Citizens Bank, N.A. v.

• Court: Court of Appeals for the Third Circuit
• Decided: October 5, 2021
• Docket: 19-3046
• Status: Published

Opinion

PRECEDENTIAL

UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT _____________

No. 19-3046

_____________

In re: CITIZENS BANK, N.A., Petitioner

On Petition for a Writ of Mandamus from the United States District Court for the Western District of Pennsylvania Related to District Court No. 2-15-cv-01541 District Judge: The Honorable Arthur J. Schwab

Submitted Pursuant to Third Circuit L.A.R. 34.1(a) ________________

Before: SMITH, Chief Judge, SHWARTZ and SCIRICA, Circuit Judges

(Filed: October 5, 2021) Thomas E. Hill Holland & Knight 400 South Hope Street 8th Floor Los Angeles, CA 90071

Brian A. Sutherland Reed Smith 101 Second Street Suite 1800 San Francisco, CA 94105

Christina Tellado Holland & Knight 400 South Hope Street 8th Floor Los Angeles, CA 90071

Kim M. Watterson Reed Smith 225 Fifth Avenue Suite 1200 Pittsburgh, PA 15222

Counsel for Petitioner

Justin L. Swidler Joshua S. Boyette Swartz Swidler 2 1101 Kings Highway North Suite 402 Cherry Hill, NJ 08034

Daniel A. Horowitz O’Brien Belland & Bushinsky 509 South Lenola Road Building 6 Moorestown, NJ 08057

Robert D. Soloff 7805 Southwest 6th Court Plantation, FL 33324

Counsel for Respondents

_____________________

OPINION OF THE COURT _____________________

SMITH, Chief Judge.

Twelve current and former mortgage loan officers (MLOs) claim that Citizens Bank forced them—and more than a thousand of their colleagues—to work over forty hours a week without paying them the overtime they were due under state and federal law. They filed a single complaint bringing a collective action under the Fair 3 Labor Standards Act (FLSA), 29 U.S.C. §§ 207, 216, and parallel state-law claims that they wished to pursue as a class action under Rule 23 of the Federal Rules of Civil Procedure. 1

The District Court scheduled a trial on the primary factual issue in the FLSA opt-in collective action but left unresolved whether it would certify a class for the state- law opt-out Rule 23 action. Because the FLSA collective action and the Rule 23 class action turn on the same facts, Citizens strongly objected to that procedural order of business. Yet the District Court essentially ignored Citizens’ objections.

With a trial date looming, Citizens filed a petition in our Court for a writ of mandamus. We stayed the case to decide that petition. This opinion explains our decision to issue the stay.

1 The District Court had jurisdiction over the FLSA claims under 28 U.S.C. § 1331 and had supplemental jurisdiction over the state-law claims under 28 U.S.C. § 1367. Because Plaintiffs styled this as a putative class action for over $5 million in damages, and because at least one MLO is a citizen of a state different from Citizens Bank, the District Court also had jurisdiction over the state law claims under 28 U.S.C. § 1332(d)(2)(A). 4 I.

Before discussing the specifics of the parties’ dispute, we will first compare and contrast the two types of aggregate litigation that Plaintiffs are simultaneously pursuing: an FLSA opt-in collective action and a Rule 23(b)(3) opt-out class action.

We begin with the FLSA, which provides a private cause of action against an employer for failing to pay overtime for a workweek of more than forty hours. 29 U.S.C. § 207(a). An FLSA action may be brought “by any one or more employees for and [on] behalf of himself or themselves and other employees similarly situated.” 29 U.S.C. § 216(b). The FLSA thus “provides a vehicle for managing claims of multiple employees against a single employer.” Halle v. W. Penn Allegheny Health Sys., Inc., 842 F.3d 215, 223 (3d Cir. 2016). However, “[n]o employee shall be a party plaintiff to any such action unless he gives his consent in writing to become such a party and such consent is filed in the court in which such action is brought.” 29 U.S.C. § 216(b).

Accordingly, after a “modest factual showing” by the named plaintiffs, a district court must first conditionally certify that the proposed collective action

5 plaintiffs are “similarly situated.” 2 Halle, 842 F.3d at 224 (citation omitted); 29 U.S.C. § 216(b). Conditional certification permits the dissemination of a court-approved notice to all potential plaintiffs, who are then given the opportunity to affirmatively opt in as plaintiffs to the lawsuit. Halle, 842 F.3d at 224.

Once the FLSA plaintiffs provide written notice that they have opted in to the collective, the parties conduct certification-related discovery and, eventually, the group of plaintiffs moves for final certification. Id. at 225. The group then bears the burden of demonstrating by a preponderance of the evidence that they are all “similarly situated” under the FLSA. Id. at 226. If the District Court finally certifies the FLSA collective action, the matter goes forward with the participation of all plaintiffs who have opted in. Id. If, on the other hand, final certification is denied, the opt-in plaintiffs are dismissed as plaintiffs and only the original named plaintiffs proceed to trial. Id.

