The Yale Law Journal

May 2015

Class Ascertainability

Geoffrey C. Shaw

abstract. In recent years, federal courts have been enforcing an “implicit” requirement for class certification, in addition to the explicit requirements established in Rule 23 of the Federal Rules of Civil Procedure. The ascertainability requirement insists that a proposed class be defined in “objective” terms and that an “administratively feasible” method exist for identifying individual class members and ascertaining their class membership. This requirement has generated considerable controversy and prevented the certification of many proposed classes. The requirement has taken a particular toll on consumer class actions, where potential class members are often unknown to the representative plaintiffs, often lack documentary proof of their injury, and often do not even know they have a legal claim at all.

This Note explores the ascertainability requirement’s conceptual foundations. The Note first evaluates the affirmative case for the requirement and finds it unpersuasive. At most, Rule 23 implicitly requires something much more modest: that classes enjoy what I call a minimally clear definition. The Note then argues that the ascertainability requirement frustrates the purposes of Rule 23 by pushing out of court the kind of cases Rule 23 was designed to bring into court. Finally, the Note proposes that courts abandon the ascertainability requirement and simply perform a rigorous analysis of Rule 23’s explicit requirements. This unremarkable approach to class certification better reflects what the Rule says and better advances what the Rule is for.

author. Yale Law School, J.D. expected 2016. I am overwhelmingly grateful for the help I have received from friends, colleagues, and mentors in writing this Note. I thank Kristin Collins, for expanding and clarifying my thinking on civil procedure in ways ascertainable and not, and for providing extensive feedback on several drafts. I am grateful to the participants in the Advanced Civil Procedure Seminar in Spring 2014, who workshopped a draft of the Note and generously offered helpful comments, and to all my colleagues in Judith Resnik’s Procedure small group in Fall 2013. I thank Mark Jia for reading a draft, Robby Nightingale, Meng Jia Yang, and Rachel Bayefsky for fantastic feedback at numerous stages of the project, and all the editors of the Yale Law Journal for their meticulous editing. Finally, I am deeply indebted to Judith Resnik for her assistance and inspiration.


“Modern society,” wrote Harry Kalven, Jr. and Maurice Rosenfield, two legal visionaries who conceptualized the class suit, “seems increasingly to expose men to . . . group injuries for which individually they are in a poor position to seek legal redress.”1 Rule 23 of the Federal Rules of Civil Procedure responds to this problem. The Rule, which took its modern form in 1966, creates a class action mechanism to aggregate the claims of people who “are isolated, scattered, and utter strangers to each other.”2 In doing so, the Rule aims to bring about regulatory effects far beyond what is possible with individual litigation alone and to break from the old formalisms that kept claims out of court. In the words of its principal drafter, Benjamin Kaplan, the Rule “intended to shake the law of class actions free of abstract categories contrived from . . . bloodless words . . . and to rebuild the law on functional lines responsive to . . . recurrent life patterns which call for mass litigation through representative parties.”3 Today, “modern society” still “expose[s]” men and women to “group injuries for which. . . they are in a poor position to seek legal redress” as individuals. But a “judicially created”4 change to the law of class certification, untethered to the carefully engineered text of Rule 23, has disrupted class action procedure and made it harder for them to “seek legal redress” as groups. This development deserves a critical and “rigorous analysis.”5

Rule 23 establishes specific criteria for class certification.6 The proposed class must be so numerous that joinder of each individual plaintiff is “impracticable”;7 the members of the class must have common claims;8 the claims of the representative plaintiffs must be typical of the class;9 and the representative plaintiffs must be able “adequately” to represent the absent class members.10 If these conditions are satisfied, the proposed class must also fit into one of three functional categories.11A party seeking class certification must affirmatively demonstrate his compliance with the Rule—that is, he must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc.12 And a court must perform a rigorous analysis to ensure that the proposed class meets the Rules requirements.13

Recently, however, a growing number of federal courts have identified an additional, implicit requirement for class certification: the class must be ascertainable.14 Although this “implied requirement of ascertainability” does not appear in the text of Rule 23 and “is judicially created,”15 courts deploy it as an independent bar to class certification.16 The general idea is that there ought to be an objective and administratively feasible way to determine exactly who is in the class. As the Court of Appeals for the Third Circuit put it, “[A]scertainability entails two important elements. First, the class must be defined with reference to objective criteria. Second, there must be a reliable and administratively feasible mechanism for determining whether putative class members fall within the class definition.”17 Courts disagree, however, about exactly what the requirement means and how it should be applied. Some courts have placed greater emphasis on the objectivity of the class’s definition, which is said to protect against excessive administrative burdens over the course of the litigation.18 Other courts have directly scrutinized the administrative feasibility of identifying individual members, requiring plaintiffs to propose and defend methods for identifying the class’s membership.19

A controversial20 case, Carrera v. Bayer Corp., offers an illuminating example.21 Gabriel Carrera sued a large pharmaceutical company, Bayer, on behalf of customers who had purchased an over-the-counter weight loss pill. Not surprisingly, most of these purchasers had not lost weight, and Carrera and the class members sought to recover small dollar amounts in compensation for misleading advertising. As with many consumer class actions, each class member’s monetary claim was so low that the case would likely never have been brought except as a class action.22 Bayer argued that it was too difficult to figure out who was a member of the class and who was not because neither Bayer nor the plaintiffs had any documentary records to prove class membership.23 Bayer had no records because it had simply sold the pills to intermediary retailers, and those intermediary stores kept no records of who bought what, only statistics about what was bought and revenue trends. Individual purchasers might have been given receipts, but most of them were lost, and even so, many receipts did not affirmatively identify purchasers. The Court of Appeals for the Third Circuit sided with Bayer and decertified the class. The class failed the ascertainability test: determining who was a member of the class and who was not could require “mini-trials” and would rely too heavily on the subjective “say-so” of potential members.24 As one writer put it, “class dismissed.”25

Ascertainability doctrine is “one of the most contentious issues in class action litigation these days.”26 On the one hand, it seems sensible for classes to be defined in a clear way that permits the court to identify the class members. Why should the court or the defendant not be able to ask “who is in the class?” and receive a definite answer? How can a court provide notice to class members if it cannot ascertain their identities? Who would share in a damage award if the claim succeeds? Who would be bound by the outcome of a case? On the other hand, what is the fate of small-dollar consumer class actions in a world with ascertainability tests? In this kind of litigation, specific evidence of individual class membership is often hard to obtain, and the value of individual claims is frequently too low to incentivize individual suits. If the criteria for class certification become harder to satisfy, will laws protecting consumers against fraud, deception, dangerous products, false advertising, breach of contract, and many other harms be sufficiently enforced? How will wrongdoers be deterred? Further, are judges supposed to create new implicit requirements to supplement the Federal Rules of Civil Procedure? And what does the ascertainability requirement actually require?

I write against the backdrop of a debate among lawyers and judges about what the ascertainability requirement means–a debate that has generated a lot of heat, but not much light. The accelerating application of the requirement in court has been haphazard at best. Efforts to enforce the requirement in different contexts have produced differences among circuits,27 splits within individual circuits,28 and even differences of opinion among different judges on the same district court.29 A few judges have voiced their concern about the requirement’s evolution. One district court judge, for example, believes that “[i]f class actions could be defeated because membership was difficult to ascertain at the class certification stage, there would be no such thing as a consumer class action.30The Carrera decision was particularly controversial, even among judges on the Court of Appeals for the Third Circuit.31 When plaintiffs petitioned for en banc review, the full Court declined rehearing,32 but not without the dissent of four judges. Judge Thomas Ambro wrote that even though he believes “an essential prerequisite of a class action . . . is that the class must be currently and readily ascertainable . . . . [o]ur Courts opinion in Carrera gives the impression to many that we now carry that requirement too far.33 Judge Ambros dissent highlighted the importance of the issue:

Even if, as I believe, the ability to identify class members is a set piece for Rule 23 to work, how far we go in requiring plaintiffs to prove that ability at the outset is exceptionally important and requires a delicate balancing of interests. It merits not only en banc review by our Court but also review by the Judicial Conference’s Committee on Rules of Practice and Procedure.34

The ascertainability requirement also merits academic criticism. Although the requirement has generated an intensifying flurry of professional commentary and debate within the judiciary,35 it has not been the subject of sustained academic discussion36—which is unfortunate, because the lack of scholarly attention deprives the courts of important critical and theoretical perspectives. Accordingly, in this Note I analyze and critique the foundations of ascertainability doctrine.

