Luanna Scott v. Family Dollar Stores, Inc.

WILKINSON, Circuit Judge,

dissenting:

I cannot join the majority’s decision, because it fails to respect the two other levels of the federal judiciary, namely the Supreme Court and the district courts. First as to the Supreme Court. The decision is Wal-Mart Stores, Inc. v. Dukes, — U.S.-, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), and the majority opinion has drained it of meaning. The defendant here, as in Wal-Mart, relies on what plaintiffs admit are multitudinous, discretionary decisions by middle and lower management, which would seem to render class action treatment under Wal-Mart impermissible and ineffectual. Notwithstanding this, the majority has unloaded on the *120district court the prospect of a massive, nationwide class action whose administrability would in all likelihood prove impossible.

In the majority’s view, Wal-Mart applies only where decisions are left to the complete discretion of low-level managers, maj. op. at 117, and are implemented on an “individual store level.” Id. The fact that a company delegates extensive discretion to 95 vice presidents and 400 district managers, Appellee’s Br. at 3 (citing Grace v. Family Dollar Stores, Inc. (In re Family Dollar FLSA Litig.), 637 F.3d 508, 510 (4th Cir.2011)), does not, in the majority’s view, bring this case within the ambit of Wal-Mart and still permits nationwide class action treatment. The majority assumes that nearly 500 middle managers somehow all exercise their discretion in lockstep. That cannot be. The fact that some middle managers would promote from within, and others recruit from without, as they are given the discretion to do, does not, in the majority’s view, preclude nationwide class action treatment. The fact that many managers would elevate women from either inside or outside the company, as they are perfectly free to do, would hardly seem discriminatory, but it would be contrary to the commonality Wal-Mart requires for a nationwide class action to proceed.

The majority responds to this point by citing the fact that exceptions to corporate salary ranges may be granted by a corporate vice president. Maj. op. at 117. But this fact only confirms the assertion that placements within the ranges are determined by middle managers. The fact that exceptions to corporate limits on raises are made by regional managers and senior vice presidents is similarly unavailing to the majority’s position — regional managers, by definition, do not make decisions on a national level. In the majority’s view, middle managers at Family Dollar are purely robotic with respect to those they supervise, but no American company operates in such a way.

The majority plainly believes Wal-Mart does not apply to middle managers exercising delegated discretion under guidelines such as these because if it believed Wal-Mart applied, the district court’s denial of nationwide class certification would be promptly affirmed. The majority’s insistence that Wal-Mart does not apply to middle management (but only to lower-level store managers) suggests not so subtly that it wants this class to be certified. But the commonality Wal-Mart insists is necessary for class action certification is plainly absent here, though the majority purports to find it in some centralized policy. The fact that a company sets pay ranges or values prior experience or performance as factors in compensation is not sinister. Vast numbers of companies do just that. A policy with an obvious business justification may occasionally produce some statistical disparity nationwide. But Wal-Mart makes clear that the fact that a policy may have some statistical disparity nationwide does nothing to dispel the fact that in many districts, the policy will not have a statistical imbalance, but indeed may work to the decided advantage of the putative class. 131 S.Ct. at 2555.

The policies cited by plaintiffs are not “built-in headwinds,” maj. op. at 116 (internal quotation marks omitted), but rather common management techniques that make common sense. If centralized delegations of discretion such as these are enough for a nationwide class action to get rolling, then few companies will be exempt. The law is punishing companies for nothing more than being companies, which is apparently the new status offense.

In reaching its decision, the majority faults the district court for denying plain*121tiffs leave to amend their complaint. But if this is an abuse of discretion, and these findings are clearly erroneous, then class action litigation will almost never end. Not content with finding the district court “abused its discretion,” maj. op. at 119, the majority holds its factual findings “clearly erroneous” as well. Id. at 117. The district judge should be commended, not condemned. The amended complaint severely prejudiced the defendants by forcing them to defend a wholly different suit three years after the original complaint was filed. The amended complaint contradicted assertions in the original complaint to such an extent as to do violence to the values of forthrightness and fair dealing that the district court had every right to expect from the litigants before it. It was also every bit as irreconcilable with the Supreme Court’s decision in Walr-Mart as the original, making denial of leave fully justifiable on futility grounds.

In sum, the district court has been brought up short and found to have abused its discretion for doing nothing more than faithfully following a Supreme Court decision and for attempting to ensure a small measure of candor and consistency in the filings of that court. It is our obligation to respect the Supreme Court’s preeminent place in a hierarchical judicial system, as well as the trial court’s discretion and experience in matters explicitly entrusted by both logic and precedent to its competence. This decision does neither.

I.

Federal Rule of Civil Procedure 15(a)(2), which governs pretrial requests for leave to amend, advises that “[t]he court should freely give leave when justice so requires.” The Supreme Court has accordingly required some “justifying reason” in support of the rejection of a party’s request to amend. Foman v. Davis, 371 U.S. 178, 182, 83 S.Ct. 227, 9 L.Ed.2d 222 (1962).

Nevertheless, the Supreme Court has repeatedly recognized that “the grant or denial of an opportunity to amend is within the discretion of the District Court.” Id,.; see also Krupski v. Costa Crociere S.p.A., 560 U.S. 538, 130 S.Ct. 2485, 2496, 177 L.Ed.2d 48 (2010). Denying leave to amend is appropriate when at least one of three circumstances exists: (1) “the amendment would be prejudicial to the opposing party;” (2) “there has been bad faith on the.part of the moving party;” .or (3) “the amendment would have been futile.” Laber v. Harvey, 438 F.3d 404, 426-27 (4th Cir.2006) (internal quotation, marks omitted). For the reasons that follow, it is abundantly clear that the district court was justified in denying plaintiffs’ motion for leave on all three grounds — prejudice, bad faith, and futility.

II.

A.

