D&G, Inc. v. Supervalu, Inc. et al

Filing 138

REDACTED MEMORANDUM OPINION AND ORDER previously signed by Judge Ann D. Montgomery on 7/16/2012. Associated Cases: 0:09-md-02090-ADM-AJB, 0:09-cv-00983-ADM-AJB, 0:09-cv-02940-ADM-AJB(GS)

UNITED STATES DISTRICT COURT
DISTRICT OF MINNESOTA

In re Wholesale Grocery Products
Antitrust Litigation

MEMORANDUM OPINION
AND ORDER
Court File No. 09-MD-2090 ADM/AJB

This Order Relates to All Actions

Richard B. Drubel, Esq., Boies, Schiller & Flexner LLP, Hanover, NH, and W. Joseph Bruckner,
Esq., and Elizabeth R. Odette, Esq., Lockridge, Grindal, Nauen, PLLP, Minneapolis, MN, on
behalf of Plaintiffs.
Stephen P. Safranski, Esq., and Heather M. McElroy, Esq., Robins, Kaplan, Miller & Ciresi,
LLP, Minneapolis, MN, on behalf of Defendant SuperValu, Inc.
Charles A. Loughlin, Esq., and David S. Shotlander, Esq., Baker Botts LLP, Washington, DC;
and Todd A. Wind, Esq., Fredrikson & Byron, PA, Minneapolis, MN, on behalf of Defendant
C&S Wholesale Grocers, Inc.
______________________________________________________________________________
I. INTRODUCTION
On May 22, 2012, the undersigned United States District Judge heard oral argument on
Plaintiffs’ Motion for Class Certification [Docket No. 202] and Defendants’ Motion to Strike
Report and Testimony of Dr. Jeffrey Leitzinger [Docket No. 212] (the “Daubert Motion”).1 This
matter consists of two2 putative class actions against Defendants SuperValu, Inc. (“SuperValu”)

1

Oral argument was also heard on Plaintiffs’ Motion to Add Lockridge Grindal Nauen
P.L.L.P. as Plaintiffs’ Co-Lead Counsel [Docket No. 253], Defendants’ Motion to Compel or in
the Alternative to Strike [Docket No. 260], Plaintiffs’ Motion to Compel 30(b)(6) Testimony
Regarding Defendant C&S’s Distribution Center Profit and Loss Documents [Docket No. 265],
Plaintiffs’ Motion to Compel Production of Documents Responsive to Plaintiffs’ Fifth Set of
Document Requests [Docket No. 267], Plaintiffs’ Motion to Compel Testimony Concerning
Defendants’ Transaction Data [Docket No. 269]. The Court ruled on each of these motions at
oral argument and they will not be addressed in this Order.
2

At one time this matter included four lawsuits: D&G, Inc., et al. v. SuperValu, Inc., et
al., Civil No. 09-983; DeLuca’s Corp. v. C&S Wholesale Grocers, et al., Civil No. 09-2940;
Charles W. Prather Company, Inc. v. C&S Wholesale Grocers, Inc., et al., Civil No. 09-2932;

and C&S Wholesale Grocers, Inc. (“C&S”) alleging violation of federal antitrust laws. The
actions were originally brought in the U.S. District Court for the District of New Hampshire and
the U.S. District Court for the District of Minnesota, but were consolidated for pretrial
proceedings as multidistrict litigation in the District of Minnesota. Transfer Order [Docket No.
1]. For the reasons set forth below, Plaintiffs’ Motion for Class Certification is denied and
Defendants’ Daubert Motion is denied.
II. BACKGROUND
Plaintiffs are retail grocers who allege Defendants SuperValu and C&S, two of the
largest grocery product wholesalers (measured by sales volume) in the United States, conspired
by way of an Asset Exchange Agreement (the “AEA”) to allocate territory and customers. 2d
Consolidated Am. Class Action Compl. [Docket No. 99] ¶¶ 1–3, 11–12, 31–44. Plaintiffs allege
the conspiracy to allocate territory and customers has injured their business and property, and
they seek treble damages and attorney’s fees under the Clayton Act, 15 U.S.C. § 15.
A.

The Wholesale Grocery Industry
Grocery wholesalers act as “middlemen” in the grocery industry–wholesalers purchase

products from manufacturers, store those products at distribution centers, and then resell
products to retail grocers. Id. ¶ 16. Retail grocers use wholesalers because wholesalers’ large
purchase volume from manufacturers leads to discounts on products, and wholesalers manage

and Blue Goose Super Market, Inc. v. C&S Wholesale Grocers, Inc., et al., Civil No. 09-3191.
By Order dated October 19, 2010 [Docket No. 84], Charles W. Prather Company, Inc.’s claims
were dismissed and by Order dated July 5, 2011 [Docket No. 141], Blue Goose Super Market,
Inc’s claims were dismissed. Therefore, this matter presently includes two lawsuits, Civil Nos.
09-983 and 09-2940. The remaining Plaintiffs are D&G, Inc. d/b/a Gary’s Foods (“Gary’s
Foods”) and DeLuca’s Market Corporation (“DeLuca’s”).
2

the coordination and logistics of sorting, ordering, pricing, and delivery of products. Expert
Report of Jeffrey J. Leitzinger, Ph.D. in Supp. of Pls.’ Mot. for Class Certification with Attachs.
& Exs. [Docket No. 204] (“Leitzinger Report”) ¶ 19; Expert Report of Dr. John H. Johnson, IV
Related to Class Certification with Exs. 1-253 [Docket No. 217] (“Johnson Report”) ¶ 15.
Wholesalers sometimes offer related services such as shelf-planning or accounting services.
Johnson Report ¶ 15.
Defendants (SuperValu and C&S, collectively “Defendants”) are “full-line” wholesalers.
2d Consolidated Am. Class Action Compl. ¶ 16. Full-line grocery wholesalers serve their clients
through their networks of distribution centers and trucking fleets. Id. ¶ 17. Full-line wholesalers
are distinguished from “partial-line” and other wholesalers by their wide product offerings,
typically offering tens to hundreds of thousands of items. Johnson Report ¶ 20. To effectively
provide the number of products for a full-line, and to manage the logistics of providing
wholesale services for those products, full-line wholesalers must invest in large distribution
centers costing $25 to $100 million. Johnson Report ¶ 89, at 47 n.177 (citations omitted).
However, full-line wholesalers are not the sole source of products for retail grocers. For
example, some larger chain retailers are able to operate their own distribution centers and “selfdistribute.” Id. ¶ 19; Johnson Report ¶ 16. “Partial-line” wholesalers also exist that provide
retailers with a limited set of product categories, often meat or produce, and require smaller
distribution facilities. Leitzinger Report ¶ 25; Johnson Report ¶ 16. Additionally, retailers may
pool their resources to own and operate “co-operative” wholesalers. Johnson Report ¶ 16.
SuperValu’s principal place of business is in Eden Prairie, Minnesota, and it operates
distribution centers in twenty-one states. 2d Consolidated Am. Class Action Compl. ¶ 11.