This Note presents a fundamentally skeptical view of the ascertainability requirement and proposes that we rethink whether it is really “implicit,”37 really justified, or really necessary. In Parts I and II, I analyze the affirmative case for the ascertainability requirement. In Part I, I examine the most common defenses of the requirement and conclude that they are unpersuasive. Courts argue that the requirement is (1) necessary to enable the delivery of notice at the beginning of a case and to enable potential class members to opt out of an action, (2) necessary to facilitate the disbursement of damages at the end of a case, and (3) necessary to clarify who is bound by a judgment after a case is over. While these considerations speak to persistent difficulties in aggregate litigation, they do not support the conclusion that ascertainability is an implicit requirement in Rule 23. In fact, the Rule as written specifically endorses unascertainable classes.

In Part II, I consider a different angle of argument. Courts and commentators have argued that the dangers of subjective and vague classes and classes with problematic scope—specifically, overbroad classes in which many members have no genuine claim and failsafe classes in which the class is defined in terms of the merits of the claim being pled—warrant an ascertainability requirement. But these class malfunctions are not best understood as failures of ascertainability, and Rule 23 already supplies the resources we need to prevent them.38

At most, the considerations at stake in Parts I and II imply a much narrower requirement: the need for a minimally clear definition of the class. A minimally clear definition is indeed a minimal demand, speaking only to the conceptual discreteness of the category that the complaint seeks to identify. It does not speak to the practical difficulties involved in identifying class members or to the distinction between “objectively” and “subjectively” defined classes, the two concerns most often raised by courts when imposing an ascertainability requirement. A minimally clear definition distinguishes a class of “young people” from a class of people under eighteen, but it does not distinguish a class of people under eighteen who show ID from a class of people under eighteen who simply “say so.” This modest idea that a class definition must be intelligible and clear is nothing new,39 and insisting on such a definition need not serve as an independent bar to certification except in unusual cases. It would not, for instance, block a class of Snapple purchasers,40 Marlboro smokers,41 or users of a weight loss supplement42—three examples of classes dismissed for lack of ascertainability.43

In Part III, I argue that the application of the ascertainability requirement frustrates the distinctive purposes of Rule 23. The requirement clashes with the fundamental point of class actions—that class members may be “isolated, scattered, and utter strangers to each other”44; that they may be unknown to the representative plaintiffs, the defendant, and the court; that they may not even know that they are injured; and that they may lack the power to “advanc[e] en masse on the courts”45 to secure their rights of their own individual initiative. The ascertainability requirement has the potential to convert the class action device into something like a procedure for mass joinder. Yet unlike the joinder rule—Rule 19—which enables present, known individuals to unite their claims,46 Rule 23 envisions a different aggregation mechanism that brings together absent, unknown parties, and makes possible suits to punish and deter illegal action that individual litigation or joinder, however many times iterated, simply cannot reach. In light of these considerations, I argue that another justification for the ascertainability requirement—that it protects a right of class action defendants to challenge each individual claim against them47—fundamentally misses the mark.

I argue in Part IV that courts should dispense with the ascertainability requirement and simply perform a “rigorous analysis” of Rule 23’s explicit requirements. As written, Rule 23 already safeguards the interests that the ascertainability requirement supposedly protects and adequately guards against the problems that the requirement supposedly forestalls. Yet a rigorous analysis of the written Rule invites courts to consider the burdens of and alternatives to class aggregation with clearer eyes and enables courts to make certification decisions in a way that reflects the inclusive ambitions of Rule 23–without making consumer class actions “almost impossible to bring.48

My argument is not that it is a bad thing for classes to be ascertainable, but simply that there ought not to be an implicit ascertainability requirement for class certification. It takes convincing affirmative reasons to support an implicit requirement, and in this debate, the ascertainablity requirement needs to earn its own keep. If the conceptual justification for the requirement is less than persuasive, if the explicit parts of the Rule can already limit problematic forms of class litigation, and if the requirement clashes with the fundamental mission of the Rule that supposedly implies it, then it is time to rethink whether the “implicit requirement” is really a requirement at all.

I. the life of the case: notice, remedies, res judicata

In this Part, I consider the most common arguments for the ascertainability requirement: that class ascertainability is“essential . . . for a court to decide and declare who will receive notice, who will share in any recovery, and who will be bound by the judgment.49 Let us start with the beginning of a case—the delivery of notice – then move to the end—the disbursement of damages—and finally consider the cases preclusive effects in the future.

A. Notice

At the beginning of a Rule 23(b)(3) class action, the court must deliver “the best notice practicable under the circumstances” to the absent class members,50 which will in turn give them the opportunity to opt out of the litigation. At first glance, this responsibility seems to provide support for an ascertainability requirement. If the identities of class members cannot be ascertained, how can they be notified? The ascertainability requirement, wrote the Court of Appeals for the Third Circuit in Marcus v. BMW of North America, LLC “protects absent class members by facilitating the ‘best notice practicable’ under Rule 23(c)(2).”51 And in Carrera, the court argued that “ascertainability . . . allow[s] potential class members to identify themselves for purposes of opting out of a class.”52

The text of Rule 23’s notice requirement, however, specifically envisions unascertainable classes. Rule 23(c)(2)(B)’s notice provision requires the “best notice practicable under the circumstances, including individual notice to all members who can be identified through reasonable effort.”53 Which is to say, the Rule specifically envisions that some class members might not be able to be identified through reasonable effort. This language is in tension with the ascertainability requirement; the Rule assumes that the class might not be ascertainable.54

The drafters of Rule 23 openly debated whether requiring the “best notice practicable under the circumstances, including individual notice to all members who can be identified through reasonable effort,”55 as the Rule now reads, was sufficient, or whether the Rule ought to include a more demanding notice provision.56 They selected a standard that would further the basic goal the drafters wished to achieve–to include people who could not necessarily “be identified through reasonable effort” in litigation–without frustrating the Constitution’s due process requirements. A decade and a half before, in Mullane v. Central Hanover Bank & Trust Co., the Supreme Court considered what kind of notice was due to unknown beneficiaries of a trust when the trust was settled.57 The case raised a recurring issue: it is unfair to subject people to a binding judicial proceeding in which their rights or interests are at stake without telling them about it first, but how far must a court go in this endeavor when it does not know exactly whose interests and rights are at stake? The Court struck a balance. It required direct notice by mail to known beneficiaries but permitted newspaper advertisements to reach unknown beneficiaries. Mullane called for notice “reasonably calculated” to inform the relevant parties under the circumstances.58 It did not require direct notice to each individual potentially affected by a judicial proceeding. Neither does Rule 23 require direct notice to each individual class member.

Adequate notice in a class action, therefore, has never required ascertainability. Class members can be notified adequately whether or not the class is ascertainable. For the same reason, a class does not need to be ascertainable in order to give absent class members sufficient opportunity to opt out.59 The adequacy of one’s opportunity to opt out depends on the adequacy of notice,60 and if notice can be adequate independent of ascertainability, so too can the opportunity for opt-out.61

B. Remedies

At the end of litigation, in the event of settlement or judgment, a court must consider remedies. A second common argument suggests that ascertainability is necessary in order to facilitate the disbursement of damages, which would be needlessly burdensome if the class were not ascertainable. How can a court ensure that payment is made to someone whose identity it does not know? Ascertainablity, one court wrote, is “essential . . . for a court to decide and declare . . . who will share in any recovery.62As Judge Ambro wrote in his opinion in Marcus v. BMW of North America, LLC, the requirement “eliminates serious administrative burdens that are incongruous with the efficiencies expected in a class action”63–namely, the crucial task of paying the plaintiffs in the event that they prevail.64 Carrera v. Bayer Corp. was preoccupied with this issue too: for the purposes of remedies, among other things, “a trial court should ensure that class members can be identified without extensive and individualized fact-finding or mini-trials, a determination which must be made at the class certification stage.65

