Federal Trade Commission v. Sysco Corporation

UNITED STATES DISTRICT COURT

FOR THE DISTRICT OF COLUMBIA

)

Federal Trade Commission, et al., )

)

Plaintiffs, )

)

v. ) Civil No. 1:15-cv-00256 (APM)

)

Sysco Corporation, et al., )

)

Defendants. )

~~~~~~~~~~~~~~~~ )

MEMORANDUM OPINION

TABLE OF CONTENTS

INTRODUCTION ......................................................................................................................... 1

BACKGROUND ........................................................................................................................... 3

I. The Foodservice Distribution Industry ............................................................................... 3

A. Overview ................................................................................................................. 3

B. Channels of Foodservice Distribution .................................................................... 4

I. Broadfine Distriln1tors ................................................................................ 4

2. .s:vstems Distributors ................................................................................... 6

3. Specialty Distributors ................................................................................. 7

4. Caslz~aml~Can:r and Cl11h S'torcs .............................................................. 7

C. Foodservice Distribution Customers ...................................................................... 8

I. Group Purchasing Orga11i::atio11s ............................................................... 8

Foodsen'ice Managemcnt Companies ........................................................ 9

3. Hospitality Chains ...................................................................................... 9

4. Restaurant Chains ..................................................................................... I 0

5. Govermnem Agencies ............................................................................... I 0

6. ''Street" C'usto1ners ................................................................................... 10

II. Case History ...................................................................................................................... 11

A. Sysco and USF ...................................................................................................... 11

B. History of the Merger. ........................................................................................... 11

C. History of these Proceedings ................................................................................ 12

LEGAL STANDARD ................................................................................................................. 14

I. Section 7 of the Clayton Act ............................................................................................. 14

II. Section 13(b) Standard for Preliminary Injunctions ......................................................... 14

III. Baker Hughes Burden-Shifting Framework ..................................................................... 16

DISCUSSION .............................................................................................................................. 17

I. The Relevant Market. ........................................................................................................ 17

A. Broadline Distribution as a Relevant Product Market .......................................... 19

1. Legal Principles A:lf'ecting the Definition r~f the

Relevant Product Market .......................................................................... 19

., The Brown Shoe "Practical Indicia" ....................................................... 23

3. E'lq>erl Testfn101~v ...................................................................................... 33

4. Conclusion as to the Broadline Product Market ...................................... 41

B. National Broadline Distribution as a Relevant Product Market ........................... 41

1. Legal Basis for De.fining Relevant Product Market

Based on Custo1ner Type .......................................................................... 42

2. Evidence Supporting a National Broadline Product Market .................... 44

C. Product Market Summary ..................................................................................... 60

D. Relevant Geographic Market ................................................................................ 60

l. iVational 1'v.larker ........................................................................................ 62

2. Local lt4arkets ....... ,, ...................... ,, ........................ ,, ...... ,, ..................... ., .. 62

II. The Probable Effects on Competition ............................................................................... 66

A Concentration in the National Broadline Customer Market ................................. 67

/. Dr. lsme! 's National Broadline Customer

Market Shares Calcu/at;ons ...................................................................... 67

2. Defimdants 'Arguments ............................................................................. 69'

3. 711e ('ourt 's Finding as to Nmfonal Broadline Customer

i\1arket 5J1ares ........................................................................................... 72

B. Concentration in the Local Markets ...................................................................... 72

/. Dr. Israel's Locc1! Broadline Customer

lv!arket 5Juires Calculations ...................................................................... 72

2. DeJi!mlants ·A rgumenrs .... ......................................................................... 74

.).

')

The Court's F'inding as to Local Broarlline Customer

lvfarket Shares ........................................................................................... 81

c. Additional Evidence of Competitive Harm .......................................................... 81

1. Unilateral E!Jects i\fationaf Cus1omer Market .......................... ,_ ........... 81

2. Merger Simulation Modef. . . . . Natio11al Cuswmer Marker .......................... 89'

3. Unilateral lijfects-Local lvfarkets .,,, ....................................................... 92

4. Local Event Studies ........................ ,, ...... ,. ................................................. 97

5. • . ., c19

,Slt1111na1J ............ ,, ... ,. ................................................. ,. ............................. ::1

III. Defendants' Rebuttal Arguments .................................................................................... 100

A. PFG Divestiture .................................................................................................. 100

!. Competitive Pressure E'xerted hy Post-Divestiture PFG ........................ 102

2. Addi1fonal Disadvantages Faced hy Post-Merger PFG .................... ,..,. 107

3. Posr-A1erger PFG as an independent Competitor .................................. l 09

B. Existing Competition .......................................................................................... 110

1. Regionalization ....................................................................................... l lO

3. Conclusion as to Existing Competition,,,.,, ......................................... ,, .. 114

C. Entry of New Firms and Expansion of Existing Competitors ............................ 114

D. Efficiencies ......................................................................................................... 117

!. Requirememfor Merger-Specffic and Verifiable E,Yficiencies ................ t 17

2. !nsttfJiciemy of Estimated lvferger-Specific Savings ............................... 123

E. Conclusion .......................................................................................................... 124

IV. The Equities .................................................................................................................... 125

CONCLUSION ......................................................................................................................... 127

11

INTRODUCTION

Americans eat outside of their homes with incredible frequency. The U.S. Department of

Commerce, for instance, recently reported, for the first time since it began tracking such data, that

Americans spent more money per month at restaurants and bars than in grocery stores. 1 Of course,

Americans eat out at many other places, too-sports arenas, school and workplace cafeterias,

hotels and resorts, hospitals, and nursing homes, just to name a few. The foodservice distribution

industry supplies food and related products to all of these locations. Foodservice distribution is

big business. In 2013, the market grew to $231 billion. By some estimates, there are over 16, 000

companies that compete in the foodservice distribution marketplace.

The two largest foodservice distribution companies in the country are Defendants Sysco

Corporation ("Sysco") and US Foods, Inc. ("USF"). Both are primarily "broadline" foodservice

distributors. As the name implies, a broadline foodservice distributor sells and delivers a "broad"

array of food and related products to just about anywhere food is consumed outside the home.

In 2013, Sysco's broadline sales were over sm billion and USF's were over sm billion.

In December 2013, Sysco and USF announced that they had entered into an agreement to

merge the companies. Fourteen months later, in February 2015, Sysco and USF announced that

they intended to divest 11 USF distribution facilities to the third largest broadline foodservice

distributor, Performance Food Group, Inc., if the merger received regulatory approval.

On February 20, 2015, the Federal Trade Commission ("FTC") and a group of states filed

suit in this court seeking an injunction to prevent the proposed merger. Specifically, under Section

13(b) of the Federal Trade Commission Act, the FTC asked this court to halt the proposed merger

1

Michelle Jamrisko, Americans ' Spending on Dining Out Just Overtook Grocery Sales for the First Time Ever,

Bloomberg Business (Apr. 14, 2015), http://www.bloomberg.com/news/articles/2015-04-14/americans-spending-on-

dining-out-just-overtook-grocery-sales-for-the-first-time-ever.

1

until the FTC completes an administrative hearing-scheduled to begin on July 21, 2015-to

determine whether the proposed combination would violate Section 7 of the Clayton Act.

The precise question presented by this case is whether the court should enjoin Sysco and

USF from merging until the proposed combination is reviewed by an FTC Administrative Law

Judge. The real-world impact of the case, however, is more consequential. Sysco and USF have

announced that they will not proceed with the merger ifthe court grants the requested injunction.

The proceedings in this case have been extraordinary. The FTC investigated the proposed

merger for more than a year before filing suit. Then, within a two-month period, the parties worked

tirelessly to exchange millions of documents, depose dozens of witnesses, and secure over a

hundred declarations. The court heard live testimony for eight days in early May 2015. Counsel

for the parties have done all of this work while exhibiting the highest degree of skill and

professionalism.

Congress passed the Clayton Act to enable the federal government to halt mergers in their

incipiency that likely would result in high market concentrations. Congress was especially

concerned with large combinations that would impact everyday consumers across the country.

The court has considered all of the evidence in this case and has reached the following conclusion:

The proposed merger of the country's first and second largest broadline foodservice distributors is

likely to cause the type of industry concentration that Congress sought to curb at the outset before

it harmed competition. The court finds that the FTC has met its burden under Section l 3(b) of the

Federal Trade Commission Act of showing that the requested injunction is in the public interest.

The court, therefore, grants the FTC's motion for preliminary injunctive relief.

2

BACKGROUND

I. THE FOODSERVICE DISTRIBUTION INDUSTRY

A. Overview

Defendants operate in a $231 billion foodservice distribution industry, where over 16,000

companies battle daily to sell food and related products to restaurants, resorts, hotels, hospitals,

schools, company cafeterias, and so on-everywhere food is served outside the home.

Hr' g Tr. 1324; DX-00329 at 17. The types of customers served by the foodservice distribution

industry come in all shapes and sizes. They range from independent restaurants, to well-known

quick-service and casual dining chains (e.g., Five Guys, Subway, and Applebee's), to hospitality

procurement companies and hotel chains (e.g., Avendra, Hilton Supply Management, and

Starwood Hotels and Resorts), to government agencies (e.g., the U.S. Department of Veterans

Affairs), to foodservice management companies (e.g., Aramark, Sodexo, and Compass Group), to

healthcare group purchasing organizations (e.g., Premier, Novation, and Navigator).

The industry recognizes four general categories of foodservice distribution companies:

(i) broadline distributors, (ii) systems distributors, (iii) specialty distributors, and (iv) cash-and-

carry and club stores. Customers commonly purchase from foodservice distributors in one or more

of these different categories, or "channels," mixing and matching to suit their needs. For example,

customers may purchase products directly from a broadline distributor; they may contract with a

brand-named food manufacturer (e.g., Tyson Foods for chicken or Kellogg's for cereal) and use a

broadline or systems distributor for warehousing and delivery; they may use specialty distributors

for select items such as produce or seafood; or they may make their purchases at a cash-and-carry

or club store (e.g., Restaurant Depot or Costco).

3

Understanding these different channels of distribution and the different customers they

serve is central to the antitrust analysis that this case demands. The court, therefore, describes

below the sellers and buyers of foodservice distribution in the United States.

B. Channels of Foodservice Distribution

1. Broadbne Distributors

Broadline distribution is characterized by several key features, including: (i) product

breadth and depth; (ii) availability of private-label products; (iii) frequent and flexible delivery,

including next-day service; and (iv) "value-added" services, such as menu and nutrition planning.

Broadline distributors offer thousands of distinct items for sale-known as "stock keeping

units" ("SKUs") for inventory management purposes-in a wide array of product categories,

including canned and dry goods, dairy, meat, poultry, produce, seafood, frozen foods, beverages,

and even janitorial supplies such as chemicals, cleaning equipment, and paper goods. Broadliners

also sell "private label" goods, which are akin to "Trader Joe's" or "Safeway" brand products

found in those grocery stores. "Private label" products are often comparable in quality to their

name-brand counterparts, but are cheaper in price. Because they are able to offer such a diverse

array of products, broadline distributors market themselves to customers as a "one-stop shop," by

virtue of their ability to supply most-if not all-food and related products needed by their

customers. Customers value the breadth of product offerings and the opportunity to aggregate a

substantial portion of their purchases with one distributor, allowing them to save costs. They also

appreciate broadliners' high level of customer service, which usually includes next-day and

emergency deliveries. Focusing heavily on individualized customer service, broadline distributors

employ much larger salesforces than the other channels.

4

Broadline distributors come in different sizes. The largest, by any measure, are Sysco and

USF. In 2013, Sysco and USF made. billion a n d . billion in broadline sales, respectively.

PX09350-236, Table 44. The next largest broadliner made less than $6 billion. Id Sysco and

USF are also the only two broadliners with true nationwide service capability. Sysco and USF

have 72 and 61 distribution centers, respectively-each with more than twice the number of

distribution centers operated by the next-largest broadliners. Because of their nationwide

footprint, Sysco and USF are often referred to as "national" broadliners. Combined, Defendants

employ over 14,000 sales representatives. No other broadliner employs more than 1,600.

Defendants together operate over 13,000 trucks. The next largest broadliners have just over 1,600.

The next tier of companies are "regional broadliners," so called because their distribution

capabilities are concentrated in discrete regions of the United States. The largest regional

broadliner, Performance Food Group ("PFG"), is the country's third-largest broadliner in terms of

sales. PFG operates 24 broadline distribution facilities, mainly in the eastern and southern parts

of the country and, in 2013, earned $6 billion in broadline revenue. The next five largest regional

broadline distributors, in order of 2013 revenues, are: (i) Gordon Food Service, which has

I 0 distribution centers mainly in the Midwest, Florida, and Texas; (ii) Reinhart Foodservice, which

has 24 distribution centers, primarily in the East and Midwest; (iii) Ben E. Keith Company, which

has seven distribution centers in Texas and bordering states; (iv) Food Services of America, which

has I 0 distribution centers, concentrated in the Northwest; and (v) Shamrock Foods, which has

four distribution centers in the Southwest and southern California. These regional broadliners had

2013 revenues ranging from approximately 4111 billion to 4111 billion.

5

The last tier of broadliners have five or fewer distribution centers and 2013 revenues of

less than $1.1 billion. Many of these operate in a single locality or region, like Shetakis

Wholesalers, which has one distribution center in Las Vegas, Nevada.

Regional broadline distributors have formed consortiums to compete for customers with

multi-regional distribution needs. The largest consortium is Distribution Market Advantage

("DMA"). DMA is a supply chain sales and marketing cooperative owned by nine independent

regional distributors, which are also its members, including Gordon Food Service, Ben E. Keith,

and Reinhart Foodservice. DMA does not own any trucks or distribution facilities; rather, its

purpose is to coordinate the bidding, contracting, and operational processes of its members to meet

the needs of large customers that require a distributor with extensive geographic coverage.