3

C&S’s principal place of business is in Keene, New Hampshire, and it operates distribution
centers in twelve states. Id. ¶ 12. Throughout the 1990s and early 2000s, SuperValu began
aggressively expanding into New England,3 a geographic market dominated by C&S. Id. ¶¶
21–23. By the early 2000s, C&S was also expanding, and began looking to enter the Midwest
market.4 Id. ¶ 26. Specifically in 2003, C&S planned to enter markets in Wisconsin and Ohio
through acquisition of three distribution centers from Fleming Companies, Inc. (“Fleming”),
which had filed for bankruptcy. Id. ¶¶ 26–27.
B.

The Asset Exchange Agreement
In September 2003, SuperValu and C&S entered into the AEA. Id. ¶ 35. Through the

AEA, C&S acquired all of SuperValu’s distribution centers, and the clients they served, in New
England, and SuperValu acquired all of C&S’s distribution centers, and the clients they served,
in the Midwest.5 Id. The AEA included a non-compete provision wherein each Defendant
agreed not to supply former customers served from a distribution center exchanged in the
agreement for two years, and each Defendant agreed not to solicit those customers for a period of
five years. Id. ¶ 36. Within eight months of the closing of the AEA, all the exchanged
distribution centers were closed. Id. ¶ 38.

3

For purposes of this Order, New England consists of Connecticut, Maine,
Massachusetts, New Hampshire, Rhode Island, and Vermont. See Pls.’ Mot. for Class
Certification [Docket No. 202] 2 (defining New England Class with reference to those states).
4

For purposes of this Order, the Midwest consists of Illinois, Indiana, Iowa, Michigan,
Minnesota, Ohio, and Wisconsin. See Pls.’ Mot. for Class Certification 1.
5

C&S operated distribution centers in the Midwest, one in Maple Heights, Ohio and the
others in Michigan; however, those centers were dedicated to individual customers and did not
serve the public. Leitzinger Report ¶ 106. The customers of those stores are excluded from the
proposed class definitions, and are not relevant to the analysis here.
4

Plaintiffs allege the purpose of the AEA was to allocate customers and territory in
violation of federal antitrust laws. Id. ¶¶ 1-3, 33, 76-83. Plaintiffs aver the elimination of
competition between SuperValu and C&S in regional markets allowed each to charge supracompetitive prices to retail customers. Id. ¶ 39.
C.

Wholesale Grocery Products and Services Pricing
C&S does not charge a single set price for products. Rather, the ultimate price paid by

each customer varies. The starting point for price is the cost to C&S to buy a product from the
manufacturer. See Johnson Report, Ex. 9 (“Saia Dep.”) 38:13-16 (describing use of cost to C&S
as starting point for dry grocery products). C&S then charges a premium to its customers,
referred to as an “upcharge.” See id. 38:17-19 (describing process for dry groceries). For many
products, the upcharge includes the difference between the price paid by C&S to the
manufacturer and the wholesale price the manufacturer suggests C&S charge its clients, referred
to as the “list price.” Aff. of W. Joseph Bruckner in Supp. of Pls.’ Sur-Reply in Supp. of Their
Mot. for Class Certification [Docket No. 325] (“Bruckner Class Certification Sur-Reply Aff.”),
Ex. A. (“Riley Dep.”) 105:16-21. List prices are established by manufacturers on a national
basis. Decl. of Stephen P. Safranski in Supp. of Defs.’ Mot. to Strike Report & Testimony of Dr.
Jeffrey Leitzinger & in Opp. to Pls.’ Mot. for Class Certification [Docket No. 279] (“Safranski
Decl.”), Ex. I 73:3-7. For other products, such as produce and meat products, C&S buys the
product at variable instant prices and gives retailers a single quote based on those acquisition
prices. Safranski Decl. Ex. I 73:25–77:23. C&S then assesses an upcharge for a client using
either a per unit basis (e.g. ten cents per case or per pallet) or as a percentage of an order (e.g. ten
percent of the total list prices of the products). Johnson Report, Ex. 10 at 30:9-20. These

5

upcharges are individually negotiated with customers. Saia Dep. 38:19–39:1. Factoring into the
individual negotiations are size of a customer’s orders, frequency of orders, local market
conditions, and distance from a customer’s store to C&S’s distribution center. Saia Dep. 43:1725, 45:8-10. Finally, C&S negotiates certain price concessions,6 such as rebates, with customers.
Saia Dep. 46:4–47:2.
SuperValu also does not charge any single price for its products and services.
SuperValu, however, incorporates a more formulaic approach for some customers known as
Activity Based Sell (“ABS”) pricing. SuperValu began implementing ABS pricing in Denver in
1995, and expanded it into the Midwest and other regions from 1998-2000. Johnson Report, Ex.
137 at SV00130800.008. ABS pricing was intended to be profit-neutral for SuperValu. Johnson
Report, Ex. 133 at SV00129374.057. However, Plaintiffs allege SuperValu could not expand
ABS pricing into New England due to competition from C&S. 2d Consolidated Am. Class
Action Compl. ¶ 25. SuperValu avers ABS pricing allows it to keep prices lower than its
competitors because ABS creates incentives for customers to act cost-effectively. Johnson
Report, Ex. 137 at SV00130800.002. ABS pricing sells products to retail grocers at each
product’s actual average acquisition cost to SuperValu plus certain fees. Johnson Report, Ex.

6

Dr. Leitzinger disputes the extent to which the price concessions claimed by Defendants
are actually reductions in price versus investments by Defendants in their customers’ businesses.
Leitzinger Decl. ¶ 212. Regardless, there is no dispute that at least some actual price
concessions were given. Id. ¶ 213. However, the nuance of negotiated price concessions need
not be further considered because, for the purposes of this Order, the Court will accept Plaintiffs’
contention that impact may be proven notwithstanding individually negotiated price concessions
if Plaintiff can prove negotiations took place from an inflated starting point for all class members
with common evidence. See In re Potash Antitrust Litig., 159 F.R.D. 682, 696 (D. Minn. 1995)
(certifying class where list prices increased dramatically and uniformly following price-fixing
conspiracy and those list prices served as a common starting point for negotiation).
6

133 at SV00129374.015–22. SuperValu assesses two types of fees–operating fees, which
account for about eighty-five percent of sales, and service fees, which account for about fifteen
percent of sales. Id. at SV00129374.025.
Operating fees are determined based on summation of nine factors, referred to as “fee
drivers:” (1) receiving, (2) storage, (3) finance, (4) selecting, (5) loading, (6) customer minimum,
(7) item minimum, (8) overhead, and (9) percent of purchase. Id. at SV00123374.026. The
additional fee for each factor is typically sufficient to both recoup the cost of the factor and
generate a profit margin for SuperValu. Id. Many of the fee drivers are calculated using a set
formula; generally, fees are determined using average costs (including overhead) multiplied by a
customer’s number of cases bought, or time stored, or whatever relevant metrics drive costs. See
id. at SV00129374.027–45 (describing calculations for each fee factor). Because the factors are
influenced by conditions particular to each distribution center, such as local wage rates for
selecting items or average cost of storing an item at a particular warehouse, the precise fee
amounts vary by distribution center. Id. at SV00129374.046. However, the finance fee and item
minimum fees for product groups are the same across all distribution centers. Id. at
SV00129374.031, SV00129374.042.
Service fees, on the other hand, are fixed fees that apply when orders with certain
minimum order requirements are placed more than twenty-one days in advance. Id. at
SV00129374.050. Service fees are used only for promotional products as these are the products
most often ordered in advance. Id. at SV00129374.047–49. Service fees are calculated using a
set per-unit charge plus a percentage charge. Id. at SV00129374.051. Service fees vary across
product categories and across services. Id. Service fees do not vary between distribution