2 As we observed in Halle, the procedures for determining whether employees are “similarly situated” for purposes of an FLSA collective action have been fashioned by courts, as Congress has not promulgated any framework nor are there any specifically applicable procedural rules. 842 F.3d at 223. 6 Turning to the class action device, Rule 23(a) provides that a matter may proceed as a class action only if: the class is so numerous that joinder of all members is impracticable; there are questions of law or fact common to the class; the claims or defenses of the representative parties are typical of the claims or defenses of the class; and the representative parties will fairly and adequately protect the interests of the class. Further, under Rule 23(b)(3), the District Court must find that questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.

These issues must be resolved when a District Court decides whether or not to certify a class. Certification requires a two-step analysis in which the District Court must first determine whether the putative class satisfies the numerosity, commonality, typicality, and adequacy of representation provisions of Rule 23(a), and then proceeds to analyze the predominance and superiority provisions of Rule 23(b)(3). Reinig v. RBS Citizens, 912 F.3d 115, 124– 25 (3d Cir. 2018). Class certification is appropriate only if, after “rigorous analysis,” the District Court concludes that plaintiffs satisfy each and every element by a preponderance of the evidence. Id. at 125 (quoting Marcus v. BMW of N. Am., 687 F.3d 583, 591 (3d Cir. 2012)); see also Gen. Tel. Co. of the Sw. v. Falcon, 457 U.S. 147, 161 (1982). The class certification decision 7 must happen “at an early practicable time.” Fed. R. Civ. P. 23(c)(1)(A). And such a decision—unlike a certification in the FLSA context—may be immediately appealed under Rule 23(f).

Certification of a class means that all individuals falling within the class definition will be bound by the judgment unless they affirmatively request to be excluded. Fed. R. Civ. P. 23(c)(2)(B). Accordingly, after a class is certified, court-approved notice is provided to all class members to advise them, inter alia, of the binding effect of a judgment and to permit them an opportunity to opt out of the class. Fed. R. Civ. P. 23(c)(2). This opt-out opportunity under Rule 23 stands in sharp contrast to the FLSA’s opt-in requirement, and “is the most conspicuous difference between the FLSA collective action device and a class action under Rule 23.” Halle, 842 F.3d at 225; see also Reinig, 912 F.3d at 132.

Despite these marked differences, we have held that an FLSA opt-in collective action is not, by its nature, incompatible with a parallel state law Rule 23 opt-out class action. Knepper v. Rite Aid Corp., 675 F.3d 249, 261 (3d Cir. 2012). Yet the present mandamus petition brings into sharp relief some of the potential challenges of trying a case that simultaneously includes both forms of aggregate litigation.

8 II.

With these differences between the FLSA opt-in collective action and the Rule 23 opt-out class action firmly in mind, we return to the procedural history of the parties’ dispute.

In November 2015, three current and former MLOs, Alex Reinig, Ken Gritz, and Bob Soda, filed a complaint alleging that Citizens maintains an unofficial policy or practice requiring MLOs to work off the clock in excess of forty hours per week without paying overtime wages due in accordance with the FLSA, 29 U.S.C. § 207, and Pennsylvania law, 43 Pa. Stat. Ann. § 260.1 et seq., and § 333.101 et seq. 3

Plaintiffs moved for conditional certification of an FLSA opt-in collective action, which the District Court granted in May 2016. Mem. Order May 3, 2016. The District Court concluded that, for FLSA certification purposes, Plaintiffs “established that the ‘off the clock’

3 Plaintiffs also claim that Citizens structures payment of commissions and bonuses to recapture overtime payments in violation of the FLSA and analogous state law. Because the District Court granted summary judgment to Citizens on the “recapture” claims, they are not relevant to the issues before us and will not be discussed in this opinion. 9 claims are related to Citizens’ policy. . . .” Id. at 5. Shortly thereafter, the District Court scheduled a trial for September 25, 2017.

After granting conditional FLSA certification, the District Court ordered Plaintiffs to notify the potential members of the collective that they would have 100 days to opt in to the FLSA action. In accordance with the District Court’s order, Plaintiffs sent notice to over 1,000 current and former MLOs. Of those, 351 provided the requisite consent forms allowing them to opt in.

After the 100-day period expired, Plaintiffs filed an amended complaint adding nine named plaintiffs to the lawsuit, alleging state law claims arising from the laws of Connecticut, Illinois, Massachusetts, Michigan, New Hampshire, New York, North Carolina, Ohio, and Rhode Island as well as Pennsylvania. 4 Shortly after filing the amended complaint, Plaintiffs filed a motion for class certification under Rule 23, seeking certification of ten classes, each of which would pursue claims under the overtime laws of a particular state. Citizens responded with two related submissions: one opposing the class

4 In August 2017, with the District Court’s leave, Plaintiffs filed a second amended complaint to add two additional named plaintiffs. It is, for our purposes, the operative complaint. 10 certification motion and the other seeking decertification of the FLSA collective action. The parties also filed cross- motions for summary judgment.

