skip to main content

The Paul, Weiss Antitrust Practice advises clients on a full range of global antitrust matters, including antitrust regulatory clearance, government investigations, private litigation, and counseling and compliance. The firm represents clients before antitrust and competition authorities in the United States, the European Union, the United Kingdom and other jurisdictions around the world.

Class Certification Case Law Updates

November 5, 2020 Download PDF

The past year has seen several cases delve into the role that both representative evidence and uninjured class members play in determining whether predominance has been satisfied within the broader class certification analysis. In this installment of client alerts focused on class certification, we discuss some of the noteworthy developments within these cases below.

Representative Evidence—or the Lack Thereof

Several recent cases have reconsidered the role of representative evidence in their class certification analyses, and in what circumstances it is sufficient to show predominance. Most notably, the Third Circuit in In re: Lamictal Direct Purchaser Antitrust Litigation[1] strongly affirmed the rule that outside of a Fair Labor Standards Act case (such as in Tyson Foods[2]) where representative evidence is the only way to show liability, plaintiffs must still demonstrate that their claims are capable of common proof at trial by a preponderance of the evidence. Some background first: 

Tyson Foods. In 2016, the Supreme Court held in Tyson Foods that the plaintiffs were permitted to use a representative study to show that the class members in an FLSA action had worked over forty hours per week without being properly compensated for that time. There, the defendant-employer did not record (or compensate for) the time pork processing plant employees spent donning and doffing protective gear.[3] As a result, the employees had to rely on representative evidence, such as employee statements, video recordings of gearing up and gearing down, and an expert study to demonstrate the amount of time for which they were uncompensated.[4] In evaluating whether this evidence could be used to satisfy the predominance requirement of class certification, the Supreme Court noted that the district court could have denied class certification for lack of predominance “only if it concluded that no reasonable juror could have believed that the employees spent roughly equal time donning and doffing” their equipment once such evidence was ruled admissible.[5] However, the Court expressly declined to issue a “broad and categorical rule” governing the use of representative evidence, explaining that determinations on whether a representative sample could be used to establish class wide liability would depend on the purpose and the underlying cause of action.[6]

Lamictal. On April 22, 2020, the Third Circuit clarified that the “no-reasonable-juror” evidentiary standard described in Tyson Foods is unique to the context of an FLSA claim in In re: Lamictal Direct Purchaser Antitrust Litigation.[7] The opinion highlights that class action plaintiffs outside of the FLSA context of Tyson Foods must still show, by a preponderance of the evidence, that they can establish class-wide injury through common proof at trial.

The claim at issue in Lamictal is an antitrust claim known as a “pay-for-delay.” Plaintiffs, a class of purchasers of the epilepsy drug lamotrigine, alleged that a patent litigation settlement between manufacturers of the brand name and generic versions of lamotrigine, GSK and Teva respectively, unlawfully incentivized Teva to delay the launch of its generic version.[8] The plaintiffs further alleged that, as a result, consumers were charged artificially high prices for brand Lamictal and moved to certify a class of all companies that purchased brand Lamictal directly from GSK or generic lamotrigine from Teva, on the basis that, “on average, the price of a generic is lower when there are two generics rather than just one.”[9] GSK and Teva challenged certification only as to the class members who purchased generic lamotrigine from Teva.[10] Following the district court’s decision to certify the class, GSK and Teva appealed.

In their appeal, the plaintiffs contended that they did need not prove antitrust injury at the certification stage to meet predominance requirements, but merely had to establish that injury is capable of common proof at trial.[11] They further contended, based on a comment in Tyson Foods, that such predominance is established unless no reasonable juror could believe the common proof at trial[12] and that, so long as their evidence of class-wide antitrust injury could survive a jury finding, they had met the predominance requirement.[13]

The appeals court disagreed with the plaintiffs and stated, “contrary to the Direct Purchasers’ assertion, Tysons Foods does not control our case, and its no-reasonable-juror statement certainly does not overturn our longstanding rule announced in Hydrogen Peroxide” that district courts are required to resolve factual determinations by a preponderance of the evidence at the class certification stage.[14]

