3d Circuit en banc on settlement and predominance

Sullivan v. DB Investments, Inc., 667 F.3d 273 (3d Cir. 2011) (en banc)

Believe it or not, this is a relatively brief summary of an important en banc opinion from the Third Circuit interpreting Dukes as of limited importance to the consumer class action context.  The district court certified two nationwide settlement classes of direct and indirect purchasers of De Beers diamonds to settle antitrust and consumer protection claims against De Beers.  A panel reversed the certification, but the en banc court of appeals reinstated the certification, ruling that “the predominance inquiry should be easily resolved here based on De Beers's conduct and the injury it caused to each and every class member, and that the straightforward application of Rule 23 and our precedent should result in affirming the District Court's order certifying the class.”  Contrary to the dissent’s contentions, it is not true that a district court must ensure that each class member possesses a viable claim in order to certify a class.

Here, differences in state consumer protection statutes didn’t override the commonalities of the class.  As the Supreme Court has stated, “predominance is a test readily met in certain cases alleging consumer or securities fraud or violations of the antitrust laws.” Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 625 (1997).  Common questions included basic elements of antitrust liability and whether the class had been injured, as well as a common jurisdictional question about De Beer’s refusal to submit to the jurisdiction of US courts.

The panel that initially vacated the certification noted variations in state consumer protection laws: “differences in whether indirect purchasers may invoke consumer protection and unjust enrichment statutes to gain antitrust relief; variations in the extent of elements of proof necessary to establish unjust enrichment or consumer fraud; and dissimilarities in whether a plaintiff must lack an adequate remedy at law to bring an equitable claim.”  Based on its conclusion that evidence of price-fixing doesn’t always give rise to an unjust enrichment or consumer protection claim, the panel concluded that common issues didn’t predominate, even though De Beers was willing to stipulate to liability in all 50 states in order to achieve a nationwide settlement—the only basis on which it was willing to settle.  The panel concluded that the certification “wrongly allowed the sovereignty of the states to be subordinated to De Beers's desire to resolve all indirect purchaser claims simultaneously.”

The en banc court identified three “guideposts” for predominance: (1) “commonality is informed by the defendant's conduct as to all class members and any resulting injuries common to all class members”; (2) “variations in state law do not necessarily defeat predominance”; and (3) “concerns regarding variations in state law largely dissipate when a court is considering the certification of a settlement class.”  Where a broad, unified course of conduct by the defendant is the foundation of liability, commonality depends on that and not on the conduct of individual class members, even if state laws differ. 

The dissent misread Dukes to support “its thesis that an inquiry into the existence or validity of each class member's claim is required at the class certification stage. To the contrary, Dukes actually bolsters our position, making clear that the focus is on whether the defendant's conduct was common as to all of the class members, not on whether each plaintiff has a ‘colorable’ claim.… [C]ommonality is satisfied where common questions generate common answers ‘apt to drive the resolution of the litigation.’ That is exactly what is presented here, for the answers to questions about De Beers's alleged misconduct and the harm it caused would be common as to all of the class members, and would thus inform the resolution of the litigation if it were not being settled.”

Indeed, the court continued, “the presence of these questions stemming solely from De Beers's asserted behavior and the fact that all class members purchased diamonds is an apt illustration of why the predominance test is “readily met in certain cases alleging consumer [ ] fraud or violations of the antitrust laws” (quoting Amchem again, and citing the advisory committee notes on Rule 23(b)(3)).

Relatedly, state law variations are insufficient to defeat certification: “We have never required the presentation of identical or uniform issues or claims as a prerequisite to certification of a class. Rather, our jurisprudence evinces a pragmatic response to certifications of common claims arising under varying state laws.”  Where there are a limited number of predictable patterns, and where deviations could be dealt with by grouping/subclassing at trial, relatively minor differences in state law can be subsumed in a single class.  “This tactic in litigation advances the laudatory purposes of the class action device, ‘preserv[ing] the resources of both the courts and the parties by permitting issues affecting all class members to be litigated in an efficient, expedited, and manageable fashion.’”

Separately, a settlement class diminishes concerns over variations in state law, because manageability is no longer an issue.

The dissent argued that, despite the predominance of shared issues of fact and law, states’ inconsistent treatment of indirect purchaser damages claims overwhelmed the commonalities.  Because a large number of states don’t allow such claims, and some don’t allow “an end-run” around antitrust using consumer protection laws, the dissent contented, a large proportion of the indirect purchaser class lacked any valid claims, and thus couldn’t “predominantly” share common issues of law or fact with indirect purchasers who had valid or at least colorable claims.

