Rule 23(a) Prerequisites

Rule 23(a) Prerequisites

Rule 23(a) allows one or more members of a class to sue or be sued as representative parties on behalf of all members “only if” four prerequisites are met.¹ Each prerequisite has its own clause in the Rule:

• Numerosity: the class is so numerous that joinder of all members is impracticable.²
• Commonality: there are questions of law or fact common to the class.³
• Typicality: the claims or defenses of the representative parties are typical of the claims or defenses of the class.⁴
• Adequacy: the representative parties will fairly and adequately protect the interests of the class.⁵

All four must be satisfied in every federal class action, and failing any one is fatal to certification.⁶ A class that fits a Rule 23(b) category perfectly still fails if one prerequisite is missing.

The prerequisites reflect the functional approach of the 1966 amendments, which replaced an arbitrary and conceptualistic division of class actions with practical questions about whether a class suit is proper.⁷ Those questions fall into two pairs. Numerosity and commonality ask whether class treatment will pay off: numerosity looks for enough claimants that ordinary joinder would be unworkable, which supplies the efficiency gain that justifies a special procedure, and commonality asks whether the class suit will realize that gain by resolving shared questions of law or fact.⁸ A case that attacks the defendant’s “upstream” conduct, such as a defective product, a discriminatory practice, or a securities fraud, fits that test because the answer for one member is likely the answer for all. A case that turns on what happened to each person “downstream,” or on each person’s own understanding of a transaction, resists the “if as to one, then as to all” framework.⁹ A class suit that saves nothing invites skepticism, and one court of appeals voiced reluctance to let actions proceed where they are not likely to benefit anyone but the lawyers who bring them.¹⁰

Typicality and adequacy ask a different question: whether the named plaintiff is a safe proxy for the absent members. The question matters because the results of a class action generally bind absent members, which is why procedural protections such as notice exist.¹¹ Typicality is the gentler check, a non-exacting inquiry that ensures the class has a titular head with an actual grievance against the defendant. Adequacy does more work in practice. When members are far removed from one another and each stake is small, the class lawyer is effectively the class representative: the lawyer assembles the claim, funds the investigation, and makes the decisions, while members are “rationally indifferent” to the day-to-day course of the case, perhaps even to its outcome.¹²

The two pairs overlap. Commonality and typicality tend to merge,¹³ and typicality overlaps adequacy because both look to the potential for conflicts within the class.¹⁴ In Wal-Mart Stores, Inc. v. Dukes, the Court decided the case on commonality and found it unnecessary to resolve typicality and adequacy.¹⁵ One set of facts can therefore bear on several prerequisites, and a good answer still addresses each one by name, in the Rule’s order.

Numerosity

Rule 23(a)(1) asks whether “the class is so numerous that joinder of all members is impracticable.”¹⁶ The requirement makes the class action a fallback for groups too large for ordinary joinder under Rules 19 and 20. If joinder is practicable, each aggrieved person can sue individually or as part of a non-class collective case, and no representative action is needed.¹⁷ The logic is efficiency: class treatment must save enough over simpler forms of joinder to justify itself.¹⁸ A handful of people who can all be named in one complaint offer no such saving.

The standard is impracticability, not impossibility. Courts ask whether joining every member would be difficult or inconvenient, and they do not insist that joinder be out of the question.¹⁹ An impossibility standard would almost never be met, since any group of identifiable people can in theory be named as coplaintiffs. The trial judge has considerable discretion in making the call and is reversed only for abuse of that discretion.²⁰

No fixed number answers the question, and courts recognize no hard and fast minimum.²¹ They nonetheless reach for benchmarks. Joinder has been found impracticable with as few as 25 members, some courts presume numerosity at forty, and others place the line somewhere between twenty and forty.²² The benchmarks leak at both ends. Classes have been certified with fewer than twenty known members, while other courts have held more than 300 members insufficient.²³ Numbers below thirty may be insufficient, yet thirty or forty franchisees would be hard enough to join that, with the other prerequisites met, they would make an appropriate class.²⁴ The count is therefore evidence of how heavy the joinder burden will be, not the whole test.