The Lamictal court distinguished the case before it from Tyson Foods by noting that FLSA cases are unique because often—due to inadequate record-keeping—a “representative sample [of employees] may be the only feasible way to establish liability.”[15] The Third Circuit further distinguished Tyson Foods because there, the court was asked to, but declined to, decertify a class after the jury had rendered a verdict in favor of the plaintiffs, whereas the district court in the Lamictal litigation reviewed the class certification motion “on a blank slate.”[16]

Accordingly, the appeals court vacated the class certification decision because the district court accepted plaintiff’s proposed common proof that relied on averaging, without a rigorous analysis. The appeals court noted that inquiry was needed into whether the market was “characterized by individual negotiations and a discounted-brand competition strategy,” which could mean that using averages would mask the fact that maybe plaintiffs suffered no overcharge.[17] The appeals court conceded that “averages may be acceptable where they do not mask individualized injury,” but held that it was impossible to know whether the averaging was acceptable here absent further analysis into the potential pricing strategies in the but-for world and a more rigorous analysis of the experts’ competing—and conflicting—evidentiary sources.[18] The certification was vacated and remanded.

Ferreras. Even in the context of FLSA cases, Tyson continues to undergo examination and refinement at the district and intermediate appellate court level. In another recent Third Circuit case, Ferreras v. American Airlines, Inc.,[19] the Third Circuit reversed certification of three subclasses because the classes could not establish commonality or predominance.

The district court certified three subclasses of airline employees who alleged that defendant American’s Newark Liberty International Airport station failed to compensate them for all their time spent clocked in to work. American’s timekeeping system was programmed to calculate pay for employees only for the duration of their assigned shifts, less an automatically deducted thirty-minute meal break.[20] Plaintiffs, however, alleged that because they often worked for periods that the timekeeping system did not account for, they should be compensated for all of their time, including: (1) time clocked in but which was pre- and post- their scheduled shift times; (2) scheduled meal breaks; and (3) time spent working but off-the-clock.[21]

The district court, relying on Tyson Foods, concluded that individualized variations between workers would not defeat class certification.[22] In the court’s view, plaintiffs met the Rule 23(b) commonality and predominance requirements because all plaintiffs shared questions of (1) whether they were not, in fact, compensated for their total hours reflected by the timekeeping system; and (2) whether they were discouraged from seeking an exception to allow them to be paid for time outside of their shift.[23]

The Third Circuit reversed, holding that the district court had erred in both its commonality and predominance analyses.

Commonality: The Third Circuit held that the district court erred in finding commonality by accepting that common questions existed, but not considering whether “a class-wide proceeding [could] generate common answers apt to drive the resolution of the litigation” and whether common issues “are more prevalent or important than the non-common, aggregation-defeating, individual issues.”[24] Regardless of any common evidence about American’s timekeeping system or their overtime policy, plaintiffs still needed to prove that each individual employee worked uncompensated overtime.[25] Further, despite the evidence that only one subset of employees suffered from a policy of not paying employees submitting requests for overtime, the district court still certified subclasses consisting of all hourly employees.[26] Because there was evidence in the record that not all employees worked during all meal breaks, and no easy measure–like time clock records–by which to compute the extent to which any particular employee worked off the clock existed, the individualized evidence needed to prove both claims defeated any commonality among the class.[27]

Predominance: The appeals court noted that because the plaintiffs failed the commonality requirement, they automatically failed the more challenging predominance requirement as well.[28] The Third Circuit also made a point of distinguishing Ferreras from Tyson Foods for purposes of predominance, noting that in Tyson Foods, the issue was how to compensate plaintiffs for the varied amount of time it took them to participate in the same act of putting on and removing their protective equipment.[29] In Ferreras, it was unclear whether the employees were working during the claimed uncompensated time and there was also variability in the acts plaintiffs performed during that time. As a result, the Third Circuit disagreed with the district court’s reliance on Tyson Foods and found that representative evidence was insufficient to prove the Ferreras plaintiffs’ case.[30]

Ridgeway. The Ninth Circuit also recently considered what constitutes a representative sample in another FLSA case in Ridgeway v. Walmart Inc.[31] In Ridgeway, long-haul truckers alleged that Wal-Mart was violating state meal and rest break laws, and that, as a result of Wal-Mart’s payment practices failing to account for time spent on required rest breaks and layovers, it failed to pay drivers minimum wage.[32] A jury found for the plaintiffs, and Wal-Mart appealed, arguing that the district court erred in certifying the class, and in awarding damages to plaintiffs based on layovers, rest breaks and inspections.[33]