As to the first ground, “[wjhether an amendment is prejudicial will often be determined by the nature of the amendment and its timing.” Laber v. Harvey, 438 F.3d 404, 427 (4th Cir.2006). With respect to the amendment’s nature, “[a] common example of a prejudicial amendment is one that ‘raises a new legal theory that would require the gathering and analysis of facts not already considered by the defendant.’ Id. (quoting Johnson v. Oroweat Foods Co., 785 F.2d 503, 510 (4th Cir.1986)) (alterations omitted). By contrast, “[a]n amendment is not prejudicial ... if it merely adds an additional theory of recovery to the facts already pled.” Id.

The majority acts as a cheerleader for the amended complaint, glossing over its gross incompatibility with the original and casually dismissing the threat of prejudice *122as “overstated.” Maj. op. at 118-19. A comparative analysis of the two complaints makes recognition of the night-and-day differences between them unavoidable. The majority’s statement that appellants do not allege a new theory, id. at 118, finds support neither in the record nor in the law. The text of the two complaints speaks— nay, screams — this conclusion for itself.

1.

At its core, the original complaint attacks Family Dollar for maintaining a supposedly subjective and decentralized decision-making structure for determining store manager compensation, which plaintiffs alleged produced illegal discrepancies between male and female pay. A crucial paragraph, in particular, levels the following accusation with great force:

Defendant’s pay decisions and/or system includes subjectivity and gender stereotyping that causes disparate impact to compensation paid to female store managers. Plaintiffs are aware, at this time, of no other criteria which causes such disparate impact other than gender bias, subjectivity and stereotyping. Plaintiffs are unaware, at this time, of any other specific criteria that are capable of separation and analyses.

Compl. ¶ 22 (emphases added).

The import of that paragraph is crystal clear: according to plaintiffs themselves, any actionable discrimination derived solely from “subjectivity and gender stereotyping” — nothing less, nothing more.1 Where subjectivity and gender stereotyping translate directly into discriminatory employment outcomes for a nationwide group of employees (as alleged here), the contested decisions must necessarily have occurred outside the corporation’s core. Any centralized employment policy — even if rooted in the prejudicial predilections of a particular officer or group of officers— could result in generally unfavorable consequences for plaintiffs only if implemented through objective standards, such that the lower-level decision-makers who determine individual store managers’ salaries have little personal power to deviate from the commands dictated by corporate headquarters. But plaintiffs’ complaint was that lower-level managers had too much discretion to deviate, not too little.

Nor does the original complaint specify any other aspect of Family Dollar’s compensation policies as a source of plaintiffs’ injury. In light of prior litigation involving Family Dollar, it should come as no surprise that the original complaint is rooted exclusively in allegations of permissive “subjectivity and gender stereotyping.” In a previous suit brought by plaintiffs’ counsel against Family Dollar, for instance, plaintiffs (some of whom are also parties to the instant action, Appellee’s Br. at 4) alleged that “[djespite [gender-based disparities in pay, Family Dollar] continues to allow its District Managers to subjectively decide what a Store Manager should earn.” Opponent’s Responsive Submission in Resp. to Ex. B of the Ct.’s Order at 12, Collins v. Family Dollar Stores, Inc., No. 7:04-cv-00553-VEH (N.D.Ala. Nov. 17, 2006), ECF No. 235.

Plaintiffs seek to avoid the thrust of their original complaint by clinging to a single sentence repeated (with immaterial variations) several times in their original complaint — that “[defendant engages in centralized control of compensation for *123store managers at the corporate level of its operations.” Compl. ¶¶ 18, 37, 46, 53. This uninformative bit of boilerplate seeks to subject corporations to nationwide class actions by virtue of their mere existence. Plaintiffs’ reasoning in this respect would penalize a company for little more than operating on a national scale under the same corporate name. Even if taken as true, the fact that some centralized directive comes from some corporate headquarters is entirely unremarkable. Surely corporations of national scope cannot flourish in the modern economy without some “centralized control of compensation” for their many thousands of employees. At the very least, corporate headquarters must allocate resources and articulate certain general policies to guide regional or other mid-level managers in setting individual salaries and wages.

The alternative would operate to inhibit the most basic tools of management and result in budgetary chaos. The question, therefore, is how much “centralized control of compensation” the original complaint actually alleges with respect to the challenged employment decisions. The answer is, clearly, not much. If this bare, conclusory statement in the complaint is given weight, then nationwide class action suits are off and running, notwithstanding Wal-Mart and the pleading standards laid down in Ashcroft v. Iqbal, 556 U.S. 662, 129 S.Ct. 1937, 173 L.Ed.2d 868 (2009), and Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 127 S.Ct. 1955, 167 L.Ed.2d. 929 (2007).

2.

The amended complaint, in stark contrast to the original, pivots 180 degrees to assert that Family Dollar’s compensation scheme actually operates in an objective and centralized manner. The amended complaint is not, as the majority contends, a “mere[] elaboration]” on the original. Maj. op. at 118. It is what the district court says it is: a bald attempt to assert a completely new theory. It backtracks on the earlier assertion that the flaw in Family Dollar’s compensation scheme was too much decentralized, subjective decision-making, by alleging that the system “requires pay to be set by uniform, company-wide criteria.” Am. Compl. ¶ 29. The complaint now decides to challenge the purported lack of subjectivity inherent in the company’s supposedly centralized compensation scheme — the very subjectivity that plaintiffs had earlier insisted was the hallmark of Family Dollar’s corporate structure.

As for plaintiffs’ allegations concerning the implementation of Family Dollar’s compensation criteria, the following passage is typical:

Store Managers’ compensation is not set by managers who have unfettered discretion to use their own judgment without regard to any corporate-imposed criteria or standards. All Store Managers’ salaries ... are subject to the same corporate-administered pay system and policy established by corporate headquarters; all Store Manager’s salaries are subject to store payroll budgets established at corporate headquarters; and all Store Managers have the same job description which sets forth a common set of duties and responsibilities regardless of location. There is no policy against having uniform employment practices at Family Dollar.

