely on a standard centered around judicial efficiency, rather than predominance, as the requirement for class certification under Rule 23(b)(3). It also fails to grasp the implications of Comcast; indeed, at one point the court asks, puzzled, “[b]ut if we are right that this is a very different case from Comcast, why did the Supreme Court remand the case to us for reconsideration in light of that decision?”157 Further, when the Seventh Circuit does address individualized damages issues, it sets up what appears to be a faulty dichotomy:

If the issues of liability are genuinely common issues, and the damages of individual class members can be readily determined in individual hearings, in settlement negotiations, or by creation of subclasses, the fact that damages are not identical across all class members should not preclude class certiﬁcation. Otherwise defendants would be able to escape liability for tortious harms of enormous aggregate magnitude but so widely distributed as not to be remediable in individual suits.158

This result, Judge Posner warns, “would drive a stake through the heart of the class action device.”159 But this conclusion only follows if one asserts that, after Comcast, satisfying predominance requires that plaintiffs suffer identical damages a premise neither the Comcast majority nor the Butler defendants ever suggest.À Instead, the Comcast majority implies that where endless permutations of plaintiffs’ damages would result in “labyrinthine individual calculations,” individual damage issues may predominate over common questions. The Sixth and Seventh Circuits may not believe such

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) labyrinthine damages calculations are necessary in these cases, but their failure to properly address the issue is difficult to reconcile with the Supreme Court’s directive that the cases be reconsidered “in light of Comcast.” More recently, however, the D.C. Circuit adopted the approach many anticipated the Sixth and Seventh Circuits would take. In In re Rail Freight Fuel Surcharge Antitrust Litigation, the D.C. Circuit exercised its jurisdiction over an interlocutory appeal from a class certification decision and vacated and remanded on certification grounds, in part due to the Comcast ruling.160 In that case, shippers brought a class action against the four major freight railroads alleging a price-fixing conspiracy regarding the railroads’ imposition of rate-based fuel surcharges.161 Plaintiffs sought to certify a class of shippers who paid the purportedly inflated fuel surcharges.162 Plaintiffs’ expert prepared two regression models that would work in conjunction to demonstrate predominance under Rule 23(b)(3).163 The From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 41

“common factor model[ ] attempted to isolate the common determinants of prices shippers paid,” and the “damages model” attempted to quantify the overage due to the alleged price fixing.164 The district court accepted the models as “plausible” and “workable.”165 The D.C. Circuit found plaintiffs’ model defective because it proved too much. While the model appeared to show injury to all class members, it also “detect[ed] injury where none could exist.”166 This was because one group of shippers had entered in “legacy contracts” with defendants.167 These contracts guaranteed fuel surcharges would be subject to formulae that predated the allegedly anticompetitive surcharge changes.168 The expert’s damages model showed similar damages for legacy contract shippers and class members.169 This critique of the expert’s model, the court explained, was “sharpen[ed]” by Comcast. After Comcast, “[i]f the damages model cannot withstand this scrutiny then, it is not just a merits issue. [Plaintiffs’ expert’s] models are essential to plaintiffs’ claim they can offer common evidence of classwide injury. No damages model, no predominance, no class certification.”170 Thus, when a damages model includes damages for parties that could not have been harmed by defendants’ actions, the model must fail. As the court stated, “we have no way of knowing the overages the damages model calcu- lates for class members is any more accurate than the obviously false estimates it produces for legacy shippers.”171 The court’s analysis in Rail Freight further exposes the flaws in Whirlpool and Butler upon remand. As the D.C. Circuit pointed out, “[i]t is not enough to submit a questionable model whose unsubstantiated claims cannot be refuted through a priori analysis.”172 Were that the case, “at the

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) class-certiﬁcation stage any method of measurement is acceptable so long as it can be applied classwide, no matter how arbitrary the measurements may be.”173 This would contradict the import, if not the holding, of Comcast. A proper damages model must not only prove the possibly of damages, it must be limited to prove damages for only those plaintiffs who did or could possibly suffer harm. As courts continue to weigh in on the standard for damages models under Rule 23(b)(3), we expect courts to embrace the logic of the D.C. Circuit in Rail Freight and further examination of the Sixth and Seventh Circuits’ decisions, which seem wholly inconsistent with Comcast.

Cases Applying Comcast’s Rigor Requirement to Proof of Damages. Applying the Comcast rigor requirement to proof of damages in putative class actions, several courts have held that class treatment under 42 JAMES KEYTE ET AL.

Rule 23(b)(3) was not warranted because individualized issues on damages would predominate over any common issues.174 For example, in Roach v. T.L. Cannon Corp.,175 employees sued Applebee’s franchises in New York and Connecticut alleging violations of the New York Labor Law (“NYLL”) and the Fair Labor Standards Act (“FLSA”). Plaintiffs sought to certify several classes, including a class of employees who were allegedly denied overtime wages when they worked more than 10 hours in any particular day (“spread of hours”). A magistrate judge recommended that a class be certified with respect to plaintiffs’ spread of hours claim, but the district court rejected the recommendation due to the subsequently issued Comcast decision. The court described Comcast as holding that the failure of a class proponent “to offer a damages model that is ‘susceptible of measurement across the entire class for purposes of 23(b)(3)’ was fatal to the certification question.”176 Specifically, the court found that the plaintiffs’ failure to offer a damages model susceptible of measurement across the entire class “[was] in contravention of the holding in [Comcast].” Evidence that “some employees, on various occasions, were denied their 10-hour spread payments” indicated to the court that damages in the putative class would require highly individualized proof.177 Thus, the court concluded that questions of individual damage calculations would overwhelm common questions, and as such, Rule 23 certification had to be denied. Similarly, in Cowden v. Parker & Associates, Inc.,178 the district court refused to certify a class because individual questions on damages would predominate over any questions common to the entire class. There, plaintiff insurance agents sued Parker & Associates, an insurance agency, for fraud,

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) negligent misrepresentation, breach of contract, unjust enrichment, conver- sion and promissory estoppel, alleging that defendant failed to pay plaintiffs commissions for their sales of Medicare Advantage (“MA”) Plans. Plaintiffs sought to certify a class of insurance agents who had worked for defendant as insurance agents selling MA Plans.179 Plaintiffs recognized that their claims would require individual analyses for each agent’s sales and expenses. Plaintiffs’ compensation was commission-based, but could not be reduced to a formula or simple calculation. Instead, commissions were shared within a hierarchy of agents, through which an agent’s superiors shared in the commissions of any agent lower on the hierarchy. It appeared that compensation was decided based on oral representations to each agent. The defendant was also permitted to deduct from each agent’s commission check expenses and fees including mailing expenses, fees for leads, commissions advanced to the agent but From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 43

not earned because the carrier ultimately rejected the policy and policy pre- miums owed to the defendant. Citing In re Whirlpool, plaintiffs argued that “ ‘no matter how individualized the issue of damages may be, these issues may be reserved for individual treatment with the question of liability tried as a class action.’ ”180 The court noted that Whirlpool had been vacated for further consideration in light of Comcast, and found that, as in Comcast and Roach, plaintiffs “offered no manageable way to calculate damages across the entire class, and the individual damages calculations that would be required will inevitably overwhelm any questions common to the entire class.”181 In re Montano v. First Light Federal Credit Union,182 another post- Comcast decision, involved an adversary proceeding in bankruptcy court where the debtors sought class injunctive, declaratory, and monetary relief based on a credit union’s alleged violation of a discharge injunction through continued reporting of discharged debt as “past due” rather than “discharged.” The district court denied plaintiffs’ motion for reconsideration of the bankruptcy court’s decision to decertify the class. It cited several bases for this decision, including Comcast, which the Supreme Court handed down one week after the bankruptcy court’s decertiﬁcation ruling. “Comcast,” the court stated, “bolsters the Court’s decision to decertify [the damages class], because there is no evidence [that] the damages they suffered would be susceptible of class-wide measurement.”183 Any damages would be individualized, “especially given the evidence that payment of the discharged debt likely would not be an element of damages in more than a few cases.” And even if a model existed that was capable of proving classwide damages, the court held that after Comcast, plaintiffs are not allowed

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) “the luxury of waiting until trial” to come forward with proof that damages could be measured on a classwide basis.184

Cases Purporting to Follow the Comcast Dissent A few courts have found the dissent persuasive and followed its approach on damages.185 The Ninth Circuit has strongly resisted the Comcast opinion, instead siding with the dissent. The leading Ninth Circuit case on the issue is Leyva v. Medline Industries Inc., another employment class action seeking damages for violation of state labor laws.186 The case concerned alleged employer rounding down of employee time and improper application of nondiscretionary bonuses to overtime pay. The court rejected a broad reading of Comcast and found that in the Ninth Circuit, “the presence of individualized damages cannot, by itself, defeat class certiﬁca- tion.”187 The court found that Comcast requires only that plaintiffs show 44 JAMES KEYTE ET AL.

their damages “stemmed from the defendant’s actions that created the legal liability.”188 One key fact in this case, which arguably distinguishes the case from most other Rule 23(b)(3) class actions, was that defendant’s removal notice relied on its own electronic databases to calculate the amount in controversy for each claim and totaled the exposure on all claims. The defendant’s apparent admission conﬁrmed to the court that damages calculations would only entail a straightforward, mechanical application of readily available data to individual plaintiffs. The ability to calculate individualized damages through such a process likely would prevent individual questions from overwhelming questions common to the class. That interpretation of Leyva was relied on and adopted in Parra v. Bashas’, Inc., where the court found that plaintiffs’ methodology for calculating back pay showed that damages were capable of measurement on a classwide basis. The court acknowledged that Comcast requires proof that damages could be determined on a classwide basis, but found that the plaintiffs’ methodology for calculating damages correlated the legal theory of harm with the economic impact of that event. Further, similar to Leyva, the court found that once liability is established, calculating damages would be “a purely mechanical process” and thus there was no concern that individual questions would overwhelm common questions. Martins v. 3PD, Inc.,189 a post-Comcast case outside the Ninth Circuit, interpreted Comcast in a similar fashion. The court noted that even before Comcast, courts of appeal have found that predominance could be defeated where “questions of damage calculation are so complex or implicate so many potential class members, that they present a ‘Herculean task’ and 190

