Puffer v. Allstate Insurance

675 F.3d 709 | 7th Cir. | 2012

675 F.3d 709 (2012)

Katherine PUFFER, Plaintiff, and
Karen Pell, et al., Intervening Plaintiffs-Appellants,
v.
ALLSTATE INSURANCE COMPANY, Defendant-Appellee.

No. 11-1273.

United States Court of Appeals, Seventh Circuit.

Argued February 23, 2012. Decided March 27, 2012.

Paul R. Garry (argued), Attorney, Meckler Bulger Tilson Marick & Pearson LLP, Richard C. Godfrey, Attorney, Kirkland & Ellis LLP, Chicago, IL, for Defendant-Appellee.

Linda Debra Friedman, Senior Attorney, Stowell & Friedman, Shona B. Glink (argued), Attorney, Meites, Mulder & Glink, Chicago, IL, for Intervenors-Appellants.

Before FLAUM and TINDER, Circuit Judges, and SHADID, District Judge.^[*]

FLAUM, Circuit Judge.

Katherine Puffer sued Allstate Insurance Company ("Allstate") on behalf of herself and a putative class, alleging that Allstate carried out a nationwide pattern or practice of sex discrimination in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., and the Equal Pay Act of 1963, 29 U.S.C. § 206(d). In her complaint, she alleged gender-based earning disparities rooted in differential treatment and disparate impact theories based on Allstate's salary, promotion, and training policies, which left significant discretion in the hands of individual managers. The district court denied plaintiff's motion for class certification. In response to the concerns expressed by the court and in response to the enactment of the Lilly Ledbetter Fair Pay Act of 2009, Pub. L. No. 111-2 (the "Ledbetter Act"), plaintiff again moved for class certification. This time, she focused exclusively on Allstate's uniform compensation policies. The court again denied certification, citing flaws in plaintiff's evidence and the lack of common issues.

Puffer then settled her individual claims with Allstate. Karen Pell, Gail Howells, and Mary Keith, members of the putative class, intervened to appeal the denials of class certification. These intervenors now allege only a disparate impact theory, claiming that Allstate's policy of awarding merit increases based on a percentage of base pay and Allstate's policy of comparing salaries to its competitors caused gender-based disparities in earnings.

We shall not reach the merits of this claim, however. We conclude that intervenors have waived their disparate impact claim by failing to meaningfully develop it before the district court. Although plaintiff nominally mentioned a disparate impact claim in her complaint, she developed and argued only her pattern-or-practice claim—a type of intentional discrimination. We therefore affirm the district court's denial of class certification.

I. Background

A. Allstate's Corporate Structure and Policies

Allstate Insurance Company is headquartered in Illinois and employs 38,000 people throughout the country. In 2002, Allstate created an operating unit called Allstate Protection, which employs approximately 80% of Allstate's personnel. Allstate Protection is organized into different Areas of Responsibility ("AORs"). The four largest AORs are Claims, Product Operations, Protection Finance, and Distribution.

Responsibilities, supervisors, office locations, and salary grades vary among these AORs. The Claims AOR handles insurance claims and employs approximately 17,000 people as attorneys, human resources professionals, information technology specialists, claims investigators, claims and subrogation service managers, and project managers. The Product Operations AOR creates insurance products, prices products, assesses risk, updates Allstate's insurance policies, and communicates with customers. Among its employees are pricing directors, actuaries, product operations managers, and personal line directors and managers. The Protection Finance AOR provides financial support and typically employs managers with finance or accounting degrees. The Distribution AOR works with the insurance agencies that sell Allstate's products by developing recognition plans, directly supporting agents, and promoting business.

The hierarchical structure among the AORs is similar: entry-level managers are at Salary Grades ("Grades") 63 and 64; senior managers are at Grade 77; directors are at Grade 78; appointed officer positions (Assistant Vice President and Assistant Field Vice President) are at Grade 80; and elected officer positions (Vice President, Senior Vice President, and Senior Vice President) are at Grade 90 and above. At the end of 2003, women comprised 61% of Protection employees but only 36% of the managerial positions and 24% of the officer positions.