Because the touchstone is the impracticability of joinder, not mere numbers, courts weigh other facts alongside size.²⁵ The facts that matter most often are these:

• A class scattered across many states is harder to join than one concentrated in a single place, so twenty members in fifteen states can present a harder joinder problem than two hundred members in one city.²⁶
• When each member’s claim is small, or members lack the resources to sue alone, they are unlikely to litigate at all unless a class is certified, which makes joinder without a class all but impossible. These “negative-value” suits favor a finding of numerosity.²⁷
• Members who cannot be named, because their identities are unknown or because the challenged practice will reach them later, cannot be joined, and they also add to the class’s size.²⁸
• Courts generally accept smaller numbers when the plaintiffs seek only injunctive relief, because the benefits of an injunction often run to people who are not formally in the suit.²⁹

Numerosity, like the other prerequisites, must be shown with evidence, but the evidence need not fix an exact count. Courts do not require proof of the precise size of the class or the identity of each member, though a speculative showing will not do.³⁰ That flexibility matters most where the defendant alone holds the records from which class size can be estimated.³¹ An estimate grounded in the defendant’s own records, such as roughly 12,000 customers drawn from data on a breach, is evidence of size; a guess is not. Data-breach litigation applies the standard this way. In re Hannaford Brothers Data Security Breach Litigation, the rule was that numerosity has no strict numeric test, an exact member count need not be proved, a speculative showing is not enough, and plaintiffs must also demonstrate that joinder is impracticable.³² At the other extreme, the plaintiffs in Edwards v. National Milk Producers Fed’n offered evidence of approximately 46 million class members, and the court found the class sufficiently numerous.³³

The opposite result follows when the group is small, known, and close together.

A group that small, known, and local gains nothing from a representative, because the joinder rules already let each member appear as a party in their own name.³⁴

Commonality

Rule 23(a)(2) requires “questions of law or fact common to the class.”³⁶ The requirement serves two purposes: it ensures that absent members are fairly and adequately represented, and it ensures practical and efficient case management.³⁷ Both purposes rest on one fact. When members’ claims share a core dispute, the representatives’ case is the absent members’ case, and one trial replaces many. When the claims share only a label, the representatives’ proof says little about anyone else, and the class device saves nothing while binding people who never appeared.

Courts once read the prerequisite generously. The Ninth Circuit in Rodriguez v. Hayes held that not every question need be common, and it accepted shared legal issues with divergent factual predicates, or a common core of salient facts coupled with disparate legal remedies, as enough.³⁸ The Rule’s plural wording does not demand more than one common question; as the Supreme Court later agreed, “even a single [common] question” will do.³⁹ The generous reading had a weakness, however: any competently drafted class complaint can recite common questions, so a test satisfied by reciting them screens out nothing.⁴⁰

Wal-Mart Stores, Inc. v. Dukes made the prerequisite bite. A class of approximately 1.5 million female employees alleged that the discretion their local supervisors exercised over pay and promotion discriminated against women in violation of Title VII.⁴¹ The Court held commonality unmet: because the respondents offered no convincing proof of a companywide discriminatory pay and promotion policy, they established no common question.⁴²

The Court’s test turns on answers, not questions. Reciting questions, such as whether the plaintiffs all work for Wal-Mart, whether their managers have discretion over pay, and whether that is unlawful, does not obtain certification; commonality requires the plaintiffs to show that the members “have suffered the same injury.”⁴³ Their claims must depend on a common contention, such as the assertion of discriminatory bias by the same supervisor, and the contention must be capable of classwide resolution. That means determining its truth or falsity will resolve an issue central to the validity of each claim in one stroke.⁴⁴ What matters is “not the raising of common ‘questions’—even in droves—but, rather the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation,” and dissimilarities within the class are what can impede those answers.⁴⁵ A common question, in short, is a dispute of fact or law whose resolution will advance the determination of the class members’ claims.⁴⁶

The shift follows from what a class judgment does. It binds absent members and gives the defendant one trial in place of many, which is fair and efficient only if the court’s answer is an answer that every member’s claim needed. If the claims turn on separate reasons, a classwide trial merely bundles individual suits under one caption. A shared statute or a list of shared questions therefore cannot carry the burden.

The fourth step explains why commonality cannot always be separated from the merits. The crux of an individual Title VII claim is “the reason for a particular employment decision,” so a class suing over millions of decisions at once must supply some glue holding the reasons together before a classwide answer to the question why each member was disfavored can exist.⁴⁸ In Dukes, proof of commonality therefore overlapped with the plaintiffs’ merits contention that Wal-Mart engaged in a pattern or practice of discrimination, and the Court examined it for that reason alone.