By stipulation, the parties agreed to the appointment of a Special Master, who recommended, inter alia, certifying a class for Plaintiffs’ state law claims under Rule 23, denying Citizens’ motion for decertification of the FLSA collective action, and granting final FLSA certification. In its objections to the Special Master’s recommendations, and central to the matter before us, Citizens argued that the scheduled FLSA trial date must be postponed because the putative class had not yet been notified of the Rule 23 certification decision, and therefore had not been given a chance to opt out.

The District Court adopted the Special Master’s report and recommendations in full. In addition to certifying the Rule 23 state law classes, the District Court granted final FLSA collective action certification, concluding that the FLSA opt-in plaintiffs are similarly situated because “the MLOs share the same job description with similar (if not identical) job duties, are paid pursuant to the same compensation plan(s), are subject to the same policies, and assert the same claims for unpaid off-the-clock overtime wages in this lawsuit.” Reinig v. RBS Citizens, 2017 WL 3599489, at *3 (W.D. Pa. Aug. 22, 2017).

11 The District Court also rejected Citizens’ objection to proceeding with the previously scheduled FLSA trial:

The remainder of Defendant’s arguments are procedural and regard matters within the sound discretion of the District Court to manage litigation before it. Trial of a single issue regarding Plaintiffs’ FLSA off-the- clock claims is scheduled to commence, and will commence, on September 25, 2017. This in no way interferes with the state subclasses[’] right to receive notice of the pending state-law claims and to opt-out of the action if they so choose.

Id. (citation omitted). The scheduled FLSA trial would address whether “Citizens Bank had a policy or practice that caused [MLOs] to not report all of the hours they worked.” JA 255.

Citizens timely filed a Rule 23(f) petition objecting to class certification, and we granted that petition. 5 In

5 Citizens also filed a petition for a writ of mandamus, arguing that the District Court’s planned trial on the main factual issue presented in both the collective and class actions, without first providing notice to the Rule 23 class members and giving them an opportunity to opt out, would violate both Rule 23 and due process. That petition 12 resolving the petition, we discerned numerous flaws in the District Court’s consideration of the Rule 23 class certification issues. For instance, we concluded that the District Court had not adequately defined any class or sub- classes as required by Rule 23(c)(1)(B), leaving us “to comb through and cross-reference multiple documents in an attempt to cobble together the parameters defining the class and a complete list of the claims, issues, and defenses to be treated on a class basis.” Reinig, 912 F.3d at 126.

In addition, we could not determine what evidence the District Court relied on to conclude that Plaintiffs had satisfied Rule 23(b)’s predominance and commonality requirements. Its “barebones” analysis was insufficient to allow us to conclude that it had conducted the “rigorous” review that Rule 23 requires, particularly because it did not reconcile “contradictory testimony and other evidence” undermining Plaintiffs’ claim of the existence of a company-wide policy to discourage or underpay overtime. Id. at 129. Accordingly, we reversed and remanded the District Court’s Rule 23 class certification order, expressing “serious doubts whether” Plaintiffs’ evidence was “sufficiently representative of the class as a whole,” and instructing the District Court to “conduct a

ultimately was dismissed as moot because, in resolving the Rule 23(f) petition, our Court vacated the underlying class certification decision. 13 ‘rigorous’ examination of the factual and legal allegations underpinning [the] claims before deciding . . . class certification.” Id. at 130.

In its Rule 23(f) petition, Citizens also asked us to exercise pendent appellate jurisdiction to consider the District Court’s FLSA collective action certification. We declined to do so. Although we acknowledged “that some of the factors and evidence necessary to satisfy the prerequisites of Rule 23 and § 216(b) may overlap,” id. at 132, we determined that “Rule 23 class certification and FLSA collective action certification are fundamentally different creatures.” Id. at 131. Each is governed by its own legal standard—for the class action, Rule 23 applies, while for the FLSA collective action, § 216(b)’s “similarly situated” language controls. So the two certification decisions, while related and potentially resting upon the same evidence, were not “inextricably intertwined” for purposes of the narrow doctrine of pendent appellate jurisdiction. See id. at 131–32. Because we lacked jurisdiction to review it in the interlocutory Rule 23(f) proceeding, the FLSA certification remained in place. See id. at 133.

On remand, despite our express instruction to the District Court to “conduct a rigorous examination of the factual and legal allegations underpinning Plaintiffs’ claims,” id. at 130, the District Court chose not to return to the question of class certification. Instead, it pressed on

14 with Plaintiffs’ FLSA collective action, reiterating its plan to proceed with a single-issue FLSA trial without first deciding whether to certify a Rule 23 class. The issue to be placed before the jury was to be: “Did Plaintiffs prove by a preponderance of the evidence that Citizens Bank had a policy or practice that caused mortgage loan officers to not report all of the hours they worked (i.e., to work ‘off the clock’)?” Reinig v. RBS Citizens, 386 F. Supp. 3d 602, 608 (W.D. Pa. June 25, 2019).

Citizens raised numerous objections to the District Court’s planned FLSA trial, moving to stay it until after a Rule 23 class certification decision had been made. The District Court declined to stay the trial and failed to meaningfully address the merits of the objections. Instead, it characterized Citizens’ position as a delay tactic and an effort to undermine the District Court’s ability to manage its own docket.