But what, precisely, are the anticipated administrative burdens? When, exactly, will the court have to perform a “mini-trial”? In the event of final judgment, the Rule as written specifically permits the individualized inquiry that the requirement supposedly forestalls. Rule 23 envisions a conceptual division between the determination of liability and the assignment of damages. In particular, a defendant’s total liability can be “determined at the aggregate level . . . with individual awards worked out in subsequent proceedings.”66 After determining the “defendant’s monetary liability to the class67 as a whole, the court chooses how and to whom to disburse the money. At this stage, the Rule allows individualized inquiry to determine which claimants get what. Judge Posner recently explained the way the Rule works clearly. Rule 23 permits “[a] class action limited to determining liability on a class-wide basis, with separate hearings to determine —if liability is established—the damages of individual class members.”68 In fact, Judge Posner wrote, this “will often be the sensible way to proceed.”69 In those separate hearings, which can be “elaborate,”70 potential claimants come forward and present their own evidence for class membership. This scheme derives from Rule 23(c)(4), which enables courts to certify classes for “particular issues.”71 The Advisory Committee noted that “in a fraud or similar case the action may retain its ‘class’ character only through the adjudication of liability to the class; the members of the class may thereafter be required to come in individually and prove the amounts of their respective claims.”72 Kaplan’s account of the drafting of the Rule assumes that of courseafter determination of common questions the individual claimants might have to come in and prove their damages . . . . [S]uch follow-up proceedings will occur in various (b)(3) actions.73In these “follow-up proceedings,” the court would assess the would-be claimant’s individual evidence for membership in the class.

In the event of settlementthe much more common outcomesomething similar takes place. According to Janet Alexander, “Most settlements simply specify the total class recovery,”74 an amount determined in the aggregate. The settlement agreement then establishes a negotiated method for delivering the remedy to the class members:

The settlement agreement sets out the recovery that will go to the class, the method of allocating the recovery among class members, the requirements for qualifying to receive a share of the recovery, the content of the class notice and the means of giving notice, and often an agreement that the defendants will not object to a fee request by class counsel of up to an agreed amount. After the settlement is approved, someone is normally appointed to administer the claims process. This task would include processing the class notice, sending claim forms, receiving, verifying, and processing claims, calculating the amount of the recovery for each claimant, preparing and sending checks, and taking care of numerous administrative details.75

Here, the claims process is likely to be chaotic, imperfect, and burdensome, calling for extensive individualized determination and “elaborate” administration. Nobody expects the claims process to be perfect: of course it is possible that there will be errors; of course it is common for only a fraction of the class members to ever file claims.76

But ascertainability need not be a prerequisite for certification for the damages process to work, whether damages flow from final judgment or from settlement. After class-wide liability is determined, individual claimants must come forward and identify themselves, making their own case for membership in the class regardless of whether the court knew, or could have known, their identity or their membership in advance. In the notice context, it is the court’s responsibility to reach out to potential class members, but only in a general, imperfect way. No ascertainability needed: the court does not need to know who exactly is in the class, because it only needs to reach out to the class generally. When it comes to remedies, it is time for the potential class members to reach out to the court, and here they make their own individual case for membership and eligibility for payment. Here again, the court does not need to know who exactly is in the class, because potential class members bear the responsibility of coming forward to offer their own individual evidence for class membership.

Whether it will be too difficult for the court to determine whether someone is a member of the class after he has made his case for membership is a separate and important questionbut this question also fails to provide support for the ascertainability requirement. The explicit text of Rule 23 invites the court to consider practical difficulty. If the court decides that the individualized inquiry after a finding of liability will be too difficult, then it may deny certification on “manageability” grounds: Rule 23 permits the court to consider the manageability of the litigation at the time of certification.77 Or the court may decide that so much effort would have to go into the individualized inquiry that the class action mechanism is no longer “superior” to individualized litigation78—that is, aggregation in that case would not generate the necessary economy of scale. Carrera v. Bayer Corp. claimed that“[i]f a class cannot be ascertained in an economical and administratively feasible manner[,] significant benefits of a class action are lost.79 This may very well be true in some situations—but why not simply say that as a result, the class action mechanism is no longer superior to other methods of resolution?80Finally, perhaps most obviously, if too much individual inquiry is required, a court could find that class issues do not predominate over individual issues.81 The framers of Rule 23 expected the manageability, superiority, and predominance provisions to help courts marshal their resources in a way that responds to the actual demands of litigation. The question for courts, Kaplan wrote, is not whether “after determination of common questions the individual claimants might have to come in and prove their damages, for such follow-up proceedings will occur in various (b)(3) actions;” the question is whether “the realities of litigation . . . suggest that the class procedure is not ‘superior’ to more commonplace devices” and whether “individual questions of liability and defense will overwhelm the common questions.”82

Still, the question of difficulty does not end the analysis, because difficulty depends upon the choice of remedy–that is, upon what exactly the court is trying to do. On this score, it is important to remember that some remedies–whether flowing from settlement or from final judgment–never involve identifying individual class members at all. Alexander notes, for example, that “non-monetary” remedies are “common, especially in consumer class actions.”83 For example, “[d]efendants may agree to cease or modify certain business practices, or to adopt safeguards for consumers.”84 Alternatively, sometimes the court orders a monetary payout to a non-party, such as a charity whose mission is related to the harm addressed by the suit. These cy pres remedies can achieve some of the same policy objectives as other remedial schemes.85 They are used sometimes because the potential claimants cannot be identified individually,86 and sometimes because “the class damages cannot be distributed to the class at all.87 This was the case with an $8.5 million settlement for a class of 37 million Gmail users contesting privacy issues raised by Googles 2010 rollout of its now-defunct Buzz social networking program.88 Another option, so-called fluid recovery, is used in cases where the harmed class members are so difficult to locate with specificity that it is economically more feasible to make the class action damages available to a substitute group of similarly situated persons.89The point is not that non-monetary, cy pres, or fluid recovery remedies are desirable in every case, but simply that remedies exist that do not require courts to identify individual class members.90

In principle, therefore, a class need not be ascertainable to deliver remedies. Liability is determined in the aggregate and claimants must identify themselves to receive a payout. The claims process is inevitably complicated and individualized, but Rule 23 as written permits courts to evaluate their ability to manage that process. And class action law allows for alternative remedies that can be appropriate even if paying damages to the class would indeed be overly burdensome.

C. Res Judicata

Another argument we must consider contends that the ascertainability requirement is necessary to clarify the preclusive effects of class action litigation. If the class is not ascertainable, how can it be clear who is bound by the case’s outcome? Judge William Alsup, for instance, explains that “[a]scertainability is needed for properly enforcing the preclusive effect of final judgment. The class definition must be clear in its applicability so that it will be clear later on whose rights are merged into the judgment, that is, who gets the benefit of any relief and who gets the burden of any loss.”91 In Judge Ambro’s phrasing, the ascertainability requirement “protects defendants by ensuring that those persons who will be bound by the final judgment are clearly identifiable.”92 According to a district judge in Illinois, ascertainability “is necessary” to “ensure the binding effect of judgment on class members.”93 Judge Pauley of the United States District Court for the Southern District of New York wrote that “class membership must be readily identifiable such that a court can determine who is in the class and bound by its ruling without engaging in numerous fact-intensive inquiries.”94 And according to a district judge in California, “[r]equiring an objectively ascertainable class is important because without one, it will be unclear who is bound by the judgment.”95

Yet consider how preclusion works in practice. Res judicata is a defense to liability. Claims that a plaintiff is barred from suit arise after a new complaint is filed. In fact, the notes to the 1966 adoption of Rule 23 presented it as obvious that “the court conducting the [class] action cannot predetermine the res judicata effect of the judgment; this can be tested only in a subsequent action.”96 If in the damages context, potential claimants identify themselves and claim to be class members, in the res judicata context, potential plaintiffs identify themselves and claim not to have been class members. In the event that the defendant in the “subsequent action” argues that the new claim is precluded, the judge must determine whether the plaintiff had in fact been a member of the putatively preclusive class. For this determination, the ascertainability of the first class is not relevant. Res judicata determinations by design involve individual inquiry, and what matters for such determinations is only whether the putatively preclusive class displays a minimally clear definition.

