The opinion of the court was delivered by: Judge John J. Tharp, Jr.
MEMORANDUM OPINION AND ORDER
The plaintiffs in this case, on behalf of themselves and a class, allege that defendant Generation Mortgage Company employs a pricing policy that discriminates against African-Americans and single females. Generation is a "reverse mortgage lender," meaning that it makes loans to borrowers that are not repaid until the home is sold; effectively, Generation makes installment payments on a portion of the equity in a home and recoups its principal, with interest, from the proceeds when the home is sold.
The plaintiffs allege that Generation has a national pricing policy that sets a minimum price for loans but vests its brokers with discretion to charge higher prices and incentivizes them to do so. They claim that Generation's brokers have applied this policy in a way that has resulted in higher interest rates on loans to African-Americans and single females. They also allege, in the alternative, that Generation engages in "reverse redlining." "Redlining" refers to an unlawful practice of denying services (such as offering mortgage loans) to defined ("redlined") neighborhoods within a metropolitan area; the plaintiffs allege that Generation practices "reverse redlining" by "targeting unfair, predatory and/or grossly unfavorable loan terms" to African-Americans and single females.*fn1
The dispute addressed here concerns discovery scheduling. The plaintiffs seek broad discovery in support of their claims, and their initial discovery requests sought information relating to every loan Generation has originated since August 2007, a request that would require review, redaction, and production of data from (according to Generation) more than 20,000 loan files comprising between four and six million pages of data. Seeking to forestall the burden and expense of such class-wide discovery, Generation has moved to bifurcate discovery, arguing that the class-wide discovery plaintiffs seek should be deferred until the plaintiffs have presented a meritorious motion for class certification. The plaintiffs maintain, however, that it is "simply not possible" to bifurcate discovery in this case because "class and merits discovery not only overlap but are indistinguishable." Without class-wide discovery, the plaintiffs concede that they "will not be able to discover . . . even whether the individual plaintiffs suffered discrimination."
There are two reasons that class and merits discovery completely overlap in this case. First, the plaintiffs have defined the class by reference to the merits of their claims. As to each of their claims (pricing discrimination and reverse-redlining), the plaintiffs have identified four subclasses: (i) African-Americans in Illinois; (ii) African-Americans nationally; (iii) Single females in Illinois; and (iv) Single females nationally. The Second Amended Complaint ("SAC") defines each of these eight subclasses by reference to the claim of discrimination allegedly suffered by that sub-class. The first sub-class, for example, is defined as comprising "African-Americans in Illinois who received disparate pricing"; another comprises "Single females nationwide who experienced disparate treatment/reverse redlining." No wonder, then, that the plaintiffs assert that their class claims overlap entirely with their discrimination claims-proof of class membership requires proof of discrimination.
The Seventh Circuit has highlighted one problem with defining class membership by reference to the substance of a claim: "Using a future decision on the merits to specify the scope of the class makes it impossible to determine who is in the class until the case ends, and creates the prospect that, if the employer should prevail on the merits, this would deprive the judgment of preclusive effect: any other former worker could file a new suit, given that the losing 'class' lacked any members." Bolden v. Walsh Constr. Co., 688 F.3d 893, 895 (7th Cir. 2012). This case presents another, related, problem with this practice; if allowed, it would permit plaintiffs to forestall class certification until the end of the discovery process, rather than promoting consideration of the question of whether a class can be certified at "an early practicable time," as Rule 23(c)(1)(A) requires.
That the plaintiffs have defined the putative class improperly is a pleading defect that can, presumably, be fixed. More problematic, though, is the plaintiffs' concession that absent class-wide discovery, they cannot establish that there are questions of law or fact common to the class as required by Federal Rule of Civil Procedure 23(a)(2). Here, the plaintiffs' claims on behalf of the putative class implicate thousands of decisions about the loan terms extended to African-Americans and to single women; to move forward as a class action, Rule 23(a)(2) requires that there be "some glue holding the alleged reasons for all those decisions together . . . produc[ing] a common answer to the crucial question why was I disfavored." Wal-Mart Stores, Inc. v. Dukes, 131 S. Ct. 2541, 2552 (2011) (emphasis in original).
The plaintiffs allege that Generation's nationwide policy of affording pricing discretion to its brokers provides the commonality-"the glue"-necessary to obtain class certification. "Commonality requires the plaintiff to demonstrate that the class members 'have suffered the same injury.'" Id., 131 S. Ct. at 2551 (quoting Gen. Tel. Co. of Sw. v. Falcon,457 U.S. 147, 157 (1982)). The plaintiffs argue that they will be able to show common injury if analysis of comprehensive loan-level data shows that African-Americans and/or single women received loan terms that were worse, by a statistically significant margin, than those received by white men. And they maintain that such an analysis is "indispensable" to the viability of their class claims because only a statistical analysis of all of Generation's loans will uncover whether Generation's policy of affording its brokers pricing discretion had a common effect on class members. The plaintiffs stake their entire claims on the statistical analysis they seek to conduct, maintaining that "the evidence of the common effect of Generation's nationwide policy lies in the loan data." Surreply at 2 (emphasis in original).