Id. ¶ 32. We are now explicitly told, moreover, in a complete about-face from the original complaint, that “Family Dollar is not operated in a decentralized, subjective manner. Nor is the pay-setting process for Store Managers based on decentralized, subjective decisionmaking.” Id. ¶ 34. The demons in the original complaint were *124those runaway lower-level managers. The demon in the amended complaint is a controlling “corporate headquarters.” Id. ¶ 35.

3.

Given all of the foregoing, it should be plain that the amended complaint is not some mere modification of the original, as the majority contends. Instead, it is manifestly, substantively different from the original. The two are utterly irreconcilable. They describe two different companies. By transforming their claims from a frontal assault on an excessively subjective and decentralized compensation system into an intricate attack on a purportedly objective and centralized scheme, plaintiffs have done far more than “raise[] a new legal theory.” Laber, 438 F.3d at 427 (internal quotation marks omitted).

The majority breezily dismisses these concerns, asserting in conclusory terms that “[t]he legal theory remains the same.” Maj. op. at 118. My colleagues would be wise to pay some modest heed to the opinion of the district judge, who was better situated to evaluate the actual implications of the transfigured complaint. The new complaint, by virtue of its novel allegations, would require significant “gathering and analysis of facts not already considered by the defendant.” Laber, 438 F.3d at 427 (internal quotation marks and alterations omitted). As the district judge emphasized: “Plaintiffs wish to pursue extensive discovery to support and clarify then-new theories, which will require the parties to re-open and conduct new expert discovery based on plaintiffs’ changed version of the facts.” J.A. 418.

Thus, the district court was correct to conclude that granting leave to amend would be prejudicial to Family Dollar. Id. at 417-18. “The proof required to defend against this new claim would be of an entirely different character than the proof which the defendant [was] led to believe would be necessary. Belated claims which change the character of litigation are not favored.” Deasy v. Hill, 833 F.2d 38, 42 (4th Cir.1987). The district court acted well within its discretion by denying plaintiffs’ motion for leave to amend.

B.

The timing of a proposed pleading amendment also bears on whether the change would prejudice the opposing party. Laber, 438 F.3d at 427. In particular, whereas an amendment “offered before any discovery has occurred” is unlikely to be prejudicial, “the further [a] case [has] progressed ..., the more likely it is that the amendment will prejudice the defendant.” Id.; see also United States ex rel. Nathan v. Takeda Pharm. N. Am., Inc., 707 F.3d 451, 461 (4th Cir.2013) (affirming denial of motion for leave to amend in light, inter alia, of a two-year gap between filing of complaint and dismissal); Mayfield v. Nat’l Ass’n for Stock Car Auto Racing, Inc., 674 F.3d 369, 379 (4th Cir.2012) (affirming denial of motion for leave to amend where “a significant amount of discovery had already been conducted”).

Here, plaintiffs’ attempt to amend their complaint came three years after the case was initially filed — and only when Family Dollar appeared poised to succeed on its motion to dismiss and/or strike the original complaint’s class allegations. Moreover, the district judge observed that plaintiffs had been given “adequate time to conduct discovery,” that they had in fact “conducted significant discovery,” and that plaintiffs’ own counsel had even admitted “that discovery is mostly completed.” J.A. 414-15. Hence, the district court concluded that any “additional discovery would be ... prejudicial to defendant.” Id. at 415. The court proceeded to hold that:

*125[Allowing plaintiffs to amend the complaint would prejudice defendant. Since the filing of the complaint three years ago, the parties have pursued discovery ... and have attempted to mediate claims under the original complaint. Here, plaintiffs chose not to file their proposed amended complaint until the briefing on defendant’s motion to dismiss was nearly complete.... Plaintiffs wish to pursue extensive discovery to support and clarify their new theories, which will require the parties to re-open and conduct new expert discovery based on plaintiffs’ changed version of the facts.

Id. at 417-18.

Plaintiffs attempt to blame the three-year delay in filing for leave to amend on the various motions and objections that were exchanged between the parties during the course of discovery. Such tit for tat, however, is not peculiar to this litigation; every complex class action of this variety will have just this sort of pretrial motion exchange. The majority’s adoption of plaintiffs’ reasoning in this respect, maj. op. at 117-18, thus comes close to establishing a per se three-year grace period for motions for leave to amend. Such a protracted interval is excessive and susceptible to manipulative conduct. The new rule established by today’s opinion endorses filing delays that patently prejudice opposing parties.

Moreover, the delay in this particular case is especially unjustifiable. Plaintiffs’ counsel have extensive experience with defendant’s corporate structure: by their own admission, they have sued Family Dollar over labor and employment matters “approximately 15” times since 2001. See Pis.’ Reply to Def.’s Resp. to Pis.’ Opp’n to Terry Price Serving as Local Counsel and Req. for Emergency Hr’g at 3 n.2, Scott v. Family Dollar Stores, Inc., No. 3:08-cv-00540-MOC-DSC (W.D.N.C. Nov. 4, 2008), ECF No. 15. As the district court noted, plaintiffs were plenty familiar through their multiple prior lawsuits with defendant’s corporate organization. J.A. 417. Although plaintiffs assert in a conclusory footnote that defendant’s compensation policies have changed since the time of these many prior suits, Appellants’ Br. at 53 n.7, they provide no substantiation for this claim nor do they identify any specific ways in which the policies have been altered.

The majority inexplicably focuses on the fact that plaintiffs’ motion for leave to amend was made prior to trial. Maj. op. at 118-19. The crux of this dispute, however, is class certification. That issue is routinely decided pretrial. See Fed. R.Civ.P. 23(c)(1)(A) (“At an early practicable time after a person sues or is sued as a class representative, the court must determine by order whether to certify the action as a class action.”). Any potential source of prejudice, therefore, lies not in inconveniences at trial but rather in the superfluous or additional discovery costs imposed on defendant as a result of plaintiffs’ fluctuating class action theories. If the majority’s misplaced emphasis on trial represents a new standard for identifying prejudice in class certification proceedings, prejudice will almost never be found.