To see how this works, consider a series of hypotheticals. First, imagine a fully ascertainable class of people employed by Safeway in 2014. Assume that there are very clear records of whom Safeway employs, and that virtually all individual members of the class have documents establishing their employment.97 The employees sue Safeway for a violation of employment law, have their class certified, and lose on the merits. Final judgment; res judicata. The next year, a Safeway employee brings the same claim again, but as an individual. Predictably, Safeway says res judicata. The court takes the new plaintiff’s argument for independent injury (“I was a Safeway employee in 2014,” he says, “and I have a legal claim”), holds it up to the definition of the putatively preclusive class, and realizes something remarkable: the very claim the plaintiff is making to establish his own injury in Case 2 is exactly what demonstrates his class membership in Case 1. Accordingly, the court grants Safeway’s motion to dismiss.

Now consider the unascertainable class of people who purchased a takeaway roast chicken from Safeway some time in 2014. Let’s assume that Safeway keeps no records of who bought chickens, only how many were sold. Let’s assume that most people don’t keep the receipts to prove their chicken purchases.98 And let’s assume that, even if they did, the receipts said “Food To-Go,” which hardly identifies chicken or the purchaser with specificity. An enterprising plaintiff files suit on behalf of this group, claiming that the chickens were negligently mislabeled. The court grants certification99 and finds that Safeway labeled its chickens with the greatest care. Plaintiffs lose on the merits. Final judgment; res judicata. Then suppose someone comes forward the next year with exactly the same claim, but as an individual. (It turns out that this plaintiff is enough of a “lunatic or a fanatic” to sue for the small value of the individualized claim.100) Safeway, again, says res judicata. The court looks at the new plaintiff’s claim to injury (which starts with evidence that he really bought a chicken from Safeway in 2014), holds this claim up against the definition of the putatively preclusive class, and realizes something remarkable: the very assertion the plaintiff is making to establish his own injury in Case 2 is exactly what demonstrates his class membership in Case 1. Case precluded; case dismissed.

Notice that the ascertainability, or lack thereof, of the first class had nothing to do with the analysis of preclusion in the second case. What does make a difference is whether the class definition delimits a discrete category.101 Consider what happens when the definition of the first class lacks a minimally clear definition. Let’s say the concern was that Safeway’s advertising harmed children, and that the class was defined as young people who purchased roast chickens from Safeway sometime in 2014. Suppose that somehow this class were certified, and that the case led to a final judgment. If, the next year, a nineteen-year-old sued Safeway on the same theory and demonstrated convincingly that he did in fact purchase a chicken in 2014, Safeway might say “res judicata.” The court would then have to determine whether nineteen counted as “young.” By the same token, it has to be clear what chickens are if the next court is to determine whether someone bought one. The class of customers who bought “organic vegetables” or “heart-healthy foods” could pose difficulties on this score unless the meaning of “organic” and “heart-healthy” were fixed in the prior class definition. “[I]n determining who was bound by an earlier ruling arising from a class case,” amici told the Carrera court, “what matters is the clarity of the class definition.”102

The need for a minimally clear definition of this sort is starkly different from a requirement that the class be ascertainable. Unlike ascertainability, minimally clear definition is unrelated to administrative feasibility. No amount of administrative expense, individual scrutiny, or new information will conclusively determine whether some individuals fit into a class without a minimally clear definition; conversely, a minimally clear definition ensures only that the class will be ascertainable before an omniscient court. And clear res judicata does not require that it be “administratively feasible” for the first court to identify the first class—because the next plaintiff herself initiates a new proceeding (by definition a “mini-trial”) and bears the burden of pleading her supposedly precluded claim. In addition, clear res judicata does not require that the proposed method of identifying members of the first class be “objective.”103 Even if proof of membership in the class in the first case was based entirely on “say-so,” it is not possible for someone who fits into the minimally clear class definition and is therefore genuinely barred from suit to re-state the same claim in a new case and at the same time declare that she was not a member of the first class. If the class definition is clear and the claim is truly precluded, one cannot “say so” to independent injury in Case 2 and at the same time “say no” to class membership in Case 1. One’s “say-so” that one was a member of the class—the issue that defendants worry about because it could result in illegitimate damage payouts—can only result in an affirmative finding of res judicata for genuinely precluded claims.

Finally, throughout this discussion we must not forget that ascertainability has mainly been used to block small dollar consumer claims–exactly the sort of case one would probably not bring on ones own.104 So it is a bit strange to say that ascertainability is necessary to clarify res judicata when the reality is that future cases in which res judicata would arise as a defense would in many instances not be brought – cases in which legal preclusion takes a back seat to economic preclusion.

II. definitional malfunctions: subjectivity, vagueness, and scope

Beyond the considerations of notice, remedies, and res judicata discussed in Part I, courts and commentators worry that there is something wrong with subjective or vague classes and classes with problematic scope–whether “overbroad” or “failsafe”–and suggest that the ascertainability requirement answers these concerns. In this Part, I argue that using the ascertainability requirement to address these class malfunctions is both confusing and unnecessary.

A. Subjectivity and Vagueness

A central worry underlying ascertainability doctrine is that there is something wrong with vagueness or subjectivity in class definition, and that the ascertainability requirement roots out vague and subjective classes. But concerns about subjectivity and vagueness do not justify an implicit ascertainability requirement. To see why, we need to observe that there are two kinds of subjectivity and two kinds of vagueness at stake in ascertainability cases–an ontological and an epistemic version of each–that merit different treatment.105

1. Subjective Classes

Consider two different ways in which a class can be “subjective.” First, the class definition can build in inherently subjective factors. Here, a person’s membership in the class is a function of his or her own mental states.106 Imagine the class of people “who were offended” or “deceived” by an advertising campaign.107 This type of subjectivity arose when a district court refused to certify the class of “all individuals who consumed [D]iet Coke from the fountain, deceived by the marketing practices employed by Coca–Cola Company into believing that fountain [D]iet Coke does not contain saccharin.108 Class definitions that incorporate states of mind can be dealt with in several ways, none of which requires an ascertainability requirement. First, in most cases, these subjective definitions violate the minimally clear definition requirement because they present classically vague predicates. What it means to be “offended” or “deceived” simply is not clear. Also, perhaps because ontologically subjective classes fail to establish a minimally clear definition, these classes will often fall short on Rule 23(a)’s explicit requirements. For instance, one concern with a class of people who were “offended” by a statement (assuming that there is a cause of action that makes such a class definition relevant) is not that there is something the matter with awarding liability for offendedness per se; it is that being offended or deceived means different things for different people. Precisely because being offended is subjective, it varies. So a class of offended or deceived people could have a hard time proving commonality.109 By the same token, it would be hard for the named plaintiffs to demonstrate that they were offended or deceived in the same way as the absent class members, which invites a typicality challenge.110 Sometimes, courts have tackled ontologically subjective definitions under the heading of manageability. In Simer v. Rios, in 1981, the Court of Appeals for the Seventh Circuit considered a class of people “discouraged from applying for [financial] assistance because they were not delinquent in the payment of their fuel bills for 1979111 and asked whether the issue of each individual plaintiffs state of mind makes the class action unmanageable.112 The court turned to the concept of manageability113 and found that the issue of state of mind [did make the] case difficult to manage as a class action.114 The central point is simply that we do not need an implicit ascertainability requirement to block classes based on state of mind.

The second form of subjectivity is different. Whether or not the definition builds in subjective factors, the proposed means of determining membership sometimes involve subjective evidence. This was the problem in Carrera. Whether one purchased a weight loss pill is an objective fact; the subjectivity arose because the court was asked to accept a potential class member’s “say-so” as the criterion for membership. Carrera provides a nice reminder that this epistemic form of subjectivity and the ontological form of subjectivity are not tethered together. For example, one could use “say-so” to determine the membership of a class that would also be ascertainable from objective records: “Raise your hand if you subscribe to the Yale Law Journal.” “Say-so” may be the most efficient way to identify classes with objective but elusive definitions: “Raise your hand if you read the Yale Law Journal.” And, of course, “say-so” is often the only way to identify membership in classes with genuinely subjective definitions: “Raise your hand if you enjoy the Yale Law Journal.” This epistemic form of subjectivity is not related to class definition: it is a problem of the difficulties of managing the litigation. As I explained in the discussion of remedies above, Rule 23’s text explicitly empowers courts to take management difficulties into consideration at the certification stage and to ask, in light of those difficulties, whether the class action is a superior method of resolving the claims. While in many cases, relying on “say-so” may result in overwhelming difficulties relating to the provision of notice and the disbursement of damages, rising to the level of genuine unmanageability, in many cases it may not. In Part IV, I address how an analysis rooted in manageability and superiority works and explain its advantages. The point for now is that neither form of subjectivity calls for an implicit ascertainability requirement.