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) overwhelm liability issues.” The court interpreted Comcast to leave open the possibility of class certification where some individual issues on the calculation of damages remain as long as those determinations “will neither be particularly complicated nor overwhelmingly numerous.”191 The Ninth Circuit has been criticized by commentators for its narrow interpretation of Comcast in the Leyva case. For example, on the issue of rounding, a plaintiff would need to establish that he clocked in and started working before his official start time. He would also need to establish the defendant rounded up to the official start time, illegally denying him compensation. As one author notes, “there could be significant individualized issue[s] as to when any particular employee started working on a given day.”192 Another article suggested that Leyva, and not Comcast, is limited to its facts. Leyva had “straight-forward damages at issue” and “did not involve From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 45

any argument that individual damages calculations would be expensive, extensive, time-consuming or complex.”193 The authors note that Leyva relies on a pre-Comcast Ninth Circuit decision to establish that damage calculations alone cannot defeat certiﬁcation.194 According to the authors, that premise may still be valid where damages involve “mere mathematical computation[s] of damages based on known data.”195 But, if the Leyva court meant that individualized damage inquiries can never predominate over common liability issues, that opinion “cannot be squared with Comcast.”196

Ways to Attack Damages Evidence Under the New Rigor

Defendants opposing class certification should make use of the Supreme Court’s language in Comcast and seize on situations where individualized questions might overwhelm common questions. The following section lays out possible ways to attack motions for class certification based on failure to offer common proof of damages. This advice is drawn from Supreme Court guidance, post-Comcast case law and commentary as well as case law predating Comcast that has been thrust back into the mainstream following the decision. In the sections that follow, the Article suggests defendants should first evaluate whether plaintiffs have provided a developed damages model. Then, assuming the plaintiffs have set forth a model, defendants should analyze whether damages modeled are susceptible to mathematical or formulaic calculation and attempt to identify individualized issues among the plaintiffs or the negotiation for the product. Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Does the Plaintiff Provide a Formula to Calculate Damages? At the most basic level, plaintiffs cannot establish that damages are susceptible of measurement across the entire class for purposes of Rule 23(b)(3) without providing a formula that endeavors to accomplish the task. After Comcast, plaintiffs cannot promise that they will develop a damages model; they must have a model in place at the class certification stage. One commentator noted: “[After Comcast,] if a case is too big to devise a damages process that matches actual damages to actual litigants, a majority of the current court is likely to treat it as ‘too big’ for litigation.”197 Roach v. T.L. Cannon Corp., a post-Comcast case discussed above, adopted this approach. Plaintiffs argued that they could address the liability question first without supplying a damages model because the damages issue “is separate from the question of liability.”198 Even if such damages 46 JAMES KEYTE ET AL.

might be highly individualized, Plaintiffs “contend[ed] that damages need not be considered for Rule 23 certification.”199 The court flatly rejected this argument and found plaintiffs’ failure to offer a damages model was fatal to class certification, following guidance from Comcast.200 Similarly, Rodney v. Northwest Airlines,201 a Sixth Circuit opinion that predates Comcast, denied certification partially due to plaintiffs’ failure to offer a formula to calculate damages. In Rodney, plaintiffs’ experts admitted they had not yet identified a methodology for calculating damages. Instead, they argued it was sufficient that they had a methodology that “could be used to calculate damages.”202 The court did not agree. Without a formula in place to compute damages, plaintiffs could not prove that class issues predominated.203 One commentator suggested that post-Comcast, a methodology capable of providing common proof of damages would be necessary even if plaintiffs wanted to separate the liability and damages portions of a class action. The author noted:

[T]he opinion in Comcast is silent as to the availability of Rule 23(c)(4) to bifurcate the problematic damages portion of a class action. … Although Rule 23(c)(4) should remain available as a management tool in cases where common liability questions are found to predominate over individualized damages questions, an attempt to circumvent the holding of Comcast by carving out the individualized damages questions in order to satisfy Rule 23(b)(3) predominance would be at odds with the Supreme Court’s holding in Comcast.204

Are Damages Susceptible to “Mathematical” or “Formulaic” Calculation? If plaintiffs have at least attempted to provide a formula to establish Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) damages across the entire class, defendants should next consider whether damages can be properly calculated under the proposed methodology in a straightforward fashion. Courts have identiﬁed three major areas where plaintiffs methodology fails to prove damages on a classwide basis: (1) where the product involves individualized negotiations; (2) where individual plaintiffs differ in ways that make proof of damages highly individualized; and (3) where plaintiffs offer a methodology to prove damages, but that methodology fails to account for differences between plaintiffs.

Does the Product at Issue Involve Individualized Negotiations? The post- Comcast decisions are nearly universal that where the product at issue involves individualized negotiations on price, damages issues are likely to overwhelm any common issues. This approach was taken by the Fifth From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 47

Circuit nearly a decade ago in Piggly Wiggly and is a key consideration when attempting to defeat class certification.205 In Piggly Wiggly, plaintiffs brought a civil antitrust case alleging defendants conspired to fix the prices of bread and cake products in Texas and Louisiana.206 The proposed class included both wholesale purchasers and bid purchasers. Wholesale purchasers, typically small grocery and conveni- ence stores, made purchases based on a price list provided by defendants, while bid purchasers, typically large grocery stores, school districts, and local government entities, received bids from defendants.207 For bid purchasers, the court accepted plaintiffs’ allegation that bids were based on the wholesale price lists given to wholesale purchasers.208 However, the court found that plaintiffs could not establish that class members suffered injury to their business or property as a result of the conspiracy through common proof and therefore denied certification under Rule 23(b)(3).209 Even starting from the wholesale price list, as plaintiffs alleged, the final price ultimately was arrived at after negotiation. The court noted that the final price resulted from many factors, including the amount of product purchased, geographic market, the particular services included, delivery costs, the discount negotiated, and the negotiating skill of the parties involved. Because no two negotiations were exactly the same, “it [would] be impossible to present evidence in a common manner as to the price each Plaintiff would have paid but for the alleged conspiracy.”210 The court further noted that predominance and manageability may be destroyed “solely by the complexity of determining damages when that determination does not lend itself to a mathematical calculation that can be applied to all class members.”211 Plaintiffs’ “mere assertions” that there are

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) a number of methods or methodologies that could be utilized to prove damages were not sufficient.212 The court concluded, “Where the plaintiffs’ damage claims ‘focus almost entirely on facts and issues specific to individuals rather than the class as a whole’, the potential exists that the class action may ‘degenerate in practice into multiple lawsuits separately tried’.”213 “In such cases, class certification is inappropriate.”214 The Fifth Circuit affirmed the district court’s ruling,215 holding that “[t]he necessity of calculating damages on an individual basis, by itself, can be grounds for not certifying a class.”216 Although it acknowledged that in antitrust cases there is a relaxed burden for proving the amount of damages once the fact of damage is proven, the court agreed with the district court that damages may not be merely speculative. The Fifth Circuit agreed with the district court that individualized issues such as negotiating ability and geographic market could not be included in a general formula. 48 JAMES KEYTE ET AL.

Similarly, in Mekani v. Miller Brewing Co.,217 beer retailers brought a civil antitrust case alleging a conspiracy among brewers and beer distributors to fix prices and territories in violation of the Sherman Act as well as price discrimination in violation of the Robinson Patman Act.218 Plaintiffs alleged that six major brewers and over 180 distributorsÀ engaged in horizontal and vertical conspiracies to fix the price of beer and establish and maintain noncompetitive geographic territories.219 The proposed class consisted of beer retailers in Michigan. The court denied certification of both the Sherman Act and Robinson Patman Act claims. On the Sherman Act claims, the court found that evenÀ if common questions would predominate on liability, “plaintiffs are still required to prove fact of injury.”220 The damages allegedly suffered by the class were not a function of a single event or decision made by the defendants, but rather resulted from a combination of decisions and factors specific to a given retailer distributor transaction: here, prices and costs could vary depending on theÀ local markets, the different brands purchased, and the different sales packages chosen.221 These individualized facts made it impossible for plaintiffs to prove amount of injury on a classwide basis and rendered the case “unmanageable as a class action.”222 Dry Cleaning & Laundry Institute of Detroit, Inc. v. Flom’s Corp.,223 another pre-Comcast decision, also found individualized negotiations created an impediment to plaintiffs’ proof of classwide damages. In Dry Cleaning, Michigan dry cleaners and a trade association alleged Sherman Act price-fixing violations against dry cleaning and laundry supply companies. The court cited the same problems that doomed class certification in Mekani and Piggly Wiggly:

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) In the instant action, thousands of transactions were involved over many years; each transaction was different; various plaintiffs may have purchased the dry cleaning sup- plies, which varies by brand and type, at different prices and varying quantities, in different ways under different credit terms. No formula proposed would be even-handed among class members or fair to defendants.224

Do Individual Plaintiffs Differ in Ways That Make Proof of Damages Highly Individualized? Even if plaintiffs do not engage in individualized negotiations, there may be important differences between the plaintiffs themselves that prevent mathematical or formulaic calculation of damages. Bell Atlantic Corp. v. AT&T Corp. provides one example of these types of differences.225 In Bell Atlantic, discussed above, plaintiffs alleged AT&T attempted to monopolize the market for caller-ID service by blocking the From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 49

free passage of caller-ID data over its long-distance network.226 Plaintiffs sought to certify two classes: a “reverse charge” class comprised of businesses that purchased AT&T reversed billed long-distance service (e.g., “800” numbers) and caller-ID service and a “call recipient” class of businesses and organizations that were actual or potential purchasers of caller- ID service for long-distance calls.227 The basic allegations were that but for AT&T’s monopolization of caller ID, plaintiffs could have avoided answering certain calls and spent less time on answered calls through the use of caller ID. Plaintiffs claimed damages due to wasted employee time on phone calls and greater long-distance charges assessed on calls that would otherwise not have been answered.228 The court found class treatment is not appropriate in cases “where the calculation of damages is not susceptible to a mathematical or formulaic calculation.”229 Class members ranged in size from primarily local sole pro- prietorships to large interstate corporations with call centers, used different types of equipment, and had different volumes of repeat business, which would, in turn, change how effective caller ID would be for call screening.230 These individualized differences demonstrate the damages are “not susceptible to a mathematical or formulaic calculation,” and therefore individualized proof overwhelms the common questions.231 The Supreme Court appears to agree that differences between individual plaintiffs can defeat common proof of damages. In Comcast, the court cautioned against certiﬁcation where “permutations” of harm amongst plaintiffs “are nearly endless.”232 Further, both GVR’d cases arguably contain such endless permutations of harm amongst plaintiffs. Key considerations on remand in those cases should include how plaintiffs differed in their use

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) of washing machines and the differences among the 21 types and 9 model years of allegedly defective machines.233 Although the defendants in those cases do not argue plaintiffs negotiated differently, they can allege that the plaintiffs differ in ways which overwhelm common proof of damages.