7

centers. See id. at SV00129374.051 (listing “current” fees without reference to distribution
centers).
Another aspect of ABS pricing is the Category Management Allowance Program
(“CMAP”). CMAP passes manufacturers’ promotional funds to retailers based on the
performance of certain conditions, such as keeping specific retail prices on an item, putting an
item on display, or enrolling in certain programs. Johnson Report, Ex. 2 (“Chew Dep.”)
168:4–169:10. Some of SuperValu’s customers use many CMAP promotions, others use some,
still others use none. Id. 169:11-17.
Not every customer of SuperValu buys products under the ABS pricing regime. Some
larger retailers such as [redacted] negotiate contracts with prices determined without ABS.
Johnson Report Exs. [redacted]. Retailers may also negotiate credits to offset ABS fees.
[redacted]. Moreover, ABS pricing does not apply to every product sold by SuperValu. ABS
pricing is used only for the dry grocery, general merchandise, dairy, and frozen product
categories. Johnson Report ¶ 33.
D.

Expert Reports
Plaintiffs and Defendants have hired economists to provide expert testimony in this case.

The experts have opined about the extent to which common evidence could be used to determine
if pricing was supra-competitive during the “class period”–the time period covered by the class
action allegations here. Plaintiffs have retained Dr. Jeffrey J. Leitzinger (“Dr. Leitzinger”), who
has provided an expert report.
In his report, Dr. Leitzinger discusses two methods he avers could show class-wide
impact of supra-competitive prices charged by C&S. The first has been dubbed the “contrary

8

hypothesis test” by the parties. In the contrary hypothesis test, Dr. Leitzinger sought to test the
hypothesis contrary to class-wide impact, i.e. test whether impact was localized to a subset of
class members. Leitzinger Report ¶ 116. Dr. Leitzinger decided the subset of class members
most likely to be impacted if impact were local rather than class-wide were retailers closest to
SuperValu distribution centers in New England. Id. Using C&S transaction data, Dr. Leitzinger
compared upcharges for C&S customers that were within fifty miles of a SuperValu distribution
center in New England prior to the AEA with those that were not. Id. ¶ 117. Rather than finding
upcharges lower for those customers closer to SuperValu distribution centers (the result that
would be consistent with the initial “contrary hypothesis”), Dr. Leitzinger found upcharges were
marginally higher for those customers. Id. Dr. Leitzinger repeated his analysis using ten-mile
increments from a zero to a fifty mile radius and found no difference. Id. ¶ 118. Dr. Leitzinger
concluded, therefore, that SuperValu’s competitive impact on the New England market was not
localized prior to the AEA and the impact of the reduced competition after the AEA would be
market-wide. Id.
The second method Dr. Leitzinger avers can show class-wide impact has been dubbed the
“variance test” by the parties. In conducting the variance test, Dr. Leitzinger assumed that were
competition from SuperValu in New England localized rather than market-wide prior to the
AEA, C&S would already be charging higher prices to customers out of range of SuperValu
distribution centers. Id. ¶ 119. Based on that assumption, Dr. Leitzinger hypothesized that the
variance of C&S upcharges would be reduced after the AEA as prices in areas formerly subject
to SuperValu’s localized affect rose to match those outside of SuperValu’s range. Id. Dr.
Leitzinger conducted a statistical analysis and found no evidence that the variance of C&S

9

upcharges declined after the AEA. Id.
The results of Dr. Leitzinger’s analyses are equally consistent with SuperValu having a
market-wide effect on wholesale grocery prices and SuperValu having no effect at all on
wholesale grocery market prices. Id. ¶ 120. Dr. Leitzinger disregards the latter possibility,
however, as implausible based on the size of SuperValu, the insistence of C&S for the noncompete provisions, and C&S’s identification of SuperValu prior to the AEA as its “largest” and
“closest” competitor. Id. ¶ 121.
Dr. Leitzinger avers he can use the “same kind” of common evidence to show class-wide
impact for the Midwest class. Leitzinger Report ¶ 123. With respect to the contrary hypothesis
test, however, Dr. Leitzinger could not use the same method. In New England, Dr. Leitzinger
compared upcharges for C&S customers within fifty miles of a SuperValu distribution center
with those further away. Id. ¶ 117. When he did the same calculation for SuperValu customers
nearby Fleming distribution centers (which would have been owned by C&S but for the AEA),
he found ABS fees were higher for SuperValu customers both closest and farthest away from
Fleming distribution centers. Bruckner Aff. [Docket No. 245], Ex. A., Leitzinger Decl. ¶ 224.
Deeming this result “nonsensical,” Dr. Leitzinger attributed this result to the fact that under
SuperValu’s ABS pricing, fees varied between distribution centers and customers were assigned
particular distribution centers based on geography. Id.; Johnson Report, Ex. 25 (“Leitzinger
Dep.”) 228:8-229:13. Therefore, Dr. Leitzinger’s contrary hypothesis test for the Midwest uses
the distance of a SuperValu distribution center, rather than the distance of a SuperValu customer,
to a Fleming distribution center. Leitzinger Decl. ¶ 223; Leitzinger Dep. 230:6-9. Dr. Leitzinger
determined which SuperValu distribution centers were “within range” (using a 160 mile metric)

10

of a Fleming distribution center, and then compared ABS fees of those SuperValu distribution
centers with the ABS fees of the other SuperValu distribution centers in the Midwest. Leitzinger
Dep. 230:10-5.
In his expert report, Dr. Leitzinger also opines that (1) a product market for full-line
wholesale grocery products and related services can be proven with common evidence,
Letizinger Report ¶¶ 40-59, (2) relevant geographic markets can be proven with common
evidence, id. ¶¶ 60-80, (3) Defendants’ market power can be proven using common evidence, id.
¶¶ 81-93, and (4) Plaintiffs’ damages are reasonably ascertainable using common evidence, id.
¶¶ 135-159. Defendants have retained Dr. John H. Johnson, IV (“Dr. Johnson”), who provided
an expert report disputing much of Dr. Leitzinger’s analysis.
On October 31, 2011, Plaintiffs timely filed their Motion for Class Certification and Dr.
Leitzinger’s expert report in support of their motion. On January 31, 2012, Defendants timely
filed their memorandum in opposition to class certification and filed their Daubert Motion
seeking to strike Dr. Leitzinger’s report and exclude his testimony. On May 22, the Court heard
oral argument on both motions. In addition, direct examination and cross-examination was
conducted of both Dr. Leitzinger and Dr. Johnson.
III. DISCUSSION
A.

Daubert Motion
Rule 702 of the Federal Rules of Evidence governs the admission of expert testimony.

