Schuman et al v. Energizer Holding, Inc. et al

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
UNITED STATES DISTRICT COURT
NORTHERN DISTRICT OF CALIFORNIA
DON COPELAND, et al.,
Plaintiffs,
v.
ENERGIZER HOLDINGS, INC., and
WALMART, INC.,
Defendants.
Case Nos. 5:23-CV-02087-PCP-SVK,
5:23-CV-02091-PCP-SVK,
5:23-CV-02093-PCP-SVK
ORDER DENYING MOTIONS FOR
CLASS CERTIFICATION WITHOUT
PREJUDICE AND DENYING MOTIONS
TO EXCLUDE PLAINTIFFS’ EXPERTS
PORTABLE POWER, INC., et al.,
Plaintiffs,
v.
ENERGIZER HOLDINGS, INC., and
WALMART, INC.,
Defendants.
KIMBERLY SCHUMAN et al.
Plaintiffs,
v.
ENERGIZER HOLDINGS, INC.; and
WALMART, INC.,
Defendants.
In these three putative class actions, consumers of Energizer batteries sue defendants
Energizer and Walmart for allegedly conspiring to keep the prices of Energizer batteries high and
eliminate price competition, particularly from online retailers. Plaintiffs in the proposed Portable
Power class now seek to certify a class of wholesale purchasers who bought batteries directly from
Energizer from September 1, 2019, to December 31, 2023, as well as a California subclass. No.
5:23-cv-02091. Plaintiffs in the proposed Schuman class seek to certify a class of consumers who
bought Energizer batteries directly from brick-and-mortar Walmart stores from September 1,
2019, to November 21, 2025. No. 5:23-cv-02093. And plaintiffs in the proposed Copeland class
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 1 of 32

2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
seek to certify a class of consumers who bought affected batteries indirectly from September 1,
2019, through November 21, 2025. No. 5:23-cv-2087. Defendants move to exclude plaintiffs’
experts Dr. Hal Singer and Dr. Kevin Caves. For the following reasons, the Court denies the
plaintiffs’ motions for class certification without prejudice to a possible future motion supported
by different evidence and denies defendants’ motions to exclude plaintiffs’ experts.
BACKGROUND
In the late 2010s, Walmart was Energizer’s largest customer. See Class Certification
Report of Hal J. Singer, Ph.D., Exh. 2, Dkt. No. 133-4 ¶ 14 n.12 (Nov. 21, 2025) [hereinafter
Singer Report].1 But Walmart faced the threat of competition from online retailers such as
Amazon who could offer end consumers lower prices on Energizer products relative to brick-and-
mortar retailers like Walmart. Id. ¶ 13–15. In late 2017 and early 2018, Walmart pressured
Energizer to monitor online retailers and prevent them from undercutting Walmart on prices for
Energizer’s headlamp, a type of flashlight one can wear on one’s head. Id. ¶ 15.
In late 2018, Energizer created a “Pricing Preservation Project,” which it later renamed to
“Project Hermes.” Id. ¶ 16; Exh. 33, Dkt. No. 133-35. Project Hermes sought to keep Energizer’s
retail prices high. Id. According to internal Energizer emails, in April 2019, Energizer learned that
Walmart’s “frustration is extremely high regarding their pricing position vs. [Amazon]” for
headlamps pricing. No. 5:23-cv-02087, Exh. 3, Dkt. No. 136-5. Walmart complained to Energizer
that third-party retailers continued to undercut Walmart on headlamp prices, culminating in
Walmart instituting a July 2019 “rollback,” or mandated retail price cut. See Exh. 3, Dkt. No. 133-
5. Energizer compensated Walmart for the loss it would sustain under the rollback by paying the
retailer the difference. Id. Walmart promised that if Energizer could keep other retailers’ prices of
Energizer’s headlamps above Walmart’s desired price for 13 weeks, then Energizer could stop
subsidizing Walmart’s rollback. See id.
In 2019, Walmart and Energizer allegedly expanded their efforts to Energizer’s batteries,
1 Unless otherwise indicated, docket number references are to the entry number in No. 5:23-cv-
02091, Portable Power, Inc. v. Energizer Holdings, Inc.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 2 of 32

3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
specifically Energizer’s AA, AAA, C, D, and 9V batteries. Energizer reported internally that
Walmart “threatened to put us on rollback at holiday if we did not get a handle on our
eComm[erce] pricing,” which would entail Energizer paying Walmart until that point. No. 5:23-
cv-02087, Exh. 65, Dkt. No. 136-67, at 311; No. 5:23-cv-02087, Exh. 31, Dkt. No. 136-33, at 362.
Energizer called its effort to ensure that online retailers’ prices matched Walmart’s the “Walmart
Challenge”. No. 5:23-cv-02087, Dkt. No. 136-33, at 362. Energizer calculated how much it would
have to pay Walmart for Walmart to match online retailers’ prices and estimated that doing so
would cost around $55 million per year. Id. at 363.
Energizer took several steps to discipline retailers who were selling below Walmart, an
endeavor internally referred to as “Project Atlas.” For example, Energizer cut off Amazon from
selling several of Energizer’s products, Exh. 39, Dkt. No. 133-41, at 00576088, and within
months, Energizer celebrated that some battery prices had increased on Amazon, Direct Purchaser
Plaintiffs’ Notice of Motion and Motion for Class Certification, Dkt. No. 133-1, at 11 [hereinafter
DPP Motion]; Exh. 53, Dkt. No. 133-55, at 00191640–41. Energizer also monitored and flagged
Walmart competitors who undercut Walmart on price and tracked “actions being taken at each
retailer to address pricing,” Exh. 41, Dkt. No. 133-43, at 00325524–31, including by raising
wholesale prices to those competitors, see, e.g., Exh. 42, Dkt. No. 133-44; Exh. 43, Dkt. No. 133-
45. Energizer also established an authorized reseller policy that required smaller retailers to apply
for the ability to sell Energizer products online. Exh. 33, Dkt. No. 133-35; No. 23-cv-2087, Exh.
45, Dkt. No. 136-47, at 00001745. Energizer started policing orders and considered price increases
or refusing to ship products to them. See, e.g., Exh. 46, Dkt. No. 133-48; Exh. 47, Dkt. No. 133-
49. Plaintiffs maintain that Energizer weighed the risk of retailers walking away from Energizer
against the benefit of satisfying Walmart. DPP Motion at 12.
After Energizer’s efforts in Project Atlas, Walmart appears to have declined to pursue a
rollback on Energizer batteries. Exh. 61, Dkt. 133-63. By November 2020, Energizer employees
understood Project Atlas, and its alleged ability to maintain higher prices, as “part of our steady
state business,” with an Energizer employee explaining, “‘Atlas Project’ is a term that signals that
there is an ‘end’ date to this initiative and that is not how we want people thinking about it. Atlas
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 3 of 32

