Tiered Certification

Introduction

Class actions serve an important regulatory tool, enabling victims to vindicate rights that are otherwise not worth pursuing. However, the magnified stakes also invite abuse. The success and potential failure of class actions stem from their unique design. In their classic form, class actions enable one representative, in exchange for monetary incentives, to enlist the power of the class and litigate on its behalf, binding its members although they are not physically in court. The heart of class actions is the so-called certification stage — at which courts decide which cases are worthy of class adjudication.

The idea of “certification,” however, is not self-evident. Modern class actions are a relatively new phenomenon, and the concept of certification is even newer. The original, 1966 version of the relevant American Rule 23 did not refer to “certification” of a class.^[1] While courts did use the term, it formally appears in the text of Rule 23 only after the 1998 and the 2003 amendments.^[2] Currently, the “certification” terminology appears pervasively in the text of Rule 23.^[3] Simultaneously, the certification process has become more rigorous,^[4] constituting a sharper border between class actions and regular litigation.

This Article challenges the rigid concept of “certification.” Instead, there may be good reasons to “smooth” out the process of certification, in order to break the binarity of the current rules.^[5] Along these lines, this Article suggests a thought experiment — a split certification process, which I refer to as “tiered certification” — and shows how this proposal can open up new opportunities to use class actions as a regulatory vehicle.

I focus on the potential of tiered certification to solve the problem of serial certification attempts. According to conventional perceptions, only a certified class precludes class members’ rights; the upshot is that a denial of class certification does not prevent class members from subsequent certification attempts. The result is undesirable relitigation. However, this outcome is an artifact of the binary view of class certification. Tiered certification, on the other hand, provides a different, non-binary regime. First, the class is preliminarily certified, under a relaxed standard, at the outset of the litigation, with scaled-back notice and opportunity to opt-out. If the class survives the preliminary certification stage, it proceeds to the second, more rigorous certification phase. A class that has passed both certification stages would bind its members, preventing them from litigating their claims; a class that passed the first but failed the second stage would only preclude the members’ procedural right to litigate as a collective — leaving them free to litigate their individual claims.^[6] In essence, the tiered-certification proposal attempts to bar costly relitigation of denials of (full) certification by maintaining minimal procedural safeguards — preliminary certification, scaled-back notice, and early opportunity to opt-out. As a by-product, the early inspection of these procedural safeguards should improve representation from the inception of a case.

Flattening the certification process, then, allows a richer set of solutions to the problem of multiple certification attempts. More generally, a smoother process opens up new directions of handling class litigation. Think of class litigation that entails some, but meager societal value — for instance, a claim that the defendant does not comply with a trivial regulatory rule and/or that the members of the class suffered only nominal damages. The onerous certification process is too heavy a tool for litigating such claims. Moreover, the binary concept of “certification” enables either class- or individual-litigation. The tiered-certification proposal, by contrast, anticipates a new category: cases that pass the preliminary certification but fail the second. This procedural design can fit class actions that entail a modest social value — semi class-actions. I briefly reflect on the application of this idea to injunction-only and statutory-damages class actions — the defendant could cease its wrongful behavior while the members of the class retain their individual (mostly trivial) damages claims. True, vindication of individual rights would be difficult without a fullblown class action; however, in many cases comprehensive class treatment seems unwarranted. A richer concept of certification, then, makes it possible to break the current one-size-fits-all treatment of class litigation.