Under Rule 702, an expert may testify if (1) the expert’s scientific, technical, or other specialized
knowledge will help the trier of fact understand the evidence or determine a fact in issue, (2) the
testimony is based on sufficient facts or data, (3) the testimony is the product of reliable

11

principles and methods, and (4) the expert has reliably applied those principles and methods to
the facts of the case. Fed. R. Evid. 702. Therefore, district courts faced with a proffer of expert
scientific testimony must make a preliminary assessment of “whether the reasoning or
methodology underlying the testimony is scientifically valid” and “whether that reasoning or
methodology properly can be applied to the facts in issue.” Daubert v. Merrell Dow Pharms.,
Inc., 509 U.S. 579, 592–93 (1993).
Here, the Daubert analysis is adapted to fit the procedural posture. See In re Zurn Pex
Plumbing Prods. Liab. Litig., 644 F.3d 604, 611–14 (8th Cir. 2011) (holding that district courts
may properly apply a “focused” or “tailored” Daubert inquiry at class certification stage). “The
main purpose of Daubert exclusion is to protect juries from being swayed by dubious scientific
testimony.” Id. at 613. At the class certification stage, the Court, not a jury, is the decision
maker, and therefore a less stringent analysis is required. Id. Furthermore, class certification is a
“tentative” and “preliminary” determination. Id. The Daubert inquiry at this procedural stage,
therefore, only scrutinizes the reliability of expert testimony in light of the criteria for class
certification and the current state of the evidence. Id. at 614. Expert disputes at class
certification are resolved only to the extent necessary to determine the nature of the evidence that
would be sufficient, if the plaintiff’s general allegations were true, to make out a prima facie case
of class liability. Id. at 611 (quoting Blades v. Monsanto Co., 400 F.3d 562, 567 (8th Cir.
2010)). The Daubert inquiry at the class certification stage is to guard against certification of a
class based on an expert opinion so flawed that it is inadmissible as a matter of law. In re Visa
Check/Mastermoney Antitrust Litig., 192 F.R.D. 68, 76 (E.D.N.Y. 2000).
Given these principles, the criteria for Plaintiffs to prevail on their Motion for Class

12

Certification are to establish the prerequisites of Rule 23(a) of the Federal Rules of Civil
Procedure and establish that one of the three requirements of Rule 23(b) applies. Fed. R. Civ. P.
23. Here, Plaintiffs allege Rule 23(b)(3) applies, which requires that 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 adjudicatory methods. Fed. R. Civ. P. 23(b)(3). To
satisfy Rule 23(b)(3)’s predominance requirement where, as here, plaintiffs seek treble damages
under the Clayton Act as a class, plaintiffs must show both conspiracy and impact (the fact of
injury) can be proven on a systematic, class-wide basis. Blades, 400 F.3d at 569.
As will be discussed below, class certification in this case turns on whether common
impact can be shown through common evidence. The focus of a Daubert analysis, therefore, is
the reliability of Dr. Leitzinger’s proposed methods for showing common impact. In particular,
Defendants contend that Dr. Leitzinger’s contrary hypothesis and variance tests are unreliable,
and move for their exclusion. However, even accepting the reliability of the contrary hypothesis
and variance tests, Plaintiffs have not shown common impact can be proven through common
evidence. Therefore, the purpose of a Daubert inquiry at class certification articulated in Visa
Check/Mastermoney, preventing a class from being certified based on dubious expert evidence,
192 F.R.D. at 76, is not implicated here.
The U.S. Court of Appeals for the Eighth Circuit in Zurn Pex instructs that expert
disputes should be resolved only to the extent necessary to make a determination regarding the
nature of the evidence sufficient for class certification. 644 F.3d at 611. In this case Dr.
Leitzinger’s methods, even if unchallenged, would be insufficient and therefore it is not
necessary to resolve disputes regarding the reliability of his methods. Indeed, in Blades, the

13

district court did not resolve any expert disputes, but instead considered all evidence, only to
deny certification on the merits. 400 F.3d at 569–71. Similarly, here because the expert
evidence, even if accepted as urged by Plaintiffs, cannot demonstrate common impact, there is
no dispute that need be resolved and the Court declines to conduct a full Daubert analysis.
Rather, the Court will resolve any disputes between Drs. Johnson and Leitzinger as needed
without excluding Dr. Leitzinger’s testimony or reports in toto. Defendants’ Daubert motion is
denied.
B.

Motion for Class Certification
1. Standard of Review
District courts are “clothed with a good deal of discretion in determining the

appropriateness of a class action.” Wright v. Stone Container Corp., 524 F.2d 1058, 1061 (8th
Cir. 1975) (citations omitted). As noted above, to avail itself of class action treatment, a plaintiff
must meet the four threshold requirements of Rule 23(a) and must then establish that the class
fits within one of the three types of class actions listed in Rule 23(b). Avritt v. Reliastar Life Ins.
Co., 615 F.3d 1023, 1028 (8th Cir. 2010).
Plaintiffs argue their proposed classes are proper under Rule 23(b)(3). See generally
Pls.’ Mem. of Law in Supp. of Their Mot. for Class Certification [Docket No. 205] 11–29.
Under Rule 23(b)(3), a class action is appropriate if 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)(3). Whether a question of law or fact is
common or individual is determined by reference to the nature of the evidence that will suffice to

14

resolve the question. Blades, 400 F.3d at 566. “If, to make a prima facie showing on a given
question, the members of a proposed class will need to present evidence that varies from member
to member, then it is an individual question.” Id. “If the same evidence will suffice for each
member to make a prima facie showing, then it becomes a common question.” Id.
In this case, Plaintiffs allege Defendants allocated territory and customers in violation of
Section 1 of the Sherman Act, 15 U.S.C. § 1, and bring suit for damages under Section 4 of the
Clayton Act, 15 U.S.C. § 15. 2d Consolidated Am. Class Action Compl. ¶¶ 76-83. To prevail,
Plaintiffs must prove (1) Defendants violated federal antitrust laws, (2) Plaintiffs suffered some
resulting injury, and (3) a measure of damages. Blades, 400 F.3d at 566; In re Hydrogen
Peroxide Antitrust Litig., 552 F.3d 305, 311 (3d Cir. 2009); In re Visa Check/Mastermoney
Antitrust Litig., 280 F.3d 124, 136 (2d Cir. 2001) (Sotomayor, J.). For the proposed classes to
be certified, therefore, Plaintiffs must demonstrate that common evidence can establish each of
these elements. Visa Check/Mastermoney, 280 F.3d at 136. The fact of injury is often referred
to as impact. See Blades, 400 F.3d at 569 (using both “fact of damage or injury” and “impact”
interchangeably).
“Impact” means that the alleged antitrust violation caused injury to the plaintiff. State of
Ala. v. Blue Bird Body Co., 573 F.2d 309, 317 (5th Cir. 1978). “An antitrust injury is an ‘injury
of the type that antitrust laws were intended to prevent and that flows from that which makes
defendants’ acts unlawful.’” Se. Mo. Hosp. v. C.R. Bard, Inc., 645 F.3d 608, 624 (8th Cir. 2011)
(quoting Blue Shield of Va. v. McCready, 457 U.S. 465, 482 (1982)). “[A]n increase in price
resulting from a dampening of competitive market forces is assuredly one type of [antitrust]
injury.” McCready, 457 U.S. at 482. This is the injury alleged by Plaintiffs. Plaintiffs allege