Citizens then came to us seeking a writ of mandamus. Specifically, Citizens asked us (1) to direct the District Court to refrain from proceeding with the FLSA collective action trial until the Rule 23 class certification motion is decided; (2) if a Rule 23 class is certified, to direct the District Court to refrain from proceeding with trial until after class members have been notified and given an opportunity to opt out; and (3) to reassign the case to a new District Judge. Citizens also sought a stay pending our decision, which we granted.

15 The day we issued our stay, the District Judge filed a supplemental response to the mandamus petition, joining Citizen Bank’s request that the case be reassigned. 6 Because the District Judge joined Citizens’ request for relief, we will now dissolve our stay so that the Chief Judge of the United States District Court for the Western District of Pennsylvania may reassign this case. We will dismiss the mandamus petition in part as moot insofar as it requests reassignment. And given that pending reassignment, we expect that the District Judge who will assume management of this litigation will take into account our ensuing discussion of the interests at stake here and the factors that motivated our grant of a stay in the first place. We are confident that the District Court will heed our prior direction to “conduct a ‘rigorous’ examination of the factual and legal allegations underpinning [the] claims before deciding . . . class certification.” Reinig, 912 F.3d at 130. Such analysis is at the heart of Rule 23 practice and procedure. See In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305, 316 (3d Cir. 2008). Accordingly, we will deny the remainder of

6 The District Judge indicated that he had taken senior status under 28 U.S.C. § 371 and favored a “generational- shift to another trial judge to manage the case in the future.” Supp. Response 1. He reiterated his view that Citizens’ mandamus petition is meritless. 16 the petition without prejudice because it is unnecessary for us to consider at this time.

III.

The parties agree that the standard framework for a stay pending appeal governs Citizens’ application for a stay pending resolution of its mandamus petition. We therefore considered:

(1) whether Citizens was sufficiently likely to obtain mandamus relief;

(2) whether Citizens would suffer irreparable injury absent a stay;

(3) whether a stay would substantially injure Plaintiffs; and

(4) where the public interest lies.

See Nken v. Holder, 556 U.S. 418, 434 (2009) (treating this test as the standard inquiry “whenever a court order may allow or disallow anticipated action before the legality of that action has been conclusively determined”). 7 “The

7 We had jurisdiction over the stay application under the All Writs Act, 28 U.S.C. § 1651. See Nken, 556 U.S. at 426. 17 first two factors of th[is] traditional standard are the most critical.” Id. Here, both of those factors supported a stay, and the third factor did as well. The fourth did not, at the time we entered our order, impact our weighing process. Because the balance of the factors preponderated strongly in favor of relief, we granted the stay.

1. Citizens Had a Sufficient Likelihood of Success on its Mandamus Petition, and Mandamus Is the Only Relief Available

To prevail on the merits of a mandamus petition, the petitioner must show that the district court clearly and indisputably erred, and that no other adequate alternative remedy exists. 8 See In re Howmedica Osteonics Corp., 867 F.3d 390, 401 (3d Cir. 2017). Although “the remedy of mandamus is a drastic one, to be invoked only in extraordinary situations,” Allied Chem. Corp. v. Daiflon, Inc., 449 U.S. 33, 34 (1980), the bar is set lower when we consider whether to grant a stay pending resolution of a petition. The stay applicant need show only “a reasonable chance, or probability, of winning” mandamus relief in

8 Mandamus also requires a showing of irreparable injury. See In re Howmedica Osteonics Corp., 867 F.3d 390, 401 (3d Cir. 2017). But because the test for a stay already accounts for irreparable injury, we address it within that framework. See infra Section III.2. 18 order to prevail. In re Revel AC, Inc., 802 F.3d 558, 568 (3d Cir. 2015). In other words, we may grant a stay even if the ultimate likelihood of granting the mandamus petition is below 50 percent, so long as it is “significantly better than negligible.” Id. at 571.

Here, for the reasons we will describe below, Citizens cleared that hurdle. In this hybrid wage-and-hour suit, the District Court refused to meaningfully engage with Citizens’ objections to the Court’s proceeding with trial in the FLSA opt-in collective action without first considering whether to certify the related state-law Rule 23 opt-out class action—even though the planned trial would resolve a fact issue that is central to all the claims, and even though our Court had remanded the case specifically to require the District Court to conduct a rigorous analysis on Rule 23 class certification. Furthermore, mandamus is the only relief available to Citizens that can resolve that error.

A. Likelihood of Success

Even if this case had been filed as a Rule 23(b)(3) opt-out class action without the presence of an FLSA collective action, we would view a trial-before- certification approach with the utmost skepticism. Such a procedural sequence would be ignoring Rule 23’s text and history, flouting Supreme Court precedent, and departing from the case law of seven circuits while undercutting four

19 others. Yet the District Court’s plan to try the main factual question in the FLSA collective action—that is, whether “Citizens Bank had a policy or practice that caused [MLOs] to not report all of the hours they worked,” JA 255—would inevitably encroach on the merits of a Rule 23 class action that, thus far, is without a certified class. Such a course of action triggers all of the same weighty concerns as a trial-before-certification approach. Moreover, complexities unique to hybrid wage-and-hour actions like this one compound the potential pitfalls. Given the various red flags raised by Citizens in response to the District Court’s planned approach, as well as the District Court’s decision to plow ahead without acknowledging or engaging with any of them, we concluded without hesitation that Citizens had a reasonable probability of prevailing on the merits of its mandamus petition.