I see no reason whatsoever to usurp the district court’s essential case management functions or to question the accuracy of its characterizations. It was entirely proper for the district court to conclude that permitting plaintiffs to amend their complaint so substantially and at such a late stage of the game would impermissibly prejudice Family Dollar. It was altogether sound for the trial court to hold that Family Dollar should not be forced to defend a new suit three years after the original complaint was filed. See Newport News *126Holdings Corp. v. Virtual City Vision, Inc., 650 F.3d 423, 439-41 (4th Cir.2011) (affirming on prejudice grounds denial of motion for leave where amendment was filed at the “eleventh hour,” “would probably have necessitated additional discovery,” and would have “substantially change[d] the nature and scope” of litigation) (internal quotation marks omitted); Equal Rights Ctr. v. Niles Bolton Assocs., 602 F.3d 597, 603-04 (4th Cir.2010) (same).

III.

As to why plaintiffs wanted to undertake such an extensive overhaul of their complaint in the first place, the majority opinion points to the Supreme Court’s decision in Wal-Mart, 131 S.Ct. 2541. The majority, however, misapprehends the import of Wal-Mart with respect to the final two grounds on which a district court may deny a motion for leave to amend a pleading — the lack of good faith, to which I now turn, and futility, discussed in Part IV. See Laber v. Harvey, 438 F.3d 404, 426-27 (4th Cir.2006).

A district court’s refusal to permit a pleading amendment on bad faith grounds is justified where “the plaintiffs first theory of recovery is based on his own reading of ... cases and it turns out that he misinterpreted how that theory would apply to the facts of his case.” Id. at 428 (emphasis omitted). That situation is precisely what occurred here. Plaintiffs misinterpreted how certain class action precedents would apply to their case and then sought to construct an entirely new set of facts to overcome their error. Their willingness to adopt contradictory factual positions in order to match their evolving legal theories evidences a degree of bad faith sufficient to warrant denial of leave to amend. To the old-fashioned view that prior representations to a court actually count for something, the majority answers: Not much.

Plaintiffs were wholly content to ride the coattails of the proposed class in WalMart while that class was enjoying success in the lower federal courts. In consenting to a transfer of venue in 2008, plaintiffs explicitly stated that, with respect to a then-recent round of the Wal-Mart litigation, “[t]he Ninth Circuit ... affirmed certification of ... a nationwide class having virtually identical claims of sex discrimination in pay to those brought in this case.” J.A. 221 (citing Dukes v. Wal-Mart, Inc., 509 F.3d 1168 (9th Cir.2007)) (emphasis added). Later in the litigation, plaintiffs argued that “[t]he evidence is expected to show that this case is more like ... the Ninth Circuit’s decision in” Wal-Mart than the cases cited by defendant. S.A. 527.

Then plaintiffs adopted a dramatically different stance after the Supreme Court reversed the Ninth Circuit’s certification decision in 2011. See 131 S.Ct. 2541. In their briefing before this court, for instance, plaintiffs contend that “Family Dollar’s salary system is the opposite of that in Wal-Mart,” Appellants’ Br. at 5; that “[t]he current case has never alleged any store-level decisionmaking similar to that in Wal-Mart,” id. at 16, 21-22 (emphasis omitted); that “[t]he Wal-Mart decision was limited to localized decisionmaking within each store that was not subject to any centralized policies or control similar to those alleged here,” id. at 20-21; and that “Wal-Mart simply does not apply to [the] Complaint [here],” id. at 52 (emphasis omitted).

Statements made at oral argument help to illustrate the gross incompatibility between the factual allegations made by plaintiffs’ original and amended complaints. The court inquired: “Don’t we have a big difference ... between your *127complaint and your amended complaint ... in terms of the substantive allegations?” Plaintiffs’ counsel responded: “No, I do not believe so.” He later elaborated:

We say that this case involves centralized criteria ... and that we can show that that centralized criteria is what’s causing the disparity, not ... anything localized.... That’s our complaint from Day 1. If you read our original complaint, it says that we are attacking a centralized system. It says nothing but that.

(emphasis added). Despite these protestations to the contrary, the original complaint actually says precisely the opposite. It states explicitly that “Plaintiffs are aware, at this time, of no other criteria which causes such disparate impact other than gender bias, subjectivity and stereotyping.” Compl. ¶22 (emphasis added).

To be sure, counsel must enjoy latitude in amending complaints to address intervening developments in the law and to incorporate factual material uncovered since the original filing. Some evolution of a plaintiffs approach to a case is to be expected, for good advocacy is adaptive in some measure. It is a matter of degree, however, and the district court was right to spot in plaintiffs’ new attack a bridge too far.

For the instant plaintiffs do not merely present a new legal argument predicated on their original factual allegations, or some modification based upon new revelations. Instead, they seek to invent an entirely new set of facts tailored to their revised theory of recovery. The corporate defendant described in the amended complaint bears no more than a nominal relationship to that described in the original. The proposed amendment is “not merely clerical or corrective. It [establishes] an entirely new factual basis for the plaintiffs’ claims.” Little v. Liquid Air Corp., 952 F.2d 841, 846 (5th Cir.1992), reinstated in relevant part, 37 F.3d 1069, 1073 & n. 8 (5th Cir.1994) (en banc); see also Cornell & Co., Inc. v. OSHRC, 573 F.2d 820, 824-25 (3rd Cir.1978) (denying leave to amend where plaintiff “changed the factual basis for the charge as well as his legal theory”) (internal quotation marks omitted).

This is more than some commonplace doctrinal point. Complaints must bear some relationship to the external reality which they purport to describe. When a corporation is reinvented from one employing decentralized, subjective decision-making to one with rigid, entirely centralized policies, law’s relationship to reality is stretched too thin. See Bradley v. Chiron Corp., 136 F.3d 1317, 1324-26 (Fed.Cir.1998) (disregarding “sham” facts in an amended complaint that contradicted the factual allegations pled in the original and represented “a transparent attempt to conform the facts to the requirements of the cause of action”). Law is not a mere set of expressions to be manipulated toward a given end. It is a system designed to ascertain truth as far as possible in order to produce justice, to the extent possible. To do this, law must maintain some concrete relationship with facts as they exist. Plaintiffs’ contradictory pleadings, which treat reality as a plastic entity to be molded to their purposes, run directly counter to this principle. See Reddy v. Litton Indus., Inc., 912 F.2d 291, 296-97 (9th Cir.1990) (“Although leave to amend should be liberally granted, the amended complaint may only allege other facts consistent with the challenged pleading.”) (internal quotation marks omitted).