4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
did a great job setting the foundation …, but that project team has been dissolved and we have
instead in permanent resources to carry it forward.” No. 23-cv-02087, Exh. 60, Dkt. No. 136-62, at
00076319.
LEGAL STANDARDS
Federal Rule of Civil Procedure 23 governs class certification. A party seeking class
certification must first satisfy the requirements of Rule 23(a) by demonstrating that
(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.
In addition to satisfying the requirements of Rule 23(a), the proposed class “must satisfy at least
one of the three requirements listed in Rule 23(b).” Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338,
345 (2011). Plaintiffs seek certification under Rule 23(b)(3), which requires that common
questions of law or fact “predominate over any questions affecting only individual members, and
that a class action [be] superior to other available methods for fairly and efficiently adjudicating
the controversy.” Fed. R. Civ. P. 23(b)(3).
The party seeking class certification “must affirmatively demonstrate [its] compliance with
Rule 23.” White v. Symetra Assigned Benefits Serv. Co., 104 F.4th 1182, 1192 (9th Cir. 2024)
(citation modified). Thus, plaintiffs “must actually prove—not simply plead—that their proposed
class satisfies each requirement of Rule 23, including ... the predominance requirement of Rule
23(b)(3).” Halliburton Co. v. Erica P. John Fund, Inc., 573 U.S. 258, 275 (2014). Plaintiffs must
show entitlement to class certification by a preponderance of the evidence. Olean Wholesale
Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 664–65 (9th Cir. 2022) (en banc).
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 4 of 32

5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
ANALYSIS
I. Portable Power, 23-cv-2091, and Schuman, 23-cv-2093
Plaintiffs in Portable Power, No. 5:23-cv-02091, and Schuman, No. 5:23-cv-02093,
together seek to certify four classes: a wholesale nationwide class, a wholesale California subclass,
a Walmart nationwide class, and a Walmart California subclass. DPP Motion at 1–2. The proposed
wholesale nationwide class is defined as:
All persons and entities other than the “Excluded Entities” that
purchased Energizer Class Batteries directly from Defendant
Energizer Holdings, Inc. (“Energizer”) in the United States at any
time from September 1, 2019, to December 31, 2023 (the “Wholesale
Class Period”). “Class Batteries” are defined as disposable AA, AAA,
C, D, 9V, or CR2032 batteries. “Energizer Class Batteries” are
defined as Energizer-manufactured Class Batteries. The Wholesale
Plaintiffs seek certification of the Wholesale Nationwide Class to
pursue damages claims under Section 1 of the Sherman Act, 15
U.S.C. § 1, against Energizer and Defendant Wal-Mart, Inc.
(“Walmart”). The Wholesale Nationwide Class and California
Subclass (referred to as the “Wholesale Class”) seeks damages for
purchases made between September 1, 2019, and the start of trial,
March 27, 2028 (“Wholesale Damages Period”).
The wholesale California subclass is defined as:
All Wholesale Class members that made their qualifying purchases as
billed to, shipped to, or paid for from California. The Wholesale
Plaintiffs seek certification of the Wholesale California Subclass to
pursue damages claims under Section 1 of the Sherman Act and the
Cartwright Act, Cal. Bus. & Prof. Code §§ 16720, et seq., against
Energizer and Walmart.
The Walmart nationwide class is defined as:
All persons and entities that purchased Energizer or Duracell Class
Batteries directly from a Walmart brick-and-mortar store in the
United States at any time from September 1, 2019, to November 21,
2025 (the Walmart “Class Period” and “Damages Period”) without
using Walmart Pay or Mobile Scan & Go. “Duracell Class Batteries”
are defined as Duracell-manufactured Class Batteries. The Walmart
Plaintiffs seek certification of the Walmart Nationwide Class to
pursue damages and injunctive relief claims under the Sherman Act
against Energizer. For the avoidance of doubt, the Walmart
Nationwide Class and California Subclass (referred to as the
“Walmart Class”) is not seeking damages for purchases made online,
via Walmart+, for in-store pick up, or from Sam’s Club.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 5 of 32

6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
The Walmart California subclass is defined as:
All Walmart Class members that made their qualifying purchases
from a Walmart brick-and-mortar store in California. The Walmart
Plaintiffs seek certification of the Walmart California Subclass to
pursue damages and injunctive relief claims under the Sherman Act
and the Cartwright Act against Energizer.
A. The Portable Power and Schuman plaintiffs satisfy 23(a).
1. Numerosity
The Portable Power and Schuman plaintiffs have demonstrated that the proposed classes
are “so numerous that joinder of all members is impracticable.” Fed. R. Civ. P. 23(a)(1). Rule 23’s
numerosity requirement is generally satisfied when the proposed class contains 40 or more
members. See J.L. v. Cissna¸ 2019 WL 415579, at *8 (N.D. Cal. Feb. 1, 2019); Corley v. Google,
Inc., 316 F.R.D. 277, 290 (N.D. Cal. 2016). Plaintiffs argue, and defendants do not contest, that
the proposed classes contain thousands of members. Singer Report ¶ 114 (“There are 2,303
Wholesale Class members that made at least one purchase of Energizer Class Batteries during the
Class Period.”), ¶ 120 (estimating over 378 million transactions for Walmart products).
2. Commonality
“Commonality mandates there be a common question of law or fact among the class
members where the same evidence will suffice for each member to make a prima facie showing or
the issue is susceptible to generalized, classwide proof.” Small v. Allianz Life Ins. Co. of N. Am.,
122 F.4th 1182, 1198 (9th Cir. 2024) (citation modified). In other words, commonality exists
where “the evidence establishes that a common question is capable of class-wide resolution.”
Noohi v. Johnson & Johnson Consumer Inc., 146 F.4th 854, 863 (9th Cir. 2025) (quoting Olean,
31 F.4th at 666–67). “To satisfy commonality, even a single common question is enough.” Small,
122 F.4th at 1198 (citation modified) (quoting Wal-Mart, 564 U.S. at 359).
Here, every proposed class member’s claim turns on several common questions including
whether Walmart and Energizer entered into an agreement, whether their actions violated the
antitrust laws, and whether their actions resulted in higher prices to plaintiffs. The answers to
those factual and legal questions “will resolve [] issue[s] that [are] central to the validity of each
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 6 of 32

7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
one of the claims in one stroke.” Wal-Mart, 564 U.S. at 350. Defendants do not contest
commonality. Thus, the commonality requirement is also satisfied.
3. Typicality
Defendants do not dispute that the Schuman plaintiffs’ claims are typical of those of the
class. Defendants argue, however, that plaintiffs Portable Power and Micropower are not typical or
adequate class representatives for the Portable Power wholesale class and subclass because they
“were small, grey-market sellers who repeatedly violated Energizer’s policies to obtain
competitive advantages over the large, professionally managed retailers like Home Depot and
Target who make up the vast majority of the proposed Wholesale Class.” Opposition at 46.
Defendants allege that Portable Power illegally obtained internal Energizer communications and
documents and have filed counterclaims against Portable Power and certain individual defendants
on that basis. See Portable Power, Inc. v. Energizer Holdings, Inc., 2026 WL 2227359 (N.D. Cal.
Aug. 3, 2026). Defendants also maintain that Portable Power and Micropower violated Energizer’s
policies by buying Energizer batteries in bulk and re-packaging those batteries in their own
packaging. Opposition at 47 (citing Portable Power Dep., Exh. 16, Dkt. No. 185-19, at 111:5–21).
And defendants argue that Micropower continued to sell Energizer batteries online despite being
prohibited from doing so by Energizer because Micropower failed to enroll in Energizer’s online
authorization program. Opposition at 47. Thus, defendants argue that both Portable Power and
Micropower will have to address specific issues and potential defenses that distinguish them from
other retailers. Opposition at 47.
“Although [defendants] name[] both typicality and adequacy in [their] argument, [their]
contention that Plaintiffs will be preoccupied with unique defenses falls within our typicality
caselaw.” DZ Reserve v. Meta Platforms, Inc., 96 F.4th 1223, 1238 (9th Cir. 2024). A class
plaintiffs’ claims should be “reasonably co-extensive with those of absent class members; they
need not be substantially identical.” Id. (quoting Hanlon v. Chrysler Corp., 150 F.3d 1011, 1020
(9th Cir. 1998), overruled on other grounds by Wal-Mart, 564 U.S. at 338). But a “named plaintiff
is not typical if there is a danger that absent class members will suffer if their representative is
preoccupied with defenses unique to it.” Id. (quoting Hanon v. Dataproducts Corp., 976 F.2d 497,
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 7 of 32