We begin, of course, with the text of Rule 23 which obliges district courts to decide certification “[a]t an early practicable time after a person sues or is sued as a class representative.” Fed. R. Civ. P. 23(c)(1)(A). When used as a modifier, “early” means “absolutely or relatively near to the beginning of a period of time”; it “designat[es] the first part or stage” of something. Early, Oxford English Dictionary (3d ed. 2015).

True enough, “practicable” means “feasible,” so a district court has discretion in its choice of timing.

20 Practicable, Oxford English Dictionary (3d ed. 2015). But the District Court’s plan to conduct a full trial cannot satisfy any definition of “early.” The District Court’s planned FLSA trial on whether “Citizens Bank had a policy or practice that caused [MLOs] to not report all of the hours they worked,” JA 255, would resolve many, if not all, of the factual issues pertaining to the Rule 23 class. Yet under the District Court’s approach, Rule 23 class certification vel non would remain unresolved until after trial. The period after a trial on the parties’ core factual dispute hardly comprises a case’s “first part or stage.” See Am. Pipe & Constr. Co. v. Utah, 414 U.S. 538, 547 (1974).

Rule 23’s history confirms that a post-trial certification decision is strongly disfavored. Originally, Rule 23 was essentially “an invitation to joinder”; it lacked a mechanism for adjudicating absent class members’ claims. Id. at 546 (internal quotation marks omitted). And that invited a very particular form of abuse:

[M]embers of the claimed class could in some situations await developments in the trial or even final judgment on the merits in order to determine whether participation would be favorable to their interests. If the evidence at the trial made their prospective position as actual class members appear weak, or if a judgment precluded the possibility of a favorable determination, such

21 putative members of the class who chose not to intervene or join as parties would not be bound by the judgment.

Id. at 547. The “unfair” upshot allowed class members “to benefit from a favorable judgment without subjecting themselves to the binding effect of an unfavorable one.” Id.

To end the unfairness of what came to be known as “one-way intervention,” Rule 23 was amended to require that courts must determine certification “[a]s soon as practicable after the commencement of [the] action.” Fed. R. Civ. P. 23(c)(1) (amended 2003). The point could not have been more clear: The “amendment[] w[as] designed, in part, specifically to mend this perceived defect in the former Rule and to assure that members of the class would be identified before trial on the merits and would be bound by all subsequent orders and judgments.” Am. Pipe & Constr. Co., 414 U.S. at 547.

In 2003, the language was altered to require certification “at an early practicable time.” Fed. R. Civ. P. 23(c)(1)(A). But the change was meant to better “reflect[] prevailing practice,” Fed. R. Civ. P. 23(c)(1) advisory committee’s note to 2003 amendment, and guide judges to “consider carefully all relevant evidence . . . before certifying a class,” In re Hydrogen Peroxide, 522 F.3d at 320, not to encourage unjustified delay of a certification

22 decision. 9 And it definitely did not “restore the practice of ‘one-way intervention’ that was rejected by the 1966 revision.” Fed. R. Civ. P. 23(c)(1) advisory committee’s note to 2003 amendment.

In the 1970s, the Supreme Court indicated a preference for early-in-the proceeding class certification in Eisen v. Carlisle & Jacquelin. 417 U.S. 156, 178 (1974). Eisen concerned a plaintiff who wished to “obtain a determination on the merits of the claims advanced on behalf of the class without any assurance that a class action may be maintained.” Id. at 177–78. The Eisen court observed, with disapproval, that addressing the merits prior to certification “allow[s] a representative plaintiff to secure the benefits of a class action without first satisfying

9 The advisory committee recognized that “[t]ime may be needed to gather information necessary to make the certification decision” on an “informed basis” and to take into account other considerations, such as the designation of class counsel. Fed. R. Civ. P. 23(c)(1) advisory committee’s note to 2003 amendment. Accordingly, the amendment reflects a delicate balance: a certification decision must be careful and well-informed, yet it also must not be unjustifiably delayed. See id.

23 the requirements for it.” Id. at 177. 10 And, in another case decided that same term, the Supreme Court succinctly reiterated: judges must decide certification questions early “to assure that members of the class [are] identified before trial on the merits and [] bound by all subsequent orders and judgments.” Am. Pipe & Constr. Co., 414 U.S. at 547.

Admittedly, the necessary “rigorous” certification analysis may “entail some overlap with the merits of the plaintiff’s underlying claim.” Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 351 (2011); see also In re Hydrogen Peroxide, 552 F.3d at 316–18. But “[m]erits questions may be considered to the extent—[and] only to the extent—that they are relevant to determining whether the Rule 23 prerequisites for class certification are satisfied.” Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 466 (2013).