At the beginning of each year, managers meet with their supervisors to discuss major job responsibilities and individual performance goals. At the end of the year, supervisors fill out performance development summaries for the managers they oversee. In a process known as "one-over-one" review, supervisors submit the performance summaries to their own superiors for input, review, and approval before turning them over to the evaluated managers. Allstate also evaluates its managers based on 17 critical success factors and on Quality Leadership Management Surveys. These evaluations factor into decisions about merit or promotional salary increases for managers.

Allstate Protection is allotted a percentage amount of money annually for merit or promotional salary increases. Allstate's Human Resources department establishes an overall range of salary increases in terms of a percentage amount for each performance and salary level. These guidelines are suggestions, leaving supervisors with the discretion to grant more or less of an increase based on performance. Salary increases for officers follow a similar process except that the senior management team makes the decisions.

B. Procedural Background

Katherine Puffer, the original plaintiff in this suit, worked at Allstate from 1977 to 2003. She primarily worked in the Finance AOR, but she also worked in the Production Operations AOR. By 1993, Puffer had been promoted eight times.

Puffer alleged that she was sexually harassed by her supervisor from late-1998 to mid-1999 and that she was not promoted because she rejected her supervisor's advances and because of her gender. She filed a charge of gender discrimination, sexual harassment, and retaliation with the Equal Employment Opportunity Commission ("EEOC") in January 2001, but she did not file a lawsuit within the required time period after receiving a right to sue letter. In 2003, Puffer was terminated as part of a reorganization and officer reduction in force. Fourteen officers, including Puffer, were terminated at that time: eleven were male and three were female. Puffer filed a new charge with the EEOC, alleging gender discrimination and retaliation, and she then timely filed the complaint in this case.

In her amended complaint, Puffer asserted three claims. First, she alleged that Allstate discriminated against her and a class of female managerial employees because of their gender and allowed the discrimination "to exist and go unremedied for so long that it amounts to a policy or practice and constitutes Allstate's standard operating procedure." Am. Compl. ¶ 38. She alleged this claim under both "differential treatment and disparate impact theories of liability under Title VII." Id. ¶ 53. Second, Puffer alleged a class claim under the Equal Pay Act, contending that Allstate pays putative class members "lower wages than male employees in substantially equal jobs even though Plaintiffs and all others similarly situated performed similar duties requiring the same skill, effort, and responsibility of male employees" and performed "substantially equal work." Id. ¶¶ 57, 59. Third, Puffer alleged an individual claim of retaliation under Title VII and the Equal Pay Act.

At the time Puffer filed her complaint, case law supported the "paycheck accrual rule," such that a new claim under Title VII arose each time an employer issued a paycheck reflecting discriminatory wages, even if the discriminatory decision setting the amount occurred outside the limitations period (i.e., more than 300 days before the filing of the first EEOC charge). But in 2007, in Ledbetter v. Goodyear Tire & Rubber Co., the Supreme Court held that present effects of a discriminatory decision that predated the limitations period could no longer form the basis for Title VII liability. 550 U.S. 618, 632, 643, 127 S.Ct. 2162, 167 L.Ed.2d 982 (2007).

On November 8, 2007, after substantial discovery, Puffer moved for class certification pursuant to Federal Rules of Civil Procedure 23(b)(2) and 23(b)(3) of her Title VII claims. She sought to certify a class of approximately 1700 female managers in the position of division or department manager (Grade 63) and above who worked or continue to work at Allstate Protection since May 9, 2001. Members of the putative class held 275 different jobs in a hundred offices across the country. Eighty percent of the putative class was employed in entry-level management positions, with the remainder of the class in higher management and officer positions. Plaintiff argued that Allstate has a "strong paternalistic culture" and further that "Allstate Protection's organizational structure vests all authority to choose, advance and compensate staff in the male-dominated, senior manager ranks." Pl.'s Class Cert. Mem. 1.

The parties briefed the class certification motion for nearly a year. They offered the opinions of two statisticians, two sociologists, two industrial psychologists, and 84 fact witnesses. Plaintiff's main expert, Dr. Janice Madden, presented statistical evidence indicating that, from 2001 through 2005,^[1] women earned substantially lower wages than men did at every salary grade—even after accounting for job code, tenure, and other variables. Plaintiff also presented the expert report of Dr. Cristina G. Banks, an industrial and organizational psychologist, and Dr. Barbara Reskin, a sociologist. Allstate countered with analysis from its own experts, including labor economist Dr. Robert Topel, industrial psychologist Dr. Kathleen K. Lundquist, and sociologist Dr. Christopher Winship.

