Shirley

Full title: Shirley

Court: United States District Court, N.D. California

Date published: Sep 25, 2012

285 F.R.D. 492 (N.D. Cal. 2012)

Opinion

[Copyrighted Material Omitted] [Copyrighted Material Omitted]

Bill Lann Lee, Julia Campins, Lindsay Polastri Nako, Lewis, Feinberg, Lee, Renaker & Jackson, P.C., Oakland, CA, Daniel M. Hutchinson, Kelly M. Dermody, Lieff Cabraser Heimann & Bernstein, LLP, Elizabeth A. Lawrence, Steve Stemerman, Davis Cowell & Bowe, James M. Finberg, Altshuler Berzon LLP, San Francisco, CA, Brad Seligman, Jocelyn Dion Larkin, Impact Fund, Berkeley, CA, for Plaintiffs.

Annette Tyman, Gerald L. Maatman, Jr., Seyfarth Shaw LLP, Chicago, IL, David B. Ross, Seyfarth Shaw LLP, New York, NY, Thomas J. Wybenga, Issaquah, WA, William Owen Kampf, Law Office of William O. Kampf, David D. Kadue, Seyfarth Shaw LLP, Los Angeles, CA, for Defendant.

ORDER GRANTING PLAINTIFFS' MOTION FOR CLASS CERTIFICATION; AND DENYING DEFENDANT'S MOTION TO ELIMINATE CLASS CLAIMS

EDWARD M. CHEN, District Judge.

TABLE OF CONTENTS

I.

INTRODUCTION

496

II.

FACTUAL AND PROCEDURAL BACKGROUND

497

A.

Factual Background

497

1.

Ellis

499

2.

Horstman

499

3.

Sasaki

500

B.

Judge Patel's Prior Order

500

1.

Standing

500

2.

Numerosity

500

3.

Commonality

502

4.

Typicality

502

5.

Adequacy

502

6.

Rule 23(b)(2)

502

C.

Ninth Circuit's Opinion

502

D.

Plaintiffs' New Class Certification Proposals

503

III.

DISCUSSION

503

A.

Evidentiary Dispute Re Costco's Declarations

503

B.

Motion for Class Certification

504

1.

Rule 23(a)

506

a.

Numerosity

506

b.

Commonality

506

i.

Dukes

507

509

ii.

Ellis II

iii.

Disparate Treatment

510

A.

Classwide Policies & Practices— Costco'sPromotion System

511

1.

High Level Management Involvement

511

2.

Common Criteria

514

3.

Costco's Own Evidence

515

4.

Discretion

518

5.

Summary

519

B.

Costco's Culture

520

C.

Classwide Effects— Statistical Evidence

521

1.

National vs. Regional

521

2.

Competing Explanations for Statistical Disparity

524

3.

Challenges to Methodology

528

D.

Conclusion

530

iv.

Disparate Impact

531

c.

Typicality

533

d.

Adequacy

535

2.

Rule 23(b)(2)— Injunctive Relief Class

535

3.

Rule 23(b)(3)— All Monetary and IndividualEquitable Relief

537

a.

Predominance

537

b.

Superiority

539

c.

Punitive Damages

540

d.

Rule 23(c)(4)

544

4.

Rule 23(g)— Class Counsel

545

IV.

CONCLUSION

545

I. INTRODUCTION

Plaintiffs Shirley " Rae" Ellis, Leah Horstman, and Elaine Sasaki (" Named Plaintiffs" ), current and former employees of defendant Costco Wholesale Corporation (" Costco" ), brought a putative class action alleging gender discrimination in Costco's promotion and management practices. Pursuant to Title VII of the Civil Rights Act of 1964, 42 U.S.C. section 2000e, et seq., Plaintiffs allege that Costco has engaged in a pattern or practice of discrimination against women in promotions to two management positions and that Costco's promotion system has a disparate impact on female employees. In addition, Plaintiff Ellis brings a claim for retaliation, and the Named Plaintiffs bring pendent causes of action alleging gender discrimination in violation of the Fair Employment and Housing Act, California Government Code section 12940, et seq.

On August 28, 2006, Plaintiffs sought certification of a nationwide class consisting of current and former female Costco employees who were denied promotion to General Manager (" GM" ) or Assistant General Manager (" AGM" ) positions since January 3, 2002. Plfs' Mot. for Class Cert., Docket No. 127. On January 11, 2007, Judge Patel granted class certification. Order Re: Mot. for Class Cert., Docket No. 494, Stip. Re: Class Definition, Docket No. 511; Ellis v. Costco Wholesale Corp. (" Ellis I " ),¹ 240 F.R.D. 627 (N.D.Cal.2007). After Costco sought and received permission to file an interlocutory appeal, on September 16, 2011, the Ninth Circuit affirmed in part, vacated in part and remanded the certification order for reconsideration in light of Wal-Mart Stores, Inc. v. Dukes, ---U.S. ----, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). See Ellis v. Costco Wholesale Corp. (" Ellis II " ), 657 F.3d 970 (9th Cir.2011).

1 The Court will refer to Judge Patel's prior ruling in this case as Ellis I, and the Ninth Circuit's ruling as Ellis II .

On remand, Defendant has filed a motion for an order eliminating class claims, and Plaintiffs have filed a cross-motion for class certification. Docket Nos. 543, 664. Plaintiffs seek to certify two classes: (1) An injunctive relief class of all women who are currently employed or who will be employed at any Costco warehouse in the U.S. who have been or will be subject to Costco's system for promotion to Assistant General Manager and/or General Manager positions; and (2) A monetary relief class of all women who have been employed at any Costco warehouse store in the U.S. since January 3, 2002 who have been subject to Costco's system for promotion to Assistant General Manager and/or General Manager positions. The parties' motions are now pending before the Court.

II. FACTUAL AND PROCEDURAL BACKGROUND

A. Factual Background

Defendant Costco is a corporation headquartered in Issaquah, Washington, which operates cash-and-carry membership warehouses throughout the United States. Costco's nationwide operations are divided into three divisions (Southwest, Eastern, and Northern/Midwest), each governed by an Executive Vice President (" EVP" ). See Docket No. 665, Ex. 25 (organizational chart). These divisions are in turn divided into regions— a total of eight companywide— managed by Senior Vice Presidents (" SVPs" ).² Id. Each Costco region is broken down into districts, led by District VPs; the districts are in turn composed of numerous Costco warehouses. Id.

2 When evidence in the record refers to vice presidents without specifying their rank, the Court will refer to them simply as " VPs."

Costco's top management— from Senior VPs up— meets once every four weeks at company headquarters in Washington. Zook Depo., Docket No. 665, Ex. 11, at 21. In addition to other matters, personnel and potential candidates for promotion are " frequently discussed among top-level managers, both at weekly meetings and the monthly meetings at Costco headquarters in Issaquah, Washington." Omoss Decl., Docket No. 611, ¶ 9 (VP of Texas region). ³ Regional executives also have direct contact with store-level staff through regular floor walks averaging approximately once per month, during which they can evaluate potential candidates for promotion and offer feedback to store-level managers. See, e.g., Rosolino Decl., Docket No. 618, ¶ 32 (Regional VP participates once per month); Ward Decl., Docket No. 631, ¶ ¶ 29-31 (VPs participate twice per month); Cafiso Decl., Docket No. 559, ¶ 18 (regional manager participates once per month); Nierstheimer Decl., Docket No. 610, ¶ 28 (Regional VPs participate once per month); Webb Decl., Docket No. 652, ¶ 6 (as Regional VP, tries to visit each warehouse once per month); Cline Decl., Docket No. 563, ¶ 21 (VPs participate every three months); Bejarano Decl., Docket No. 640, ¶ 25 (regional manager participates once per month); Alvarez Decl., Docket No. 551, ¶ 9 (regional managers observe and discuss candidates for promotion); Asch Decl., Docket No. 554, ¶ 15 (VPs participate and evaluate employees); D'Agostino Decl., Docket No. 565, ¶ 25 (received " constant feedback" from VPs based on floor walks). Although less frequent, higher management— up to and including the CEO— also participate in " floor walks" at each warehouse to meet and evaluate store-level management. See, e.g., Eringer Decl., Docket No. 567, ¶ ¶ 44-45 (regional and district executives, VPs, and CEO); Evans Decl., Docket No. 568, ¶ 30 (floor walks " are conducted by various levels of management, from the warehouse level all the way up to the CEO" ); Hinds Decl., Docket No. 584, ¶ 24 (VPs and CEO participate in walks). As relevant to this matter, the Ninth Circuit described Costco's store-level structure as follows:

3 Mr. Omoss does not specify whether the potential candidates to whom he refers includes both AGM and GM candidates, but the context of his declaration (discussing both AGM and GM prospects) suggests he was referring to candidates for promotion generally, including AGM and GM positions.

Costco operates over 350 warehouse-style retail establishments (warehouses). These warehouses sell items ranging from groceries to electronics. Within each Costco warehouse, the management structure consists of a General Manager (GM), two to three Assistant General Managers (AGM), and three to four Senior Staff Managers. A Costco GM is responsible for the entire operation of his or her respective warehouse and earns an average salary of approximately $116,000, plus stock and bonuses. Costco AGMs are second in command within each warehouse and earn an average salary of approximately $73,000, plus stock and bonuses. Costco's Senior Staff Managers are divided into four categories: Front End Managers, Administration Managers, Receiving Managers, and Merchandise Managers.⁴ Front End Managers oversee cashiers, membership/marketing personnel, cart staff, and other employees who deal directly with Costco members. Administration Managers manage administrative functions such as payroll and human resources. Receiving Managers oversee stocking of all incoming items from the receiving dock to the shelves. Merchandising Managers oversee lower level managers and are responsible for planning floor displays to maximize sales.

4 Most warehouses employ four Senior Staff Managers— one for each of the four Senior Staff Manager positions. However, in some larger warehouses, one of the AGMs assumes the role of Administration Manager.

Costco promotes almost entirely from within its organization. Only current Costco AGMs are eligible for GM positions. Costco does not have any written policy explaining to employees the criteria to be considered for promotion to GM or AGM, though candidates are promoted from a list of promotable candidates. Costco does not have written guidelines explaining how candidates should be selected for the promotable lists and does not regularly inform employees about the existence of such lists. Costco does not require that more than one candidate be considered for any particular opening or that a performance evaluation or any other documents be reviewed before a recommended candidate is approved. Costco also lacks a consistent practice for interviewing potential candidates for GM and AGM openings. Costco does not keep records regarding the selection process.

Costco employs a different promotion procedure for Senior Staff Managers. Costco fills the majority of Senior Staff openings by rotating managers among the four Senior Staff positions. This rotation is part of Costco's philosophy and, in Costco's opinion, trains and develops managers for future advancement by exposing them to different aspects of Costco's operations. Like the GM and AGM promotion procedures, Costco has no written guidelines regarding rotation of Senior Staff Managers.

Ellis II,

Despite the lack of written guidelines, as observed by the Ninth Circuit, Costco nonetheless imposes uniform policies and practices with regard to its promotion system. In addition to promoting from within, requiring that GMs come from the roster of AGMs and requiring AGMs to come from the roster of SSMs, Costco also requires merchandising experience— generally in the form of the Merchandise Manager (" MM" ) position— before Senior Staff Managers (" SSMs" ) are eligible for promotion to AGM. See Hoover Decl., Docket No. 586, ¶ 7; Kadue Decl., Docket No. 544, ¶ 4 (collecting citations to testimony across regions); Sinegal Depo., Docket No. 665, Ex. 8, at 42-43. Costco does not allow posting for either position. Sinegal Depo. at 123-24; see also id. at 128. The company maintains lists of promotable candidates to AGM at the regional level, and lists for GM at the regional and higher levels. Zook Decl., Docket No. 653, ¶ 19(f) (AGM list at regional level); id. ¶ 20(d) (GM list at regional level); Portera Depo., Docket No. 665, Ex. 6, at 86-87 (EVP receives promotable list of GM candidates); Sinegal Depo., Docket No. 665, Ex. 8, at 65 (GM promotable list maintained at headquarters in Green Room). Costco's upper management closely tracks promotion into both of these important positions, and " officers at the regional and corporate levels are involved in promotion decisions" for both AGM and GM. Ellis I, 240 F.R.D. at 639. For example, Costco's former CEO, Jim Sinegal, ⁵ testified that he is directly involved in the process of recruiting and selecting candidates for promotion, and that he provides instructions to his staff as to criteria for promotion to both AGM and GM. Sinegal Depo., Docket No. 665, Ex. 8, at 23-24, 29-30, 42-43. This is especially the case for the GM positions, each of which he personally reviews. Sinegal Depo. at 23-24. For AGM promotions, regional management are the final decisionmakers as to specific promotions, with direct input from local GMs. See Schutt Depo., Docket No. 544, Ex. U, at 76, 95-96 (AGM decision made at the regional level, with the GM, Regional Manager, Regional VP, and Senior VP involved, and " [m]any times" the Senior VP will approve it); Portera Depo., Docket No. 544, Ex. O, at 122 (GMs make the decision and recommendation for AGM promotions " in conjunction with their regional vice presidents, and then the ultimate approval is given by a senior vice president of the region" ); Zook Decl., Docket No. 653, ¶ ¶ 7, 14-15 (stating that EVP is informed of the decision and EVP " oversee[s] the promotion process involving ... AGMs," and that GMs are the primary decisionmakers with respect to AGM promotions in conjunction with Regional Operations Managers); see also Hoover Depo., Docket No. 544, Ex. I, at 91 (GM, VP, and Senior VP make decision, EVP informed); Webb Decl., Docket No. 652, ¶ 6 (Regional VP makes decision with input from GM); Webb Depo., Docket No. 665, Ex. 10, at 82 (same).

5 Although Mr. Sinegal is no longer Costco's CEO as of 2012, the Court refers to him as " CEO" and uses the present tense with respect to testimony by or about him because the evidence in the record concerns his tenure.

1. Ellis

The Ninth Circuit described Plaintiff Ellis's factual background as follows:

Costco hired Shirley Ellis as an AGM in 1998. Prior to joining Costco, Ellis worked for nearly 20 years in retail management, including five years as a general manager for Sam's Club (Costco's chief competitor). According to Ellis, she left Sam's Club, because she was actively recruited by Costco and promised promotion to GM within a year. On the other hand, Costco claims that it recruited Ellis because she misrepresented herself as a star at Sam's Club, when she had, in fact, lost her job for poor performance.

In Ellis's first year with Costco, she transferred locations twice in order to further her goal of promotion to GM. During this time, several GM positions became available, but she did not learn of the openings until after they were filled. In 2000, Ellis transferred to Colorado to assist her sick mother. According to Ellis, a supervisor told her that it would not hurt her chances for promotion. After six months, Ellis notified Costco that she was again able to relocate anywhere as a GM.

In 2002, Ellis sent a letter to her supervisors expressing a " burning desire" to help Costco be successful and advance within the company, asking how the GM selection process worked, where she stood as a candidate for promotion, and what she needed to do to become a GM. In October 2002, Ellis, while still employed with Costco as an AGM, filed a gender discrimination charge with the Equal Employment Opportunity Commission (EEOC), alleging that she had been passed over for promotion to GM because she was female. Ellis left Costco in November 2004.

Ellis II, 657 F.3d at 976.

2. Horstman

The Ninth Circuit described Plaintiff Horstman as follows:

Leah Horstman worked for Costco for more than 23 years beginning in 1981. In 1996, after 15 years with Costco, Horstman was promoted to be a Senior Staff Manager. By 2000, Horstman had rotated through the Administrative Manager, Merchandise Manager, and Receiving Manager positions. She had earlier worked as an Assistant Front End Manager, but did not rotate to the Front End Manager position because of scheduling conflicts and her duties as a single mother with two young daughters.

Throughout her career with Costco, Horstman repeatedly expressed her interest in advancing to AGM and GM and questioned supervisors about the requirements for both positions. Heeding the advice of a supervisor, Horstman also transferred to a high-volume store and expressed a willingness to move from California to Texas in order to become an AGM. However, in her final three annual self-performance reviews, Horstman indicated that her goal was to stay in a position similar to that which she held at the time for three to five years so that she could balance her family life and then to continue her advancement to AGM and GM. Horstman filed a discrimination charge with the EEOC in October 2003 and resigned in June 2004.

Id.

3. Sasaki

The Ninth Circuit described Plaintiff Sasaki as follows:

Elaine Sasaki began working for Costco in 1985. Sasaki advanced to become a Senior Staff Manager within four years. She received consistently high performance reviews, and her GM first indicated that she was ready to be promoted to AGM in 1993. Although Sasaki offered to transfer to places as far away as Hong Kong, she was not promoted to AGM until 1996. Sasaki is currently an AGM in Visalia, California.

Since Sasaki was promoted to AGM in 1996, she has not been selected for at least eight GM positions. She claims she was not aware of any of these openings until after they were filled. Sasaki has relocated four times to improve her chances of promotion to GM. In September 2003, Sasaki wrote to Costco's director of human resources expressing her concern that she had not been promoted because of her gender. At least some of her concern stems from an incident in which she claimed to rebuff the advances of her regional Senior Vice President in a hotel elevator and was later told by him that he holds her to higher standards than other AGMs. According to Costco, Sasaki has not been promoted because both her performance appraisals and her self-evaluations identify areas for improvement. Further, she has never ranked high on Costco's GM promotable list. Sasaki filed a gender discrimination charge with the EEOC on March 1, 2005. She remains employed as an AGM at Costco.

Id. at 976-77.

B. Judge Patel's Prior Order

On the previous motion for class certification, Judge Patel made the following rulings.

1. Standing

First, the court determined that Plaintiffs had standing to bring suit for injunctive relief, even though two of the Plaintiffs were former, rather than current, employees. Ellis I, 240 F.R.D. at 637. Although the Ninth Circuit partially vacated that finding, standing is no longer at issue in this case as Plaintiffs seek to have Plaintiff Sasaki, a current employee, represent the injunctive relief class. Defendant raises no challenge on standing grounds. See Plfs' Reply at 2 n. 2 (so noting).

2. Numerosity

Second, the court concluded that Plaintiffs met the numerosity requirement for class certification under Rule 23(a), because the potential class was approximately 700 people. Ellis I, 240 F.R.D. at 638. That finding was not challenged on appeal, nor does Defendant raise it here.

3. Commonality

Third, Judge Patel found that Plaintiffs met the commonality requirement. The court considered the parties' " battle of the experts," including, on Plaintiffs' side, statistical comparisons of women's promotion rate at Costco as compared to their male counterparts and as compared to other similar companies, as well as social science and cognitive bias research showing that " the Costco culture and subjective promotion processes discriminate against women." Id. at 638-40. Defendant, on its side, presented expert testimony " that women are not underrepresented at Costco and that any gender disparities, if they exist, are confined to two regions of Costco." Id. at 638. Defendant presented further expert testimony " that gender disparities, if they exist, are based upon factors, such as women's lack of interest in jobs requiring early morning hours, which are unrelated to Costco's culture and promotion processes." Id. In addition, Defendant moved to strike Plaintiffs' experts' testimony, arguing that it was irrelevant and unreliable. Id.

Considering these competing accounts, the court found that " plaintiffs have provided sufficient evidence of gender disparities in the promotion of women to GM and AGM to raise a common issue of triable fact." Id.; see also id. at 639 (" [P]laintiffs have presented compelling evidence of gender disparities at this time sufficient to demonstrate class-wide impact." ). The court found that Defendant's challenges to Plaintiffs' evidence (including, e.g., Plaintiffs' use of aggregate rather than regional data) " attack the weight of the evidence and not its admissibility." Id.; see also id. (" Plaintiffs['] expert declarations, while questioned by defendants, are strong enough to establish commonality" as to classwide gender disparities).

In addition to the evidence of classwide disparities, Judge Patel found that Plaintiffs had presented sufficient evidence of a uniform promotion practice that caused this classwide gender disparity. Specifically, the court found that

Costco has a consistent promotion-from-within policy. Costco has no written selection criteria for promotion to GM and AGM nor any formal notification or application procedures for those positions. This process involves decision-making by central management, including the maintenance of a Green Room at corporate headquarters where the photographs of potential promotees are displayed. Indeed, there was a conscious decision made not to post vacancies for GM and AGM. See Sinegal Dep. 123:22-124:2, 124:9-16, Exh. 1. Defendants argue that promotion decisions to AGM and GM vary by region and that promotion decisions to AGM are made at the store level. However, these assertions are unsupported and do not undermine commonality. Rather, the evidence before the court indicates that officers at the regional and corporate levels are involved in promotion decisions. The absence of written criteria for promotion as well as other evidence that the process is subjective satisfies the court that there is a policy common to the class in this regard. See Bates v. UPS, 204 F.R.D. 440, 448 (N.D.Cal.2001) (Henderson, J.).

Id. The court further found that Plaintiffs had presented evidence through their social science expert (Dr. Reskin) that this common policy " has a gender-differentiated impact on the promotion to AGM and GM." Id. The court concluded that " the defendants' objections to Dr. Reskin's study do not convince the court that it is irrelevant or unreliable." Id. at 640.

The court rejected Costco's argument that any gender disparities are " attributable to the gender-differentiated supply in the pool of qualified employees for AGM and GM positions ... [which] is reflective of women's differential interest in jobs requiring works hours incompatible with family responsibilities." Id. On this point, the court concluded that " [w]hat defendant has presented is an argument that is common to the class, and, if anything, supports the commonality factor. In evaluating all of the evidence presented, the court finds that plaintiffs have presented strong evidence of a common culture at Costco which disadvantages women." Id.

Finally, the court concluded that, based on Plaintiffs' evidence that Costco itself regarded evidence of gender disparities in the company as a company-wide issue, " Costco's treatment of gender disparities in promotion to Senior Staff as well as AGM and GM positions as a company-wide issue establishes a common question of fact." Id. (discussing Costco's companywide BOLD Initiative and efforts to increase diversity in the ranks of management companywide). Accordingly, the court determined that Plaintiffs had demonstrated " common issues of fact and theories of law as to gender disparities in promotions to AGM and GM, the nature of Costco's culture and its effect on women." Id.

Relevant to the court's analysis of commonality, the court largely denied Defendant's motion to strike Plaintiffs' experts under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). The court found that each of Plaintiffs' experts met the Daubert test, with one partial exception. The court struck " Dr. Drogin's analysis of the average years to reach AGM found in ¶ 22 and Table 8 of the Drogin Declaration," because he had improperly " truncated the time period analyzed in order to reach the conclusion that a statistically significant disparity exists in the average number of years men and women take to be promoted from Senior Staff positions to AGM positions." Ellis I, 240 F.R.D. at 648. The court rejected Defendant's remaining Daubert challenges to Plaintiffs' experts' conclusions. Defendant did not challenge the court's Daubert findings on appeal, and the Ninth Circuit specifically noted that the court's rulings on this issue survived review. Ellis II, 657 F.3d at 981 n. 6 (" [W]e affirm the district court's decision to not strike Plaintiffs' experts' declarations and consider Costco's challenges to Plaintiffs' experts only in our review of the district court's Rule 23(a) analysis." ).

4. Typicality

Fourth, the district court determined that Plaintiffs had satisfied the typicality requirement. The court " reject[ed] Costco's contention that the named plaintiffs have unique claims," such as sexual harassment or breach of contract claims that would differ from the class as a whole. Ellis I, 240 F.R.D. at 641. The court further concluded that " Costco's allusions to unique defenses [are not] sufficient to defeat typicality. The court need not address the merits of each of the proposed defenses; rather, it is enough to say that as a general matter, individualized defenses do not defeat typicality." Id. (citing Hanon v. Dataproducts Corp., 976 F.2d 497, 508-09 (9th Cir.1992)).