15

Defendants’ allocation of customers and territories allowed Defendants to charge supracompetitive prices in the Midwest and New England. 2d Consolidated Am. Class Action Compl.
¶¶ 39, 70(a), 77, 83. Therefore, to proceed as a class action, Plaintiffs must be able to show with
common evidence that supra-competitive prices were in fact charged by Defendants.
In making a class certification determination, the Court must conduct a “rigorous
analysis.” Avritt, 615 F.3d at 1029 (citation omitted). In conducting this analysis, the Court
may look behind the pleadings to the extent necessary to determine whether, given the factual
setting and assuming Plaintiffs’ general allegations are true, common evidence could suffice to
establish a prima face case for the class. Blades, 400 F.3d at 566. The closer a dispute comes to
the heart of Plaintiffs’ claims, the greater the precaution that must be exercised in resolving such
disputes to determine the nature of the evidence required to state a prima facie claim. Id. at 567.
Assuming without deciding that Plaintiffs have met the Rule 23(a) requirements, neither
the New England Class nor Midwest Class is certified because after rigorous analysis of the
nature of the evidence in this case, Plaintiffs have failed to establish that common issues
predominate with respect to impact. The proposed New England Class and proposed Midwest
Class are each analyzed in turn below.
2. New England Class
Plaintiffs aver common evidence can show class-wide impact in New England through
the contrary hypothesis test, the variance test, and list prices. As discussed above, the impact
claimed to have been incurred by New England Class members is paying supra-competitive
prices. None of the methods proposed by Plaintiffs would be able to show through common
evidence that the New England Class members paid supra-competitive prices.

16

To understand the nature of evidence required to prove impact through common
evidence, it is helpful to consider the prototypical case where class certification would be
appropriate as well as cases where individual issues predominate with respect to impact. For
many goods price is invariable, it is a fixed-value for all customers. For example, gas stations
typically charge a single price for diesel fuel, the price does not vary among customers.
Therefore, if a gas station were able to monopolize the diesel market and elevate the price of
diesel above its competitive level, proving common impact would be straightforward. A plaintiff
seeking class treatment would need only offer a method to prove the price of diesel but for the
gas station’s monopoly would have been lower than the price actually charged. Then, any
customer who purchased diesel at that elevated price would have suffered injury, and common
issues would predominate over individual issues with respect to impact.
The analysis here, however, is complicated because prices vary by customer. When
prices vary, each customer may have to establish its own “but for” price as well as price
ultimately paid–both individual inquiries. See Blades, 400 F.3d at 574 (affirming denial of class
certification where proving injury to purchasers of genetically modified seeds with negligible or
zero price premiums–as compared to non-genetically modified seeds–would require calculating
hypothetical competitive prices that would vary among types of seeds and geography); see also
In re Flash Memory Antitrust Litig., No. C 07-0086, 2010 WL 2332081, at *8 (N.D. Cal. June 9,
2010) (“As a general matter, antitrust claims predicated on negotiated transactions, as opposed to
purchases based on list prices, often entail consideration of individualized proof of impact.”)
(citations omitted). Variable pricing, however, does not absolutely bar class certification.
Courts have found class certification warranted notwithstanding individually negotiated prices

17

where negotiation took place from an inflated, supra-competitive starting point. In re Potash
Antitrust Litig., 159 F.R.D. 682, 696 (D. Minn. 1995); In re Wirebound Boxes Antitrust Litig.,
128 F.R.D. 268, 271–72 (D. Minn. 1989) (certifying class despite individually negotiated prices
on industrial boxes where expert testified he could demonstrate “uniform minimum level of
inflation” of prices).
With respect to the New England Class, Plaintiffs aver three distinct methods for
showing impact to all class members through common evidence–the contrary hypothesis test, the
variance test, and list prices.
a. Contrary Hypothesis Test
The first method proposed by Plaintiffs is the contrary hypothesis test. The contrary
hypothesis test purports to demonstrate that SuperValu’s influence on C&S’s pricing in New
England was not limited geographically. Even accepting the test’s validity and reliability (a
decision the Court has deferred), the geographic range of SuperValu’s competitive effect does
not prove prices were actually raised by C&S to supra-competitive levels following the AEA.
Plaintiffs’ logic is as follows: (1) the contrary hypothesis test proves SuperValu
restrained prices for C&S throughout New England (as opposed to restraining only the subsets of
New England customers closest to SuperValu distribution centers), (2) economic theory suggests
that the elimination of competition facilitates a raise in prices, (3) the AEA eliminated
competition, and (4) therefore prices were raised throughout New England. Absent from this
logic chain is any method for proving that prices were actually raised above their competitive
level for every New England class member as required. See, e.g., Blue Bird, 573 F.2d at 327
(“[E]ach plaintiff must still prove that this [price-fixing] conspiracy was actually implemented in

18

his state and that it did in fact cause him injury.”). Plaintiffs cannot rely on theory alone for
point (2), they must produce a method of confirming that theory matched reality for each class
member because they bear the burden of proving impact. See Amerinet, Inc. v. Xerox Corp.,
972 F.2d 1483, 1493–94 (8th Cir. 1992) (noting that plaintiff’s burden of proof with respect to
antitrust injury is not relaxed as with regard to amount of antitrust damages).
Here, the contrary hypothesis test does not address the price levels after the AEA and
whether each member of the New England Class was in fact charged a supra-competitive price.
If accepted as valid, the contrary hypothesis test proves that SuperValu’s presence in New
England influenced C&S’s prices throughout that region. However, even accepting that premise,
the contrary hypothesis shows nothing about prices for C&S customers after the AEA (and in
fact does not analyze upcharges after the AEA at all). Without the presence of SuperValu in
New England, some C&S customers might have been able to resist supra-competitive pricing
because other variables beyond SuperValu’s presence affect competition.
C&S’s prices are all the result of individual negotiations. Many factors influence those
negotiations. For example, the size of orders, frequency of orders, and transportation costs all
influence upcharge levels. Saia Dep. 43:17-22, 45:8-10. The local competitive market also
factors into negotiations. Saia Dep. 43:23-25. C&S’s largest competitor after the AEA was
Bozzuto’s. Saia Dep. 191:6-9. Close proximity to Bozzuto’s or another strong local competitor
may effect the local competitive market and influence upcharges. C&S’s general business
strategy, which includes serving customers for a lower price with the expectation of expanding
services to that customer, also factors into the negotiated price. See Saia Dep. 70:19-72:13
(discussing situations where C&S would consummate negotiated deal for strategic reasons