Another consortium is Multi-Unit Group ("MUG''), an alliance of 19 broadline distributors who

are part of UniPro Foodservice, a larger consortium that includes distributors in different channels.

As explained later, these regional consortia have had mixed results in competing for large,

geographically dispersed customers.

2. Systems Distributors

Systems distributors, also referred to as "custom" or "customized" distributors, primarily

serve fast food, quick service, fast casual, and casual chain restaurants (e.g., Burger King,

Wendy's, and Applebee's), which have fixed or limited menus. Unlike broadliners, systems

distributors do not carry a large, diverse number of SKUs. Rather, their inventory profile is a small

number of proprietary SKUs, which are manufactured specifically for the customer. For instance,

the systems distributor for Wendy's carries and delivers the food products needed for Wendy's'

menu and does not make those products available to others. As a result, systems distributors

typically provide only warehousing and transport services. They do not offer private label products

6

or value-added services such as menu planning, and they have very small salesforces, if any.

Systems distributors make large, limited-SKU deliveries on a fixed, limited schedule, and typically

do not offer next-day or emergency deliveries.

Some foodservice distribution companies operate both systems and broadline divisions.

For instance, Sysco operates SYGMA, a systems distribution division. SYGMA is run by a

different set of executives and, for the most part, operated out of separate distribution centers. PFG

offers systems distribution through PFG Customized, which is run separately from its broadline

division.

3. Specialty Distributors

Specialty distributors offer a limited and focused grouping of products within one or more

product categories-typically fresh produce, meat, seafood, dairy or baked goods. Other specialty

distributors focus on a specific type of cuisine, such as Italian fare. Many customers, especially

independent restaurants, use specialty distributors to supplement their purchases from broadline

distributors because the specialty distributor offers higher quality or fresher products than the

broadline distributor or provides unique products that the broadline distributor does not carry, such

as products from local farmers. Both in terms of number of SKUs and geographic coverage,

specialty distributors are typically smaller than broadline distributors.

To compete with specialty distributors, some broadliners operate specialty divisions.

Sysco, for instance, operates several specialty divisions separately from its broadline division. So,

too, does PFG, which operates Roma, a specialty division for Italian food products.

4. Cash-and-Carry and Club Stores

Cash-and-carry stores offer a "self-service" model of food distribution, in which customers

make purchases at the store and transport the purchased goods themselves. Club stores like Costco

7

and Sam's Club also fall within this distribution channel. With limited exceptions, cash-and-carry

stores do not deliver. They also offer fewer products than broadline distributors. For example,

the largest cash-and-carry store, Restaurant Depot, only carries up t o . SKUs. Additionally,

cash-and-carry stores do not have sales personnel dedicated to individual customers. Because of

these features, the prices offered by cash-and-carry stores are significantly lower than those offered

by broadliners. The typical cash-and-carry customer is an independent restaurant that either does

not meet broadline distributors' minimum purchase requirements or needs to supplement its

broadline deliveries.

C. Foodsenrice Distribution Customers

Foodservice distribution customers are a heterogeneous group. The largest customers, such

as group purchasing organizations and foodservice management companies, buy hundreds of

millions of dollars of product a year, whereas a single independent restaurant buys a small fraction

of that amount. Some customers choose to buy from a single line of distribution; others mix

distribution channels. Some customers demand fixed pricing, whereas others buy based on daily

market rates. Generally speaking, however, customers can be grouped into several categories.

/, Group Purchasing Organizations

Group purchasing organizations, or GPOs, are entities that, through the collective buying

power of their members, obtain lower prices for foodservice products. GPOs negotiate direct

contracts with food manufacturers and thereby secure lower prices than a member could

individually.

GPOs do not have their own distribution capabilities. Rather, they contract with broadline

distributors for warehousing, delivery, and operational services. When a member purchases a

GPO-contracted good, the member pays the broadliner on a "cost-plus" basis: it pays for the "cost"

8

of the product based on the GPO' s contract with the manufacturer, "plus" the distributor's markup,

which is negotiated between the GPO and distributor. GPOs also contract with broadliners to

allow their members to purchase products from breadline distributors (rather than from

manufacturers), in which case they pay the breadline distributor both the distribution margin

(markup) and the cost for the product set by the distributor. GPO members also buy from specialty

distributors.

GPOs are prominent in the healthcare and hospitality industries. The largest healthcare

GPOs include Premier, Novation, and Navigator. One of the largest hospitality GPOs is Avendra.

These companies annually spend hundreds of millions of dollars on breadline distribution.

2. Foodservice Management Companies

Foodservice management companies operate cafeterias or other dining facilities at

educational institutions, sports venues, and workplaces. Like GPOs, foodservice management

companies negotiate contracts with food manufacturers and rely on broadliners for storage and

delivery; they also purchase directly from broadliners and specialty distributors. Sodexo, Compass

Group, and Aramark are among the country's largest foodservice management companies. Those

three companies each spend approximately ti billion annually on breadline distribution.

3. Hospitality Chains

Hospitality chains are also large purchasers. Hilton Hotels, for example, uses a system

similar to a GPO. It has a subsidiary, Hilton Supply Management LLC, which negotiates contracts

on behalf of over 4,000 members to obtain food and related items at a discounted price. Other

hospitality companies, such as Hyatt Hotels, purchase most of their foodservice products through

Avendra, the largest hospitality GPO. Starwood Hotels and Interstate Hotels & Resorts, on the

other hand, directly manage food procurement and distribution contracts for their properties.

9

Regardless of the food purchasing model, hospitality chains also buy food directly from

broadliners and rely on them for their storage and delivery needs. These companies spend

hundreds of millions of dollars annually on broadline distribution. Individual hotels and resorts

also buy directly from specialty distributors, as needed.

4. Restaurant Chains

Restaurant chains come in many sizes with a wide variety of characteristics. This customer

category includes nationwide fast food or quick service restaurants such as Burger King and

Subway, each with thousands of locations in all regions of the country. It also includes regional

fast casual restaurant chains such as Culver's (primarily in the Midwest) and Zaxby's (primarily

in the Southeast), as well as nationwide sit-down restaurant chains, such as Applebee's and

Cheesecake Factory. The channel of distribution a chain restaurant uses depends, in part, on the

number of locations and menu variety. The greater the number of locations and the fewer the

menu items, the more amenable the chain restaurant is to systems distribution.

5. Government Agencies

Some government agencies, notably the Defense Logistics Agency and the U.S.

Department of Veterans Affairs, are large buyers of broadline distribution services. Those

agencies, for instance, spend hundreds of millions of dollars each year on broadline foodservice.

6. "Street" Customers

Customers with only one location, or a handful of locations, are referred to in the industry

as "street," "local," or "independent" customers. Examples of this type of customer include

independent restaurants and resorts. Unlike the types of customers identified above, street

customers usually do not have written contracts with broadliners; instead, they negotiate prices on

a weekly or other short term basis. They also tend to diversify their purchases among multiple

10

distribution channels. Indeed, according to a study conducted by an industry trade group, the

International Foodservice Distributors Association, the typical independent customer uses up to

twelve different supply sources. DX-00293 at 29.

II. CASE HISTORY

A. Sysco and USF

Defendant Sysco is a publicly-traded corporation headquartered in Houston, Texas. As the

largest North American foodservice distributor, Sysco distributes food to approximately 425,000

customers in the United States, generating sales of about $46. 5 billion in fiscal year 2014. CompI.

for TRO and Prelim. lnj. Pursuant to Section 13(b) of the FTC Act, ECF No. 3 at if 24 [hereinafter

Compl.]. Sysco's business is divided into three divisions: (i) Broadline (81 percent of revenue);

(ii) SYGMA, which provides systems distribution (13 percent ofrevenue); and (iii) "Other," which

provides, among other things, specialty produce distribution (6 percent of revenue). Id. if 25.

Sysco's broadline division operates out of 72 distribution centers located across the United States.

Id

Defendant US Foods, Inc., is a privately-held corporation based in Rosemont, Illinois, and

is a wholly owned subsidiary of Defendant USF Holding Corp. USF is controlled by the

investment funds of Clayton, Dubilier & Rice, Inc., and KKR & Co., L.P. The second-largest

foodservice distributor in the United States, USF operates 61 broadline distribution centers across

the country and serves over 200,000 customers nationwide. Id. if 27. In fiscal year 2013, USF

generated approximately $22 billion in revenue. Id

B. History of the Merger

On December 8, 2013, Sysco and USF signed a definitive merger agreement, whereby

Sysco agreed to acquire all shares ofUSF for $500 million in cash and $3 billion in newly issued

11

Sysco equity. Sysco also agreed to assume $4.7 billion in USF's existing debt, for a total

transaction value of $8.2 billion. The merger agreement expires on September 8, 2015.

After announcing the merger, Defendants filed a notification regarding the merger as

required by the Hart-Scott-Rodino Antitrust Improvements Act, 15 U.S.C. § 18a. As a result of

this filing, the FTC commenced an investigation to determine the effects of the proposed

combination. The FTC is an administrative agency of the United States federal government that

derives its authority from the Federal Trade Commission Act ("FTC Act"), 15 U.S.C. §§ 41 et seq.

Among other duties, the FTC is vested with authority and responsibility for enforcing Section 7 of

the Clayton Act, 15 U.S.C. § 18, and Section 5 of the FTC Act, 15 U.S.C. § 45.

During the FTC's investigation, and with the hope of gaining regulatory approval, on

February 2, 2015, Sysco and USF announced an asset purchase agreement with regional broadline

distributor Performance Food Group, Inc. ("PFG"), to sell 11 of USF's 61 distribution centers to

PFG, contingent upon the successful completion of the merger. The 11 USF distribution centers-

intended to increase PFG' s geographic footprint-are, for the most part, located within the western

half of the country, where PFG at present has only one distribution center. Currently, the

11 distribution centers account for approximately $4.5 billion in broadline sales. PX09250-0l l.

The parties also executed a Transition Services Agreement. Under the two agreements, PFG

would acquire all assets and employees at the 11 distribution centers, all customers under those

contracts (assuming the customers consent), and the right to use USF private label products at

those facilities for up to three years.

C. History of these Proceedings

On February 19, 2015, the Commissioners of the FTC voted 3-2 to authorize the filing of

an administrative complaint in the FTC's Article I court to block the proposed merger, based on a

12

finding that there was reason to believe that the merger would violate Section 7 of the Clayton Act,

15 U.S.C. § 18, and Section 5 of the FTC Act, 15 U.S.C. § 45. Trial before an Administrative Law

Judge is scheduled to begin on July 21, 2015.

Also, on February 19, 2015, the Commission authorized the FTC staff to seek a preliminary

injunction in federal court under Section 13(b) of the FTC Act in order to prevent Defendants from

completing the merger. The FTC filed this action on February 20, 2015, seeking a temporary

restraining order ("TRO") and preliminary injunction to maintain the status quo until the

conclusion of the administrative trial. The FTC is joined in this action by the District of Columbia

and the following states: California, Illinois, Iowa, Maryland, Minnesota, Nebraska, North

Carolina, Ohio, Tennessee, Pennsylvania, and Virginia (collectively, the "Plaintiff States"). By

and through their respective Attorneys General, the Plaintiff States have joined with the FTC in

this action pursuant to Section 16 of the Clayton Act, 15 U.S.C. § 26, in their sovereign or quasi-

sovereign capacities as parens patriae on behalf of the citizens, general welfare, and economy of

each of their states.

On February 24, 2015, Defendants stipulated to a TRO, agreeing not to merge until three

calendar days after this court rules on the FTC's Motion for Preliminary Injunction. The court

entered the stipulated TRO on February 27, 2015. Defendants have since represented that they

will abandon the transaction if this court grants the preliminary injunction.

On March 4, 2015, the court scheduled a preliminary injunction hearing to start on May 5,

2015. The parties' counsel accomplished an extraordinary amount of work in the two months

leading up to the evidentiary hearing. They exchanged approximately 14.8 million documents and

took 72 depositions. Moreover, in addition to the more than 90 industry participant declarations

that accompanied the FTC's motion for preliminary injunction, Defendants obtained 65 new

13

declarations or counter declarations, while the FTC obtained an additional 25 new or counter

declarations. During the eight-day evidentiary hearing, the court heard testimony from 20

witnesses, either live or via video deposition. The parties submitted a total of 185 declarations

into evidence, as well as over 3,500 exhibits and excerpts of over 70 depositions. The court heard

closing arguments on May 28, 2015.

LEGAL STANDARD

I. SECTION 7 OF THE CLAYTON ACT

Section 7 of the Clayton Act prohibits mergers or acquisitions "the effect of [which] may

be substantially to lessen competition, or to tend to create a monopoly" in "any line of commerce

or in any activity affecting commerce in any section of the country." 15 U.S.C. § 18. When the

FTC has "reason to believe that a corporation is violating, or is about to violate, Section 7 of the

Clayton Act," it may seek a preliminary injunction under Section 13 (b) of the FTC Act to "prevent

a merger pending the Commission's administrative adjudication of the merger's legality." FTC v.

Staples, Inc., 970 F. Supp. 1066, 1070 (D.D.C. 1997) (citing 15 U.S.C. § 53(b)). "Section 13(b)

provides for the grant of a preliminary injunction where such action would be in the public

interest-as determined by a weighing of the equities and a consideration of the Commission's

likelihood of success on the merits." FTC v. HJ Heinz Co., 246 F.3d 708, 714 (D.C. Cir. 2001)

(citing 15 U.S.C. § 53(b)).

II. SECTION 13(B) STANDARD FOR PRELIMINARY INJUNCTIONS

The Section 13(b) standard for preliminary injunctions differs from the familiar equity

standard applied in other contexts. As the Court of Appeals explained in Heinz: "Congress

intended this standard to depart from what it regarded as the then-traditional equity standard, which

it characterized as requiring the plaintiff to show: (1) irreparable damage, (2) probability of

14

success on the merits and (3) a balance of equities favoring the plaintiff." 246 F.3d at 714 (internal

citation omitted). The court continued: "Congress determined that the traditional standard was

not 'appropriate for the implementation of a Federal statute by an independent regulatory agency

where the standards of the public interest measure the propriety and the need for injunctive relief.'"