Seven Courts of Appeals have gone further and held that, for Rule 23(b)(3) actions, “Rule 23 requires class certification prior to a trial on the merits.” Peritz v. Liberty Loan Corp., 523 F.2d 349, 353 (7th Cir. 1975) (emphasis

10 Although Eisen and other cases cited below predate Rule 23(c)(1)(A)’s 2003 amendment, that amendment did not mean “to alter the standard itself[,] so . . . prior caselaw remains instructive.” Mary Kay Kane, 7AA Federal Practice and Procedure § 1785.3 (3d ed. 2019). 24 added). 11 Although our Court and three other Circuits have not gone quite that far, and have occasionally blessed

11 See also Danny B. ex rel. Elliot v. Raimondo, 784 F.3d 825, 838 (1st Cir. 2015) (“[W]e are aware of no precedent authorizing a district court, over objection, to conduct a full-blown trial on the merits without pausing to take up a timely motion for class certification. . . . The bottom line is that staging a case in this manner puts the cart before the mule.”); Philip Morris v. Nat’l Asbestos Workers Med. Fund, 214 F.3d 132, 135 (2d Cir. 2000) (declining to “foreclose the possibility of a post-trial class certification in another case,” but noting “it is difficult to imagine cases in which it is appropriate to defer class certification until after decision on the merits,” especially for (b)(3) classes like that one, which “certainly” required pretrial certification); Nance v. Union Carbide Corp., 540 F.2d 718, 723 n.9 (4th Cir. 1976) (noting “[t]he language of Rule 23(c) makes it quite clear that the determination of class status is to be made ‘before the decision on the merits’” (quoting Peritz, 523 F.2d at 354)), vacated on other grounds by 431 U.S. 952 (1977); Paxton v. Union Nat’l Bank, 688 F.2d 552, 558 (8th Cir. 1982) (“The subsequent decision to delay certification until after the trial was completed, notwithstanding the apparent acquiescence of the parties, ‘is directly contrary to the command of subdivision (23)(c)(1) . . . .’” (footnote omitted) (quoting Eisen, 417 U.S. at 178)); Horn v. 25 a trial-before-certification approach, we have cabined it to cases in which the defendant consents. See Katz v. Carte Blanche Corp., 496 F.2d 747, 762 (3d Cir. 1974) (en banc) (“[W]e are dealing only with the defendant who declines the protection against one-way intervention . . . which [R]ule 23(b)(3) was designed to afford.”). 12 We have consistently spurned a forced trial-before-certification

Associated Wholesale Grocers, Inc., 555 F.2d 270, 273 (10th Cir. 1977) (“[D]elay in making a decision on certification of the class until after the trial on the merits appears to be a procedure which is not in harmony with the literal terms of Rule 23(c)(1) or with many of the cases.”); Cohen v. Off. Depot, Inc., 204 F.3d 1069, 1078 (11th Cir. 2000) (“Certification under Rule 23(b)(3) would require that the class members receive notice of the suit ‘well before the merits of [it] are adjudicated.’” (quoting Schwarzschild v. Tse, 69 F.3d 293, 295 (9th Cir. 1995) (alteration in original))). See generally William B. Rubenstein, 3 Newberg on Class Actions § 7:11 (5th ed. 2013) (“[C]ourts have generally held that class certification is inappropriate following a trial on the merits, at least in cases adjudicated under Rule 23(b)(3).”). 12 See also Floyd v Bowen, 833 F.2d 529, 534–35 (5th Cir. 1987); Wright v. Shock, 742 F.2d 541, 544 (9th Cir. 1984); Postow v. OBA Fed. Sav. & Loan Ass’n, 627 F.2d 1370, 1383 (D.C. Cir. 1980).

26 procedure: “If a class action defendant insists upon early class action determination and notice, he is, under the rule, entitled to it.” Id.

At all events, no Court of Appeals has approved foisting trial-before-certification on an unwilling Rule 23(b)(3) defendant. Yet that is precisely what the District Court’s planned FLSA trial threatened to do to Citizens. 13

Failure to rule on class certification creates “an atmosphere of confusion.” Philip Morris, 214 F.3d at 134. And we think any confusion is compounded when what is

13 The District Court conceivably could have determined that the planned FLSA collective action trial on whether “Citizens Bank had a policy or practice that caused [MLOs] to not report all of the hours they worked,” JA 255, somehow would not impinge on the factual issues in the Rule 23 class action. But the District Court failed to do so. That court did not provide us with any reasoning to consider, and we discern no basis for it to have reached such a conclusion. Rather, it appears that the planned FLSA collective action trial would have addressed the primary merits issue pertaining to both the FLSA and the state-law class action claims. See Mand. Pet. 1–2 (“The federal and state-law claims are substantively identical; the only difference is procedural.”).