19

notwithstanding low profit margin). A C&S customer endowed with a large amount of
bargaining power by virtue of any these significant factors could be able to resist an attempt by
C&S to raise prices after the AEA. Admittedly, as Dr. Leitzinger explains, years of economic
research and theory corroborate the link between competition and pricing. Leitzinger Decl. ¶¶
50-51. Indeed, it may very well be that every New England class member was actually charged
a supra-competitive price. However, proving that fact cannot be done with the contrary
hypothesis test, which merely assumes that prices will be raised after the AEA because
competition is reduced. The facts and circumstances of this case–with the potential for
numerous factors affecting both the price each customer would receive in a competitive
environment and the price each customer actually received–require a more searching analysis
than merely assuming prices would rise for every class member because competition was
reduced overall. The contrary hypothesis test has no evidentiary value in establishing a prima
facie case of injury for each New England Class member.
b. Variance Test
The variance test fares no better than the contrary hypothesis test in proving class injury.
The variance test compares the variance in C&S upcharges in New England before and after the
AEA. Because the variance in upcharges did not decrease, Plaintiffs aver SuperValu’s
competitive influence was not particularized to any subset of customers and therefore any injury
would be class-wide. Plaintiffs’ logic is as follows: (1) owing to the relationship between prices
and competition, if SuperValu’s competitive influence were particularized to some subset of
C&S customers, C&S would already be charging higher prices to those customers, (2) after
SuperValu exited the market, the price differences between the customers affected by

20

SuperValu’s influence and those that were not would converge to a single supra-competitive
price, (3) if prices converge variance will decrease, and (4) variance did not decrease, therefore
SuperValu’s influence was not particularized to any subset of C&S customers. As with the
contrary hypothesis test, however, the variance test does not establish supra-competitive prices
were actually paid by any class members.
Again, at this procedural stage the focus is on the nature of the evidence, and whether
impact can be proven can be proven for all class members with the same evidence. Blades, 400
F.3d at 566. The evidence offered here is an interpretation of the change in variance of C&S’s
upcharges before and after the AEA. Therefore, the nature of variance must be understood.
Variance is a statistical measure of the spread of data calculated by comparing each data
point’s value with the average value of the data set. See Encyclopedia Britannica Online
Academic Edition, Encyclopedia Britannica Inc., (July 11, 2012),
http://www.britannica.com/EBchecked/topic/562938/standard-deviation (defining standard
deviation as the square root of variance, and defining variance as the square of the difference
between each value and their mean divided by the number of values). A high variance means the
data points in a set are spread apart relative to each other. A lower variance means the data
points are closer together relative to each other. Because variance measures the spread of values
relative to each other, variance has several properties relevant to this discussion. First, if each
value is increased by the same amount, variance will be unchanged. For instance, if all
upcharges by C&S were increased by $1, the variance would remain constant. Second, because
variance is not affected by whether values are greater or lesser than average (variance only
measures how far a value is from average), variance will also remain constant if the change in

21

values offset each other. As a rudimentary example, if one upcharge was $1 above average and
another was $1 below average, if the first decreased by $2 and the second increased by $2,
variance would remain constant.
With those principles in mind, whether the variance test can show impact for all New
England class members will be considered. Dr. Leitzinger’s interpretation of the variance test
rests on his expectation that variance will decrease, which in turn rests on the premise that some
C&S customers were already being charged supra-competitive prices before the AEA. However,
once that premise is rejected, the change in variance of upcharges does nothing to inform the
factfinder whether prices increased for all class members. For example, an increase in variance
could mean that all upcharges increased for all class members but some increased more than
others. An increase in variance could also mean that some customers experienced increased
upcharges while other upcharges remained constant, a result which could preclude class
certification. See Avritt, 615 F.3d at 1034 (“[A] class cannot be certified if it contains members
who lack standing.”) (citation omitted). Plaintiffs offer no method for determining whether the
change in variance is the result of across-the-board price increases or something else.
Decrease and increase in variance are not the only possible results—variance can also
remain constant. Dr. Leitzinger’s report is unclear whether variance increased or remained
constant, he merely avers there was no statistically significant decrease. Leitzinger Report at
64–95 n.242. A constant variance would be most beneficial to Plaintiffs’ case because if it could
be proven that average upcharges increased then a constant variance would be consistent with a
uniform increase in prices. However, even assuming that variance was indeed constant,
alternative explanations exist that must be rebutted before impact could be proven by a

22

preponderance of the evidence. If the average upcharge increased and variance remained
constant, that result would also be consistent with some class members experiencing no impact
due to the possibility that variance remained constant owing to an offset by the increase in
upcharges for other class members.
For example, if there are two class members, one that paid a lower than average upcharge
before the AEA and another that paid a higher than average upcharge before the AEA. The
spread of the upcharges for those customers relative to average (i.e. the variance) will remain
constant if, after the AEA, the first class member’s upcharge was unchanged and the other’s
upcharge increased less than average. In that case, as the average increased the first class
member’s upcharge would get farther from average while the second’s would become closer to
average. Therefore, the sum of those distances from average, the variance, could be the same
before and after. The converse situation would also cause a static variance. If a customer paid a
lower than average upcharge before the AEA but then its upcharge increased dramatically
afterwards, that swing could be offset by a higher-paying customer that had its upcharge increase
only a little or not at all. Again, in that case the spread of the data relative to its average could be
the same. The variance could remain constant because the overall spread of the data from the
average upcharge remained the same. In fact, there are nearly limitless ways that the variance
could remain constant (or increase or decrease) given the high number of data points.
Constant variance could only prove upcharges increased for all customers if we assume
(or infer from some evidence) that upcharges behaved in a uniform manner. However, Plaintiffs
cannot merely assume what they must prove, see In re Wirebound Boxes Antitrust Litig., 128
F.R.D. at 271–72 (finding impact provable through common evidence where expert would testify

23

to “uniform minimum level of inflation” for all purchasers), and no evidence has been offered
that upcharges behaved in a uniform manner after the AEA. To the contrary, all evidence
suggests that upcharges varied widely.
Dr. Johnson, the defense expert, states that upcharges varied across customers, across
product categories, and across time. Johnson Report ¶¶ 111-12. Dr. Leitizinger disputes much
of Dr. Johson’s analysis, but nonetheless does not dispute that upcharges varied across
customers, categories, and time. Indeed, Dr. Leitzinger’s analysis confirms that consistently
across categories, some C&S customers experienced an increase in upcharges, while some
experienced a decrease, and some experienced no change. Leitzinger Decl. Attach. 9. This
evidence is probative that upcharge increases after the AEA, if any, would not be uniform.
Indeed, evaluating the evidence prospectively, it may be that a certain product category would be
less susceptible to a price increase, for example if there were the presence of a strong partial-line
supplier. Or, it may be the case that certain customers were less susceptible to a price increase
because of the presence of another local wholesaler, the ability to self-distribute, or bargaining
power derived from high-volume purchases. In light of the nature of the wholesale grocery
industry and the evidence presented, a uniform increase in upcharges across the New England
cannot be inferred.
Finally, as Plaintiffs aver, a price may be supra-competitive without actually increasing if
a competitive price would still be lower than the price charged. Therefore, whatever the
variance is before or after the AEA, it could still be consistent with a supra-competitive price
because upcharges do not necessarily have to increase for an antitrust injury to be present if
upcharges would have decreased or decreased to a greater degree but for the alleged antitrust