In its review of the class certification and damages analysis of the trial court, the Ninth Circuit agreed with the district court finding that common issues predominated. Rejecting Wal-Mart’s argument that the plaintiffs could not use representative evidence to prove the elements of their case, the appeals court held that liability was suitable for class treatment on the basis of substantial supporting evidence showing that Wal-Mart owed class members minimum payment during layovers.[34] And once the jury found that minimum wages were owed, the varying amount of time spent on each task “went to the question of damages.”[35]

Wal-Mart also argued that classwide damages could not be established because of the “broad range of experiences among drivers” and the variation of time they spent on each task.[36] Although the appeals court said Wal-Mart’s argument as to damages was “more compelling,” it reiterated that “[t]ime and time again, this court has reaffirmed the principle that the need for individualized damages calculations does not doom a class action.”[37] Nonetheless, noting that plaintiffs still had to prove their damages, the court considered whether plaintiffs had done so through representative evidence.

Evaluating the issue through the lens of Tyson Foods,[38] the Ninth Circuit held that the plaintiffs put forth adequate evidence in this case, where a representative sample was the only practicable way for plaintiffs to present their evidentiary data. As the court held in Tyson Foods, representative evidence may include testimony, expert studies and other evidence, all of which the plaintiffs presented in this case.[39] Many plaintiffs testified about the length of their breaks and inspections, which the appeals court found to be “ample evidence” for the average time for each put forth by plaintiffs.[40] Additionally, the expert evidence submitted by plaintiffs could have been presented as corroborative evidence in any plaintiff’s individual suit had they moved forward individually rather than as a class.[41] The appeals court also noted that Wal-Mart failed to convince a jury not to extrapolate from the expert testimony.[42] Finally, Wal-Mart was unable to show a methodological error in the expert’s work, which was based on payroll data, data from Department of Transportation inspections, driver logs, questionnaires of a random sample of class members and a good deal of other data.[43] Plaintiffs therefore met their burden to show enough representative evidence to allow a jury to draw a reasonable inference about the unpaid time worked, and the district court did not err in granting class certification.[44]

As seen in Lamictal, Ferreras and Ridgeway, applications of Tyson continue to demonstrate a cautious approach to the use of representative evidence, and the outcome often turns on both the procedural posture of the case and the nature of the underlying claim.

Uninjured Class Members

In addition to developments regarding the use of representative evidence, courts continue to weigh in on other notable issues regarding class certification. The Eleventh Circuit and the Southern District of New York have both recently weighed in on the question of how courts should view the presence of uninjured class members with respect to Rule 23(b) predominance analyses. In Cordoba v. DIRECTV, LLC,[45] the Eleventh Circuit provided extensive analysis of how the uninjured class members’ lack of standing could undermine the predominance of the class issues. In In re Aluminum Warehousing Antitrust Litigation,[46] the Southern District of New York reiterated that faulty statistical models that failed to isolate uninjured class members could not meet the predominance requirement for class certification.

Cordoba. The plaintiffs in Cordoba alleged that DIRECTV and its telemarketing contractor, Telecel Marketing Solutions, violated the Telephone Consumer Protection Act by failing to maintain an internal Do-Not-Call list.[47] The United States District Court for the Northern District of Georgia certified two classes. One class included all individuals whose names were on the National Do-Not-Call Registry but who nonetheless received more than one DIRECTV telemarketing call from Telecel, comprising some 926 individuals. The second class included 16,870 people who simply received more than one DIRECTV telemarketing call from Telecel.

The defendants sought interlocutory appeal. The Eleventh Circuit Court of Appeals granted the appeal, but only as to whether persons who received a telemarketing call but who had never asked to be placed on a do-not-call list had standing, and whether that affected certification of the second class. The appeals court concluded that any member of the putative class that had neither joined the National Do Not Call Registry nor notified DIRECTV or Telecel that it did not want to receive marketing calls lacked standing because they were not harmed—to them, “it doesn’t make any difference that Telecel hadn’t maintained an internal do-not-call list.”[48] The court then had to address what the effect of having putative class members that lacked standing had on the predominance analysis.