5. Adequacy

Fifth, the court concluded that the Named Plaintiffs were adequate representatives of the class. Id. at 641-42. As discussed further below, although the Ninth Circuit vacated this finding in part, adequacy is no longer at issue in this case based on Plaintiffs' new proposed 23(b) class structure.

6. Rule 23(b)(2)

Sixth, the court found that Plaintiffs met the requirements for certification under Rule 23(b)(2). The court concluded that Plaintiffs' injunctive relief claims predominated, and that therefore Plaintiffs' compensatory and punitive damages claims could properly be considered within a(b)(2) class as well. Id. at 642-43. The court found that " [t]he need for individualized determinations of compensatory damages need not defeat certification under Rule 23(b)(2); rather, the court can accommodate this need by bifurcating the trial into different phases." Id. at 643. The court further concluded that " [p]unitive damages claims are suitable for certification under Rule 23(b)(2) because such a claim focuses on the conduct of the defendant and not the individual characteristics of the plaintiffs." Id. (citation omitted).

C. Ninth Circuit's Opinion

On appeal, the Ninth Circuit affirmed in part and vacated in part Judge Patel's decision. Ellis II, 657 F.3d 970. As relevant to the current motions, the Ninth Circuit vacated the district court's findings regarding commonality, finding that the district court had not engaged in the required " rigorous analysis" of Plaintiffs' contentions regarding commonality, including considering the merits as necessary, but instead had erroneously applied a Daubert standard to the parties' proffered evidence without examining its persuasiveness. Id. at 981. The court further found that the district court had erred in its typicality analysis by failing to address Defendant's purported unique defenses to the Named Plaintiffs' claims. Id. at 984-85. Under Rule 23(b), the Ninth Circuit found that the district court had erred in certifying the class under Rule 23(b)(2) because Dukes had called into doubt the notion that monetary relief could be available through a(b)(2) class. Id. at 986. Accordingly, the Ninth Circuit vacated the district court's (b)(2) finding and remanded for the court to apply the Dukes standard and determine whether the class could be certified under (b)(2) and/or (b)(3). Id. at 987. Because these specific rulings are crucial to the Court's task on remand, the Court will describe the Ninth Circuit's rationale with respect to each decision in more detail in the sections below.

The Ninth Circuit made other rulings that are not directly raised by the parties in the current round of motions. First, it found that at least one plaintiff, Elaine Sasaki, had standing. Id. at 974. As noted above, Defendant does not raise any challenge regarding standing before this Court. Second, it found that Ms. Sasaki was an adequate class representative to pursue injunctive relief, but that the other two Named Plaintiffs could not adequately represent an injunctive relief class because they were former employees. Id. at 974-75. However, Defendant does not currently challenge the adequacy of any Plaintiff before this Court, because Plaintiffs now seek certification of a(b)(2) injunctive relief class represented by Sasaki, and a(b)(3) damages class represented by all three Named Plaintiffs. Accordingly, adequacy is not at issue in the parties' current motions.

D. Plaintiffs' New Class Certification Proposals

On remand, Plaintiffs have made certain changes to their class certification proposals. Specifically, Plaintiffs propose a hybrid class certification approach under Rule 23(b), and request that the Court certify (1) an injunctive relief class of current employees only under (b)(2) represented by Plaintiff Sasaki, through which the Court would determine liability and injunctive relief; and (2) a monetary relief class of both current and former employees under (b)(3) represented by all Named Plaintiffs, through which the Court would adjudicate all remedies, including punitive damages. Plaintiffs alternatively request that the Court certify injunctive and classwide liability claims as " particular issues" under 23(c)(4).

Plaintiffs propose the following trial plan. In Stage One (Part One) of the proceedings, Plaintiffs propose that the jury decide: (1) Whether Costco has engaged in a pattern or practice of discrimination (liability for classwide disparate treatment); (2) Whether Costco's conduct meets the standard for an award of punitive damages; (3) If liable for punitive damages, the aggregate amount of punitive damages owed to the class (with the Court retaining discretion to adjust said damages after Stage Two); and (4) Whether Costco is liable to the named Plaintiffs for gender discrimination and, if so, in what amount. Plaintiffs propose that the Court will then decide whether Costco's employment practices have had an adverse impact on the class (prima facie case of disparate impact). In Stage One (Part Two) of the proceedings, Plaintiffs propose for the Court to determine (1) Whether Costco's employment practices were justified by business necessity (defense to the disparate impact claim), and if so, whether there was a less discriminatory alternative; and (2) In the event of a liability finding, appropriate injunctive relief. Finally, in Stage Two, Plaintiffs propose individual hearings to determine back pay and compensatory damages and to adjudicate individual defenses. The Court would then adjust the punitive damages award to reflect any due process concerns as to its proportionality to actual damages awarded.

Defendant challenges Plaintiffs' proposed class certification (or recertification) on several grounds. First, Costco argues that Plaintiffs cannot meet the commonality or typicality requirements of Rule 23(a). Defendant does not raise a challenge as to numerosity or adequacy. Second, it argues that Plaintiffs do not meet the requirements of either (b)(2) or (b)(3) for class certification. Third, it argues that Plaintiffs' punitive damages claims are not certifiable. Finally, it argues that certification under 23(c)(4) would be inappropriate.

III. DISCUSSION

A. Evidentiary Dispute Re Costco's Declarations

As a preliminary matter, the Court notes that Costco submits numerous declarations from its employees in support of its motion to eliminate class claims. However, Plaintiffs inform the Court that such declarations were the subject of considerable dispute during the prior round of class certification briefing before Judge Patel. See Plfs' Cross-Mot., Docket No. 664, at 3 n. 1. Specifically, Defendant procured these declarations after discovery had closed and without notice to counsel or disclosure of the declarants. In response to Plaintiffs' motion to strike the declarations, Judge Patel ordered Defendant to produce the declarants for cross-examination or withdraw them. The parties subsequently stipulated that Defendant would withdraw portions of the declarations and that the remaining portions would be considered for Rule 23 purposes only. Specifically, the parties stipulated, in relevant part, to the following:

(1) Language in putative class member declarations (including the heading " NO GENDER DISCRIMINATION" ) stating the conclusion that the declarant or anyone else has never experienced or observed gender discrimination or that Costco has never discriminated against women or words to that effect is withdrawn and will not be considered by the Court.

(2) Language in putative class member declarations that plaintiffs do not represent the declarant's interests or that she does not wish to represented by plaintiffs or their counsel or words to that effect is withdrawn and will not be considered by the Court.

(3) The remainder of the declarations may be cited for Rule 23 purposes only, and neither side waives the right to argue the weight to be given to said declarations. In particular, the declarations may not be used to impeach the testimony of any class member or to limit her remedies or representation in this case.

Stipulation Re: Plfs' Mot. to Strike, Docket No. 484. Judge Patel accepted the stipulation and further ordered that the court may require testimony from the declarants if necessary, and that it may order testimony from a Costco 30(b)(6) witness regarding the preparation of Costco's declarations. Order Re: Plfs' Mot. to Strike, Docket No. 486, at 2.

Despite this stipulation, Defendant cites to withdrawn portions of these declarations in its briefing and in other declarations by, e.g., Defense counsel. Plaintiffs request that the Court strike those portions of the declarations already withdrawn under the parties' previous stipulation. Defendant does not respond to Plaintiffs' argument. Accordingly, the Court STRIKES those portions of the declarations Defendant was already ordered to withdraw, and will not consider Defendant's references to declarants' statements denying discrimination.

B. Motion for Class Certification

Rule 23 of the Federal Rules of Civil Procedure permits Plaintiffs to sue as representatives of a class only if

(1) the class is so numerous that joinder of all members is impracticable;

(2) there are questions of law or fact common to the class;

(3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and

(4) the representative parties will fairly and adequately protect the interests of the class.

Fed.R.Civ.P. 23(a). In addition, a purported class must be certified under Rule 23(b) by satisfying any one of its prongs. The two alternative prongs under which Plaintiffs seek certification permit classwide treatment if

(2) the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole; or

(3) the court finds that the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.

Fed.R.Civ.P. 23(b).

As noted above, Plaintiffs seek a hybrid certification of both a Rule 23(b)(2) (for liability and injunctive relief) and (b)(3) (for monetary relief) class. Plaintiffs seek to utilize a multi-staged bifurcated trial system, in which the parties would first litigate classwide liability and injunctive relief based on Plaintiffs' claims for pattern-or-practice intentional discrimination, as well as disparate impact.

As the Supreme Court reaffirmed in Dukes, pattern-or-practice cases alleging disparate treatment under Title VII typically follow a bifurcated, burden-shifting structure laid out by Int'l Broth. of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977):

In a pattern-or-practice case, the plaintiff tries to " establish by a preponderance of the evidence that ... discrimination was the company's standard operating procedure[,] the regular rather than the unusual practice." [ Teamsters, 431 U.S. at 336, 97 S.Ct. 1843]; see also Franks v. Bowman Transp. Co., 424 U.S. 747, 772, 96 S.Ct. 1251, 47 L.Ed.2d 444 (1976). If he succeeds, that showing will support a rebuttable inference that all class members were victims of the discriminatory practice, and will justify " an award of prospective relief," such as " an injunctive order against the continuation of the discriminatory practice." Teamsters, supra, at 361, 97 S.Ct. 1843.

....

We have established a procedure for trying pattern-or-practice cases that gives effect to [Title VII's] statutory requirements. When the plaintiff seeks individual relief such as reinstatement or backpay after establishing a pattern or practice of discrimination, " a district court must usually conduct additional proceedings ... to determine the scope of individual relief." Teamsters, 431 U.S., at 361, 97 S.Ct. 1843. At this phase, the burden of proof will shift to the company, but it will have the right to raise any individual affirmative defenses it may have, and to " demonstrate that the individual applicant was denied an employment opportunity for lawful reasons." Id., at 362, 97 S.Ct. 1843.

Dukes, 131 S.Ct. at 2552 n. 7, 2561. Using this framework, Plaintiffs propose to adjudicate liability for their pattern-or-practice disparate treatment claim in Stage One of the proceedings.

Similarly, Plaintiffs propose to adjudicate the initial phase of their disparate impact claim in Stage One, in which they would seek to " ‘ establish by a preponderance of the evidence that the employer " uses a particular employment practice that causes a disparate impact on the basis of race, color, religion, sex, or national origin." ’ " United States v. City of New York, 276 F.R.D. 22, 34 (E.D.N.Y.2011) (discussing Teamsters approach for both disparate impact and disparate treatment claims) (quoting Robinson v. Metro-N. Commuter R.R. Co., 267 F.3d 147, 158 (2d Cir.2001), abrogated on other grounds by Dukes, 131 S.Ct. at 2560-62; 42 U.S.C. § 2000e-2(k)(1)(A)(i)). Defendant would then have the opportunity to demonstrate that the challenged practices are " consistent with business necessity" and that there was no less discriminatory alternative. § 2000e-2(k)(1)(A)(i)-(ii).

If Plaintiffs prevail on either claim, the Court could fashion classwide injunctive relief. In addition, in Stage Two of the case, if Plaintiffs prevail on either of their claims, Plaintiffs propose Teamsters hearings in order to adjudicate individual claims for backpay or particularized injunctive relief and compensatory damages, as well as the individual's share of any punitive damages.⁶

6 While Defendant challenges Plaintiffs' citation to Teamsters and attempts to distinguish it on the basis that it " is a government-initiated case that did not implicate Rule 23," Teamsters has frequently been employed in the Title VII class action context, with the Supreme Court's approval. See, e.g., Dukes, 131 S.Ct. at 2552 & n. 7, 2556, 2561 (discussing Teamsters in the context of Rule 23 analysis); Cooper v. Federal Reserve Bank of Richmond, 467 U.S. 867, 876 n. 9, 104 S.Ct. 2794, 81 L.Ed.2d 718 (1984) (" Although Teamsters involved an action litigated on the merits by the Government as plaintiff under § 707(a) of the Act, it is plain that the elements of a prima facie pattern-or-practice case are the same in a private class action." ) (citation omitted); Chin v. Port Auth. of New York & New Jersey, 685 F.3d 135, 147 (2d Cir.2012) (describing " the burden-shifting framework set out in Teamsters and available both to the government in § 2000e-6 litigation and to class-action plaintiffs in private actions alleging discrimination" as recognized by Dukes and previous cases) (citations omitted); Hohider v. United Parcel Serv., Inc., 574 F.3d 169, 179 n. 11 (3d Cir.2009) (" Since Cooper, courts of appeals have used the Teamsters two-stage framework to analyze pattern-or-practice claims brought as private-plaintiff class actions under Title VII." ) (internal citations omitted); Robinson v. Metro-N. Commuter R.R. Co., 267 F.3d 147, 158 (2d Cir.2001) (using Teamsters approach in connection with Rule 23 motion for class certification); Lowery v. Circuit City Stores, Inc., 158 F.3d 742, 760 (4th Cir.1998) (" The courts of appeals have ... permitted pattern or practice class action suits using the Teamsters method of proof." ), vacated on other grounds, 527 U.S. 1031, 119 S.Ct. 2388, 144 L.Ed.2d 790 (1999); Easterling v. Connecticut Dept. of Correction, 278 F.R.D. 41, 48 (D.Conn.2011) (using Teamsters framework in disparate impact class action); Karp v. CIGNA Healthcare, Inc., 882 F.Supp.2d 199, ----, CIV.A. 11-10361-FDS, 2012 WL 1358652, at *8 (D.Mass. Apr. 18, 2012) (describing Teamsters pattern-or-practice claims as " a method of proof [under Title VII] with two components: an evidentiary component (that is, a plaintiff may prove unlawful discrimination through evidence that the company engaged in a pattern or practice of discrimination, and that the pattern or practice caused her injury) and a burden-shifting component (that is, when a plaintiff has proved the existence of a pattern or practice and an adverse employment action, the burden shifts to the defendant to disprove causation)" ).

Plaintiffs contend that each of their claims satisfies the Rule 23(a) factors and either the Rule 23(b)(2) or (b)(3) factors.

1. Rule 23(a)

Rule 23(a) requires a purported class to satisfy four criteria for certification: numerosity, commonality, typicality, and adequacy of the representatives. Fed. R. Civ. Pro. 23(a). Failure to meet any of the criteria defeats the motion. Rutledge v. Elect. Hose & Rubber Co., 511 F.2d 668, 673 (9th Cir.1975) (citation omitted).

a. Numerosity

Plaintiffs satisfy the numerosity requirement if " the class is so large that joinder of all members is impracticable." Hanlon v. Chrysler Corp., 150 F.3d 1011, 1019 (9th Cir.1998). In this case, Plaintiffs easily meet the numerosity requirement. See Wang v. Chinese Daily News, 231 F.R.D. 602, 607 (C.D.Cal.2005), vacated on other grounds, --- U.S. ----, 132 S.Ct. 74, 181 L.Ed.2d 1 (2011) (100 or more plaintiffs leads to a presumption of numerosity); Ikonen v. Hartz Mountain Corp., 122 F.R.D. 258, 262 (S.D.Cal.1988) (finding a purported class of forty members sufficient to satisfy numerosity) (citation omitted). Defendant does not contest this factor.

b. Commonality

Plaintiffs must demonstrate that there are " questions of law or fact common to the class" in order to satisfy Rule 23(a)(2). Plaintiffs need not demonstrate that all questions are common to the class; rather, it is sufficient if either " shared legal issues with divergent factual predicates" or " a common core of salient facts coupled with disparate legal remedies within the class" are present. Hanlon, 150 F.3d at 1019-20; see also, e.g., Johnson v. Gen. Mills, Inc., 278 F.R.D. 548, 551 (C.D.Cal.2012) (citing above passage from Hanlon post- Dukes ). " Even a single [common] question" will suffice to satisfy Rule 23(a). Dukes, 131 S.Ct. at 2556 (citation omitted). However, Plaintiffs must " demonstrate that the class members have suffered the same injury," not " merely that they have all suffered a violation of the same provision of law." Id. at 2551. Moreover, " What 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 omitted) (emphasis in original).

In order to assess commonality, " it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question." General Tele. Co. of S.W. v. Falcon, 457 U.S. 147, 160, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982) (" [T]he class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiff's cause of action." ) (quotation omitted); see also Dukes, 131 S.Ct. at 2551. " Here, the question of commonality overlaps with Plaintiffs' claim that Costco's system of promotion and corporate culture constitutes a pattern or practice of discrimination." Ellis II, 657 F.3d at 980. Thus, the court must consider the merits to the extent necessary to determine commonality. However, in peeking at the merits to determine whether commonality is present, " [t]he court may not go so far ... as to judge the validity of these claims." USW v. ConocoPhillips Co., 593 F.3d 802, 808-09 (9th Cir.2010) (quoting Staton v. Boeing Co., 327 F.3d 938, 954 (9th Cir.2003)); Ellis II, 657 F.3d at 983 n. 8 (" The district court is required to examine the merits of the underlying claim in this context, only inasmuch as it must determine whether common questions exist; not to determine whether class members could actually prevail on the merits of their claims.... To hold otherwise would turn class certification into a mini-trial." ) (internal citations omitted); In re Whirlpool Corp. Front-Loading Washer Products Liab. Litig., 678 F.3d 409, 417-18 (6th Cir.2012) (" [A] district court must resolve factual disputes necessary to class certification, but [ ] ‘ the court should not turn the class certification proceedings into a dress rehearsal for the trial on the merits.’ " ) (quoting Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 811 (7th Cir.2012)); see also id. (citing Third Circuit decision post- Dukes articulating same principle).

Because the parties' central dispute concerns the impact of Dukes on the viability of Plaintiffs' class claims, the Court first reviews Dukes's holding on commonality. The Court then reviews the Ninth Circuit's instructions in Ellis II to be applied on remand before proceeding to examine the parties' competing arguments as to commonality.

i. Dukes

Dukes addressed " one of the most expansive class actions ever." 131 S.Ct. at 2546. The Dukes class consisted of " one and a half million plaintiffs, current and former female employees of petitioner Wal-Mart who allege that the discretion exercised by their local supervisors over pay and promotion matters violates Title VII by discriminating against women." Id. These employees were spread across thousands of stores, in varying positions inside and outside of management. Indeed, Plaintiffs purported to represent all female Wal-Mart employees and challenged Wal-Mart's pay and promotion policies throughout its hierarchy. Id. at 2557 (commenting that class members " held a multitude of different jobs, at different levels of Wal-Mart's hierarchy, for variable lengths of time, in 3,400 stores, sprinkled across 50 states, with a kaleidoscope of supervisors (male and female), subject to a variety of regional policies that all differed" ) (quoting Dukes, 603 F.3d at 652 (Kozinski, J., dissenting)). Plaintiffs' theory of commonality was " that a strong and uniform ‘ corporate culture’ permits bias against women to infect, perhaps subconsciously, the discretionary decisionmaking of each one of Wal-Mart's thousands of managers— thereby making every woman at the company the victim of one common discriminatory practice." Id. at 2548.

Addressing the standard courts must apply in assessing commonality, the Supreme Court explained that " [c]ommonality requires the plaintiff to demonstrate that the class members have suffered the same injury. This does not mean merely that they have all suffered a violation of the same provision of law." Id. at 2551 (internal citations and quotation marks omitted). Instead, plaintiffs' " claims must depend upon a common contention ... 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. Dissimilarities within the proposed class are what have the potential to impede the generation of common answers." Id. (internal quotation omitted) (emphasis in original).

In other words, " [w]ithout some glue holding the alleged reasons for all those 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. (emphasis in original). In the context of pattern-or-practice discrimination cases alleging disparate treatment, the Supreme Court specifically identified the " glue" that might create commonality as relevant to this case: " [s]ignificant proof that an employer operated under a general policy of discrimination conceivably could justify a class of both applicants and employees if the discrimination manifested itself in hiring and promotion practices in the same general fashion, such as through entirely subjective decisionmaking processes." Id. at 2553 (quoting Falcon, 457 U.S. at 159 n. 15, 102 S.Ct. 2364).

In Dukes, the Court found such " significant proof" " entirely absent," because the " only evidence of a ‘ general policy of discrimination’ respondents produced was the testimony of" their sociological expert as to Wal-Mart's " strong corporate culture." Id. at 2553. Instead, the Court found that the only common " policy" the plaintiffs had demonstrated was one of allowing discretion, which was merely " a policy against having uniform employment practices." Id. at 2554 (emphasis in original). The Court reasoned that the exercise of discretion itself is not evidence of any discriminatory policy. While such a policy could, in certain instances, constitute a discriminatory practice, " the recognition that this type of Title VII claim ‘ can’ exist does not lead to the conclusion that every employee in a company using a system of discretion has such a claim in common." Id. at 2554. Because some people may exercise discretion in permissible ways and others may not, Plaintiffs must " identif[y] a common mode of exercising discretion that pervades the entire company." Id. at 2554-55.

The Court found that the Dukes class had failed to identify such a " common mode" because " [i]n a company of Wal-Mart's size and geographical scope, it is quite unbelievable that all managers would exercise their discretion in a common way without some common direction. " Id. at 2555 (emphasis added). Plaintiffs presented no direct evidence of such a " common direction," nor did they produce any persuasive statistical or anecdotal evidence indicating such common direction. Specifically, because decisions were made at the store level (and not directed from higher levels of the management hierarchy), the Court found that statistical evidence of regional disparities were not probative of commonality because such disparities might be explained by only a small subset of stores within each region. In addition, the Court found that even if there was a disparity in every store, " that would still not demonstrate that commonality of issue exists" unless plaintiffs successfully identified a " specific employment practice" or " common direction" from upper management that tied the claims of all members of the class together. Id.

As Dukes acknowledged, the Supreme Court's own prior precedent has recognized that the exercise of discretion may provide the basis of a Title VII claim. This is especially so in the context of a disparate impact claim. The disparate impact context highlights the rationale behind such a claim, as an employer's policy allowing for discretion could lead to lower-level managers' discriminatory implementation of said discretion without any proven animus on the part of those who granted the discretion. Dukes acknowledged that " in appropriate cases,' giving discretion to lower-level supervisors can be the basis of Title VII liability under a disparate-impact theory— since ‘ an employer's undisciplined system of subjective decisionmaking [can have] precisely the same effects as a system pervaded by impermissible intentional discrimination.’ " 131 S.Ct. at 2554 (quoting Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 990-91, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988) (" If an employer's undisciplined system of subjective decisionmaking has precisely the same effects as a system pervaded by impermissible intentional discrimination, it is difficult to see why Title VII's proscription against discriminatory actions should not apply.... We conclude, accordingly, that subjective or discretionary employment practices may be analyzed under the disparate impact approach in appropriate cases." )). However, even in the disparate impact context, the Court concluded, a mere " policy" of allowing discretion is insufficient to raise a common question for class action purposes. Instead, a proposed class must identify a " specific employment practice" under which said discretion operates to state a disparate impact claim. Dukes, 131 S.Ct. at 2555 (citing Watson, 487 U.S. at 994, 108 S.Ct. 2777 (" [T]he plaintiff's burden in establishing a prima facie case goes beyond the need to show that there are statistical disparities in the employer's work force. The plaintiff must begin by identifying the specific employment practice that is challenged." )). Without such a specific challenged practice, nothing ties the discretionary decisions together. Id. at 2556.