24

violation. Plaintiffs, however, have not identified how they will establish what competitive
upcharges would have been on a systematic basis. With all the variability of upcharges–some
product categories increasing, others decreasing, others remaining the same, customers paying
different upcharge amounts, and upcharges varying for a single customer across time–no
evidence of uniform movement exists and no evidence exists from which to uniformly calculate
a competitive price. Seemingly, the only method to determine whether all class members were
injured would be to conduct an individual inquiry into the upcharges actually paid by customers
compared to some competitive benchmark. This cannot be accomplished with the variance test,
and it is not common evidence that can establish a prima facie case of impact for each New
England class member.
c. Contrary Hypothesis and Variance Test Plus Benchmarking
In summary, neither the variance test nor the contrary hypothesis test presents common
evidence showing impact, i.e. the fact of an antitrust injury, for each class member. Even
accepting the methods for arriving at those tests’ conclusions as sound and reliable (which again
the Court has not yet done), the results of the test are immaterial. With respect to the contrary
hypothesis test, that SuperValu may have had competitive influence across New England prior to
the AEA does not establish that C&S used the lack of that influence after the AEA to actually
raise prices for all New England class members. Plaintiffs merely assume that C&S did in fact
raise prices in response to the absence of SuperValu as a competitor. With respect to the
variance test, the results are consistent with some class members suffering no injury, and
therefore cannot prove injury in fact to all class members. With both tests, the problem is no
evidence proves upcharges actually increased or, if upcharges did not increase, that upcharges

25

were supra-competitive.
However, Plaintiffs aver that the damage calculation used by Dr. Leitzinger proves that
upcharges increased, and therefore when combined with the contrary hypothesis and variance
tests, show that the increase was class-wide. Plaintiffs’ analysis is flawed. Dr. Leitzinger
proposes using benchmarking to determine damages. Benchmarking compares either (1)
Defendants’ profits during the time period with no competition with profits during time periods
with competition–specifically the time period before the AEA and the time period after the noncompete provisions of the AEA expired, or (2) compares Defendants’ profits in New England
and the Midwest with profits from distribution centers outside of New England and the Midwest
during the time period of the alleged antitrust violation. Leitzinger Report ¶ 147. However,
benchmarking only compares averages. Id. ¶¶ 148-153. That profits may have increased on
average, does not mean that monopolist profits were extracted from each class member.
Furthermore, as discussed above, the contrary hypothesis and variance tests cannot establish that
prices or upcharges or profits actually increased for each class member. Therefore, while profits
may have increased on average, Plaintiffs still do not identify any method for showing that this
average increase was due to a supra-competitive price actually being charged to each New
England Class member. Neither the contrary hypothesis test nor the variance test are capable of
so doing.
d. List Prices
Having dispelled the contrary hypothesis test and variance test, one more method of
proving impact through common evidence has been offered by Plaintiffs–“list” prices. Plaintiffs
argue that list prices are common across all class members, and therefore if they can show list

26

prices were inflated, impact has been shown through common evidence. Plaintiffs’ basic legal
premise is correct—if individual negotiations took place from a list price that was inflated due to
an antitrust violation, impact has been established. See, e.g., In re Potash Antitrust Litig., 159
F.R.D. at 696 (certifying class where prices increased “dramatically and uniformly” creating
inflated list prices that served as common starting point for negotiations). However, Plaintiffs’
legal premise does not match the facts of this case. Plaintiffs cannot show that the AEA inflated
list prices beyond their competitive level because the parties to the AEA, C&S and SuperValu,
do not set list prices. List prices are set by manufacturers. Safranski Decl. Ex. I 73:3-7.
Furthermore, list prices are nationwide. Id. Plaintiffs have not alleged that manufacturers
conspired with C&S to raise list prices, or that such a conspiracy extended across the entire
United States. Therefore, while list prices are certainly common across the New England Class
plaintiffs, they do not establish any causal link between the prices paid by class members and the
alleged antitrust violation.
Without the ability to prove impact through evidence common to class members, the New
England Class does not meet the predominance requirements of Rule 23(b)(3) and the New
England Class will not be certified. Plaintiffs’ Motion for Class Certification as it relates to the
New England Class is denied.
3. Midwest Class
Plaintiffs contend that, as with New England Class, the contrary hypothesis and variance
tests are common evidence that can prove impact to each Midwest Class member. Plaintiffs
further contend that ABS pricing can serve as common evidence proving impact to each class
member.

27

As with the proposed New England Class, the antitrust injury alleged to have been
inflicted on the proposed Midwest Class is the charging of supra-competitive prices. 2d
Consolidated Am. Class Action Compl. ¶¶ 39, 70(a), 77, 83. Like C&S in New England,
SuperValu does not charge any single price for its products and services. In the Midwest,
however, SuperValu used a formulaic approach to pricing–known as ABS pricing–during the
relevant time period. See 2d Consolidated Am. Class Action Compl. ¶ 25. Not every SuperValu
customer in the Midwest paid for products and services under ABS, some negotiated other fee
arrangements and some bought only products not priced using ABS. See, e.g., Johnson Report,
Exs. 141, 180 (showing some customers negotiated alternative fee arrangements). Whether the
contrary hypothesis test, variance test, or analysis of ABS pricing can be common evidence
showing impact to each Midwest class member under this pricing regime will be considered in
turn below.
a. Contrary Hypothesis and Variance Tests
Just as in the New England market, Plaintiffs propose using the same statistical tests
developed by Dr. Leitzinger–the contrary hypothesis and variance tests, discussed above–to
show impact in the Midwest. The tests suffer from the same deficiencies as applied in the
Midwest to SuperValu’s pricing regime as they do in New England for C&S, and therefore
cannot establish a prima facie case of impact for each Midwest class member.
As with New England, the contrary hypothesis test does not establish any supracompetitive prices were actually paid by class members in the Midwest. At most, the contrary
hypothesis test disproves that Fleming’s competitive impact in the Midwest prior to the AEA
was geographically localized. That evidence, however, does not establish whether prices after

28

the AEA were actually supra-competitive. Likewise, the variance test cannot show impact for all
class members in the Midwest. As discussed above with respect to the New England class,
changes in variance can be caused by many different changes in data. If variance increased after
the AEA, it would be consistent both with ABS fees increasing for all customers at different
rates and with ABS fees increasing for some customers but not all customers. Furthermore, if
variance remained constant, it would be consistent with ABS fees remaining constant, ABS fees
increasing (or decreasing) uniformly, or with ABS fees increasing (or decreasing) in variable,
offsetting ways. Therefore, the analysis of the contrary hypothesis and variance tests as
discussed above is incorporated by reference, including the discussion of benchmarking. Rather
than rehashing the same analysis, the unique aspects of those tests as they relate to the proposed
Midwest Class will be discussed.
As with the proposed New England Class, the problem for the proposed Midwest Class is
there is no evidence from which to infer ABS fees increased in a uniform manner. Indeed, the
evidence is to the contrary. Operating fees under ABS Pricing include fee drivers that are static
across distribution centers and fee drivers that vary across distribution centers. Johnson Report
Ex. 133 at SV00129374.031, SV00129374.042, SV00129374.046. The precise fee ultimately
charged to each customer varies depending on customer ordering behavior. Id. at
SV00129374.046. Likewise, service fees under ABS vary across product categories and services.
Id. at SV001239374.051.
Therefore, just as with C&S pricing in New England, the variability of ABS pricing
precludes any inference of uniform movement. The evidence of record is that SuperValu’s ABS
fees in the Midwest increased, decreased, and stayed the same for various customers and in

