By Andrew J. Trask, Of Counsel with Shook, Hardy & Bacon L.L.P. in its San Francisco, CA office.

Following on the heels of the U.S. Court of Appeals for the First Circuit’s groundbreaking opinion in In re Asacol Antitrust Litigation, 907 F.3d 42 (1st Cir. 2018), the District of Columbia Circuit has also held that classes that contain definite numbers of non-injured members cannot be certified, because individual questions related to injury will predominate over common issues.

In re Rail Freight Fuel Surcharge Antitrust Litigation, 934 F.3d 619 (D.C. Cir. 2019), was part of a multi-district litigation in which the plaintiffs alleged that alleged that railway companies had engaged in a price-fixing conspiracy.  The alleged conspiracy concerned “fuel surcharges,” imposed above the base shipping price when the cost of fuel rises sufficiently high.

The putative class contained roughly 16,000 shippers.  At certification (in the U.S. District Court for the District of Columbia), the plaintiffs had sought to show common issues would predominate over individualized issues by hiring an economics expert to perform a regression analysis.  Id. at 621.  Regression analysis—for those of us who went to law school to avoid taking statistics—takes different variables (like supply, demand, weather) that contribute to a result (like a price), and measures the effect of each variable.  It is a common way to show relationships between causes and effects.  In this case, the plaintiff’s regression model, which sought to measure the effect of the alleged conspiracy, resulted in “negative damages” for 2,000 members (an eighth, or about 12.5%) of the proposed class.  “Negative damages” as a result meant that a large part of the class either were simply not damaged at all, or somehow made money on the challenged transactions.

The plaintiffs’ regression model had a checkered history: the plaintiffs had been using this model to seek certification of this class since 2012.  Originally, the defendants had challenged certification because the model had serious flaws, but the district court had certified a class anyway, finding the model was both “plausible” and “workable.” In re Rail Freight Fuel Surcharge Antitrust Litig., 287 F.R.D. 1, 43 (D.D.C. 2012).  The D.C. Circuit had vacated that certification on interlocutory review, because the model’s “propensity toward false positives” made it difficult to tell whether individual class members had actually been harmed.  In re Rail Freight Fuel Surcharge Antitrust Litig., 725 F.3d 244. 254 (D.C. Cir. 2013). Common questions, it reasoned, “cannot predominate where there exists no reliable means of proving classwide injury in fact.” Id. at 253.  As a result, Rule 23 requires a “hard look at the soundness of statistical models that purport to show predominance.” Id. at 255.

On remand from that first appeal, the district court ordered supplemental discovery.  In re Rail Freight Fuel Surcharge Antitrust Litigation,  934 F.3d at 621.  Then it denied certification, finding three flaws with the model: (1) it measured “highly inflated damages” for traffic with additional modes of travel (e.g., a train and a ship); (2) it erroneously measured damages for shipments under legacy contracts (a flaw the defendants had complained about since the first certification battle); and finally (3) it had measured “negative damages” for 2,000 class members.  In re Rail Freight Fuel Surcharge Antitrust Litig., 292 F. Supp. 3d 1,4, 122-41 (D.D.C. 2017).

This time, the plaintiffs filed interlocutory appeal under Rule 23(f).  But the existence of these uninjured class members placed the plaintiffs in a strategic bind.  In briefing the issue, they launched two lines of argument.  First, they attacked the credibility of their own damages model, arguing that the negative damages were a result of “normal prediction error.”  In re Rail Freight Fuel Surcharge Antitrust Litigation,  934 F.3d at 624.  They also argued that “predominance does not require common evidence extending to all class members.”  Id.

The D.C. Circuit took the case again.  Quoting Tyson Foods, Inc. v. Bouaphakeo, the panel identified an “individual question” as “one for which ‘members of a proposed class will need to present evidence that varies from member to member.’”  Id. at 622, quoting 136 S. Ct. 1036, 1045 (2016).  For the sake of argument, the panel assumed the damages model was sufficiently reliable—even though that was one of the most hotly contested issues at certification.  Id. at 623.  Then it noted that, even assuming the model was reliable, “that still leaves the plaintiffs with no common proof of those essential elements of liability for the remaining 12.7 percent” of the proposed class.  Id. at 624.

The panel did not find either of plaintiffs’ arguments persuasive.  It noted that the district court did not believe the model was flawed, and if it had, that would make the model too inaccurate to calculate classwide damages.  Id. 

The panel also rejected the plaintiffs’ argument that predominance does not need to reach all members of the proposed class as inconsistent with its prior ruling.  It held that “[u]ninjured class members cannot prevail on the merits, so their claims must be winnowed away as part of the liability determination.  And that prospect raises the —when does the need for individualized proof of injury and causation destroy predominance?”  Id.  It noted the district court’s observation that the “few reported opinions” tackling this question “suggest that 5% to 6% constitutes the outer limits of a de minimis number” of uninjured class members.  Id. at 625, quoting Rail Freight II, 292 F. Supp. 3d at 137.  That was a problem for certification, because “the 12.7 percent figure in this case is more than twice that approximate upper bound.”  Id.

Equally important, the plaintiffs had shown no way to separate that 12.7 percent from the 87.3 percent of allegedly injured class members.  “The absence of any winnowing mechanism sharply distinguishes Nexium, the plaintiffs’ best case.”  Id.  (In re Nexium Antitrust Litigation, 777 F. 3d 9 (1st Cir. 2015), was the case that the First Circuit distinguished when it held that a class with 10 percent uninjured members could not establish a predominance of common issues. See In re Asacol.)  Since the regression analysis establishing damages was “essential” to the plaintiffs’ case, this meant they were out of luck.  The panel affirmed the denial of certification.

This case, combined with the First Circuit’s In re Asacol, shows a burgeoning trend against significant no-injury cases.  Two circuit courts of appeal have now held that proposed classes with significant percentages of non-injured members cannot be certified under Rule 23(b)(3), because too many individual inquiries are required to separate the actually injured from the non-injured.  There is a countervailing trend, however.  Jurisdictions like the Ninth Circuit still allow no-injury cases by invoking different logic, pushing off the question of whether anyone has been injured as part of a “merits” determination they refuse to engage in.  Nguyen v. Nissan North America, Inc., 932 F. 3d 811 (9th Cir. 2019).  However, the clear logic of In re Asacol and now In re Rail Freight Surcharge provides practitioners with powerful tools to fight no-injury cases at the certification stage.  And, given the Supreme Court’s clear interest in the intersection between class certification and merits inquiries, it is likely we could see this issue resolved once and for all in the near future.