Id. (quoting HR. Rep. No. 93-624 at 31 (1971)); see also FTC v. Exxon Corp., 636 F.2d 1336,

1343 (D.C. Cir. 1980) ("In enacting [Section 13(b)], Congress further demonstrated its concern

that injunctive relief be broadly available to the FTC by incorporating a unique 'public interest'

standard in 15 U.S.C. [§] 53(b), rather than the more stringent, traditional 'equity' standard for

injunctive relief.").

Under Section 13(b)'s "public interest" standard, "[t]he FTC is not required to establish

that the proposed merger would in fact violate section 7 of the Clayton Act." Heinz, 246 F.3d at

714. Rather, to demonstrate the likelihood of success on the merits, "the government need only

show that there is a reasonable probability that the challenged transaction will substantially impair

competition." Staples, 970 F. Supp. at 1072 (citation omitted) (internal quotation marks omitted).

A trial court evaluating a demand for injunctive relief therefore must "measure the

probability that, after an administrative hearing on the merits, the Commission will succeed in

proving that the effect of the [proposed] merger 'may be substantially to lessen competition, or to

tend to create a monopoly' in violation of section 7 of the Clayton Act."' Heinz, 246 F.3d at 714

(quoting 15 U.S. C. § 18). The FTC satisfies this standard if it "has raised questions going to the

merits so serious, substantial, difficult and doubtful as to make them fair ground for thorough

investigation, study, deliberation and determination by the FTC in the first instance and ultimately

by the Court of Appeals." Id. at 714-15 (citations omitted) (internal quotation marks omitted).

This standard reflects Congress' use of the words "may be substantially to lessen competition" in

15

Section 7, as Congress' concern "was with probabilities, not certainties" of decreased competition.

Id at 713 (citing Brown Shoe Co. v. United States, 370 U.S. 294, 323 (1962)) (other citations

omitted).

Though more relaxed than the traditional equity injunction standard, Section 13(b)' s public

interest standard nevertheless demands rigorous proof to block a proposed merger or acquisition.

"[T]he issuance of a preliminary injunction prior to a full trial on the merits is an extraordinary

and drastic remedy." Exxon, 636 F.2d at 1343 (citations omitted) (internal quotation marks

omitted). That is because "the issuance of a preliminary injunction blocking an acquisition or

merger may prevent the transaction from ever being consummated." Id "Given the stakes, the

FTC's burden is not insubstantial .... " FTC v. Arch Coal, 329 F. Supp. 2d 109, 123 (D.D.C.

2004), case dismissed, No. 04-5291, 2004 WL 2066879 (D.C. Cir. Sept. 15, 2004). "[A] showing

of a fair or tenable chance of success on the merits will not suffice for injunctive relief." Id

(citation omitted) (internal quotation marks omitted).

III. BAKER HUGHES BURDEN-SHIFTING FRAMEWORK

In United States v. Baker Hughes, Inc., 908 F.2d 981, 982-83 (D.C. Cir. 1990), the Court

of Appeals established a burden-shifting framework for evaluating the FTC's likelihood of success

on the merits. See Heinz, 246 F.3d at 715 (applying Baker Hughes "to the preliminary injunctive

relief stage"). Under the Baker Hughes framework, the FTC bears the initial burden of showing

that the merger would lead to "undue concentration in the market for a particular product in a

particular geographic area." Baker Hughes, 908 F.2d at 982; see also Heinz, 246 F.3d at 715

(quoting United States v. Phi/a. Nat'! Bank, 374 U.S. 321, 363 (1963)) ("[T]he government must

show that the merger would produce 'a firm controlling an undue percentage share of the relevant

market, and [would] result[ ] in a significant increase in the concentration of firms in that

16

market.'"). Such a showing establishes a "presumption" that the merger will substantially lessen

competition. Baker Hughes, 908 F.2d at 982.

The burden then shifts to the defendant to rebut the presumption by offering proof that "the

market-share statistics [give] an inaccurate account of the [merger's] probable effects on

competition in the relevant market." Heinz, 246 F.3d at 715 (quoting United States v. Citizens &

S Nat 'l Bank, 422 U.S. 86, 120 (1975)) (internal quotation marks omitted); see also Baker Hughes,

908 F.2d at 991 ("[A] defendant seeking to rebut a presumption of anticompetitive effect must

show that the prima facie case inaccurately predicts the relevant transaction's probable effect on

future competition."). "The more compelling the prima facie case, the more evidence the

defendant must present to rebut it successfully." Baker Hughes, 908 F.2d at 991. "A defendant

can make the required showing by affirmatively showing why a given transaction is unlikely to

substantially lessen competition, or by discrediting the data underlying the initial presumption in

the government's favor." Id.

"If the defendant successfully rebuts the presumption, the burden of producing additional

evidence of anticompetitive effect shifts to the government, and merges with the ultimate burden

of persuasion, which remains with the government at all times." Id. at 983. "[A] failure of proof

in any respect will mean the transaction should not be enjoined." Arch Coal, 329 F. Supp. 2d at

116. The court must also weigh the equities, but if the FTC is unable to demonstrate a likelihood

of success, the equities alone cannot justify an injunction. Id.

DISCUSSION

I. THE RELEVANT MARKET

Merger analysis starts with defining the relevant market. United States v. Marine Bancorp.,

418 U.S. 602, 618 (1974) (Market definition is "'a necessary predicate' to deciding whether a

17

merger contravenes the Clayton Act.") (quoting United States v. E.J. Du Pont De Nemours & Co.,

353 U.S. 586, 593 (1957)); see also FTC v. Swedish Match, 131 F. Supp. 2d 151, 156 (D.D.C.

2000). The relevant market has two component parts. "First, the 'relevant product market'

identifies the product and services with which the defendants' products compete. Second, the

'relevant geographic market' identifies the geographic area in which the defendant competes in

marketing its products or service." Arch Coal, Inc., 329 F. Supp. 2d at 119; see also FTC v. CCC

Holdings Inc., 605 F. Supp. 2d 26, 37 (D.D.C. 2009) (same). "Defining the relevant market is

critical in an antitrust case because the legality of the proposed merger[] in question almost always

depends upon the market power of the parties involved." FTC v. Cardinal Health, Inc.,

12 F. Supp. 2d 34, 45 (D.D.C. 1998).

Market definition has been the parties' primary battlefield in this case. According to the

FTC, the relevant product market is broadline foodservice distribution. Compl. ~ 40. Because

broadline distribution is defined by a number of distinct attributes-such as a vast array of product

offerings, private label offerings, next-day delivery, and value-added services-the FTC contends

that the other modes of distribution are not reasonable substitutes for broadline distribution and

thus must be excluded from the product market.

The FTC further contends that, within the product market for broadline distribution, there

is another product market for foodservice distribution sold to "national" customers. Id ~ 44. These

customers, the FTC asserts, are distinct from "local" or "street" customers in multiple respects.

National customers have a nationwide or multi-regional footprint and, because of that footprint,

typically contract with a broadliner that has geographically dispersed distribution centers; they

usually make purchases under a single contract that offers price, product, and service consistency

across all facilities; and they award contracts through a request for proposal or bilateral

18

negotiations. National customers include, among others, GPOs, foodservice management

companies, hospitality chains, and national chain restaurants. By contrast, the FTC says, the

typical "local" or "street" customer is an independent restaurant, which does not require multiple,

geographically dispersed distribution centers; purchases in smaller quantities; and ordinarily does

not have a contract with its foodservice distributor(s) as it negotiates purchases on a weekly or

other short-term basis. The FTC contends that for national customers the geographic market is

nationwide. For local customers, it argues that the geographic market is localized near Defendants'

distribution centers.

Defendants counter that the foodservice distribution market cannot be sliced and diced as

advocated by the FTC. According to Defendants, the relevant market is the entire $231 billion

foodservice distribution industry, consisting not only of broadline food distributors, but also

specialty distributors, systems distributors, and cash-and-carry stores. All of these modes of

distribution, Defendants argue, compete for foodservice distribution customer spending. Based on

this market definition, Defendants assert that together, they make up approximately 25 percent of

total foodservice distribution sales. They also dispute that there is a product market for "national

customers," asserting that such a market has been created by the FTC out of whole cloth to

artificially inflate Defendants' market shares. According to the FTC, Defendants combined have,

at least, a 59 percent share of the national customer product market.

A. Broadline Distribution as a Relevant Product Market

1. Legal Principles Affecting the Definition of the Relevant Product Market

The Supreme Court in Brown Shoe set forth the general rule for defining a product market:

"The outer boundaries of a product market are determined by the reasonable interchangeability of

use or the cross-elasticity of demand between the product itself and substitutes for it." Brown

19

Shoe, 370 U.S. at 325. Stated another way, a product market includes all goods that are reasonable

substitutes, even though the products themselves are not entirely the same. Cardinal Health,

12 F. Supp. 2d at 46; Staples, Inc., 970 F. Supp. at 1074 (stating the question as "whether two

products can be used for the same purpose, and if so, whether and to what extent purchasers are

willing to substitute one for the other").

Whether goods are "reasonable substitutes" depends on two factors: functional

interchangeability and cross-elasticity of demand. "Functional interchangeability" refers to

whether buyers view similar products as substitutes. See id ("Whether there are other products

available to consumers which are similar in character or use to the products in question may be

termed 'functional interchangeability.'"). "If consumers can substitute the use of one for the other,

then the products in question will be deemed 'functionally interchangeable.'" Arch Coal, 329

F. Supp. 2d at 119; see also United States v. E.l du Pont de Nemours & Co., 351 U.S. 377, 393

(1956)) ("Determination of the competitive market for commodities depends on how different

from one another are the offered commodities in character or use, how far buyers will go to

substitute one commodity for another."). "Courts will generally include functionally

interchangeable products in the same product market unless factors other than use indicate that

they are not actually part of the same market." Arch Coal, 329 F. Supp. 2d at 119.

As for cross-elasticity of demand, there the question turns in part on price. E.l Du Pont

De Nemours, 351 U.S. at 400 ("An element for consideration as to cross-elasticity of demand

between products is the responsiveness of the sales of one product to price changes of the other.").

If an increase in the price for product A causes a substantial number of customers to switch to

product B, the products compete in the same market. See id ("If a slight decrease in the price of

cellophane causes a considerable number of customers of other flexible wrappings to switch to

20

cellophane, it would be an indication ... that the products compete in the same market."); Arch

Coal, 329 F. Supp. 2d at 120. Price is not, however, the only variable in determining the cross-

elasticity of demand between products. Cross-elasticity of demand also depends on the "ease and

speed with which customers can substitute [the product] and the desirability of doing so." FTC v.

Whole Foods Market, Inc., 548 F.3d 1028, 1037 (D.C. Cir. 2008) (Brown, J.). Thus, substitution

based on a reduction in price will not correlate to a high cross-elasticity of demand unless the

switch can be accomplished without the consumer incurring undue expense or inconvenience.

See Phila. Nat'l Bank, 374 U.S. at 358 (observing that "[t]he factor of inconvenience localizes

banking competition as effectively as high transportation costs in other industries").

Three other established principles are critical to defining the relevant product market in

this case. The first is that the "product" that comprises the market need not be a discrete good for

sale. As the Supreme Court has made clear: "We see no barrier to combining in a single market

a number of different products or services where that combination reflects commercial realities."

United States v. Grinnell Corp., 384 U.S. 563, 572 (1966); Phila. Nat'l Bank, 374 U.S. at 356

(citation omitted) (finding that "the cluster of products ... and services ... denoted by the term

'commercial banking' ... composes a distinct line of commerce"). Thus, what is relevant for

consideration here is not any particular food item sold or delivered by Defendants, but the full

panoply of products and services offered by them that customers recognize as "breadline

distribution."

Second, "the mere fact that a firm may be termed a competitor in the overall marketplace

does not necessarily require that it be included in the relevant product market for antitrust

purposes." Staples, 970 F. Supp. at 1075; Cardinal Health, 12 F. Supp. 2d at 47 (same). That is

because market definition hinges on whether consumers view the products as "reasonable

21

substitutes." Cardinal Health, 12 F. Supp. 2d at 46. So, for example, fruit can be bought from

both a grocery store and a fruit stand, but no one would reasonably assert that buying all of one's

groceries from a fruit stand is a reasonable substitute for buying from a grocery store. See Whole

Foods, 548 F.3d at 1040 (Brown, J.) ("The fact that a customer might buy a stick of gum at a

supermarket or at a convenience store does not mean there is no definable groceries market.").

Thus, as applicable here, the fact that buyers may cross-shop between modes of food distribution

does not necessarily make them part of the same market for the purpose of merger analysis.

Third, market definition is guided by the "narrowest market" principle. Arch Coal, 329

F. Supp. 2d at 120. That is, "a relevant market cannot meaningfully encompass [an] infinite range

[of products]. The circle must be drawn narrowly to exclude any other product to which, within

reasonable variations in price, only a limited number of buyers will tum." Times-Picayune Puhl g

Co. v. United States, 345 U.S. 594, 612 n.31 (1953). Judge Bates inArch Coal succinctly described

the "narrowest market" principle in practice as follows:

The analysis begins by examining the most narrowly-defined product or group of

products sold by the merging firms to ascertain ifthe evidence and data support the

conclusion that this product or group of products constitutes a relevant market. If

not, the analysis shifts to the next broadest product grouping to test whether that is

a relevant market. This process continues until a relevant market is identified.

Arch Coal, 329 F. Supp. 2d at 120; see also United States v. H&R Block, Inc., 833 F. Supp. 2d 36,

58-60 (D.D.C. 2011) (explaining "the principle that the relevant product market should ordinarily

be defined as the smallest product market that will satisfy the hypothetical monopolist test").