Thus, Dukes concluded that the plaintiffs had demonstrated no common issues with respect to either their pattern-or-practice disparate treatment or their disparate impact claims. Id. at 2557.

As explained further below, the proposed classes in the instant case differ from that examined in Dukes in several material ways. First, the size of the class at issue is a mere fraction of that in Dukes . Although class size has no per se bearing on commonality, when the claims focus in part on the exercise of managerial discretion, it is reasonable to suspect that the larger the class size, the less plausible it is that a class will be able to demonstrate a common mode of exercising discretion. See Chen-Oster v. Goldman, Sachs & Co., 877 F.Supp.2d 113, ----, 10 CIV. 6950 LBS JCF, 2012 WL 2912741, at *3 (S.D.N.Y. July 17, 2012) (" The Supreme Court suggested (when not explicitly stating) that the sheer size of the [ Dukes ] class and the vast number and diffusion of challenged employment decisions was key to the commonality decision. This makes a great deal of sense when the purpose of the commonality enquiry is to identify ‘ some glue holding the alleged reasons for all of [the challenged] employment decisions together.’ " ) (quoting Dukes, 131 S.Ct. at 2552) (emphasis omitted); Ross v. RBS Citizens, N.A., 667 F.3d 900, 909 (7th Cir.2012) (" In Dukes, 1.5 million nationwide claimants were required to prove that thousands of store managers had the same discriminatory intent in preferring men over women for promotions and pay raises." ). As this case only involves applicants to GM and AGM positions, the class size of approximately 700 is a fraction of the 1.5 million claimants in Dukes .

Second, the scope of the proposed class is far more limited and focused than that in Dukes . Here, the class covers only two closely-related, management-level positions (AGM and GM) that share a uniform job description across the class, and presents a targeted challenge to the failure to promote women into those positions. In contrast, Dukes covered women in all positions at Wal-Mart in thousands of stores and raised claims based on both pay and promotion policies. Thus, the scope of this class and its claims are worlds away from Dukes .

Third, and most important, Plaintiffs in this case identify specific employment practices Costco implements companywide under the influence and control of top management. Unlike in Dukes, which the Supreme Court concluded merely identified the delegation of discretion ( i.e., the absence of a policy), here Plaintiffs identify specific practices and a common mode of guided discretion directed from the top levels of the company. Cf. Dukes, 131 S.Ct. at 2555 (" In a company of Wal-Mart's size and geographical scope, it is quite unbelievable that all managers would exercise their discretion in a common way without some common direction. " ) (emphasis added). It is this " common direction" and the identification of specific practices (other than the mere general delegation of authority), in addition to the smaller size and scope of the class, that separates this case from Dukes . The Court details Plaintiffs' persuasive evidence of these practices that distinguish Dukes below.

ii. Ellis II

In Ellis II, the Ninth Circuit vacated Judge Patel's holding regarding commonality and remanded for the Court to conduct the required " rigorous analysis" under Rule 23 to determine whether Plaintiffs have provided significant proof of common employment practices. The Ninth Circuit directed the district court to determine whether Plaintiffs had identified " a common question that will connect many individual promotional decisions to their claim for class relief," examining the merits as necessary to make that determination. Ellis II, 657 F.3d at 981. The court also identified some points of agreement with Judge Patel's decision in Ellis I . Specifically, the court stated, " we agree that the district court was not required to resolve factual disputes regarding: (1) whether women were in fact discriminated against in relevant managerial positions at Costco, or (2) whether Costco does in fact have a culture of gender stereotyping and paternalism." Id. at 983. Nonetheless,

the district court was required to resolve any factual disputes necessary to determine whether there was a common pattern and practice that could affect the class as a whole. If there is no evidence that the entire class was subject to the same allegedly discriminatory practice, there is no question common to the class. In other words, the district court must determine whether there was " significant proof that [Costco] operated under a general policy of discrimination."

Id. (emphasis in original) (quoting Dukes, 131 S.Ct. at 2553 (alteration omitted)). The court found that, in examining this evidence, " the district court seems to have confused the Daubert standard it correctly applied to Costco's motions to strike with the ‘ rigorous analysis' standard to be applied when analyzing commonality.... [T]o the extent the district court limited its analysis of whether there was commonality to a determination of whether Plaintiffs' evidence ... was admissible, it did so in error." Id. at 982. Instead, " the district court was required to resolve any factual disputes necessary to determine whether there was a common pattern and practice that could affect the class as a whole." Id. at 983 (emphasis in original).

A key factual question on remand, the Ninth Circuit noted, was whether disparities existed across Costco's eight regions or whether such disparities were merely confined to a few regions. Disparity across the regions would lend credence to Plaintiffs' allegations that " promotion decisions are based on the biased attitudes of the CEO and upper management." Id. Because the parties' experts disputed the existence of disparities across regions, the Ninth Circuit instructed the district court on remand to " examin[e] the merits to decide this issue." Id. at 984. The Court's task on remand is therefore to " judg[e] the persuasiveness of the evidence presented" on the issue of commonality. Id. at 982.

The Ninth Circuit's opinion did not explicitly consider any differences between Plaintiffs' disparate treatment and disparate impact claims as they relate to commonality. However, because there are differences with respect to the way Dukes might be applied to disparate treatment, as opposed to disparate impact claims, the Court will address these claims separately below. First, the Court concludes that, with respect to the Plaintiffs' disparate treatment claim, Plaintiffs have provided " significant proof" that Defendant operates under a general policy of discrimination and that Defendant's management utilizes a " common mode of exercising discretion," and have therefore satisfied the standard for certification under Rule 23(a). Dukes, 131 S.Ct. at 2553, 2554. Second, the Court further concludes that commonality is even clearer with respect to disparate impact, as Plaintiffs have identified specific employment practices they allege have caused the gender disparity in promotions to AGM and GM positions.

iii. Disparate Treatment

Plaintiffs argue that the Ninth Circuit only vacated and remanded for reconsideration of the specific issue it mentioned in the context of commonality: whether the gender disparity in promotions is consistent across the company, rather than a problem isolated to certain regions. See Plfs' Cross-Mot., Docket No. 664, at 20. Thus, Plaintiffs contend the following finding from Judge Patel still stands and demonstrates that Costco has uniform personnel and promotion policies:

Defendants argue that promotion decisions to AGM and GM vary by region and that promotion decisions to AGM are made at the store level. However, these assertions are unsupported and do not undermine commonality. Rather, the evidence before the court indicates that officers at the regional and corporate levels are involved in promotion decisions. The absence of written criteria for promotion as well as other evidence that the process is subjective satisfies the court that there is a policy common to the class in this regard. See Bates v. UPS, 204 F.R.D. 440, 448 (N.D.Cal.2001) (Henderson, J.).

Ellis I, 240 F.R.D. at 639. Although one could read Ellis II to support Plaintiffs' argument— indeed, the fact section of Ellis II indicates that Costco operated under common, companywide promotion policies and practices— other portions of the Ninth Circuit's opinion suggest this Court should reexamine commonality more broadly than simply the regional-versus-national disparity debate. See Ellis II, 657 F.3d at 983 n. 7 (commenting on the parties' dispute over " whether promotional decisions for AGMs were made by the GM of the local warehouse, or were made or strongly influenced by upper management at Costco headquarters" as relevant to commonality). Thus, out of an abundance of caution, the Court will examine the parties' evidence pertaining to commonality as a whole without treating Judge Patel's disputed findings as conclusive on this question except where affirmed by the Ninth Circuit.

In the instant case, Plaintiffs have produced " significant proof" that " the entire class was subject to the same allegedly discriminatory practice[s]." Ellis II, 657 F.3d at 983. This proof comes in three general categories. First, Plaintiffs produce persuasive evidence of numerous common policies and practices under which Costco conducts promotions to AGM and GM. Second, Plaintiffs similarly demonstrate a pervasive companywide culture that, along with the common policies and practices, guide Costco managers' discretion in making promotion decisions. Plaintiffs allege that this culture generates and reinforces discriminatory outcomes. Third, Plaintiffs demonstrate classwide effects purportedly caused by said policies and practices affecting all regions (the specific question identified by the Ninth Circuit for the Court to address on remand).

A. Classwide Policies & Practices— Costco's Promotion System

1. High Level Management Involvement

As to the first category, Plaintiffs identify a common, companywide promotion system within Costco that is made up of numerous common components. Specifically, reviewing the record as a whole, there are several companywide policies and practices that comprise Costco's promotion system, most of which Defendant concedes.

First, Costco imposes the same recruitment and selection process for promotion to AGM and GM across the company, both in terms of the persons generally involved in promotion decisions and the process by which they make those decisions. Costco has a policy and practice of promoting " virtually 100%" from within. Sinegal Depo., Docket No. 665, Ex. 8, at 42. Further, it has imposed a conscious policy and practice against job posting for AGM and GM openings. Matthews Depo., Docket No. 665, Ex. 5, at 150-51 (in response to recommendation that company impose job posting for all positions, company established posting only for positions below AGM and GM); Sinegal Depo., Docket No. 665, Ex. 8, at 123-24 (CEO Sinegal has " always felt very ... strongly and very adamantly, that [AGM and GM positions] were not the types of jobs that should be up for posting" ); id. at 128 (acknowledging having described the idea of posting for AGM and GM positions as " bull shit" ). ⁷ Costco similarly does not post for rotations within Senior Staff Manager (" SSM" ) positions, the positions from which AGM candidates are promoted. Posting is only required to fill an open SSM position. See, e.g., Salaried Job Posting Policy, Docket No. 665, Ex. 27. Costco does not use an application or interview process for AGM and GM positions; employees are tapped on the shoulder for advancement. See, e.g., Glenn Depo., Docket No. 665, Ex. 3, at 22; see also Petty Decl., Docket No. 615, ¶ ¶ 37-38; Sasaki Decl., Docket No. 143, ¶ 15. Costco does not use internal written procedures for selecting promotion candidates. See, e.g., Zook Depo., Docket No. 665, Ex. 11, at 78-79.

7 Unlike in Dukes, which addressed the mere absence of a posting requirement— and the fact that local discretion led to inconsistent job posting— here Plaintiffs offer unrebutted evidence that Costco has an affirmative companywide policy against posting for the jobs at issue. Compare Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137, 149 (N.D.Cal.2004) (describing the lack of consistent job posting), with Salaried Job Posting Policy, Docket No. 665, Ex. 27 (Costco job posting policy stating, " All salaried positions in the warehouse EXCEPT the [AGM] and the [GM] must be posted" ).

Top management's involvement in the promotion process is also consistent, and pervasive, classwide. For GM promotions, the Regional and Senior VPs are the primary sources of candidate recommendations, but the decisions must be approved all the way up the chain of command. See Sinegal Depo., Docket No. 665, Ex. 8, at 23-24 (CEO reviews candidates for GM and " in most instances I know who they are" ); DiCerchio Depo., Docket No. 665, Ex. 1, at 48-50 (COO and CEO approve GM selections, with input from district manager, regional manager, and EVP); Portera Depo., Docket No. 544, Ex. O, at 84-85 (the regional and senior VPs provide EVP recommendations for GM promotions); Schutt Depo., Docket No. 544, Ex. U, at 74-75 (EVP receives recommendations from Senior VPs for GM promotions, discusses possible alternative candidates, and reports recommendation to his boss and to Sinegal); Zook Depo., Docket No. 544, Ex. X, at 133-35 (EVP receives notice of GM promotions, along with DiCerchio and Sinegal); Zook Decl., Docket No. 653, ¶ 14 (EVP " oversee[s] the promotion process involving both GMs and AGMs" ); Larkin Decl., Docket No. 665, Ex. 20, ¶ 8 (Letter from Defense counsel to Plaintiffs' counsel stating, " For GM promotions, the persons personally informed before final decision would be District VP, Senior VP, Executive VP, CEP" ). As part of upper management's involvement, the company uses a GM " promotables" list, generated and maintained at the regional level and reviewed at higher executive levels, as the pool from which candidates are chosen. See Def's Mot. to Eliminate Class Claims, Docket No. 543, at 7; Portera Depo., Docket No. 544, Ex. O, at 84 (EVP requests lists of promotables from AGM to GM once or twice a year); id. at 108-09 (evaluation process for GM candidates is ongoing at all levels of the company, and " these individuals are constantly monitored" ); Zook Decl., Docket No. 653, ¶ 19; Booth Decl., Docket No. 558, ¶ 3; Hoover Decl., Docket No. 586, ¶ 11; Omoss Decl., Docket No. 611, ¶ ¶ 5, 14. Candidates for GM are also displayed in the Green Room at company headquarters. Sinegal Depo., Docket No. 665, Ex. 8, at 38-39, 65-69. More broadly, Costco senior management, including the CEO and EVPs as well as District and Senior VPs, maintains a constant ongoing dialogue with warehouse-level management about candidates for promotion, and CEO Sinegal reviews lists of promotable candidates. Matthews Depo., Docket No. 665, Ex. 5, at 148-50.⁸

8 Although Mr. Matthews does not specify, the context of his testimony as to upper management's " ongoing dialogue" regarding personnel who may be ready for promotion indicates that he was testifying about management's involvement in the promotion process as a whole, including both AGM and GM candidates. See Matthews Depo., Docket No. 665, Ex. 5, at 148-50.

With respect to AGM selection, Costco characterizes these decisions as made locally, and reviewed regionally. Based on the Court's review of the record, this understates the extent of regional and senior executives' involvement in such decisions and the consistency of the challenged policies across AGM and GM positions. See Ellis II, 657 F.3d at 985 (" The Costco policies and culture challenged by Plaintiffs apply equally to AGM and GM promotion decisions." ). Instead, the weight of the evidence supports the view that senior management oversees and directs the promotion process for AGM candidates. First, and most importantly, senior executives— including EVPs and the CEO— describe their continued oversight of and involvement in the AGM promotion process. See, e.g., Sinegal Depo., Docket No. 544, Ex. V, at 41-42 (AGM decisions are made at the regional level, and as CEO he has given " [a] lot" of instructions as to how to fill AGM positions); Schutt Depo., Docket No. 544, Ex. U, at 76, 95-96 (AGM decision made at the regional level, with the GM, regional manager, regional VP, and Senior VP involved, and " [m]any times" the Senior VP will approve it); Portera Depo., Docket No. 544, Ex. O, at 122 (GM makes the " decision and recommendation" for AGM promotions " in conjunction with their regional vice presidents, and then the ultimate approval is given by a senior vice president of the region" ); Zook Depo., Docket No. 544, Ex. X, at 132, 135 (characterizing AGM promotions as being made by GM and District Manager, but also stating that EVP has a role in making sure his regional managers and GMs have a good pool of people promotable to AGM); Zook Decl., Docket No. 653, ¶ ¶ 7, 14-15 (stating that EVP is informed of the decision and EVP " oversee[s] the promotion process involving both GMs and AGMs" ); Hoover Depo., Docket No. 544, Ex. I, at 91-92 (GM, regional VP, and Senior VP are involved in the AGM promotion decision, and the EVP is informed of the decision; stating further that " [b]y the time you're appointed assistant manager, everybody knows that person really well" ); Omoss Decl., Docket No. 611, ¶ 5 (characterizing GMs as the primary decisionmakers with respect to AGM promotions, but admitting that as Texas VP he is " actively involved in the promotions of both GMs and AGMs," that GMs consult with him about candidates for promotion as he knows them from his visits to the warehouses, and that he keeps and updates lists of promotable candidates for AGM based on his own observations as well as those of GMs and other Costco employees); see also Larkin Decl., Docket No. 665, Ex. 20, ¶ 8 (Letter from Defense counsel to Plaintiffs' counsel stating that the District VP, Senior VP, and Executive VP are all personally informed before a final AGM decision).

Second, promotable lists for AGMs are maintained at the regional level, indicating regional executives' involvement in the process of identifying and selecting candidates for those positions. Def's Mot. to Eliminate Class Claims, Docket No. 543, at 7 (" Costco has written lists to identify managers promotable to GM and AGM (as well as other positions)." ); Mar Decl., Docket No. 598, ¶ 15 (Regional VP used an AGM promotable list); Hickey Decl., Docket No. 583, ¶ 23 (as a GM, provides list of promotables for AGM and GM to higher ups); Gaherty Depo., Docket No. 665, Ex. 2, at 58-59 (promotable list for AGM maintained in the region); Zook Decl., Docket No. 653, ¶ 19(f) (promotable list for AGM maintained at regional level).

Third, class member declarations submitted by Costco indicate that decisions occur at the regional level and that upper management is involved in the promotion process. See, e.g., Morgan Decl., Docket No. 608, ¶ 20 (promoted to AGM by her Regional VP in Bay Area); Hagemeyer Decl., Docket No. 580, ¶ 17 (GMs and VPs evaluate candidates for promotion to AGM during floor walks); Mendoza Decl., Docket No. 604, ¶ 24 (GMs and VPs evaluate AGM promotables); Loveland Decl., Docket No. 594, ¶ 20 (AGM decision made by Regional Manager or VP based on GM's recommendation); D'Agostino Decl., Docket No. 565, ¶ 30 (VPs make final decisions for AGM promotions); Bolger Decl., Docket No. 557, ¶ 10 (received rotation to MM position after conversation with District VP/Regional Operations Manager about advancement into AGM position); Peterson Decl., Docket No. 614, ¶ 15 (did not make final decision as GM about AGM promotions; decision occurred at the regional level); Ward Decl., Docket No. 631, ¶ ¶ 32, 35 (SVP was informed of her interest in AGM position; GMs report information about people who are interested in warehouse management (AGM or GM) to SVPs); Hickey Decl., Docket No. 583, ¶ ¶ 3-4, 15, 23 (selected for AGM by District and Regional VPs; senior management team evaluates candidates); Laureano Decl., Docket No. 592, ¶ ¶ 12, 20 (Regional VP is involved in AGM promotions); Greek Decl., Docket No. 578, ¶ 17 (Regional Managers evaluate AGM candidates).

Moreover, Costco management repeatedly describe the fact that they are part of a close-knit, centralized management team, and that regional personnel work closely with each other and with national management. See, e.g., Portera Depo., Docket No. 665, Ex. 6, at 110-11 (" The way we run our organization, although it's a more regional and division oriented, work closely together as a group and a team in developing policies and procedures, in developing individuals. There is quite a bit of interaction between the [EVPs] and the [SVPs] of the company." ); Zook Depo., Docket No. 665, Ex. 11, at 21 (operations meetings are held every four weeks at headquarters with everyone from Senior VP and up). As discussed further below, this close collaboration results in uniform standards for both AGM and GM promotions, according to the CEO himself. See Sinegal Depo., Docket No. 665, Ex. 8, at 42-44.

Indeed, as part of senior management's participation in, and influence over the promotion process to both AGM and GM, Costco uses " floor walks" as an additional opportunity for senior management to assess possible promotion candidates in stores on a regular basis and to engage in informal mentorship. See, e.g., Cafiso Decl., Docket No. 559, ¶ 18; Cline Decl., Docket No. 563, ¶ 21; Rosolino Decl., Docket No. 618, ¶ 32; Ward Decl., Docket No. 631, ¶ 29; Bejarano Decl., Docket No. 640, ¶ 25; Alvarez Decl., Docket No. 551, ¶ 9; Asch Decl., Docket No. 554, ¶ 15; D'Agostino Decl., Docket No. 565, ¶ 25. Top executives— including VPs and up to and including the CEO— participate in these walks to meet and evaluate store-level management. See, e.g., Def's Mot. to Eliminate Class Claims, Docket No. 543, at 7 (regional managers " visit warehouses frequently" ); see, e.g., Matthews Depo., Docket No. 665, Ex. 5, at 148-50 (the CEO and COO " travel with each of the regionals and their district VP's, and the EVP's are usually along in that same process, [through which they are] not only getting to meet and talk with the individuals that they're referring to, but in addition dealing with the district VP's about who it is that is most prepared at that point in time for additional responsibilities" ); Eringer Decl., Docket No. 567, ¶ ¶ 44-45 (regional and district managers, VPs, and CEO participated); Evans Decl., Docket No. 568, ¶ 30 (floor walks " are conducted by various levels of management, from the warehouse level all the way up to the CEO" ); Hinds Decl., Docket No. 584, ¶ 24 (VPs and CEO participated in walks). Executives explain that Costco's process of continual evaluation negates the need for formal, written selection processes because candidates are well-known to upper management by the time they are ready for promotion. See, e.g., Portera Decl., Docket No. 617, ¶ 5; Omoss Decl., Docket No. 611, ¶ ¶ 7-8; Hoover Decl., Docket No. 586, ¶ 12.

2. Common Criteria

Beyond the procedural ground rules governing promotion decisions, Costco imposes consistent substantive criteria for promotion into the AGM and GM positions that are well-known at least among those making the promotion decisions. For example, candidates for GM must be AGMs, and candidates for AGM must be SSMs. Sinegal Depo., Docket No. 665, Ex. 8, at 44; Zook Depo., Docket No. 544, Ex. X, at 69-71. Potential AGMs typically must have merchandising experience in the MM position, and generally should have experience in most if not all of the four SSM positions. See, e.g., Mulligan Decl., Docket No. 647, ¶ 31 (Defendant's expert asked to assume MM experience was most valuable for promotion); Kadue Declaration, Docket No. 544, ¶ ¶ 3, 4 (describing merchandising as positions that are critical to advancement, and listing declarations in support); Zook Decl., Docket No. 653, ¶ 13; Drogin Decl., Docket No. 666, ¶ 12 (over 85% of people promoted to AGM have MM experience). Costco also places a premium on schedule flexibility and the ability and willingness to relocate. See, e.g., Mulligan Decl., Docket No. 647, ¶ 32 (asked to so assume for purposes of defense expert report); Vachris Decl., Docket No. 627, ¶ ¶ 6, 9; Bolger Decl., Docket No. 557, ¶ ¶ 8, 16; Eringer Decl., Docket No. 567, ¶ 43; Evans Decl., Docket No. 568, ¶ 31; Hagemeyer Decl., Docket No. 580, ¶ ¶ 14-15; Moore Decl., Docket No. 607, ¶ ¶ 18-19; Spence Decl., Docket No. 624, ¶ 17; Ward Decl., Docket No. 631, ¶ 39; Whitney Decl., Docket No. 632, ¶ 25.

CEO Sinegal personally instructs his staff as to the criteria they should employ in making promotion decisions for AGM and GM and states that said criteria is uniform. Sinegal Depo., Docket No. 665, Ex. 8, at 29-30, 42-43 (describing criteria for both AGM and GM positions as mandating candidates with " people skills," " merchandising skills," and " the ability to be adroit with the numbers" ); id. at 45 (criteria for AGMs " are the same throughout Costco in the United States" ); id. at 30 (standards for GM promotion are " generally the same" throughout the company, and " we have a pretty clear understanding of what is required to be a" GM); see also Matthews Depo., Docket No. 665, Ex. 5, at 150 (stating that management rejected a recommendation to establish clearer guidelines of job criteria for warehouse positions because the positions were already well-defined throughout the company). Indeed, Defendant's senior executives dispute any assertion that the promotion process is undisciplined, haphazard, or varied based on the whims of local managers. See, e.g., Schutt Decl., Docket No. 623, ¶ ¶ 7-11 (describing orderly process and stating in paragraph 11, " Plaintiffs falsely suggest that promotions to GM and AGM are decided without any consistent standards at all.... [A]s I testified, Costco's standards are straightforward, readily evaluated, and well understood throughout the Company." ); Matthews Decl., Docket No. 600, ¶ 3 (disputing idea that there is no guidance on promotions and averring that there are " various forms of guidance and instruction Costco's Human Resources Department provides to its employees regarding promotions ..., including the Costco Employee Agreement and the code of ethics therein, the Rothman Workplan, and employee newsletters, which describe the company culture, and career paths and opportunities" ); Zook Decl., Docket No. 653, ¶ 19 (describing uniform promotion criteria).