The answer: “a powerful problem.”[49] It fed right into the Rule 23 inquiry of “how the class will prove causation and injury and whether those elements will be subject to class-wide proof.”[50] The appeals court noted that standing would be an “individualized issue” and that it was “one that the district court did not account for or consider in any way in deciding whether issues common to the class actually predominated over issues that were individualized to each class member.”[51]

In this regard, the Eleventh Circuit did not draw any bright lines regarding numbers or proportions or procedure, but focused instead on the underdeveloped record. It noted a need for more information on “two key questions: First, how many class members (or what proportion of them) asked Telecel not to call them anymore . . . [a]nd second, how do class members intend to prove that they made these requests?”[52] Notably, the appeals court noted that its decision did not mean “that a substantially similar class cannot be drawn or certified . . . or even that the class as certified by the district court was necessarily too broad.”[53] However, the court was troubled by the lack of information, particularly with regard to the number of uninjured putative class members. “A plaintiff need not prove that every member of the proposed class has Article III standing prior to certification, and in some cases a court might reasonably certify a class that includes some putative members who might not have satisfied the requirements of Lujan and decide to deal with that problem later . . . . But there is a meaningful difference between a class with a few members who might not have suffered an injury traceable to the defendants and a class with potentially many more, even a majority, who do not have Article III standing.”[54]

Aluminum. Most recently, the Southern District of New York also addressed the issue of uninjured class members by examining whether statistical models could accurately be used as common proof to prove classwide injury. In Aluminum, plaintiffs alleged that financial institutions and aluminum warehouse owners conspired to artificially increase the value of aluminum derivatives, causing plaintiffs to buy aluminum at inflated prices.[55]

Plaintiffs provided an expert report that used a chain of statistical models to prove that defendants engaged in anticompetitive behavior to lengthen loading queues for aluminum at warehouses, which in turn, caused a delay which increased the value of the aluminum derivatives and, ultimately, that increased benchmark price was passed on to aluminum purchasers.[56] Plaintiffs also provided anecdotal evidence in the form of statements from industry experts and employees. But the district court found that neither method could properly be used as common proof required to show predominance under Rule 23(b) and denied class certification.

In performing its analysis, the court examined the statistical models provided by plaintiffs’ expert. Under Comcast, when a statistical model is purported to prove classwide impact and causation, courts must “rigorously examine the soundness of that model at the class certification stage.”[57] The court in Aluminum found several flaws in plaintiffs’ statistical model. For example, the model “fail[ed] to isolate the effects of the conspiracy” and instead only isolated the effect of a London Metals Exchange load-out rule two years into the alleged six-year conspiracy, meaning plaintiffs “lack[ed] classwide proof that the alleged conspiracy lengthened queues throughout the relevant period.”[58] The model also impermissibly relied on averaging, as the expert “estimated an aggregated excess load-out figure for the entire time period and averaged it as a monthly number for the entire conspiracy” while simultaneously “fail[ing] to distinguish products of the alleged conspiracy from economic background events.”[59]

The district court paid particular attention to the expert’s use of averaging, noting that “courts have disdained models that have found classwide price impact by means of averaging impact across a class period,” because the method does not guarantee that each class member suffered an individual injury.[60] In Aluminum, this issue was poignant due to the “significant variation in the level of purported conspiratorial activity at any given moment during the class period.”[61] The district court ultimately found that the statistical model concealed uninjured class members by “yield[ing] false positives,” thereby failing to meet the common proof requirements for predominance.[62]

In addition, the court rejected plaintiffs’ argument that anecdotal evidence should suffice by itself as common proof. While the court acknowledged that the statements from industry experts and employees provided important context in evaluating the expert’s statistical models, documentary evidence is merely supplemental:

[T]he statements by market participants on which plaintiffs rely are anchored to particular industry circumstances and moments in time. They do not, in terms, profess to opine on the necessary proposition here: that defendants’ conspiracy, as alleged in this litigation, unitarily worked antitrust pricing injury on all entities and persons now defined to fall within the putative class.[63]

