Can you have a class action if class members can’t reliably be found? That question is at the heart of the debate over ascertainability—one that has divided the federal courts. Earlier this week, the Ninth Circuit weighed in, holding in Briseno v. ConAgra Foods, Inc. (pdf) that plaintiffs need not demonstrate “an administratively feasible way to identify class members [as] a prerequisite to class certification.”

That conclusion is disappointing.

In our view, due process and the Rules Enabling Act require taking ascertainability seriously. That means that, first, plaintiffs must demonstrate that there is an administratively feasible way to tell who is in the class, without plaintiff-by-plaintiff fact finding and an unmanageable series of mini-trials; and second, the use of barebones affidavits by would-be class members attesting that they fall within the class definition are inadequate, because due process entitles defendants to challenge an individual’s claim of membership in the class.

The leading case adopting these principles is the Third Circuit’s opinion in Carrera v. Bayer Corp., 727 F.3d 300 (3d Cir. 2013), which we have previously discussed extensively. Several other circuits have followed Carrera’s lead and likewise required plaintiffs to demonstrate a feasible way to tell who is in the class—including the Second Circuit in Brecher v. Republic of Argentina, 806 F.3d 22 (2d Cir. 2015), the Fourth Circuit in EQT Production Co v. Adair, 764 F.3d 347 (4th Cir. 2014) (our colleague reported on EQT here), and the Eleventh Circuit (albeit in an unpublished decision) in Karhu v. Vital Pharmaceuticals, Inc., 621 F. App’x 945 (11th Cir. 2015).

Other courts of appeals, however, have refused to follow Carrera and rejected the “requirement that plaintiffs prove at the certification stage that there is a ‘reliable and administratively feasible’ way to identify” class members. Mullins v. Direct Digital, LLC, 795 F.3d 654, 657 (7th Cir. 2015); see also Rikos v. Procter & Gamble Co., 799 F.3d 497 (6th Cir. 2015) (citing Mullins with approval and “see[ing] no reason to follow Carrera”); Sandusky Wellness Ctr., LLC v. Medtox Sci., Inc., 821 F.3d 992 (8th Cir. 2016) (declining to clearly “outline[] a requirement of ascertainability” but citing Mullins with apparent approval).

The Ninth Circuit’s decision in Briseno places that court firmly in the latter camp, deepening the circuit split on this critical issue. (We filed an amicus brief on behalf of the Chamber of Commerce of the United States in a separate Ninth Circuit case raising the same issue.) The panel questioned whether Rule 23 contains an independent “ascertainability” requirement at all, but held that, in any event, such a requirement does not compel plaintiffs to propose a feasible method for identifying “who is in the class.”

In its opinion, the Ninth Circuit did not deny that it would be difficult—if not impossible—to identify potential class members in an objective way: the proposed class encompasses residents of eleven states who purchased Wesson-brand cooking oil products at certain times. Identification of these individuals cannot realistically take place through documentation, as consumers typically do not keep receipts or packaging from food products that likely were purchased and consumed years ago; and the makers of Wesson oil do not sell directly to consumers, so have no records of the purchasers of their products. The Ninth Circuit panel recognized as much, noting that it “would be difficult to demonstrate” an administratively feasible way of identifying class members and that strictly adhering to the ascertainability requirement would be “outcome determinative”—in other words, it would require reversal of the district court’s certification order. But the court then used that observation as a reason not to impose a strict ascertainability requirement, lamenting that “[c]lass actions involving inexpensive consumer goods in particular would likely fail at the outset if administrative feasibility were a freestanding prerequisite to certification.”

That approach to class certification treats consumer class actions as the norm, thus (at least in our view) turning on its head the Supreme Court’s repeated admonition that class actions are “an exception to the usual rule” that claims must be litigated individually. Califano v. Yamasaki, 442 U.S. 682, 700-01 (1979).

Moreover, there are several fundamental flaws in the Ninth Circuit’s analysis.