Senior management's involvement in AGM and GM promotions makes sense given the importance of these positions and their placement in the upper echelons of Costco. Far from mere mid-level supervisors, AGMs and GMs oversee warehouses with an average of over 200 employees and over $120 million of merchandise sales per year. See Mulligan Decl., Docket No. 647, ¶ 24; Nelson Decl., Docket No. 609 (describing the characteristics of warehouses and importance of warehouse management). CEO Sinegal confirms that he " consider[s] the jobs of assistant manager and manager to be top management jobs" because of the responsibility involved. Sinegal Depo., Docket No. 544, Ex. V, at 124; see also id. at 42-43 (describing " how important we consider growing warehouse managers" in the context of a discussion about AGM promotions, and noting that " [w]e want the same strengths that we would find in a warehouse manager" in an AGM); Schutt Depo., Docket No. 544, Ex. U, at 75 (" [I]t's a big job being a warehouse manager for Costco, so it's a very well thought out, fully evaluated decision, and it's important enough to require the ultimate approval of the [CEO] of the company." ); Vachris Decl., Docket No. 627, 7 (" The AGM positions also involve a tremendous amount of responsibility, as these people run the warehouse when the GM is not there." ). Thus, the record reflects that the level of scrutiny by top executives is commensurate with the importance Costco attaches to these positions.

3. Costco's Own Evidence

Indeed, Costco's own anecdotal evidence and admissions further evidence the existence of common practices that apply to the class as a whole. For example, Costco's internal diversity studies, focusing on barriers to women's advancement, treat its promotion process (and the gender disparities therein) as a common, companywide problem with a companywide solution. In 2001, Costco convened a series of focus groups under the BOLD ⁹ Initiative to examine the company's recruitment and promotion process. See, e.g., Matthews Depo. at 111-13, 129, 145; Larkin Decl., Ex. 18 (BOLD Initiative memo to Steering Committee regarding recruitment and promotion recommendations). ¹⁰ The Recruitment and Promotion Project Team found that while the current employee base overall reflected the U.S. population by race and gender, " serious breakdowns occur at the management level." Larkin Decl., Docket No. 665, Ex. 18, at 1. In response to a series of team recommendations for companywide policies and initiatives, Costco implemented some proposals and declined to implement others. See, e.g., Matthews Depo. at 145-46. Importantly, for example, Costco rejected the team's recommendation to implement job posting " for all positions regardless of level." Matthews Depo. at 150. Instead, the company limited posting to positions below the AGM level. Id. at 151. The company's actions through the BOLD Initiative and subsequent efforts demonstrate that " Costco's senior management had knowledge of the gender disparity in the AGM and GM promotion processes and treated this disparity as a company-wide issue." Ellis I, 240 F.R.D. at 640. Indeed, Costco's own briefing touts its companywide efforts as effective in increasing diversity through various companywide policies. See Def's Mot. to Eliminate Class Claims, Docket No. 543, at 8. Costco thus implicitly concedes that its promotion practices are subject to companywide control and adjustment depending upon senior management's goals and instructions.¹¹

9 BOLD denotes Business Opportunities Leadership Diversity. See Matthews Depo., Docket No. 665, Ex. 5, at 111-13.

10 Notes from some of those focus groups support Plaintiffs' allegations of bias as a barrier to advancement. See, e.g., Larkin Decl., Docket No. 665, Ex. 12.

11 The company's decisions that came out of the BOLD process were memorialized in the Rothman Workplan. Matthews Depo. at 113. However, by design, the bulk of Costco's implemented changes did not directly affect the AGM/GM promotion system. Ellis II, 657 F.3d at 979 n. 5 (" [T]he Rothman Workplan only changed promotion practices as to the four Senior Staff jobs, but not as to AGM or GM positions." ); Docket No. 548-49 (Rothman Workplan in Costco Today magazine). Thus, as a result of a conscious decision, Costco's practices have remained largely consistent with respect to AGM and GM promotion practices during the measured time period for which the parties have data. Given Costco's decision to standardize lower-level promotion practices, it seems implausible to suggest that Costco exercises no centralized control over such important, high-level positions as AGM and GM.

As for other anecdotal evidence of commonality, Mr. Kadue's Declaration, Docket No. 544, summarizes Defendant's proffered declarations from putative class members and identifies several common characteristics of Costco's promotion process that, Defendant asserts, hold true across regions, including the fact that: (1) Merchandising positions are critical to advancement, id. ¶ ¶ 3, 4; (2) GMs make recommendations as to promotions, but those decisions are ultimately made or reviewed by higher executives at the regional and VP level, id. ¶ 5 (stating that local managers recommend promotion candidates, but that regional executives actually make the decisions); (3) Employees sometimes forego promotional opportunities for personal reasons, id. ¶ 6; (4) Promotion criteria are clear to Costco employees, id. ¶ 7 (describing criteria such as schedule flexibility, warehouse experience, hard work, and people management); and (5) Costco fosters the promotion of women across all regions, id. ¶ 8. The Court has already addressed issues 1, 2, and 4 above, and concluded that they support commonality. In addition, factor 5 (to the extent it is not based solely on withdrawn evidence),¹² while introduced to negate a finding of discrimination, actually supports commonality as it indicates that Costco acts with respect to the class as a whole.

12 The Court notes it is questionable whether this characteristic should be included here, as the bulk of declarations to which Costco points in support of this claim would appear to be covered by the parties' previous stipulation to withdraw portions of the class member declarations Costco obtained. See Docket Nos., 484, 486. Nonetheless, to the extent any of the language to which Costco points in support of this claim have not been stricken by the parties' stipulation, such a claim remains a common, classwide claim that does not undermine commonality. Rather, it is simply a classwide denial of the merits of Plaintiffs' claims that Costco does not foster the promotion of women and indeed, hinders their advancement.

As for factor 3, Costco argues that its anecdotal evidence of women " defer[ring] advancement for family reasons ... refutes ... any defensible notion of commonality." This argument is unavailing. While Costco presents personal/family obligations and their potential hindrance on career advancement as an issue unique to isolated, individual class members (and therefore incapable of classwide resolution), Costco's own proffered evidence suggests the issue of family obligations and its effect on career advancement is universal, far from being an isolated concern. See Kadue Decl., ¶ 6 (summarizing dozens of declarants' testimony on the subject). Costco's internal documents substantiate this. See, e.g., Larkin Decl., Ex. 14 (Diversity in the Workplace presentation identifying " Managing Work & Family" as a barrier to achieving merchandising experience that " affects both genders" ).¹³ Indeed, testimony from Costco executives and employees indicates that not only is such a concern not unique to a mere isolated subset of women, but it is common to people in general, regardless of gender. ¹⁴ In any event, to the extent Defendant argues that women are, on average, less interested in the positions that lead to promotion, less interested in promotion in general, or less able to perform any of the duties necessary for promotion based on personal circumstances, such arguments are premised on general assertions about the Class as a whole and how it differs, on average, from men employed at Costco.¹⁵ Indeed, Plaintiffs argue that in advancing these justifications, Costco's rationale itself evidences sexist stereotypes and assumptions about women's personal obligations relative to their professional ambition. See, e.g., Plfs' Cross-Mot., Docket No. 664, at 8 (arguing that the view that " the shortfall of women in higher level warehouse management positions is not attributable to any failing of Costco's decision-making process but instead results from women's preference for jobs that allow them to accommodate their families" reflects " stereotyped perceptions about the roles for women and men" ) (citing Sinegal Depo., Docket No. 665, Ex. 8, at 141, 146-47 (describing women as " hav[ing] a tendency to be the caretakers and have the responsibility for the children and for the family" and as tending to fill administrative positions at Costco " since the beginning of time" )). Regardless of the merits of this argument, the generality of this assertion in the record demonstrates that any defensive rationale based on family or personal circumstances is a common issue best resolved through classwide treatment, rather than one that undermines commonality.¹⁶

13 As discussed below in the context of the parties' expert evidence, some of the parties' experts— including Defense expert Dr. Saad— offer common, classwide theories purporting to explain the gender disparity in promotions as caused in whole or in part by such family obligations. These again support, rather than undermine, commonality.

14 Omoss Decl., Docket No. 611 (" I testified at length that the challenges facing employees, including their personal family situations, are not gender-specific." ) (Regional VP); Vachris Decl., Docket No. 627, ¶ 9 (" His commitment to his family or his current living situation is greater than his commitment to his career advancement." ) (emphasis added); Johnson Decl., Docket No. 588, ¶ 34 (describing male GM's personal hardship as being similar to hers); Arredondo Decl., Docket No. 553, ¶ 14 (she and husband both face work-family conflict and " stagger their fulltime hours to accommodate childcare needs" ); Auerbach Decl., Docket No. 555, ¶ 13 (husband stayed home with kids); Cline Decl., Docket No. 563, ¶ 3 (she and husband took opposite work shifts to share childcare responsibilities); Evans Decl., Docket No. 568, ¶ 26 (both parents work fulltime, use family help for childcare); Foster Decl., Docket No. 569, ¶ 13 (coordinated work hours with husband); Gonzales Decl., Docket No. 576, ¶ ¶ 3, 5, 7 (described give-and-take between husband's career priorities and hers, and that husband eventually stayed home); Manooa Decl., Docket No. 597, ¶ ¶ 7-9 (both husband and wife received promotions and transfers to new location); Moore Decl., Docket No. 607, ¶ 17 (" Although I have never deferred or declined a position because of my family responsibilities, I've factored it in. I've never had to say no, though, because my husband (who is also in Costco Management) and I have always worked together to figure out a solution to any issues which arise." ); Vlady Decl., Docket No. 629, ¶ ¶ 10, 14-15 (described husband's accommodation of her schedule); Borgner Decl., Docket No. 636, ¶ 20 (neither she nor husband, also a Costco employee, wants to relocate); Bejarano Decl., Docket No. 640, ¶ 12 (she and husband, who also works at Costco, work opposite shifts); Churillo Decl., Docket No. 642, ¶ 35 (husband also works for Costco, describes challenge of coordinating their shifts); Cruz Decl., Docket No. 643, ¶ 29 (describes working with husband to coordinate responsibilities); Poser Decl., Docket No. 646, ¶ 21 (" [M]y husband worked for (and still works for) Costco, and the company has allowed us to work opposite schedules so that we could raise our family." ); id. ¶ 29 (husband worked evenings while she worked MM shifts).

15 Indeed, Defendant offers supporting expert testimony and statistical analysis on this subject, including job posting and application data, discussed below in the context of the Court's evaluation of the parties' expert reports.

16 The Court notes that even post- Dukes, class members need not be identical with no distinctions in their personal employment history in order to present common issues affecting the class as a whole. Were that the standard, Rule 23 would cease to exist altogether. See Stinson v. City of New York, 282 F.R.D. 360, 369-70 (S.D.N.Y.2012), reconsideration denied, 10 CIV. 4228 RWS, 2012 WL 2952840 (S.D.N.Y. July 19, 2012) (" A court may find a common issue of law even though there exists some factual variation among class members' specific grievances." ) (quoting Dupler v. Costco Wholesale Corp., 249 F.R.D. 29, 37 (E.D.N.Y.2008)). Far from defeating commonality, Defendant's proffered class member declarations affirm and reinforce Plaintiffs' argument that common issues of fact and law exist in this case, the answers to which will drive the litigation forward toward resolution.

Furthermore, Costco's employee declarations actually support Plaintiffs' claim that employment and promotion practices are uniform throughout the company. Many of these declarations describe several transfers to various stores in each of the eight regions throughout the country, yet they do not describe significant differences between stores and the way they operate. See, e.g., Ellis Decl., Docket No. 654, ¶ 25 (" During my career at Costco, I worked in five buildings in two states (Michigan and Colorado). From my experience, Costco policies were the same in all five warehouses." ); Barnwell Decl., Docket No. 674, ¶ 21 (six locations in three states); Laureano Decl., Docket No. 592, ¶ 5 (worked in three of eight regions). Accordingly, the Court concludes that each of the issues described in the Kadue Declaration and accompanying employee declarations identify common, classwide issues relevant to Plaintiffs' claims of discrimination.¹⁷

17 Costco's own briefing further highlights this fact, as it summarizes the common practices it employs in the promotion process with no attempt to argue regional or local variation. See, e.g., Def's Mot. to Eliminate Class Claims, Docket No. 543, at 7 (" Costco has written lists to identify managers promotable to GM and AGM (as well as other positions). The managers evaluating GM and AGM candidates do so on the basis of regular interactions, spanning several years. Costco exemplifies ‘ management by walking around’ : GMs and AGMs walk their warehouses daily, and ROMs [Regional Operations Managers] visit warehouses frequently. They acquire evaluative data on promotion candidates, formally memorialized at least annually in written performance reviews." ) (emphasis in original) (internal citations omitted).

4. Discretion

Defendant further argues that its challenged practices allow for a certain degree of discretion among managers in making promotion decisions, and that this negates commonality. This argument is not persuasive. Even under Dukes, the fact that some degree of discretion is exercised among managers does not in and of itself preclude class certification. In fact, under the Teamsters pattern or practice method of proof, a prima facie case of a pattern or practice of discrimination is established by evidence that " racial discrimination was the company's standard operating procedure— the regular rather than the unusual practice." Cooper, 467 U.S. at 876, 104 S.Ct. 2794 (quoting Teamsters, 431 U.S. at 336, 97 S.Ct. 1843). Thus, Plaintiffs need not prove absolute uniformity, but only a " regular" practice. Accordingly, " at 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." Id. (citations omitted) (emphasis added); see also Karp v. CIGNA Healthcare, Inc., 882 F.Supp.2d 199, ----, CIV.A. 11-10361-FDS, 2012 WL 1358652, at *6 (D.Mass. Apr. 18, 2012) (" A plaintiff can establish a pattern or practice of discrimination by relying on statistical evidence and other similarly situated employees' accounts of discrimination." ) (citing, e.g., Dukes, 131 S.Ct. at 2555-56). As demonstrated above, the focus of Plaintiffs' claims— and Defendant's refutation of those claims— is on Defendant's companywide policies and practices. Even though there is, of course, some degree of residual discretion when individual hiring decisions are made, the common practices and policies challenged herein affect and guide that discretion. There is, as discussed herein, persuasive evidence of a " common mode of exercising discretion that pervades the entire company." Dukes, 131 S.Ct. at 2554-55. In addition, there are specific employment practices affecting the recruitment as well as selection processes that tie the claims of class members together. Id. at 2555.

Given the extensive involvement by senior management in the promotion process for these positions described above, Plaintiffs have produced significant and persuasive proof that the discretion exercised within that process occurs " under the rubric of a company-wide employment practice." Chen-Oster v. Goldman, Sachs & Co., 877 F.Supp.2d 113, 118, 10 CIV. 6950 LBS JCF, 2012 WL 2912741, at *3 (S.D.N.Y. July 17, 2012) (citing, e.g., Dukes, 131 S.Ct. at 2554). Indeed, as described above, CEO Sinegal and other senior executives view Costco's promotion practices as a cohesive whole, and opine that criteria for candidates are clear and readily applied by local management with the input and oversight of the senior team. See Floyd v. City of New York, 283 F.R.D. 153, 162-63 (S.D.N.Y.2012) (fact that policy and stop-and-frisk program was discussed regularly at top-level management meetings and Chief of Police's office discussed the program with lower-level staff supported commonality). Unlike in Dukes, in which " Wal-Mart permits store managers to apply their own subjective criteria when selecting candidates," Dukes, 131 S.Ct. at 2547, here the criteria Plaintiffs allege to be subjective and unvalidated derive from top management's own instructions, and senior management oversees the exercise of discretion in implementing said criteria. Costco's promotion system guides discretion, and thus differs from the " fragmented discretion untethered to any companywide policy and procedure" the Supreme Court found problematic in Dukes . Chen-Oster, 877 F.Supp.2d at 119, 2012 WL 2912741 at *3; ¹⁸ see Floyd, 283 F.R.D. at 174 (" [D]efendants confuse the exercise of judgment in implementing a centralized policy [— here, a police department stop and frisk policy— ] with the exercise of discretion in formulating a local store policy or practice." ) (emphasis in original); Stinson v. City of New York, 282 F.R.D. 360, 370 (S.D.N.Y.2012), reconsideration denied, 10 CIV. 4228 RWS, 2012 WL 2952840 (S.D.N.Y. July 19, 2012) (" Unlike in Dukes where the plaintiffs alleged a corporate policy of discretion to local managers and a corporate culture hostile to the advancement of women, Plaintiffs here have alleged a specific policy promulgated by Defendants, namely, that Defendants have established a practice by which NYPD officers issue summonses without probable cause in order to meet a summons quota." ); Morrow v. Washington, 277 F.R.D. 172, 192 (E.D.Tex.2011) (finding commonality based in part on the fact that the challenged practice " was conceived and implemented by a small number of Tenaha police officers and city officials working in concert during a specified time period" ); Cf. Bolden v. Walsh, 688 F.3d 893, 896 (7th Cir.2012) (finding no commonality between workers at different worksites where supervisors and policies differed and " [d]ifferent sites had materially different working conditions, as most plaintiffs conceded in their depositions" ).

18 In Chen-Oster, the court found that a companywide policy or policies incorporating discretion may nonetheless satisfy Dukes in a pattern-or-practice case. In considering a motion to strike class allegations in a pattern-or-practice and disparate impact case, the court found that

5. Summary

Independently viewing the record as a whole, the Court rejects Defendant's argument " that promotion decisions to AGM and GM vary by region and that promotion decisions to AGM are made at the store level." Ellis I, 240 F.R.D. at 639. " Rather, the evidence before the court indicates that officers at the regional and corporate levels are involved in promotion decisions." Id. ¹⁹ Furthermore, the fact that the scope of this suit is confined to the selection of GMs and AGMs by a much smaller group of decisionmakers than in Dukes — decisionmakers who form a relatively small coherent group— makes it far less " unbelievable" that these managers would exercise their discretion in a common way. Dukes, 131 S.Ct. at 2555. To the contrary, Plaintiffs have offered persuasive evidence of a common direction emanating from Costco's upper management. B. Costco's Culture

19 Although the Court identifies specific components of this process, Plaintiffs need not separate each component part from the overall promotion process on the merits if they " can demonstrate to the court that the elements of a respondent's decisionmaking process are not capable of separation for analysis," in which case " the decisionmaking process may be analyzed as one employment practice" for purposes of causation. See 42 U.S.C.A. § 2000e-2(k)(1)(B)(i).

In addition to the above-identified policies and practices, Plaintiffs contend through their sociological expert, Dr. Reskin, that Costco's culture fosters and reinforces stereotyped thinking, which allows gender bias to infuse the promotion process from the top down. See Reskin Decl., Docket No. 670, ¶ 9 (summarizing findings). Dr. Reskin conducted a social framework analysis, examining Costco's personnel and promotion policies and practices in the context of social science literature and her expertise in workplace discrimination and " organizational policies and practices that can mitigate conscious and unconscious stereotyping, automatic and conscious ingroup favoritism, and sex bias." Reskin Decl., Docket No. 670, ¶ 5. Such bias includes, as noted above, Plaintiffs' and Dr. Reskin's charge that the CEO and other top executives employ stereotyped thinking regarding women's roles in society. See Reskin Decl., Docket No. 670, ¶ ¶ 53-60 (citing testimony from CEO Sinegal and others that women's caretaking role causes them to be less interested in promotion opportunities, as well as other stereotyped assumptions about, e.g., women's ability to be forklift drivers). She concluded that " [c]entralized control, reinforced by a strong organizational culture, creates and sustains uniformity in the personnel policies and practices throughout Costco's operational units. This common culture is characterized by unwritten rules and informal, undocumented personnel practices featuring discretion by decision makers." Reskin Decl., Docket No. 670, ¶ 9. Dr. Reskin opined that these informal, yet cohesive, practices are " likely to be tarnished by biases that operate against women." Id. Dr. Reskin contrasted Costco's practices with the more formal practices that, social science research indicates, " sustain or reduce barriers to women's career success." Id. ¶ 10. Costco has rejected such policies. Id.

The Court finds Dr. Reskin's testimony persuasive. It is consistent with and provides further support for Plaintiffs' claim that Costco operates under a common, companywide promotion system. See Price Waterhouse v. Hopkins, 490 U.S. 228, 235-36, 109 S.Ct. 1775, 104 L.Ed.2d 268 (1989) (accepting expert testimony on gender stereotyping and its " likely influence[ ]" on partnership selection process).

This evidence of corporate culture also underscores the fact that adjudication of Plaintiffs' claims will require answering common questions with respect to the class as a whole. In this regard, Defendant does not disagree— or at least, presents no evidence— that Costco has no strong corporate culture that guides managers' promotion decisions companywide. See Landy Decl., Docket No. 655, at 51 (" It is neither remarkable, nor the subject of expert opinion that Costco has a culture. Fact witnesses for Costco have repeatedly acknowledged and described the Costco culture." ). To the contrary, CEO Sinegal and other Costco employees and documents place strong weight on Costco's culture as an important influencer for its promotion practices. See, e.g., Matthews Decl., Docket No. 600, ¶ 3 (disputing the notion that promotion requirements have been passed down only through oral culture because such an argument " falsely impl[ies] that Costco has no written documentation regarding promotions" ; contending instead that numerous written documents provide guidance by " describ[ing] the company culture, and career paths and opportunities" ); Sinegal Depo., Docket No. 665, Ex. 8, at 30 (describing the Costco culture as important for potential GMs to understand, including the open door policy and merchandising concepts); id. at 123-24 (rejecting job posting for AGM and GM because Costco has a strong " system in our company, we had a culture built into our company that was recognizing on a regional basis who was qualified" for those jobs); see also, e.g., April 2005 Costco diversity report, Docket No. 665-13 (Costco's internal report on diversity, which leads with a description of its culture— including its open door policy and commitment to, inter alia, trust, personal initiative, and mutual respect— and is infused with references to the Costco way of doing business). Defendant and its experts merely disagree with Dr. Reskin as to the characteristics of that culture. See Sinegal Depo., Docket No. 544, Ex. V, at 155 (disputing the idea that managers might be discriminating in certain employment decisions because " [i]t's so clearly ingrained in the culture of our company, I can't imagine how they would miss that message" ). In other words, Defendant disagrees as to " whether Costco does in fact have a culture of gender stereotyping and paternalism," an issue the Ninth Circuit explicitly held the Court need not resolve at the class certification stage. Ellis II, 657 F.3d at 983. While the Court finds Dr. Reskin's testimony persuasive for purposes of Rule 23(a) analysis, regardless of which party is correct on the merits, the content of the companywide culture is a common question amenable to resolution on a classwide basis.