Do Plaintiffs Propose a Damages Formula Clearly Inadequate to Calculate Individual Damages? Alternatively, defendants can attack plaintiffs’ methodology for failing to adequately calculate individual damages. As discussed above,234 the use of averages can be a fatal ﬂaw in a plaintiff’s attempt to prove a common impact. While Defendants should challenge Plaintiffs’ methodologies that rely on averaging under Daubert, they can also make damages-speciﬁc challenges outside of the Daubert context. Averages can provide even greater problems in trying to prove damages across an entire class. For example in Bell Atlantic, the court stated, “[c]lass treatment is 50 JAMES KEYTE ET AL.

not appropriate in cases … where the formula by which the parties propose to calculate individual damages is clearly inadequate.”235 The plaintiffs’ method relied primarily on averaging. Damages were calculated based on “average number of seconds saved per call [both long-distance and local] through the use of caller ID, an average wage rate for the typical employee answering and processing telephone calls, and the total number of AT&T calls to class members made during the class period.” The formula was adjusted for the reverse charge class “using AT&T’s billing records, to include recovery of any long-distance charges assessed against class members that might have otherwise been avoided through the use of caller ID.”236 The court noted that plaintiffs’ averaging formula “ma[de] no effort to adjust for the variegated nature of businesses included in the classes” and thus could not reasonably approximate actual damages suffered by class members.237 Dry Cleaning, also discussed above, addressed similar methodological problems. In that case, plaintiffs attempted to supply a method for calculating damages that would overcome the problems of common proof caused by individualized negotiations. The method relied on “benchmarking,” which the court rejected, correctly noting that this method failed to account for the variation in markets, products, negotiating power and buyer size.238 As a result, plaintiffs could not establish that common questions predominated over individualized inquiry on damages.239 Finally, Reed v. Advocate Health Care faulted the plaintiffs’ expert for relying on averages.240 In Reed, registered nurses (“RNs”) alleged health care organizations conspired to depress wages and exchange compensation information. Although plaintiffs proposed a formula to calculate damages,

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) the court found the formula “unacceptably masks the signiﬁcant variation in RN base wages during the Class Period.” The court sided with defendants’ expert, who noted:

[T]he relative movements of mere averages (means) do not prove common impact to individual RNs. For example, mean wages for Defendants’ RNs could move together even though particular Defendants gave larger increases to certain, hard to ﬁnd nurses, and smaller increases to others. The issue is the feasibility of common proof regarding individual nurses, not a hypothetical ‘average’ nurse.241

CONCLUSION

Comcast conﬁrmed that there is a new rigor in Rule 23(b)(3) class actions, and it is not limited to issues of impact. And, in the wake of Comcast, both From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 51

practitioners and economists must adjust their tactics and proof to this new environment, which itself will evolve fairly rapidly as lower courts continue to grapple with the requirements that the Supreme Court has imposed.

NOTES

1. In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 318 (3d Cir. 2008). 2. See, for example, Reed v. Advocate Health Care, 268 F.R.D. 573 (N.D. Ill. 2009); In re Plastics Additives, 03-CV-2038, 2010 WL 3431837 (E.D. Pa. August 31, 2010); In re Intel Corp. Microprocessor Antitrust Litig., MDL 05-1717 JJF, 2010 WL 8591815 (D. Del. July 28, 2010). See also Thompson, Bloch, and Guerin- Calvert (2010) (“In our view, Hydrogen Peroxide has done for class certification what Twombly and Iqbal did for the pleading standards plaintiffs must satisfy when faced with a motion to dismiss, and what Daubert did with respect to the standards for evaluating expert evidence.”). 3. Daubert v. Merrell Dow Pharm., 509 U.S. 579 (1993). 4. See, for example, Piggly Wiggly Clarksville, Inc. v. Interstate Brands Corp., 215 F.R.D. 523 (E.D. Tex. 2003), aff’d, 100 F. App’x 296 (5th Cir. 2004). 5. In Story Parchment, for example, the Court recognized that “although damages may not be determined by mere speculation or guess,” the damage calculations need not be exact. Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 563 (1931). Instead, evidence that supports a “just and reasonable inference” will be accepted even if the result is “only approximate.” Ibid. See also Olden v. LaFarge Corp., 383 F.3d 495, 508 (6th Cir. 2004) (affirming class certification where liability could be determined as to the entire class but individual issues sur- faced during the determination of damages). 6. Comcast Corp. v. Behrend, 133 S. Ct. 1426 (2013). 7. See infra section IV.B.1(a). Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) 8. Ibid. 9. See Comcast, 133 S. Ct. at 1432. 10. See Bogosian v. Gulf Oil Corp., 561 F.2d 434, 454 (3d Cir. 1977). 11. See ibid. at 1433. 12. Ibid. at 1432 (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 615, 617 (1997)). 13. In Comcast, the Court held that respondents could not bring a Daubert challenge to expert testimony after the class certification hearing occurred. By refusing to allow respondents to do so, the Court hinted that parties must raise, and courts must consider, Daubert issues at the class certification stage. Although Dukes also did not hold that trial courts must apply Daubert scrutiny at the class certification stage, the Court suggested it much more explicitly than it did in Comcast: “The parties dispute whether Bielby’s testimony even met the standards for admission of expert testimony … The District Court concluded that Daubert did not apply to expert testimony at the certification stage of class-action proceedings. We doubt that this is so, but even if properly considered, Bielby’s testimony does nothing to 52 JAMES KEYTE ET AL.

advance respondent’s case.” Wal-Mart Stores, Inc. v. Dukes, 113 S. Ct. 2541, 2553 2554 (2011). 14.À 600 F.3d 813 (7th Cir. 2010). 15. Ibid. at 814. 16. Ibid. at 816. 17. Ibid. at 816 817. 18. Ibid. À 19. Ibid. 20. 419 F. App’x 887 (11th Cir. 2011). 21. Ibid. at 890. 22. Comcast, 133 S. Ct. 1426, 1431 n.4. 23. This was precisely the result in In re Lineboard Antitrust Litigation, a 2002 case decided by the Third Circuit before Hydrogen Peroxide. In re Lineboard Antitrust Litig. 305 F.3d 145, 154 155 (3d Cir. 2002). There, the plaintiffs’ expert assured the court that she could relyÀ on several accepted statistical or mathematical approaches to prove harm, including benchmarking, the yardstick approach, and simply comparing prices during the conspiracy period to prices during the noncon- spiracy period. Ibid. Accepting the expert’s methods at face value methods that have been rejected by courts after Hydrogen Peroxide the courtÀ held that plaintiffs eventually could establish injury on a classwide basis.À Ibid. at 155. 24. There are indications that, at least in the past, an unsuccessful Daubert challenge at the class certification stage can reduce the likelihood of successfully challenging that expert at a later phase. Accordingly, there might be some instances where a Daubert challenge at the class certification stage could be counterproductive for defendants. See James Langenfeld & Christopher Alexander, Daubert and other Gatekeeping Challenges of Antitrust Experts, Antitrust, Vol. 25, pp. 21, 24, 25 (2011). 25. The “Philadelphia cluster,” the relevant cluster in Comcast, was made up of 16 counties located in Pennsylvania, Delaware and New Jersey. Comcast, 133 S. Ct. at 1430. 26. The four theories were that (1) clustering acquisitions prevented or made it difficult for customers to compare prices; (2) direct broadcast satellite companies, one set of potential competitors, found it more difficult to gain rights to local sports con- Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) tent and decided not to enter the Philadelphia market as a result; (3) Comcast’s ability to obtain programming material at lower prices allowed it to raise prices; and (4) the “overbuilder deterrence theory” discussed above. Comcast, 133 S. Ct. at 1439. 27. In short, the model presented by plaintiff’s expert created a “but-for” baseline price a price the market would have settled on absent the effects of all four allegedÀ theories of liability. 28. See infra section IV.B.1. 29. Federal Rule of Civil Procedure 23(c)(1)(A) (“Time to Issue. At an early prac- ticable time after a person sues or is sued as a class representative, the court must determine by order whether to certify the action as a class action.”). See also Hydrogen Peroxide, 552 F.3d 305, 318 (3d Cir. 2008) (discussing 2003 Amendments to Rule 23(c)(1)(A) and their implications for the timing of class certification). 30. Compare M. Sean Royall, Disaggregation of Antitrust Damages,65 Antitrust L.J. 311 (1997) with Conwood Co., L.P. v. U.S. Tobacco Co., 209 F.3d 768 (6th Cir. 2002). 31. See Comcast, 133 S. Ct. at 1440 1441. À From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 53

32. American Honda Motor Co. v. Allen, 600 F.3d 813, 818 (7th Cir. 2010). 33. Ibid. 34. Despite the expert’s report being published in a journal for forensic engineers who testify as motorcycle experts, there is no indication that the standard itself was accepted by anyone other than the expert. In addition, the published article acknowledged there was no standard articulated by the government, the industry, or the Society of Automotive Engineers that determined acceptable response characteristics for motorcycles once they go into “wobble mode.” 35. Plaintiff’s expert had never conducted any “rider confidence studies” to determine when or how motorcycle riders perceive wobble nor did he perform a test to determine the minimum wobble amplitude at which a rider would detect oscillation. 36. The expert’s study relied on a test of a single motorcycle ridden by a single test rider and then extrapolated those findings to an entire fleet of motorcycles produced from 2001 to 2008. 37. Ibid. at 818 819. 38. 400 F.3d 562À (8th Cir. 2005). 39. Ibid. at 569 570. 40. See ibid. atÀ 569 571. 41. See In re ScrapÀ Metal Antitrust Litig., 527 F.3d 517, 535 (6th Cir. 2008) (“Thus, even when there are individual variations in damages, the requirements of Rule 23(b)(3) are satisfied if the plaintiffs can establish that defendants conspired to interfere with the free-market pricing structure.”). 42. In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 313 (3d Cir. 2008) (rejecting this type of simplistic analysis of market pricing). 43. No. 03-CV-2038, 2010 WL 3431837 (E.D. Pa. August 31, 2010). 44. Ibid. at *13. 45. Ibid. 46. Ibid. at *14. 47. Ibid. 48. Ibid. 49. No. 09-23187-CIV, 2012 WL 27668 (S.D. Fla. January 3, 2012). 50. Ibid. at *10. Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) 51. Ibid. 52. See, for example, Burtis and Neher (2011). 53. Johnson and Leonard (2008); see also Fisher (1986) (describing correlation analysis as “pseudoscience” and “poor statistical and econometric testimony mas- querading as serious science”). 54. Reed v. Advocate Health Care, 268 F.R.D. 573, 591 (N.D. Ill. 2009) (quoting ABA Section of Antitrust Law, Econometrics: Legal, Practical, and Technical Issues 220 (2005)); see also In re Graphics Processing Units Antitrust Litig., 253 F.R. D. 478, 493 (N.D. Cal. 2008) (hereinafter “In re GPU”) (rejecting “correlation analyses” based on averages rather than specific “prices paid by individual consumers” and finding that plaintiffs’ expert “evaded the very burden that he was supposed to shoulder”); Forister and Hussain (2010) (“Because not all prices exactly follow the trend of the ‘average’ price, this aggregation or pooling of data precludes one from determining ways in which individual price series are related and how the alleged price-fixing agreement would affect different prices.”). 55. 339 F.3d 294 (5th Cir. 2003). 54 JAMES KEYTE ET AL.