29

various product categories. Johnson Report figs. 25, 26.7 With such variability, a constant
variance is better explained by a rearrangement of the data points relative to the mean and not
with a uniform increase in ABS fees. Indeed, as Plaintiffs recognize, the motivation of ABS
pricing was to allow customers to manipulate their fees so ABS fees for some customers could
have been lower whether or not a supra-competitive fee was charged. Without any uniform
movement, variance does not indicate any systemic relationship between an increase in ABS fees
generally or injury to each class member. Therefore, the variance test is not common evidence
of impact. However, as Plaintiffs contend, ABS fees could have contained embedded supracompetitive elements, even if customers were able to manipulate fees down overall. Whether
analysis of the ABS fees themselves can provide common evidence of impact will now be
considered.
b. ABS Pricing
Plaintiffs contend that if they can show that ABS fees were inflated due to the lack of
competition from the AEA, then all class members buying products under ABS8 pricing would

7

Dr. Leitzinger takes issue with Dr. Johnson’s use of ABS fees as a percentage of costs
in calculating whether ABS fees increased, decreased, or had no change. Leitzinger De1cl. ¶
171. However, it appears Dr. Leitzinger also used percentages in conducting the contrary
hypothesis and variance tests. Leitzinger Report at 64 n.242; see also Leitzinger Dep. 232:1-5
(indicating contrary hypothesis test was conducted to look at ABS fees “as a percentage of
sales”). Dr. Leitzinger is also critical of Dr. Johnson’s use of data dating back to 2003.
Leitzinger Decl. ¶ 174. However, the variance test was also calculated comparing 2003 data
with post-AEA data. Leitzinger Report at 64 n.242. Regardless the distinctions are immaterial
because there is no dispute that some putative class members experienced a decrease in ABS
fees. [redacted].
8

Defendants aver ABS pricing cannot be common evidence because not all class
members bought products and services under the ABS regime. Plaintiffs aver the analysis is
unaffected because SuperValu’s non-ABS deals were equivalently profitable. Leitzinger Report
¶ 124. The dispute need not be resolved at this juncture because even assuming the analysis
30

have suffered injury. The presence of ABS pricing and its formulaic nature makes a better case
for certification of the Midwest Class than for the New England Class. However, Plaintiffs’
argument fails because they cannot articulate a method for showing with the same evidence that
ABS fees were inflated. See Blades, 400 F.3d at 566 (“If the same evidence will suffice for each
[class] member to make a prima facie showing, then it becomes a common question.”) (citation
omitted). Rather, analyzing ABS fees for class members would require individual inquiry.
Plaintiffs’ argument that ABS fees can be common evidence of impact rests on the faulty
premise that ABS pricing does not vary by customer. To be sure, the ABS formula itself does
not vary. The ABS operating fees ultimately paid by customers vary due to differing behavior,
each distribution center uses the same formula and each distribution center applies that formula
using the same unit prices, the “fee drivers,” with only the ordering behavior and transportation
distances varying between customers. See Johnson Report Ex. 133 at SV001239374.026–46
(explaining ABS operating fees). Therefore, notwithstanding the fact that customers may pay
different total fees, if it could be shown the ABS operating fees were supra-competitive due to
the inflation of one or more of the “fee drivers” then impact would be shown to all customers.
However, ABS operating fees vary by customer, even those with identical behavior, in one
important regard–distribution center. Some fee drivers vary across distribution centers.
Different Midwest class members purchased products from different distribution centers. Any
one of the fee drivers could be the vehicle for imposing supra-competitive prices. Therefore, an
analysis of ABS operating fees to determine which components are in fact supra-competitive

were unaffected, ABS pricing is not common evidence capable of establishing a prima facie case
of impact for each class member.
31

would require an analysis of each distribution center’s fees and each distribution center’s
competitive conditions–an individual inquiry that precludes the use of the same evidence for
establishing injury.
For example, assume that analysis of SuperValu’s Hopkins, Minnesota distribution center
reveals that every fee driver of ABS operating fees unique to that distribution center was supracompetitive. Because each distribution center has different fees, there would be no reason to
believe SuperValu’s Fort Wayne, Indiana distribution center would share the same supracompetitive fee drivers. It could be the case that the Fort Wayne distribution center (with its
location near the intersection of Indiana, Ohio, and Michigan) competed for a substantial number
of customers with Spartan Stores, which has distribution centers nearby in Grand Rapids and
Plymouth, Michigan, Johnson Report fig. 42, and therefore had no supra-competitive fee drivers.
Likewise, it could be that the Fort Wayne distribution center did charge supra-competitive prices
but that labor is cheaper in Fort Wayne than Hopkins and therefore SuperValu concentrated the
supra-competitive aspects of its operating fees in Fort Wayne in labor-intensive fee drivers, such
as selecting, in order to mask them as compared to other distribution centers. The point is these
are all individualized inquiries. Individual inquiries must be made of each distribution center.
Then, the distribution center used by each class member must be identified. The same evidence
cannot show impact for operating fees.
Service fees are a closer question still because they do not vary across distribution
centers. However, because service fees do not vary across distribution centers, they are the same
regardless of whether a distribution center was within the Midwest, where anticompetitive
affects are alleged, or in other regions unaffected by the AEA. Therefore, on their face, service

32

fees do not include any anticompetitive price “premium;” service fees are the same in regions
with and without competition. Without any such premium, the only way to show service fees to
be supra-competitive would be to engage in an analysis of what the price would have been “but
for” any anticompetitive affect in the Midwest. See Blades, 400 F.3d at 573 (noting that impact
for alleged antitrust violation concerning genetically modified seeds that had no price premiums
compared to corresponding non-genetically modified seeds could only be shown by comparing
actual prices to hypothetical competitive prices). Such an analysis would be a daunting task, and
could not be done through common evidence. The entire impetus motivating of ABS service
fees is to provide certainty and predictability in a industry and marketplace rife with uncertainty.
Johnson Report Ex. 133 at SC00123974.047–50. No evidence has been adduced that any other
wholesaler offers a similar service with fixed rates. To the contrary, all evidence of record
suggests that the nature of the wholesale grocery pricing is individual negotiation. Therefore, to
determine what price a customer buying products with ABS service fees would have paid but for
the AEA would require an individual inquiry into that customer’s bargaining power and the
results of hypothetical negotiations in a competitive market. The same evidence cannot show
impact for class members charged ABS service fees.
In summary, ABS pricing cannot provide common evidence to establish a prima facie
case of class impact. As such, class certification is not warranted for the Midwest Class and
Plaintiffs’ motion is denied with respect to those Plaintiffs.
IV. CONCLUSION
Based upon the foregoing, and all the files, records, and proceedings herein, IT IS
HEREBY ORDERED that:

33

1.

Plaintiffs’ Motion for Class Certification [Docket No. 202] is DENIED;

2.

Defendants’ Daubert Motion [Docket No. 211] is DENIED;

3.

This Order shall be temporarily filed under seal, the seal shall be lifted at the
close of business on July 26, 2012 to protect possible trade secret information.

BY THE COURT:

s/Ann D. Montgomery
ANN D. MONTGOMERY
U.S. DISTRICT JUDGE
Dated: July 16, 2012.

34