General Telephone Co. of Southwest v. Falcon had earlier identified two ways to bridge the gap between one employee’s claim and a class of people injured in the same way. The first is proof that the employer used a biased testing procedure to evaluate applicants and incumbent employees alike, because one finding about the test answers a central question for everyone it touched. The second is significant proof that the employer operated under a general policy of discrimination that showed up in hiring and promotion in the same general fashion, such as through entirely subjective decisionmaking.⁴⁹ Dukes held that the plaintiffs could use neither route. Wal-Mart had no testing procedure or other companywide evaluation method that could be charged with bias, and the whole point of permitting discretionary decisionmaking is to avoid evaluating employees under a common standard.⁵⁰

The plaintiffs’ remaining evidence fell short of the second route. Their sociologist could say that Wal-Mart’s corporate culture made it vulnerable to gender bias, but he could not estimate what percent of employment decisions stereotypical thinking determined, so his testimony was “worlds away” from significant proof that Wal-Mart operated under a general policy of discrimination.⁵¹ The anecdotes did no better: set beside the accounts in Teamsters v. United States, they were sparse relative to the size of the class, and the Court quoted a dissenting description of the class members as people who “have little in common but their sex and this lawsuit.”⁵² In a company of Wal-Mart’s size, it is unlikely that all managers would exercise their discretion in a common way without some common direction.⁵³ Managers may decide for various reasons, some discriminatory and some not, so a pattern of discretionary decisions does not form a pattern or practice.⁵⁴

The first company’s case matches the Court’s own example of a common contention, which was bias on the part of the same supervisor.⁵⁵ Some lower courts have accordingly allowed classes to proceed where a small group of high-level personnel exercised discretion to produce disparate results, in contrast to the much larger number of low-level managers in Dukes.⁵⁶

Dissenting Justices objected that the Court “disqualifies the class at the starting gate” by importing into Rule 23(a) concerns properly addressed in the Rule 23(b)(3) assessment.⁵⁷ Their sharper charge was structural. The majority’s emphasis on differences between class members mimics the predominance inquiry, and asking whether those differences “impede” common adjudication duplicates the Rule’s question whether a class action is superior to other modes of adjudication; the “dissimilarities” inquiry the majority endorsed was developed in the context of Rule 23(b)(3).⁵⁸

The majority replied that it considered dissimilarities not to decide, as Rule 23(b)(3) requires, whether common questions predominate, but to decide, as Rule 23(a)(2) requires, whether there is even a single common question, and there was none.⁵⁹ Both sides agree that one common question is enough. They disagree about how far differences among members may be weighed at this threshold. For exam purposes, state the majority’s test as the governing rule and remember that dissimilarities matter under (a)(2) only to the extent they prevent a common answer. A majority of federal courts apply the heightened standard to classes under (b)(1)(A), (b)(1)(B), (b)(2), and (b)(3) alike, although classes continue to be certified after Dukes.⁶⁰

A challenge to a uniform policy shows the other pole. In Floyd v. City of New York, the plaintiffs challenged department-wide stop-and-frisk policies and practices, and their questions concerned that policy: either the department maintained it as described or it did not, and either answer applies to every class member.⁶¹ No answer to those questions explains why a particular person was stopped, which is the individualized question that sank the Dukes class. The difference lies in what the claim attacks and what it seeks. The Floyd plaintiffs sought injunctive relief against the policies, not individualized backpay or a personal injunction, so a common answer about the policies is exactly what their claims need.⁶²

Judge Posner drew a similar line in McReynolds v. Merrill Lynch, Pierce, Fenner & Smith, Inc. He concluded that even after Dukes, Rule 23(b)(2) suits remain appropriate for obtaining injunctive relief where a centralized policy is alleged to affect a large class, even when the magnitude and existence of the impact vary by class member.⁶³ Commonality looks at the group and its shared contention. The next prerequisite, typicality, turns to the representative’s own claim.

Typicality

Rule 23(a)(3) requires that “the claims or defenses of the representative parties are typical of the claims or defenses of the class.”⁶⁵ Commonality looks at the group and asks whether its members share a contention that one proceeding can answer. Typicality looks at the named plaintiff and asks whether that person’s claim resembles the claims of the people represented. The prerequisites do different work, because members whose claims are vastly dissimilar can still share a common question of law or fact, and typicality assures that the representatives will “feel the pain” of the class members.⁶⁶ A class can therefore clear commonality and still fail here if the person chosen to lead it was injured in a different way from everyone else.