8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
508 (9th Cir. 1992)) (cleaned up).
Portable Power’s and Micropower’s claims are “reasonably co-extensive with those of
absent class members” despite the risk that defendants will face issues that are unique to them and
not “substantially identical” to those of other class members. Id. Portable Power and Micropower
are accused of violating Energizer’s rules for resellers. But plaintiffs’ theory is that Energizer and
Walmart “agreed on a scheme to stifle price competition and inflate wholesale and retail battery
prices.” DPP Motion at 3. To the extent that their purported policy violations actually involved
pretextual attempts by Energizer to discipline maverick resellers outcompeting Walmart, they
continue to be typical plaintiffs under their theory of the case. Therefore, Portable Power and
Micropower have shown that they are typical plaintiffs under Rule 23(a).
4. Adequacy
Rule 23(a)(4) requires that the plaintiffs “will fairly and adequately protect the interests” of
the proposed class. “The adequacy inquiry is addressed by answering two questions: (1) do the
named plaintiffs and their counsel have any conflicts of interest with other class members and
(2) will the named plaintiffs and their counsel prosecute the action vigorously on behalf of the
class? If either answer is no, the representative is inadequate.” Small, 122 F.4th at 1202 (citation
modified). To satisfy these criteria, a “class representative must be part of the class and possess the
same interest and suffer the same injury as the class members.” Amchem Prods. v. Windsor, 521
U.S. 591, 625 (1997) (citation modified). Defendants do not challenge class certification on
adequacy grounds, aside from challenging the typicality of Portable Power and Micropower as
lead plaintiffs. See supra.
As to the proposed class representative, the moving plaintiffs state that they do not have
any conflicts of interest with absent class members. DPP Motion at 15. Plaintiffs bought the class
product batteries from defendants and share an interest with absent class members in establishing
defendants’ liability. As to the proposed class counsel, the Court is satisfied that they will
“prosecute the action vigorously on behalf of the class” given their experience in class action
litigation. Small, 122 F.4th at 1202.
Consequently, plaintiffs satisfy the adequacy requirement of Rule 23(a)(4).
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 8 of 32

9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
B. The Portable Power and Schuman plaintiffs have not yet satisfied 23(b)(3).
Rule 23(b)(3) requires that the common questions of law or fact identified above
“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.”
Rule 23(b)(3)’s predominance inquiry “tests whether proposed classes are sufficiently cohesive to
warrant adjudication by representation.” Amchem Prods., 521 U.S. at 623. “The commonality and
predominance inquiries overlap.” White, 104 F.4th at 1191. After separating questions common to
the legal claims of the class from those that “present individualized issues,” courts “analyze
whether the common questions predominate over the individual questions.” DZ Reserve, 96 F.4th
at 1233. In essence, this inquiry “asks whether the common, aggregation-enabling, issues in the
case are more prevalent or important than the non-common, aggregation-defeating, individual
issues.” Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442, 453 (2016) (citation modified).
“[T]o prove there is a common question of law or fact that relates to a central issue in an
antitrust class action, plaintiffs must establish that ‘essential elements of the cause of action,’ such
as the existence of an antitrust violation or antitrust impact, are capable of being established
through a common body of evidence, applicable to the whole class.” Olean, 31 F.4th at 666
(citation modified). At the class certification stage of proceedings, plaintiffs must show
predominance by providing “evidence … capable of showing that the [plaintiffs] suffered antitrust
impact on a class-wide basis.” Id. at 681. In Olean, the en banc Ninth Circuit affirmed the district
court’s certification of an antitrust class of direct purchasers of tuna because plaintiffs’ economic
model was capable of showing that the tuna suppliers’ conspiracy raised prices for the entire class.
Id. at 671, 677.
The Portable Power and Schuman plaintiffs rely on the report of economist Dr. Hal Singer
to argue that they can establish antitrust injury or impact “through a common body of evidence,
applicable to the whole class.” Id. at 666. Defendants contend that Dr. Singer’s report is
methodologically unsound and therefore does not satisfy plaintiffs’ burden at class certification.
1. Difference-in-Differences (DiD) Regression
The Ninth Circuit has explained that econometric tools and methods like the ones used by
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 9 of 32

10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
plaintiffs’ experts can demonstrate classwide antitrust injuries. In Olean, the plaintiffs’ expert
“first performed a pricing correlation test” which showed that prices of the defendant tuna
suppliers’ products “moved up or down together regardless of product or customer type.” Id. at
671. The shared movement of tuna product prices “thus supported the proposition that the Tuna
Suppliers’ collusion had a common, supra-competitive impact on their prices,” which in turn
supported an inference that the suppliers’ collusion “would affect direct purchasers on a class-
wide basis.” Id.
The Olean plaintiffs’ expert then conducted a multiple regression analysis. Multiple
regression analysis is a “statistical tool for understanding the relationship between two or more
variables,” which represent events or attributes that can have two or more values. Daniel L.
Rubenfeld & David Card, Reference Guide on Multiple Regression and Advanced Statistical
Models, in National Academies of Sciences, Engineering, and Medicine & Federal Judicial Center,
1 REFERENCE MANUAL ON SCIENTIFIC EVIDENCE 577, 676 (4th ed. 2025) [hereinafter FJC].
Multiple regression analyses estimate relationships between “a variable to explain [] called the
dependent variable” and “additional explanatory variables that are thought to produce or be
associated with changes in the dependent variable.” Id. at 580. Multiple regression analyses, and
the models they use, can calculate correlations between explanatory and dependent variables while
controlling for other explanatory variables. See id. at 586. “If a regression model uses appropriate
independent or explanatory variables, it can test and isolate the extent to which the actual prices
paid by plaintiffs are higher because of a defendant’s collusive behavior.” Olean, 31 F.4th at 671
(cleaned up). Thus, valid multiple regression models can provide evidence of antitrust injury on a
classwide basis. See id.
One type of multiple regression analysis relies upon “difference-in-differences.”
Difference-in-differences (DiD) research designs study two groups over time, one of which
receives the treatment or intervention of interest while the other does not. The group that receives
the treatment is the “treatment group” while the group that does not is the “comparison group.”
The analyst studies the difference in outcomes between the treatment and comparison groups. See
FJC at 624. The analyst should choose two groups that are similar enough such that studying them
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 10 of 32