2. Vague Classes

Like we saw with “subjective” classes, there are actually two distinguishable problems being discussed under the label of vagueness, one ontological and one epistemic. First, some classes–the class of “bald people” or “young people”–are inherently vague because the class definition builds in vague terms. The problem with these classes is that they lack a minimally clear definition, not that they are unascertainable. What, for instance, would we say about the class of bald purchasers of Rogaine, all armed with receipts, suing for false advertising? What would we say about the class of young people suing Safeway? Vaguer still, what would we say about the class of “slithy toves” who “gyre[d] and gimble[d] in the wabe”?115 The problem with these classes is that even if the court had all the information in the world, and even if it performed numerous “mini-trials,” there would still be unresolvable disputes about membership. The problem is with the conceptual discreteness of the category in question, not with the information needed to establish membership in that category. As a result, it would be inaccurate to say that the class of bald purchasers or young people is uncertifiable because of the administrative feasibility of establishing one’s baldness or youth. Instead, it makes much more sense to say simply that the class lacks a minimally clear definition.116 In fact, before ascertainability doctrine came into vogue, courts did just this. Take, for example, DeBremaecker v. Short,117 a case from 1970 that is sometimes cited as an early precursor to ascertainability doctrine.118 In DeBremaecker, the Court of Appeals for the Fifth Circuit blocked certification of a class defined as “residents of this State active in the ‘peace movement.’”119 In refusing to certify the class, the court pointed to the “uncertainty of the meaning of ‘peace movement.’”120 This class is probably not ascertainable by today’s standards. How would a court determine whether a person was active in the peace movement without significant administrative expenditure or an appeal to overly subjective factors? But that is not the core problem. The core problem is that the terms used in the definition are inherently vague. In principle, no court, no matter how much information it received, could determine the full membership of the class.

Other classes are vague in the sense that it is difficult to identify individual class members even though it is logically possible to do so. The class of “all bald people” is inherently vague because there are necessarily disputable cases of baldness.121 But there may also be disputable cases of membership in the class of people who have fewer than a thousand hairs if the process of counting hairs is imperfect. While the class of “heavy smokers” is inherently vague, the class of people who “smoked Marlboro cigarettes for at least twenty pack-years”–a real class, dismissed on ascertainability grounds122–successfully delimits a discrete category. The potential vagueness lies in the difficulty of determining who fits the definition, not in the definition itself. Carrera fits into this category too: whether someone purchased a weight loss pill is not inherently vague; one’s purchase is a fact in the world. The difficulty lies in information gathering. This epistemic form of vagueness is not a problem of class definition; it is a problem of the difficulties of managing the litigation. And here, Rule 23’s text supplies directly relevant resources. As I explained in the discussion of remedies above,123 the Rule’s manageability and superiority provisions equip courts to evaluate likely information gathering difficulties and dismiss classes presenting overwhelming burdens. But while the ascertainability requirement does not offer much beyond throwing these classes out of court, regardless of whether the difficulties of identifying the class members would actually derail the litigation, analysis of manageability and superiority invites courts realistically to consider the likely difficulties and to make a better tailored certification decision. In Part IV, I explain how this analysis works and argue that it is superior to an ascertainability test. The point for now is that neither form of vagueness justifies an implicit ascertainability requirement.

B. Problems of Scope

The ascertainability requirement is sometimes hailed as a way to block classes that suffer from two different problems of definitional scope: overbroad definition and failsafe definition. Neither problem justifies the ascertainability requirement.

1. Overbroad Classes

The first problem of scope is overbreadth. Overbroad classes contain too many people who would have no legitimate claim if they sued on their own. Overbreadth therefore raises concerns for defendants, who should not have to pay illegitimate claims, and problems for genuinely harmed members of the class, whose claims would be diluted. Courts have perceived a close connection between the ascertainability requirement and overbreadth.124 A defendant recently told the Northern District of California that a proposed class is “unascertainable because it is overbroad.”125 And the Central District of California recently “agree[d] . . . that [a] proposed class is not sufficiently ascertainable because it is overbroad.126

Overbroad classes can be blocked under existing Rule 23 requirements. If a class is defined in an overbroad way, the claims of the members of the class will by definition not be common, and the claims of the named plaintiffs will not be typical of the class as a whole. General Telephone Co. of the Southwest v. Falcon provides a classic example of precisely this scenario. In that case, plaintiffs sought to certify a class of “all hourly Mexican American employees who have been employed, are employed, or may in the future be employed and all those Mexican-Americans who have applied or would have applied for employment had the Defendant not practiced racial discrimination in its employment practices.”127 As Rubenstein argued, the “class definition would arguably have failed” to satisfy the ascertainability requirement because “it would be somewhat difficult to ascertain” the “membership” of the class. But the “Court made no mention” of ascertainability because it did not need to: it “rest[ed] its decision on the class’s failure to meet the commonality and typicality requirements of Rule 23(a)(2) and Rule 23(a)(3).”128 Like commonality and typicality, predominance and superiority are helpful resources here too. For example, a judge of the Northern District of California ultimately concluded that “[t]he overbreadth question” in an allegedly unascertainable class “is better addressed in the context of 23(b)(3) predominance.”129 Courts can also narrow the definition of the class to cure the overbreadth problem; it is not always a problem that must block certification outright.130

Unlike courts that deploy the ascertainability requirement as a cure for overbroad classes, other courts sometimes fixate on ascertainability without recognizing that the real problem is overbreadth. Recently, a court wrote that “[t]he ascertainability requirement is not satisfied when the class is defined simply as consisting of all persons who may have been injured by some generically described wrongful conduct allegedly engaged in by a defendant.131 Although the court viewed the problem in terms of ascertainability, the real problem is overbreadth, best addressed through the commonality and typicality prongs of Rule 23(a). It may be impossible to show common injury for everyone who may have been injured (unless, of course, that potential injury was actually the injury being alleged in the suit, as would be the case for a class of people placed in danger, for example). And it would be difficult to establish that the claims of a named plaintiff are typical of the class as a whole if the only claim being made is that everyone may have suffered the allegedly typical injury.

More importantly, ascertainability does not always root out overbroad classes. In fact, an overbroad class is often more ascertainable than a narrowly tailored one. Return to the hypothetical chickens and assume that everyone paid with a credit card, so there were personally identifiable records of who shopped at Safeway in 2014 in at least three different places: Safeway’s records, credit card records, and personal records. But let’s also assume that the receipts and records identified takeaway chickens as “food to go,” the same designation as numerous other products at around the same price point. The real, narrowly tailored plaintiffs class in this case is unascertainable. But an overbroad class—people who shopped at Safeway in 2014 and bought things around the same price as chickens documented as “Food To-Go”—is entirely ascertainable. Far from efficiently blocking overbroad classes, the ascertainability requirement can bias class definition toward overbreadth.132

2. Failsafe Classes

The second problem of scope is not under-breadth, but something more slippery: the “failsafe” class, the situation where class membership is defined partly in terms of the merits of the claim being advanced.133 If we return to the Safeway example, a class of all persons who were sold illegally mislabeled chickens in 2014 is problematic because membership in the class depends upon the resolution of the claim on the merits. If the plaintiffs win, and the court finds that the chickens were indeed illegally mislabeled, they are paid and precluded. But a ruling for the defendants, finding that the labeling was legal, transforms the class into an empty set. Nobody gets paid, but nobody is precluded. Hence a failsafe class definition is a classic example of “heads I win, tails you lose.”134 The temptation to define a class this way probably arises less from dreams of defeating res judicata and more from fears of succumbing to overbreadth. After all, the scope of a failsafe class is tailored perfectly to the alleged injury—by definition. Yet whatever the motivation, failsafe classes distort the incentives involved in class litigation, in which plaintiffs take on the risk of preclusion in return for potential payout, which could result in excessive future litigation.