On January 15, 2009, Magistrate Judge Schenkier denied plaintiff's motion for class certification. The 42-page opinion discussed the pattern-or-practice claim almost exclusively, mentioning the disparate impact claim in a footnote but providing no separate analysis. Although the court found that the putative class satisfied Rule 23(a)'s numerosity and adequacy requirements, the court found that it failed to satisfy the commonality requirement. The court stressed the heterogeneity of the class members and the individual decisionmakers. The court found plaintiff's expert reports only minimally probative of commonality and found the declarations to demonstrate a lack of commonality. The court discounted Dr. Madden's statistical analysis because it included wage differentials that predated the class period (contrary to Ledbetter), because it failed to account for "measures of productivity," and because it revealed significant variations in the standard deviations between salary grades. The court also determined that plaintiff failed to meet the typicality requirement because the claims and defenses would necessitate individualized inquiries.

The district court concluded that, even if plaintiff could satisfy commonality and typicality, she failed to show that certification was proper under either Rule 23(b)(2) or (b)(3). The court rejected the propriety of a Rule 23(b)(2) class since plaintiff sought primarily money damages. The court also found a Rule 23(b)(3) class to be inappropriate, concluding that common issues did not predominate based on the individualized inquiries that would be necessary to determine the reason for each employment action. The court determined that a class action was not the superior means to adjudicate the claim because 1700 individual jury trials would be needed to decide damages, constitutional questions would arise by having different juries examine the liability issue, and the individual class members had sufficiently large stakes to be able to afford to litigate on their own.^[2]

Soon after the district court's decision, President Obama signed the Ledbetter Act, effectively overruling the Supreme Court's Ledbetter decision and retroactively reinstating the paycheck accrual rule.

On September 11, 2009, plaintiff moved again for class certification and argued that the court should reconsider the evidence submitted in light of the Ledbetter Act. She substantially narrowed her argument in this motion, arguing only that Allstate's uniform annual salary review and competitive pay analysis disadvantaged female managers. Although she used the terms "disparate impact" and "pattern-or-practice" to describe her claims in her motion for class certification, her supporting memorandum discussed factual allegations and legal arguments only in support of the pattern-or-practice claim.

On October 8, 2009, Magistrate Judge Schenkier issued an extensive oral decision denying plaintiff's second motion for certification. The court acknowledged that plaintiff had made a "genuine effort" to address his concerns but concluded that the renewed motion still failed to satisfy Rule 23. Regarding its commonality concerns, the court criticized Dr. Madden's analysis for failing to control for all of the variables, failing to consider the influence of performance assessments on salary differences, and failing to control for job location (relevant for looking at the actions of particular supervisors). The court expressed that the significant variations in standard deviations by salary grade suggested that differentiations existed that were not adequately explained by the statistical analysis. The court was not persuaded by plaintiff's narrowing of the class certification claim to external review of salary decisions because the court found it to be inseparable from internal issues and because Dr. Madden's analysis had examined both. The court did not address the effect of the Ledbetter Act since unrelated deficiencies convinced the court that plaintiff failed to establish commonality. Although plaintiff eliminated her claim for compensatory damages, the court still determined that the putative class failed to meet Rule 23(b)(2) because backpay would be substantial (possibly $85 million) and because declaratory relief would be only backwards-looking. The court also expressed concern about the lack of an opt-out right. The court likewise found Rule 23(b)(3) to be an inappropriate basis for certification based on its previous concerns about predominance and superiority.

Puffer's case was then transferred to Magistrate Judge Gilbert. On January 7, 2011, Puffer settled her individual claims with Allstate. The court dismissed her case with prejudice, thus making the denials of class certification final and appealable.