56. Ibid. at 299 300. 57. Ibid. at 300.À According to plaintiffs, the business benefits of caller ID are screening out unwanted calls (and thereby reducing long-distance expenses), return- ing calls hours later (even in cases where the caller does not leave a message), tracking call volume (because some units can record information for later recall), and faster and more efficient customer service. 58. Ibid. at 300. 59. Ibid. 60. Ibid. at 307. 61. Ibid. at 306, 307. 62. Ibid. 63. Ibid. 64. Reed v. Advocate Health Care, 268 F.R.D. 573, 577 578 (N.D. Ill. 2009). 65. See ibid. at 582 584. À 66. Ibid. at 589. À 67. Ibid. at 591. 68. Ibid. at 592. 69. Ibid. 70. Ibid. at 593. 71. Ibid. at 594. 72. In re Wholesale Grocery Prod. Antitrust Litig., No. 09 MD 2090 ADM/AJB, 2012 WL 3031085 (D. Minn. July 25, 2012). À À 73. Ibid. at *10. 74. Ibid. 75. Ibid. at *13. 76. Ibid. 77. Ibid. at *10. 78. 863 F. Supp. 2d 966, 972 (C.D. Cal. 2012). 79. See ibid. at 974. 80. Ibid. at 975. 81. No. 09-ML-2048-C, 2011 WL 6826813 at *1 (W.D. Okla. December 28, 2011). Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) 82. Ibid. at *13. 83. Ibid. at *15. The court acknowledged that the test was “promulgated by serious economists who are experts in their field, peer reviewed, published and discussed in economic literature, and previously used in other litigation.” Thus, the court determined that whether the GRS test passes Daubert muster “is not easily resolved.” Ibid. 84. Ibid. 85. See ibid. at *15 16. 86. See, for example,À Blades v. Monsanto Co., 400 F.3d 562, 569 (8th Cir. 2005) (deferring ruling on Daubert motion in order to consider Plaintiffs’ arguments on class certification); Bacon v. Honda of Am. Mfg., Inc., 205 F.R.D. 466, 471 (S.D. Ohio 2001), aff’d, 370 F.3d 565 (6th Cir. 2004) (denying Daubert motion in order to consider class certification motion); In re Visa Check/MasterMoney Antitrust Litig., 192 F.R.D. 68, 77 (E.D.N.Y. 2000) (discussing preliminary role of Daubert inquiry in class certification analysis) aff’d, 280 F.3d 124 (2d Cir. 2001). From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 55

87. See, for example, Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 625 (1997) (predominance is “a test readily met in certain cases alleging … violations of the antitrust laws.”). 88. In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 313 (3d Cir. 2008) (“Private damage claims by numerous individuals arising out of concerted antitrust violations may or may not involve predominating common questions.” (citing Fed. R. Civ. P. 23(b)(3) advisory committee’s note, 1966 Amendment)); see also Robinson v. Tex. Auto. Dealers Ass’n, 387 F.3d 416, 420 421 (5th Cir. 2004) (“There are no hard and fast rules … regarding the suitabilityÀ of a particular type of antitrust case for class action treatment.”). 89. See Wal-Mart Stores, Inc. v. Dukes, 113 S. Ct. 2541, 2551 (2011); see also Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 156 (1982). 90. Dukes, 131 S. Ct. at 255; see also Rutstein v. Avis Rent-A-Car Sys., Inc., 211 F.3d 1228, 1234 (11th Cir. 2000) (collecting cases that considered merits when deciding a motion for class certification); Coopers & Lybrand v. Livesay, 437 U.S. 463, 467 & n.12 (1978) (“[T]he class determination generally involves considerations that are ‘enmeshed in the factual and legal issues comprising the plaintiff’s cause of action’ … ‘The more complex determinations required in Rule 23(b)(3) class actions entail even greater entanglement with the merits’.” (citation omitted)). Notably, for class certification, the burden of proof rests with plaintiffs to put forth common proof of impact by a preponderance of the evidence. See Hydrogen Peroxide, 552 F.3d at 320; see also In re Checking Account Overdraft Litig., Nos. 1:09-MD-02036-JLK et al., 2011 WL 3158998, at *2 (S.D. Fla. July 25, 2011) (to be published in F.R.D.) (“A district court may certify a class only if, after ‘rigorous analysis’, it determines that the party seeking certification has met its burden of a preponderance of the evidence.”). 91. See, for example, Blades v. Monsanto Co., 400 F.3d 562, 569 570 (8th Cir. 2005) (court denied class certification where it “considered all expertÀ testimony offered by both sides” and “afforded that testimony such weight as [the Court] deemed appropriate”); In re New Motor Vehicles Can. Export Antitrust Litig., 522 F.3d 6, 17 (1st Cir. 2008) (“It would be contrary to the ‘rigorous analysis of the pre- requisites established by Rule 23 before certifying a class’ to put blinders on as to Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) an issue simply because it implicates the merits of the case.”); West v. Prudential Sec., Inc., 282 F.3d 935, 938 (7th Cir. 2002) (same); Unger v. Amedisys Inc., 401 F.3d 316 (5th Cir. 2005) (same); In re Evanston N.W. Healthcare Corp. Antitrust Litig., 268 F.R.D. 56 (N.D. Ill. 2010) (citing Hydrogen Peroxide, 552 F.3d at 324) (“[r]esolving expert disputes to determine whether a class certification requirement has been met is always a task for the court no matter whether a dispute might appear to implicate the ‘credibility’ of one orÀ more experts, a matter resembling those usually reserved for a trier of fact”). Note, however, that before Comcast, a minority of circuits held that it is inappropriate to resolve merits issues related to expert testimony and would certify classes on the basis of plaintiffs’ expert opinions so long as the experts’ methodologies were not “worthless” or “inherently faulty.” See In re Polypropylene Carpet Antitrust Litig., 996 F. Supp. 18, 25, 29 (N.D. Ga. 1997); Drayton v. W. Auto Supply Co., No. 01 10415, 2002 WL 32508918, at *6 n.13 (11th Cir. Mar. 11, 2002) (affirming the districtÀ court’s refusal to conduct a Daubert analysis or to resolve a 56 JAMES KEYTE ET AL.

“battle of the experts” and finding that expert testimony proffered was sufficient at class certification stage to demonstrate that plaintiffs could show “common” proof of impact); In re Commercial Tissue Prods., 183 F.R.D. 589, 596 (N.D. Fla. 1998) (“This case presents the familiar ‘battle of the experts’. The certification stage of this litigation is not, however, the proper forum in which to resolve this battle.”). 92. See, for example, In re Domestic Air Transp. Antitrust Litig., 137 F.R.D. 677, 689 (N.D. Ga. 1991) (inflated fares resulted in artificial “base” price which became benchmark for discounted or negotiated fares); J.B.D.L. Corp. v. Wyeth-Ayerst Labs., Inc., 225 F.R.D. 208, 217 (S.D. Ohio 2003) (finding that price negotiations with each customer do not undermine plaintiffs’ arguments related to common impact because conspiracy allegedly raised the starting price for negotiations); In re Flat Glass Antitrust Litig., 191 F.R.D. 472, 486 (W.D. Pa. 1999) (finding that “even though some plaintiffs negotiated prices, if plaintiffs can establish that the base price from which these negotiations occurred was inflated, this would establish at least the fact of damage, even if the extent of the damage by each plaintiff varied”); see also In re Indus. Diamonds Antitrust Litig., 167 F.R.D. 374, 383 (S.D.N.Y. 1996) (collecting cases); In re Catfish Antitrust Litig., 826 F. Supp. 1019 (N.D. Miss. 1993); Hedges Enters., Inc. v. Cont’l Grp. Inc., 81 F.R.D. 461, 475 (E.D. Pa. 1979). 93. See, for example, Kenett Corp. v. Mass. Furniture & Piano Movers Ass’n, Inc., 101 F.R.D. 313, 314, 316 (D. Mass. 1984) (finding that although moving service established standard fee schedules, individual questions predominated because the bundle of moving services provided differed from mover to mover and customer to customer “in light of the costs facing the company at the time, its competitive market, its reputation, and the particular services required for the move”); Am. Custom Homes v. Detroit Lumberman’s Ass’n, 91 F.R.D. 548, 550 (E.D. Mich. 1981) (finding that although defendant provided prices lists, individual questions predominated because actual purchases of defendant’s product involved individualized negotiations that varied based on produce design, credit terms, negotiating power, and specific purchase arrangements); In re Beef Indus. Antitrust Litig., 1986 2 Trade Cas. (CCH), 1986 WL 8890 at *2 (S.D. Tex. 1986) (finding that althoughÀ defendant alleging part of cattle price was tied to “Yellow Sheet,” individual questions predominated because the price for cattle depended on a “multitude of varying factors Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) differing from one area to another” including weather, minimum cattle requirements, price competition, and the beef by-product market). 94. In re Plastics Additives Antitrust Litig., No. 03-CV-2038, 2010 WL 3431837, at *13 (August 31, 2010). 95. Ibid. 96. Ibid. 97. No. 09-23187-CIV, 2012 WL 27668, at *9 (S.D. Fla. January 3, 2012). 98. Ibid. 99. 253 F.R.D. 478, 480 (N.D. Cal. 2008). 100. Ibid. 101. Ibid. (“First, cards and chips were sold to original equipment manufacturers (OEMs), such as Dell or Hewlett-Packard. The OEMs then installed the cards or chips in their own computers for later resale to individual consumers. Second, chips were sold to add-in-board manufacturers (AIBs) who in turn incorporated them into their own cards. These cards could then be sold as standalone products for a From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 57