Strangely, failsafe class definitions have been diagnosed as ascertainability problems.135 The intuitive idea behind this diagnosis is that it is impossible to identify the members of the class in advance of the judgment because the very criterion for membership, what the court will hold, has not yet come into being. Ascertainability provides a clever but ultimately unsatisfying diagnosis of what is wrong with failsafe classes. Thinking that the fundamental problem with a failsafe class definition is that the class is not ascertainable has the deliciously ironic implication that adjudication on the merits is neither objective nor administratively feasible. In fact, the core problem with a failsafe definition is unrelated to ascertainability. On a basic level, suing on behalf of a failsafe class violates a basic procedural principle that one must identify the plaintiff and the defendant before one states a legal claim.136 It is as absurd to define a class as all those who purchased illegally mislabeled chickens as it is to sue he who wronged me and then expect the court to figure out just who he really is. Identifying “the plaintiff” is more complicated in a class action—the procedures established in Rule 23 form a template for how to do so. But the fact that an action is brought on behalf of a class rather than an individual does not in any way permit departing from the sequence of pleading.137

Fundamental diagnosis aside, Rule 23 already provides the necessary equipment to block failsafe class definitions. Courts have ruled that a failsafe class fails to establish numerosity, given that there will be no class members if the plaintiffs lose.138Some courts point out that failsafe classes make adequate notice impossible.139 As a result, some courts have dismissed failsafe classes on manageability grounds.140One could also argue that because a failsafe class sidesteps the basic incentive structure of aggregate litigation—where plaintiffs assume the risk of preclusion in exchange for a potential payout—class adjudication fails to be a superior method of adjudication.141

III. keeping faith with rule 23

In Part I and Part II, I argued that the ascertainability requirement is unnecessaryneeded neither to facilitate notice, remedies, and preclusion, nor to answer concerns about subjectivity, vagueness, and scope. In this Part, I argue that the requirement is counterproductive. The requirement clashes with the original vision of Rule 23 and frustrates its goals.

“The historic mission” of Rule 23 is “protecting the small guy.”142 By incentivizing private attorneys to identify and represent groups with legitimate injuries and by encouraging active “judicial initiative and management”143 of suits, the Rule aimed to “enhance the [litigation] opportunities of hitherto powerless groups.”144 “Individuals today, probably more than they realize,” one commentator wrote in 1969, shortly after the Rule was amended to look much like it does today,

are suffering from and being damaged by the practices of large corporate bodies. . . . How may the small claimant be compensated for these damages? What alternative is there to the effective use of the class action? Asked thirty years ago these questions might have been unanswerable. But today Rule 23 exists in an amended and improved form.145

The “improved” Rule 23 empowered the “small claimant” by harnessing economies of scale. “The policy at the very core of the class action mechanism, the Supreme Court explained thirty years later, is to overcome the problem that small recoveries do not provide the incentive for any individual to bring a solo action prosecuting his or her rights.”146 As Alexander has explained,

Class actions [are] a way of leveling the playing field for poor or economically less powerful individuals. Normally an individual, especially a poor person, is at a great disadvantage in a court case against a well-financed corporate opponent who can afford high-priced lawyers. But when claims are brought together in class action form, the aggregate amount may be large enough to make it possible to engage the services of equally skilled counsel. The effectiveness of class actions in empowering the economically powerless against the economically powerful may be one reason that they have been under almost constant attack by business interests and conservative politicians.147

A second goal for the class action was law enforcement—something Kalven and Rosenfield understood as a cardinal benefit of aggregation. “If each is left to assert his right alone if and when he can,” they wrote, “there will at best be a random and fragmentary enforcement, if there is any at all. This result is not only unfortunate in the particular case, but it will operate seriously to impair the deterrent effect of the sanctions which underlie much of contemporary law.”148 “By ‘enabling’ claims,” Alexander explains, “the class action device can provide appropriate incentives for corporations, assuring that they pay the true costs of their own conduct, rather than passing the costs on to consumers while retaining the benefits as profit.”149

Enthusiastic though they were about its potential, the drafters of Rule 23 recognized problems endemic to class suits. These benefits of class aggregation did not come without risks. How would aggregate litigation actually work, given the disparities of information, resources, incentives and motivation among potential claimants?150 What about potential class members who did not want their claims litigated en masse?151 How could class action procedure ensure that absent class members were aware of the proposed suit and their rights in the first place?152 And if one managed to solve these problems, others lurked close behind. How would a court dole out damages to absent class members?153 Who would be bound by the outcome of a representative suit?154 What about fairness to defendants?155

In light of these concerns, the drafters built procedural safeguards into Rule 23. For example, the Rule requires commonality, typicality, and adequate representation in addition to mere numerosity,156 provides for notice and the opportunity to opt out,157 and requires judicial approval of settlements.158 But in crafting these safeguards, the drafters were creating a procedural mechanism categorically different from individual litigation or joinder. The goal was to include and aggregate the claims of parties who were left behind in other procedural schemes. Rule 19, the joinder Rule, enables present, known individuals to unite their claims;159 Rule 23 provides for a different aggregation mechanism that unites the claims of absent, unknown parties.

Accordingly, the drafters of the Rule avoided safeguards that would require the identification of each individual class member. Here is an example. Foreshadowing the current literature on choice architecture,160 the drafters of Rule 23 openly debated whether Rule 23(b)(3) should include an opt-out procedure or an opt-in procedure. That is, should absent class members have to identify themselves to join the class, or should they be considered part of the class automatically, unless they chose affirmatively to opt out? If the drafters had answered this question the other way and gone for opt-in, we would not be debating ascertainability doctrine today. In that alternative world, Rule 23(b)(3) classes would be formed by joining together known individuals, one by one. Classes would be ascertainable by design. The Rule would look a lot like a joinder provision. The drafters did consider the argument that an opt-in system was fairer to defendants. Here is how Kaplan described the debate:

It was suggested that the judgment in a (b)(3) class action, instead of covering by its terms all class members who do not opt out, should embrace only those individuals who in response to notice affirmatively signify their desire to be included (others might perhaps be included if considered essential by the court). It is unfair to a defendant opposing the class, so the argument goes, to subject him to possible liability toward individuals who remain passive after receiving notice or who may, indeed, have had no notice of the proceeding: under the previous law, some, perhaps many, of those persons might simply have foregone any claims against the defendant; they might in fact have remained ignorant of having any possible claims. Running through this argument was the idea that litigation should be a matter for distinct action by each individual.161

But the drafters found these considerations unpersuasive. Here is what they concluded: “[R]equiring the individuals affirmatively to request inclusion in the lawsuit would result in freezing out the claims of people–especially small claims held by small people–who for one reason or another, ignorance, timidity, unfamiliarity with business or legal matters, will simply not take the affirmative step.”162 Bringing the “claims held by small people” into court, regardless of whether their identities were identifiable, was too important. Accordingly, the drafters settled on opt-out. “[I]t seems fair, Kaplan wrote, for the silent to be considered as part of the class. Otherwise the (b)(3) type would become a class action which was not that at all.163 Put differently, the drafters affirmatively chose to create an aggregation mechanism that would include the claims of unknown people even considering the challenges that could result.

The drafters hoped courts would charge forward into this uncharted territory with a spirit of accommodation. Specifically, although the word “ascertainability” was not part of the class action vocabulary in the 1960s, it was not unpredictable that questions would arise about whether and how to identify unknown class members. In a passage that eerily evokestoday’s ascertainability cases, Tom Ford, a commentator on the 1966 amendments, identified the basic problem of ascertainability:

It is assumed that five individuals have evidence that certain suppliers over a period of years have fixed the prices of a product which each of these individuals use. After determining their possible damages (which for the five total approximately 2500 dollars for the period), and after retaining and consulting with counsel, they institute a class action on behalf of themselves and all other purchasers of this product. Plaintiffs, of course, feel that there are a great many members of the class. However, they have no means for securing the names and addresses of these class members. In a case where the defendants can supply this information, the major barrier to the plaintiffs in the definition of the class is whether, and under what circumstances, the court will permit discovery to secure this data. On the other hand, it may be supposed that these names and addresses cannot be furnished by defendants. How will other members of the class learn of the pendency of this action? How will the plaintiffs firm up the class? These questions are not easily answered now; but, for the amended Rule to be effective in these kinds of class actions, the courts will have to fashion rules of accommodation.164

Ford expected courts to accommodate[e] what we now call unascertainable classes. Otherwise, the amended Rule would not be effective.