Putative class members Karen Pell, Mary Keith, and Gail Howells moved to intervene to appeal the January 15, 2009 and October 8, 2009 denials of class certification. Pell worked for Allstate from 1973 to 1989 and again from 1995 to 2006, in three northeastern states. She briefly worked as a Frontline Performance Leader (Grade 63) before being demoted to a desk position in 2005. Keith worked for Allstate from 1980 to 2009, predominantly in the Finance AOR. She last worked as a Director in Finance (Grade 78). Howells worked for Allstate from 1984 to 2006, primarily in the Distribution AOR in the South West Region. She worked in several Grade 64 management positions. Magistrate Judge Gilbert granted their motion to intervene on January 28, 2011.

Intervenors appeal only the denial of certification for the Title VII disparate impact challenge under Rule 23(a), (b)(3), and (c)(4). They claim that Allstate's salary administration process, which determines merit increases based on a percentage of base pay and an external market comparison, had a disparate impact on female managers between 2002 and 2005. They do not appeal the denial of class certification as to the pattern-or-practice claim of intentional discrimination, in light of the Supreme Court's decision in Wal-Mart Stores, Inc. v. Dukes, ___ U.S. ___, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). They rely on the statistical reports from Dr. Madden, but they do not rely on the report of Dr. Banks, who had stated that bias stems from the exercise of discretion and that performance evaluations are not conducted consistently between regions or supervisors, or on the report of Dr. Reskin, who has opined that Allstate has a paternalistic culture and that the exercise of discretion contributes to the systematic disadvantage of the female managers.

II. Discussion

A. Standard of Review

We review the district court's denial of a motion for class certification for an abuse of discretion, but we review the court's purely legal determinations de novo. Andrews v. Chevy Chase Bank, 545 F.3d 570, 573 (7th Cir.2008). A plaintiff who moves for class certification must satisfy the numerosity, commonality, typicality, and adequacy of representation requirements of Rule 23(a), as well as at least one subsection of Rule 23(b). FED. R.CIV.P. 23; see also Arreola v. Godinez, 546 F.3d 788, 797 (7th Cir.2008).

Intervenors argue for de novo review, contending that the district court applied the wrong standard under Lewis v. City of Chicago, ___ U.S. ___, 130 S.Ct. 2191, 176 L.Ed.2d 967 (2010), and that we should reconsider their claim in light of Wal-Mart Stores, Inc. v. Dukes, ___ U.S. ___, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). In particular, intervenors argue that the district court failed to analyze plaintiff's disparate impact claim separately from her pattern-or-practice claim. We fully recognize this argument, but, as we explain in the following section, we conclude that the district court's approach resulted from plaintiff's insufficient presentation of the disparate impact claim.

B. Waiver of Disparate Impact Claim

Allstate contends that plaintiff failed to adequately present a disparate impact claim based on a uniform salary administration policy to the district court such that intervenors are now barred from making such a claim on appeal. After reviewing the motions and memoranda filed by plaintiff, we agree with Allstate that plaintiff did not adequately present her disparate impact claim to the district court. Because plaintiff did not preserve this claim, intervenors may not rely on it as the basis for their appeal. The waived disparate impact claim is the only claim that intervenors appeal, and therefore we must affirm the district court's denial of certification.

1. Legal Framework

Disparate impact claims require no proof of discriminatory motive and "involve employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another and cannot be justified by business necessity." Int'l Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). By contrast, differential treatment claims, also known as disparate treatment claims, require plaintiffs to prove discriminatory motive or intent. See id. at 335-36, 97 S.Ct. 1843.

Pattern-or-practice claims, like differential treatment claims, represent a theory of intentional discrimination. Council 31, Am. Fed'n of State, Cnty. & Mun. Emps., AFL-CIO v. Ward, 978 F.2d 373, 378 (7th Cir.1992). Pattern-or-practice claims require a "showing that an employer regularly and purposefully discriminates against a protected group." Id.; see also King v. Gen. Elec. Co., 960 F.2d 617, 623 (7th Cir.1992). Plaintiffs must prove that discrimination "was the company's standard operating procedure— the regular rather than the unusual practice." Int'l Bhd. of Teamsters, 431 U.S. at 336, 97 S.Ct. 1843; see also id. at 360, 97 S.Ct. 1843 ("[A]t the liability stage of a pattern-or-practice trial the focus often will not be on individual hiring decisions, but on a pattern of discriminatory decisionmaking."). As one circuit has observed, "[i]n every instance in which the phrase has been used by the Supreme Court, a `pattern or practice' claim under Title VII refers to a pattern or practice of disparate treatment, rather than disparate impact." Wessmann v. Gittens, 160 F.3d 790, 817 n. 20 (1st Cir.1998); see also Allison v. Citgo Petroleum Corp., 151 F.3d 402, 409 (5th Cir.1998) (referring to a pattern-or-practice claim as a "systemic disparate treatment theory"). We have referred to a pattern-or-practice case as a "means of proving intentional discrimination [that] is distinct from a disparate impact case, where the plaintiff need not make any showing of the employer's intent." Council 31, 978 F.2d at 378.