11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
can control for all or nearly all other relevant explanatory variables besides the treatment effect the
analyst seeks to isolate and study. DiD studies provide an estimate of the causal effect of the
treatment by subtracting the outcome for the comparison group from the outcome for the treatment
group. See id. Because the two groups are identical (or similar enough) except for the treatment,
the difference in outcomes can be attributed to the effect of the treatment itself and not to other
explanatory variables. The premise that the two groups are identical or similar enough to each
other is foundational for the DiD research design and called the “parallel trends assumption.” Id. at
625. The parallel trends assumption requires that, if the treatment group did not receive the
treatment, the outcomes for the two groups would have moved in parallel over time. See id. In a
valid DiD study, the two groups’ outcomes should generally move in parallel before the treatment.
See id.
In seeking to certify their Rule 23(b)(3) wholesale and Walmart classes and subclasses,
plaintiffs rely on an expert report from Dr. Hal J. Singer. See Singer Report, Dkt. No. 133-4.
Singer conducts a DiD regression model taking September 2019 as the date demarcating the
before-treatment period from the after-treatment period. Singer Report ¶ 97. Singer sets the
dependent variable as class battery prices and the treatment as the presence of the scheme. Id.
¶ 98.
Singer sets as the comparison group “Energizer headlamps and lighting products.” Id. ¶ 99.
He does so for two reasons. First, Singer reasons that Energizer did not have market power in
headlamps and lighting products, so Energizer and Walmart’s scheme would not lead to antitrust
injury in those markets. Without market power, the theory goes, Energizer and Walmart’s scheme
would be ineffective to affect the headlamps and lighting products markets’ prices. As Singer
notes, “Energizer had sought to increase the price of headlamps in response to Walmart’s
demands, yet had failed to do so.” Id. Second, Singer asserts that “headlamps and lighting product
prices are influenced by similar supply-and-demand factors as battery prices, thereby implying
that headlamps and lighting product prices serve as a reasonable competitive benchmark to what
battery prices would have looked like absent the Scheme.” Id. Exhibit 2 of Singer’s report graphs
indexed prices of batteries and lighting products to the wholesale and Walmart classes. Id. at 45.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 11 of 32

12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Singer estimates the overcharges to Wholesale Class members using transaction-level sales
data of the prices paid by all wholesale and retail direct purchasers from January 2013 to
December 2023, amounting to over nine million transactions and over $10 billion in revenue. Id.
¶ 101. The resulting regression estimates that defendants’ scheme “resulted in an overcharge of
approximately 11.1 percent to the Wholesale Class.” Id. ¶ 102.
Singer also estimated the effect of the scheme on the Walmart class. Using transaction-
level sales data for Walmart retail customers from 2013 to 2023, Singer used the same DiD model
as he did for the wholesale class, included Duracell class batteries because they are part of the
Walmart class, and found that defendant’s scheme resulted in a 13.9 percent overcharge to the
Walmart class. Id. ¶¶ 104, 105.
At the second step of his analysis, Singer sought to establish that the generalized
overcharge affected all or almost all class members. Id. ¶ 109. To do so, Singer used an “in-sample
prediction” and a “price structure.” Id. ¶ 110. The in-sample prediction method uses the first step’s
DiD estimates of a generalized overcharge to derive a but-for price that direct purchasers would
have paid without the scheme. Id. ¶¶ 110–112. Singer then compared the price that the class
members would have paid with what class members actually paid and then concluded that any
overpayments, either by transaction or on net, establish antitrust injury attributable to the scheme.
Id. ¶¶ 111, 114. Singer’s in-sample prediction found that 99.7 percent of the wholesale class were
subject to at least one transaction affected by defendants’ scheme and 95.5 percent harmed on net.
Id. ¶ 114. He also found that 99.7 percent of the wholesale California subclass suffered harm at
least once and 93.6 percent suffered harm on net. Id. ¶ 116. As to the Walmart class, Singer found
that 98.3 percent of transactions during the class period were affected by higher prices stemming
from the scheme as well as 98.4 percent of the Walmart California subclass. Id. ¶¶ 120, 121.
Singer argues that the few wholesale and Walmart class members for which the in-sample
method does not find harm nonetheless likely were harmed because his method would not account
for idiosyncratic sources of lower prices such as customer-specific discounts, the “in-sample
prediction is inherently conservative” because it attributes downward price shocks not accounted
for by the model to the scheme, and because Energizer made use of a common price structure that
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 12 of 32

13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
further suggests that all class members would have been affected. Id. ¶¶ 114, 120. A pricing
structure, according to Singer, refers to how a change in prices “would have propagated to prices
paid by all or nearly all Class Members.” Id. ¶ 123. Singer attempts to establish the existence of a
pricing structure to show that the generalized overcharge affected all members of the proposed
classes. Energizer documents establish that it puts brands into different groups and seeks to
enforce differences in pricing between those brands. Id. ¶¶ 124–130. Singer also uses multiple
variable regressions to conclude that “prices tend to rise and fall together when general factors
change.” Id. ¶ 131.
Though defendants contest both Singer’s general overcharge model, his first step, as well
as how he showed that the overcharged affected all class members, his second step, the Court
needs only address the first step. To establish a persuasive DiD identification strategy, the choice
of lighting products must satisfy the parallel trends assumption—i.e., that the average outcomes
for the treated population (class batteries) and comparison population (lighting products) would
have evolved in parallel without the treatment effect. See Jonathan Roth, Pedro H.C. Sant’Anna,
Alyssa Bilinski & John Poe, What’s Trending in Difference-in-Differences? A Synthesis of the
Recent Econometrics Literature, 235 J. OF ECONOMETRICS 2218, 2221 (2023). Both sides explain
that the whether the assumption is met depends upon whether one can plausibly infer that the
treatment and comparison groups are sufficiently similar as to justify using the comparison group
to isolate the causal effect of the treatment. See Class Certification Reply Declaration of Hal J.
Singer, Ph.D., Exh. 93, Dkt. No. 209-3, ¶¶ 40–42; Rebuttal Expert Report of Catherine Tucker,
Ph.D., Exh. B, Dkt. No. 182-5 ¶ 113 (Mar. 6, 2026) [hereinafter Tucker Report]. Put differently,
the parallel trends assumption is plausible only if Singer can persuasively establish that, if
Energizer and Walmart had not engaged in its allegedly anticompetitive scheme, the mean
outcomes in Energizer batteries prices would have changed in parallel to the mean outcomes in
lighting products. See FJC at 626.
Singer and defendants’ economist Dr. Catherine Tucker use several approaches to consider
whether Singer’s comparison and treatment groups satisfy the parallel trends assumption here,
including visual inspections, statistical tests, and the broader market context. Each consideration
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 13 of 32

14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
undermines the plausibility of the parallel trends assumption as to the choice of comparison and
treatment groups and thus fatally undermines the reliability of the overcharge models at issue.
First, a visual inspection of pre-treatment trends for both batteries and lighting products
does not support the parallel trends assumption. Singer provides an exhibit that he claims shows
that “headlamps and lighting products exhibit similar pricing behavior” to batteries “prior to
September 2019 (which is marked with a dashed, vertical line).” Singer Report ¶ 99.
The problem, as defendants’ economist Tucker points out, is that a visual inspection of
pre-treatment outcomes does not appear to show that batteries and lighting products prices moved
in parallel before September 2019. See Tucker Report ¶ 123. Tucker observes that, unlike
batteries, lighting products “exhibit[] downward spikes and, from around 2016, a substantial
decrease in level.” Tucker Report ¶ 124.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 14 of 32