Were plaintiffs permitted to substitute contradictory factual narratives every time an intervening opinion cast doubt upon their claims, they could hold defendants hostage by indefinitely postponing final *128judgment. The majority finds the original complaint deficient because it alleged only a subjective decision-making structure. Maj. op. at 115-16. But when plaintiffs sought to run from their prior representations and assert a highly controlled decision-making apparatus, the majority says no problem. I regret that the majority encourages litigants to approach courts in such a manner.

IV.

Finally, the district court’s rejection of plaintiffs’ motion for leave to amend was warranted on a third ground: futility. See Laber v. Harvey, 438 F.3d 404, 426-27 (4th Cir.2006). “Futility is apparent if the proposed amended complaint fails to state a claim under the applicable rules and accompanying standards” — that is, if it “fails to satisfy the requirements of the federal rules.” Katyle v. Penn Nat’l Gaming, Inc., 637 F.3d 462, 471 (4th Cir.2011) (internal quotation marks omitted). WalMart itself expounded “the requirements of the federal rules” — specifically, Federal Rule of Civil Procedure 23’s commands concerning the certification of class actions. It is plain that the amended complaint fails to state a claim by virtue of that decision.

A.

As in Wal-Mart, “[t]he crux of this case is commonality — the rule requiring a plaintiff to show that ‘there are questions of law or fact common to the class.’ ” 131 S.Ct. at 2550-51 (quoting Fed.R.Civ.P. 23(a)(2)). Wal-Mart’s central teaching is that the claims of each class member “must depend upon a common contention.” Id. at 2551. That common contention, in turn, must “be of such a nature that it is capable of classwide resolution — which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.” Id. Thus, “[w]hat matters to class certification ... is not the raising of common ‘questions’ — even in droves — but, rather the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation.” Id. (internal quotation marks omitted).

Applying these principles to employment discrimination claims, the Wal-Mart Court made clear that “[wjithout some glue holding the alleged reasons for [each of the challenged] decisions together, it will be impossible to say that examination of all the class members’ claims for relief will produce a common answer to the crucial question why was I disfavored.” Id. at 2552. As relevant here, plaintiffs must show “ ‘[significant proof that an employer operated under a general policy of discrimination’ ” in order to demonstrate the existence of the requisite “glue.” Id. at 2553 (quoting Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 159 n. 15, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982)) (alteration in original). For two reasons, plaintiffs have failed to satisfy this standard. First, the claims in the amended complaint fail on their face. Second, even if the claims were not facially deficient, the proffered evidence would still be incapable of supporting such claims on a classwide basis.

1.

In light of the stunning similarities between this case and Wal-Mart, the allegations in the amended complaint — just as in the original complaint — are legally insufficient from the outset. In both cases, defendants are large corporations operating nationwide chains of consumer-goods stores. From Wal-Mart:

Petitioner Wal-Mart is the Nation’s largest private employer. It operates four types of retail stores throughout the country.... Those stores are divid*129ed into seven nationwide divisions, which in turn comprise 41 regions of 80 to 85 stores apiece. Each store has between 40 and 53 separate departments and 80 to 500 staff positions. In all, Wal-Mart operates approximately 3,400 stores and employs more than one million people.

131 S.Ct. at 2547. And from the briefing here: “Family Dollar operates a chain of over 7,000 stores in more than 40 states” and “ ‘has divided its operations into 95 regions, each run by a vice president, and then into districts, each run by a district manager.’” Appellee’s Br. at 3 (quoting Grace v. Family Dollar Stores, Inc. (In re Family Dollar FLSA Litig.), 637 F.3d 508, 510 (4th Cir.2011)). There are approximately four hundred districts, each of which includes between ten and thirty stores. Id.

In both cases, the proposed class encompassed many thousands of retail-level, female employees and former employees. In both cases, plaintiffs challenged various pay and promotion decision procedures as improperly gender-related. Compare Wal-Mart, 131 S.Ct. at 2547 (“The named plaintiffs in this lawsuit, representing the 1.5 million members of the certified class, are three current or former Wal-Mart employees who allege that the company discriminated against them on the basis of their sex by denying them equal pay or promotions, in violation of Title VII of the Civil Rights Act of 1964.... ”), with Am. Compl. ¶ 5 (“The plaintiffs bring this action on behalf of themselves and all female Store Managers pursuant to Title VII of the 1964 Civil Rights Act ... and § 216(b) of the Equal Pay Act of 1963 ... to redress the defendant’s widespread and pervasive gender discrimination in employment opportunities.”).

And most significantly, in both cases, all of the contested employment actions derived from the same type of decision-making structure. In Wal-Mart, the Supreme Court stated that mid-level managers were allowed to exercise “discretion” within “limits” imposed and enforced by “corporate oversight,” with such oversight including “preestablished ranges” and “certain objective criteria” for pay and promotions. 131 S.Ct. at 2547. The district court in Wal-Mart provided even greater detail, explaining that “the company maintains centralized corporate policies that provide some constraint on the degree of managerial discretion over in-store personnel decisions.” Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137, 152-53 (N.D.Cal.2004). For instance, “there is a basic compensation structure that applies similarly to all in-store salaried management positions across all types of Wal-Mart stores, in that the computation begins with a base salary within a range set by the corporation ..., with adjustments allowed for profit incentives and/or merit increases.” Id. at 148.