For disparate impact claims, a plaintiff must establish that a particular employment practice causes a disparate impact on a member of a protected class. 42 U.S.C. § 2000e-2(k); Adams v. City of Chicago, 469 F.3d 609, 613 (7th Cir.2006). To satisfy this burden, the plaintiff is "responsible for isolating and identifying the specific employments practices that are allegedly responsible for any observed statistical disparities." Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 994, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988); Bennett v. Roberts, 295 F.3d 687, 698 (7th Cir.2002). Notably, "it is not enough to simply allege that there is a disparate impact on workers, or point to a generalized policy that leads to such an impact." Smith v. City of Jackson, 544 U.S. 228, 241, 125 S.Ct. 1536, 161 L.Ed.2d 410 (2005). Failure to identify the specific practice could lead to employers being held liable for "the myriad of innocent causes that may lead to statistical imbalances." Watson, 487 U.S. at 994, 108 S.Ct. 2777. The plaintiff must also establish causation by "offer[ing] statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotions because of their membership in a protected group." Id. at 994-95, 108 S.Ct. 2777. Only if the plaintiff succeeds in establishing disparate impact does the burden shift to the defendant-employer to demonstrate that the employment practice is "job related for the position in question and consistent with business necessity." 42 U.S.C. § 2000e-2(k)(1)(A); see also Lewis, 130 S.Ct. at 2198-99. If the employer satisfies this requirement, the burden shifts back to the plaintiff to show that an equally valid and less discriminatory practice was available that the employer refused to use. Adams, 469 F.3d at 613.

2. Intervenors' Claim on Appeal

After undergoing numerous transformations, intervenors' claim now asserts that Allstate's annual award of merit increases based on a percentage of base pay and its external salary comparison policy, although neutral on their face, had a disparate impact on women during the class period. Intervenors claim that undisputed evidence shows that gender-based disparities in earnings existed at the beginning of the class period and that these two policies failed to eliminate these disparities. With respect to the internal policy, intervenors allege that the award of merit increases based on percentage pay caused female managers to receive smaller increases in terms of actual dollars than similarly situated male managers. With respect to the external policy, intervenors allege that the market analysis compared the average salary by job code to the external market and adjusted where the average salary fell below 10% of the market rate. Intervenors claim that this policy failed to correct gender-based earnings disparities because the average salary was skewed by disproportionately high salaries for men.

3. Waiver Analysis

It is a well-established rule that arguments not raised to the district court are waived on appeal. See Brown v. Auto. Components Holdings, LLC, 622 F.3d 685, 691 (7th Cir.2010); Robyns v. Reliance Standard Life Ins. Co., 130 F.3d 1231, 1238 (7th Cir.1997). Moreover, even arguments that have been raised may still be waived on appeal if they are underdeveloped, conclusory, or unsupported by law. See United States v. Berkowitz, 927 F.2d 1376, 1384 (7th Cir.1991) ("[P]erfunctory and undeveloped arguments, and arguments that are unsupported by pertinent authority, are waived...."); United States v. Dunkel, 927 F.2d 955, 956 (7th Cir.1991) ("A skeletal `argument,' really nothing more than an assertion, does not preserve a claim."). We have also recognized that raising an issue in general terms is not sufficient to preserve specific arguments that were not previously presented. See Fednav Int'l Ltd. v. Cont'l Ins. Co., 624 F.3d 834, 841 (7th Cir.2010). Similarly, we have announced that "[i]t is the parties' responsibility to allege facts and indicate their relevance under the correct legal standard." Econ. Folding Box Corp. v. Anchor Frozen Foods Corp., 515 F.3d 718, 721 (7th Cir.2008) (quoting APS Sports Collectibles, Inc. v. Sports Time, Inc., 299 F.3d 624, 631 (7th Cir.2002)) (internal quotation marks omitted) (finding waiver where plaintiff argued for reversal based on a ground not presented below). Although plaintiff is not a party to this appeal, intervenors stand in plaintiff's shoes and can only present arguments that she could have appealed. See Champ v. Siegel Trading Co., 55 F.3d 269, 273-74 (7th Cir.1995). Because plaintiff neglected to argue the disparate impact theory before the district court, it is waived on appeal.