15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
In Singer’s reply declaration attached to plaintiffs’ joint reply in support of class
certification, Singer argues that when one looks at pre-treatment prices for the Walmart class,
lighting products and batteries moved in parallel, especially in the three-and-a-half-year period
leading up to September 2019:
Class Certification Reply Declaration of Hal J. Singer, Ph.D., Exh. 93, Dkt. 209-3, at 35 (May 15,
2026) [hereinafter Singer Reply Declaration]. But this inspection again does not support the
inference that batteries and lighting products moved in parallel during the shortened pre-treatment
period. At multiple points, prices for batteries and lighting products moved in opposite directions,
or one stayed constant while the other changed. For example, in March 2016, the price index for
batteries fell while the index for lighting products increased sharply. And in the months leading up
to September 2019, batteries prices stayed roughly steady or increased while lighting products
prices fell sharply. Singer responds to Tucker’s “focus[ing] on the high degree of volatility in
lighting product prices that begins around the first quarter of 2016” by arguing that the “presence
of volatility does not invalidate my model, so long as that volatility would be expected to continue
during the treatment period in the but-for world.” Singer Reply Declaration ¶ 52. Singer argues
that such an expectation—that volatility in lighting product prices would continue during the
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 15 of 32

16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
treatment period—does indeed “continue into the treatment period.” Id. Singer’s argument is
unpersuasive, however, because the volatility of lighting products prices matters insofar as those
prices are volatile relative to the prices of class batteries. If both batteries and lighting products
exhibited similar volatility, then that fact would suggest that the two are plausible groups to use in
a DiD identification strategy. But the volatility of lighting products prices relative to battery prices
undermines the validity of Singer’s models because lighting products prices appear to be more
volatile than batteries prices and thus suggests that the prices of the two product groups do not
move together but in fact are affected by different, unobserved factors. See Reply Declaration of
Catherine Tucker, Ph.D., Exh. T, Dkt. No. 204-4 ¶¶ 81–85 [hereinafter Tucker Reply Decl.].
To be sure, the fact that the prices do not move together according to a visual inspection
does not necessarily imply that lighting products and batteries cannot be used in a DiD
identification strategy. The disjunction merely fails to support that they can. The Court must
therefore consider other evidence of parallel trends. Tucker provides another such check by
plotting the difference between price indices for class batteries and lighting products during the
pre-treatment period. See Tucker Report ¶¶ 123–125. If the parallel trends assumption held, then
one would expect that prices for batteries and lighting products would move together and that the
difference between the two sets of prices would be roughly constant. But Tucker’s Figure 26
shows that the difference between the prices changed somewhat over time. See Tucker Report
¶ 124. Similarly, her Figure 63 shows varying differences between batteries and lighting products
prices for the wholesale class in both of Singer’s DiD models. See Tucker Reply Decl. ¶ 84.
Tucker does not explain how to assess whether the changes in differences between the price
indices are significant, but the y-axis scale, focusing on tenths of the price indices, is essentially
the same scale Singer used for his visual inspections.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 16 of 32

17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Tucker Report ¶ 124.
Tucker Reply Decl. ¶ 84.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 17 of 32

18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
The plotted difference between batteries’ and lighting products’ price indices thus does not
support the parallel trends assumption. The Court acknowledges that the parallel trends
assumption need not be established statistically. Declaration of Hal J. Singer, Ph.D. in Support of
Opposition to Motion to Exclude, Exh. 76, Dkt. No. 202-3 ¶¶ 10, 11 (Apr. 15, 2026). But even if
not strictly required, such statistical tests help the Court understand whether the parallel trends
assumption is plausible, which Singer agrees is the test for his model’s validity. Id. ¶ 7.
Even where a visual and statistical inspection of pre-treatment trends fails to support a
parallel trends assumption, additional context could establish that the two groups are similar
enough for a DiD approach. But the different contexts for batteries and lighting products further
undermine the plausibility of comparing batteries to lighting products because they suggest that
unobserved variables are not addressed by Singer’s approach. First, “disposable batteries account
for approximately 15 to 20 percent, and in many instances less, of the cost of Energizer Lighting
Products, which renders battery-cost fluctuations unlikely to be a major driver of Lighting-Product
cost or price fluctuations.” Tucker Report ¶ 98. Second, defendants report that the “primary driver
and largest component of the cost to manufacture lighting products is the cost to build the body of
the lighting product.” Declaration of Jeffrey Stoll, Dkt. No. 182-23 ¶ 3. Energizer’s Director of
Lighting, Global Marketing, explained that in May 2020, Energizer outsourced the “manufacturing
and packaging of lighting products” and that, between May 2020 and December 31, 2025, “the
third-party manufacturers did not implement significant price increases to Energizer for the
packaged lighting products that they assembled.” Id. ¶ 4. Energizer’s production of lighting
products may therefore be meaningfully different from its production of batteries. Unobserved
variables that affect the cost of lighting products would not be shared across both lighting products
and batteries. And though cost is merely one of multiple variables that affect Energizer’s ultimate
pricing decisions, if unobserved variables relevant to cost differ between lighting products and
batteries, the strength of Singer’s DiD approach is undermined.
Finally, different post-treatment period events undermine the choice of lighting products as
a comparison group. Analysts looking to satisfy the parallel trends assumption typically look at the
pre-treatment period to confirm that the two groups behave mostly similarly. But, as mentioned
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 18 of 32

19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
previously, the theory of the DiD research design assumes that, “in the absence of the intervention
the mean potential outcomes for the two groups would [have] move[d] in parallel.” See FJC at
625. The theory thus implies that in the post-treatment period the two groups would have moved
in parallel except for the effect caused by the treatment studied. But if the comparison group
experienced a meaningfully different event in addition to the treatment studied, that could further
detract from the validity of the DiD model because it would provide an alternative explanation for
the different outcomes in the groups.
Here, Tucker uses the measure of costs from the reports of Singer and the indirect
purchaser plaintiffs’ expert economic Dr. Kevin Caves to show that the difference in the cost to
produce batteries and lighting products decreased during the post-treatment period, suggesting that
the cost function for lighting products changed relative to the cost function for batteries. Tucker
Report ¶¶ 152–155. Tucker argues that Singer cannot add an explanatory variable to account for
costs “because prices do not shift with every movement of costs.” Id. ¶ 157. Instead, as Tucker
notes, most Energizer customers had contractual limitations that would have delayed Energizer’s
ability to pass through cost increases for at least one to three months from. Tucker Report ¶ 73;
Tucker Reply Decl. ¶¶ 59–72. The result, Tucker points out, are implausible statistical results. For
example, Singer’s new DiD model finds that “when Energizer’s costs double, holding other
factors constant, its prices decrease by approximately 7.0 percent,” which would be very unlikely
given “the upward pressure that cost increases exert on prices.” Tucker Reply Decl. ¶ 71.
The differences between batteries and lighting products also help explain why Singer’s
strategy of using Energizer’s market power in batteries but not lighting products as the treatment is
unpersuasive. As Tucker points out, the possession or lack of market power can affect its bearer’s
response to exogenous cost increases. As a result, unobserved variables, such as those that might
have been relevant during the COVID-19 economic crisis, may have differing effects on class
batteries and lighting products in both the pre-September 2019 and post-September 2019 period.
Tucker Report ¶ 106. Given that Energizer had market power over batteries and did not have
market power over lighting products, that variation may have affected the dependent variable, the
price index, but may also have affected the two groups of products differently in either the pre- or
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 19 of 32