Following the precedent established in Comcast and Lamictal, the Southern District of New York held that because the expert’s statistical model of classwide injury fails due to design flaws, and because no other alternative adequately demonstrated that there could be common proof of classwide impact, class certification must be denied.[64]

Cordoba and Aluminum are additional cases, similar to the previously discussed In re Rail Freight Fuel Surcharge Antitrust Litig.,[65] that may be relevant to defendants considering arguments related to the effect uninjured class members have on certification. As Cordoba, Aluminum and Rail Freight demonstrate, courts agree that a class cannot contain too many uninjured class members, although they have taken different approaches to their analyses.

                                                                                                                       *    *    *


[1]       957 F.3d 184 (3d Cir. 2020).

[2]       Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036 (2016).

[3]       Tyson Foods, 136 S. Ct. at 1043.

[4]       Id.

[5]       Id. at 1049.

[6]       Id.

[7]       957 F.3d 184 (3d Cir. 2020).

[8]       Lamictal, 957 F.3d at 189.

[9]       Id.

[10]      Id.

[11]      Id. at 191.

[12]      Id.

[13]      Id.

[14]      Id. at 191 (citing In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 307 (3d Cir. 2009)).

[15]      Id. at 191 (quoting Tyson Foods, 136 S. Ct. at 1040).

[16]      Id.

[17]      Id. at 193–95.

[18]      Id. at 194–195.

[19]      946 F.3d 178 (3d Cir. 2019).

[20]      Id. at 181.

[21]      Id. at 181–82.

[22]      Id. at 186.

[23]      Id. at 185–86.

[24]      Ferreras, 946 F.3d at 185 (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011) and Tyson Foods, 136 S. Ct. at 1045)(emphasis added). Notably, the Third Circuit also found that the district court erred by applying an improper standard and effectively conditionally certifying a class—which is impermissible under Rule 23—under a “pleading” and “initial evidence” standard rather than by a preponderance of the evidence. Id. When rejecting the no-reasonable-juror standard from Tyson, the court in Lamictal cited to Ferreras (among other cases) for the proposition that factual determinations still must be proven by a preponderance of the evidence. See Lamictal, 957 F.3d at 192.  

[25]      Id. at 185.

[26]      Id. at 186.

[27]      Id.

[28]      Id.

[29]      Id. at 186 (citing Tyson Foods, 136 S. Ct. at 1041–42).

[30]      Id. at 187.

[31]      946 F.3d 1066 (9th Cir. 2020).

[32]      Id. at 1072.

[33]      Id. at 1075.

[34]      Id. at 1083.

[35]      Id. at 1086.

[36]      Id.

[37]      Id.

[38]      Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036 (2016).

[39]      Id. at 1088.

[40]      Id.

[41]      Id. at 1087–88.

[42]      Ridgeway, 946 F.3d at 1087.

[43]      Id. at 1088.

[44]      Id.

[45]      942 F.3d 1259 (11th Cir. 2019).

[46]      No. 14-cv-3116 (PAE), 2020 WL 4218329 (S.D.N.Y. July 23, 2020).

[47]      Id. at 1264.

[48]      Id. at 1272.

[49]      Id. at 1273–74.

[50]      Id. at 1273 (citations omitted).  

[51]      Id. at 1274.

[52]      Id. at 1275.

[53]      Id. at 1277.

[54]      Id.

[55]      Aluminum, 2020 WL 4218329.

[56]      Id. at *16.

[57]      Id. at *38.

[58]      Id. at *45.

[59]      Id. at *48, *54.

[60]      Id. at *48.

[61]      Id.

[62]      Id. at *54.

[63]      Id. at *41.

[64]      Id. at *58.

[65]      Dakota v. BNSF Ry. Co. (In re Rail Freight Fuel Surcharge Antitrust Litig.), 934 F.3d 619 (D.C. Cir. 2019). For our prior summary and analysis on that case, see D.C. Circuit Upholds Denial of Class Certification When Economic Model Showed Uninjured Members, Aug. 23, 2019, See also Class Certification Developments, Sept. 27, 2019,

© 2024 Paul, Weiss, Rifkind, Wharton & Garrison LLP

Privacy Policy