Plaintiff repeatedly and almost exclusively argued to the district court that she was seeking class certification under a pattern-or-practice theory. Although her complaint alleged both "differential treatment and disparate impact theories of liability," Am. Compl. ¶ 53, she only provided factual allegations and legal arguments to support her pattern-or-practice claim.^[3] She did not argue that Allstate utilized a facially neutral employment policy that resulted in a disparate impact,^[4] and she did not cite any disparate impact cases.

In her original motion for class certification, plaintiff consistently framed her claim as one of intentional discrimination: "Puffer brings this case as a pattern-or-practice case, and has the prima facie burden of showing that unlawful discrimination based upon sex was the employer's regular policy." Pl.'s Class Cert. Mem. 24; see also id. at 1 ("Allstate steered women into these lower-paying, female-dominated jobs."). Plaintiff emphasized that "the disparities in job assignment and compensation result from the strong paternalistic culture at Allstate." Id. at 1. She argued that the "organizational structure vests all authority to choose, advance and compensate staff in the male-dominated, senior manager ranks" and that the "employment policies allow management to apply their own subjective views, without explanation or accountability, of whom to promote or develop." Id. Plaintiff referred to her case as a "pattern-or-practice" case repeatedly when arguing that her proposed class satisfies the commonality, predominance, and superiority requirements of Rule 23. See, e.g., id. at 24, 26 ("[C]lass-wide proceedings would definitively settle the common pattern-or-practice issue."), 27-29; Pl.'s Class Cert. Reply 25-26. But plaintiff's first class certification motion and memorandum are void of any allegations related to a theory of unintentional discrimination, which is the only theory that intervenors raise on appeal.^[5]

Similarly, in her second motion for class certification, plaintiff asserted in one sentence that she was alleging both disparate impact and differential treatment claims,^[6] but she failed to support this bare and conclusory assertion. In the memorandum, she focused her allegations on Allstate's uniform compensation policies, but she nowhere labeled this claim as a disparate impact claim and she nowhere cited to disparate impact case law. Instead, she repeatedly characterized her claim as a pattern-or-practice claim, see, e.g., Pl.'s Corr. Second Class Cert. Mem. at 12-18, and she relied on cases that exclusively discuss intentional discrimination. See, e.g., id. at 17.

The district court acknowledged that plaintiff alleged both disparate treatment and disparate impact claims in her complaint, and the court briefly addressed the legal distinctions between these two claims. The district court, however, almost exclusively analyzed plaintiff's disparate treatment claim. The district court did not distinguish the intentional and unintentional discrimination theories when finding a lack of commonality and predominance. The court stated that the proposed class members would need to prove the individualized issue of discriminatory motive, even though the court had noted elsewhere that "motive is irrelevant" in disparate impact claims and that business necessity—a non-individualized issue—is the only defense available to rebut a prima facie disparate impact claim. The minimal attention that the district court gave to plaintiff's disparate impact claim can be directly attributed to the scant support that plaintiff provided for this claim. Plaintiff only sufficiently developed her claim of intentional discrimination—thus, this is the only claim that the district court fully addressed and the only claim that is preserved for appeal.