As discussed in Part II, the factual and legal allegations contained in the amended complaint in this case were so novel as to warrant a finding of prejudice. The fact that certain allegations are new, however, does not indicate that they are viable. Here, despite plaintiffs’ efforts to allege extensive centralized control, the amended complaint reveals the existence of a corporate decision-making structure parallel to that described in Wal-Mart. As the district court here explained, “Although plaintiffs [now] purport to deny that class members’ pay is set through a discretionary, subjective process, ... the discretionary pay of managers, within uniformly established parameters, remain[s] the only source of discrimination alleged.” J.A. 417. That pattern of dispersed managerial discretion within centralized parameters is precisely that of Wal-Mart.

*130While plaintiffs fail to so much as identify the source of many of the supposed nefarious corporate parameters, see, e.g., Am. Compl. ¶ 51, even if we were to accept these dubious assertions at face value, plaintiffs’ proffered amendment would still be futile. In an effort to identify a “specific employment practice” responsible for the alleged pay discrepancies, WalMart, 131 S.Ct. at 2555 (internal quotation marks omitted), plaintiffs (and the majority) point to four corporate policies. With respect to each claim, plaintiffs’ own brief gives away the ballgame. First, plaintiffs challenge the corporate-imposed salary ranges for store managers. Appellants’ Br. at 13-14; Am. Compl. ¶ 35. A salary range, however, intrinsically imparts discretion to those charged with administering it. As the district judge noted, “a large number of decision-makers, ... located around the country, exercise individual discretion in placing Store Managers within the established pay ranges.” J.A. 419. Discretion cabined by broad corporate policies — including salary ranges — is precisely the structure that Wal-Mart found not to be susceptible to class action treatment. 131 S.Ct. at 2547 (denying class certification despite defendant’s use of salary ranges). If the existence of such discretion defeated class action commonality in Wal-Mart, it must do so here.

Second, plaintiffs decry the alleged corporate-imposed cap on pay raises and contend that exceptions to this cap, which may only be granted by Regional Managers and Divisional Vice Presidents, are granted disproportionately to males. Appellants’ Br. at 13; Am. Compl. ¶ 36. Regional Managers and Divisional Vice Presidents, however, as their respective titles indicate, are middle managers. See J.A. 419. By definition, they are incapable of dictating corporate-wide policies. As the district judge noted, plaintiffs’ allegations in this respect again converge with the facts in Wal-Mart: both cases involve dispersed decision-makers exercising discretion (e.g., granting exceptions) free of direct corporate control and oversight. Id. at 417, 419.

Third, plaintiffs argue that defendant’s criteria for determining compensation— criteria which include prior experience and performance evaluations — disparately impact women. Appellants’ Br. at 32; Am. Compl. ¶¶ 40, 51. The use of such criteria is hardly remarkable; the only thing that would be remarkable is if Family Dollar failed to find prior experience and prior performance relevant. The business justification for this practice is obvious.

Plaintiffs do not allege, moreover, that these criteria constitute a rigid formula; instead, the criteria appear to be simple guideposts listing multiple factors designed to channel the discretionary decisions of those middle managers charged with setting store manager salaries. It is the business equivalent of a judicial totality-of-the-circumstances test, with the weight and relevance of the factor or circumstance to be determined individually. Indeed, it would be senseless to set rigid salaries for every store manager at corporate headquarters, both because it would strip the system of incentives and because the performance of each manager.simply is not identical. As noted above, this type of broad corporate constraint on what is fundamentally a discretionary determination does not satisfy the commonality requirement. Wal-Mart, 131 S.Ct. at 2547 (denying certification despite defendant’s use of “objective criteria” in making promotion decisions).

Fourth, plaintiffs complain that corporate policies require that store managers promoted from within be paid less than those who are hired laterally. Appellants’ *131Br. at 9; Am. Compl. ¶ 52, 54. This policy allegedly produces a disparate impact insofar as female store managers are disproportionately promoted in-house. Appellants’ Br. at 10; Am. Compl. ¶ 55. Plaintiffs do not allege, however, that either method of selection is centrally mandated, nor do they allege that any central policy is even responsible for the supposed tendency of females to be promoted from within rather than hired laterally.

Given that the alleged policy does not dictate the internal or external route of store manager selection, any disparate impact that arises will necessarily be the result of decentralized choices by middle managers. Whether women are disproportionately hired from within will vary from region to region. In short, “[i]n a company of [Family Dollar’s] size and geographical scope, it is quite unbelievable that all managers would exercise their discretion in a common way without some common direction.” Wal-Mart, 131 S.Ct. at 2555. Consequently, the existence of any disparate impact resulting from this particular policy will be resistant to coherent analysis at the national level.

The business justification for allotting a slight premium to lateral hires is hardly obscure. It may well take such an allowance to persuade an employee to switch companies. Furthermore, the Supreme Court has noted that the mere fact that a business practice produces some statistical disparity is, standing alone, insufficient to conclude that a class action will be viable. Id. at 2555-56. Under the lateral hire policy at issue here, for example, some middle managers will hire women from outside, or promote men from within. In other cases, the hiring party may herself be female. The alleged policy could very well work to the benefit of women in certain districts. In short, the results will vary by district and by region; nationwide patterns are inadequate to justify an inference of discrimination at the subnational level. The variable results produced by this particular practice — which is neutral on its face and supported by an obvious business justification — are precisely what Wal-Mart envisioned as inimical to class action commonality.

The fact that each of plaintiffs’ key claims ultimately reduces to an allegation of cabined discretion should be unsurprising. There is nothing inherently discriminatory about delegated discretion. Companies must rely on delegated discretion. It would be virtually impossible, as a matter of sheer practicality, for a company as extensive in scope as Family Dollar to micromanage store manager compensation via centralized policies. It is simply unfathomable that Family Dollar’s corporate headquarters could afford to dictate the compensation paid to managers in each of its 7,000 stores. Some discretion is intrinsic to this type of national business. See id. at 2554 (noting that an employment policy of decentralized decision-making is “a very common and presumptively reasonable way of doing business”).

