On another level, however, class definitions can make or break a motion for class certification. Sometimes they do so because, when examined critically, they reveal why a proposed class cannot satisfy the certification requirements of Rule 23. One test of whether a plaintiff can meet his or her burden under Rule 23 is whether a problem with one class definition can be fixed by substituting a different class definition in the same case. If not, there may be an underlying issue with one or more of the Rule 23 requirements. In such a case, the class definition is not, as some would claim, being wrongfully exploited by those who would limit the availability of the class action device. Rather, it is a lens that can bring into focus the flaws that make the case unsuitable for class treatment no matter how the class is defined.
For example, a plaintiff in a wage-and-hour case might propose the following class definition:
All persons who worked as hourly employees of XYZ Corporation from 2008 to the present.
The beauty of this proposed definition lies in its simplicity. In most cases, membership in the class can readily be ascertained from the defendant’s business records. Those records will permit the preparation of a class list, complete with names and current and last-known addresses, which a notice administrator can use in mailing notice to all class members. By eliminating guesswork and uncertainty, the proposed definition will result in ease of administration, one facet of establishing that a class action is manageable and, hence, superior.
But when one examines the class definition in the light of the claims being advanced, the cracks in the picture emerge. A typical case might allege that XYZ Corporation forces its employees to work through their meal periods and rest breaks. Often, the plaintiffs will not claim, and would not be able to prove, that every employee has been denied his or her breaks. Part of the company’s defense may be that, while there appear to be instances of individual employees being forced to skip their breaks, there are also many employees who received all of their breaks or, if they missed them, chose to skip them for any of a variety of personal reasons (such as wanting to leave work early). If there is evidence that some employees were forced to skip breaks, while other employees received all of their breaks, and still others missed breaks only by their own choosing, the class of all employees is overly broad: It includes people who have claims against XYZ and others who do not.
Courts have held that an overly broad class cannot be certified. See, e.g., Oshana v. Coca-Cola Co., 472 F.3d 506 (7th Cir. 2006), cert. denied, 551 U.S. 1115 (2007). There are both legal and practical reasons for this rule. Legally, when a class includes members who have not been injured by the defendant’s alleged conduct and therefore have no claim, there is a problem with standing. In federal court, this problem may be considered to be of constitutional proportions—there is no case or controversy for those class members for purposes of Article III standing. State courts may recognize the same problem, although not as a constitutional issue. Some courts deal with this concern by trying to gauge the magnitude of the overbreadth issue; if only a small number or percentage of class members are uninjured, they permit the class action to go forward even with the overly broad definition. See, e.g., Messner v. Northshore HealthSystem, 669 F.3d 802 (7th Cir. 2012); Aspinall v. Philip Morris Cos., Inc., 442 Mass. 381, 398 n. 21 (2004). But such an approach may be too forgiving, and it will not solve the overbreadth problem in all cases.
On a practical level, an overbroad class definition may lead to confusion among the members of the proposed class. A person who receives notice of the class action but who does not have a claim may not understand why he or she received the notice, why he or she is being included in the case, or what his or her rights are as a member of the class. If the case results in a plaintiffs’ verdict, are the uninjured class members entitled to relief? And if the case settles, will the uninjured class members participate in the settlement? If uninjured class members are allowed to claim money from a common fund, how does the resulting dilution of the rights of the truly injured class members affect the fairness of the settlement and the adequacy of the class representatives and their counsel to represent the injured people? Perhaps of greatest concern is the possible unfairness of holding that persons who should not have been included in the class to begin with are bound by the judgment in the case.
The obvious solution to these problems is to define the class more narrowly to exclude the uninjured class members. So, again using the wage-and-hour case as an example, plaintiffs’ counsel might posit the following definition:
All persons who worked as hourly employees of XYZ Corporation from 2008 to the present and who were forced to work through their meal periods and rest breaks.
This definition would seem to align membership in the class perfectly with the claims being advanced. There is no standing issue because only people whose rights were being denied are included. Theoretically, there also is no risk of confusing uninjured people by sending them notices of their membership in the class action, and no risk of diluting the rights of injured people by giving their uninjured counterparts a claim to a common fund.
So what’s the problem? Unless the defendant had a uniform corporate policy or practice of forcing all of its workers to miss their breaks (which also might have avoided the overbreadth concern), the problem is figuring out who meets the definition for class membership. The class definition raises at least two questions: Who missed their meal periods and rest breaks? And of that group, who missed them because they were forced?