To begin with, the Ninth Circuit concluded that because ascertainability is not explicitly mentioned in Rule 23, any concerns about identifying class members in an administratively feasible way should be addressed entirely under the rubric of Rule 23(b)(3)’s superiority requirement—in particular, its mandate that courts consider “the likely difficulties in managing a class action.” The court made clear that this shift from ascertainability to manageability will blunt the force of arguments that class members cannot be feasibly identified, because the court went on to say that “courts should not refuse to certify a class merely on the basis of manageability concerns.” That is a bold proposition: outside of the settlement context, the fundamental question that Rule 23 asks is whether a class-wide trial is appropriate, and it seems self-evident that if a class trial is not manageable, then a class should not be certified.

The panel’s opinion instead appears to replace the relatively concrete questions of manageability and ascertainability with a vague standard authorizing courts to “balance” the difficulty (or impossibility) of identifying class members with their views of “the benefits of class adjudication.”

This nebulous balancing approach gives short shrift to the due process underpinnings of the ascertainability requirement. Because the Rules Enabling Act prohibits the use of the Federal Rules to “enlarge, abridge, or modify any substantive right,” a Rule 23 class action is not—and cannot be—anything more than the sum of the individual class members’ claims. And because in an individual case a plaintiff would have to prove that he purchased the challenged product and the defendant would have the right to challenge the plaintiff’s evidentiary showing, cross-examine the plaintiff, and have a court or jury resolve any factual disputes, the defendant must maintain the same due process rights in a Rule 23 class action. Yet if class members cannot be identified, defendants will have no way to bring such challenges.

The Ninth Circuit brushed aside these due process concerns by concluding that defendants can bring individualized challenges to would-be class members’ right to recovery as part of a claims administration process, after a class has been certified and liability determined.

But this certify-first, ascertain-later approach is deeply flawed. While parties may agree in settlement class actions to have a claims administrator assess claims forms (that are sometimes the equivalent of affidavits) as to whether an individual qualifies for relief, that agreement reflects one of the compromises of settling a case. In a litigated case, however, the administrative determinations of an outside third party in a high-stakes class action do not satisfy a defendant’s due process right to cross-examine its opponents and for judicial resolution of factual disputes.

The Ninth Circuit panel also problematically endorsed the plaintiffs’ proposal to avoid identifying actual people who are class members by calculating the defendant’s total liability on an aggregate basis based on total sales (without knowing to whom the product was sold)—an approach sometimes referred to as “fluid recovery.” Yet numerous courts have rejected the use of such an approach in a litigated class action (as opposed to a settlement), including the Ninth Circuit over four decades ago in In re Hotel Tel. Charges, 500 F.2d 58 (9th Cir. 1974). Moreover, the Ninth Circuit case cited by the panel, Six Mexican Workers v. Arizona Citrus Growers, 904 F.2d 1301 (9th Cir. 1990), held only that cy pres recovery might be used for the distribution of unclaimed damages when plaintiffs bring claims for statutory damages that “are not dependent on proof of actual injury,” and thus there is less concern about “the impermissible circumvention of individual proof requirements.” The Unfair Competition Law claims at issue in Briseno, however, do require proof of actual injury. (As we have previously discussed in connection with Spokeo, Inc. v. Robins, there are serious questions about whether Article III permits federal courts to certify classes that include persons who have not suffered concrete harm.)

Finally, given that the Ninth Circuit oversees the so-called “Food Court,” its rejection of a meaningful ascertainability requirement as a prerequisite to class certification will likely encourage the filing of food-labeling class actions that have been the subject of substantial criticism. Yet the evidence suggests that the principal beneficiaries of such class actions are the lawyers (on both sides, to be sure) rather than the class members, most of whom cannot be identified and thus do not actually receive any benefit.

* * *

In short, the courts of appeals are now deeply divided on whether there must be an administratively feasible way to identify actual class members before a class may be certified. In an appropriate case, the issue is ripe for Supreme Court review.