The inference is therefore inescapable that Family Dollar relies on middle managers — who have greater and more intimate knowledge of facts on the ground than the members of top management — to attend to the details of store manager compensation within the broad constraints imposed by corporate headquarters. Family Dollar expects its intermediate executives to be more than mere automatons. See Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 990, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988) (noting that “it may be customary and quite reasonable simply to delegate employment decisions to those employees who are most familiar with the jobs to be filled and with the candidates for those jobs”). Plaintiffs’ inventive *132pleadings simply cannot disguise the economic and managerial realities associated with running a national corporation. “[L]ocal discretion cannot support a company-wide class no matter how cleverly lawyers may try to repackage local variability as uniformity.” Bolden v. Walsh Constr. Co., 688 F.3d 893, 898 (7th Cir.2012). The presence of such variability makes it difficult to establish the commonality necessary for class action treatment because, among other things, business managers in many regions and districts will exercise delegated discretion in favor of the plaintiff class.

Plaintiffs’ argument, therefore, continues to boil down to the contention that an “exercise of discretion results in disparities in pay based on gender.” J.A. 419. WalMart, of course, found challenges based on such a decision-making structure largely resistant to class action treatment. 131 S.Ct. at 2555-56. The district court nicely summarized this aspect of the futility analysis when it concluded that “the proposed amended complaint appears to be an attempt to recast plaintiffs’ class claims simply to avoid dismissal under [Wal-Mart ], but even the allegations in the amended complaint ultimately point to subjective, individualized decisions rather than pointing to any uniform company-wide policy that discriminates against [female] Store Managers.” J.A. 417.

2.

Plaintiffs’ amended complaint is deficient for an additional reason: the evidence plaintiffs have offered fails to satisfy the standards suggested by Wal-Mart. As the Court in that case made clear, “Rule 23 does not set forth a mere pleading standard.” 131 S.Ct. at 2551. Rather, “[a] party seeking class certification must affirmatively demonstrate his compliance with the Rule — that is, he must be prepared to prove that there are in fact ... common questions of law or fact, etc.” Id. (emphasis omitted); see also Comcast Corp. v. Behrend, — U.S.-, 133 S.Ct. 1426, 1432, 185 L.Ed.2d 515 (2013). In light of the fact that plaintiffs have already “conducted significant discovery in this and other similar cases against defendant in other jurisdictions,” J.A. 415, the data that they have gathered is inadequate to satisfy the evidentiary standard imposed by Wal-Mart.

Wal-Mart emphasized that “left to their own devices most managers in any corporation ... would select sex-neutral, performance-based criteria for hiring and promotion that produce no actionable disparity at all.” 131 S.Ct. at 2554. Furthermore, with respect to a large national corporation like Wal-Mart or Family Dollar, “it is quite unbelievable that all managers would exercise their discretion in a common way without some common direction.” Id. at 2555. And while WalMart did not foreclose the theoretical possibility that such coordinated, discriminatory, discretionary activity might one day be demonstrated, it concluded that the “statistical and anecdotal evidence” in that case fell “well short.” Id.

Here, the only real evidence that plaintiffs have provided is numerical in nature, and it fails for the same reason as that in Wal-Mart. The amended complaint supplies figures purporting to show “statistically significant disparities in what Family Dollar pays men and women for the same job of Store Managers.” Am. Compl. ¶ 24. Plaintiffs point to an alleged salary gap amounting to approximately $2,500 per year between 2008 and 2010, which they peg at twenty-two to twenty-three standard deviations above “what would be expected in the absence of gender-based discrimination” when “controlling] for non-gender factors that may affect pay such as *133store, district, region, store type, store size, store location, store volume, education and prior work history, and length of service.” Id. ¶¶ 24-27. But this lengthy enumeration of “controlled” elements itself belies plaintiffs’ claim that any alleged discrepancy in store manager pay is the product of a rigid collection of centralized corporate policies.

Plaintiffs’ statistical evidence is insufficient under Wal-Mart on two counts, both stemming from the fact that it is national in scope. First and fundamentally, the Supreme Court specifically underscored the “failure of inference” inherent in attempting to draw particularized conclusions from national statistical data. 131 S.Ct. at 2555. That is, “[information about disparities at the regional and national level[s] does not establish the existence of disparities at individual stores,” or within individual districts, “let alone raise the inference that a company-wide policy of discrimination is implemented by discretionary decisions at the store and district level.” Id. (internal quotation marks omitted); see also Bolden, 688 F.3d at 896 (“If [the defendant employed] 25 superintendents, 5 of whom discriminated in awarding overtime, aggregate data would show that black workers did worse than white workers — but that result would not imply that all 25 superintendents behaved similarly, so it would not demonstrate commonality.”); Bennett v. Nucor Corp., 656 F.3d 802, 815-16 (8th Cir.2011) (“[A] bottom-line [statistical] analysis is insufficient to demonstrate that any disparate treatment or disparate impact present in one department was also common to all the others.”).

Second, nationwide data fails to account for various nondiscriminatory conditions that may have produced divergent results from one area to another. For instance, as Wal-Mart tells us, “[s]ome managers will claim that the availability of women, or qualified women, or interested women, in their stores’ area does not mirror the national or regional statistics.” 131 S.Ct. at 2555. The controls that plaintiffs claim to have factored into their statistical conclusions here do not account for those factors, nor could their crude statistics possibly comprehend the myriad other conceivable circumstances that may affect comparative compensation levels in specific locales.

B.

Plaintiffs have thus failed to provide “convincing proof’ of any policy that discriminates in a “companywide” manner; as a result, “they have not established the existence of any common question.” WalMart, 131 S.Ct. at 2556-57. The amended complaint suffers from the same fatal flaw as the original, rendering plaintiffs’ attempt to reboot the litigation futile, and rendering the district court’s decision to refuse the amendment on that ground an entirely proper exercise of its discretion. Even without reaching the patent inadequacies of the amended complaint under Rule 23(b) — including the obvious further difficulties raised by plaintiffs’ request for backpay in light of the Court’s remedial holding in Wal-Mart, 131 S.Ct. at 2557— the entire class action fails for a lack of commonality under Rule 23(a)(2).