The usual formulation asks whether the representative’s claim arises from the same course of events as the class members’ claims and whether the representative would make similar legal arguments to prove the defendant’s liability.⁶⁷ The comparison concerns the nature of the claim rather than the specific facts from which it arose, and it is satisfied when the representative has the incentive to prove every element of the cause of action on behalf of the members.⁶⁸ Courts that compare the representative’s claim with the others nearly always hold that the claims need not be identical, only substantially similar, and that factual differences do not destroy typicality unless they cast serious doubt on the representative’s ability to prosecute the other members’ claims.⁶⁹ Injuries that derive from a unitary course of conduct by a single system can be typical even when their precise nature varies.⁷⁰

The Rule asks for alignment, not identity, for a practical reason. A representative who suffers the same kind of injury from the same conduct has every incentive to prove the common case, so pursuing the individual claim advances the class’s claims too. When the representative’s claim differs in kind, or faces a defense that applies to that person alone, the representative’s effort and the court’s attention drift toward issues the absent members do not share. Because every individual has some different facts, a demand for identity would disqualify nearly every representative, so the inquiry stays undemanding and aims to ensure that the class has a titular head with an actual grievance against the defendant.⁷¹ Within that low bar, the questions that matter are about conflict: whether anything significant differentiates the representative’s claim to a degree that could set the representative’s interests against the class’s.⁷²

General Telephone Co. of Southwest v. Falcon shows the comparison failing. Relying on typicality, the Court raised serious doubts about “across-the-board” class actions, in which a plaintiff claiming discrimination in one kind of employment decision, such as hiring, seeks to represent employees who claim discrimination in promotion, termination, or pay.⁷⁴ The injuries differ in kind, because a hiring decision and a promotion decision are different decisions, and proof that one was discriminatory says little about the other. The Court indicated how the groups could be joined: through a biased test used to evaluate both applicants and incumbent employees, or through significant proof that a general policy of discrimination showed up in hiring and promotion in the same general fashion, such as through entirely subjective decisionmaking.⁷⁵

Differences in the extent of harm stand on the other side of the line. Class members “need not share identical claims” so long as the same alleged conduct affects them, so variation in how much each member was hurt by one course of conduct leaves typicality intact.⁷⁶ Falcon isolates the variable. There, the difference between representative and class went to what the defendant did to each person, so winning the representative’s claim would have left the class’s claims unproved. Where one course of conduct injures everyone, differences in the degree of harm do not matter, because the representative’s proof of the conduct is the class’s proof.

The same logic governs defenses. A lack of typicality may be found where the named plaintiff “was not harmed by the [conduct] he alleges to have injured the class” or where the named plaintiff’s claim is subject to “specific factual defenses” atypical of the class.⁷⁷ A defense available against the whole class, or a difference that varies only in detail, leaves the alignment of interests intact. Membership in the class is part of the comparison too. In Floyd v. City of New York, the defendants argued that the named plaintiffs could not represent Latinos, and the court rejected the argument because the cases the defendants cited denied certification where the named plaintiffs fell outside the subclass they sought to represent, while the Floyd plaintiffs fell squarely inside the Fourteenth Amendment subclass of Blacks and Latinos stopped because of their race.⁷⁸

A defense matters to typicality when it is peculiar to the representative, because only then does it pull the representative’s litigation effort away from the class’s common case.

Typicality overlaps its neighbors, and courts and commentators describe its independent work differently. One account treats typicality as largely ensured by the required commonality and adequacy of representation,⁷⁹ yet courts that compare the representative’s claim with the class’s give the requirement a role of its own.⁸⁰ An exam answer should therefore run the comparison separately and then note where the same facts bear on adequacy, the next prerequisite, since a representative with a unique defense or a diverging interest raises a question under both.

Adequate Representation

Rule 23(a)(4) requires that “the representative parties will fairly and adequately protect the interests of the class.”⁸² Typicality asks whether the representative’s claim resembles the class’s claims. Adequacy asks whether the representative and class counsel can be trusted to pursue those claims for people who never appeared in the case. The question carries the weight of the class device because the results of a class action generally bind absent members.⁸³ Rule 23(a)(4) rests on due process: a ruling cannot bind absent class members if their representatives were inadequate, as Hansberry v. Lee held. The requirement reaches both the class representatives and class counsel, but it demands adequate representatives and counsel, not the best possible ones.⁸⁴

Three conditions follow from that due process principle. The absent members’ interests must be of the same class as the representatives’ interests, meaning that they share common interests. The absentees’ interests must in fact be adequately represented by the parties present. And courts must adopt procedures that insure those conditions are met and that the litigation is conducted so as to insure full and fair consideration of the common issue.⁸⁵ Courts apply the principle through two questions: whether the representative’s interests conflict with the class members’ interests, and whether the class has an attorney capable of prosecuting the action vigorously. A representative who satisfies typicality generally satisfies the first question, and counsel who are qualified, experienced, and capable satisfy the second.⁸⁶ In Gonzales v. Cassidy, the court asked whether the representative had interests in common with the unnamed members and whether he would vigorously prosecute the class’s interests through qualified counsel. It found both met, pointing to counsel’s successful appeal to the Supreme Court after an adverse decision below, and it therefore declined to deny res judicata effect to the earlier class judgment.⁸⁷