retail price or to other computer manufacturers (e.g., OEMs) for incorporation into whole computers. Third, distributors could purchase chips or cards, which they in turn sold to other entities along the chain of distribution (i.e., AIBs, OEMs, or other distributors). Fourth, retailers, such as Best Buy, could purchase graphics cards for later sale to individual consumers. Fifth, original design manufacturers (ODMs) purchased graphics cards or chips that would be incorporated into parts that would later be branded by another firm along the chain of distribution for sale. Sixth, ATI sold graphics cards online through its website ATI.com directly to individual consumers.”). 102. Ibid. at 490 491. 103. Nichols v. MobileÀ Cnty. Bd. of Realtors, No. 76-619-P, 1980 WL 1975, at *7 (S.D. Ala. May 16, 1980) (collecting and comparing antitrust cases in which products were homogeneous or distinct); In re Infant Formula Antitrust Litig., No. MDL 878, 1992 WL 503465, at *5 (N.D. Fla. January 13, 1992) (“Contentions of infinite diversity of product, marketing practices, and pricing have been made in numerous cases and rejected. Courts have consistently found the conspiracy issue the overriding, predominant question”). 104. Nichols, 1980 WL 1975 at *7. 105. In re Plastics Additives Antitrust Litig., No. 03-CV-2038, 2010 WL 3431837, at *8 (E.D. Pa. August 31, 2010). 106. Ibid. 107. In re Fla. Cement & Concrete Antitrust Litig., 09-23187-CIV, 2012 WL 27668 (S.D. Fla. January 3, 2012). 108. Ibid. at 13. 109. Ibid. 110. Ibid. 111. Another argument defendants can make is that there are variations across the geographic regions included in the plaintiffs’ proposed class. Although plaintiffs generally do not need to define a relevant geographic market in a per se price-fixing case, issues related to geographic market may be relevant where defendants show that localized markets and local pricing exist because these markets may raise individual questions about impact and damages. See Behrend v. Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Comcast, No. 10-2865, 2011 WL 3678805, at *7 8 (E.D. Pa. August 24, 2011) (citing Areeda & Hovenkamp, 1998 2007 & supp.À 2008, ¶398b); see also Blades v. Monsanto Co., 400 F.3d 562 (8thÀ Cir. 2005) (denying class certification because of the localized nature of the markets for genetically modified corn and soybean seeds that exhibited substantial price variation); Behrend, 2011 WL 3678805, at *7 8 (Jordan, J., concurring) (finding that geographic market is a relevant inquiryÀ in a per se case where “the class definition includes a geographic component”). 112. Comcast Corp. v. Behrend, 133 S. Ct. 1426, 1434 (2013). 113. Ibid. 114. See Blades, 400 F.3d at 562 (“Simply put, plaintiffs presume class-wide impact without any consideration of whether the markets or the alleged conspiracy at issue here actually operated in such a manner as to justify that presumption. Dr. Leitzinger assumes the answer to this critical question …”). 115. Ibid. 58 JAMES KEYTE ET AL.

116. See Comcast, 133 S. Ct. at 1435. 117. Ibid. 118. See ibid. 119. In re Graphics Processing Units Antitrust Litig, 253 F.R.D. 478, 508 (N.D. Cal. 2008). 120. Ibid. at 489 491. 121. Ibid. at 490;Àsee also In re Indus. Diamonds Antitrust Litig., 167 F.R.D. 374, 383 385 (S.D.N.Y. 1996) (“[W]e are persuaded that despite the wide range of pro- ductsÀ and prices involved, common proof of impact is possible on behalf of purchasers who bought list-price products. Common proof of impact is not possible, however, on behalf of those purchasers who bought non-list price products.”). 122. See generally Comcast Corp. v. Behrend, 133 S. Ct. 1426 (2013). 123. Ibid. at 1432. 124. Ibid. (citation omitted). 125. Ibid. at 1432 1433. 126. Ibid. at 1433.À 127. Ibid. at 1435 n.6. This Article does not contend that plaintiffs must demonstrate that they suffered the same amount of harm to establish that common questions predominate after Comcast. The Court’s indication that plaintiffs must offer proof of damages that establishes the “requisite level of commonality” does, however, underscore the Court’s rigorous analysis of common proof of the amount of damages. 128. Ibid. at 1433. 129. See ibid. at 1434 (discussing how plaintiffs’ “damages” may not have been tied to actual harm in Philadelphia and thus it was possible that many individuals in the Philadelphia DMA did not suffer actionable harm). 130. See ibid. at 1434 1435. 131. See ibid. at 1433.À 132. Ibid. at 1436 (citation omitted). 133. Ibid. 134. Ibid. 135. Ibid. at 1437. The dissenters noted that, instead of refusing to certify a class Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) action, courts can divide classes into subclasses, certify classes for liability purposes only, or alter or amend cases as they develop. Ibid. at 1437 n.*. 136. 678 F.3d 409 (6th Cir. 2012) (GVR’d April 1, 2013). 137. 702 F.3d 359 (7th Cir. 2012) (GVR’d June 3, 2013). 138. In the third, Ross v. RBS Citizens, N.A., 667 F.3d 900 (6th Cir. 2013) (GVR’d April 1, 2013), Defendant RBS Citizens appealed solely on the ground that Plaintiffs’ proposed class did not comply with Rule 23(c)(1)(B). As such, the opinion does not discuss damages or Rule 23(b)(3). 139. Whirlpool, 678 F.3d at 415. 140. Ibid. at 412. 141. Ibid. at 415. 142. Ibid. at 419 (citation omitted). 143. Ibid. at 421. 144. Butler, 702 F.3d at 360 361. The defects all concerned Whirlpool- manufactured machines, but SearsÀ was the defendant because it could still be From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 59

strictly liable for breach of warranty, with a possible contribution or indemniﬁca- tion claim against Whirlpool. See ibid. at 363. 145. Ibid. at 362. 146. Ibid. at 363. 147. Ibid. at 362, 363. 148. Ibid. at 363. 149. Ibid. at 362. 150. Beisner, Miller, and Wyatt (2013). 151. McLaughlin (2013) at 2. 152. Ibid. 153. Mass Tort Defense (2013). 154. In re Whirlpool Corp. Front-Loading Washer Products Liab. Litig., 722 F.3d 838, 860 (6th Cir. 2013). 155. See ibid. at 855 856 (relying on an unsupported state law analysis that would allow class membersÀ to recover damages even when the defect had not manifested for plaintiffs). 156. Wal-Mart Stores, Inc. v. Dukes, 131 S. Ct. 2541, 2550 (2011). 157. Butler v. Sears, Roebuck & Co., 11-8029, 2013 WL 4478200 at *4 (7th Cir. August 22, 2013). 158. Ibid. at *5. 159. Ibid. 160. In re Rail Freight Fuel Surcharge Antitrust Litig.-MDL No. 1869, 12-7085, 2013 WL 4038561 at *8 (D.C. Cir. August 9, 2013). 161. Ibid. at *1. Fuel surcharges are added on top of the base rate railroads charge customers for freight. Ibid. Historically, fuel surcharges have taken two forms: “mileage-based fuel surcharges,” which raise total freight rates in proportion to shipping distances, and “rate-based fuel surcharges,” which impose a surcharge as a function of the base rate once fuel prices exceed a “strike” or “trigger” price. Ibid. In 2003 2004, all four major railroads had fully switched to rate-based surcharges, and allÀ four lowered the trigger prices. Ibid. The Surface Transportation Board ended the practice, expressing concerns that, “the rationale for the fuel surcharges fuel cost and recovery” was disconnected from the base Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) rates on which the surchargesÀ were based. Ibid. at *2. Following decision, plaintiffs here alleged the defendants engaged in price ﬁxing under §1 of the Sherman Act. Ibid. 162. Ibid. at *2. 163. Ibid. at *3. 164. Ibid. 165. Ibid. (quoting In re Rail Freight Fuel Surcharge Antitrust Litig. (Fuel Surcharge II), 287 F.R.D. 1, 67 (D.D.C. 2012). 166. Ibid. at *5. 167. Ibid. at *1. 168. Ibid. 169. Ibid. at *5. 170. Ibid. at *6 (internal citations omitted). 171. Ibid. at *7. 172. Ibid. 60 JAMES KEYTE ET AL.

173. Ibid. (citing Comcast, 133 S. Ct. at 1433). 174. To date, Rail Freight is the most defendant-friendly antitrust case applying post-Comcast rigor to proof of damages in class certification cases. There are also a number of precedents in other areas of the law discussed below, which involve rigorous damages analysis and which we expect will inform antitrust class certification analysis. 175. No. 3:10-CV-0591, 2013 WL 1316452 (N.D.N.Y. March 29, 2013). 176. Ibid. at *3. 177. Ibid. 178. No. 5:09-323-KKC, 2013 WL 2285163 (E.D. Ky. May 22, 2013). 179. Ibid. at *1. 180. Ibid. at *6 (citation omitted). 181. Ibid. 182. Bankruptcy No. 7-04-17866-TL, 2013 WL 2244216 (D.N.M. May 21, 2013). 183. Ibid. at *7. 184. Ibid. See also Smith v. Family Video Movie Club, Inc., No. 1 CV 1773, 2013 WL 1628176, at *5 (N.D. Ill. April 15, 2013) (“[After Comcast] damages must be susceptible to measurement across the entire class, and individual damage calculations cannot overwhelm questions common to the class.”); Phillips v. Asset Acceptance, LLC, No. 09 C 7993, 2013 WL 1568092, at *3 (N.D. Ill. April 12, 2013) (“[Comcast] may portend a tightening of class certification standards more generally, particularly as to the circumstances under which the task of measuring damages sustained by absent members destroys predominance under Rule 23(b)(3).”). 185. See Munoz v. PHH Corp., 2013 WL 2146925, at *24 (E.D. Ca. 2013) (“The Comcast decision does not infringe on the long-standing principle that individual class member damage calculations are permissible in a certified class under Rule 23(b)(3).”); In re High-Tech Employee Antitrust Litig., 2013 WL 1352016 (N.D. Cal. 2013) (quoting the Comcast dissent for the principle that “individual damage calculations do not preclude class certification under Rule 23(b)(3)” and finding that Plaintiffs satisfied their burden on damage calculations by providing a method of calculating damages consistent with their theory of liability); Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) In re Diamond Foods, Inc., Sec. Litig., 2013 WL 1891382 (N.D. Ca. 2013) (withhold- ing judgment on whether Comcast requires class certification to be denied absent affirmative evidence that damages are susceptible of measurement across the entire class); Harris v. comScore, Inc., 2013 WL 1339262 (N.D. Ill. 2013) (finding that “factual damages issues do not provide a reason to deny class certification when the harm to each plaintiff is too small to justify resolving the suits individually.”). In Harris, however, the Court cited Butler v. Sears, Roebuck & Co., 702 F.3d 359, 362 (7th Cir. 2012), for the principle that efficiency concerns trump factual damages questions. Butler, as discussed in section IV.B.1.a, supra, was vacated and remanded by the Supreme Court on June 3, 2013, two months after the Harris opinion. It is unclear if the Court would have resolved the issue the same way if it had known the Supreme Court would remand Butler. Compare Wang v. Hearst Corp., 2012 WL 1903787, at *9 (S.D.N.Y. 2013) (“Indeed, although one could certainly quibble with the significance of a grant, vacate, and remand (‘GVR’) order from the Supreme Court, one must pause at least for a moment when one sees that the Supreme Court, From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 61