Ascertainability doctrine reflects exactly the opposite impulse. It pushes out of court the very classes that Rule 23 was designed to bring in to court and as a result makes the Rule less effective. It does so because it misses the point, expressed by the Supreme Court, that “[t]he class action is an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.165

The distinctive character of class aggregation under Rule 23 illuminates the fundamental error in the argument that ascertainability protects the right of defendants to challenge each individual claim in a class action.166As Judge Scirica wrote in Carrera, “Ascertainability provides due process by requiring that a defendant be able to test the reliability of the evidence submitted to prove class membership.”167 This is both because “[a] defendant in a class action has a due process right to raise individual challenges and defenses to claims, and a class action cannot be certified in a way that eviscerates this right or masks individual issues,”168 and because “[a] defendant has a similar, if not the same, due process right to challenge the proof used to demonstrate class membership as it does to challenge the elements of a plaintiff’s claim.”169 This argument misses the point, which was put to the court by amici, “that class actions offer a form of representative litigation distinct from traditional joinder mechanisms.”170 To say that due process requires the opportunity to challenge the claims of each individual class member—or to say that interpreting Rule 23 to exclude such a right would “abridge” or “modify” the substantive rights of defendants171—would represent a radical departure from established practice and question the foundational logic of the class action mechanism. If in ordinary litigation, defendants are entitled to challenge the legitimacy of the individual plaintiff’s individual claim, in a class action, defendants are entitled to the “rigorous”172 application of the written requirements of Rule 23 and the “rigorous” enforcement of all the Rule’s carefully engineered procedural safeguards. By design, these safeguards do not envision the specific identification of each class member.

Carrera was also concerned, for the sake of the absent class members, about the dilution of claims. The court observed that “[i]t is unfair to absent class members if there is a significant likelihood their recovery will be diluted by fraudulent or inaccurate claims”173 and suggested that the ascertainability requirement was necessary to guard against dilution. Dilution, however, is best understood as an overbreadth problem, which is best addressed in the context of the commonality and typicality requirements of Rule 23(a).174 More to the point, however, defeating class certification based on dilution concerns squarely conflicts with the goals of class aggregation. Amici in Carrera put it perfectly: “[T]o deny class certification based on a fear of dilution of claims would, in effect, deprive potential class members of any recovery as a means to ensure they do not recover too little.”175

Today, some scholars of civil procedure are concerned that the original purposes of Rule 23—and the broader ambitions of the Federal Rules of Civil Procedure as a whole—are under attack from multiple angles.176 As John Coffee and Stefan Paulovic wrote, “For better or worse, it is today clear that the tide has turned against class certification, and new barriers have arisen across a variety of contexts where formerly class certification had seemed automatic.”177 Ascertainability doctrine is part of the same story.178 This Note has documented numerous instances of classes dismissed on ascertainablility grounds. The requirement has defeated the class of Snapple drinkers who alleged that the “All Natural” label on the iced tea was misleading,179 the class of Sam’s Club shoppers who purchased “Service Plan[s] to cover as-is products,”180 the class of people who bought fraudulently marketed weight loss supplements,181 the class of people who had paid for run-flat tires that turned out to be defective,182 and many more. The list goes on. The point is not that the claims of these consumers were meritorious, or even that they might have been. The point is that these groups could never even state their claims in court because they missed out on class certification. In these and similar cases, no compensation is available for victims, the law goes unenforced, and no judgment guides lawful behavior for others.

A small number of courts see the ascertainability requirement in this light. Judge Rakoff wrote that the ascertainability requirement has the potential to render class actions against producers almost impossible to bring.183 Judge Jon Tigar wrote that

[a]dopting the Carrera approach would have significant negative ramifications for the ability to obtain redress for consumer injuries. Few people retain receipts for low-priced goods, since there is little possibility they will need to later verify that they made the purchase. Yet it is precisely in circumstances like these, where the injury to any individual consumer is small, but the cumulative injury to consumers as a group is substantial, that the class action mechanism provides one of its most important social benefits. In the absence of a class action, the injury would go unredressed.184

And a decision in the Southern District of California recently explained that “[i]f class actions could be defeated because membership was difficult to ascertain at the class certification stage, there would be no such thing as a consumer class action.185 These small dollar consumer class actions are precisely the cases that the drafters of Rule 23 had in mind and wanted to enable: cases in which individuals “are in a poor position to seek legal redress, either because they do not know enough or because such redress is disproportionately expensive.”186 In fact, “[W]hen what is small is not the aggregate but the individual claim; indeed that’s the type of case in which class action treatment is most needful.”187

IV. a return to the text

In this section, I defend a simple alternative to ascertainability analysis–not a novel way of interpreting Rule 23, but an uncomplicated, ordinary way rooted in the text and purpose of the Rule. When a court confronts what we currently understand as an ascertainability problem, it should turn to faithful old friends: the requirements explicitly in Rule 23. In the previous Parts of this Note, I argued that Rule 23’s explicit requirements, combined with a demand for a minimally clear class definition, are powerful enough to ensure the smooth functioning of litigation and to protect against class malfunctions. In this Part, I argue that a return to the text makes possible certification decisions that better reflect the actual realities of class litigation, that take into consideration the realistic alternatives to class actions, and that better vindicate the purposes of Rule 23. This approach will secure the benefits that the ascertainability requirement claims to provide without making consumer class actions “almost impossible to bring.”188

A rigorous analysis of “manageability” invites courts to evaluate the actual burdens a class action is likely to entail. How will notice be delivered? Is a suitable remedy available? Certification decisions rooted in these questions will likely be considerably more forgiving than decisions applying an ascertainability requirement. Determining membership by “say-so,” for example, need not block certification in every case. In some cases it might be the most efficient way to determine membership. Imagine the class of people who purchased REI sneakers. Imagine that each purchaser kept her receipt, but somewhere inconvenient, and that REI had records of each purchase, but only inaccessibly located in a faraway storage facility. If the court felt that the explicit requirements of Rule 23(a) were satisfied, the risk of error of “say-so” for small damage awards might be worth it if the costs of documentary identification were too high. The court would ask, with regard to the specific case before it, whether it really could deliver adequate notice and facilitate opt-out and whether an appropriate remedy was available. The question must be whether the actual challenges entailed in the administration of [the] class are . . . so burdensome as to defeat certification.189

In this analysis, the court must remember that it is the manageability of a class action it is evaluating. As the drafters of Rule 23 and lawyers since have understood, class actions inherently involve administrative burdens, individual inquiry, and some uncertainty. In Ford’s words, courts should be skeptical of “dismal specters paraded before them by defendants with regard to the action’s manageability” and instead proceed “in accord with the Rule’s purposes.”190

By the same token, a rigorous analysis of “superiority” invites courts to consider the realistic alternatives to the class action. To illustrate, the Carrera court claimed that“[i]f a class cannot be ascertained in an economical and administratively feasible manner significant benefits of a class action are lost.191 By contrast, the court could have asked whether, in light of the economics of the situation, the class action mechanism is nevertheless superior to other methods of resolution. This analysis would force the court to reckon with what, exactly, those other methods of resolution might be. If [the] realistic alternative to a class action is not 17 million individual suits, but zero individual suits—as Judge Posner suggested was often the case—then a class action with some administrative burden may still be superior to nothing at all.192 This analysis permits courts to decline to certify classes so difficult to administer that the demand for individualized procedure would drown out any hope of aggregate efficiency. But part of this analysis must be to consider what alternatives are really available. In many situations “it would be specious to argue that” individual suits or “wholesale joinder would be either practicable or desirable.”193

One might prefer the stability and predictability of a fixed ascertainability rule to the flexibility and uncertainty of case-by-case certification analysis and therefore opt for an ascertainability requirement all the same. After all, ascertainability is the axe to manageability’s scalpel. But the ascertainability requirement’s attempt to achieve stability and predictability comes with an exacting price: forestalling precisely the kind of litigation the Rule was written to enable–the kind that protects “the small guy” where other procedural mechanisms cannot.194

Consider Safeway and its chickens one last time. The class of people who purchased a chicken in 2014 is likely unascertainable, and on that basis a court today might deny certification. But imagine if the court performed a rigorous analysis of Rule 23’s text instead. The court might find that the named plaintiffs could indeed establish that the class is numerous, that there are common claims, that the named plaintiffs’ claims are typical of the class’s claims, and that those plaintiffs can adequately represent the absent members. The court could consider whether the definition of the class is minimally clear–that is, whether it is clear what Safeway is, what chickens are, and when 2014 began and ended. The court could forecast the actual difficulties the litigation will present and ask, in light of those difficulties, whether the case is truly manageable. The court could likely find a way of providing adequate notice. Because of the value of the individual damage awards, there would likely not be overly burdensome mini-trials over who actually bought a chicken; there would likely not be vast numbers of illegitimate claims. Finally, the court could consider whether there is any other way to compensate potential victims and any other way to deter illegal action in the future. This analysis would likely point toward certification. What the court should not do is end the case just because Safeway kept no relevant records195 and customers lost their receipts. Rule 23 demands more of courts than that.