The conflict question produces the hardest adequacy cases. Amchem Products, Inc. v. Windsor shows the structural problem. The settlement there bundled diverse groups of asbestos-exposed persons together, and each named party served generally as representative for the whole, not for a separate constituency. The Court found that the settling parties had reached a global compromise with “no structural assurance of fair and adequate representation for the diverse groups and individuals affected.”⁸⁸ The diversity mattered because, as the Third Circuit observed, the settlement did more than create a fund; it made “important judgments on how recovery is to be allocated among different kinds of plaintiffs, decisions that necessarily favor some claimants over others.”⁸⁹ A fund is finite and its terms divide it, so a representative who bargains hard for one group’s priorities gives up the other group’s, and one representative cannot be loyal to both sides of that trade.

The Court reached a similar result in the limited-fund settlement in Ortiz v. Fibreboard Corp. It found disparate interests within the certified class, among them that claimants whose claims arose before 1959 held more valuable claims than later claimants, because the insurance policy supplying most of the settlement money expired that year. These conflicts were “well within the requirement of structural protection recognized in Amchem,” although “at some point there must be an end to reclassification with separate counsel.”⁹⁰ The proponents argued that a later fairness hearing would protect the class. The Court answered that Rule 23 requires its protections against inequity at the precertification stage, independently of the postcertification fairness review under subdivision (e).⁹¹ The point holds for every settlement class: the requirements that protect absentees by blocking unwarranted or overbroad class definitions “demand undiluted, even heightened, attention in the settlement context.”⁹²

Not every difference among members is a disabling conflict. Divergent wishes alone cannot defeat adequacy, because in almost any class some members would prefer not to assert their rights. Stockholder derivative suits, brought by minority stockholders to challenge action that a majority approves, and taxpayer challenges to proposed public spending routinely proceed as class suits.⁹³ The practical question is whether the divergence affects how the representative would handle the claims at issue. Most courts require serious deficiencies before they find a representative inadequate.⁹⁴ When a case has several representatives, the prerequisite is satisfied if even one is adequate, although a representative who fails the standard cannot continue in that role, and the quality of representation counts for more than the number of representatives.⁹⁵

Where subgroups really do diverge, the structural cure is to separate them. When appropriate, a class may be divided into subclasses, each treated as a class under the Rule.⁹⁶ Because each subclass must satisfy the prerequisites itself, each needs its own representatives, and, as the reference in Ortiz to “reclassification with separate counsel” indicates, its own counsel.⁹⁷ The mechanics of forming subclasses return below with the certification order.

Conflicts can also arise from the representative’s own situation.

Personal circumstances can matter even when the claim is typical. Adequacy of the representatives often overlaps with typicality, which is why Floyd treats them together, but not always, and a court may ask whether a terminally ill plaintiff can adequately represent the class.⁹⁸

The bar for finding inadequacy stays high, so a court weighs such facts as bearing on the representative’s ability to pursue the class’s claims and does not treat every personal difference as disqualifying.

The second question concerns counsel, and it matters more than the text of Rule 23(a)(4) suggests. When individual stakes are small and members have no pre-existing relation to one another, the lawyer is the real representative: the lawyer assembles the claim, funds the investigation, and makes the decisions that matter, while members are rationally indifferent to the day-to-day course of the case.⁹⁹ A named plaintiff who nominally controls the case therefore offers the class little protection unless the lawyer is competent and loyal. Unless a statute provides otherwise, a court that certifies a class must appoint class counsel, and in choosing counsel the court looks to the lawyers’ preparation, experience, knowledge of the law, and resources.¹⁰⁰ Once appointed, class counsel must fairly and adequately represent the interests of the class.¹⁰¹ The lawyer’s own interests matter as well, since conflicts between class members and their attorneys are not uncommon.¹⁰²

The procedure around adequacy has features that exam answers should reflect. The defendant has its own stake in the question: if the defendant wins a judgment against inadequate representatives, absent members will have a strong argument that they are not bound, and the victory can be attacked.¹⁰³ The court’s first finding does not end the matter. The court has a continuing duty to monitor the adequacy of representatives and counsel, and Rule 23(d)(1)(B)(iii) lets it give notice inviting class members to signify whether they consider the representation fair and adequate.¹⁰⁴