‘in light of Comcast’, has issued an order vacating and remanding a Seventh Circuit’s decision affirming the district court’s certification in an overtime misclassification case.” (quoting RBS Citizens, N.A. v. Ross, ——— U.S. ———, 133 S. Ct. 1722, ——— L.Ed.2d ——— (2013)). 186. 2013 WL 2306567 (9th Cir. 2013). 187. Ibid. at *514. 188. Ibid. 189. CIV.A. 11-11313-DPW, 2013 WL 1320454 (D. Mass. March 28, 2013). 190. Ibid. at *8 n.3 (citing Newton v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 259 F.3d 154, 187 (3d Cir. 2001)). See also Broussard v. Meineke Discount Muffler Shops, Inc., 155 F.3d 331, 342 343 (4th Cir. 1998); Windham v. Am. Brands, Inc., 565 F.2d 59, 68 (4th Cir. 1977).À 191. Ibid. 192. Kaufman (2013). 193. Paley and Ryan (2013). 194. Ibid. 195. Ibid. 196. Ibid. 197. Mason, Little, and Yeroshalmi (2013). 198. No. 3:10-CV-0591, 2013 WL 1316452 at *3 (N.D.N.Y. March 29, 2013). 199. Ibid. 200. Ibid. 201. 146 F. App’x 783 (6th Cir. 2005). 202. Ibid. at 791. 203. Ibid. 204. Kazan and Vasquez (2013). 205. Piggly Wiggly Clarksville, Inc. v. Interstate Brands Corp., 100 F. App’x 296 (5th Cir. 2004). 206. Piggly Wiggly Clarksville, Inc. v. Interstate Brands Corp., 215 F.R.D. 523 (E.D. Tex. 2003). 207. Ibid. at 527. 208. Ibid. at 528. Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) 209. Ibid. at 530. 210. See ibid. at 531. 211. Ibid. (citing Windham v. Am. Brands, Inc., 565 F.2d 59, 67 68 (4th Cir. 1977); In re Beef Indus. Antitrust Litig., 1986 WL 8890, *1 (N.D. Tex.À 1986); Dry Cleaning & Laundry Inst. of Detroit, Inc. v. Flom’s Corp., 1993 WL 527928, *3 (E.D. Mich. 1993)). 212. On this point, see also section IV.C.1, supra. 213. Ibid. (quoting Castano, 84 F.3d at 745 n.19) (citation omitted). 214. Ibid. 215. Piggly Wiggly Clarksville, Inc. v. Interstate Brands Corp., 100 F. App’x 296 (5th Cir. 2004). 216. Ibid. at 297 (citations omitted). 217. 93 F.R.D. 506 (E.D. Mich. 1982). 218. Ibid. at 507. 219. Ibid. at 509. 62 JAMES KEYTE ET AL.

220. Ibid. at 511. 221. Ibid. 222. Ibid. 223. 91-CV-76072-DT, 1993 WL 527928 (E.D. Mich. October 19, 1993) (hereinafter “Dry Cleaning”). 224. Ibid. at *4. 225. 339 F.3d 294 (5th Cir. 2003). For discussion of Bell Atlantic in the Daubert context, see section II.B.4(b), supra. 226. Ibid. at 297. 227. Ibid. at 299 300. 228. Ibid. at 300.À 229. Ibid. at 307. 230. Ibid. at 306, 307. 231. Ibid. at 307. 232. Comcast, 133 S. Ct. 1426, 1434 1435 (2013). 233. See generally In re Whirlpool Corp.À Front-Loading Washer Products Liability Litigation, 678 F.3d 409 (6th Cir. 2012); Butler v. Sears Roebuck & Co., 702 F.3d 359 (7th Cir. 2012). 234. See section II.B.4(b), supra. 235. Bell Atl. Corp. v. AT&T Corp., 339 F.3d 294, 307 (5th Cir. 2003). 236. Ibid. at 300. 237. Ibid. at 307. 238. Dry Cleaning & Laundry Inst. of Detroit, Inc. v. Flom’s Corp., 91-CV-76072- DT, 1993 WL 527928, at *5 (E.D. Mich. October 19, 1993). 239. Ibid. at *6. 240. 268 F.R.D. 573, 590 (N.D. Ill. 2009). For a discussion of Advocate Health Care in the Daubert context, see section II.B.4(B), supra. 241. Ibid. at 592 (citation omitted).

Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) ACKNOWLEDGMENTS

The authors would like to thank Marissa Troiano and Mike Folger, associates in the group, for their invaluable assistance in preparing this article.

REFERENCES

Areeda, P. E., & Hovenkamp, H. (1998 2007). Antitrust law: An analysis of antitrust principles À and their application (3rd ed., supp. 2008 ¶398b). Beisner, J. H., Miller, J. D., & Wyatt, G. M. (2013). From cable TV to washing machines: The Supreme Court cracks down on class actions. Class action litigation Report, 14 Class 616. Burtis, M. M., & Neher, D. V. (2011). Correlation and regression analysis in antitrust class certiﬁcation. Antitrust Law Journal, 77, 495, 511 512. À From Hydrogen Peroxide to Comcast: The New Rigor in Antitrust Class Actions 63

Fisher, F. M. (1986). Statisticians, econometricians, and adversary proceedings. Journal of the American Statistical Association, 81, 277, 282 285. À Forister, E. F., & Hussain S. (2010). Empirical approaches in assessing class certification in direct purchaser price-fixing cases, The Antitrust Review of the Americas, 22. Johnson, J. H., & Leonard, G. K. (2008). In the eye of the beholder: Price structure as junk science in antitrust class certification proceedings. Antitrust, 22, 108, 111. Kaufman, T. (2013, May 29). Ninth circuit rules that Comcast does not kill wage and hour class actions, Labor & Employment Blog. Retrieved from http://www.laboremploymen- tlawblog.com/wage-and-hour-ninth-circuit-rules-thatcomcast-does-not-kill-wage-and- hour-class-actions.html Kazan, B. M., & Vasquez, G. Y. (2013). Viability of Rule 23(b)(3) Cases after “Dukes,” “Amgen” and “Comcast”: Court carves narrow path to class certification. New York Law Journal, June 10. Mason, C. M., Little, D. H., & Yeroshalmi, S. (2013). Supreme Court addresses problems of size: “Too big” and “too small” cases pose a struggle, New York Law Journal, June 10. Mass Tort Defense (2013, June 13). Supreme Court remand two class actions in light of Comcast. McLaughlin, J. M. (2013). Litigating the effect of “Comcast v. Behrend”, New York Law Journal, June 13. Retrieved from http://www.newyorklawjournal.com/PubArticleNY. jsp?id=1202603956620&Litigating_the_Effect_of_Comcast_v_Behrend&slreturn=20130 631142303 Paley, A., & Ryan, P. (2013). Ninth circuit reverses denial of class certification where determining damages is a purely mechanical exercise. Wage & Hour Litigation Blog, June 4. Retrieved from http://www.wagehourlitigation.com/rule-23-class-certification/ninth- circuit-reverses-denial-of-class-certification-wheredetermining-damages-is-a-purely- mechanica-1/ Thompson, M. J. (Moderator), Bloch, R. E., & Guerin-Calvert, M. E. (2010). Anatomy of an information exchange healthcare antitrust case. AHLA Seminar Papers (Vol. 2010, Issue 1), Antitrust in Healthcare Conference, May 24. Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) This article has been cited by:

1. Joshua P. DavisThe class cert games: Coach, commentator, or critic? 65-76. [Abstract] [Full Text] [PDF] [PDF] Downloaded by 162.90.146.67 At 08:15 16 July 2015 (PT) Remedies Under Frederic Depoortere the EUMR

Skadden, Arps, Slate, Meagher & Flom Giorgio Motta

An analysis of the European Commission merger decisions in the Freescale Semiconductor) and seven out of 12 Phase II conditional period 2014-2015 confrms and strengthens some important trends in approvals (PRSfM/GEMA/STIM – JV, Hutchison 3G UK/Telefonica the EU’s approach to remedies. On the one hand, the Commission Ireland, Liberty Global/Ziggo, Telefonica Deutschland/E-Plus, demonstrated its willingness to approve very complex merger cases Liberty Global/Corelio/W&W/De Vijver Media, DEMB/Mondelez/ by imposing often equally complex remedies. On the other hand, Charger OPCO, Orange/Jazztel) involved remedies that fell short the scope of the remedies imposed remains very broad, most often of full divestitures or were a combination of divestitures and other exceeding the areas of competitive concern in order to ensure the non-structural or quasi-structural remedies. viability of the divested business. The Commission has a wide For instance, in Chiquita Brands International/Fyffes, the parties margin of discretion in the design and evaluation of merger remedies offered to terminate an exclusivity clause with a shipping company and readily uses it. A more recent trend is that the Commission is and refrain from agreeing similar exclusivity provisions for a period showing a growing preference for up-front buyer solutions, obliging of ten years, in order to prevent the alleged foreclosure of competing the merging parties to postpone closing of the notifed transaction banana importers’ access to third-party shipping services. Similarly, until a buyer has been found and approved for the divestiture business in PRSfM/GEMA/STIM – JV, in order to address the Commission’s offered as a remedy for the Commission’s approval. concern that the proposed joint venture between the three music In the period 2014-2015 (1 January 2014-17 September 2015), the collecting societies would raise barriers to competitors’ entry into Commission issued 826 Phase I and 15 Phase II decisions. Of these the market for copyright administration services provided to ‘Option decisions 34 involved conditions (22 in Phase I and 12 in Phase II, 3 publishers’, the parties committed that the joint venture would representing 3% and 80% of Phase I and II decisions, respectively). not enter into exclusive contracts with its customers for copyright No mergers were prohibited during this period, although one administration services. notifed concentration, Teliasonera/Telenor/JV, was withdrawn on In a number of cases, the Commission accepted commitments to 11 September 2015, less than a month before the Commission’s provide access to infrastructure or services on FRAND terms deadline for issuing a decision. (SNCF Mobilities/Eurostar International Limited, PRSfM/GEMA/ In contrast, in the period 2012-2013, the Commission issued more STIM – JV) or to terminate non-compete clauses (Liberty Global/ than 860 Phase I decisions, of which 20 (or 2%) were conditional Ziggo) as a means to lowering barriers to entry for competitors. approvals, and 14 Phase II decisions, of which eight (or 57%) were Commitments to supply or license were accepted to remedy vertical conditional approvals. Three decisions were Article 8 (3) prohibitions. foreclosure concerns. For instance, in Airbus/Safran, Airbus Therefore, the most recent period marks a substantial increase of committed to supply satellite and space vehicle components to Phase II conditional approvals but no (or fewer) prohibition decisions. competing satellite prime contractors; in IMS Health/Cegedim Business, IMS offered access to its “brick structure” to competitors active in the downstream market for primary market research; and in Preference for divestitures, but fexible Liberty Global/Corelio/W&W/De Vijver Media, the parties agreed approach to non-structural remedies to license two popular TV channels to competing downstream TV distributors in Belgium. The Commission continued to show a very strong preference for structural remedies, and more specifcally, divestitures. Acting Deputy Director General for mergers Carles Esteva Mosso recently Broad discretion in the remedy design stressed the basic rule in the Remedies Notice that structural remedies are the Commission’s preferred solution, as they allow a “singular, one-time intervention” on the regulator’s part. (Carles Viability of the divested business trumps Esteva Mosso, speech in New York, 18 September 2015). proportionality of the remedy Nevertheless, a number of non-structural remedies were accepted According to the Commission’s Remedies Notice, the divested in recent Phase I and Phase II decisions. More specifcally, in the business has to include all the assets and personnel necessary 2014-2015 period, seven out of 22 Phase I conditional approvals to ensure its viability and competitiveness. Even though the (Chiquita Brands International/Fyffes, CSAV/HGV/Kuhne/ Commission is bound by the principle of proportionality when Maritime/Hapag-Lloyd AG, ALITALIA/ETIHAD, IMS Health/ assessing a remedy (see in particular recital 30 of the EU Merger Cegedim Business, Airbus/Safran/JV, SNCF Mobilities/Eurostar Regulation, according to which “…commitments should be International Limited, IAG/AER Lingus and NXP Semiconductors/ proportionate to the competition problem and entirely eliminate