Similarly, Defense psychological expert Dr. Landy's attack on Dr. Reskin's testimony is unpersuasive. While Dr. Landy purports to highlight regional differences in HR and promotion practices, the variations he highlights are largely unconnected to the AGM and GM promotion process. See Landy Decl., Docket No. 655, at 44-47. Nor does Dr. Landy's testimony negate commonality. To the extent his collected quotes reveal certain collateral differences between regions ( e.g., whether a region implements a Manager-in-Training program), such differences do not negate the overarching uniformity of the process, as well as senior management's involvement in that process, as discussed at length above. In addition, while Dr. Landy criticizes Dr. Reskin's description of Costco's culture, his argument challenges her conclusions on the merits about the characteristics and effects of that culture. See Landy Decl., Docket No. 655, at 53 (" What is missing from Dr. Reskin's observations about the Costco culture is any evidence that the culture is somehow pathological and devoted to discriminating against women. On the contrary, various depositions of Costco representatives and the exhibits to those depositions confirm an organization determined to directly address any possible barriers to the advancement of women and minorities." ); see also id. at 53-58 (describing further the influential aspects of Costco's culture and concluding that their effects are positive, not negative). Again, as to commonality, both sides do not dispute that there are policies and practices common to the class as a whole. See Landy Decl., Docket No. 655, at 69-89. Thus, the experts' dispute as to the merits and impact of Defendant's promotion policies merely confirm that such a dispute is one that has an impact on the class as a whole, and its resolution on a classwide basis is apt to drive the litigation.

The Court finds Plaintiffs' evidence of culture is persuasive in light of the CEO's admitted direct involvement in company promotion practices, the relatively high level of seniority for the positions at issue in this case, and the small number of top executives involved in GM and AGM recruitment and selection. However, the Court does not place undue or dispositive weight on this factor. Unlike in Dukes, the evidence of Costco's culture is just one component among many pieces of persuasive evidence of companywide practices and policies that support a finding of commonality. Cf. Dukes, 131 S.Ct. at 2553 (rejecting social framework analysis where it was " [t]he only evidence of a general policy of discrimination" ) (quotation marks omitted).

C. Classwide Effects— Statistical Evidence

1. National vs. Regional

Finally, Plaintiffs provide persuasive statistical evidence of gender disparities throughout Costco sufficient to refute the notion that " [d]issimilarities within the proposed class ... [might] impede the generation of common answers." Dukes, 131 S.Ct. at 2551 (quotation omitted). As noted by both Judge Patel and the Ninth Circuit, a key dispute between the parties is whether there is a nationwide gender disparity in promotions to AGM and GM positions within Costco, or whether any purported disparity is merely confined to certain regions and therefore not common to the class as a whole. Plaintiffs' expert, Dr. Drogin, " concluded that there were statistically significant internal gender-based disparities in both the AGM and GM positions based on his analysis of aggregate, nationwide data." Ellis I, 240 F.R.D. at 639 (citing Drogin Decl. ¶ 25). In contrast, Defendant's expert, Dr. Saad, found no gender disparity in AGM and GM promotions when he conducted a study by region, rather than in the aggregate, and controlled for certain variables for which Dr. Drogin did not control.

On balance, the Court finds that Plaintiffs' expert testimony is more persuasive at least insofar as it pertains to commonality. With respect to AGM promotions, Dr. Drogin demonstrates that, although women form nearly 27.7% of the available pool of SSMs from which AGMs are drawn, they received only 18.4% of AGM promotions during the pre-filing period. Drogin Decl., Docket ¶ 19, Table 7a. The gender disparity in promotions was statistically significant for each year from 1999 to 2004 (before the lawsuit was filed), and highly significant for the period as a whole. Id. Indeed, in raw numbers, the comparison is stark: women received only 103 of the 561 promotions to AGM between 1999 and August 2004. Id. Dr. Drogin also concluded that the results were statistically significant even when he controlled for the employee's number of years experience at the SSM level, as well as the region in which the promoted person worked. Id. ¶ 20 & Table 7b, ¶ 21 & Table 7c.

Dr. Saad, Defendant's expert, challenges Dr. Drogin's disparity findings, arguing that one must separate out the promotion numbers by region and evaluate each region separately. Using this formula, Dr. Saad concludes that there is no statistically significant disparity within most of the regions. See Saad Supp. Decl., Docket No. 620, Exs. R2-R8 (showing statistically significant disparity in only Los Angeles and the Northeast).²⁰ However, looking at the tables Dr. Saad provides for each region, the number of promotions per year is so small per region that it would be difficult to attach much importance to the lack of statistical significance. For example, for most regions in most years, the total number of promotions to AGM is under 20, and frequently under 10. Compounding the problem, women are already underrepresented in the pool of AGM candidates, averaging under one-third of the employees in the SSM pool from which AGMs are promoted. Thus, to demonstrate statistical significance as to promotions from SSM to AGM within such a small pool for each region would require a severe disparity in raw numbers.²¹

20 Dr. Drogin contends there is a statistically significant disparity in the Southeast as well, but Dr. Saad's table indicates that region is just shy of statistical significance. See Saad Supp. Decl., Docket No. 620, Ex. R8.

21 For example, the LA region yields a statistically significant gender disparity because women received only 7 promotions to AGM out of 53 between 1999 and July 31, 2004. See Saad Supp. Decl., Docket No. 620, Ex. R3.

Therefore, Dr. Drogin persuasively explains that given the small number of total promotions to AGM companywide, only 659 in total and 561 before the lawsuit was filed,²² separating the data by region " would tend to obscure the overall pattern of promotion rates." Drogin Decl., Docket No. 641, at 14 n. 21; see also Drogin Supp. Decl., Docket No. 667, ¶ 16. ²³ Dr. Reskin similarly opines that the lack of apparent statistically significant disparity in certain regions can be the result of a small number of cases, which increase the chance that random error has an effect on the data. Reskin Depo., Docket No. 680, Ex. 1, at 213. Courts have recognized that aggregate data can be more probative in these circumstances. See Paige v. California, 291 F.3d 1141, 1148 (9th Cir.2002) (aggregation " is particularly appropriate where small sample size may distort the statistical analysis and may render any findings not statistically probative." ); Compare, e.g., Coates v. Johnson & Johnson, 756 F.2d 524, 541-42 (7th Cir.1985) (explaining that pooling data is sometimes warranted where small sample sizes can tend to obscure disparities, citing case in which pooling was appropriate given sample sizes from 2 to 39 per category, and describing sample sizes of 490, 509, and 815 as sufficiently large for analysis), with Saad Supp. Decl., Docket No. 620, Exs. R2-R8 (listing regional data in which total promotions from 1999-2004 range from 8 to 150).²⁴ Defendant offers no persuasive response as to the issue of sample size.²⁵ Thus, the lack of statistically significant disparities at the regional level says little about the classwide nature of Defendant's practices given the small sample sizes.

22 Dr. Saad's data indicate only 553 promotions during the pre-lawsuit period from 1999 to July 31, 2004, which does not change the above point. See Saad Supp. Decl., Docket No. 620, Ex. 5.

23 Contrary to Defendant's suggestion, Dr. Drogin did not " admit" in deposition that the regional sample sizes were large enough for meaningful analysis that would be equally probative to the aggregate data. Instead, he merely discussed the fact that three regions yielded statistically significant gender disparities based on his calculations, while the remaining four yielded raw disparities that did not reach statistical significance. See Drogin Depo., Docket No. 544-1, Ex. B, at 27-30 (stating that regional samples are " not a very large number," and stating that certain regions produce statistically significant results while others do not). This is consistent with his overall conclusion, which the Court finds persuasive, that the small numbers per region make it difficult to draw meaningful conclusions from that data as compared to the nationwide pool. As discussed infra, Dr. Saad similarly acknowledges this difficulty with small sample sizes in a different context. Dr. Drogin also bases his nationwide statistics on the companywide locus of decisionmaking, which the Court addresses below.

24 While neither Coates nor this Court purports to specify a floor under which sample sizes are per se too small (or a ceiling over which samples are per se large enough), the Court lists the sample sizes described in Coates as an example of what courts have considered sufficient.

25 Indeed, Defendant appears to largely offer a tautology— that disaggregated data is more probative simply because it yields results favorable to Costco. The Court is not persuaded. Similarly, the mere fact that certain regions yielded statistically significant results does not demonstrate that the sample size for each region is sufficient to meaningfully compare them and draw conclusions as to their purported differences. Indeed, elsewhere in his report, Dr. Saad acknowledged that small sample sizes might call into question the meaning of certain disparities, even when they technically yield statistically significant results. See Saad Decl., Docket No. 619, ¶ 72 & n. 12, Ex. 25 (listing applications to Meat Manager position as yielding a statistically significant gender disparity, but stating that " given the relatively small sample of total applicants ..., the statistics may not be reliable indicators of differential job preferences or economic constraints" ); see also id. ¶ 76 & n. 13, Ex. 28 (same).

The Court finds there is good reason to rely on nationwide statistics. Not only do the larger aggregate numbers allow for a robust analysis and yield more reliable and more meaningful statistical results, Costco's own promotion practices support a nationwide statistical analysis. As discussed above, CEO Sinegal avers that promotion policies and practices are uniform across the company. Costco executives are informed of and involved in the recruitment and selection process from the top levels of the company. Thus, unlike in Dukes, here Plaintiffs' statistical analysis conforms to the level of decision for the challenged practices, including the adoption of the many companywide policies described above. Cf. Dukes, 131 S.Ct. at 2555. In addition, the fact that candidates are often promoted from across regions lends further support to treating the gender disparity as a national, rather than regional, issue, and adjusting the level of focus for statistical analysis accordingly. See, e.g., Mulligan Decl., Docket No. 647, ¶ 32 (asked to assume for defense expert report that willingness to relocate is a factor in promotions); Kadue Decl., ¶ 7 (describing willingness to move as a factor in promotions based on class member declarations); Vachris Decl., Docket No. 627, ¶ ¶ 6, 9.

Even if the data were examined strictly on a region by region basis, the fact that all seven non-Texas regions show a raw gender disparity in promotions is telling. Drogin Depo., Docket No. 544, Ex. B, at 27; Saad Supp. Decl., Docket No. 620, Exs. R2-R8. Examining the data by region, ²⁶ the data supports Plaintiffs' contention that gender disparities extend across all regions, and the absence of statistical significance within each individual region is of limited value for the reasons discussed above. See Watson, 487 U.S. at 997 n. 3, 108 S.Ct. 2777 (" We have emphasized the useful role that statistical methods can have in Title VII cases, but we have not suggested that any particular number of ‘ standard deviations' can determine whether a plaintiff has made out a prima facie case in the complex area of employment discrimination.... Instead, courts appear generally to have judged the ‘ significance’ or ‘ substantiality’ of numerical disparities on a case-by-case basis." ); Chin v. Port Auth. of New York & New Jersey, 685 F.3d 135, 153 (2d Cir.2012) (upholding jury's reliance on non-statistically significant disparity based on small sample size and non-statistical evidence of discrimination and noting that " [c]ourts should take a case-by-case approach in judging the significance or substantiality of disparities, one that considers not only statistics but also all the surrounding facts and circumstances" ) (internal citations and quotation marks omitted).

26 Both experts agree Texas is too small in terms of promotions to produce reliable data.

Thus, in answer to the Ninth Circuit's directive to determine whether disparities exist across all regions such that Plaintiffs may be able to " show that ‘ discrimination manifested itself in ... promotion practices in the same general fashion’ ... throughout Costco," the Court finds that Plaintiffs have demonstrated such a companywide disparity for purposes of establishing commonality under Rule 23. Id. (quoting Dukes, 131 S.Ct. at 2553). See Ellis II, 657 F.3d at 983 (" If, as Plaintiffs allege, promotion decisions are based on the biased attitudes of the CEO and upper management, one would expect disparities in all, or at least most, regions." ).

In sum, the Court finds there is good and persuasive evidence of disparities in all regions and that the lack of a statistically significant disparity in each region does not disprove the companywide nature of the disparity, given the small sample sizes. Further, Dr. Drogin's decision to aggregate the data nationwide is reasonable because the larger sample size allows for more robust statistical analysis and matches the relevant level of decisionmaking within Costco. Thus, the Court finds there is persuasive evidence that gender disparities are present classwide, not just in isolated regions.

2. Competing Explanations for Statistical Disparity

Beyond the debate over regional differences, Defendant's general attacks on Plaintiffs' statistical evidence are largely unpersuasive at least insofar as they bear generally on commonality. The theme of Costco's arguments addressed in this section is that Costco offers competing control variables to explain the observed gender disparity in promotions, variables Costco contends refute any claim of discrimination. However, such classwide variables are precisely the sorts of arguments subject to classwide resolution. As discussed below, these disputes amount to competing classwide explanations for the observed raw disparity and therefore support, rather than undermine, a finding of commonality.

For example, for GM promotions, both experts agree there is no statistically significant disparity in the rate of promotion from AGM to GM. Dr. Drogin provides a persuasive explanation for why that is so. He explains that the pool of women is already artificially low due to the low rates of women promoted into the AGM position. Drogin Decl., Docket No. 641, ¶ 23. In other words, the representation of women as AGMs is itself a tainted variable, and because virtually all GMs come from the AGM pool, the lack of statistical significance does not detract from the overall gender disparity in both positions. See Morgan v. United Parcel Serv. of Am., Inc., 380 F.3d 459, 470 (8th Cir.2004) (" Regression analyses in discrimination cases attempt to control for the legitimate reasons for pay disparities through the use of explanatory variables. But, illegitimate reasons -reasons themselves representative of the unlawful discrimination at issue- should be excluded from the regression (or otherwise dealt with) to avoid underestimating the significance of a disparity. " ) (emphasis added). Defendant's response on this point is unpersuasive and irrelevant to commonality. It does not argue that there are any intra-class differences in the rate of AGM or GM promotions; rather, it makes a classwide argument that there is in fact no disparity in the promotion rate to GM. The Court need not resolve this factual question at the certification stage; rather, it presents a common question suitable for classwide resolution. See Ellis II, 657 F.3d at 983 n. 8 (court may not " determine whether class members could actually prevail on the merits of their claims" ); see also id. (rejecting Costco's argument that " [t]here is no commonality absent (a) statistical proof of under-promotion of women and (b) a plausible link between the practice and the impact," because it would essentially " turn class certification into a mini-trial" ) (emphasis omitted).

Similarly, Defendant offers competing explanations for the gender disparity in promotions to the AGM position from the Senior Staff (SSM) positions, and for the disparity in placement in merchandising positions within the Senior Staff level which feed the AGM promotion. Yet, each of Defendant's competing variables are classwide variables, subject to common resolution. For example, Dr. Saad advocates for, and implements in his report, a control variable for the four specific types of SSMs from which AGMs receive promotions, rather than merely using the SSM category as a whole as the pool from which AGMs are drawn. See, e.g., Saad Decl., Docket No. 619, ¶ ¶ 21, 31-50. Dr. Saad argues that given the importance of the Merchandise Manager (" MM" ) position among the four SSM positions, controlling for experience in this position yields more reliable results as to whether there is any discrimination in promotion. His data indicate that once he controls for MM experience, there is no remaining gender disparity in AGM promotions. See id. Ex. 8.

Like the AGM-to-GM promotion discussed above, however, Plaintiffs dispute this method as merely concealing the disparity through a tainted variable, as the rate of promotion into the MM position is much lower for women than any other SSM position. Plaintiffs also introduce qualitative evidence that women have been blocked from the MM position due to gender bias, to support their theory that MM experience is a tainted variable. See, e.g., Larkin Decl., Docket No. 665, Ex. 23; Ellis I, 240 F.R.D. at 648 (" Plaintiffs have presented substantial evidence, beyond mere allegations, that raises the inference that MM is tainted by Costco's preference for hiring men for that position." ). Moreover, Costco's senior management is aware of the gender disparity in merchandising experience and the fact that reliance on such experience prevents women from ascending the management ranks. See, e.g., Sinegal Depo., Docket No. 665, Ex. 8, at 146-47 (opining that women tend not to be in merchandising positions); Larkin Decl., Docket No. 665, Ex. 23 (2005 Rothman Workplan Session Comments describing stereotypes that women cannot do merchandising positions), at CRE 0142527, CRE 0142538; id., Ex. 21, at CRE 0142691 (2006 Diversity Meeting notes describing dearth of women rotating into merchandising positions). See generally Stender v. Lucky Stores, Inc., 803 F.Supp. 259, 333 (N.D.Cal.1992) (findings of fact and conclusions of law, finding " that the lack of statistically significant disparities in promotions of women to Third Person, Assistant Store Manager and Store Manager is caused by women being blocked from upper management positions at the lower rungs of the promotional ladder" ); Segar v. Smith, 738 F.2d 1249, 1283 (D.C.Cir.1984) (finding plaintiffs had presented sufficient evidence to demonstrate liability for discriminatory promotions where there was a statistically significant disparity in promotions from GS-11 to GS-12 and, although disparity for promotions above GS-12 was not statistically significant, there was other evidence to raise an inference of discrimination). In any event, this dispute is a quintessential example of a common question: whether MM experience is properly included as a control in a statistical analysis of gender disparities in promotions to AGM. Costco thus offers a generalized defense applicable to classwide allegations. This dispute does not require individualized treatment, but instead is subject to a classwide resolution.

As another example, Dr. Saad recommends additional changes to the data as a whole, such as removing employees who were on leaves of absence at the time of promotion. See Saad Decl., Docket No. 619, ¶ ¶ 41 n. 8, 43, 50, Ex. 9; Saad Supp. Decl., Docket No. 620, ¶ ¶ 40-43, Exs. 6, R1-9. Dr. Saad contends that doing so yields promotion disparities that are not statistically significant for the class as a whole in all years, although the results are still statistically significant for the measured time period as a whole. Saad Decl., Docket No. 619, Ex. 6. While the proper composition of the data pool is certain to be a contested topic on the merits, Dr. Saad's contention is not one that affects commonality as he makes no argument, nor does his data demonstrate, any divergent effects between regions. Instead, Dr. Saad's argument presents another common question capable of classwide analysis. See also Saad Decl., Docket No. 619, ¶ ¶ 78-85 (presenting statistical analysis of women's leaves of absence as compared to men's across the company).²⁷

27 The Court also notes that excluding any employee who had taken a leave of absence at any point during the year prior to the promotion year from the pool of eligible promotees, as Dr. Saad has done, would appear to unnecessarily remove people eligible for promotion during much, if not all, of the year the promotion actually occurred. See Saad Decl. ¶ ¶ 43, 50 (explaining that the leave of absence control removes people who took a leave of absence in the year prior to the promotion year). In addition, as Dr. Saad's own data suggests people who take leaves of absence do so for much less than a full year. Id. ¶ 30 (describing average duration of leave by female SSMs as 136.2 days). Thus, it is unclear why, for example, a woman who took a leave of absence from January-March of 2002 should be excluded from the pool of eligible candidates for a promotion occurring in September 2003. In any event, while the parties are free to debate the merits of this control variable on summary judgment or at trial, inclusion or exclusion of this control variable is a classwide issue subject to classwide resolution, as Defendant's own data demonstrate.

Defendant next argues Dr. Saad's study of job posting data across the company (for positions below the AGM level) demonstrates that women are less interested in the merchandising positions (such as MM) that make one promotable to AGM and GM. Costco Reply at 20-21; see Saad Decl., Docket No. 619, ¶ ¶ 63-77. Therefore, Defendant argues, inclusion of the MM control variable is warranted, and once included, the purported gender disparity in promotions to AGM disappears. However, the merits of Dr. Saad's job posting analysis and its implications for the persuasiveness of Plaintiffs' evidence of gender disparities is a merits question that does not defeat commonality. Again, Defendant has identified a quintessential common question of fact: Did Costco's practices cause the gender disparity in promotions, or did women's differential preferences as a group determine that disparity? Costco proposes a classwide, common method of proving its proposed answer to that question, as it claims job posting data reveals that " differential gender preferences affected the number of females interested in merchandising management jobs." Costco Reply at 21. In other words, it claims that women, on average, are less interested in the prerequisites for promotion, and if this fact were controlled for, there would be no gender disparity. This is precisely the kind of " common answer [ ] apt to drive the resolution of the litigation." Dukes, 131 S.Ct. at 2551 (internal quotation omitted) (emphasis in original).²⁸

28 Indeed, it is telling that the only cases Defendant cites in support of this argument, see Costco Reply at 20 n. 111, 112, concern the sufficiency of plaintiffs' substantive allegations and evidence at the motion to dismiss, summary judgment, or trial stage, not class certification.

Beyond the dispute between Drs. Drogin and Saad as to AGM and GM promotion data, the parties also dispute the merits of Dr. Bendick's benchmarking analysis on behalf of Plaintiffs. Dr. Bendick compared the representation of women in Costco management with women's representation in other purportedly comparable retail establishments. Bendick Decl., Docket No. 668; Bendick Supp. Decl., Docket No. 669. The utility of benchmarking is that, as Dr. Bendick explains, it can help to isolate an employer's unique policies and practices from general societal factors affecting women's representation in certain positions. Because such factors would be expected to affect all employees and employers not unique to Costco ( e.g., personal or family constraints, or differential interest in certain positions cause by factors outside Costco's control), benchmarking allows one to isolate internal from external effects. Dr. Bendick opines that using several benchmarks based on different assumptions about the relevant comparable labor pool, each benchmark yields a disparity between women's representation at Costco and their representation in other retail establishments. See generally Bendick Decl., Docket No. 669, ¶ ¶ 29-34.²⁹ Dr. Bendick concludes that Costco's personnel and promotion policies and practices are responsible for the disparity. Id. ¶ 35. The Court finds such benchmarking analysis persuasive insofar as it illuminates the common question of whether Costco's companywide policies are responsible for the gender disparity in promotions to AGM and GM.

29 Each individual benchmark is inherently imperfect, as no data pool provides a perfect comparison. This is the reason, as Dr. Bendick explains, for diversifying among several different benchmarks and synthesizing their results, so as to account for their relative strengths and weaknesses. Bendick Supp. Decl., Docket No. 669, ¶ ¶ 33-38.

Dr. Saad provided rebuttal testimony arguing, inter alia, that Dr. Bendick's analysis was flawed because it does not control for earnings, and does not use the appropriate comparison retailers. Saad Decl., Docket No. 619, ¶ ¶ 99-100; see also Saad Supp. Decl., Docket No. 620, ¶ ¶ 13-25. Dr. Bendick's comparison pool therefore, according to Dr. Saad, compares the relevant Costco positions to the wrong positions in the labor market. He also contends Dr. Bendick uses the wrong Census code to compare Costco SSMs, and the wrong subset of comparison retailers. Saad Supp. Decl., Docket No. 620, ¶ ¶ 26-32. Using Dr. Saad's choice of a comparison data pool, Dr. Saad generates benchmarks that are favorable to Costco and show that women's representation in the relevant positions at Costco is actually higher than in comparable retailers. Saad Decl., Docket No. 619, ¶ 105, Exs. 43-44.³⁰ Another defense expert, Dr. Mulligan, faults Dr. Bendick's benchmarking analysis because, she contends, it fails to properly account for the different characteristics of female workers that affect their labor supply (for example, their greater family responsibilities). See, e.g., Mulligan Decl., Docket No. 647, ¶ ¶ 16-21, 34-35, 45, 51-56, 61-62, 96, 103, 106.