The critical question here, therefore, is whether broadline food distribution qualifies as the

relevant product market, or whether the product market should be expanded to include other modes

of distribution.

22

2. The Brown Shoe "Practical Indicia"

Courts look to two main types of evidence in defining the relevant product market: the

"practical indicia" set forth by the Supreme Court in Brown Shoe and testimony from experts in

the field of economics. The court turns first to the Brown Shoe factors.

According to Brown Shoe, "[t]he boundaries of [a product market] may be determined by

examining such practical indicia as industry or public recognition ... , the product's peculiar

characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity

to price changes, and specialized vendors." Brown Shoe, 370 U.S. at 325. "These indicia seem to

be evidentiary proxies for direct proof of substitutability." Rothery Storage & Van Co. v. Atlas

Van Lines, Inc., 792 F.2d 210, 218 (D.C. Cir. 1986); H&R Block, 833 F. Supp. 2d at 51. Courts

have relied on the Brown Shoe factors in a number of cases to define the relevant product market. 2

See, e.g., Staples, 970 F. Supp. at 1075-80; Cardinal Health, 12 F. Supp. 2d at 46-48; Swedish

Match, 131 F. Supp. 2d at 159-64; CCC Holdings, 605 F. Supp. 2d at 39-44; H&R Block,

833 F. Supp. 2d at 51-60.

The court finds that the Brown Shoe factors support the FTC's position that broadline

foodservice distribution is the relevant product market for evaluating the proposed merger.

As discussed below, an analysis of those factors demonstrates that other modes of foodservice

distribution are not functionally interchangeable with broadline foodservice distribution.

a. Product breadth and diversity

The most distinguishing feature of broadline distribution is its product breadth and

diversity. Broadliners stock thousands of SKUs across every major food and food-related category

2

The Brown Shoe practical indicia may indeed be "old school," as Sysco's counsel asserted at oral argument, Closing

Arg. Hr'g Tr. 44, and its analytical framework relegated "to the jurisprudential sidelines," see Whole Foods, 548 F.3d

at 1059 (Kavanaugh, J., dissenting). But Brown Shoe remains the law, and this court cannot ignore its dictates.

23

in their distribution centers. See Staples, 970 F. Supp. at 1078 (comparing SKU selections among

different sales outlets). The average Sysco or USF distribution center carries o v e r - SKUs.

Regional broadliners cany fewer SKUs than Defendants, but still maintain between 6,000 to

19,000 SKUs in their distribution centers. PX09350-215, Table 22. Broadliners also offer "private

label" products, which are a broadliner's branded products. Sysco has o v e r - private-label

SKUs, and USF has over • . PX09350-219, Table 32. This product breadth and diversity

enables broadliners to serve a wide variety of customers and to be a one-stop shop, if the customer

wishes. As USF's Executive Vice President of Strategy David Schreibman testified at the FTC's

Investigational Hearing: "[W]e have such a broad selection of SKUs because that is a key

consideration of our customer base, you have to have what they want." Investig'I Hr'g Tr.,

PX00590-006 at 24.

The other distribution channels pale in comparison to broadline in terms of product breadth

and diversity. Systems distributors cany a limited number of SKUs-usually only a few

thousand-in their distribution centers. PX09350-215, Table 22. These SKUs are ordinarily

proprietary in nature and used only by the customers for which they were developed, meaning that

systems products are not readily sellable to other customers. Specialty distributors also carry a

limited number of SKUs, usually for niche products-such as fresh produce, meat, seafood, dairy,

or bakery items-which tend to complement broadline offerings. As Sysco's CEO William

DeLaney explained: "We own [specialty] to create great traction with our customers, ... we felt

we had some gaps in our [broadline] product offerings, whether it was special produce, special cut

steaks .... " Investig'l Hr'g Tr., PX00580-010 at 38. Cash-and-carry stores likewise do not have

the same breadth and diversity of products as broadline distributors. One of the largest cash-and-

carry stores, Restaurant Depot, carries. SKUs. USF's CHEF' STORE carries less than 4,000.

24

PX09350-216, Table 26. A number of customer declarants stated that cash-and-carry store

products tended to be less uniform and inferior in quality to products carried by broadliners.

b, Distinct facilities and operations

No one entering a systems, specialty, or cash-and-carry outlet would mistake it for a

broadline distribution facility. See Staples, 970 F. Supp. at 1079 ("No one entering a Wal-Mart

would mistake it for an office superstore .... You certainly know an office superstore when you

see one."). Broadline distribution centers are massive. The average size of a Sysco distribution

center is over 380,000 square feet; for USF, it is over 270,000 square feet. Some regional

distributors also have distribution centers ranging from 200,000 to 400,000 square feet. PX09350-

215, Table 25. Non-broadline facilities are generally smaller in size and cannot readily be

converted into a broadline facility or accommodate broadline customers.

Broadline facilities also have large salesforces attached to them. Broadline facilities

typically have dozens of sales representatives, while systems distributors have few sales

representatives at their facilities. PX09350-215, Table 23. Cash-and-carry stores generally do not

have dedicated account representatives at all. Because the model of distribution is self-service,

cash-and-carry sales representatives do not learn the individualized needs of their customers in a

systematic manner.

Additional proof that broadline foodservice distribution is a separate product market comes

from the corporate structure oflarge foodservice distributors. Major foodservice distributors offer

distribution in other channels besides broadline, but they run those businesses separately from their

broadline businesses. See, e.g., H&R Block, 833 F. Supp. 2d at 56 (observing that digital do-it-

yourself tax preparation was a distinct product market from assisted tax preparation because H&R

Block ran them as "separate business units"). Sysco runs its systems distribution business,

25

SYGMA, as a separate division. So, too, does PFG, which runs a systems business known as PFG

Customized. Sysco also runs separate specialty divisions, such as Fresh Point, a fresh produce

supplier. So, too, does PFG, which has its own specialty division, Roma, which supplies Italian

restaurants and pizza parlors. And USF runs a separate cash-and-carry operation, CHEF' STORE.

This type of corporate structuring shows that those who run and manage foodservice companies

view broadline as distinct from other modes of distribution.

c. Delivery

Timely and reliable delivery is critical in the food distribution industry. Unless customers

can get the food they want when they need it, their businesses are at risk of losing clients and

money. Broadliners have the capacity-due in large part to their extensive fleet of service

vehicles, PX09350-217, Table 29-to offer frequent and flexible delivery schedules to meet

customer needs, including next-day delivery. Ample evidence shows that, for a wide array of

broadline customers-from large GPOs to individually-owned restaurants-next-day delivery is

crucial to meeting their needs.

Neither systems distributors nor cash-and-carry stores offer the same degree of frequency

and flexibility of delivery as broadliners. 3 Systems distributors tend to make large, limited-SKU

deliveries on a fixed schedule. Also, systems fleets, on average, travel longer distances than

broadline fleets to make deliveries. Carry-and-carry stores, for the most part, do not deliver.

Rather, their primary model is self-service-that is, the customer transports the merchandise on

her own. Some cash-and-carry outlets do offer delivery options. Costco, for example, offers

limited-mileage delivery from some of its stores, and Restaurant Depot leases refrigerated trucks

3

There was little evidence presented about the delivery capabilities of specialty distributors, aside from the fact that

they have a limited geographic range of delivery. See PX00427-002 (Sodexo declarant indicating that specialty

distributors covered a limited geographic range); PX00594-012 at 45 (MedAssets stating the same); PX00407-002

(Amerinet stating the same).

26

to its best customers. But those programs are quite limited and cannot substitute for the

comprehensive and flexible delivery networks offered by broadliners to all of their customers.

d. Customer service and value-added services

Another distinguishing feature of broadline distributors is their high degree of customer

service and value-added service offerings. For example, broadliners offer menu and nutritional-

meal planning services to, among others, healthcare, hospitality, and restaurant customers. They

also offer value-added services at their distribution facilities, such as food safety training and new

product updates. Other modes of delivery do not generally offer comparable value-added services.

e. Distinct customers

Due in large part to the breadth of their product and service offerings, broadliners are

capable of serving a wide range of customers, including classes of customers that the other

channels cannot reach. Systems is a more efficient and cost-effective mode of distribution for fast

food and quick service restaurants. Specialty distributors can provide higher quality and fresher

products in certain categories, but have limited product offerings and charge higher prices than

broadliners. Cash-and-carry stores are less expensive and more accessible for buyers such as

independent restaurants, but their lack of delivery service makes them unsuitable for the large

majority of foodservice customers.

These other channels, therefore, simply cannot and do not serve as wide an array of

customers as broadliners do. The largest broadline customers, such as GPOs, foodservice

management companies, and hospitality providers, cannot use systems or cash-and-carry for their

needs. They purchase only modest quantities of product from specialty distributors. Even most

independent restaurants cannot use cash-and-carry stores as a reasonable substitute for their

broadliner, even though such stores offer lower prices.

27

f Distinct pricing

Broadliners generally compete only against other broadliners on pricing. PFG's President

and CEO, George Holm, who has over 37 years of industry experience, testified that systems and

specialty distributors do not significantly affect the pricing and services that PFG offers to its

customers. Hr' g Tr. 575-76, 643. And, although broadliners recognize that cash-and-carry stores

provide lower prices, the record does not show broadliners benchmarking their prices against cash-

and-carry stores or lowering prices to compete with them. To the contrary, as USF's Executive

Vice President of Strategy David Schreibman succinctly stated in an email comparing pricing

between USF as a broadliner and its own cash-and-carry division, CHEF' STORE: "In the store,

we will be competitive with on a similar cost model. On the truck, we will be

competitive with broadline distributors on a similar cost model." PX03l14-003.

g. Industry or public recognition

Overwhelmingly, the evidence shows that players m the foodservice distribution

industry-both its suppliers and customers-recognize broadline, systems, specialty, and cash-

and-carry to be distinct modes of distribution. See Rothery Storage, 792 F.2d at 219 n.4 ("The

'industry or public recognition of the submarket as a separate economic' unit matters because we

assume that economic actors usually have accurate perceptions of economic realities."). The court

received both live and out-of-court sworn testimony from Defendants' executives; executives from

other broadline distributors; officers of non-broadline companies; and customers, large and small.

They uniformly observed that these modes of distribution are distinct in the variety of ways

described above. In short, the industry widely recognizes that broadline distributors offer a unique

cluster of products and services that is not functionally interchangeable with other modes of

distribution.

28

h, Defendants' response to Brown Shoe "practical indicia"

Defendants do not, for the most part, contest the above-described distinctions between

broadline and other channels of distribution. Instead, Defendants contend that defining the

relevant market to include only broadliners "misunderstands consumer behavior." Memo ofDefs.

Sysco Corp., USP Holding Corp. and US Foods, Inc., in Opp'n to Pis.' Mot. for A Prehm Inj.,

ECF No. 130 at 19 [hereinafter Defs.' Opp'n Br.]. They argue "customers simultaneously can,

and routinely do, choose to patronize competitors of all stripes offering fungible goods through

different but overlapping distribution channels." Id. What matters, Defendants claim, is that non-

broadliners are able to constrain a broadliner's pricing by competing for customers who are able

to move their entire purchasing, or portions of their purchasing, between channels. Id. at 19

("Whether a substitute channel is a 'comprehensive' substitute is irrelevant to that question.").

Defendants offer as one compelling example the burger chain Five Guys, which recently re-

allocated over $300 million in annual business from USP to a collection of regional broadliners

and systems distributors.

Defendants are indisputably correct that customers buy across channels, especially

independent restaurants. They are also unquestionably correct that some customers, particularly

quick service and fast food restaurant chains, are capable of moving large segments of business

from broadline to systems. But the fact that Defendants sometimes compete against other channels

of distribution in the larger marketplace does not mean that those alternative channels belong in

the relevant product market for purposes of merger analysis. See Staples, 970 F. Supp. at 1075

("[T]he mere fact that a firm may be termed a competitor in the overall marketplace does not

necessarily require that it be included in the relevant product market for antitrust purposes."); see

also Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law: An Analysis ofAntitrust Principles

29

and Their Application~ 565b (4th ed. 2014) ("[I]t would be improper to group complementary

goods into the same relevant market just because they occasionally substitute for one another.

Substitution must be effective to hold the primary good to a price near its costs[.]").

Two key decisions from this jurisdiction, Whole Foods and Staples, support this

conclusion. In Whole Foods, the question was whether there existed a product market for premium

natural and organic supermarkets ("PNOS") separate from ordinary supermarkets. The Court of

Appeals' ultimate decision was fractured-each judge issued a separate opinion, leaving no

controlling opinion from the Court. Two judges, however, concluded that PNOS is a separate

product market from ordinary supermarkets, even though there was evidence that customers

"cross-shopp[ed]" between the two. 548 F.3d at 1040 (Brown, J.); id. ("But the fact that PNOS

and ordinary supermarkets 'are direct competitors in some submarkets ... is not the end of the

inquiry."') (quoting United States v. Conn. Nat. Bank, 418 U.S. 656, 664 n.3 (1974)); id. at 1048

(Tatel, J.) ("That Whole Foods and Wild Oats have attracted many customers away from

conventional grocery stores by offering extensive selections of natural and organic products thus

tells us nothing about whether [they] should be treated as operating in the same market as

conventional grocery stores."). Both judges agreed that just because customers were able to buy

some categories of grocery products from both outlets-similar to how broadline customers are

able to purchase some products from other modes of distribution-did not mean that PNOS was

in the same product market as grocery stores. See id. at 1040 (Brown, J.) (citing testimony that

"Whole Foods competes actively with conventional supermarkets for dry groceries sales, even

though it ignores their prices for high-quality perishables"); id. at 1049 (Tatel, J.) ("As Judge

Brown's opinion explains, this suggests that any competition between Whole Foods and

30

conventional retailers may be limited to a narrow range of products that play a minor role in Whole

Food's profitability.").