20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
post-treatment period in ways that the model may not have captured.2
In short, Singer has not adequately justified his choice to use lighting products as the
comparison group for his DiD analysis. The evidence before the Court fails to justify any
assumption that the prices for lighting products and batteries moved in parallel prior to the
implementation of the scheme challenged here or that the two products are so similar that their
prices would be expected to have moved in parallel.
2. Before-During Regression
Singer’s report also provides a “before-during” model, which he calls a “robustness test.”
Such a model could at least theoretically provide evidence of predominance in an antitrust class
action, as it did in Olean. See 31 F.4th at 671. A before-during design analyzes outcomes for a
subject that receives a treatment both before and after that treatment and incorporates explanatory
variables so as to produce an estimate of the effect that the treatment had on the dependent
variable. See FJC at 622–23. Singer acknowledges some of the differential strengths of the before-
during approach as compared to DiD. See Singer Report ¶ 106 & 48–49 n.200. In particular, the
“[v]alidity of a pre/post comparison requires that the average outcome for the group that received
the treatment would have remained constant in the absence of treatment.” FJC at 623.
Defendants note many reasons to suspect that the average outcome for batteries would not
have remained constant in the absence of the scheme. For example, zinc, steel, and lithium
carbonate are raw materials that make up a significant portion of the cost of Energizer’s alkaline
batteries. Tucker notes that, after December 2020, input prices for these raw materials increased
substantially. Tucker Report ¶ 55. Lithium carbonate prices increased 440 percent from 2021 to
2022. Tucker Report at 41. Figure 9 from the Tucker Report shows the sharp increase in prices for
zinc, steel, and copper, which occurred at the end of 2020. Copper prices are relevant because they
are an important input into the production of manganese dioxide, which is itself a key input into
the production of alkaline batteries. Id. ¶ 56.
2 A more persuasive analysis might have chosen a comparison group in a market where Energizer
had market power but did not exercise it in an allegedly anticompetitive manner.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 20 of 32

21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Tucker also points out that Energizer announced a price increase in batteries in June 2021
that took effect in October 2021. Tucker Report ¶ 58. An internal May 2021 presentation at
Energizer reported that most of the increase in price was due to the increase in raw material costs
as well as labor and transportation costs. Tucker Report ¶ 60. Tucker tracked additional post-2019
price increases and noted they occur at approximately the same time and are proportional to the
increase in the cost of raw materials. See id. ¶¶ 61–64.
While these changes may raise questions about the validity of Singer’s before-during
regression model, the model does attempt to control for variables including costs and
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 21 of 32

22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
transportation issues such as fuel or supply chain pressure. Singer Report ¶¶ 151–152.3 Ultimately,
defendants’ criticisms of Singer’s before-during regression model may go to the weight that
should properly be accorded to Singer’s opinion, rather than to its admissibility or capability of
establishing antitrust impact on a classwide basis. See Olean, 31 F.4th at 679 (“A lack of
persuasiveness is not fatal at certification.”).
The problem for plaintiffs, however, is that Singer did not offer his model as a standalone
basis for satisfying plaintiffs’ burden to prove that antitrust impact can be established across the
class. Singer initially described his before-during model as a “robustness test,” relegating it to the
fourth and final appendix in his class certification report and describing it as “an alternative
modeling approach to verify that [his] DiD results are robust to changes in specification.” Singer
Report ¶ 150. As defendants note, plaintiffs’ “motion [for class certification] never mentioned Dr.
Singer’s before-during model.” Defendant Walmart Inc.’s Sur-Reply, Dkt. No. 225, at 6. It was
only in plaintiffs’ joint reply brief that they argued that Singer’s before-during model offers an
independent means of showing that the defendants’ scheme increased battery prices across the
class. See Plaintiffs’ Joint Reply in Support of Class Certification, Dkt. No. 209-1, at 2, 6. Given
that plaintiffs did not offer the before-during model as a standalone method of establishing
classwide antitrust impact in their initial motion for class certification, and given that Singer
himself did not expressly offer it as sufficient evidence of that impact, plaintiffs cannot rely on it
for that purpose with respect to the pending motion for class certification, and the Court must do
the same.4
* * *
In short, Dr. Singer’s difference-in-differences regression analysis is the only standalone
3 Singer acknowledges that a valid before-during approach should control for all potentially
confounding, relevant explanatory variables. Singer Report at 48 n.200; see also Tucker Report
¶ 157 (noting that given the lag in costs affecting prices, “an approach which simply tried to
control for costs in a statistical specification would be flawed because prices do not shift with
every movement of costs”).
4 Because the Court declines to consider whether Singer’s before-during model provides a
standalone basis for certification, it need not conclusively determine on the existing record
whether defendants’ challenges to that model render it admissible or instead merely involves its
persuasiveness.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 22 of 32

23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
evidence that the Portable Power and Schuman plaintiffs have properly offered to show that they
can establish antitrust impact on a classwide impact. For the reasons discussed, Singer’s analysis
fails to do so. Accordingly, the Portable Power and Schuman plaintiffs have not satisfied their
burden, and their motion for class certification is therefore denied. Because it is possible that the
Portable Power and Schuman plaintiffs could satisfy their burden on a different evidentiary
record, the denial of plaintiffs’ motion is without prejudice to a future motion supported by
different evidence.5
III. Copeland, No. 23-cv-02087
The plaintiffs in the Copeland action, No. 5:23-cv-02087, move to certify a Rule 23(b)(3)
class of indirect purchasers of affected batteries, for appointment as class representatives, and for
appointment of Cohen Milstein Sellers & Toll PLLC as class counsel. No. 5:23-cv-02087,
Copeland Plaintiffs’ Notice of Motion for Class Certification, Dkt. No. 136-1 [hereinafter IPP
Motion].6 Plaintiffs propose a class consisting of:
All persons and entities in Repealer Jurisdictions who indirectly
purchased Energizer-, Rayovac-, or Eveready-branded disposable
household batteries for their own use and not for resale, at any time
during the period from September 1, 2019 through November 21,
2025 (the “Class Period”). Household batteries are defined as
batteries of cell size AA, AAA, C, D, or 9V.
“Repealer Jurisdictions” are states that have passed state laws permitting indirect purchasers to
pursue state antitrust claims notwithstanding the Supreme Court’s conclusion that such purchasers
lack federal claims. See Illinois Brick Co. v. Illinois, 431 U.S. 720 (1977).7
5 Energizer filed a supplemental opposition to the Schuman and Copeland plaintiffs’ motion for
class certification, arguing that both sets of plaintiffs and class members would be subject to
arbitration agreements and class waivers. See No. 5:23-cv-02093, Dkt. No. 127. Though the Court
notes that neither Walmart’s attached terms of service nor those of third-party retailers appear to
require arbitration of the disputes here, see, e.g., No. 5:23-cv-02093, Dkt. 126-7, at 2, 41, because
the Court denies class certification on independent grounds, the Court need not address the
grounds advanced in Energizer’s supplemental opposition at this time.
6 The Copeland plaintiffs are Don Copeland, Joseph Murray, David Freifeld, Patrick Whitney,
Phillip Hague, Denise Foris, Roxann Doriott, Bruce Mims, Lori Ably, Timothy Brown, Peter
Costas, and Mike Ballard.
7 The Copeland plaintiffs clarify that, for purposes of their class certification motion, the
“Repealer Jurisdictions” are Arizona, California, Connecticut, District of Columbia, Florida,
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 23 of 32