It bears reemphasis that the employment decision-making structure at issue here — in which a business articulates certain centralized policies but also imparts to mid-level managers some discretion to implement them — is not only common to Wal-Mart and Family Dollar. It is typical of most national corporations. See McReynolds v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 672 F.3d 482, 488 (7th Cir.2012) (noting that large corporations may grant discretion to local managers “as a matter of necessity”). The result *134is a substantial variety of outcomes attributable to the disparate management philosophies, priorities, and circumstances of each decentralized decision-maker — exactly what one would expect in a company staffed by human beings.

The majority fails even to suggest why the challenged policies might be legally suspect. Indeed, the corporate guidelines targeted by plaintiffs — such as the use of salary ranges, the purported bonuses for lateral hires, and the inclusion of prior experience and performance as factors in pay decisions — are among the most anodyne in the corporate world. Permitting a class action suit to proceed on such a slender basis exposes a large swath of companies to class-action liability simply for adopting perfectly ordinary, plain vanilla policies. These policies do, however, share one relevant feature: they delegate discretion.

Wal-Mart recognized the difficulty of accounting for regional discrepancies and individual exercises of discretion through the blunderbuss of class action litigation. The gravamen of that decision is that nationwide classes face a steep climb to certification under Rule 23. 131 S.Ct. at 2554 (holding that under the circumstances discussed, “[a] party seeking to certify a nationwide class will be unable to show that all the employees’ Title VII claims will in fact depend on the answers to common questions”). Given the managerial nightmares encountered by district judges assigned these unwieldy pieces of litigation, no other conclusion would be possible.2

It is also important to note that denial of nationwide class certification here would not leave plaintiffs without a path forward. Each could continue to pursue a personal claim of discrimination, as the district court made clear. J.A. 420. Or, should plaintiffs choose to take a different tack on remand, class certification could perhaps be suitable for more modest — and thus more manageable — groups, such as dis*135trict-level clusters, where the differences in the circumstances faced by each member may be less pronounced. See Bolden, 688 F.3d at 899 (denying class certification but suggesting that smaller subclasses might be certifiable). While plaintiffs have chosen to bite off more than they can chew thus far, smaller morsels may prove more palatable in the end.

V.

In holding Wal-Mart inapplicable to the manifold discretionary decisions of middle managers, the majority has hollowed out that case. Moreover, the district court engaged in a sound exercise of discretion on any one of the three grounds commonly recognized as reasons for denying leave to amend. The majority’s decision is unjustifiable under the straightforward application of governing precedent.

In a larger sense, though, the majority’s ruling is more damaging even than the disregard of precedent. It impairs the judicial process in three significant ways. First, it prolongs disputes far past the point of reason. It requires companies to defend completely different cases no less than three years after the filing of the complaint. No other court has gone this far. In so doing, the majority fails to address even the rudimentary managerial realities of modern national corporations. The more’s the pity, because in many places and under many managers, the chief beneficiary would have been the plaintiff class.

Second, the majority pulls up curbside and dumps on the district court an utterly unwieldy, unmanageable piece of litigation. It is a truism that unpleasant tasks roll downhill, and it is also worth the observation that the majority will not have to deal with the many problems it has wrought. We use an abuse of discretion standard in this context for a reason. The district judge is best situated to make the type of determinations at issue on this appeal. See Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 630, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997) (Breyer, J., concurring in part and dissenting in part) (noting in the class action context that a district court “is far more familiar with the issues and litigants than is a court of appeals”). Given the standard, this is a rude reversal, as it would be even for a trial court opinion less well reasoned than the one reversed.

Third, the majority has subverted a Supreme Court decision that, whether congenial or not, was written precisely for a dispute such as this one. We count upon district courts to faithfully apply our decisions and precedents. The Supreme Court should be able to count upon us to do the same.

I yield to no one in my respect for the truly fine judges in the majority. But let this much be clear. Even the above unfortunate consequences pale in comparison to the incentives today’s ruling creates for future parties. The plaintiffs in this case played fast and loose with the district court, offering not an “amended complaint,” but rather a completely contradictory one. They assumed that the allegations in a complaint need bear no discernible relationship to any external reality but reflect only the limitless malleability of lawyers’ verbal skills. The district court recognized that the system was being gamed and moved to instill respect for the integrity of the process over which it had the duty to preside. That we should not only reverse the trial court, but do so as clearly erroneous and an abuse of discretion, is simply wrong.

The abuse was committed on appeal.

. The concurring opinion of my good colleague, which ignores this reality, is notable chiefly for its silences. It advances an analysis even more cursory than that of the majority on the theory that some vague, soothing assurance about ordinary Rule 15 motions will obscure the extraordinary steps that have been taken. Granted, it is in the nature of a concurrence to be brief in relative terms, but surely some revelatory engagement with appellee’s claims should be forthcoming. The concurrence neglects to address which of the district court’s factual findings were clearly erroneous, or which of its discretionary judgments ran afoul of the abuse-of-discretion standard of review. It declines to say exactly what new information was supposedly discovered during mediation, or why that information was not known to plaintiffs’ counsel as a result of their fifteen previous suits against Family Dollar. It fails to justify the irreconcilability of the various pleadings or the changed thrust of the factual allegations contained therein. It neglects to address the district court's finding that this entirely new case severely prejudiced defendant three years after the filing of the original complaint. It refuses to explain why Wal-Mart's commonality holding, by its plain language, does not apply to middle managers. It further refuses to explain why 500 vice presidents and district managers who concededly made discretionary decisions within delegated ranges are anything other than middle management, or why a system in which discretion is channeled by broad corporate guidelines does not fall within Wal-Mart's literal terms. It does not state why it is justifiable to rope regions and districts with progressive hiring practices into nationwide litigation, or how this national class action, with all its disparate and moving parts, is supposed to be administered, or what the district court is even supposed to do upon remand. It fails, finally, to illuminate for courts and litigants why this decision does not subject every company in America with similarly unremarkable policies to the prospect of class-action liability (and the reality of interminable class certification disputes) merely for existing. Perhaps my fine colleagues will some day provide some answers to some of these questions, but for now they are doing what football teams usually do on fourth down.