Some courts recognize the advantages of this approach. In a recent case in the Northern District of California, plaintiffs sought to certify a class of “[a]ll persons in California who purchased an Arizona brand beverage from March 17, 2006 until the present time which contained High Fructose Corn Syrup or citric acid . . . which were marked, advertised or labeled as being All Natural, or 100% Natural[.]’”196 Prior to a rigorous analysis197 in which the court found that the Rule 23(a) requirements were satisfied,198 the court considered Arizonas ascertainability objection. Here is what the court said:

Defendants suggest that simply because most members of the proposed class will not have retained all of their receipts for AriZona Iced Tea over the past few years, the administration of this class will require “fact-intensive mini trials” to establish whether each purported class member had in fact made a purchase entitling them to class membership. This is simply not the case. . . . The challenges entailed in the administration of this class are not so burdensome as to defeat certification.199

Instead of dismissing the class on ascertainability grounds, the court examined what challenges it would face in administering the litigation. It then asked whether it was up to the job. The same form of analysis would likely have pointed toward certification in Carrera v. Bayer Corp., Marcus v. BMW of North America, LLC, Weiner v. Snapple Beverage Corp., and many other cases, though that of course would be a question for a court revisiting the cases and reading the Rule afresh. The very smallness of the individual claims in these cases makes it more likely that the class action is superior to other methods of resolution and less likely that administrative burdens would overwhelm the court.

Finally, I need to consider and rebut the argument that the ascertainability requirement somehow follows from the text to which I propose we return. There are two main arguments that the ascertaiability requirement is anchored to the text of Rule 23, each unpersuasive. First, it would be a mistake to read Rule 23(c)(1)(B)’s mandate that the certification order “define” the class as support for the ascertainability requirement. “Defin[ing]” the class at certification merely requires a statement of the parameters of the class sufficient to guide the delivery of notice and to trigger opt-outs–the tasks required immediately after certification.200 As we have seen, the text of the notice provision assumes that some class members cannot be “identifed through reasonable effort.”201 At the end of the case, in a judgment, the Rule insists that the court “specify or describe” the people “the court finds to be class members.” 202 The court may do so by restating the parameters of the class and subtracting people who opted out from “those to whom . . . notice was directed.”203 In other words, the Rule envisions the class definition evolving over time. At first, the court must establish the parameters of the class in order to provide notice; at the end of the case, the court must refine the definition to reflect that some potential class members opted out. These definitional requirements are fundamentally different from ascertainability. If the Rules Committee had indeed intended to codify an ascertainability requirement with the word “define,” it could and would have stated the exact kind of definition it had in mind.204 Such an interpretation of the word today would throw the Rule’s definitional logic out of balance.

A second argument claims that ascertainability is part of what it takes to be a “class” in the first place.205In this vein, the requirement looks like a threshold to the thresholdrequirements, a characteristic of classes presupposed by the threshold tests.206 This argument sees ascertainability as a feature of a cohesive group. Ascertainability, however, is a feature of the relationship between a group and the person or institution doing the ascertaining–in this case, the court–and is not the sort of property that would change the group’s ontological status. The group of people who purchased Safeway chickens in 2014 and kept their receipts is every bit as much a “class” as the group of people who purchased Safeway chickens in 2014 and did not. A stronger position might be that a “class” requires a minimally clear definition establishing a discrete category. The fundamental difference between the class of young people and the class of people under eighteen does relate to the nature of the group itself. That said, what matters is not the semantic sting of the word “class,” but what kind of classes are appropriate for adjudication under Rule 23–and that question brings us back to the arguments discussed in this Note.


This Note has analyzed the implicit ascertainability requirement and concluded that the arguments to justify it are unpersuasive. If one is concerned about notice, one’s concern is misplaced: Rule 23 does not require individual notice to each class member. Concerns about the remedies process speak to the manageability of the litigation – something Rule 23 explicitly invites courts to consider. The clear application of res judicata requires merely a minimally clear class definition, something far different from ascertainability. Similarly, vague or subjective classes sometimes pose practical difficulties beyond the ability of the court to manage (which the Rule explicitly invites courts to consider) and sometimes fail to establish a minimally clear definition (which is not an ascertainability problem). The real problem with overbroad classes has to do with commonality and typicality, and the real problem with failsafe classes is more fundamental than a lack of ascertainability. Finally, if one is concerned about the defendant’s right to challenge each individual claim, one has misunderstood the nature of class aggregation, which is fundamentally different from joinder or individual litigation.

To put all this together, the ascertainability requirement lacks a coherent conceptual justification and is not necessary to the smooth functioning of class action litigation. The procedural malfunctions ascertainability supposedly protects against are better understood, and better attacked, in the context of the Rule’s written requirements. And the ascertainability requirement frustrates the purposes of Rule 23 by undercutting the basis for the very suits the Rule was written to enable. Courts should return their attention to the Rule’s text. Rather than asking whether a proposed class displays a property that is foreign to the text of Rule 23, courts should perform a “rigorous analysis” of the Rule’s written factors – appreciating the specific difficulties the litigation is likely to present and appreciating the realistic alternatives to class aggregation. The result of this simpler process will be a cleaner, more theoretically coherent law of class certification that better resonates with the vision of Rule 23.

Dissenting from the denial of en banc review in Carrera, Judge Ambro suggested that the Judicial Conference’s Committee on Rules of Practice and Procedure look into this matter: “Rule 23 explicitly imposes limitations on the availability of class actions. [Recent cases add] another—that class membership is reasonably capable of being ascertained. If the Committee agrees with that, how easy (or how hard) must this identification be?”207 On this point I agree with Judge Ambro. The ascertainability requirement has done its work in the shadows of implication, like a ghost in a carefully engineered machine. If we decide, as a society, that we really do want a strong ascertainability requirement and that we really do want fewer consumer class actions, the change should be established in the text of the Rule for all the world to see. The task of implementing a markedly narrower vision for class aggregation should not ride the coattails of a convoluted theory of what is “implicit” in existing law.

Within the caselaw, if we resign ourselves to the fact that the word “ascertainability” might be here to stay and predict that it will become a permanent fixture of class certification analysis, then we should reclaim its meaning and understand it in a way that enhances, rather than compromises, Rule 23’s mission. For example, a court might stick with the word “ascertainability,” even with an implicit ascertainability requirement, but construe it only to require a minimally clear class definition. This would be a major improvement on current doctrine–more faithful to the objectives of Rule 23 and arguably more consistent with the precedents on which ascertainability doctrine has been based. But the need for a minimally clear definition is not best understood as ascertainability-lite. It would be far better to call it what it is: a minimally clear definition.208 From a different perspective, we might embrace the concept of ascertainability, but regard it only as a noteworthy characteristic of many well-formed classes—not in any way as a general requirement for certification. Nobody disputes that if a class is ascertainable, it is frequently consistent with Rule 23’s other requirements. But we must not confuse what is nice for what is necessary.

These alternatives remain hypothetical. The ascertainability requirement is gaining momentum in the federal courts when it belongs on the defensive, if not on the run. Our task must be to confront ascertainability doctrine with clear eyes and to rebuild the jurisprudence of Rule 23 to reflect what the Rule says and what the Rule is for. “To permit the defendants to contest liability with each claimant in a single, separate suit, would, in many cases give defendants an advantage which would be almost equivalent to closing the door of justice to all small claimants.209 So wrote one of the first courts to interpret the amended Rule, in 1967. This is what we think the class suit practice was to prevent . . . . Its correct limitations must be ascertained by the experiences which brought it into existence.210