Before proceeding, two introductory notes are in order. First, this Article focuses on the American law of class actions, which is based on Rule 23 of the Federal Rules of Civil Procedure. Additional examples are taken from the Israeli regime, which is based on the Israeli Class Action Law of 2006 (Class Action Law). The discussion centers in particular on “classic” class actions, which allow class members the right to opt-out of the collective proceedings.^[7]However, the Article intends to be more general, with reference to any legal system that uses a designated procedure, “certification,” to distinguish between class actions and individual litigation.^[8] Second, the tiered-certification proposal can be conceived of as a thought-experiment, but the changes it advances are, from another perspective, actual, even old-fashioned. The proposal is timely because the threshold for certification has been raised in recent years, shifting efforts to early stages of litigation and amplifying the importance of adequate representation early on. However, the law has not been modified to ensure, at the outset, that the representatives are adequate; tiered certification provides such a tool, through preliminary certification. Relatedly, the proposal in fact echoes previous practices. As aforementioned, a single “certification” is not obvious; indeed, in the past judges in the United States have experimented with conditional certification — a preliminary approval of class litigation, under a more lenient standard. Furthermore, originally, the federal rules envisioned a relatively low certification threshold, early in the proceedings.^[9] Certification has become more fact-intensive, costly, and complex.^[10] The idea of preliminary certification restores, to some extent, the original intention of Rule 23. It makes (early) certification simple again.

The Article proceeds as follows. Part I lays out the basic requirements for certifying a class, the trend to raise those standards and to shift litigation efforts to early phases, and the doctrinal importance of certification. Part II sketches the tiered-certification proposal — a preliminary certification on a lenient standard, followed by a complete, rigorous certification, where each stage entails different repercussions. Part III discusses the relitigation problem, i.e., the absence of a preclusive effect of certification denials in the first forum. It analyzes the relevant policy considerations, and surveys the main directions that courts, scholars, and policymakers have offered for tackling the problem. It then offers tiered certification as a superior solution. Part IV suggests extending tiered certification to address cases that carry small social value, which falls short of complete, rigorous certification. The last Part concludes.

I Class Certification and Its Effects

Certification is a crucial stage in class litigation. Unlike individual litigation, class proceedings require a preliminary step, in which the court has to “certify” a lawsuit as a class action. The certification requirements are an artifact of the well-known difficulties in conducting class proceedings in general courts. On the one hand, class litigation can fulfill an important role in facilitating claims that are otherwise not worth bringing. On the other hand, class litigation enables a representative — the named plaintiff — to preclude the individual rights of the class of victims, opening up a host of agency problems. Moreover, as class actions typically magnify the monetary stakes, they can be abused to threaten defendants and extract undesired settlements. With these dangers in mind, certification constitutes the gatekeeper to the world of class litigation — or, a minimum threshold to trigger the unconventional weapon of class actions. Indeed, the idea of an intermediate “certification” stage in class actions now seems axiomatic, at least in the United States, Israel, and other legal systems that draw on the American experience.^[11] This Part briefly presents the basic certification requirements and their constitutive role.

Broadly, one can identify three types of requirements for plaintiffs to pass the certification threshold — those that relate to adequate representation; the fit between the case to class-wide treatment; and minimal evidentiary limits.

First, to minimize agency problems between the class and its representatives, legal systems seek to guarantee that “the representative parties will fairly and adequately protect the interests of the class.”^[12] As a certified class binds its members, who are not in court, “adequacy-of-representation” is “the cornerstone of class action litigation.”^[13] While the capacity of courts to prevent representatives from sacrificing the interests of class members seems limited,^[14] courts and policymakers have devised several mechanisms to ensure adequate representation of class members’ interests.^[15]

Second, at the certification stage courts ask whether, given the inherent difficulties, it is appropriate to adjudicate the case on a class-wide basis. Presumably, class-wide proceedings are futile if there are no “questions of law or fact common to the class.”^[16] Likewise, class litigation need be “superior to other available methods for … adjudicating the controversy.”^[17] Concerns regarding the excessive scope of class actions have pushed courts to make these inquiries more demanding.^[18] To illustrate, a few years ago the U.S. Supreme Court refused to authorize class litigation on behalf of the female employees of Wal-Mart, holding that these women do not have enough in common to join together in a single suit.^[19]

Third, and related to the two previous sets of conditions, legal systems also require the class to present at the preliminary certification stage a minimal factual threshold. The Israeli Act specifies that at the certification stage the court should find that there is a “reasonable probability” that the class will win.^[20] While the federal rules do not contain an explicit factual threshold, it is clear that American courts have required a higher standard in the recent years. Judges are now directed to “rigorously” consider and resolve issues of fact necessary to determine certification, even if the very same issues would be decided on the merits, after certification.^[21] As a result, litigation efforts, including experts’ opinions, have shifted to the pre-certification stage — “frontloading.”^[22] Israeli courts have similarly mandated more onerous certification process.^[23]