The problem with this argument is that a statistical disparity in the loan terms received by African-Americans and/or single women says nothing about the cause of the disparity-which is the reason that the Supreme Court rejected the very same argument in Dukes. There, the Court addressed the propriety of certification of a class comprised of female Wal-Mart employees in a case alleging that Wal-Mart managers exercised the discretion the company afforded them over pay and promotions in a manner that discriminated against the female employees. 131 S. Ct. at 2547. As here, the Dukes plaintiffs sought to establish commonality between the huge number of individual decisions at issue (there, decisions affecting pay and promotions; here, decisions as to loan terms) by means of statistical evidence showing that the class members fared worse than others. But the Court rejected the notion that statistical analysis could suffice as evidence that class members were, in common, treated in a discriminatory fashion, holding that "merely proving that the discretionary system has produced a racial or sexual disparity is not enough" to establish discrimination. Id. at 2555 (emphasis in original) (citing Watson v. Fort Worth Bank & Trust, 497 U.S. 977, 994 (1988)).
Though it did not use the term, the Dukes Court identified "the fallacy of division"-the deductive error inherent in substituting evidence of the general for evidence of the particular-to explain the flaw in the plaintiffs' reasoning. See THE NEW ENCYCLOPAEDIA BRITANNICA, vol. 23 p. 252 (15th ed.) (stating that the fallacy of division "occurs when the premise that a collective whole has a certain nature is improperly used to infer that a part of this whole must also be of this nature"); Adrian Vermeule, Foreword: System Effects and the Constitution, 123 Harv. L. Rev. 4, 8 (2009) (defining fallacy of division as "the assumption that what is true of the aggregate must be true of the members"); Rosen v. Unilever U.S., Inc., No. 09 C 2563, 2010 WL 4807100, *6 (N.D. Cal. May 3, 2010) (providing example of the fallacy of division: "since a blend of oils is represented as being nutritious, if partially hydrogenated oil is part of the blend, it must also be nutritious"). In the context of Dukes, that meant that a general disparity between pay and promotion opportunities in one region could not be taken as evidence that all store managers within that region exercised their discretion uniformly in a manner that disadvantaged female employees; as the Court noted, a regional pay disparity "may be attributable to only a small set of Wal-Mart stores, and cannot by itself establish the uniform, store-by-store disparity upon which the plaintiffs' theory of commonality depends." 131 S. Ct. at 2555. Similarly here, higher average loan prices for the putative class members may reflect only that some brokers charge higher prices, not necessarily that all do. Indeed, if only one broker charged substantially higher prices based on race, that would raise the average price charged by all brokers (though the marginal increase would of course be a function of the size of the price differential and the number of brokers). See also Bolden, 688 F.3d at 896-97 (making the same point in reversing class certification orders predicated on statistical evidence of commonality).
More fundamentally, as the Court further explained in Dukes, higher average loan prices for putative class members would do nothing to establish that any of the brokers who did charge putative class members more did so because of the borrowers' race or gender, rather than lawful factors that otherwise distinguish reverse mortgage customers (for example, the myriad factors that can affect the value of a home, such as its age, condition, location, and size, or factors that would effectively determine the length of the loan, such as the age of the borrowers). Rather, there must be a common policy or practice that supports an inference that the statistical disparity is the product of discrimination and, as the Court noted, a policy that vests managers (or, here, brokers) with pricing discretion is the antithesis of a common policy or practice; "it is a policy against having uniform . . . practices." 131 S. Ct.at 2554 (emphasis in original).
The plaintiffs attempt to avoid the import of Dukes by arguing that Generation cabins the discretion it affords brokers by allowing the brokers to exercise their discretion in only one way-to increase, rather than reduce, prices charged for mortgages. That argument is a makeweight, as it in no way implies that the company requires, or even encourages, brokers to charge higher rates to African-Americans or single women. Presumably every company would like to charge as much as it can for its products, but a policy to that effect says nothing about whether the company discriminates among its customers on the basis of race or gender.
The plaintiffs' argument for resisting bifurcation, then, is not
compelling: the statistical evidence that they hope to develop during
the course of merits discovery will not-even if it shows, as
plaintiffs expect, that African-Americans and single women paid more
for their loans with Generation than did others-suffice to "glue"
their claims together in a way that satisfies Rule 23(a)(2)'s
commonality requirement. And if the plaintiffs' prospects for
obtaining class certification are dim, regardless of the scope of
discovery, that is plainly a factor that supports bifurcation-it would
be utterly inefficient and unjust to subject a defendant to months, if
not years, of onerous and expensive discovery so that the plaintiffs
may continue a quixotic undertaking destined to fail.*fn2
Bifurcation is warranted where it promises to promote
efficiency and economy without working substantial prejudice. See
Harris v. comScore, Inc., No. 11 C 5807, 2012 WL 686709, *3 (N.D. Ill.
Mar. 2, 2012); Fed. R. Civ. P. 23, 2003 Advisory Committee Notes
(stating that bifurcated discovery may be appropriate where it enables
a court to efficiently rule on class certification). Cf. Houseman v.
U.S. Aviation Underwriters, 171 F.3d 1117, 1121 (7th Cir. 1999)
(addressing bifurcation in the context of Rule 42(b)). And while the
plaintiffs claim that bifurcation will prejudice them-indeed, they
maintain that bifurcating class and merits discovery "would, in
effect, unfairly end the entire case"-there is no prejudice in
accelerating that outcome if it is indeed inevitable. If the only
evidence that the plaintiffs can offer to prove that the putative
class members each suffered discrimination is statistical evidence of
a disparity between the loan terms that the class obtained
collectively, then their claim is destined to fail in the wake of
For now, though, the Court rules only on what is before it-namely, Generation's bifurcation motion. That motion is granted; discovery may proceed at this point only as to evidence pertinent to the question of whether Generation's brokers exercised their pricing discretion in a common manner (or other issues relevant to class certification). Statistical evidence of disparate pricing is not relevant to the commonality inquiry, but it remains open to the plaintiffs and their counsel to ...