24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
A. The Copeland plaintiffs satisfy 23(a).
1. Numerosity
The proposed class is numerous because it would consist of “millions of retail battery
purchasers.” IPP Motion at 13–14.
2. Commonality
The proposed class shares common questions of law and fact, such as whether defendants
engaged in conduct forbidden by state antitrust law.
3. Typicality
The named plaintiffs are “consumers who purchased Energizer household batteries for
personal use at non-Walmart retailers during the class period.” IPP Motion at 14. They allege that
defendants acted together to inflate the prices that they and absent class members paid for class
batteries. Their claims are therefore typical.
4. Adequacy
Plaintiffs do not appear to have conflicts of interest with absent class members and
promise to “prosecute the action vigorously.” IPP Motion at 15.
B. The Copeland plaintiffs have not yet satisfied 23(b)(3).
Plaintiffs’ expert Dr. Kevin W. Caves uses record evidence and an econometric analysis to
show that defendants’ conduct led to the indirect purchaser plaintiffs suffering an antitrust injury.
Class Certification Report of Kevin W. Caves, Ph.D., No. 5:23-cv-02087, Dkt. No. 136-3 (Nov.
21, 2025) [hereinafter Caves Report]. Similarly to Singer, Caves uses a DiD econometric analysis
to show that the defendants’ conduct “resulted in anticompetitive effects in the form of inflated
wholesale and retail prices for disposable household batteries.” Caves Report ¶ 139. Like Singer,
Caves then goes on to present evidence “that those inflated prices caused class-wide injury that
impacted all or nearly all members of the Class.” Id.
As with Singer’s expert report, Caves has not shown that his DiD approach is a valid
Hawai‘i, Idaho, Illinois, Iowa, Kansas, Maine, Maryland, Massachusetts, Michigan, Minnesota,
Mississippi, Missouri, Montana, Nebraska, Nevada, New Hampshire, New Mexico, New York,
North Carolina, North Dakota, Oregon, Puerto Rico, Rhode Island, South Carolina, South Dakota,
Tennessee, Utah, Vermont, West Virginia, and Wisconsin.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 24 of 32

25
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
method for estimating the effect that defendants’ conduct had on class battery prices.
To be certain, the record evidence that Caves marshals does indeed tend to show that
Walmart and Energizer sought to curtail price competition over battery prices and were successful
in enacting list price increases. See Caves Report ¶¶ 115–122. Caves cites emails from 2021 and
2022 wherein Energizer appeared to conclude that retailers were passing on Energizer’s price
increases on to consumers. Caves Report ¶ 124 & 62 n.255. But, in order to show predominance,
plaintiffs must provide a method “capable of answering the question whether there was antitrust
impact” as a result of the allegedly anticompetitive conduct. See Olean, 31 F.4th at 676. Though
record evidence can certainly be informative, pointing to list price increases and other measures to
maintain price levels is insufficient on its own to satisfy Rule 23(b)(3)’s predominance
requirement here because it does not provide a method to identify whether, for example, list prices
increased as a result of illegal conduct rather than increases in input costs. See Tucker Report
¶¶ 85–86.
Caves provides a DiD regression model that, like Singer’s, uses Energizer lighting
products (headlamps and lights) as a comparison group. Caves Report ¶¶ 129, 132. Caves justifies
the choice of lighting products as a comparison group “because (1) batteries are one of the primary
components of headlamps and lighting products, and (2) similar factors impact demand for
headlamps, lighting products, and batteries….” Caves Report ¶ 133. Caves’ DiD regression model
finds that the defendants’ allegedly anticompetitive conduct increased prices 11.91 percent relative
to lighting products. Caves Report ¶ 137.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 25 of 32

26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Caves’s DiD model exhibits many of the same shortcomings as Singer’s model, and for the
reasons already noted, Caves’s model also cannot establish antitrust impact on classwide basis.
The Court notes that a visual analysis of nominal and real price indices for batteries and lighting
products does not lend itself to the parallel trends assumption, especially because the price indices
for the two groups appear to intersect at multiple times. Caves Report at 71–72.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 26 of 32

27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
 See id. at 72.
Like Singer, Caves uses a before-during regression model to “test the robustness of the
result of my DID model.” Caves Report, Appendix A ¶ 1. Caves’s regression equation provides a
term that “represents additional factors that influence the supply or demand for disposable
household batteries.” Caves Report, Appendix A ¶ 4. As with Singer’s DiD model, there is a
serious question as to whether Caves’s two-period before-during model adequately accounts for
the exogenous raw materials shocks experienced by Energizer in late 2020 and 2021. But in any
event, like the Portable Power and Schuman plaintiffs, the Copeland indirect-purchaser plaintiffs
failed to adequately invoke Caves’s before-during model as an independent method of establishing
antitrust impact for purposes of the pending motion. Though their initial motion for class
certification did mention Caves’s before-during regression, it did so only in a footnote. See IPP
Motion at 23 n.104. Given both Caves’s own description of the model as a “test” of his DiD
regression analysis and plaintiffs’ relegation of the model to a footnote, defendants did not receive
proper notice of plaintiffs’ intent to rely upon such a standalone model in support of the pending
motion, and the Court therefore will not do so.
* * *
Like the Portable Power and Schuman plaintiffs, for the purposes of their pending motion
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 27 of 32

28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
the Copeland indirect-purchaser plaintiffs rely upon their expert’s DiD regression analysis to
establish that antitrust impact can be established on a classwide basis. That analysis, however, fails
to do so. The Copeland plaintiffs have therefore failed to satisfy their burden to establish Rule
23(b)(3)’s predominance requirement with competent evidence. Accordingly, the Copeland
plaintiffs’ motion for class certification is also denied. Because it is possible that the Copeland
plaintiffs could satisfy their burden on a different evidentiary record, the denial of plaintiffs’
motion is without prejudice to a future motion supported by different evidence.
IV. Article III Standing
Defendants separately argue that certification is improper because plaintiffs’ models do not
establish injury to every single class member and those uninjured absent class members lack
Article III standing. Opposition at 19. But for the purposes of class certification, “only named
class members—but not unnamed class members—must demonstrate evidence of standing.” See
Healy v. Milliman, Inc., 164 F.4th 701, 706 (9th Cir. 2026). The Court therefore need not consider
the Article III standing of absent class members at this time.
V. Defendants’ Motion to Exclude Plaintiffs’ Experts
Defendants move to exclude plaintiffs’ experts under Federal Rule of Evidence 702 for
using invalid control groups in their DiD regression model. Defendants argue that Singer’s use of
an invalid control or comparison group “render[s] Dr. Singer’s DiD analysis unreliable.” Motion
to Exclude Plaintiffs’ Expert Hal Singer, Dkt. No. 182-2, at 9 [hereinafter Singer Motion to
Exclude]. Defendants also argue that Singer’s before-during regression is flawed in its design and
produces unrealistic results. Singer Motion to Exclude at 22. And defendants argue that other
choices Caves made, such as using average and not transaction-level data, are not helpful to a
factfinder. No. 5:23-cv-02087, Motion to Exclude Expert Testimony and Opinions of Dr. Kevin
Caves, Dkt. No. 157-2.
Federal Rule of Evidence 702 provides that an expert witness can provide opinion
testimony if the proponent shows the Court that “it is more likely than not that: (a) the expert’s
scientific, technical, or other specialized knowledge will help the trier of fact to understand the
evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c)
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 28 of 32