10 WWW.ICLG.CO.UK ICLG TO: MERGER CONTROL 2016 Skadden, Arps, Slate, Meagher & Flom Remedies Under the EUMR it”), there is often a tension between these two principles, whereby engineers. Notably, the remedy package included the divestiture the remedies imposed show that viability tends to take precedence of long-term servicing agreements for 34 installed GT 26 turbines, over proportionality: the scope of the accepted remedy routinely even though the Commission’s press release on the case does not exceeds the competition issues that the commitments are designed identify any competition concerns in the market for servicing to remedy. Alstom’s gas turbines. Several decisions issued within the period 2014-2015 highlight the Last, in a number of cases in the telecommunications sector, the Commission’s generous interpretation of the principle of viability of Commission attempted not only to eliminate the competitive the divested business. overlaps, but also to effectively improve the position of actual For example, in Merck/Sigma-Aldrich, the Commission imposed competitors and facilitate entry into concentrated national telecoms the divestiture of worldwide rights and worldwide customer base markets. In a speech delivered in September 2014, former of Sigma’s Fluka and associated brands in relation to solvents and Commissioner for Competition Joaquín Almunia stressed the inorganics, although the competition concerns in these product prominent role of remedies in rectifying the lack of a single market: markets were limited to the EEA. The Commission noted that this “[i]n the absence of a genuine internal market for telecoms, when measure enhanced the long-term viability of the divested business recently reviewing some mergers we had to take account of the fact as it mitigated any risk of brand confusion. (see para. 250) that competition in this industry still takes place on a national basis and that remedies were needed to preserve competitive prices for In Holcim/Lafarge, the Commission identifed competition concerns consumers, and not only the need to fnance fresh investments.” in construction materials, such as cement, aggregates, ready-mix Carles Esteva Mosso also commented that, “[w]e can’t change concrete on customers located near Holcim and Lafarge’s production market defnitions. But we can change markets. We can build an facilities. To approve the transaction in Phase I, the Commission internal telecoms market, and an internal energy market. These will accepted a divestment package that included Lafarge’s businesses in be priorities for the next Commission. And this may eventually open Germany, Romania and the UK, and Holcim’s operations in France, the doors to increased consolidation in these sectors.” (See Joaquin Hungary, Slovakia, Spain and the Czech Republic. The remedies Almunia, speech of 23 September 2014, and Carles Esteva Mosso, comprised assets as well as the services necessary to ensure the speech of 11 September 2014). viability of the business. Commissioner Vestager admitted that, “the remedy package in this case is very substantial. But that was In particular, in the Hutchison 3G UK/Telefonica Ireland, the necessary to allow a clear cut decision already in frst phase.” This Commission concluded that the merger would bring together the statement of the Commissioner refects the position in the Remedies second and the fourth largest mobile network operators (MNO) Notice that commitments offered in Phase I, or in the initial stage of in Ireland, and remove H3G as an important competitive force on Phase II before the Commission issues a Statement of Objections, the Irish market for retail mobile telecommunications services and must be suffcient to clearly rule out ‘serious doubts’ – however, this the wholesale market for access and call origination in Ireland. To does not mean that remedies agreed at the end of a full Phase II are address the concerns at the retail level, H3G offered to sell up to more narrow in scope. 30% of the merged company’s network capacity in Ireland at fxed payments to two mobile virtual network operators (MVNOs), with In Huntsman Corporation/Equity Interests Held by Rockwood an option for one of them to become a full MNO by acquiring Holdings, Huntsman agreed to divest its global TR52 business, even spectrum from H3G at a later stage. As regards wholesale access though the Commission’s decision identifed competition concerns and call origination market, H3G offered to continue the network in the market for titanium dioxide for printing ink applications sharing agreement with Eircom (the fourth and smallest competitor only in the EEA. The divestment package (offered but rejected of the four Irish MNOs) on improved terms, allowing that operator in Phase I, and improved in Phase II) included the global TR52 to become a credible network host for MVNOs in Ireland. brand, technology and know-how, customer arrangements and key personnel. The Commission concluded that the expanded scope In Telefonica Deutschland/E-Plus, the Commission imposed similar of the remedy (coupled with an up-front buyer commitment) was clearance conditions: the transaction would combine the third and necessary for the viability of the divested business. fourth largest MNOs in Germany and impede competition among the remaining three competitors. Telefonica agreed to sell up to 30% In Ineos/Solvay/JV and GE/Alstom, the remedy packages were of the merged company’s network capacity to one or several (up to fnalized after the Statement of Objections and oral hearing. three) MVNO(s) in Germany at fxed payments. This remedy ensured Notably, in both cases the Commission noted that the objective of that up to three MVNOs would enter or expand in the German retail the remedies was not only to eliminate the competition concerns, mobile telephony market with the necessary degree of certainty. In but to “replicate” Solvay in the market for commodity S-PVC in addition, the parties offered to divest radio wave spectrum and certain North West Europe and Alstom in the heavy duty gas turbines assets either to a new MNO entrant or to the MVNO(s) who would market respectively. take up the divested network capacity, with the aim to facilitate the In Ineos/Solvay/JV, the Commission stressed that the viability of entry or enable the development of a new MNO in the German market the remedy was essentially determined by the geographic location in the future. Notably, the package also included wholesale access of the divested plants, their degree of vertical integration and access to 4G services to all interested MVNOs and service providers in the to key inputs. As a result, the parties offered to divest not only future. Again, the objective was to improve the position of German Ineos’ S-PVC plants in Germany, France and the Netherlands, but MVNOs and service providers to whom Telefonica or E-Plus granted also its upstream chlorine and ethylene dichloride production assets wholesale access to 2G and 3G services. in the UK and Belgium, and enter into a joint venture agreement Last, in Orange/Jazztel, the Commission found that the transaction with the purchaser for the production of chlorine. The divestment would impede competition in the retail markets for fxed internet package thus extended to the upstream activities and comprised a access services in Spain. The remedy extended beyond the overlap fully integrated self-standing S-PVC business. In GE/Alstom, the market, as Orange offered not only to divest an independent Fibre- parties committed to divest Alstom’s GT26 and GT36 heavy duty To-The-Home network, but also to grant to the purchaser wholesale gas turbine business, including key personnel, as well as existing access to its mobile network, including 4G services, even though no upgrades and pipeline technology for future upgrades and R&D competition concerns were identifed in the mobile telecoms market. The Commission considered this solution necessary for the viability

ICLG TO: MERGER CONTROL 2016 WWW.ICLG.CO.UK 11 Skadden, Arps, Slate, Meagher & Flom Remedies Under the EUMR