The court in Staples held much the same. There, the question was whether consumable

office supplies sold by office superstores constituted a separate product market from office

supplies sold elsewhere. See Staples, Inc., 970 F. Supp. at 1073. The court acknowledged that no

matter who sells them, office supply products-to some extent, like food products-are

"undeniably the same." Id at 1075. The court nevertheless held that the sale of office supplies

through superstores constituted the relevant product market. "[T]he unique combination of size,

selection, depth and breadth of inventory offered by the superstores distinguishes them from other

retailers." Id at 1079. Those words apply with equal force to broadline distributors relative to

other food distribution channels. See also Cardinal Health, 12 F. Supp. 2d at 47 (concluding that

the wholesale drug industry "provide[s] customers with an efficient way to obtain prescription

drugs through centralized warehousing, delivery, and billing services that enable the customers to

avoid carrying large inventories, dealing with large number of vendors, and negotiating numerous

transactions").

Defendants have not convincingly distinguished Whole Foods or Staples. 4 Instead, they

urge the court to look to United States v. Sungard Data Sys., Inc., 172 F. Supp. 2d 172 (D.D.C.

4

In neither their opposition to the FTC's motion for preliminary injunction nor their proposed findings of fact and

conclusions of law do Defendants attempt to distinguish Whole Foods or Staples. At oral argument, Defendants

distinguished Staples based on the fact that in Staples the FTC had pricing data to show that prices were lower in

markets where both merging firms were present. Closing Arg. Hr' g Tr. at 38-40. Defendants also sought to distinguish

Whole Foods on the facts, arguing that in Whole Foods the defendants could not show that in the event of a price

increase consumers of PNOS could go to a standard grocery store. Id. at 40-41. But the court finds these efforts to

distinguish Staples and Whole Foods unconvincing. It is true that there was stronger pricing data in Staples, but

pricing data alone did not lead to the court's conclusion. The factual similarities between this case and Staples,

particularly the Brown Shoe practical indicia, are otherwise strong. As for Whole Foods, it is even more factually

analogous to this case than is Staples. If anything, the proof that other channels of distribution are not reasonable

substitutes for broadline is more compelling in this case than the evidence in Whole Foods that ordinary grocery stores

are not a reasonable substitute for PNOS.

31

2001), as an analogous case. There, the question was whether different types of disaster recovery

services for computer data comprised the same product market. Id. at 183. The court rejected the

government's product market definition as limited only to shared hotsite services because "the

government's market contains an extremely heterogeneous group of customers," id. at 182, who

"are simply too varied and too dissimilar to support any generalizations," id. at 193. Here, it is

unquestionably true that foodservice distribution customers are incredibly varied in their needs,

buying habits, and price sensitivities. But Sungard differs in one critical respect. The court there

observed that "the striking heterogeneity of the market, particularly as reflected by the conflicting

evidence relating to customer perceptions and practices," undercut the government's market

definition. Id. at 182-83 (emphasis added). Here, that simply is not the case. Though the

customers may be varied, the court has little doubt that the industry, from the perspective of both

sellers and buyers, perceives broadline to be a separate mode of food distribution. Witnesses of

all stripes had little trouble distinguishing among the different channels of distribution, and

Defendants offered no evidence of any industry confusion among them. Those facts make this

case fundamentally different from Sungard. See id. at 183 ("Customer responses were also often

vague and confused" and product definitions were "consistently unclear.").

Defendants also argue that the FTC's definition of broadline as the relevant market

improperly excludes other modes based on "a small number of customers' subjective preferences

for broadline distribution." Defs.' Opp'n Br. at 17 (footnote omitted). But the evidence, as it

relates to broadline versus other distribution channels, is hardly selective. Defendants' own

executives acknowledged the fundamental differences between broadline and other modes of

32

distribution. 5 So, too, did executives of regional broadliners, such as PFG, 6 Sharnrock,7 Reinhart

Foodservice, 8 and Shetakis9 ; consortiums, such as UniPro 10 ; systems distributors, such as

Maines 11 ; and cash-and-carry stores, such as Restaurant Depot. 12 Likewise, customers of every

size recognized the differences between broadline and the other food distribution modes. In short,

this is not the kind of case in which the testimonial evidence failed to demonstrate a consensus

among the industry's players regarding the boundaries of the product market.

3. Expert Testimony

Having concluded that the Brown Shoe "practical indicia" support a product market for

broadline foodservice distribution, the court turns next to the second type of evidence that courts

consider in product market definition: expert testimony in the field of economics. One of the

primary methods used by economists to determine a product market is called the "hypothetical

monopolist test." This test asks whether a hypothetical monopolist who has control over a set of

substitutable products could profitably raise prices on those products. If so, the products may

comprise the relevant product market. See H&R Block, 833 F. Supp. 2d at 51-52. The theory

behind the test is straightforward. If enough consumers are able to substitute away from the

5

See, e.g., DX-00319 at 32-36 (Sysco's CEO, William DeLaney, explained that systems is a "tailored, customized

approach to certain types of customers" and the "model is not to serve GPO customers"); Hr' g Tr. 1369-70 (DeLaney

stated that, compared to cash-and-carry, broadline is a "value package" that includes delivery services and menu

consulting); Hr'g Tr. 1452 (David Schreibman ofUSF stated that "specialty distributors compete by having a broader

array of products within their expertise" that "broadliner[s] may not have in [their] portfolio"); Investigat'l Hr'g Tr.,

PX00580-008-010 at 32-39 (DeLaney explained that broadline and specialty are "two different businesses," whereas

broadline distribution includes "a full range of products"); Investigat' l Hr' g Tr., PX00584-060 at 239-40 (Louis Nasir,

the Pacific Market President for Sysco, maintained that cash-and-carry stores "don't have the same selection" of

products and "also don't have consistent inventory" compared with broadliners); Investigat'l Hr'g Tr., PX00590-0l l

at 42 (Schreibman stated that he was not aware of a cash-and-carry store that delivers).

6

See PX00429-002-007; Hr'g Tr. 571-73.

7

DX-00285 at 115-16, 164-66.

8

DX-00295 at 16-17, 22.

9

PX004l4-001.

10

DX-00260 at 139.

11

DX-00264 at 64, 141; PX00424-001 (Maines is predominantly systems, butl percent of2013 revenues were from

broadline sales).

12

DX-00314 at 146-47.

33

hypothetical monopolist's product to another product and thereby make a pnce mcrease

unprofitable, then the relevant market cannot include only the monopolist's product and must also

include the substitute goods. On the other hand, if the hypothetical monopolist could profitably

raise price by a small amount, even with the loss of some customers, then economists consider the

monopolist's product to constitute the relevant market.

The hypothetical monopolist test, which courts have applied, is set forth in the

U.S. Department of Justice and FTC's Horizontal Merger Guidelines. See U.S. Dep't of Justice &

FTC Horizontal Merger Guidelines § 4.1.1 (2010) [hereinafter Merger Guidelines]; H&R Block,

833 F. Supp. 2d at 51-52; CCC Holdings, 605 F. Supp. 2d at 40; Arch Coal, 329 F. Supp. 2d at 120

& n. 7. As stated in the Merger Guidelines:

[T]he test requires that a hypothetical profit-maximizing firm, not subject to price

regulation, that was the only present and future seller of those products ... likely

would impose at least a small but significant and non-transitory increase in price

("SSNIP") on at least one product in the market, including at least one product sold

by one of the merging firms.

Merger Guidelines§ 4.1.1. The SSNIP "is intended to represent a 'small but significant' increase

in the prices charged by firms in the candidate market" and is typically assumed to be "five percent

of the price paid by customers for the products or services to which the merging firms contribute

value." Merger Guidelines§ 4.1.2.

As applied to this case, the hypothetical monopolist test asks: If there was only one

broadline food distributor, could it profitably raise price by five percent, or would that price increase

result in a substantial number of customers moving enough of their spend to other modes of

distribution-systems, specialty, or cash-and-carry-such that the price increase would be

unprofitable? If the price increase would be profitable, then the relevant product market is broadline

distribution; if unprofitable, it means that the relevant market must include at least one other channel

34

of distribution. Each side presented expert testimony from economists who performed the

hypothetical monopolist test but who came to different results.

a. Dr. Mark Israel

For its expert economic evidence, the FTC presented the testimony of Dr. Mark Israel, who

received a doctorate in economics from Stanford University and now serves as Executive Vice

President at Compass Lexecon, a consulting firm. Dr. Israel's testimony served two primary

functions. First, he acted as a de facto summary witness, synthesizing the mass of testimonial and

documentary evidence gathered by the FTC. Dr. Israel's summary of that evidence parallels the

discussion in the above sub-sections, so the court does not revisit it here. Second, Dr. Israel

conducted a SSNIP test, using what is known as an "aggregate diversion analysis." Its purpose is

to determine the amount of sales that a hypothetical monopolist of broadline distribution could

lose before a price increase becomes unprofitable. See Swedish Match, 131 F. Supp. 2d at 160

(describing the related methodology of"critical loss analysis"); H&R Block, 833 F. Supp. 2d at 63

(same). A detailed recitation of Dr. Israel's aggregate diversion analysis is necessary because

Defendants challenge the basic elements of his work.

Aggregate diversion analysis has three basic steps. The first is to determine the threshold

aggregate diversion ratio, which is the percentage of customers that would need to stay within the

broadline market to make a price increase profitable. See H&R Block, 833 F. Supp. 2d at 63. This

is strictly a mathematical step, with the aggregate diversion ratio a function of the subject product's

gross margin. The gross margin is defined as the price of selling one additional product minus the

cost of selling the additional product. 13 The second step is to determine the actual aggregate

diversion-that is, the actual percentage of customers of a single broadliner that would switch to

13

Gross margin is calculated as follows: (Revenue-Cost of Goods Sold)/Revenue.

35

another broadliner after a price increase. "Since these lost sales are recaptured within the proposed

market, they are not lost to the hypothetical monopolist." Id. As will be seen, this step involved

an analysis of Defendants' actual sales data. The final step is to compare the two: if the actual

aggregate diversion is greater than the threshold ratio, then the hypothetical monopolist could

profitably raise prices and the candidate market is the relevant product market. See id. In other

words, as applied here, if the percentage of customers of a single broadliner who would switch to

another broadliner (as opposed to another mode of distribution) in response to a price increase is

greater than the percentage of customers needed to stay within the market to make a price increase

profitable, then the relevant product market is properly defined as broadline distribution.

At step one of his aggregate diversion analysis, Dr. Israel assumed a gross margin of

10 percent, a figure lower than the gross margin contained in the parties' financial reporting. 14

A 10 percent gross margin, according to Dr. Israel, yields a 50 percent threshold aggregate

diversion ratio based on a formula devised by two economists, Michael Katz and Carl Shapiro. 15

Next, Dr. Israel calculated the actual aggregate diversion based on three different data sets.

He constructed the first two data sets from national and regional requests for proposals ("RFPs")

and "bidding" summary information and documents produced by each Defendant to the FTC.

Based on this information, Dr. Israel built a database for each company that tracked, for each

bidding opportunity, the incumbent distributor, the winning distributor, and the competing bidders.

PX09350-104. Based on Sysco's RFP/bidding data, he found that, when Sysco lost a bid,

14

Dr. Israel testified that the parties' reported gross margins are between 15 and 20 percent, but to be conservative he

used a 10 percent margin. Hr'g Tr. 1004-05.

15

The Katz-Shapiro formula that Dr. Israel used is L =XI(){+ M), where L is the aggregate diversion ratio, or "critical

loss," X is the price increase, and M is the margin. PX09350-055 at n.134. For his aggregate diversion analysis,

Dr. Israel used a 10 percent price increase and a 10 percent margin, for a resulting critical loss of 50 percent, i.e., .50

= .10/(.10 + .10). Hr'g Tr. 1004-07.

36

broad.liner; the remaining losses were to another mode of distribution. PX09350-056. Based on

USF's RFP/bidding data, the percentage was even higher-USF lost to other broadliner-

Htffl of the time. Id.

Dr. Israel constructed his third data set from USF's "Linc" database. Linc is a customer

relations management tool that USF local sales representatives used until recently to track sales

opportunities. The Linc database contains fields that sales representatives can complete to describe

a sales opportunity, including a "main competition" field. Dr. Israel assumed that, if USF did not

win an opportunity, it was won by the identified "main competitor." The Linc database contained

hundreds of thousands of observations, about a third of which included information on the "main

competitor."

opportunities lost by USF (again, based on potential revenue of those sales opportunities) were

lost to other broadliners. PX09350-056.

At the third step, Dr. Israel compared the aggregate diversion ratio of 50 percent to the

actual diversion percentages derived from the three data sets. He concluded that, because each of

the three actual diversion percentages was higher than the 50 percent threshold aggregate diversion

ratio, broadline distribution was the relevant product market. In other words, Dr. Israel found that

only 50 percent of broadline customers would need to remain within the broadline market to make

a price increase profitable, while according to three different data sets, the actual percentage of

customers who would remain within the broadline market (by switching to another broadliner) was

greater than 50 percent. Therefore, Dr. Israel's calculations indicated that broadline distribution

was the relevant product market.

37

b, Defendants' experts

Defendants mounted an aggressive challenge to Dr. Israel's work through their own expert

witnesses. Defendants first presented Dr. Jerry Hausman, a professor of economics at

Massachusetts Institute of Technology. Dr. Hausman testified, in short, that Dr. Israel's aggregate

diversion analysis was wrong because (i) he used the wrong gross margin and (ii) he used the

wrong mathematical formula to calculate the threshold aggregate diversion ratio. According to

Dr. Hausman, Dr. Israel excluded certain variable costs from his gross margin. The actual gross

margin was not 10 percent, according to Dr. Hausman, but between II percent and II percent.