27 scheduled for trial is a hybrid wage-and-hour case like this one. 14

14 The District Court appears to have interpreted our decision to decline the exercise of pendent appellate jurisdiction as somehow blessing its decision to proceed with an FLSA trial before deciding Rule 23 class certification. See Dkt #283 at 5–6 (Mem. Op., June 25, 2019) (denying reconsideration of certification of the FLSA collective action). But that was hardly the case. Our decision on pendent appellate jurisdiction concerned only the limited issue of whether the District Court’s certification of the FLSA collective was “inextricably intertwined” with the certification of the Rule 23 class. See Reinig, 912 F.3d at 131 (“[T]he question of whether the potential plaintiffs had met the FLSA’s less burdensome ‘similarly situated’ standard was ‘quite distinct from the question whether plaintiffs had satisfied the much higher Rule 23 predominance threshold.’” (quoting Myers v. Hertz Corp., 624 F.3d 537, 555–56 (2d Cir. 2010) (cleaned up))). We did not—and indeed, given our limited jurisdiction over only the specific question of class certification, could not—consider whether trial on the FLSA collective action should proceed prior to class certification. Moreover, to be clear, a decision that the two certification decisions are not inextricably intertwined because the legal standards differ does not address 28 If the jury in the FLSA collective action trial had gone on to find that Citizens in fact had “a policy or practice that caused [MLOs] to not report all of the hours they worked,” JA 255, as plaintiffs allege, Citizens’ FLSA trial loss on this issue would have greatly enhanced the probability of its being liable under both the FLSA and the state wage-and-hour laws. The 350 successful FLSA plaintiffs would have proceeded to a trial on their damages claims. And by then, a jury’s finding that was favorable to the FLSA plaintiffs would have aired all the evidence about the existence of a common policy of underpaying all MLOs. 15 The existence of such a common underpayment

whether the two proceedings turn on a common issue of fact. 15 We assume that the jury’s factfinding in the FLSA trial would preclude a second trial on the identical factual issue in the Rule 23 class action. And we further assume that is what the District Court had in mind in its “case management” rulings. But because we do not have the benefit of the District Court’s reasoning, we can only speculate about how the District Court envisioned the Rule 23 class action trial would play out. Did the District Court intend that a second, class-action-specific trial would be held? If so, holding a second trial for the class action creates tension with the Seventh Amendment, which “requires that, when a court bifurcates a case, it must 29 policy may well have satisfied Rule 23(b)(3)’s requirement that “questions of law or fact common to class members predominate.” Thus, an FLSA trial reaching the conclusion that Citizens did have a policy of underpaying all MLOs would, effectively, both identify and determine the merits of the class members’ common factual question. 16

‘divide issues between separate trials in such a way that the same issue is not reexamined by different juries.’” In re Paoli R.R. Yard PCB Litig., 113 F.3d 444, 452 n.5 (3d Cir. 1997) (quoting In re Rhone-Poulenc Rorer, Inc., 51 F.3d 1293, 1303 (7th Cir. 1995)); see also Katz, 496 F.2d at 762 (observing that a defendant who elects a certification-after-trial procedure would be deemed to have waived the Seventh Amendment guarantee of a unitary trial before a single jury on all issues). 16 One serious impediment to certifying a class after an FLSA trial is Rule 23(b)(3)’s requirement of superiority. How can a district court conclude that “a class action is superior to other available methods for fairly and efficiently adjudicating the controversy” if an FLSA collective action trial has already decided the central question posed by the class action? As a practical matter, what work is left for the class action device to do? 30 If a Rule 23 class were then certified after a plaintiffs’ verdict in an FLSA trial, wouldn’t the most likely result be that the remaining 700 or so MLOs would decline to opt out of the class? See Fed. R. Civ. P. 23(c)(2)(B). What would be their incentive to do otherwise? An FLSA trial win would essentially guarantee class certification, given that the same evidence of Citizens’ policies would be offered to support the state law claims. It would seem that, with little left to be resolved factually, the 700 remaining class members would quickly seek to recover damages as well.

On the other side of the coin, a victory for Citizens in the FLSA collective action would bind only the 350 or so MLOs who opted into the FLSA collective action, not the 700 or so who chose not to participate. See Collins v. E.I. DuPont de Nemours & Co., 34 F.3d 172, 180 (3d Cir. 1994). Those roughly 700 remaining MLOs would be free to opt out of a (b)(3) class and pursue individual lawsuits. In other words, an FLSA trial victory could thwart Citizens’ hope for global peace via a class action, leave Citizens with a substantial amount of litigation remaining on its plate, and greatly increase Citizens’ litigation costs and potential liability exposure. As Citizens puts it, the planned FLSA trial before a class certification ruling “would arbitrarily deprive Citizens of the benefits of the class action device to which it is entitled under Rule 23— namely, the full preclusive effect of the class action judgment.” Mand. Pet. 33. 31 All of this leads us to consider: if the practice of holding an FLSA trial before determining Rule 23 class certification were to become the norm in hybrid wage-and- hour suits, most employees likely would never opt in to an FLSA action. Why should they? The federal and state remedies overlap, and we disfavor double recovery. Cf. Rana v. Islam, 887 F.3d 118, 123 (2d Cir. 2018). There would be no incentive for plaintiffs to opt in to the FLSA collective action. They would be better off remaining on the sidelines while the FLSA trial proceeds. If a handful of participating plaintiffs then succeed at the FLSA trial, the potential plaintiffs who are looking on could simply wait for a state-law class to be certified under Rule 23. And, by waiting, those plaintiffs would avoid any risk of being bound by an unfavorable judgment in the FLSA action. That scenario, in short, is a win-win for the employees and disadvantages Citizens at every turn.