of the new entrant, given that the vast majority of fxed internet (ii) on certain occasions involving the provision non-discriminatory contracts in Spain are bundled with a mobile component. access to infrastructure or networks (para. 64); (iii) in cases which On the other hand, the Commission has been careful to limit the cause considerable risks of preserving the competitiveness and scope of remedies to merger-specifc competition issues. In saleability of the divestment business in the interim period until particular, in SSAB/Rautaruukki, the proposed transaction would divestiture (para. 55) and (iv) in cases where there are considerable have combined the two market leaders in the production and obstacles for a divestiture, such as third party rights, or uncertainties distribution of fat carbon steel products in Nordic countries. The as to fnding a suitable purchaser. (para. 54). The last point shows parties committed to divest a signifcant part of their distribution that there is a correlation between the viability of the remedy and the network. The Commission was satisfed with the remedy package, up-front buyer solution. and dismissed customer complaints about the adequacy of the remedy, because these concerns were likely caused by the pre- Up-front buyer in recent divestiture remedies existing possible dominant position of Ruukki. The Commission noted that “[r]emoving such possible dominance, which is not Over the past few years, an increasing number of conditional merger-specifc (in the sense of not being caused by the merger), approvals contain an up-front buyer commitment, despite statements would be beyond what can be asked by way of commitments in a of Commission offcials that they remain the exception. In a recent merger control procedure.” (para. 244) panel discussion, acting Deputy Director General for mergers Carles Esteva Mosso mentioned, “[w]e continue to believe our default Limited judicial review enhances the Commission’s remedy should be a divestiture the parties can conclude after they margin of discretion close the merger,” and “Only 25 per cent of the divestitures required by DG Comp have included up-front buyers.” (Carles Esteva Mosso, speech of 18 September 2015). The general absence of judicial review of commitment decisions in favour of notifying parties allows the Commission to leverage its Up-front solutions are prevalent in Phase II decisions. In the period margin of discretion in the design, evaluation and implementation 2014-2015, the Commission accepted an up-front commitment of merger remedies. in seven out of 12 Phase II conditional approvals (or 58%), a substantial increase from the period 2012-2013 (two out of eight or First, the European Courts have confrmed that the Commission 25% of Phase II conditional approvals). The same trend is observed enjoys a wide margin of discretion in the evaluation of remedies in Phase I conditional approvals. In the period 2014-2015, four out inherent to complex assessments of an economic nature, and thus the of 22 Phase I (or 18%) involved an up-front commitment, but none scope of judicial review of remedy decisions is limited to manifest in the period 2012-2013. errors of assessment. According to established case-law, the fact that other commitments might also have been accepted, or might even Up-front buyer solutions were also offered in Phase I cases Crown have been more favourable to competition, cannot justify annulment Holdings/MIVISA, Holcim/Lafarge, IMS Health/Cegedim Business of that decision in so far as the Commission was reasonably entitled and Merck/Sigma-Aldrich, and in Phase II cases Ineos/Solvay/JV, to conclude that the commitments set out in the decision served to Hutchison 3G UK/Telefonica Ireland, Telefonica Deutschland/E- Plus, Huntsman Corporation/Equity Interests Held by Rockwood dispel the serious doubts. The General Court reiterated this position Holdings, Zimmer/Biomet, Orange/Jazztel and GE/Alstom. The up- in its judgment of 5 September 2014 in Éditions Odile Jacob SAS v front buyer clause was an additional means to ensure viability of European Commission (Case T-471/11 at para. 146) the divested business, especially in cases where the identity of the Second, the short deadlines for notifying parties to appeal a purchaser was important, and incentivise the parties to implement Commission decision, including conditional merger clearance substantial and complex remedy packages as soon as possible. A decisions, do not offer the parties concerned to fully assess the few examples are provided below. complexities associated with the implementation of the remedy or, In Huntsman Corporation/Equity Interests Held by Rockwood more importantly, the impact of the remedy on their business before Holdings, the Commission’s market test indicated that the divested deciding whether or not to appeal. business would be viable, if a suitable purchaser acquired it and Lastly and perhaps most importantly, if the notifying parties were developed its presence in the market. As a result, the parties offered successful in challenging and the Court annuls the Commission’s an up-front buyer commitment and additional purchaser criteria decision, the merging parties would lose their EUMR approval and to that effect: the producer should be a sulphate-based titanium would need to restart the Commission approval process from zero. dioxide producer, have the capacity and scope to meet current and The uncertainties surrounding such outcomes create a very strong reasonably foreseeable TR52 demand worldwide and the ability disincentive for the merging parties to attack a conditional approval to distribute TR52 in the EEA or to acquire this ability within the decision, even if they disagree with the scope or modalities of the transitional period. conditions involved. In Merck/Sigma-Aldrich, the Parties offered an up-front buyer commitment taking into account the feedback from the Commission’s Up-front buyer market test, which attached particular importance to the identity of the purchaser. More specifcally, the customer responses indicated that the purchaser of the divested business should already be active Increased number of up-front cases in the laboratory chemicals business, to ensure its credibility vis-à- vis customers, and have a product portfolio in life science, which should include but go beyond the problematic products to further With an up-front buyer commitment, the parties commit not to strengthen the sustainability of its activities on a long term basis. close the notifed transaction until they have entered into a binding agreement with a purchaser for the divested business, approved Finally, in IMS Health/Cegedim Business the parties introduced by the Commission. Under the Commission’s Remedies Notice, an up-front solution to address the Commission’s concerns about the Commission considers that an up-front solution is preferred on the effective implementation of the remedy, owing to the limited certain occasions, and in particular: (i) carve-out remedies (para. 36); interest that purchasers who met the specifed criteria showed in acquiring the divestment business.

12 WWW.ICLG.CO.UK ICLG TO: MERGER CONTROL 2016 Skadden, Arps, Slate, Meagher & Flom Remedies Under the EUMR

Working Group (MWG) is actively engaged in developing common Problems arising with the up-front approach guidelines with the goal of facilitating the agencies’ coordination in the area of remedies. Up-front provisions are designed to incentivise the parties to In May 2015, the ICN’s MWG issued the Practical Guide to implement the remedy within a short time frame from approval. At International Enforcement Cooperation in Mergers (the “Practical the same time, however, they create signifcant additional pressure Guide”). The Practical Guide is intended to serve as a voluntary on the merging parties as they further prolong the period during and fexible framework for interagency cooperation in merger which the transaction cannot be completed, resulting regularly investigations. One of the sections of the Practical Guide deals in considerable additional costs in terms of fnancing, as well as with cooperation in remedy design and remedy implementation. negative commercial, business and personnel effects. Specifcally: ■ With respect to the design of a remedy, the Practical Guide Multi-agency Cooperation in Remedy Cases invites agencies to consult with each other on issues such as the anticipated type of remedy (structural or behavioural One other important recent trend is the increasing level of remedy), the potential product and geographic scope of the remedy, the anticipated process and timelines for the cooperation between the European Commission and other agencies execution of the remedy. in the context of remedies in multi-jurisdictional mergers. ■ With respect to the implementation of a remedy, the Practical The extent of cooperation depends largely on the nature and scope Guide recommends close cooperation amongst agencies on of the antitrust concerns identifed by each agency. Cooperation issues such as the selection and approval of the monitors/ is particularly important in cases that raise issues of common trustees, the hold separate managers and the potential concerns, such as in transactions with global or cross-border purchasers of the same divestment business. regional markets. However, cooperation is also benefcial in cases It remains to be seen how these guidelines will be implemented in with distinct national or regional markets to the extent that there practice. Another question is how one agency’s practices will affect may be overlapping remedies or where remedies accepted in one other agencies, through increased coordination. In view of the wide jurisdiction may have an impact in another jurisdiction. discretion the European Commission enjoys in developing remedies This is a welcome approach: cooperation decreases the likelihood as conditions for its approval decisions, the issue of interagency of conficting or divergent remedies being presented to and accepted effects may become a critical question in the future. by the different agencies and reduces the risk of implementation of these remedies. Acknowledgment This is a well-established practice in transactions subject to EU and U.S. merger control approvals. The European Commission and the Athanasia Gavala was a co-author of this article. She is an U.S. antitrust agencies (DOJ and FTC) regularly coordinate and associate in Skadden’s Brussels offce, focusing on European keep each other informed of remedies that they are considering, Union and international competition law. She focuses on merger and of the related discussions with the parties. This is particularly control matters and has worked with clients from a diverse range relevant because there are some important differences in the remedy of industries including consumer products, energy, entertainment, procedures under the EU and U.S. merger control laws. fnancial services, pharmaceuticals, retail and telecommunications. Cooperation between the European Commission and other agencies She recently advised General Electric in the landmark acquisition of is also becoming increasingly common. These efforts are part of Alstom Power & Grid business before the European Commission. the bilateral relationships between the European Commission and other agencies, and more recently, in the context of the activities of the ICN (International Competition Network). The ICN’s Merger

ICLG TO: MERGER CONTROL 2016 WWW.ICLG.CO.UK 13 Skadden, Arps, Slate, Meagher & Flom Remedies Under the EUMR

Frederic Depoortere Giorgio Motta Skadden, Arps, Slate, Meagher & Flom Skadden, Arps, Slate, Meagher & Flom Avenue Louise 523 Avenue Louise 523 Brussels 1050 Brussels 1050 Belgium Belgium

Tel: +32 2 639 0334 Tel: +32 2 639 0314 Fax: +32 2 641 4034 Fax: +32 2 641 4014 Email: [email protected] Email: [email protected] URL: www.skadden.com URL: www.skadden.com

Frederic Depoortere is a Partner in Skadden Arps’ Brussels offce. Mr. Giorgio Motta is Counsel in Skadden Arps’ Brussels offce. Mr. Motta Depoortere joined the Brussels offce of Skadden, Arps in 1996. He joined the Brussels offce of Skadden, Arps in 1999. Mr. Motta has deals primarily with European Union (EU) and international antitrust a wide-ranging experience in European Union (“EU”), Italian and and regulatory aspects of mergers, acquisitions and joint ventures. international antitrust merger control and cartel enforcement. Mr. Depoortere has worked on numerous international transactions Mr. Motta advises clients on antitrust aspects of mergers, acquisitions requiring global antitrust and merger control analysis and notifcations and joint ventures. He has worked on numerous transactions requiring for international clients. He also deals with general EU competition international antitrust merger control approvals both in Europe and law issues such as cartels, vertical restraints and dominance. on a worldwide basis for companies in the energy, telecom, fnancial Mr. Depoortere is a graduate of the Catholic University of Leuven in services, pharmaceutical, consumer goods, and many other industries. Belgium and completed part of his law studies in Ghent (Belgium) and Mr. Motta also advises clients in cartels, as well as EU and Italian Strasbourg (France). He holds an LL.M. degree from the University competition law issues relating to vertical restraints and dominance. of Chicago Law School and is a member of the Brussels and New He has represented clients in Article 101 investigations in relation to York Bars. cartels, trade association membership, strategic alliances, distribution Mr. Depoortere repeatedly has been selected for inclusion in arrangements and other vertical agreements, both before the Chambers Global: The World’s Leading Lawyers for Business. He European Commission and the Italian Competition Authority. was included in Global Competition Review’s list of the world’s leading Mr. Motta advises clients also on a broad range of other EU law issues, 40 competition lawyers under the age of 40. Mr. Depoortere also including in the areas of EU state aid and public procurement. was named a leading practitioner in his feld by Who’s Who Legal: Competition Lawyers & Economists. Mr. Motta is also a Non-Governmental Advisor (NGA) of the Italian Competition Authority for the activities of the International Competition Network – ICN.

Skadden is one of the largest law frms in the world, serving clients in every major fnancial centre with over 1,700 lawyers in 23 locations. For almost 60 years, Skadden has been providing a wide array of legal services to the corporate, industrial, fnancial and governmental communities. Our clients range from start-up companies to the largest global corporations.

Skadden’s Antitrust and Competition Group is a global leader in its feld. Chambers Global: The World’s Leading Lawyers for Business recognises Skadden as one of the top-tier frms in the area of antitrust and competition. Skadden’s European competition law practice advises and represents clients on a wide variety of cutting edge EU competition law issues, including both conduct cases (abuse of dominance proceedings under Article 102 TFEU and cartel proceedings under Article 101 TFEU), as well as mergers and acquisitions. Our attorneys work closely with in-house counsel to advise on compliance and defend against enforcement actions brought by the European Commission or Member State authorities and, where necessary, represent clients in appeals before the European courts.

14 WWW.ICLG.CO.UK ICLG TO: MERGER CONTROL 2016