Also, Dr. Hausman testified that the aggregate diversion formula Dr. Israel used was incorrect and

led to an overly narrow market definition. 16 Using the proper margins and the correct formula,

Dr. Hausman opined, the aggregate diversion ratio is not 50 percent, but rather over 100 percent,

which is an impossibility (i.e., more than 100 percent of customers cannot switch in response to a

price increase). Thus, he concluded, the relevant product market is not broadline, but all channels

of food distribution.

While Dr. Hausman challenged Dr. Israel's calculation of the threshold aggregate diversion

ratio, Defendants' other expert, Dr. Timothy Bresnahan, a professor of economics at Stanford

University, critiqued Dr. Israel's use of the RFP/bidding and Linc data sets to calculate the actual

aggregate diversion. Regarding the RFP/bidding data, Dr. Bresnahan described the data as

contrived and unreliable-a point that Defendants consistently articulated to the FTC during the

investigation phase. Dr. Bresnahan explained that the companies do not keep comprehensive RFP

16

According to Dr. Hausman, the correct formula is L = XIM, where L is the aggregate diversion ratio, or "critical

loss," Xis the price increase, and Mis the margin. Dr. Hausman testified that this is the more appropriate formula in

an asymmetric market, like food distribution, which involves suppliers and customers with different costs, different

types of customers, and a different mix of products. Hr'g Tr. 1960-64; DFF at 285-86 (citing to DX-05028 at 11).

The formula used by Dr. Israel, on the other hand, is more appropriate in a symmetric market, that is, a market marked

by homogeneity among suppliers and customers. Hr'g Tr. 1960, 1965-66; DX-05028 at 10-11.

38

or bidding data in the ordinary course of business and that the information Dr. Israel relied upon

was pulled together at the insistence of the FTC, in part based on employees' unreliable notes and

memories. As for the Linc data, it too was flawed, Dr. Bresnahan suggested, because it is a

prospective sales database, not an actual transactions database in which USF sales personnel were

accurately recording wins and losses. Moreover, neither the RFP/bidding data nor the Linc data

describes whether Sysco or USF lost a customer for a price-based reason or some reason having

nothing to do with price.

c. The court's finding as to the expert testimony

Having weighed the competing expert testimonies and considered them in light of the

evidentiary record as a whole, the court finds Dr. Israel's aggregate diversion analysis and

conclusion to be more persuasive than that advanced by Defendants' expert, Dr. Hausman. 17

Dr. Israel's reliance on the RFP/bidding and Linc data sets for calculating the aggregate diversion

is problematic for the reasons Defendants have identified and, for those reasons, the court hesitates

to rely on Dr. Israel's precise aggregate diversion percentages. But, when evaluated against the

record as a whole, Dr. Israel's conclusions are more consistent with the business realities of the

food distribution market than Dr. Hausman's. See Cardinal Health, 12 F. Supp. 2d at 46 (stating

that "the determination of the relevant market in the end is 'a matter of business reality-[ ] of

how the market is perceived by those who strive for profit in it."' (alteration in original) (quoting

FTC v. Coca-Cola Co., 641 F. Supp. 1128, 1132 (D.D.C. 1986), vacated as moot, 829 F.2d 191

(D.C. Cir. 1987)); Arch Coal, 329 F. Supp. 2d at 116 ("[A]ntitrust theory and speculation cannot

17

In finding Dr. Israel's conclusion more persuasive than that advanced by Defendants' expert, the court might be

doing more than it is required to do. As Judge Tatel stated in Whole Foods: "Although courts certainly must evaluate

the evidence in section 13(b) proceedings and may safely reject expert testimony they find unsupported, they trench

on the FTC's role when they choose between plausible, well-supported expert studies." Whole Foods, 548 F.3d at

1048 (Tatel, J.).

39

trump facts[.]"); H&R Block, 833 F. Supp. 2d at 65 (bearing in mind the shortcomings of the

expert's analysis and treating the analysis as "another data point" in determining the relevant

market, rather than as conclusive).

The court finds Dr. Hausman's conclusion-that the actual aggregate diversion ratio is

greater than 100 percent-inconsistent with business reality. On cross-examination, Dr. Hausman

admitted that his conclusion meant that a hypothetical monopolist who had control over every

single broadline distributor in the country could not profitably impose a SSNIP on customers,

because enough customers would switch to other channels of distribution. Hr'g Tr. 2003-04. Yet

many industry leaders testified either that other channels of distribution did not constrain the prices

charged by broadliners or that other channels were not substitutes for broadline distribution. For

instance, PFG's President and CEO, George Holm, testified that systems and specialty distributors

do not significantly affect the pricing and services that PFG's broadline division offers to its

customers. Hr' g Tr. 575-76. He also testified that systems and specialty distributors were not

substitutes for broadliners. Hr'g Tr. 573. Such evidence from industry leaders, 18 which the court

credits, contradicts Dr. Hausman's conclusion that a hypothetical monopolist ofbroadline services

would not be able to impose a SSNIP because enough customers would switch to other channels

of distribution.

18

See also PX00429-004-007 (George Holm, President and CEO of PFG, explaining that systems, specialty, and cash-

and-carry distributors are not substitutes for customers needing broadline distribution); DX-00285 at 125-26 (John

Roussel, COO of Shamrock Foods, stating that it's "not possible" or "practical" for a broadline customer to use a

systems distributor); DX-00260 at 139 (Bob Stewart, interim CEO ofUnipro, explaining that a broadline customer

cannot easily switch to a systems distributor and a broadline customer's needs are different than a systems customer's

needs).

40

4. Conclusion as to the Broadline Product Market

In conclusion, based on the vast record of evidence the parties have presented, the court

finds that the FTC has carried its burden of demonstrating that broad.line distribution is the relevant

product market.

B. National Broadline Distribution as a Relevant Product Market

The FTC asserts that, within the broader product market for broad.line distribution, there is

a narrower but distinct product market for "broad.line foodservice distribution services sold to

National Customers." CompI. if 44. According to the FTC, "[ d]ue to [their] geographic dispersion,

National Customers typically contract with a broadline foodservice distributor that has distribution

centers proximate to all (or virtually all) of their locations." Id. if 42.

National Customers typically contract with a broadliner that can provide--across

all of their locations-product consistency and availability, efficient contract

management and administration (e.g., centralized ordering and reporting, a single

point of contact, and consistent pricing across all locations), volume discounts from

aggregated purchasing, and the ability to expand geographically with the same

broadline foodservice distributor.

Id. National customers include healthcare GPOs; foodservice management companies; and large

hotel and restaurant chains. Id if 41. The FTC contends that Sysco and USF "are the only two

single-firm broadline distributors with national geographic reach and, as such, are best positioned

to serve National Customers." Id. if 63.

Defendants vigorously dispute that there is such a thing as a "National Customer." They

contend that a product market built around so-called national customers is "contrived,'' Defs.'

Opp'n Br. at 16, and that the FTC's distinction between national and local customers is "factually

and economically meaningless,'' id at 13. They counter that the national-local distinction is not,

as the FTC claims, built on differentiating customer characteristics, but is improperly based on an

administrative distinction as to whether the customer prefers to be managed at the corporate level

41

(making it a "national" customer) or at the local distribution center (making it a "local" customer).

Id. at 12-15. The so-called national customer category, they also argue, is improperly based on a

"few core customers who say they prefer the merging parties." Id. at 13. In addition, Defendants

assert that Dr. Israel did not perform a SSNIP test to assess the existence of a national customer

market. Id. at 12.

!. Legal Basis for Defining Relevant Product Market Based on Customer Type

Before turning to the evidence, the court first considers the legal basis for defining a

product market based on a type of customer. Neither side comprehensively addressed this issue.

Admittedly, defining a product market based on a type of customer seems incongruous. After all,

one ordinarily thinks of a customer as purchasing a product in the market, and not as the product

market itself But, in this case, according to the FTC, the national customer and broadline product

converge to define a market for broadline products sold to national customers. Broadline

distributors must offer a particular kind of "product"-a cluster of goods and services that can be

delivered across a broad geographic area-to compete for national customers. In that sense, the

customer's requirements operate to define the product offering itself

The clearest articulation of this approach to product market definition comes from the

Merger Guidelines. The Merger Guidelines are not binding, but the Court of Appeals and other

courts have looked to them for guidance in previous merger cases. See, e.g., Heinz, 246 F.3d at

716 n.9; H&R Block, 833 F. Supp. 2d at 52 n.10. Section 4.1.4 of the Merger Guidelines provides

that "[i]f a hypothetical monopolist could profitably target a subset of customers for price

increases, the Agencies may identify relevant markets defined around those targeted customers, to

whom a hypothetical monopolist would profitably and separately impose at least a SSNIP."

Merger Guidelines § 4.1.4. Markets to serve targeted customers are also known as "price

42

discrimination markets." Id. Professors Areeda and Hovenkamp have endorsed market definition

of this kind, as well: "Successful price discrimination means that the disfavored geographic or

product class is insulated from the favored class and, if the discrimination is of sufficient

magnitude, should be counted as a separate relevant market." 2B Phillip E. Areeda & Herbert

Hovenkamp, Antitrust Law: An Analysis ofAnt;trust Prindples and Their Application~ 534d (3d

ed. 2007). The concern underlying price discrimination markets is that certain types of captured

or dedicated customers could be targeted for monopolist pricing even if a price increase for all

customers would not be profitable. See Merger Guidelines § 3; Areeda & Hovenkamp 3d ed.,

supra,~ 533d ("[S]ellers may be able to discriminate against buyers who have fewer alternatives

or for whom the product performs a more valuable function[.]").

Defining a market around a targeted customer, as the FTC urges here, is not free from

controversy, as the different opinions in Whole Foods demonstrate. 19 Relying on an earlier version

of the Merger Guidelines that recognized price discrimination against "targeted buyers,"

Judge Brown explained that "core consumers"-in that case, those committed to premium and

natural organic supermarkets-"can, in appropriate circumstances, be worthy of antitrust

protection." Whole Foods, 548 F.3d at 1037 (Brown, J.) (citing DOJ and FTC, 1992 Horizontal

Merger Guidelines§ 1.12, 57 Fed. Reg. 41,552, 41,555 (1992)). Judge Brown went on to say:

In particular, when one or a few firms differentiate themselves by offering a

particular package of goods or services, it is quite possible for there to be a central

group of customers for whom "only [that package] will do." ... Such customers

may be captive to the sole supplier, which can then, by means of price

discrimination, extract monopoly profits from them while competing for the

business of marginal customers.

19

The FTC cites to the "distinct customers" factor in Brown Shoe as support for defining a market around a targeted

customer. However, Brown Shoe only listed "distinct customers" as one of many factors for courts to consider in

defining a market. Brown Shoe, 370 U.S. at 325. It did not endorse defining a market around a group of targeted

customers.

43

Whole Foods, 548 F.3d at 1038 (Brown, J.) (quoting Grinnell, 384 U.S. at 574) (alteration in

original).

Judge Kavanaugh, in dissent, rejected defining a market around a "core customer." Whole

Foods, 548 F.3d at 1062 (Kavanaugh, J., dissenting). According to Judge Kavanaugh, "there is

no support in the law for that singular focus on the core customer. Indeed, if that approach took

root, it would have serious repercussions because virtually every merger involves some core

customers who would stick with the company regardless of a significant price increase. " 20 Id. The

relevant question for market definition, according to Judge Kavanaugh, is not whether a die-hard

group of core customers would be impacted by a substantial price increase, but whether the merged

company "could increase prices by five percent or more without losing so many marginal

customers as to make the price increase unprofitable." Id.

2. Evidence Supporting a National Broadline Product Market

Ultimately, the court here need not resolve the Whole Foods disagreement over defining a

market around a "core" customer. That is because the ordinary factors that courts consider in

defining a market-the Brown Shoe practical indicia and the Merger Guidelines' SSNIP test-

support a finding that broadline distribution to national customers is a relevant product market.

See, e.g., Areeda & Hovenkamp 3d ed., supra,~ 533d ("If the defendant can profit by charging

pharmacies a price significantly over its cost, then the pharmacy sales are a relevant market[.]").

20

The Merger Guidelines do not, for instance, set forth how a court is to distinguish a "targeted" group of customers

from customers in general. This gives rise to the question of what limiting principles or factors a court should apply

in defining a price discrimination market. Absent limitations, price discrimination against a single customer might be

used to justify blocking a merger. This is not a mere theoretical possibility. According to the Merger Guidelines, "[i]f

prices are negotiated individually with customers, the hypothetical monopolist test may suggest relevant markets that

are as narrow as individual customers." Merger Guidelines § 4.1.4 (emphasis added).

44

a. Industry and public recognition

Among the most compelling evidence supporting a product market for national customers

is the fact that regional broadliners have formed cooperatives, such as DMA and MUG, to compete

for customers with a geographically dispersed footprint. Regional distributors, because of their

limited footprints, do not have the capacity to serve customers with multi-regional needs across all

of their locations. Only Sysco and USF have that capacity. These cooperatives were formed

specifically to compete against Sysco and USF, by enabling regional competitors to combine to

provide nationwide or multi-regional delivery and, importantly, to offer a single point of contact

for the customer. Dan Cox, the President and CEO ofDMA, explained that DMA was formed in

1988 as a competitive response to Sysco' s merger with another company, Continental. See

PX00565-051 at 202. He explained that "[w]hen that industry event took place, it was the first

time that there was truly a national platform for foodservice distribution." Id. Put simply, business

ventures like DMA would not exist if there were not a separate market for customers who have

national or multi-regional distribution needs. See Rothery Storage, 792 F.2d at 218 n.4 (stating

that courts must "assume that economic actors usually have accurate perceptions of economic

realities").