The en banc majority found this argument “misdirected.”  “The question is not what valid claims can plaintiffs assert; rather, it is simply whether common issues of fact or law predominate…. [T]here is no ‘claims’ or ‘merits’ litmus test incorporated into the predominance inquiry beyond what is necessary to determine preliminarily whether certain elements will necessitate individual or common proof.”  The legal viability of asserted claims should more properly be considered through a motion to dismiss or for summary judgment, not as part of a Rule 23 certification process.  Introducing a Rule 12(b)(6)-like inquiry into certification would conflict with Rule 23 itself, which “makes clear that a district court has limited authority to examine the merits when conducting the certification inquiry…. A court may inquire whether the elements of asserted claims are capable of proof through common evidence, but lacks authority to adjudge the legal validity or soundness of the substantive elements of asserted claims. Put another way, a district court may inquire into the merits of the claims presented in order to determine whether the requirements of Rule 23 are met, but not in order to determine whether the individual elements of each claim are satisfied.”  That is, the key is whether the elements of the claim can be proved through evidence common to the class rather than individual to its members.  Again, this is especially true for a settlement class. 

The en banc majority also noted that the claimed defect was in statutory standing, but that’s not jurisdictional.  (This seems to leave some room for defeating certification due to Article III problems with class members’ claims, but the court is notably silent on that issue, which has been vigorously argued in the Ninth Circuit.)  “Here, the supposed lack of one element necessary to prove a violation on the merits—statutory standing—does not establish a concomitant absence of other predominantly common issues.”  (The majority cited a case affirming certification even though some plaintiffs couldn’t establish reliance, which was a necessary element of their state-law fraud claims—this may also have relevance to the Article III question, since at least in California the argument has generally been that lack of reliance means lack of constitutionally recognizable injury in fact.)

The dissent “mistakenly places the cart before the horse by requiring the District Court to establish the validity of the disputed elements of the asserted claims—namely, the viability of indirect purchaser actions under state substantive laws—prior to certifying the class.”  Again, Rule 23 doesn’t incorporate Rule 12(b)(6).  “In addition to exceeding the plain requirements of Rule 23, in nationwide class settlements, such as the one here, and even if limited to a statutory standing inquiry, this analysis would necessitate an intensive, fifty-state cataloguing of differences in state law at an early stage of the proceedings, and without the benefit of a developed record.”  The choice of law inquiry in particular would be too difficult so early on, and district courts would often confront unsettled matters of state law, “needlessly introduc[ing] additional legal uncertainty into a certification process that does not demand it.”

The dissent’s proposed “colorable legal claim” threshold for commonality would also imply that a court should consider every potential disqualifier from having such a claim:

[S]hould the court consider whether all potential class members complied with applicable pre-notice requirements under the relevant substantive law? … Should the court evaluate whether each class member's claim complies with the applicable statute of limitations? The answers to these questions most certainly implicate whether a litigant, in a class action or otherwise, has a ‘colorable legal claim.” These questions, moreover, show how flawed, from an administrative, logical, and practical standpoint, the dissent's and objectors' approach really is. No class would ever be certified because it would be impossible to demonstrate that every class member has a “colorable legal claim.” More than this, it would gut commonality, for, most certainly, individual issues would then predominate. There would simply be no class that could meet this commonality and predominance test.

(The dissent, in response, referred to statute of limitations and similar issues as questions of “form” that could appropriately be disregarded in a class action, unlike the basic question of legal entitlement to relief here.)  The majority continued that such a standard would “effectively rule out the ability of a defendant to achieve ‘global peace’ by obtaining releases from all those who might wish to assert claims, meritorious or not.… Here, in an effort to avoid protracted litigation and future relitigation of settled questions in federal and state courts across numerous jurisdictions, De Beers pursued a global settlement and demanded a release of potential damage claims in all fifty states.”  It’s a legitimate policy objective to facilitate such global settlements, despite the dissent’s disparagement of them.  “[A]chieving global peace is a valid, and valuable, incentive to class action settlements.… If the dissent's position were adopted, there would be no settlements, collusive or otherwise.”  The result would be an endless series of state-by-state claims.  It’s better to concentrate the litigation in a single action.