In general, then, one can observe a trend to raise certification standards. This trend has an important impact on class actions. Arthur Miller summarizes: “The certification process has become so arduous that its cost and delay — coupled with the risk of eventual failure — either deter the institution of potentially meritorious class actions or lower their settlement value. Obviously, these developments represent significant inhibitions for even the strong willed.”^[24]

Frontloading burdens plaintiffs and their lawyers,^[25] potentially chilling valuable class actions. The wisdom of the doctrinal trend to raise the certification bar is beyond the scope of this Article.^[26] It should be noted, though, that the proponents of a more “arduous” certification process stress that an erroneous certification decision entails grave consequences. Defendants who lose on certification often face some chance of paying enormous damages, and thus they are allegedly “under intense pressure to settle,”^[27] even anemic cases. In addition to pressing defendants to settle, certification also precludes the rights of absent class members, as discussed below. In short, the heightened certification standards can be seen as a natural response to the growing importance of class litigation and the belief that past courts certified it with “insufficient rigor.”^[28]

To the extent that its requirements are met, the very act of certification is constitutive of class members’ rights. Upon certification, the court should define the class^[29] and send a notice to its members. The notice advises class members regarding their opportunity to appear in court and opt-out of the proceedings.^[30] These are not mere formalities. Class actions preclude the individual rights of absent class members, infringing “a principle of general application in Anglo-American jurisprudence that one is not bound by a judgment … in a litigation in which … he has not been made a party.”^[31] At least doctrinally, the deviation from preclusion principles relies on the notice that the court sends to class members and their opportunity to opt-out without being bound by the decision. As the U.S. Supreme Court explained, minimal procedural due process safeguards are necessary. In order to bind absent class members, they “must receive notice plus an opportunity to be heard and participate in the litigation.”^[32] These conditions can roughly be summarized as follows: Notice + Opportunity to opt-out + Certification = Preclusion of individual rights.

These are well-known concepts. They guarantee that without a proper certification procedure class members will not lose the right to litigate their individual claims. But as the Introduction showed, the axiomatic idea of a single “certification” stage is not self-evident. To demonstrate an alternative, the following Part briefly sketches a proposal for a two-stage certification process, which, I argue, better tackles familiar class action dilemmas.

II An Alternative — Tiered Certification

This Part presents the contours of the tiered certification proposal, and briefly discusses its advantages and disadvantages. A tiered process could rely on two “certification” stages. The first is preliminary certification. To be preliminarily certified, the class has to pass some initial threshold — the court, for example, should be convinced that there is a prima facie case for class certification, and that the class is adequately represented. Preliminary certification can resemble a motion to dismiss: in general, courts in the United States are required to dismiss, upon a motion by the defendant and before discovery kicks in, cases that do not present at that stage “enough facts to state a claim to relief that is plausible on its face.”^[33] The leap between a motion to dismiss under the “plausible on its face” standard, and a prima facie certification, is not too wide.^[34] Regardless of the similarity to motions to dismiss, the inquiry at the preliminary certification stage should not be fact-intensive, as opposed to the current certification requirements.

Following the previous discussion, in order for the first certification to have some constitutive effect on the class, a notice to its members is presumably required. Given the low certification requirements at the first stage, the notice could be thinner than the regular notice. If the plaintiffs are easily ascertainable and the defendant can cheaply contact them, for instance via email, the notice could be sent individually to all class members. When it is harder to individually contact class members, a relatively cheap, general public announcement can suffice.^[35] Whatever form the notice takes, it should presumably provide class members the opportunity to be heard, to plead to replace class counsel, and to remove themselves from the collective at this preliminary stage. As I suggest later, class members who opted out at this preliminary stage would retain the right to bring both their individual cases and a collective action on behalf of similarly situated persons. To the extent that the preliminary stage benefitted the class, e.g., a class-wide injunction that prevented the defendant from further committing its wrongful behavior, the representatives should be awarded their fees, relative to the benefit to the class at that stage.