There are potentially at least two ways to go about answering those questions. The first, and the one most frequently proffered by plaintiffs’ counsel, is to rely on the company’s own business records. After all, employers are required to maintain accurate timekeeping records. But those timekeeping records may not accurately show missed breaks and rarely will show whether a break was missed because the employee was forced to skip it. There are nuances specific to employment cases that are illustrative of the types of issues that can arise. For example, rest breaks usually are paid breaks. Employers may not be required to keep records of rest breaks because they do not affect pay, and some employees may not bother to clock out for rest breaks for the same reason. Despite the expectation that timekeeping records will be accurate, they may be especially inaccurate when it comes to determining who took or did not take their rest breaks. Plaintiffs will often counter these arguments by citing case law that holds that, because it is the employer’s obligation to maintain accurate timekeeping records, there is a presumption that the records are accurate that overcomes the defendant’s protestations to the contrary. But the employer will counter that the presumption is rebuttable and that regardless of the accuracy of the records to show whether breaks were missed, they will not be able to establish whether the employees were forced to miss them.
The second way of establishing who belongs to this more narrow class is simply to ask each employee whether he or she was forced to miss a break. Such an undertaking is, by definition, individualized. Not only would it require an individualized inquiry of each employee, but for each employee who claims to have been forced to miss a break, the employer also would be entitled to test the employee’s claim by interviewing his or her managers and coworkers. This approach leads ineluctably to the conclusion that common issues do not predominate over individual issues and that the class cannot be certified.
The types of problems caused by this narrowed class definition are referred to variously as problems of ascertainability, a merits-based class, and a fail-safe class. Most courts recognize ascertainability as an additional, implied requirement of Rule 23. See, e.g., Marcus v. BMW of N. Am., LLC, 687 F.3d 583, 592–93 (3d Cir. 2012); Romberio v. Unumprovident Corp., 385 F. App’x 423, 431 (6th Cir. 2009); In re IPO Secs. Litig., 471 F.3d 24, 44–45 (2d Cir. 2006). Courts require that membership in a class be capable of determination based on objective factors, without the need for numerous individualized determinations or mini-trials. In our example, determining which employees are members of the more narrowly defined class requires extensive individualized inquiries, which render class certification inappropriate.
In BMW, the Third Circuit recently explained the rationale behind the requirement:
The ascertainability requirement serves several important objectives. First, it eliminates “serious administrative burdens that are incongruous with the efficiencies expected in a class action” by insisting on the easy identification of class members. . . . Second, it protects absent class members by facilitating the “best notice practicable” under Rule 23(c)(2) in a Rule 23(b)(3) action. . . . Third, it protects defendants by ensuring that those persons who will be bound by the final judgment are clearly identifiable.
687 F.3d at 593 (citations omitted); see also Northside Chiropractic, Inc. v. Yellowbook, Inc., 2012 WL 3777010 at *3–5 (N. D. Ill. 2012) (describing purposes of ascertainability requirement as alerting parties and the court to potential manageability problem in identifying members of the class, ensuring that only parties actually harmed will receive any awarded relief, and avoiding problem of tying class definition to a future decision on the merits).
A merits-based class is a class whose membership cannot be determined without conducting individualized inquiries concerning the merits of each potential class member’s claim. In our example, it is impossible to know who is in the narrowed class without determining the merits of each employee’s claim that he or she was forced to miss breaks. A merits-based class is sometimes referred to as a fail-safe class because, if the plaintiff wins on the merits, he or she wins a class-wide judgment, but if the plaintiff loses, there are no class members, and the adverse judgment is binding only on the plaintiff, leaving the defendant exposed to the possibility of additional class actions filed on behalf of other plaintiffs. See Messner, 669 F.3d at 825 (fail-safe class is “improper because a class member either wins or, by virtue of losing, is defined out of the class and is therefore not bound by the judgment”); Bolden v. Walsh Constr. Co., 688 F.3d 893, 895 (7th Cir. 2012) (same); Lindsay Transmission, LLC v. Office Depot, Inc., 2013 WL 275568 (E.D. Mo. 2013).
Whether the issue is described as one of ascertainability or of having a merits-based or fail-safe class, the underlying problem is one of Rule 23(b)(3) predominance and superiority, and, since Wal-Mart Stores, Inc. v. Dukes, 131 S. Ct. 2541 (2011), Rule 23(a)(2) commonality. Because membership in the class can be determined only through individualized inquiry requiring mini-trials, individual issues predominate and the case is unmanageable.