Equally compelling evidence of the national-local distinction comes from a report done by

the management consulting firm, McK.insey & Co., whom Sysco hired to assist with merger

integration. After closely analyzing the two companies' operations, McKinsey prepared a

presentation in July 2014, titled "National, Intermediate, and Field Coverage Models." The

presentation observed that "Sysco and US Foods have different approaches to grouping customers

and determining service models .... Both companies effectively operate two service models with

distinct capabilities to serve two types of customers." PX09010-002 (emphasis added). The

45

presentation described "National Customers" as those who "use complex contracts with margin

schedules, make online purchases of proprietary products, require auditing support, and coordinate

across multiple markets." Id. By contrast, "Field Customers" were those who "make weekly

purchases through in-person consultations, receive specialist support tailored to independent

restaurants, require minimal auditing support, and operate in I or few markets." Id. McKinsey

further observed that national customers' "requirements" included "[ s]et margin schedule

contract[s]"; "[e]fficient ordering across multiple locations"; "[l]arge number[s] of deviated,

proprietary and close-coded products"; "[r]egulatory and audit support"; "[i]n-depth reporting";

and "[c]onsistency of service, pricing and products across multiple [m]arkets." PX09010-004.

Field customers' "requirements," on the other hand, included the "[a]bility to make decisions each

week along with consultation"; "[a]ccess to national, commodity, and some proprietary products";

"[f]ull business, culinary, and product support for independent businesses"; and "minimal"

"[c]oordination across geographies." Id. McKinsey ultimately recommended that the companies

recognize and build a new service model around a third kind of customer-an "Intermediate"

customer-who would be identifiable based on five variables: (i) national contract/no contract;

(ii) nature of industry; (iii) number of markets; (iv) number of regions; and (v) size of annual sales.

PX09010-007. The McKinsey presentation identified as "conclusively" national those customers

who operate in three or more markets or two or more regions. Id.

McKinsey is not the only industry analyst or expert to acknowledge that national customers

form a market distinct from local buyers. Cleveland Research Company, an investment research

firm, produced an analyst report on Sysco after the merger's announcement and recognized that

Sysco and USF serve a distinct group of national customers. One of the report's conclusions was

that "Sysco/USP will [be] able to keep most of their larger contracted and national account

46

customers for the near- and medium-term due to national scale and existing contracts .... Based

on our research, most national operators prefer to deal with one distributor because it is more

efficient and less expensive than dealing with several regional players." PX09332-006 (emphasis

added).

The industry's trade group, the International Food Distributors Association ("IFDA"), also

recognizes a distinction between national and local customers. IFDA produces a Quarterly

Operations survey that reports separate sales figures for "national" and "street" accounts.

PX00570-004 at 78. IFDA's President, Mark Allen, explained that IFDA distinguishes between

the two because "the dynamics between the two [types of] businesses might be a little bit different.

The operating metrics might be a little bit different." Id. at 80.

Defendants' ordinary course documents also recognize the national-local distinction and

tout their strategic advantage as to the former. See H&R Block, 833 F. Supp. 2d at 52 ("When

determining the relevant product market, courts often pay close attention to the defendants'

ordinary course of business documents."). A Sysco "Investor Day" presentation from 2010

distinguishes the company's "Contract Sales (Broadline)" from "Street Sales," PX03101-010, and

separates its "Key Competitors - National," from regional competitors, PX03101-020. Similarly,

a presentation entitled "Board of Directors Strategy Sessions," dated July 2010, distinguishes

between Sysco's market size for "corporate contracts"-defined to include "major foodservice

management (FSM) sales, major group purchasing organization (GPO) sales, and major chain

sales (non FSM or GPO)"-and "Street" business. PX01008-006.

USF has similar documents. An internal USF presentation, titled "Business Overview,"

describes "[USF's] Customers" as falling into three categories: (i) "Street: Independent restaurants

or small local chains"; (ii) "National Accounts: Contracted customers located across the country,"

47

including acute and long-term healthcare facilities, hotels and the hospitality industry, schools, and

US. military and government agencies; and (iii) "National Chain Restaurants: Fast food and

quick-serve establishments." PX03122-004. See also PX03034-006 (similarly categorizing the

company's customers). A USF "Investor Presentation" from November 2012 describes USF as

the "2nd largest national broadline distributor," PX03000-006, and touts its "[a]bility to leverage

our national scale to cost effectively service customers nationally," PX03000-014. Further, it

distinguishes between "National Scale," where "US Foods is the second-largest broadline

foodservice distributor in the US.," and "Local Scale," where "US Foods is estimated #1 or #2

position in II of served markets," PX03000-014. See also PX03007-007 (internal document in

which KKR & Co., one of USF's private equity owners, distinguishes between "Street and

National Account customer segments").

Other key players in the industry also recognize that national customers are different.

For instance, the President and CEO of PFG, George Holm, agreed that "Sysco and US Foods are

the only two distributors for broadline with the capability to serve national broadline customers

with locations dispersed throughout the United States," including foodservice management

companies, GPOs, large healthcare systems, and certain restaurant chains. Hr'g Tr. 596.

Representatives of DMA and Reinhart likewise referred to national customers as those that are

geographically dispersed and need a single point of contact. See PX00412-002-003; PX00415-

004.

b. Distinct customer needs

There is ample record evidence that national customers' needs differ from those of local

customers. The McKinsey analysis described above concisely summarized those distinctions.

PX09010-004.

48

For starters, national customers, because of their dispersed geographic presence, often

require a broadliner to meet their foodservice needs in more than one region. As a result, the

number of distribution centers in a broadliner's network is often an important factor for such

customers. In sharp contrast, according to Sysco, "all, or almost all," of its "local contract

customers" are served by only one distribution center. PX01400-001.

The Defendants' ordinary course documents highlighted their comprehensive distribution

networks as a competitive advantage for serving national customers. See, e.g., PX03000-014 (USF

presentation touting its "[a]bility to leverage our national scale to cost effectively service

customers nationally"); PX00247-001-002 (USF email communication t o - describing

the "US Foods Value Proposition" as including "Privately held National Distribution footprint

company"; "Single IT operating platform nationally"; and a "Single Point of Contact"); PXO 1062-

005 (Sysco presentation to Aramark highlighting that Sysco' s "national footprint, strong service

approach and our breadth of product offerings is what differentiates us from our competition").

As USF's David Schreibman acknowledged during the evidentiary hearing, "US Foods['] leading

national market position is due to US Foods['] geographic presence that includes 62 distribution

centers across the United States." Hr'g Tr. 1520-21. He also acknowledged that Sysco was the

only company with greater scale than USF. Id. at 1522.

In addition to multi-regional distribution capabilities, national customers generally demand

a set margin contract that applies across multiple locations. As PFG' s George Holm testified, a

single contract enables customers to simplify contract administration and to reduce administrative

costs. Id. at 600-02. Additionally, national customers often use RFPs and/or bilateral negotiations

to award broadline foodservice distribution contracts. Id. at 1595-97. In sharp contrast, pricing

for local or "street" customers, according to Sysco, "[is] ultimately the result of individual

49

negotiations between the customer and [broad.liner]" and "can vary on a weekly and even daily

basis." PX06057-032.

National customers also seek a single technology platform for handling their purchases.

Consolidating purchasing through a single ordering platform creates efficiencies and cost savings,

particularly as it relates to managing direct contracts with manufacturers and administering price

changes. The importance of this feature is evidenced by DMA's development of a single ordering

platform that enables customers to purchase from its members. Indeed, DMA promotes its

technology platform as superior to Sysco's and USF's. PX00565-006 at 23-24. If national

customers had not demanded such a feature, DMA would not have developed it.

Finally, product consistency is a factor for some national customers, particularly for those

who wish-to-purchase private label products. See PX09010-004 (McKinsey report identifying as

a "Customer requirement[]" for "National" customers "consistency of service, pricing, and

products across multiple Markets"). Large customers can achieve a high degree of product

consistency through direct contracting with product manufacturers or by purchasing proprietary

brands stocked by Defendants. DX-01359 at 73 (Dr. Bresnahan report observing that "one way

customers that value consistency achieve it is through direct negotiation with manufacturers to

create propriety products" and that "[c]ustomers can also rely on national brands to ensure

consistency"). However, because private label goods offer a strong value benefit, if a national

customer wishes to purchase such goods and have them available across all of its locations, it can

do so most efficiently through a broad.liner with national geographic scope. See Hr'g Tr. 600

(George Holm of PFG stating that one reason national customers prefer to contract with Sysco or

USF is that "[w]here they have a preference for a private brand, []it is the same product [across]

their system").

50

c, Defendants' Operations

Both Sysco and USF operate dedicated sales groups from their national headquarters that

are responsible for negotiating and managing contracts with customers who use multiple

distribution centers. See Grinnell, 384 U.S. at 572-74 (holding that centralized station security

services operated on a national level is a relevant product market). Sysco refers to these customers

as "corporate multi-unit customers," or CMUs. USF refers to them as "national sales customers."

According to USF's Senior Vice President for National Sales, Tom Lynch, each national customer

in his group has a single USF representative who is responsible for that customer. The largest

customers are assigned a full-time dedicated employee to manage the account. PX00517-014-015

at 56-58.

d. SSNIP Test

Contraiy to what Defendants contend, Dr. Israel did perform a SSNIP test to determine

whether there is a separate product market for national customers. That SSNIP test was performed

as an element of the SSNIP test that Dr. Israel used to assess whether broadline distribution was a

relevant product market. As Dr. Israel testified, he applied to national customers the same

10 percent gross margin that he used to calculate the aggregate diversion ratio for all customers.

Hr'g Tr. 1005 (stating that he used a 10 percent gross margin "to both local and national

customers"). He derived the actual diversion for national customers based on the RFP/bidding

data provided by the defendant companies. Id. at 1009 (describing the "RFP/bidding data" as

"really national [customer] data"). Using the same methods discussed above, Dr. Israel calculated

USF' s national customers to be !pf§iltfJ:'I. In other words, o v e r - of the time (based

on potential revenue from sales opportunities), when Sysco or USF lost a bid opportunity for a

51

national customer, it was to another broadliner. Because these percentages were greater than the

aggregate diversion ratio of 50 percent, Dr. Israel concluded that broadline service to national

customers was a relevant market. In other words, Dr. Israel found that only 50 percent of national

broadline customers would need to remain within the broadline market to make a price increase

profitable, while the actual percentage of national customers who would remain within the

broadline market (by switching to another broadliner) was greater than 50 percent. Dr. Israel's

calculations, therefore, indicated that broadline distribution to national customers was the relevant

product market.

The court already has expressed its reservations about relying on the RFP/bidding data to

precisely calculate the aggregate diversion ratio. But, as before, the court finds that the ultimate

conclusion of the SSNIP test-that broadline foodservice to national customers is a relevant

product market-is supported by the weight of the evidence. Numerous national customer

witnesses testified that other channels of distribution were not adequate substitutes for broadline

distribution. 21 Although Defendants have shown that some national customers who were served

by broadliners are now served by systems or systems-like distributors-most notably, Subway and

Five Guys-those are the exceptions. Subway and Five Guys, because of their limited menus, are

more amenable to substituting to a systems model. The same simply cannot be said of other large

national customers, like GPOs, foodservice management companies, and hospitality chains, which

rely heavily on broadliners.

21

See Hr'g Tr. 143-145 (Christine Szrom, fact witness for U.S. Department of Veteran Affairs, explaining that she is

not familiar with systems distribution and could "absolutely not" use a cash-and-carry distributors); Hr' g Tr. 214-17

(James Thompson, Head of Procurement for Interstate Hotels and Resorts, stating that "it would be very difficult if

not impossible" to operate Interstate's foodservice distribution without a broadliner and that specialty is not a

substitute for broadline distribution); PJ<llll-002 (Joan Ralph, Group Vice President at Premier, Inc., saying that

" [e] ven if we choose one day to contract with systems distributors, specialty distributors, or cash and carry stores,

each would be as an additional, distinct service for our members who may ne_e~ck, last-minute item or two; none

~ce or serve as a substitute for broadline distribution services"); P~-002 ~

- - , noting t h a t - cannot contract with a systems distributor or use other f~

52

e, Defendants' arguments against a national customer market

Asserting that there is no separate product market for national broadline customers,

Defendants first argue that the national-local distinction is "arbitrary" because it is based on

nothing more than customer preference about account management. Defendants' executives

testified that Sysco's CMU customers and USF's national customers are so designated, not because

of any particular characteristic or group of characteristics, but purely because the customer prefers

to have its account managed by the headquarters sales team, instead of by its local distribution

center. The FTC's and Dr. Israel's reliance on the companies' administrative designation,

Defendants argue, leads to arbitrary classifications. For example, some of Defendants' customers

who use a small number of distribution centers are counted by the FTC as "national" customers.

As Dr. Hausman demonstrated, 37 percent of Sysco's CMU customers use five or fewer

distribution centers and 55 percent use ten or fewer. And, for USF, 51 percent of their national

customers use five or fewer distribution centers and 67 percent use ten or fewer. Hr' g Tr. 1976.

Additionally, similarly situated customers-in terms of size, number of distribution centers,

revenues, etc.-are sometimes treated differently. One customer may be identified as national and

another as local, simply because one prefers to be managed from headquarters and the other from

the local distribution center.

Defendants are correct that their "national" customer lists are over-inclusive-not every

customer on those lists has multi-regional distribution needs. And they are also correct that the

FTC could have more accurately defined a class of "national" customers by testing each candidate

national customer against specific "national" criteria, such as the number of distribution centers

used. But, ultimately, for the purpose of defining a product market, the court finds that the parties'

53

"national" customer designation is a useful proxy for customers requiring geographically dispersed

distribution and attendant services.

As the graphic below prepared by Dr. Israel shows, if the merger were to occur, a

significant proportion of the combined company's national customer revenues would come from

customers who use a large number of distribution centers. PX09375-077, Figure 3. National

customers using more than 3 5 distribution centers would account forl percent of a merged Sysco-

USF' s revenue; national customers using more than 24 distribution centers would account for I

percent of revenue; and national customers using at least 10 distribution centers would account for

IIpercent of revenue. Those figures demonstrate that Defendants' national-customer designations

capture those key customers (based on revenues) who use a large number of distribution centers.