29
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
the testimony is the product of reliable principles and methods; and (d) the expert’s opinion
reflects a reliable application of the principles and methods to the facts of the case.” The Supreme
Court’s decision in Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993), requires the
Court to act as a gatekeeper for expert testimony.
Whether the Court must apply “a ‘full’ or ‘limited’ Daubert analysis” to plaintiffs’ experts’
declarations “may depend on the timing of the class certification decision.” Lytle v. Nutramax
Labs., Inc., 114 F.4th 1011, 1031 (9th Cir. 2024). If discovery has not closed and the “expert’s
model has yet to be fully developed, a district court is limited at class certification to making a
predictive judgment about how likely it is the expert’s analysis will eventually bear fruit.” Id. A
district court may certify a class based on an unexecuted damages model that may eventually turn
out to be erroneous. See Noohi, 146 F.4th at 863–64.
Here, discovery has not yet closed. Though the Court has rejected Singer’s and Caves’s
existing DiD regression models, that holding is limited to the conclusion that those models do not
provide “evidence establish[ing] that” the “common question” of antitrust injury “is capable of
class-wide resolution.” Olean, 31 F.4th at 666–67. The Court does not, however, conclude that the
experts’ analyses could never bear fruit. Plaintiffs’ economists have chosen generally accepted
methods of estimating antitrust injury that could, using different specifications, produce evidence
satisfying Rule 23(b)(3). Therefore, the Court denies the defendants’ motions to exclude Drs.
Singer and Caves without prejudice.
VI. Motions to Seal
Pursuant to the Court’s order setting omnibus sealing procedure in connection with
plaintiffs’ motions for class certification, Dkt. No. 128, the parties have filed a joint motion to file
under seal material filed in connection with the motions for class certification and motions to
exclude, Dkt. No. 217.8
The parties agree that the compelling-reasons standard applies here because the filings are
“more than tangentially related to the merits of a case.” Ctr. for Auto Safety v. Chrysler Grp., LLC,
8 Again, all docket references, unless otherwise indicated, are to 5:23-cv-02091.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 29 of 32

30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
809 F.3d 1092, 1100 (9th Cir. 2016). Under the compelling-reasons standard, a party that seeks to
seal a court record “must articulate compelling reasons supported by specific factual findings …
that outweigh the general history of access and the public policies favoring disclosure, such as the
‘public interest in understanding the judicial process.’” Kamakana v. City & County of Honolulu,
447 F.3d 1172, 1178–79 (9th Cir. 2006) (cleaned up). The compelling-reasons standard “sharply
tips the balance in favor of production when a document … becomes part of a judicial record.” Id.
at 1180. “Under this stringent standard,” the Court must “conscientiously balance the competing
interests of the public and the party who seeks to keep certain judicial records secret.” Ctr. for
Auto Safety, 809 F.3d at 1096–97. “Examples” of reasons to seal a court record include when the
record “might be used to ‘gratify private spite or promote public scandal,’ to circulate ‘libelous’
statements, or ‘as sources of business information that might harm a litigant’s competitive
standing.” Id. at 1097 (quoting Nixon v. Warner Commc’ns, Inc., 435 U.S. 589, 598–99 (1978)).
The parties seek to seal nearly 200 documents in connection with plaintiffs’ motions for
class certification. The Court denies the motion to seal as to the briefing on plaintiffs’ motions for
class certification and defendants’ motions to exclude.9
The parties justify their motions to seal portions of the relevant briefing on the motions for
class certification and motions to exclude on various grounds related to potential harms to the
defendants’ businesses. The parties cite “Confidential and Strategic Business Information,”
“Transaction Sales and Purchase Data,” “Negotiation and Internal Business Strategy,” and
“Confidential Terms of Agreements” as satisfying the compelling-reasons standard to seal. The
parties argue that the information may be trade secrets or could allow competitors to undercut their
pricing. Though “sources of business information that might harm a litigant’s competitive
standing” can justify sealing under the compelling-reasons standard, Nixon, 435 U.S. at 598, such
reasons do not justify sealing the briefing here. As the Court has established, the compelling-
reasons standard requires compelling reasons to seal court records, not merely good cause. The
9 In the docket for No. 5:23-cv-2087, these documents are Dkt. Nos. 190-17, 191-13, 192-22, 193-
21, 193-25, 193-37, 193-39, 194-12, 194-14, 194-16, and 207.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 30 of 32

31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
Ninth Circuit recognizes a “strong presumption in favor of access to court records.” Foltz v. State
Farm Mut. Auto. Ins. Co., 331 F.3d 1122, 1135 (9th Cir. 2003). The reasons that the parties
present for sealing portions of the briefing do not overcome the “strong presumption in favor of
access to court records.” Id. Much of the information does not concern currently sensitive business
information but merely details and figures that do not necessarily establish a risk of competitive
harm. The information at issue almost entirely involves events occurring multiple years before the
filing of this order, such that the information would be of limited use in the present to competitors
or negotiating counterparties.
To the extent that the sealed material is discussed in this order, the motion to seal is denied.
The parties’ motion to seal the remaining exhibits submitted in support of and in opposition to
class certification is granted.10 The interim administrative motions to file under seal are denied as
moot given this order’s denial of the motions to seal the relevant briefing.
Energizer has also moved to consider whether plaintiffs Portable Power and Micropower
Battery’s materials should be sealed, Dkt. No. 147, in connection with Energizer’s amended
answer and counterclaims, Dkt No. 146. Because the answer and counterclaims asserted are
central to the merits of Energizer’s case, the compelling-reasons standard applies. See In re Meta
Pixel Tax Filing Cases, 2025 WL 1872974, at *2 (N.D. Cal. May 9, 2025). Because the
compelling-reasons standard applies, the Court denies the motion to consider whether another
party’s materials should be sealed, except to the extent that the materials disclose what appears to
be individuals’ phone numbers. Thus, the motion to consider whether another party’s materials
should be sealed is denied except as to pages 25:16, 25:22, 26:10, 26:15, 27:5, 28:2, 28:6, 28:10.
The Court also denies as moot Dkt. No. 194.
CONCLUSION
For the reasons stated above, the Court denies the motions for class certification and denies
the motions to exclude plaintiffs’ experts Singer and Caves. The parties’ sealing motions are
denied in part and granted in part. Defendants’ motion for leave to file supplemental authority is
10 This includes the motion at No. 5:23-cv-02087, Dkt. No. 211.
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 31 of 32

32
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
United States District Court
Northern District of California
granted. Dkt. No. 240.
The parties also recently stipulated to extend the stay of discovery until December 2026.
Dkt. No. 242. Because the Court has now ruled on plaintiffs’ class certification motions, a
continued stay of discovery may be unnecessary. The Court therefore declines to enter the
proposed stipulation. Instead, the Court sets a further status conference for Thursday, October 22,
2026, at 2:00 p.m. in Courtroom 8 of this Court’s San José courthouse. By no later than Thursday,
October 15, 2026, the parties in all three cases shall provide the Court with a status conference
statement setting forth their positions as to any changes to the existing case schedule that should
be made to account for the Court’s order on class certification.
IT IS SO ORDERED.
Dated: September 28, 2026
P. Casey Pitts
United States District Judge
Case 5:23-cv-02093-PCP Document 175 Filed 09/28/26 Page 32 of 32