30 Dr. Bendick, of course, disputes Dr. Saad's competing methodology.

The Court finds all experts' arguments illuminate a common question of fact relevant to Plaintiffs' claims of discrimination. Contrary to Defendant's suggestion, the Court's task at this juncture is not to determine which expert's benchmarks (Dr. Saad's or Dr. Bendick's) are correct on the merits. To do so would run afoul of the Ninth Circuit's clear instruction that the Court examine the merits only insofar as it is " necessary to determine whether there was a common pattern and practice that could affect the class as a whole. " Ellis II, 657 F.3d at 983 (emphasis in original). Instead, the issue is whether there are common questions and answers that will drive the litigation.³¹

31 In this instance, all experts' opinions regarding Dr. Bendick's benchmarking analysis support commonality. For example, Dr. Saad does not fault Dr. Bendick for failing to account for internal differences among Costco's regions or stores; nor does he opine that benchmark analysis is inappropriate on the whole because Costco's internal differences preclude any useful classwide comparison. Instead, Dr. Saad simply disagrees as to the proper classwide benchmark to use in the analysis. Nor does Dr. Mulligan contend that intra-class differences among women at Costco prevent an examination of the gender disparity in promotions; rather, Dr. Mulligan argues Dr. Bendick failed to attempt a " residual method" of analysis that would account for differences in women's labor supply that can affect their overall representation in the positions at issue. See Mulligan Decl., Docket No. 647, ¶ ¶ 96, 103, 106. Dr. Mulligan presents a competing theory to answer " [t]he relevant economic question for purposes of analyzing Plaintiffs' claims in this case" : " whether the observed statistical composition of Costco's managerial workforce results from employer discrimination against women." Id. ¶ 45. Dr. Mulligan's opinions are, according to her, derived from comparisons between women at Costco and women in the labor force in general, and are capable of classwide analysis. See, e.g., Mulligan Decl., Docket No. 647, ¶ 55 (defending her opinions based on broad national economic studies as to women's different characteristics because " it is reasonable to assume that Costco's workforce has family situations that are typical of those in America in general, and the economy-wide studies are applicable to Costco's labor force" ); id. ¶ 55 n. 18 (noting that she has seen no data suggesting that regional weighting of the economic statistics " would affect [her] opinion" ); id. ¶ ¶ 109-113 (describing characteristics unique to Costco that make it less likely discrimination is responsible for the gender disparity, without differentiating within Costco). Thus, all three experts' arguments on this point demonstrate that there are common questions and highlight " the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation." Dukes, 131 S.Ct. at 2551.

In short, Defendant offers numerous competing explanations for the observed gender disparity in promotions. None of these explanations undermine the companywide nature of the challenged policies and their disparate effects discussed above. Moreover, Defendant's remaining challenges present common proposed answers to Plaintiffs' common questions; these disputes are subject to generalized proof and " capable of classwide resolution." Dukes, 131 S.Ct. at 2545. These questions are therefore not " factual disputes necessary to determine whether there was a common pattern and practice that could affect the class as a whole." Ellis II, 657 F.3d at 983 (emphasis omitted). To the contrary, they provide further support for the Court's conclusion that Plaintiffs have demonstrated " the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation." Id. at 981 (quoting Dukes, 131 S.Ct. at 2551) (emphasis omitted).

3. Challenges to Methodology

The parties also dispute whether Plaintiffs should be permitted to use pre-statute of limitations data to demonstrate the disparity in promotions. Again, this is a dispute common to the class as a whole and unlike the statistical challenge based on purported regional differences, goes directly to the merits of Plaintiffs' classwide claims based on a defense that would apply to the class as a whole. Disputes over " whose statistical findings and observations are more credible ... [are] relevant only to the merits of plaintiffs' claim— whether plaintiffs actually suffered disparate treatment— and not to whether plaintiffs have asserted common questions of fact or law." Ellis II, 657 F.3d at 983 n. 8 (internal citations and quotation marks omitted). Indeed, nothing about Defendant's statute of limitations argument concerns " [d]issimilarities within the proposed class." Dukes, 131 S.Ct. at 2551. Instead, this argument concerns only the merits of Plaintiffs' claims and what evidence they may use to meet their burden of proof.

In any event, as Dr. Saad concedes, there is still a statistically significant disparity during the limitations period if one uses Dr. Drogin's data pool. See Saad Supp. Decl., Docket No. 620, ¶ 84, Exs. R38, R39. It is only once Dr. Saad implements other classwide manipulations to the data pool ( e.g., removing employees on leaves of absence, and other changes discussed above) that the results lose statistical significance. As the Court concluded above, these classwide disputes over how to manipulate the data pool to generate various classwide results present common questions subject to a common resolution, and the Court need not (and should not) resolve those disputes on the merits at this juncture.

Moreover, with respect to the merits of the disputed methodology, as Judge Patel previously found (and the Ninth Circuit did not challenge), there is " no legal bar to the use of pre-liability period evidence in this case, and ... Dr. Drogin's analysis [is] relevant to plaintiffs' claims." Ellis I, 240 F.R.D. at 647.³² As the Second Circuit has recently explained, pre-statute of limitations data is generally both relevant and frequently used in such matters as evidence of the alleged discriminatory practice, even when a plaintiff may not directly recover as to events occurring outside the applicable time period. See, e.g., Chin, 685 F.3d at 150 (" It is well established ... that so long as at least ‘ one alleged adverse employment action ... occurred within the applicable filing period[,] ... evidence of an earlier alleged [discriminatory] act may constitute relevant " background evidence in support of [that] timely claim." ’ " ) (quoting, e.g., Nat'l R.R. Passenger Corp. v. Morgan, 536 U.S. 101, 113, 122 S.Ct. 2061, 153 L.Ed.2d 106 (2002) (" Nor does the statute bar an employee from using the prior acts as background evidence in support of a timely claim." )); see Raad v. Fairbanks N. Star Borough Sch. Dist., 323 F.3d 1185, 1192 (9th Cir.2003), opinion amended on denial of reh'g, 00-35999, 2003 WL 21027351 (9th Cir. May 8, 2003) (" [W]hile these claims are not independently actionable, evidence about the District's refusal to hire Raad for a full-time teaching position in 1991 and 1992 is relevant and admissible insofar as it bears on her claim that she was discriminatorily refused a full-time position in August 1993." ); Lyons v. England, 307 F.3d 1092, 1108 (9th Cir.2002) (" [A]ppellants are permitted to offer evidence of the pre-limitations discriminatory detail assignment scheme in the prosecution of their timely claims." ). Although Defendant challenges Plaintiffs' use of pre-limitations data in combination with data during the limitations period, Lyons explicitly affirmed that " appropriate background evidence will be evidence, either direct or circumstantial, that, when combined with evidence of the employer's present conduct, give[s] rise to an inference of unlawful discrimination." Lyons, 307 F.3d at 1111 (internal citations and quotation marks omitted). " In particular, ... statistical studies may include data from outside the statute of limitations to prove timely discriminatory acts." Chin, 685 F.3d at 150 (citation omitted). The Court notes that such data is similarly permissible for a disparate impact claim. Paige v. California, 291 F.3d 1141, 1149 (9th Cir.2002) (" [I]t is appropriate to admit pre-liability data into evidence in a disparate impact case if promotional practices remain similar over a long period of time, as they have in this case." ) (citations omitted).

32 Indeed, more than simply " relevant," the Court finds his analysis persuasive for purposes of establishing commonality.

Pre-limitations evidence is especially warranted as a supplement to limitations data where, as here, Defendant itself contends that the promotion process for any given candidate takes place over the course of many years. See, e.g., Hoover Decl., Docket No. 586, ¶ 12 (explaining that regional management receives recommendations from GMs as to promotable candidates and develop a familiarity with potential candidates over the course of years); Omoss Decl., Docket No. 611, ¶ 8 (" [T]he walks by me as a regional manager are but one of several sets of many observations, made over many years, regarding the candidate's job performance." ). Indeed, Defendant itself provides a volume of evidence stemming from before the statute of limitations in the form of the over 80 declarations from Class Members describing their employment histories, which sometimes stretch back into the 1980s. Defendant offers no basis for why its proffered evidence outside the limitations period should be considered to rebut Plaintiffs' charges, while Plaintiffs' evidence should be confined to the limitations period.

In reply, Defendant offers a new challenge to the use of pre-limitations data, arguing that there was a change in Defendant's promotion practices under the Rothman Workplan that renders pre-limitations data inapposite. Setting aside the fact that Defendant failed to raise this argument until its reply, the Ninth Circuit expressly found to the contrary. Ellis II, 657 F.3d at 979 & n. 5 (" Costco's challenged promotion practices for GM and AGM positions have not changed.... Costco's adoption of the Rothman Workplan does not change this result." ). Judge Patel also noted that while Defendant made certain changes in response to its BOLD initiative, such changes largely did not extend to the AGM and GM promotional process. Ellis I, 240 F.R.D. at 640.

Based on the Court's own review of the record, the Court agrees with these findings. Accordingly, the Court rejects Defendants' challenge to Plaintiffs' evidence on the basis of the statute of limitations and finds Dr. Drogin's analysis persuasive. In addition, apart from the merits, Defendant's argument— essentially, that its companywide changes to its previous practices through the Rothman Workplan negates Plaintiffs' evidence of discriminatory promotions— further confirms the commonality finding of this Court. See Ellis II, 657 F.3d at 983 n. 8 (" The district court is required to examine the merits of the underlying claim in this context, only inasmuch as it must determine whether common questions exist; not to determine whether class members could actually prevail on the merits of their claims." ) (emphasis added).

Defendant also challenges Plaintiffs' evidence on the basis of post-lawsuit data showing a decrease in disparities between the promotion of men and women to AGM and GM positions. However, " post-charge hiring behavior is less probative than pre-charge conduct because a business may be improving its hiring practices to avoid liability or large damages in their pending discrimination case." E.E.O.C. v. Joe's Stone Crab, Inc., 220 F.3d 1263, 1277 n. 14 (11th Cir.2000) (citing James v. Stockham Valves & Fittings Co., 559 F.2d 310, 325 n. 18 (5th Cir.1977); Rowe v. General Motors Corp., 457 F.2d 348, 359 (5th Cir.1972)); see also Stender v. Lucky Stores, Inc., C 88-1467 MHP, 1991 WL 127073, at *4 (N.D.Cal. Apr. 4, 1991) (" The court agrees that ‘ [a]ctions taken in the face of litigation are equivocal in purpose, motive and permanence.’ " ) (quoting James, 559 F.2d at 325 n. 18); see generally Teamsters, 431 U.S. at 341-42, 97 S.Ct. 1843 (1977) (" The company's later [post-suit] changes in its hiring and promotion policies could be of little comfort to the victims of the earlier ... discrimination, and could not erase its previous illegal conduct or its obligation to afford relief to those who suffered because of it." ); E.E.O.C. v. Goodyear Aerospace Corp., 813 F.2d 1539, 1544 (9th Cir.1987) (" An employer that takes curative actions only after it has been sued fails to provide sufficient assurances that it will not repeat the violation to justify denying an injunction." ). The Court rejects Defendant's contention that post-filing data negates the disparity shown by Plaintiffs.

In any event, regardless of its probative value on the merits, Defendant's argument is once again premised on a common contention regarding the class as a whole, and therefore actually supports the finding of commonality. Indeed, the substantial increase in women promoted to AGM and GM post-filing, if anything, provides further support for Plaintiffs' claim that Defendant exercises companywide control over its promotion practices and implements a " common mode of exercising discretion." Dukes, 131 S.Ct. at 2554. If promotional decisions were truly localized and outside upper management's control, it would be implausible to suggest that each local manager suddenly decided to promote more women after Plaintiffs filed this lawsuit. See, e.g., Saad Supp. Decl., Docket No. 620, Ex. R40 (showing that women's promotions to AGM underperformed relative to their proportions in the candidate pool each year pre-filing, while their promotions overperformed relative to their proportions in the candidate pool post-filing); Drogin Decl., Docket No. 666, ¶ 20 (" In the period after the lawsuit was filed, the percent of women among those promoted nearly doubled, from 18.4% during 1999 through July 2004 to 34.7% in the period from August 2004 forward." ). ³³ Thus, far from undermining commonality, the Court concludes that Defendant's evidence of post-filing changes in the proportion of women hired into the positions at issue supports commonality.

33 Similarly, the post-filing change cuts against Defendant's other purported explanations for the gender disparity in promotions, such as MM experience, leaves of absence, and other indicators of women's purported lower interest in promotion. There is no indication that these indicators changed drastically from pre-filing to post-filing. Thus, the post-filing change renders Defendant's challenges to Plaintiffs' statistical evidence even less persuasive.

In sum, the Court rejects Defendant's arguments regarding the time period for measuring gender disparity insofar as those arguments bear on commonality.

D. Conclusion

Although Defendant likens this case to Dukes, this case is distinguishable from Dukes . First, as described above, the putative class in this case is smaller and the scope of the claims far narrower than in Dukes . Second, unlike in Dukes, here Plaintiffs point to several companywide policies and practices allegedly responsible for the disparity in promotions to AGM and GM. In Dukes, " because there was no company-wide policy to challenge ...— the only relevant corporate policies were a policy forbidding sex discrimination and a policy of delegating employment decisions to local managers— there was no common issue to justify class treatment." McReynolds, 672 F.3d at 488. The instant case presents not simply the absence of a policy, as in Dukes, but discrete companywide policies guided and supervised by a relatively small and coherent group of company executives. These policies and practices, set and enforced by upper management, provide the " glue" the Supreme Court sought— but did not find— in Dukes, sufficient 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. " Dukes, 131 S.Ct. at 2552. Third, while the Supreme Court found that Dukes had attempted to rely only on a contention that there was a common culture within the company, here Plaintiffs' cultural and cognitive bias evidence provides a supplement to their concrete evidence of companywide policies and practices. Fourth, Plaintiffs' statistical evidence demonstrates classwide— as opposed to fragmented or localized— gender disparities supporting its contention that Defendant's classwide practices yield classwide effects. In sum, although this case bears some superficial factual resemblance to Dukes, it is in reality a much different case. Because Plaintiffs have presented significant proof of companywide policies and companywide gender disparities— essentially, purported common " causes" and common " effects" — the Court finds Plaintiffs have satisfied the requirement for commonality under Rule 23(a)(2) for purposes of their disparate treatment claim.

iv. Disparate Impact

Although the above analysis applies with equal force to Plaintiffs' disparate impact claim, the Court briefly addresses disparate impact separately because under Dukes, the disparate impact analysis is even clearer. Because the question under this theory is whether Defendant's policies and practices have a discriminatory impact on the Class as a whole without regard to intent, the Dukes -identified problem of decentralized and discretionary individual managers' decisions presents less of a hurdle to certification if the plaintiffs identify specific companywide employment practices responsible for the disparate impact. See Dukes, 131 S.Ct. at 2554 (" ‘ [I]n appropriate cases,’ giving discretion to lower-level supervisors can be the basis of Title VII liability under a disparate-impact theory— since ‘ an employer's undisciplined system of subjective decisionmaking [can have] precisely the same effects as a system pervaded by impermissible intentional discrimination.’ " ) (quoting Watson, 487 U.S. at 990-91, 108 S.Ct. 2777); id. at 2555 (conditioning the ability to bring such a claim on " ‘ identifying the specific employment practice that is challenged’ " ) (quoting Watson, 487 U.S. at 994, 108 S.Ct. 2777).

As described above, Plaintiffs have identified such specific employment practices within Costco's promotion system for AGMs and GMs. This system includes inter alia, a tap-on-the-shoulder appointment process (without an application or interview), the mandated lack of posting for open positions, promotion exclusively from within, a requirement of MM experience (for AGM) and AGM experience (for GM), reliance on unwritten and informal evaluation of candidates by senior management ( e.g., floor walks), reliance on promotable lists of desired candidates, reliance on common but unvalidated criteria for assessing candidates (supplied by the CEO and other top executives), and placing a premium on schedule flexibility and ability to relocate. Plaintiffs' argument— that such companywide practices lead to disparate outcomes— is a common question subject to classwide proof and rebuttal.

The Seventh Circuit's recent decision in McReynolds is particularly instructive on this point. In McReynolds, the Seventh Circuit considered whether 700 black Merrill Lynch brokers could present claims for racially discriminatory impact and injunctive relief through class treatment. ³⁴ 672 F.3d 482. The plaintiffs challenged two company-wide practices that gave brokers autonomy within their respective branch offices. First, the company's " teaming" policy allowed brokers to self-select into teams that " share clients, and the aim in forming or joining a team is to gain access to additional clients, or if one is already rich in clients to share some of them with brokers who have complementary skills that will secure the clients' loyalty and maybe persuade them to invest more with Merrill Lynch." Id. at 488. Branch managers and regional directors did not have authority to direct brokers in how they selected team members. The plaintiffs argued that this policy permitted brokers to discriminate in who they selected to be team members, with an adverse disparate impact on black brokers. Second, the plaintiffs challenged Merrill Lynch's account distribution policy, which provided the process by which brokers competed for a departing broker's accounts. " The company establishes criteria for deciding who will win the competition[, ...] includ[ing] the competing brokers' records of revenue generated for the company and of the number and investments of clients retained." Id. at 489. The plaintiffs argued that this account distribution policy fed on and reinforced the discriminatory effects of the team policy; because black brokers were less likely to gain admittance onto good teams, they were also less likely to be in a position to win the account distribution competitions.

34 Although the McReynolds plaintiffs sought certification of their disparate impact claim only under Rule 23(c)(4) (which provides for class action treatment " with respect to particular issues" ), the court's analysis is nonetheless highly relevant for purposes of commonality, as it considered whether Merrill Lynch's employment practices presented an issue common to the class as a whole and whether resolving questions as to the alleged disparate impact of the defendant's practices in a class action would be an efficient and effective means of moving the litigation forward. These are precisely the questions raised by Dukes and by the parties in this case.

The Seventh Circuit reversed the district court's denial of class certification, and found that Merrill Lynch's company-wide policies presented common issues proper for classwide adjudication despite the fact that such policies permitted discretion on the part of brokers. In so doing, it provided a useful hypothetical:

Suppose a police department authorizes each police officer to select an officer junior to him to be his partner. And suppose it turns out that male police officers never select female officers as their partners and white officers never select black officers as their partners. There would be no intentional discrimination at the departmental level, but the practice of allowing police officers to choose their partners could be challenged as enabling sexual and racial discrimination— as having in the jargon of discrimination law a " disparate impact" on a protected group— and if a discriminatory effect was proved, then to avoid an adverse judgment the department would have to prove that the policy was essential to the department's mission. 42 U.S.C. § 2000e-2(k)(1)(A)(i); Ricci v. DeStefano, 557 U.S. 557, 129 S.Ct. 2658, 2672-73, 174 L.Ed.2d 490 (2009); Bryant v. City of Chicago, 200 F.3d 1092, 1098-99 (7th Cir.2000). That case would not be controlled by Wal-Mart (although there is an undoubted resemblance), in which employment decisions were delegated to local managers; it would be an employment decision by top management.

McReynolds, 672 F.3d at 489.

McReynolds also acknowledged that the derivative effects of a companywide policy could themselves present issues common to the class. For example, in that case, the fact that the team policy had an adverse impact on black brokers generated a derivative additional disadvantage to the class, because their inability to benefit from the team process made them also less able to benefit from the account distribution policy. See id. at 489-90 (" [I]f as a result of racial preference at the team level black brokers employed by Merrill Lynch find it hard to join teams, or at least good teams, and as a result don't generate as much revenue or attract and retain as many clients as white brokers do, then they will not do well in the competition for account distributions either; and a kind of vicious cycle will set in." ). The court concluded that " [t]his spiral effect attributable to company-wide policy and arguably disadvantageous to black brokers presents another question common to the class, along with the question whether, if the team-inflected account distribution system does have this disparate impact, it nevertheless is justified by business necessity." Id. at 490.

The same could be said of the GM promotion process and its dependency on AGM experience at Costco as well as the AGM depending on MM experience. Because the process for reaching these positions is tainted, according to Plaintiffs, it results in a dwindled pool of female candidates who would be eligible for the higher-level promotion. In McReynolds, the discriminatory selection processes turned on the fact that those who choose " tend to base decisions on emotions and preconceptions, for want of objective criteria," and that those preconceptions lead them to choose " people who are like themselves." Id. at 489. The claim of discrimination in the case at bar is even stronger than McReynolds. Here the exercise of discretion is confined and guided by companywide policies and practices and directed by upper management, and there is evidence of a specific corporate culture that disfavors women compared to men. More is alleged (and supported by evidence) than in McReynolds where the discrimination resided in the exercise of discretion by and personal preferences of brokers.

Furthermore, the court in McReynolds found commonality sufficient for certification based on the policy which merely authorized the exercise of brokers' discriminatory preferences based on the contribution effect of that structure. The court explained:

[T]he defendant's brief states that " any discrimination here would result from local, highly-individualized implementation of policies rather than the policies themselves." That is too stark a dichotomy. Assume that with no company-wide policy on teaming or account distribution, but instead delegation to local management of the decision whether to allow teaming and the criteria for account distribution, there would be racial discrimination by brokers or local managers, like the discrimination alleged in Wal-Mart. But assume further that company-wide policies authorizing broker-initiated teaming, and basing account distributions on past success, increase the amount of discrimination. The incremental causal effect (overlooked by the district judge) of those company-wide policies— which is the alleged disparate impact— could be most efficiently determined on a class-wide basis.

Id. at 490 (emphasis added).

Again the case at bar presents an even more compelling case. Not only did Costco's policy authorize the exercise of discretion at particular levels which has (allegedly) effectively " increase[d] the amount of discrimination," id., there is substantial and persuasive evidence that the discretion was guided and influenced by discrete policies, practices, and culture which disfavored women.

Accordingly, consistent with Dukes and McReynolds, the Court concludes that along with Plaintiffs' disparate treatment described above, Plaintiffs' disparate impact claim also satisfies the commonality requirement.

c. Typicality

Rule 23(a)(3) requires that the " claims or defenses of the representative parties [be] typical of the claims or defenses of the class." Courts assess typicality by determining " whether other members have the same or similar injury, whether the action is based on conduct which is not unique to the named plaintiffs, and whether other class members have been injured by the same course of conduct." Hanon v. Dataproducts Corp., 976 F.2d 497, 508 (9th Cir.1992).

Judge Patel found that Plaintiffs satisfied the typicality requirement because their claims were typical of the Class's as a whole, and that Defendant's purported unique defenses did not change that conclusion because " as a general matter, individualized defenses do not defeat typicality." Ellis I, 240 F.R.D. at 641 (citing Hanon, 976 F.2d at 508-09). On appeal, the Ninth Circuit vacated this finding because

Hanon, however, supports Costco's position. In Hanon, we stated that " a named plaintiff's motion for class certification should not be granted if ‘ there is a danger that absent class members will suffer if their representative is preoccupied with defenses unique to it.’ " 976 F.2d at 508 (quoting Gary Plastic Packaging Corp. v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 903 F.2d 176, 180 (2d Cir.1990)). We found that the named plaintiff did not satisfy Rule 23(a)'s typicality requirement because his " unique background and factual situation require[d] him to prepare to meet defenses that [were] not typical of the defenses which may be raised against other members of the proposed class." Id.