Following the first, preliminary certification, and after approving the representatives of the class and paying their interim fees, if necessary, the court should proceed to deciding the second, full certification. The full certification under this proposal is similar to existing practices — a fact-intensive process, where the class should meet the more rigorous requirements that the current doctrine mandates. Along these lines, in case the court decides to fully certify, it should notify the class again, and this notice should be more comprehensive than the first, preliminary one. The second, full notice should attempt to locate more members of the class, employing notice mechanisms similar to those that are currently used. As under the current regime, with full certification and comprehensive notice, any decision precludes those who did not opt-out from litigating their individual claims (and by extension, from bringing a collective action on behalf of others). Likewise, any benefit to the class following the second certification should trigger attorneys’ fees, calculated based on that benefit.

The tiered certification design could better handle several recurring dilemmas, such as the relitigation problem and quasi class-actions. The value of the two-stage approach stems from its ability to break the current binarity — contemporary class actions are either certified or uncertified. The proposal “flattens,” to some extent, the process. It does so by providing some — but not complete — effect to class actions that have passed the first certification but failed the second. I suggest that the first, preliminary certification should preclude class members’ rights to litigate as a collective, while the second, full certification also precludes their individual rights. Accordingly, the tiered-certification procedure also generates a new category of class actions — cases that pass the preliminary certification but fail the second. Before going into those details, the following briefly highlights additional issues concerning the tiered-certification approach.

Better monitoring of class counsel during early stages. In an era of factintensive certification, the proposed tiered certification allows better monitoring of class counsel from the get-go. Adequate representation is perhaps the most important requirement of class litigation. As more efforts are currently being shifted to the pre-certification stage,^[36] courts should ensure that adequate representation exists at those early stages.^[37] However, under the current regime courts have no formal obligation to do so before certifying.^[38]

Tiered certification is a vehicle for such early scrutiny of the representatives. Under this proposal, courts are required to affirmatively hold, at the outset, that adequate representation exists. Moreover, as (scaled-back) notice is sent, class members would have the opportunity to voice their dissatisfaction with their representatives and plead to replace them, enabling courts to select the best ones. In sum, early scrutiny of the representatives is particularly important given the current regime, which on the one hand requires more efforts in the initial stages, but on the other hand provides no adequate safeguards at that time. Tiered certification can fill this gap.^[39]

Tiered certification and existing practices. The idea of tiered, scaled-back certification coheres with existing practices. It is based on the capacity of courts, at different points in time, to scrutinize the relevant issues with varying rigor and provide different procedural safeguards, i.e., notice and opt-out.

First, courts in general are accustomed to filling a gatekeeping role, i.e., evaluating the merits of a case early, at a lower threshold.^[40] Class action judges are specifically asked to decide the same issues according to different legal thresholds — first during certification, and then when deciding the entire case. In particular, multi-certification class actions do exist. Some judges have experimented with conditional certification: e.g., a preliminary approval of a settlement, under a “more relaxed standard,”^[41] or “tentative certification,”^[42] with or without notice. In fact, until 2003, Rule 23 explicitly enabled a court to conditionally certify a class.^[43] Likewise, the practice of certifying class litigation — and then decertifying — is well known.^[44] Finally, a close twotiered certification process exists under the Fair Labor Standards Act (FLSA), which enables opt-in collective litigation.^[45] In the first stage, “the district court judge employs a lenient standard and makes an initial determination of whether potential class members should receive notice of the ongoing action,” and then, in the second phase, “[c]ourts engage in a ‘more stringent’ inquiry.”^[46] Given the experience with tiered decision-making in general, and in class actions in particular, implementing the tiered-certification proposal seems easy.