These definitional problems arise not only in wage-and-hour cases but in any case where not all persons who were employed by the defendant, who purchased the defendant’s product, or who were exposed to the defendant’s allegedly unlawful behavior were caused harm, and where there is no common way to separate the harmed from the unharmed. For example, in a consumer fraud case, a plaintiff might seek certification of all purchasers of a defendant’s product, alleging that the product was defective. If only some units of the product were defective, if the product failed to perform properly only in certain circumstances that did not apply to all purchasers, or if only some purchasers relied on the defendant’s allegedly false advertising when deciding to buy the product, then a class of all purchasers is overbroad. A more narrowly constructed class definition—for example, all persons who relied on the manufacturer’s allegedly misleading representations in deciding to purchase the product—would avoid the problem of overbreadth, but it also would require individualized inquiries on the merits of each class member’s claim to ascertain class membership.
What should be apparent in these examples is that the class definition is not the disease; rather, it is the symptom of an underlying condition that cannot be cured by redrafting. The condition is that different people were affected differently by their interactions with the defendant, and the only way to determine who was harmed is to examine the circumstances of each potential class member. This underlying problem renders futile any effort to define a class that warrants certification, and the inability to come up with a viable class definition signals the existence of the fundamental flaw in the case. It is the clue that reveals why the class action should not proceed, the window into the viability of the case as a class action, and the first place that counsel should look to develop certification strategy.
Of course, not all cases suffer from these maladies, and many class definitions can be applied in a straightforward way. If, for example, the employer did have a uniform policy that it would not allow any of its hourly workers to take their earned breaks, a class defined as all hourly employees might well be appropriate. Or if every unit of a product sold was defective in a manner that resulted in legally cognizable harm to every person who bought it, then a consumer class might properly include all purchasers. The class definition points to a certification issue only where a certification issue actually exists.
Nor are definitional issues even in problematic cases always irresolvable. The obvious solution for many of them is settlement. One reason that an overly broad class definition is troublesome is that a defendant should not be held answerable to people whom it has not harmed. Holding a defendant liable to a class that includes uninjured members, without allowing it to prove its non-liability to that subgroup, would deprive the defendant of its right to due process. Further, inclusion of uninjured people in a class that recovers a judgment would violate the Rules Enabling Act, because the application of a procedural rule (Rule 23) would have the effect of enlarging the rights of people who would be unable to maintain a claim if they were to sue solely on their own behalf.
When a defendant settles, however, it undertakes to give up these defenses in order to achieve some other goal, such as eliminating uncertainty and buying peace. Although defendants may not be forced to give up their due process or statutory rights, they should be allowed to do so voluntarily by settling with a class that has been defined broadly, so long as such a settlement does not unfairly prejudice the rights of absent class members. Alternatively, plaintiffs and defendants might choose to resolve their dispute by constructing a settlement that will include a mechanism for identifying injured persons entitled to participate in the settlement, such as through an individual claims process. The settlement may include provisions for the defendant to dispute submitted claims, but typically these provisions are less rigorous than applying burdens of proof and rules of evidence to a litigated claim.
Just as in a settlement with a broadly defined class, the defendant may, as a matter of business judgment, choose to settle with a narrowly defined class by relaxing its insistence that each individual claim be proved as it would have to be in a court of law. Although there are limitations on parties’ abilities to use the device of settlement to overcome objections to class treatment (see Amchem Prods., Inc. v. Windsor, 521 U.S. 591 (1997)), courts should be hesitant to reject consensual resolutions of class actions that are designed to provide a benefit to all class members, particularly in the absence of any objections or apparent unfairness.
Concerns about the class definition do not necessarily end when the parties agree to settle. If the settlement is reached after the court had certified a Rule 23(b)(3) class, for example, any changes in the class definition may require special notice to those affected by the change, particularly if the class is narrowed so that the settlement class excludes people who were members of the litigation class. Such a change might also create a conflict for the class representatives and class counsel, potentially resulting in an adequacy challenge or other objections to the settlement. Alternatively, a change in the class definition before any class has been certified may be easier to accomplish, but care still must be taken to ensure that the definition captures everyone who should be entitled to a share of the settlement and everyone whom the defendant expects to be bound.
Whether the case ends through litigation or settlement, getting the class definition right is critical to the proponent’s success. It is no wonder, then, that savvy plaintiffs’ counsel will take care in its crafting and that defense counsel will target vulnerable class definitions in their oppositions to class certification. Although in many cases faulty class definitions can be cured, in many others they are emblematic of fundamental reasons why the claims being advanced are not suitable for class treatment. The rigorous scrutiny required in evaluating motions for class certification, and the undesirable consequences that result when a bad class definition is approved, demand that courts maintain their vigilance when deciding whether a class has been, or indeed can be, properly defined.
Keywords: litigation, class action, derivative suits, ascertainability, class certification, merits-based class, fail-safe class, Federal Rule of Civil Procedure 23