Ellis II, 657 F.3d at 984. The Ninth Circuit did not address Costco's contentions as to its purported unique defenses on the merits; rather, it remanded for this Court to consider in the first instance " whether these defenses are typical of those that Costco may raise against other members of the class or whether they are unique such that Plaintiffs cannot satisfy Rule 23(a)'s typicality requirement." Id. at 985. Costco admits that " [a] typical defense would be that the non-selected person lacked the skill or ability, or seniority (if skill and ability were equal) possessed by the individual who got the promotion[, or] that the unsuccessful candidate lacked merchandising experience." Costco Reply at 26.

In Hanon, the Ninth Circuit found that a plaintiff in a securities fraud action would be subject to unique defenses that precluded typicality because he was subject to unique defenses which could " threaten to become the focus of the litigation." Hanon, 976 F.2d at 508 (quotation omitted). Specifically, the court found that

Hanon's unique background and factual situation require him to prepare to meet defenses that are not typical of the defenses which may be raised against other members of the proposed class. Hanon's reliance on the integrity of the market would be subject to serious dispute as a result of his extensive experience in prior securities litigation, his relationship with his lawyers, his practice of buying a minimal number shares of stock in various companies, and his uneconomical purchase of only ten shares of stock in Dataproducts.

Id. Thus, the court concluded that " [b]ecause of Hanon's unique situation, it is predictable that a major focus of the litigation will be on a defense unique to him." Id. at 509.

In the instant case, Costco's purported defenses against the Named Plaintiffs are not unique in such a way that they would create a distraction that will become a " major focus of the litigation." First, with respect to Plaintiff Horstman, Costco argues " that, due to family reasons, Horstman rejected rotation to Front End Manager and stated several times that she wished to defer her pursuit of promotion for three to five years." Ellis II, 657 F.3d at 984. As noted above, this defense is far from unique among class members, according to Defendant. Indeed, Costco presents a collection of such examples from numerous class members, and argues that women's personal choices account for a large portion of any disparity in promotion. See, e.g., Kadue Decl., Docket No. 544, ¶ 6 (providing list of dozens of class members Costco argues deferred promotional opportunities for personal or family reasons, and listing such a choice as among the basic factual propositions to be gleaned from Costco's employee declarations). Thus, far from atypical, Costco's defense against Horstman represents a key defense typical of claims it will raise against the class as a whole. Costco's defense as to Plaintiff Horstman thus supports, rather than undermines, typicality.

Second, with respect to Plaintiff Sasaki, Costco claims " that Sasaki is not an outstanding performer and that Costco's expert has concluded that there is no statistical evidence supporting a claim that females are promoted to GM at a lesser rate in Sasaki's region." Ellis II, 657 F.3d at 984. Such a claim is indistinguishable from Costco's own admission that a typical defense would constitute an argument that an employee lacked the skill or ability to obtain the promotion. Costco Reply at 26. Thus, this defense is not unique and does not undermine typicality.

Finally, with respect to Plaintiff Ellis, Costco asserts that " Ellis misrepresented her way into Costco, lacked the Costco experience of other AGMs, transferred to a market with limited promotional opportunities, and was disciplined for abusing subordinates." Ellis II, 657 F.3d at 984-85; see also Vachris Decl., Docket No. 627, ¶ ¶ 12-23, 28 (Regional Vice President of Operations for District 2 of San Diego Region, describing Ms. Ellis's discipline in 2004 for overbearing managerial conduct).³⁵ Ellis, of course, disputes these claims on the merits, and points out especially that she was only disciplined after she filed discrimination charges. See Cross-Mot. at 27 n. 17.

35 Although Costco also argues that Ellis's retaliation claim would be a distraction, Plaintiffs propose to sever her retaliation claim from the class claims and try it separately. Defendant does not oppose such a severance, and the Court agrees that Ellis's retaliation claim should be addressed through separate proceedings. Accordingly, the Court does not address whether the retaliation claim would defeat typicality.

Ellis's personal circumstances present a somewhat closer call on typicality. However, the Court agrees that Costco's defenses against Ellis largely fall into the " typical" categories Defendant has already conceded. For example, her lack of Costco-specific experience is simply another way of saying she lacked seniority, a defense Costco admits is typical. Similarly, her transfer to a limited market fits within Costco's broader claim that women, on average, choose not to put themselves in promotion positions for various reasons, and that employees willing to relocate have greater promotion potential. See Stockdale Decl., Docket No. 650, at 22-23 (women's job preferences more likely to be constrained by family demands); Mulligan Decl., Docket No. 647, ¶ 114 (same); Vachris Decl., Docket No. 627, ¶ ¶ 6, 9.

Costco's other defenses against Ellis present a closer case. Costco argues that Ellis misrepresented her previous job experience and performance at Sam's Club in order to get her job at Costco, and that had Costco known of such a representation it would not have hired her. Gaherty Decl., Docket No. 571, ¶ 8. In addition, it argues that she was subject to personal discipline for poor behavior that caused her to be removed from the promotable list. Vachris Decl., Docket No. 627, ¶ ¶ 12-23. However, at bottom, such claims are merely an extension of the typical claim that Ellis lacked the " skill or ability" necessary for the promotion, based on her job performance. Costco Reply at 26. Courts have rejected challenges to typicality even when alleged misrepresentations are involved, because such claims are merely alternative explanations for the alleged discrimination, and it is " always the defendant's contention in class action discrimination claims[ ] that the plaintiffs suffered no discrimination, or at least that any discrimination that occurred was isolated rather than systematic." Velez v. Novartis Pharmaceuticals Corp., 244 F.R.D. 243, 268 (S.D.N.Y.2007); see also id. (" Defendants cannot rebut typicality by claiming that something other than discrimination explains the named plaintiffs' experience." ); Duling v. Gristede's Operating Corp., 267 F.R.D. 86, 97-98 (S.D.N.Y.2010). Moreover, the Court is not convinced that Ellis's two offenses are likely to become a " major focus" of the litigation especially when compared to the common and typical classwide issues that must be resolved.

Accordingly, the Court concludes Defendant's defenses against the Named Plaintiffs " are typical of those that Costco may raise against other members of the class," thus satisfying typicality. Ellis II, 657 F.3d at 985.

d. Adequacy

Rule 23(a)(4) requires that the class representatives " fairly and adequately protect the interests of the class." " This factor requires: (1) that the proposed representative Plaintiffs do not have conflicts of interest with the proposed class, and (2) that Plaintiffs are represented by qualified and competent counsel." Dukes, 603 F.3d at 614. As noted above, the Ninth Circuit affirmed in part and vacated in part Judge Patel's ruling that each of the three Plaintiffs was an adequate representative, but on remand Defendant does not contest that any Plaintiff is adequate under Plaintiffs' new hybrid class approach. Thus, adequacy is not in dispute.

2. Rule 23(b)(2) — Injunctive Relief Class

In addition to the threshold requirements of Rule 23(a), plaintiffs seeking class certification must also satisfy one of the prongs of Rule 23(b). In the instant case, Plaintiffs seek to certify a class of current employees under (b)(2), represented by Plaintiff Sasaki, for purposes of adjudicating liability and securing injunctive relief on behalf of this class. Plaintiffs seek to certify a (b)(3) class for purposes of monetary and individualized equitable relief ( e.g., " rightful place" relief).

Rule 23(b)(2) mandates class certification if " the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole." " The key to the (b)(2) class is the indivisible nature of the injunctive or declaratory remedy warranted— the notion that the conduct is such that it can be enjoined or declared unlawful only as to all of the class members or as to none of them." Dukes, 131 S.Ct. at 2557 (internal quotations omitted). However, the Rule does not permit certification of individualized monetary relief claims. See id. (Rule 23(b)(2) " does not authorize class certification when each class member would be entitled to an individualized award of monetary damages" ).

As a preliminary matter, the parties dispute whether the propriety of an injunctive relief class under (b)(2) is properly at issue in this round of briefing. Costco's initial motion did not address the propriety of an injunctive relief class under 23(b)(2). Nor did the Ninth Circuit address the issue on appeal; rather, Costco's argument on appeal was that " the district court abused its discretion by certifying the class pursuant to Rule 23(b)(2), because Plaintiffs do not primarily seek injunctive relief." Ellis II, 657 F.3d at 986 (emphasis added). The Ninth Circuit vacated Judge Patel's ruling as to certification under (b)(2) not because of any issue with regard to the injunctive relief sought, but rather because the standard it employed regarding the availability of monetary relief under (b)(2) was erroneous in light of Dukes . Id. at 987. That problem has been obviated by the narrowing of relief now sought under (b)(2).

In opposition to Plaintiffs' cross-motion for class certification, Defendant raises for the first time the argument that Plaintiffs' injunctive relief claims are non-certifiable under (b)(2). First, Defendant argues that any " rightful place" relief would be subject to the same individualized determinations that plague Plaintiffs' claims for monetary relief, and that therefore such injunctive relief is impermissible because (b)(2) " does not authorize class certification when each individual class member would be entitled to a different injunction or declaratory judgment against the defendant." Dukes, 131 S.Ct. at 2557. Plaintiffs do not contest this point, and concede that any " rightful place" equitable relief should be covered by (b)(3) and adjudicated during Plaintiffs' proposed Stage 2 hearings along with individualized damages determinations. Plaintiffs' Reply at 12 n. 14.

Second, Defendant argues that Plaintiffs' classwide injunctive relief claims fail under (b)(2) because any injunctions requiring, e.g., job posting and interviews, " might help some class members, while hurting others." Costco Reply at 29. On this point, Plaintiffs contend that Defendant failed to raise this claim on appeal and that it has therefore waived any claim against certifying such injunctive relief claims at this point. See Ellis I, 240 F.R.D. at 643 (" The injunctive relief sought by the plaintiffs in this case would produce far-reaching changes at Costco and benefit class members in the same way." ); Ellis II, 657 F.3d at 975 (issuing limited ruling that " [i]n light of Wal-Mart's rejection of the ‘ predominance’ test, 131 S.Ct. at 2557-59, the district court must consider whether the claims for various forms of monetary relief will require individual determinations and are therefore only appropriate for a Rule 23(b)(3) class. Thus, we vacate the district court's certification of the class under Rule 23(b)(2)." ). The Court agrees.

Even assuming the issue is properly raised, Defendant provides no authority for its argument that injunctive relief applicable to the class as a whole is unwarranted under (b)(2) simply because some women have managed to receive promotions even under the old system. In Rodriguez v. Hayes, 591 F.3d 1105, 1125 (9th Cir.2010), the Ninth Circuit held, " The fact that some class members may have suffered no injury or different injuries from the challenged practice does not prevent the class from meeting the requirements of Rule 23(b)(2)." See also Floyd v. City of New York, 283 F.R.D. 153, 173 (S.D.N.Y.2012) (" [E]ven after Wal-Mart, Rule 23(b)(2) suits remain appropriate mechanisms for obtaining injunctive relief in cases where a centralized policy is alleged to impact a large class of plaintiffs, even when the magnitude (and existence) of the impact may vary by class member." ) (citations omitted). While Costco points to declarations from certain putative class members who were able to succeed under the current system, and thus may not receive as great a benefit from any changes to that system, it provides no logical or evidentiary support for the proposition that injunctive relief would hurt members of the class. See Costco Reply at 29 & n. 166 (listing class members' declarations describing support they received in promotions process). In any event, Costco's argument, if accepted, would be far reaching and eviscerate the availability of (b)(2) certification for injunctive relief in virtually all employment cases. Unless a discriminatory practice were to preclude, e.g., all women from receiving any promotion, there would always be employees who succeed despite a discriminatory pattern or practice. The Court rejects Defendant's argument.

Because Defendant raises no other challenges to certification under (b)(2), and because " class members complain of a pattern or practice that is generally applicable to the class as a whole," and any classwide injunctive relief would address said pattern or practice, Plaintiffs have met the requirements of (b)(2) for purposes of seeking injunctive relief. ³⁶ In addition, because liability turns on Defendant's alleged patterns or practices, adjudication of liability is also appropriate under (b)(2) in the first stage of trial. See, e.g., United States v. City of New York, 258 F.R.D. 47, 67 (E.D.N.Y.2009) (certifying class at liability stage under (b)(2) and reserving until later whether class treatment of the damages phase would be warranted). Furthermore, liability under Plaintiffs' disparate impact theory turns on Defendant's specific employment practices that, Plaintiffs contend, have classwide discriminatory effects. See McReynolds, 672 F.3d at 491-92 (certifying (b)(2) class for liability and injunctive relief in disparate impact case). Accordingly, the Court GRANTS Plaintiffs' motion to certify a Rule 23(b)(2) class for purposes of liability and injunctive relief.

36 Indeed, Plaintiffs' complaint highlights the classwide nature of their requested relief. The complaint seeks " [a] preliminary and permanent injunction against Costco and its officers, owners, agents, successors, employees, representatives, and any and all persons acting in concert with them, that requires the following:

3. Rule 23(b)(3) — All Monetary and Individual Equitable Relief

" [T]he presence of commonality alone [under 23(a)(2) ] is not sufficient to fulfill Rule 23(b)(3)." Hanlon, 150 F.3d at 1022. Rather, " [t]o qualify for certification under [Rule 23(b)(3) ], a class must satisfy two conditions in addition to the Rule 23(a) prerequisites: common questions must " predominate over any questions affecting only individual members," and class resolution must be " superior to other available methods for the fair and efficient adjudication of the controversy." Fed.R.Civ.P. 23(b)(3). In the instant case, Plaintiffs seek to certify a Rule 23(b)(3) class for purposes of monetary and equitable remedies, including back pay, rightful place relief, and compensatory and punitive damages. Because Defendant raises a number of challenges to Plaintiffs' punitive damages claim, the Court will address punitive damages separately after considering predominance and superiority with respect to the remaining claims.

a. Predominance

In evaluating whether common issues predominate, the operative question is whether a putative class is " sufficiently cohesive" to merit representative adjudication. Amchem Prod., Inc. v. Windsor, 521 U.S. 591, 623, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). Mere allegations of a pattern or practice of discrimination do not per se satisfy the predominance requirement. See Falcon, 457 U.S. 147, 157, 102 S.Ct. 2364 (" [T]he allegation that such discrimination has occurred neither determines whether a class action may be maintained in accordance with Rule 23 nor defines the class that may be certified" ). Such allegations must, as here, still be subject to generalized proof. Though common issues need not be " dispositive of the litigation," In re Lorazepam & Clorazepate Antitrust Litig., 202 F.R.D. 12, 29 (D.D.C.2001), they must " present a significant aspect of the case [that] can be resolved for all members of the class in a single adjudication" so as to justify " handling the dispute on a representative rather than an individual basis." Hanlon, 150 F.3d at 1022. Courts must thus separate the issues subject to " generalized proof" from those subject to " individualized proof" to determine whether plaintiffs have satisfied the predominance requirement. See In re Dynamic Random Access Memory (DRAM) Antitrust Litig., No. M 02-1486, 2006 WL 1530166, at *6 (N.D.Cal. June 5, 2006) (" Predominance requires that the common issues be both numerically and qualitatively substantial in relation to the issues peculiar to individual class members." ) (internal quotation omitted).

In the instant case, the Court concludes that common issues predominate and that the class is sufficiently cohesive to warrant (b)(3) certification. As discussed above with respect to commonality and typicality, Plaintiffs have presented significant proof that Costco operates under a common, nationwide promotion system for its AGM and GM positions and have identified specific employment practices that have caused a disparity in promotions. Plaintiffs contend that this system is discriminatory, both under a disparate treatment and a disparate impact theory. And as noted above, Defendant's arguments in response largely confirm the classwide nature of the parties' dispute. Resolution of Plaintiffs' challenge to those practices will resolve significant issues with respect to the class as a whole, and this dwarfs individualized issues as to particular employment decisions. Stinson v. City of New York, 282 F.R.D. 360, 382 (S.D.N.Y.2012) (" ‘ Class-wide issues predominate if resolution of some of the legal or factual questions that qualify each class member's case as a genuine controversy can be achieved through generalized proof, and if these particular issues are more substantial than the issues subject only to individualized proof.’ " ) (quoting Moore v. PaineWebber, Inc., 306 F.3d 1247, 1252 (2d Cir.2002); citing Erica P. John Fund, Inc. v. Halliburton Co., --- U.S. ----, 131 S.Ct. 2179, 2184-87, 180 L.Ed.2d 24 (2011)).

With respect to the disparate treatment claim, whether Defendant has engaged in a pattern or practice of discrimination such that all class members are entitled to a presumption of discrimination under the Teamsters method of proof is a common issue subject to classwide resolution. This " pattern and practice question predominates because it has a direct impact on every class member's effort to establish liability and on every class member's entitlement to ... monetary relief." Ingram v. The Coca-Cola Co., 200 F.R.D. 685, 699 (N.D.Ga.2001) (certifying (b)(3) class of plaintiffs alleging a pattern or practice of race discrimination in employment under Teamsters framework); see also Williams v. Mohawk Indus., Inc., 568 F.3d 1350, 1357 (11th Cir.2009) (" Common issues of fact and law predominate if they ha[ve] a direct impact on every class member's effort to establish liability and on every class member's entitlement to injunctive and monetary relief." ) (internal citations and quotation marks omitted). Similarly, with respect to the disparate impact claim, whether Defendant's facially neutral policies and practices have a disparate impact on class members, and whether those practices are nonetheless justified by business necessity, are similarly issues best addressed with respect to the entire class.³⁷ Adjudicating these issues on a classwide basis is necessary before any individualized proceeding can occur. See City of New York, 276 F.R.D. at 48 (" The [predominance] question is not one of scale; instead it is whether certification ‘ would achieve economies of time, effort, and expense, and promote uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results.’ " ) (quoting Myers v. Hertz Corp., 624 F.3d 537, 547 (2d Cir.2010)).³⁸

37 Because of the hybrid approach— certifying certain portions of Plaintiffs' claims under (b)(2) and others under (b)(3)— suggested by Plaintiffs and mandated by Dukes, the common issues relevant to Plaintiffs' claims for monetary and equitable relief overlap with the common issues present for the (b)(2) class. See Kartman v. State Farm Mut. Auto. Ins. Co., 634 F.3d 883, 895 (7th Cir.2011), cert. denied, --- U.S. ----, 132 S.Ct. 242, 181 L.Ed.2d 138 (2011) (" [I]n an appropriate case, a Rule 23(b)(2) class and a Rule 23(b)(3) class may be certified where there is a real basis for both damages and an equitable remedy." ) (citations omitted).

38 In addition, even with respect to individuals' claims for monetary relief, there may be additional common questions to resolve before the individual hearings begin. For example, " ‘ where the number of qualified class members exceeds the number of openings lost to the class through discrimination and identification of individuals entitled to relief would drag the court into a quagmire of hypothetical judgments and result in mere guesswork,’ [citation], a case ‘ may require class-wide, rather than individualized, assessments of monetary relief.’ [citation]." City of New York, 847 F.Supp.2d at 408-09 (quoting Catlett v. Mo. Highway & Transp. Comm'n, 828 F.2d 1260, 1267 (8th Cir.1987); Robinson, 267 F.3d at 161 n. 6); see also City of New York, 276 F.R.D. at 44-45. Although Dukes disapproved of a so-called " Trial by Formula" without any individualized proceedings, 131 S.Ct. at 2561, there may be common issues to be resolved before embarking on such individual hearings.

Although this case does present individualized questions with respect to any particular class member's entitlement to relief, Plaintiffs' proposed trial plan addresses these concerns by employing the Teamsters framework, in which individual class members will present their claims for relief in a second phase of trial if liability is established, and Defendant will have an opportunity to present individualized defenses with respect to each class member. See, e.g., City of New York, 276 F.R.D. at 34 (" Individual issues arise in disparate impact and pattern-or-practice disparate treatment cases only if the class establishes the employer's liability and the litigation proceeds to the remedial phase." ). The need for individualized hearings does not, on its own, defeat class certification. See, e.g., Blackie v. Barrack, 524 F.2d 891, 905 (9th Cir.1975) (" The amount of damages is invariably an individual question and does not defeat class action treatment." ); Stearns v. Ticketmaster Corp., 655 F.3d 1013, 1026 (9th Cir.2011) (citing Blackie in a case decided after Dukes and Ellis II ); In re Whirlpool Corp. Front-Loading Washer Products Liab. Litig., 678 F.3d 409, 419 (6th Cir.2012) (" [N]o matter how individualized the issue of damages may be, these issues may be reserved for individual treatment with the question of liability tried as a class action." ) (internal citations and quotation marks omitted); Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 815 (7th Cir.2012), reh'g denied (Feb. 28, 2012) (" It is well established that the presence of individualized questions regarding damages does not prevent certification under Rule 23(b)(3)." ) (citing, e.g., Dukes, 131 S.Ct. at 2558 (deeming it " clear that individualized monetary claims belong in Rule 23(b)(3)" )) (collecting additional cases); see also Johns v. Bayer Corp., 280 F.R.D. 551, 555-56, 560 (S.D.Cal.2012) (certifying (b)(3) class of consumers who purchased a multi-vitamin due to allegedly misleading advertising, even though Bayer argued some consumers " did not place any credence on the" misleading claim) (internal citations omitted); Herrera v. LCS Fin. Services Corp., 274 F.R.D. 666, 680-81 (N.D.Cal.2011) (" [I]ndividual damages issues typically do not bar class certification where common questions predominate over individual questions as to liability" ) (citing 5 James W. Moore, Moore's Federal Practice § 23.45[2] [a] (3d ed. 2010); Hazelwood v. Bruck Law Offices SC, 244 F.R.D. 523, 525 (E.D.Wis.2007) (holding that common questions as to whether a collection letter violated the FDCPA predominated over individual questions regarding class members' actual damages)). In this case, the Court concludes that the individualized hearings required are narrow in scope and significance when compared to the threshold, classwide issues subject to generalized proof.

In short, the multiple and substantial questions on such classwide issues predominate over the claims for individual relief. The Court concludes that Plaintiffs have satisfied the predominance requirement for (b)(3) certification.

b. Superiority

Rule 23(b)(3) requires that a class action be " superior to other methods for the fair and efficient adjudication of the controversy." Hanlon, 150 F.3d at 1023. The court considers the following " non-exclusive" factors, id., in evaluating superiority:

(A) the class members' interests in individually controlling the prosecution or defense of separate actions;

(B) the extent and nature of any litigation concerning the controversy already begun by or against class members;

(C) the desirability or undesirability of concentrating the litigation of the claims in the particular forum; and

(D) the likely difficulties in managing a class action.

Fed.R.Civ.P. 23(b)(3).

These factors support classwide treatment in the instant case. Given the class's common questions affecting the class as a whole at the liability stages of this matter, and given class members' ability to opt out of the monetary relief class or to pursue their claims for relief in the second state of the proceedings under (b)(3), class members have a diminished interest in individually controlling the common portions of this action. At the same time, the (b)(3) certification framework safeguards " the due process rights of those class members, i.e. the right ‘ to decide for themselves whether to tie their fates to the class representatives' or go it alone.’ " In re Motor Fuel Temperature Sales Practices Litig., 279 F.R.D. 598, 606 (D.Kan.2012) (emphasis in original) (quoting Dukes, 131 S.Ct. at 2559). Further, neither party contends other litigation concerning the class herein has already been asserted elsewhere.