Second, courts already vary the degree of procedural safeguards given to class members. In particular, “notice” is a fluid concept. Rule 23 mandates “the best notice that is practicable under the circumstances,”^[47] and leaves it to the judge to fashion the appropriate notice in each case.^[48] While Rule 23 clearly prefers individual notice,^[49] courts have employed various forms of notice, from direct mail, to publication in media outlets, to more technologically-enhanced mechanisms.^[50] Another relevant example is the authorization in Rule 23 to send notice before certification, per the discretion of the presiding judge — exercised by some courts to inform class members regarding the denial of class certification.^[51] The Israeli Act employs a similarly flexible conception of notice.^[52] As the value of the notice is already balanced against the “practicalities” of each case,^[53] scaled-back notice and opportunity to opt-out seem to be easily implemented. Indeed, in a different context, federal courts already use a staged-notification process.^[54]

Third, similar to notices, fee-shifting decisions are also very flexible, commonly taking into account the benefit to the class;^[55] hence, there is no reason to believe that judges cannot decide on attorneys’ fees within each certification stage.

“Smoothing” certification. The tiered-certification process can help plaintiffs’ attorneys to fund their cases and facilitate settlements. Currently, the certification process requires the class to pass a high threshold. As discussed before, the heightened certification standards may be desirable, but they particularly harm plaintiffs’ lawyers, who work on a contingency fee and have to finance, in advance, a costly process.^[56] A “smoother” procedure, with two “certifications,” can presumably help those lawyers — they can file the case with a smaller investment and secure more funding after passing the preliminary approval. Along the same lines, an additional holding of the judge, early in the life of the case, could reduce the informational gaps between the parties and streamline settlements.^[57] Finally, similar to the reasoning behind encouraging defendants to move to dismiss,^[58] preliminary certification could better screen out frivolous class actions.^[59]

Additional costs. The proposal can be criticized due to the additional costs that the dual certification process entails. Presumably, an additional layer of adjudication generates new expenses, such as the time the judge needs to write another decision, costs associated with another round of notices to class members, etc. These concerns are not insurmountable for the implementation of the tiered-certification procedure.

First, one may wonder who should bear the extra costs of the proposed procedure. At least with respect to the costs related to notifying class members, the solution could be similar to the rule governing regular certification. In many cases, class counsel advance these costs and can be reimbursed should the class win.^[60] This solution fits preliminary certification. Moreover, as its name suggests, scaled-back notice should be less expensive than full notice.^[61]

A second and related point is the effect of the two-stage certification process on parties’ legal expenses. While the precise effects of dual certification are complicated to analyze, there are reasons to believe that the added costs that the parties would bear from preliminary certification are minimal. Under the current doctrine, the plaintiffs in any case need to invest sizeable costs in order to pass certification, and as the foregoing has shown it may be easier for them to defer some of these costs to the second stage (should they pass the first certification).^[62] From the defendants’ perspective, a preliminary certification can weed out undesirable cases.^[63] From both parties’ perspectives, an informative judicial holding at the first certification stage should enhance the odds of settlement.^[64] Moreover, as shown below, the tiered certification process has other, more particular benefits that the parties could share, such as avoiding the relitigation problem and better handling quasi-class actions.

A third point is the added burden on judges. The preliminary certification stage requires judges to make positive findings, early on, regarding the capacity of the representing attorneys and the merits of their allegations. Moreover, to the extent that absent plaintiffs would like to voice their opinion at the preliminary certification stage, judges would have to spend additional time. Judges, however, face similar problems in the current regime. They already decide dismissals early on, and preliminary certification should not be substantially different.^[65] Likewise, judges already regulate the right-to-be-heard of litigious class members. Furthermore, I expect preliminary certification to be a relatively lenient stage, contrary to the fact-intensive full certification. Finally, early certification could save judicial costs, as the parties might settle thereafter.

To summarize, the proposed tiered certification may well add expenses. However, the practical difficulties resemble current practical difficulties, and can likewise be resolved on a case-by-case basis. Preliminary certification should not be a taxing step, and its expenses should be assessed against the benefits, e.g., early monitoring of class counsel. The next Part elaborates on a particular benefit — the utility of tiered certification in addressing the relitigation problem.