There’s also nothing troubling about such a settlement under the Rules Enabling Act or principles of federalism.  The dissent argued that the approval “violated principles of federalism by extending to the plaintiffs a substantive right that they could not have asserted in state court.”  But recognizing a private agreement, which is what a settlement is, doesn’t constitute a recognition or expansion of substantive rights.

The majority also rejected challenges to the fairness of the settlement.

Judge Scirica concurred, noting that given the claimed reason to contest a settlement—to protect absent class members—it was unclear how such absent members, all of whom claimed injury, were harmed by De Beers’ willingness to settle all potential claims.  Adding a 12(b)(6) analysis for settlement class certification “could present serious difficulties in administration and greatly increase costs and fees, and may deplete rather than increase the recovery of even successful plaintiffs.”

Judges Jordan and Smith dissented in strong terms.  The problem, they argued, was that De Beers’ conduct gave rise to a cause of action in some states, but not in others, and it was inappropriate to group the two sets of class members together: there was neither commonality in questions of law or fact nor predominance between those groups.  If all we do is look at the defendant’s conduct in establishing commonality, then no one in the world can be excluded from the class regardless of conceded lack of injury.  Common questions should have answers that affect the validity of all class members’ claims and thus are legally relevant. 

Me: The concept of overdetermination might be useful here: is it legally relevant if some fact is either necessary to or a complete defense to liability, when there is also some specific defense applicable to a particular class member?  We recognize pleading in the alternative all the time.  The dissent says “if predominance means anything, it must mean that the resolution of something will actually affect somehow the claims of all class members,” but I don’t think that can be right, at least not for what the dissent seems to mean by “affect”; otherwise a single valid defense individual to a single class member would defeat predominance.  This might be a good time to pull out the “I do not think that word means what you think it means” chestnut.

In any event, the dissent concluded, “for there to be any common questions, all class members must have at least some colorable legal claim.”  The dissent didn’t want a full 12(b)(6) determination, only exclusion of class members where, because of statutory differences, “it is clear” that they are “entirely without a cognizable claim.”  There’s a difference between a doubtful claim (which shouldn’t be excluded on this test) and a claim that clearly does not exist.  “Assuming the parties revise the class to eliminate claims clearly lacking a colorable legal basis, and assuming the class otherwise satisfies Rule 23, the district court could then certify the class.”  In a footnote, the dissent elaborated:

Claims that are of doubtful quality still have, as the adjective indicates, some doubt about them, which means they still retain at least some superficial possibility of being valid. Such claims, because they cling to that possibility, will typically not need to cause a district judge any agita in addressing the certification of a class for settlement purposes. The central problem in this case, however, goes beyond factual disputes or debatable points of law. The problem here is that there are class members who, according to the plain terms of controlling law, have no claim at all, not even a dubious one.

My take, for what it’s worth: the dissent has a point, but it’s fully rebutted by the majority’s response that there is no limit on the dissent’s principle that would keep courts from a full 12(b)(6) determination as to each class member, since there are all kinds of reasons a class member might not have a colorable legal claim.  Certainly every defendant would invent reasons why its objections to commonality fell within the dissent’s principle: asking for a class defined by reliance, for example, since in some states a consumer without reliance “clearly lacks a colorable basis” for a claim, and then arguing that the class definition is too imprecise to survive.  Subclassing would appear to be the best solution where the barriers to recovery are state-wide and based on statutory standing. 

The dissent suggested that real class members are harmed by inclusion of class members who’d otherwise lose, since they share in the settlement fund.  Both sides seem to me to be speculating ahead of data: the majority’s counterargument is that there wouldn’t be many (if any) settlements without global settlements, which seems at least equally plausible.  The majority, in a footnote, said: “the dissent assumes that the size of the settlement fund would be the same if the Indirect Purchasers who cannot recover individually were excluded from the class. Surely this cannot be the case, for the settlement amount to which De Beers has agreed must be based in large part on the number of potential class members and on securing global peace. Had those Indirect Purchasers who could not recover individually been excluded, we seriously doubt that the Indirect Purchaser settlement fund would still be $272.5 million.”

On federalism, the dissent argued that a state that has denied its citizens a particular claim shouldn’t have its judgments overridden by a settlement that allows those citizens to recover as if they had a claim.  Global peace isn’t the be-all and end-all.  “By failing to enforce the limits of Rule 23, today's decision will encourage frivolous class action claims and have the predictable consequence of weakening the incentives—the sheltering shadow—under which non-frivolous disputes would otherwise be properly resolved.”