Moreover, the size of the (b)(3) class (estimated to be approximately 700 in number) is manageable. Cf. Herrera, 274 F.R.D. at 680-81 (certifying (b)(3) class of over 4,000 borrowers subject to allegedly deceptive and unlawful debt collection practices following foreclosure of the homes for which they had borrowed the money at issue). Plaintiffs' trial plan also presents a manageable way to adjudicate these class members' claims. Because Plaintiffs point to common, classwide Costco promotion policies they contend are both intentionally discriminatory and cause a disparate impact, judicial economy favors adjudicating their claims together in one proceeding. Indeed, a classwide adjudication is far more manageable than the alternative individual proceedings on all issues, because it has the potential to resolve multiple issues in one proceeding before proceeding to individual hearings on relief. See United States v. City of New York, 276 F.R.D. 22, 49 (E.D.N.Y.2011) (" In deciding whether class certification will achieve substantial efficiencies, the proper comparison is not between class litigation and no litigation at all, but between class litigation and actions conducted separately by individual class members." ). Classwide adjudication on common issues also prevents inconsistent verdicts.

Accordingly, the Court concludes that certification is appropriate under Rule 23(b)(3) for Plaintiffs' monetary and individual equitable relief claims.

c. Punitive Damages

Defendant separately argues that Plaintiffs cannot obtain certification for their punitive damages claims under either (b)(2) or (b)(3) because it would violate both due process and the Rules Enabling Act. Mot. for Order Elim. Class Claims, Docket No. 543, at 23. See generally State Farm Mut. Auto. Ins. Co. v. Campbell, 538 U.S. 408, 426, 123 S.Ct. 1513, 155 L.Ed.2d 585 (2003) (" [C]ourts must ensure that the measure of punishment is both reasonable and proportionate to the amount of harm to the plaintiff and to the general damages recovered." ); Philip Morris USA v. Williams, 549 U.S. 346, 349, 127 S.Ct. 1057, 166 L.Ed.2d 940 (2007) (holding that due process clause prohibited jury from basing punitive damages award " in part upon its desire to punish the defendant for harming persons who are not before the court ( e.g., victims whom the parties do not represent)" ) (emphasis in original). On appeal herein, the Ninth Circuit stated the following regarding punitive damages:

The district court earlier found that claims for punitive damages are suitable for certification under 23(b)(2), " because such ... claim[s] focus [ ] on the conduct of the defendant and not the individual characteristics of the plaintiffs." Ellis, 240 F.R.D. at 643; see also Dukes, 603 F.3d at 622 (noting that plaintiffs' claim for punitive damages did " not require individualized punitive damages determinations" ); Kolstad v. Am. Dental Assoc., 527 U.S. 526, 535, 119 S.Ct. 2118, 144 L.Ed.2d 494 (1999) (noting that whether punitive damages are warranted is based on the employer's state of mind, i.e., if " [t]he employer [acted] with ‘ malice or with reckless indifference to the plaintiff's federally protected rights.’ " (quoting 42 U.S.C. § 1981a(b)(1)) (emphasis and alterations omitted)). The court may consider [on remand] whether punitive damages are an allowable " form[ ] of ‘ incidental’ monetary relief" consistent with the Court's interpretation of 23(b)(2) because they do not require an individual determination. See Wal-Mart, 131 S.Ct. at 2560.

Ellis II, 657 F.3d at 987. Plaintiffs argue the Ninth Circuit explicitly endorsed Judge Patel's finding that punitive damages did not require any individualized inquiry. However, it is not clear from the court's order whether it affirmed Judge Patel's findings on this point or whether (more plausibly) it merely summarized the issues for this Court to consider on remand. The Ninth Circuit's citation to Dukes suggests that it was not resolving the question of whether punitive damages could constitute damages that do not require individualized assessments, as Dukes merely states that it " need not decide in this case whether there are any forms of " incidental" monetary relief that are consistent with the interpretation of Rule 23(b)(2) we have announced and that comply with the Due Process Clause." 131 S.Ct. at 2560.

Assuming, in an abundance of caution, the question whether punitive damages requires some individual inquiry is still present for the Court's resolution, Defendant raises a number of challenges to Plaintiffs' proposed plan to adjudicate punitive damages. First, Defendant argues that Plaintiffs' proposal to determine punitive damages during the first phase of trial along with liability and injunctive relief, before any individualized determination as to compensatory damages or equitable monetary relief, would violate the Seventh Amendment and Supreme Court authority. Second, Defendant argues that punitive damages are not certifiable under Rule 23(b)(3) because classwide adjudication would violate the Rules Enabling Act, which prohibits the Federal Rules of Civil Procedure from " abridg[ing], enlarg[ing] or modify[ing] any substantive right." 28 U.S.C. § 2072(b). Defendant also argues that individualized issues predominate. The Court addresses Defendant's concerns below.

First, as to the stage of proceedings at which Plaintiffs can seek punitive damages, there is a debate among courts as to the best course of action for adjudicating such claims. On one side, for example, a district court recently rejected the EEOC's proposal (similar to Plaintiffs' here) to adjudicate punitive damages at the first phase of trial:

The EEOC's proposal to determine punitive damages at the conclusion of Stage I, before compensatory damages are determined, runs afoul of the principles articulated in State Farm and Philip Morris USA. Under the EEOC's proposed scheme, the Stage I jury will determine the amount of punitive damages to be awarded. Yet, at the time that the jury is being asked to make that finding, there will have been no determination as to the number of class members adversely affected by the discriminatory practice. A finding of liability at the conclusion of Stage I does not necessarily entitle all (or even any) members to compensatory damages.... Rather than ensuring a " proportional relationship" between compensatory and punitive damages, as State Farm instructs, the EEOC's plan seeks to completely divorce any relationship between those determinations.

E.E.O.C. v. Sterling Jewelers Inc., 788 F.Supp.2d 83, 89-90 (W.D.N.Y.2011) (citing State Farm, 538 U.S. 408, 123 S.Ct. 1513; Philip Morris USA, 549 U.S. 346, 127 S.Ct. 1057). There is additional authority in accord. See, e.g., id. at 90-91 (collecting cases and secondary authority supporting the proposition that punitive damages must be determined after, not before, compensatory damages); Allison v. Citgo Petroleum Corp., 151 F.3d 402, 418 (5th Cir.1998) (" [P]unitive damages must be determined after proof of liability to individual plaintiffs at the second stage of a pattern or practice case, not upon the mere finding of general liability to the class at the first stage." ); Nelson v. Wal-Mart Stores, Inc., 245 F.R.D. 358, 378 (E.D.Ark.2007) (" Individualized determinations are necessary to fully realize the extent of the harm caused by Wal-Mart's conduct and properly assess the need for punishment and deterrence." ).

Other cases, cited by Plaintiffs, take their suggested approach and determine punitive damages during the first phase. See Satchell v. FedEx Corp., C 03-02659 SI, 2005 WL 2397522, at *12 (N.D.Cal. Sept. 28, 2005) (" The first phase will address liability and relief applicable to the class as a whole, including declaratory and injunctive relief, and whether defendant is liable for punitive damages." ); E.E.O.C. v. Dial Corp., 259 F.Supp.2d 710, 712 (N.D.Ill.2003) (dividing trial into four phases, with same jury deciding phase one (liability for pattern or practice of discrimination) and phase two (whether punitive damages are available and how much), and with different juries deciding phase three (compensatory damages for individuals and the amount of punitive damages available to each individual) and the court deciding phase four (apportionment of punitive damages)); Barefield v. Chevron, U.S.A., Inc., 12 Fed.R.Serv.3d 1232, 1988 WL 188433 (N.D.Cal.1988).

These cases base their decision on several factors: (1) the focus of punitive damages is on deterrence and punishment for the defendant's conduct, so facts unique to each class member are largely irrelevant, see Barefield, 1988 WL 188433 at *4; (2) judicial economy favors adjudicating punitive damages in phase one in order to avoid repetitive proceedings and inconsistent results, see id. ; and (3) practical and equitable concerns favor determining punitive damages in phase one, because " [r]equiring each class member to litigate the appropriate amount of punitive damages due him or her in stage two proceedings would not only be inefficient but could also result in claimants with comparable actual damages receiving substantially unequal punitive awards," which could result in substantially under- or over-compensating class members, id. at *5. As the court in Dial observed, " [A] determination of an amount of punitive damages for all the persons, as a group, who ultimately are found to be aggrieved by the pattern or practice should be decided by the jury in Phase I and II who has heard all the evidence regarding the nature and scope of the pattern or practice. No jury deciding compensatory damages of an individual or small group of individuals can have the same insight on what will be needed to deter the pattern or practice on a plant-wide basis or for punishment as will the Phase I and II jury." 259 F.Supp.2d at 712-13.

Yet a third category of cases separates the availability of punitive damages (phase one) from the amount of those damages (phase two). For example, another case cited by Plaintiffs explicitly determined that the availability of punitive damages would be adjudicated in phase one because it overlapped substantially with the classwide liability determination, but that the amount of punitive damages would be determined in the second phase. That court explained,

Defendants confuse the issue of " the class's eligibility for award of punitive damages" with the issue of " determination of the amount of punitive damages in individual cases." Section " 1981a(b)(1) requires plaintiffs to make a ... ‘ demonstrat[ion]’ of their eligibility for punitive damages." Wal-Mart Stores, 187 F.3d at 1245 (citing Kolstad, 527 U.S. at 534, 119 S.Ct. at 2124). To demonstrate that an employer is liable for punitive damages, plaintiffs must at least show that the employer discriminated against them " in the face of a perceived risk that its action [would] violate federal law." Id. (citing Kolstad, 527 U.S. at 536, 119 S.Ct. at 2125). Furthermore, and decisively, " [t]he purpose of punitive damages is not to compensate the victim, but to punish and deter the defendant," and thus the focus of a punitive damages claim is " not on the facts unique to each class member, but on the defendant's conduct toward the class as a whole." Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137, 172 (N.D.Cal.2004).

Moreover, the Tenth and Ninth Circuits have allowed the same jury to determine defendant's liability and plaintiff's eligibility for punitive damages. Cf. Markham v. Nat'l States Ins. Co., 122 Fed.Appx. 392, 399 (10th Cir.2004); see also Dukes v. Wal-Mart, Inc., 509 F.3d 1168, 1190 n. 16 (9th Cir.2007) (explaining that the determination of liability in Phase I includes the determination of liability for punitive damages).

....

The question of eligibility of Plaintiffs to punitive damages will be presented to the jury in the liability phase (Phase I). The assessment regarding the amount of compensatory and punitive damages will take place in Phase II.

E.E.O.C. v. Outback Steak House of Florida, Inc., 576 F.Supp.2d 1202, 1205-07 (D.Colo.2008).

In the instant case, the Court concludes that while the availability of punitive damages should be adjudicated in Stage One of the trial, determination of the aggregate amount and individual distribution of punitive damages should be reserved for Stage Two. Such an arrangement will take advantage of the bifurcated trial procedure while safeguarding Defendant's right to ensure that any punitive damages award remains tethered to the compensatory damages actually awarded in Stage Two, consistent with State Farm.

Defendant argues that the No Re-Examination Clause of the Seventh Amendment ³⁹ bars Plaintiffs' bifurcation proposal. However, the Seventh Amendment actually favors the Court's conclusion: separating liability into two different phases (liability and punitive damages) could potentially cause the precise problems Defendant seeks to avoid. That is, the classwide liability question of whether Defendant has engaged in a pattern or practice of intentional discrimination may overlap substantially with the question of whether Defendant acted with malice or reckless indifference to Plaintiffs' protected rights, as Plaintiffs must show in order to obtain punitive damages. In particular, to obtain punitive damages, Plaintiffs will have to demonstrate that Defendant " engaged in a discriminatory practice or discriminatory practices with malice or with reckless indifference to the federally protected rights of an aggrieved individual." Kolstad v. Am. Dental Ass'n, 527 U.S. 526, 534, 119 S.Ct. 2118, 144 L.Ed.2d 494 (1999) (emphasis omitted) (quoting 42 U.S.C. § 1981a(a)(1)). As the Supreme Court explained, punitive damages under § 1981a " focus[es] on the employer's state of mind," and mere evidence of individual decisions or outrageous conduct is insufficient to obtain punitive damages; rather, " [t]he plaintiff must impute liability for punitive damages to" Costco itself. Id. at 535, 536, 119 S.Ct. 2118. Thus, given the nature of Plaintiffs' claims alleging a pattern or practice of discrimination, the punitive damages inquiry necessarily focuses on Defendant's conduct with respect to the class as a whole, rather than any individual employment decisions with respect to specific employees. See Ellis II, 657 F.3d at 987 (punitive damages claims " do not require an individual determination" ); Barefield v. Chevron, U.S.A., Inc., 12 Fed. R. Serv.3d 1232, at *3 (N.D.Cal.1988) (" Because the purpose of punitive damages is not to compensate the victim, but to punish and deter the defendant, any claim for such damages hinges, not on facts unique to each class member, but on the defendant's conduct toward the class as a whole." ) (citing Jenkins v. Raymark Industries, 782 F.2d 468, 474 (5th Cir.1986), reh. denied, 785 F.2d 1034 (5th Cir.1986) (finding, under Texas law and federal Constitution, that " [w]hile no plaintiff may receive an award of punitive damages without proving that he suffered actual damages ... the allocation need not be made concurrently with an evaluation of the defendant's conduct" )); Ellis I, 240 F.R.D. at 643 (same). Thus, trying these potentially overlapping issues of liability and entitlement to punitive damages before a single jury ensures compliance with the Seventh Amendment's prohibition on reexamination.

39 " [N]o fact tried by jury, shall be otherwise re-examined in any Court of the United States ..." U.S. Const. amend. VII. See generally Matter of Rhone-Poulenc Rorer, Inc., 51 F.3d 1293, 1303 (7th Cir.1995) ( " Of particular relevance here, the judge must not divide issues between separate trials in such a way that the same issue is reexamined by different juries." ).

Conversely, separating all classwide liability determinations (of liability and availability of punitive damages) from any quantification of damages would preserve separate questions for separate fact finders and is thus consistent with the Seventh Amendment. See Arthur Young & Co. v. U.S. Dist. Court, 549 F.2d 686, 697 (9th Cir.1977) (" [S]eparation of the trial on individual damage issues from the class trial in this securities fraud class action is not a novel procedure, nor in contravention of the Seventh Amendment." ).⁴⁰ As Judge Illston explained:

40 But see, e.g., Castano v. Am. Tobacco Co., 84 F.3d 734, 751 (5th Cir.1996).

Courts have routinely adopted the approach advocated by plaintiffs in which the first phase of the proceedings focuses exclusively on classwide claims, e.g., whether a defendant has in fact engaged in discriminatory employment practices. A jury verdict in favor of plaintiffs at this phase would result in injunctive and declaratory relief, and possibly, punitive damages. Individual compensatory damages would be resolved in the second phase of the proceedings which, since they would adjudicate individual claims, would not involve the " same issues" as did the first phase. As evidenced by the numerous cases across the country that have addressed this issue, the Seventh Amendment does not mandate that all phases of the litigation be heard by the same jury.

Butler v. Home Depot, Inc., see, e.g., Arnold v. United Artists Theatre Circuit, Inc., see also Robinson v. Metro-N. Commuter R.R. Co., abrogated in part on other grounds by Dukes,

Defendant's due process rights are adequately protected under this scheme. Defendant will have ample opportunity to present defenses to Plaintiffs' punitive damages claim on the level of its companywide policies and practices. If the Plaintiffs were to prevail and establish liability and entitlement to punitive damages despite Defendant's defenses, the remedial phase of the litigation would offer Defendant the opportunity to present evidence as to the proper amount of punitive damages as well as individualized defenses which could defeat any individual class member's claim to punitive damages. This bifurcated approach will safeguard Defendant's right to ensure that the amount of punitive damages is not disproportionate relative to compensatory damages.

Accordingly, the Court determines that (b)(3) certification of Plaintiffs' punitive damages claims is appropriate.

d. Rule 23(c)(4)

Although the Court determines that class certification is appropriate for all of Plaintiffs' claims under a(b)(2)/(b)(3) hybrid approach, even if individualized issues were to predominate with respect to Plaintiffs' monetary relief claims, the Court would utilize the mechanism under Rule 23(c)(4) to adjudicate those issues capable of classwide resolution separately. " When appropriate, an action may be brought or maintained as a class action with respect to particular issues." Fed.R.Civ.P. 23(c)(4); see generally Valentino v. Carter-Wallace, Inc., 97 F.3d 1227, 1234 (9th Cir.1996) (" Even if the common questions do not predominate over the individual questions so that class certification of the entire action is warranted, Rule 23 authorizes the district court in appropriate cases to isolate the common issues under Rule 23(c)(4)(A) and proceed with class treatment of these particular issues." ). In the instant case, for the reasons set forth above, certification would be appropriate under (b)(2) for liability and injunctive relief, and under (b)(3) for the availability of punitive damages, even if certification of the remaining issues was premature or inappropriate. See McReynolds, 672 F.3d at 491 (finding certification appropriate under (b)(2) and (c)(4) where " [t]he practices challenged in this case present a pair of issues that can most efficiently be determined on a class-wide basis, consistent with" Rule 23(c)(4)); City of New York, 276 F.R.D. at 49 (" [E]ven if the individual questions were significant enough to defeat certification under Rule 23(b)(3), the court will isolate those common questions appropriate for class treatment by narrowly certifying the non-hire victim and delayed-hire victim subclasses as to the issues susceptible to classwide proof." ) (citations omitted); Bautista-Perez v. Holder, C 07-4192 TEH, 2009 WL 2031759 (N.D.Cal. July 9, 2009) (" [I]n an abundance of caution, the Court will certify the class under Rule 23(b)(2) as to questions of non-monetary injunctive and declaratory relief, and will defer the issue of certification as to monetary relief until after the liability determination." ). 4. Rule 23(g) — Class Counsel

Having determined that Named Plaintiffs have met Rule 23's requirements for class certification, the Court must also appoint class counsel. See Fed.R.Civ.P. 23(g) (" Unless a statute provides otherwise, a court that certifies a class must appoint class counsel." ). The record reflects that proposed class counsel have experience in handling complex litigation and have done extensive work prosecuting the claims in this action, including through appeal to the Ninth Circuit. The record also shows that they have a command of the applicable law and are willing to commit resources to representing class members' claims. Accordingly, the Court appoints Impact Fund, Lewis, Feinberg, Lee, Renaker & Jackson, P.C., Davis, Cowell & Bowe LLP, Lieff, Cabraser, Heimann & Bernstein LLP, and Altshuler Berzon LLP as Class Counsel.

IV. CONCLUSION

For the foregoing reasons, having conducted a rigorous analysis of the evidence submitted by both sides, the Court finds as follows:

1. The proposed class is so numerous that joinder is impracticable.

2. There are numerous common questions of fact and law, the answers to which are apt to drive the resolution of this case.

3. The claims of the Named Plaintiffs are typical of those of the proposed class.

4. Named Plaintiffs Shirley " Rae" Ellis, Leah Horstman, and Elaine Sasaki are adequate class representatives.

5. Certification of the Injunctive Relief Class under Fed.R.Civ.P. 23(b)(2) is appropriate because Defendant is alleged to have acted and/or refused to act on grounds generally applicable to the class.

6. Certification of the Monetary Relief Class (back pay, compensatory, and punitive damages) under Fed.R.Civ.P. 23(b)(3) is appropriate because common questions predominate over individual questions and class treatment is the superior method of resolving the claims.

7. Impact Fund, Lewis, Feinberg, Lee, Renaker & Jackson, P.C., Davis, Cowell & Bowe LLP, Lieff, Cabraser, Heimann & Bernstein LLP, and Altshuler Berzon LLP will fairly and adequately represent the interests of the class.

Accordingly, the Court hereby ORDERS as follows:

The proposed classes are certified under Rule 23(b)(2) and 23(b)(3) and the classes are defined as:

Injunctive Relief Class:

All women who are currently employed or who will be employed at any Costco warehouse in the U.S. who have been or will be subject to Costco's system for promotion to Assistant General Manager and/or General Manager positions.

Monetary Relief Class:

All women who have been employed at any Costco warehouse store in the U.S. since January 3, 2002 who have been subject to Costco's system for promotion to Assistant General Manager and/or General Manager positions.

The class claims, issues and defenses are those relating to Defendant's liability and, if appropriate, relief for Plaintiffs and the classes. Third Amended Complaint, Docket No. 537, ¶ ¶ 86-113, 140.

Named Plaintiff Elaine Sasaki is appointed as the class representative for the Injunctive Relief Class defined above. Named Plaintiffs Shirley " Rae" Ellis, Leah Horstman and Elaine Sasaki are appointed as class representatives for the Monetary Relief Class defined above.

The Impact Fund, Davis, Cowell & Bowe LLP, Lewis, Feinberg, Lee, Renaker & Jackson P.C., Lieff, Cabraser, Heimann & Bernstein LLP, and Altshuler Berzon LLP, are appointed as counsel to the classes defined above pursuant to Fed.R.Civ.P. 23(g).

The tentative trial plan, subject to modification as the case progresses, is as follows:

Stage One (Part One):

The jury decides:

• Whether Costco has engaged in a pattern or practice of discrimination (liability for classwide disparate treatment); • Whether Costco's conduct meets the standard for an award of punitive damages; and

• Whether Costco is liable to the Named Plaintiffs for gender discrimination and, if so, in what amount;

The Court decides:

• Whether Costco's employment practices have had an adverse impact on the class (prima facie case of disparate impact).

Stage One (Part Two):

The Court would decide:

• Whether Costco's employment practices were justified by business necessity (defense to the disparate impact claim), and if so, whether there was a less discriminatory alternative;

• In the event of a liability finding for either pattern-or-practice discrimination or disparate impact, appropriate injunctive relief;

Stage Two:

• Individual hearings to determine rightful place, back pay and compensatory damages and adjudicate individual defenses; and

• If liable for punitive damages, the aggregate amount of punitive damages owed to the class and the share of said damages owed to each class member.

This Order disposes of Docket Nos. 543 and 664.

IT IS SO ORDERED.

Plaintiffs have identified a number of specific, companywide " employment practices" and " testing procedure[s]." These include the " 360-degree review" process, the forced-quartile ranking of employees, and the " tap on the shoulder" system for selecting employees for promotion. R & R at 13. As opposed to hiring and promotion at Wal-Mart, which was committed to " local managers' broad discretion," based on managers' " own subjective criteria," and " exercised in a largely subjective manner," Dukes, 131 S.Ct. at 2547, the employment practices in this case, together or individually, might well— with the benefit of discovery— comprise a " common mode of exercising discretion that pervades the entire company," Id. at 2554-55.

Chen-Oster at ----, 2012 WL 2912741 at *2. The employment practices to which Chen-Oster referred are similar to some at issue in this case. For example, the " tap on the shoulder" system identified in Chen-Oster is similar to Plaintiffs' described practice here, in which executives choose a list of promotables without objective criteria and without an application process through which employees can indicate interest in promotion opportunities. As in Chen-Oster, managers employed a closed system where promotion decisions are made by a select group of higher ranking managers applying criteria that allegedly disfavors women.

a. desisting from engaging in each of the unlawful practices, policies, customs, and usages set forth in this complaint;

b. adopting non-discriminatory and objective promotion standards;

c. creating a transparent and non-discriminatory job posting and application process for Assistant General Manager and General Manager positions;

d. instituting an affirmative action policy to ensure that women receive the share of Assistant General Manager and General Manage positions they would have obtained were it not for Costco's discriminatory practices; and

e. creating a monitoring and reporting system to ensure that injunctive relief is fully implemented.

Third Am. Compl., Docket No. 537, at 24-25.