III The Relitigation Problem — Denials of Class Certification

What are the repercussions of a decision not to certify under the current, single-certification regime? Obviously, without certification class members are free to litigate their own claims. However, in that case, would class members have the right to litigate collectively? Can they mount additional attempts to certify the class? The general rule is that a decision not to certify a class entails no implications. Therefore, when one plaintiff attempts to certify class litigation, but fails, each of the remaining members can mount her own certification attempt. Alternatively put, the class as a whole has multiple bites at the (certification) apple, up to the number of its members. I refer to this situation as the relitigation problem. This Part discusses the general rule, the advantages and disadvantages of relitigation, and the existing directions for coping with the problem as well as their critique. Against this backdrop, this Part shows that the tiered-certification proposal provides a reasonable solution to the problems caused by relitigation.

IV Toward Semi-Class Actions

In addition to addressing the relitigation problem, the idea of tiered certification can open up new ways to handle certain types of class actions. While the axiomatic, single certification stage creates a “discontinuity” between certified and uncertified classes, tiered certification offers a richer view of class litigation. In particular, the tiered-certification regime anticipates a new category: cases that pass the preliminary certification but fail the second. This category can include cases that merit some, but not complete, collective treatment — cases that can be referred to as semi-class actions. While a complete treatment of semi-class actions is beyond the scope of the current Article, this Part relies on two unique procedures under the Israeli Law, which resemble preliminary certification, to illustrate this idea.

Conclusion

The current conception of class actions is binary. Certification is a fact-intensive, tedious stage, and, in terms of precluding absent members, the class is either certified or not. This description calls for a richer view that “flattens” the procedural requirements and repercussions of certification. This Article offers such a thought-experiment, advancing the idea of a tiered certification — a first, preliminary certification, not precluding the right to litigate individually; and a second, complete certification. The tiered-certification mechanism could better address several class action dilemmas.

Relitigation of denials of class certification is one example. Relitigation is a problem, and the heightened certification standards that courts have demanded seem to aggravate it. Current solutions fall short of remedying the problem. They either refuse to relieve the burden of relitigation, due to adherence to the binary concept of class certification; or preclude the right of class members to litigate collectively without appropriate safeguards. The way out of the quandary lies elsewhere. The tiered-certification proposal invokes the same legal fiction that serves to bind absent members to the final outcome of class litigation — notice and opportunity to participate — to give denials a preclusive effect. Thus, it enables partial class actions for the purpose of denying subsequent class actions. True, the scaled-certification proposal is not devoid of difficulties. However, it uses familiar tools to balance the desire to avoid relitigation with the notion of protecting absent plaintiffs.

More generally, shedding the concept of a single certification stage suggests new directions for handling class actions. In particular, tiered certification envisions “semi-class actions”: class actions that should pass a preliminary threshold but fail full certification. Semi-class actions are, in essence, situations in which collective proceedings entail some social value, but triggering the apparatus of modern class litigation is too onerous and not cost-effective. Trivial violations of regulatory schemes are an example. While there is some social value in rectifying the defendant’s behavior, current certification requirements are too high to handle these regulatory violations. Instead of either certifying or denying class treatment, a middle ground could be a prospective injunction that forces the defendant to comply with the regulatory scheme without precluding individual proceedings. Tiered certification, then, breaks the one-size-fits-all nature of the current certification requirements and opens new opportunities to regulate class actions — a middle ground between precluding and not precluding, no class litigation and full-fledged proceedings.

The tiered-certification proposal also has some more immediate benefits in the current climate. Originally, the rules in the United States envisioned a process in which certification decisions are made early on, and, as a result, “certification notice [would be] sent to the class prior to the resolution of the case on the merits.”^[160] However, the trend both in the United States and Israel is front-loading, i.e., shifting efforts to pre-certification phases. As a corollary, pre-certification stages have become longer, and notice is sent later than it used to be. The proposal restores the original intention of Rule 23, as it guarantees minimal procedural safeguards early on. Therefore, it can also improve representation from the inception of a case, filling a gap in the current regime.