In Pond v. Michelin North America, Inc., we determined that plaintiff had waived her disparate impact argument because her reliance on the McDonnell Douglas framework in her briefing demonstrated that she was raising solely a claim of disparate treatment. 183 F.3d 592, 597 (7th Cir.1999). We found her two references to support in the record to be insufficient, stating that "[a]rguments not raised in district court are waived on appeal, as are arguments raised in a conclusory or underdeveloped manner." Id. (citation omitted). The present case is analogous: plaintiff repeatedly labeled her case as a pattern-or-practice case, and intervenors are unable to point to any factual or legal support that plaintiff raised in support of a disparate impact claim. The few bare assertions are simply not enough. While we recognize the difficulty presented by shifting legal doctrine, intervenors still "cannot change course on appeal to raise an argument different than the one presented to the district court." MMG Fin. Corp. v. Midwest Amusements Park, LLC, 630 F.3d 651, 659 (7th Cir. 2011). Plaintiff did not meaningfully develop an argument seeking certification based on a disparate impact theory of liability. See Econ. Folding Box Corp., 515 F.3d at 720-21. We therefore conclude that intervenors' disparate impact claim is waived.^[7]

III. Conclusion

For the foregoing reasons, we AFFIRM the judgment of the district court.

NOTES

[*] The Honorable James E. Shadid, Chief Judge for the United States District Court for the Central District of Illinois, sitting by designation.

[1] In 2006, Allstate apparently adjusted the earnings levels of women to eliminate observed gender disparities within job codes.

[2] Twenty-five members of the putative class later pursued individual claims.

[3] Although the same set of facts can give rise to claims under both intentional and unintentional theories of discrimination, see Int'l Bhd. of Teamsters, 431 U.S. at 335 n. 15, 97 S.Ct. 1843, plaintiff only used the facts to develop her pattern-or-practice claim. The factual allegations section of plaintiff's complaint refers to the "pattern-or-practice" of discrimination at Allstate at least eight times but refers not once to "disparate impact." Am. Compl. ¶¶ 4, 5, 6, 12, 15, 18, 26, 35.

[4] To the contrary, plaintiff argued that Allstate's compensation policies were not gender neutral. Pl.'s Corr. Second Class Cert. Mem. 7, 12, 18.

[5] In responding to Allstate's waiver argument, intervenors do not direct our attention to any disparate impact arguments made by plaintiff in support of her first motion for class certification. See Intervenors' Reply 22-23.

[6] Allstate asserts incorrectly in its brief that the phrase "disparate impact" does not appear a single time in plaintiff's second motion for class certification. The motion states: "Plaintiff alleges, among other things, that Allstate engaged in a pattern or practice of intentional gender discrimination and employed compensation policies and practices that had a disparate impact on women managers...." Pl.'s Corr. Second Mot. Class Cert. 2 (emphasis added). Notwithstanding Allstate's erroneous assertion, plaintiff's singular reference to "disparate impact" is not sufficient to preserve this theory of liability for appeal. See Berkowitz, 927 F.2d at 1384.

[7] Even if we had not found the disparate impact claim to be waived, this claim would fall short on its merits. Intervenors bear the burden of establishing "a causal connection between the employment practice and the statistical disparity, offering `statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotion because of their member in a protected group.'" Bennett, 295 F.3d at 698 (quoting Vitug v. Multistate Tax Comm'n, 88 F.3d 506, 513 (7th Cir.1996)). They have not met this burden, nor have they established that their proposed class satisfies Rule 23.

The record does not support intervenors' claim that Allstate's compensation policies are uniformly applied and cause gender-based earning disparities. Intervenors present us with the same expert testimony and record evidence that plaintiff previously presented when arguing that Allstate's salary decisions are decentralized, discretionary, and partially based on subjective criteria. Intervenors have not established that Allstate adheres to a uniform salary administration policy. More importantly, intervenors have not established that the merit increase and market comparison policies are the cause of the earnings disparities, as opposed to any of the other policies that were alleged originally and that the experts determined to be causes of the disparity. Furthermore, because intervenors rely on the same report from Dr. Madden that the district court criticized, they run into similar difficulties satisfying the commonality and predominance requirements. For example, Dr. Madden did not consider all relevant variables in one regression, nor did she analyze whether variations in performance levels could explain variations in earnings. Moreover, individual issues would certainly predominate at the damages phase. Bifurcation of the liability and damages phases under Rule 23(c)(4) would not resolve this concern because intervenors have not established that a uniform policy caused the disparity.