Because we are addressing the reasons behind our grant of a stay motion, we need not definitively resolve the various difficulties that arise from the District Court’s chosen approach, including whether its “order of battle” violates Rule 23 or creates a substantial danger of unfair one-way intervention. Nor must we determine whether class certification is even possible after a trial on the merits has already taken place. Moreover, we need not decide whether there could be circumstances under which an FLSA trial might reasonably proceed prior to class certification in a parallel Rule 23 action. What we do 32 conclude here is that, by compelling the FLSA opt-in collective action trial before deciding Rule 23 class certification—in contravention of our clear instruction to conduct a rigorous examination of the class certification issue and without assessing any of the procedural complexities we have discussed—the District Court elected to forge ahead, thereby creating a predicament for others to unravel. We thus conclude that Citizens had a reasonable probability of successfully showing that the District Court clearly and indisputably erred.

B. Mandamus is the only available remedy

Given the reasonable probability that Citizens could successfully show that the District Court erred, mandamus was the only avenue with the remedial force to address the District Court’s error. Citizens could not pursue an immediate appeal of the District Court’s order scheduling the FLSA collective action trial because that order is not a final decision under 28 U.S.C. § 1291. See Morton Int’l, Inc. v. A.E. Staley Mfg. Co., 460 F.3d 470, 476 (3d Cir. 2006) (“The classic definition of a ‘final decision’ is one that ‘ends the litigation on the merits and leaves nothing for the court to do but execute the judgment.’” (quoting Quackenbush v. Allstate Ins. Co., 517 U.S. 706, 712 (1996))). The collateral order doctrine also does not permit an immediate appeal, as the District Court’s procedural approach did not conclusively resolve a

33 disputed issue that is completely separate from the merits. See Cipollone v. Liggett Grp., Inc., 785 F.2d 1108, 1117– 18 (3d Cir. 1986).

Waiting to review the issues presented in Citizens’ mandamus petition until the proceedings became final as to all causes of action and all parties for purposes of § 1291 simply would not protect the interests at stake. Allowing the District Court to conduct its planned FLSA collective action trial would publicly preview the evidence common to the FLSA and state-law claims. That would give potential Rule 23 class members an enormous informational advantage in any subsequent “do-over,” even if we were to ultimately vacate the FLSA verdict and remand for a pretrial Rule 23 class certification decision. In short, an appeal that comes too late can almost never unscramble the egg. 17

17 Although Citizens could have pursued a permissive interlocutory appeal under 28 U.S.C. § 1292(b), we do not fault it for not going that route. Parties need not attempt to proceed under § 1292(b) if it is “sufficiently clear that the District Court would have refused.” In re Briscoe, 448 F.3d 201, 212 n.7 (3d Cir. 2006). Here, the District Court ruled that a stay of trial pending a decision on the Rule 23 certification would have unduly delayed trial. That decision suggests that the District Court would not be 34 2. Citizens Would Have Suffered Irreparable Injury Absent a Stay

Citizens petitioned for mandamus relief less than three weeks before the FLSA collective action trial was to begin. As already discussed, the trial would have irreparably injured Citizens by airing evidence pertaining to Citizens’ liability. So if we were to intervene at all, we had to do so before trial commenced. Yet we needed more than three weeks to consider the petition. A stay preserved the status quo and avoided the impending harm to Citizens.

3. A Stay Did Not Substantially Injure Plaintiffs

The stay did not substantially injure the MLOs. If they ultimately prevail and recover damages in the form of unpaid overtime, their damages either held constant during the stay period (for former employees) or kept growing (for current employees), and they may seek prejudgment interest as well. And if Plaintiffs are ultimately

persuaded that resolution of the issue would “materially advance the ultimate determination of the litigation,” as required by § 1292. See In re Sch. Asbestos Litig., 977 F.2d 764, 793 (3d Cir. 1992). 35 unsuccessful, the delay makes no difference at all. From a damages perspective, Plaintiffs emerge no worse off.

4. The Public Interest Did Not Weigh Against a Stay

The public interest did not weigh against a stay. If anything, because we have now had the opportunity to address the difficulties of prosecuting both an FLSA opt- in collective action and a Rule 23 opt-out class action, the public, including potential litigants, and the practicing bar benefit from the foregoing discussion of the pitfalls created by the District Court’s decision to proceed with an FLSA trial before ruling on a Rule 23 motion for class certification.

IV.

Having explained our decision to stay the case pending resolution of the mandamus petition, a stay is no longer necessary. Accordingly, we will dissolve the stay, dismiss Citizens’ mandamus petition in part as moot given the District Judge’s withdrawal, and refer the matter to the Chief Judge of the Western District of Pennsylvania for reassignment.

Given the impending reassignment, we will deny the remainder of the petition as unnecessary at this time. We expect both parties to work with the newly assigned District Judge to resolve this case fairly and expeditiously, 36 including by carefully considering the numerous procedural complexities of this hybrid action.