The "national" designation includes, among others, the largest GPOs, like Premier, Novation, and

MedAssets, each of whom uses over I distribution centers; the largest foodservice management

companies, like Sodexo, Aramark, and Compass, each of whom uses more than I distribution

centers; the largest hotel management company, Hilton, which uses I distribution centers; and

the second largest government customer, the U.S. Department of Veterans Affairs, which uses I

distribution centers (the largest is the U.S. Department of Defense, which uses I distribution

centers). PX09375-076, Table 5. Thus, for these customers, the label "national" is not merely

administrative; it accurately reflects this high revenue-generating group's actual needs. The fact

54

that some smaller customers are included among the Defendants' ·'national'' designations does not

mean that the desig;1atio11 lacks evidentia1y value for defining a market for national customers.

Figure 3

SyM:o and l'SF 2013 Re\eUUt'S IJ,v l'\uml>er ufDistriuution Centen rwd

Next, Defendants assert that defining a pnce discrimination market around national

customers is untenable because the FTC failed to show that so-called national customers shared

any objectively observable characteristics that would enable t11e combined company to price

discriminate against that group. See Merger Guidelines § 3 (stating that "differential pricing" is

an essential element of price discrimination, which "may involve" offering different pricing to

different types of customers "based on observable characteristics"), In other worcls, they argue

that this grouping of customers is so heterogeneous that there is no collllllon, identifiable

55

characteristic that could serve as a proxy for determining which customers in the broadline market

have inelastic demand.

Defendants are undoubtedly correct that, even among their largest customers, there is great

variety in the customers' servicing needs and requirements. But price discrimination can occur

even when customers do not have common observable characteristics. As the Merger Guidelines

state, markets for targeted customers may exist "when prices are individually negotiated and

suppliers have information about customers that would allow a hypothetical monopolist to identify

customers that are likely to pay ahigher price forthe relevant product." Merger Guidelines§ 4.1.4;

see also Carl Shapiro, The 2010 Horizontal Merger Guidelines: From Hedgehog to Fox in Forty

Years, 77 Antitrust L.J. 49, 93 (2010) (observing that, in markets for intermediate goods and

services, "prices typically are negotiated and price discrimination is common").

Here, the evidence is clear that Defendants engage in individual negotiations with their

national customers and possess substantial information about them. Indeed, the fact that

Defendants employ substantially more sales representatives than other broadliners, PX093 50-218,

Table 30, and assign full-time dedicated employees to some of their largest customers is indicative

of the "know-your-customer" philosophies of both firms. Defendants, therefore, already have

substantial customer information that would allow them to predict which of their customers have

inelastic demand and which do not. Price discrimination can occur in such a marketplace, even if

the targeted customers do not share specific identifiable traits.

Finally, Defendants contend that a product market of targeted national customers does not

comport with business realities. This argument has two main elements. First, they assert that,

contrary to what the FTC contends, Compl. iii! 5, 42, national customers do not require a broadline

foodservice distributor that is national in scope. Rather, they argue, even at current prices, many

56

large customers spread their distribution needs over multiple regional suppliers. For instance,

Defendants cite GPOs, l i k e - · - ' Amerinet and large government agencies. like

the Defonse Logistics Agency, as using a regional contracting approach. Defs.' Opp'n Br. at 15.

They also refer to one of the largest foodservice management companies, Sodexo. which splits its

distribution into I regions. Id. A. nd. then there is Subway and Five Guys. two large chain

restaurants that have regionalized and purchase from multiple suppliers. Id. at 15-16. Because

these tyves of customers can regionalize or credibly threaten to regionalize. Defendants argue, the

merged company would not be able to discriminate against them on price.

But Defendants' argument founders when faced with the actual purchasing habits of the

industry's largest customers. The evidence shows that the bulk of the broadline purchasing done

by most geographically dispersed breadline customers is still done through Sysco and USF.

Of Avendra's members' breadline spend, I percent is witl1 Sysco and USF. PL 's CoITected

Proposed Findings of Fact and Conclusions of Law. ECF No. 173 at 114 [hereinafter PFF].

Members of other GPOs similarly purchase a large percentage of their goods from Sysco and USF.

The total breadline spend of Premier,22 Novation, MedAssets, and HPSI members with Sysco and

USF is, respectively, I percent, II percent I percent, and I percent Id. at 113-15: FTC

Closing Arg. Slides at 35. Large foodservice management companies similarly make the bulk of

their broadline purchases from Sysco and USF. Sodexo, Aramark, Compass, and Centerplate,

respectively, spend I percent, I percent, I percent, and I percent of their broa<lline

foodservice distribution dollars with Sysco and USF. PFF at 113-16: FTC Closing Arg. Slides at

35. The story is similar for large hospitality customers. Two of tl1e largest, Hilton and Interstate,

57

allocate I percent and I percent of their broadline spend, respectively, to the two companies.

PFF at 114, 116; FTC Closing Arg. Slides at 35. Even the Defense Logistics Agency, which

contracts regionally, dedicates I percent of its broadline spend to Sysco and USF. PFF at 116;

FTC Closing Arg. Slides at 35.

The court infers from this evidence that geographically dispersed customers view Sysco

and USF as having significant comparative advantages over regional distributors, particularly

because of their far-reaching distribution networks. Though some customers have spread their

business over multiple broadliners, a significant portion (as measured by total revenues) have not.

Indeed, PFG's George Holm observed that the "clear trend amongst national broadline customers

is to move toward a single nationwide provider." Hr'g Tr. 598 (emphasis added); PX09081-002

(letter from PFG's counsel to FTC, dated November 14, 2014, stating the same). See Brown Shoe,

370 U.S. at 332 (footnote omitted) ("Another important factor to consider is the trend toward

concentration in the industry."). Mr. Holm further admitted that either Sysco or USF essentially

wins every RFP issued by a national customer. Hr' g Tr. 598-99. And PFG acknowledged by letter

to the FTC that, even as the country's third-largest broadliner, "PFG has difficulty competing for

national broadline accounts because it does not have a nationwide footprint of broadline

distribution centers." PX09081-001. Other large regional broadliners have said the same about

their own businesses models. 23 Defendants' contention-that a product market defined around

national customers does not comport with business reality because such customers have

regionalized or can regionalize-is thus belied by the record evidence.

23

See, e.g., PX00415-004 (Reinhart); PX00416-003 (Merchants); PX00434-003-004 (Labatt); PX00438-002-003

(Cash-Wa); PX00443-005 (Ben E. Keith); PX00449-003 (Jacmar); PX00451-005 (Services Group of America);

PX00458-004 (Nicholas & Co.); PX00460-002-003 (Shamrock); PX00529-047-048 at 188-89 (Gordon).

58

Second, Defendants argue that margin data shows that, as a merged entity, they would not

be able to price discriminate against national customers. Dr. Hausman demonstrated that

Defendants' margin on sales to customers who use fewer distribution centers is actually higher

than their margin on sales to those who use more. DX-01355 at 58-61. Defendants contend that

under the FTC's theory, they presently have a duopoly as to national customers, yet they do not

earn duopoly profits on that customer class. Defendants thus maintain that, just as they cannot

today price discriminate to earn duopoly profits, they would not be able to price discriminate after

the merger to earn monopoly profits.

Defendants' argument, however, is unconvincing. Defendants' present inability to earn

duopoly profits on national customers is probably because large customers can keep prices down

by leveraging the defendant companies against one another. As the Cleveland Research Company

observed: "Based on our research, we believe both Sysco and US Foods have priced each other

down competing for larger national/regional contract accounts over the last several years."

PX09332-004. The ability of large buyers to keep prices down, functioning as what is known in

antitrust literature as "power buyers,'' see Cardinal Health, 12 F. Supp. 2d at 58-59; Merger

Guidelines § 8, depends on the alternatives these large buyers have available to them, see Shapiro,

supra, at 95; Areeda & Hovenkamp 3d ed., supra,~ 943a. If a merger reduces alternatives, the

power buyers' ability to constrain price and avoid price discrimination can be correspondingly

diminished. See Merger Guidelines § 8 ("Normally, a merger that eliminates a supplier whose

presence contributed significantly to a buyer's negotiating leverage will harm that buyer."). Thus,

the fact that Defendants are currently unable to price discriminate against national customers does

not mean that they would be unable to do so as a merged firm.

59

C. Product Market Summary

Having considered and weighed the parties' arguments and evidence, the court concludes

that the FTC has carried its burden of showing that, for purposes of merger analysis, (i) broadline

foodservice distribution is a relevant product market, and (ii) broadline foodservice distribution to

national customers is also a relevant product market.

D. Relevant Geographic Market

The court now turns to the second part of defining the relevant market, which involves

determining the relevant geographic market. The Supreme Court has stated that, for Section 7 of

the Clayton Act, the relevant geographic market is "the area in which the goods or services at issue

are marketed to a significant degree by the acquired firm." Marine Bancorp., 418 U.S. at 620-21.

Stated differently, "[t]he proper question to be asked ... [is] where, within the area of competitive

overlap, the effect of the merger on competition will be direct and immediate." Phi/a. Nat. Bank,

374 U.S. at 357; see also Cardinal Health, 12 F. Supp. 2d at 49 (citation omitted) (internal

quotation marks omitted) (stating that the relevant geographic market is "the area to which

consumers can practically turn for alternative sources of the product and in which the antitrust

defendants face competition"). Like the product market, the geographic market must "correspond

to the commercial realities of the industry and be economically significant." Brown Shoe, 370

U.S. at 336-37 (footnote omitted) (internal quotation marks omitted). The Supreme Court has

recognized that an "element of 'fuzziness would seem inherent in any attempt to delineate the

relevant geographical market,"' and therefore "such markets need not-indeed cannot-be defined

with scientific precision." Conn. Nat. Bank, 418 U.S. at 669 (quoting Phi/a. Nat 'l Bank, 374 U.S.

at 360 n.37). That said, the relevant geographic market "must be sufficiently defined so that the

60

[c]ourt understands in which part of the country competition is threatened." Cardinal Health,

12 F. Supp. 2d at 49.

The FTC contends that there are two relevant geographic markets in this case. For national

broadline customers, the relevant geographic market is nationwide. For local broadline customers,

the relevant geographic markets are localized around Defendants' distribution centers.

With regard to national customers, for essentially the same reasons that the FTC asserts

that there is a product market for broadline distribution to national customers, the FTC asserts that

the geographic market for those customers is nationwide. The FTC relies on the fact that

Defendants plan on a national level and have "national account" teams dedicated to national

customers; their contractual pricing and service terms with national customers apply across

regions; and their competition for national customers is largely other broadliners with nationwide

coverage.

As for local customers, as discussed in more detail below, the FTC's local geographic

markets were constructed by Dr. Israel and are premised on customers' proximity to Defendants'

distribution centers. The basic idea is that, for local customers, distance to a distribution center is

a key service factor and, for Defendants, distance traveled from a distribution center to make

deliveries is a critical cost component. The FTC alleges that the merger threatens to harm

competition in 32 local geographic markets where Sysco and USF together currently have

dominant market shares. Compl. ~ 60.

Defendants dispute that there is a nationwide geographic market for the same reasons that

they contend that there is no national customer product market. As for the local geographic

markets, Defendants aggressively challenge the methodology that Dr. Israel used in defining local

markets. Their primary criticism is that the geographic areas are drawn so narrowly that they

61

exclude actual competition from the relevant market. This results, they contend, in local market

concentrations that artificially inflate Defendants' market shares.

l. National Market

Although the physical act of delivering food products occurs locally, for national customers

the relevant geographic area for competitive alternatives is nationwide, primarily because of their

geographically dispersed footprint. Defendants compete within this market by touting their

nationwide distribution capabilities to these customers; bidding against other broadliners with

multi-regional capabilities (which is to say, against each other and the regional cooperatives);

coordinating the marketing, negotiating, and managing of these customers through their "national

account" teams; and entering with these customers into a single contract whose terms, including

pricing, apply across regions. For these reasons, the court finds that the relevant geographic market

for broadline foodservice to national customers is nationwide. See Grinnell, 384 U.S. at 575-76

(finding a national geographic market where central station services "operated on a national level,''

and there was "national planning," a nationwide schedule of prices, and nationwide contracting

for multi-state businesses); Cardinal Health, 12 F. Supp. 2d at 50 (finding a national geographic

market where evidence showed that "GPOs negotiate contracts with several wholesalers, making

the same prices available throughout the country to all of their members-local, regional, or

national").

2. Local Markets

Defining the local geographic market presents a far greater challenge. Not surprisingly,

there is no industry standard for delineating the area that makes up a local geographic market for

broadline distribution. Each local market has its own unique attributes. Customer composition

and concentration differs across markets; so does the demand for products, with SKU variations

62

reflecting local tastes and palettes. Average driving distances for foodservice distributors vary

depending on the density of the area, with longer hauls more common in rural parts of the country

and shorter trips more prevalent in urban areas. And, of course, the competitors vary from market

to market.

The FTC tasked Dr. Israel with defining the local geographic markets. He constructed

them as follows. In his first step, Dr. Israel drew circles around the location of each Sysco and

USF distribution center. To determine the size of each circle, Dr. Israel used a radius, referred to

as the "draw distance," that, on average, captured 75 percent of the distribution center's sales to

local customers. The length of each distribution center's 75 percent draw radius differed. For

example, the 75 percent draw distance around Sysco's Billings, Montana, facility was 262 miles,

whereas the 75 percent draw distance around Sysco's Jersey City, New Jersey, facility was only

24 miles. PX09350-221-224, Table 38. What that means is Sysco drives over 200 miles further

to capture 75 percent of its local sales in Billings than it does in Jersey City. That disparity makes

sense, as more populated areas correspond to higher customer concentrations and shorter delivery

distances.

In his second step, Dr. Israel identified each company's local customers that fell within an

area of intersection between the draw circle around the Sysco distribution center and the draw

circle around the USF distribution center. This area of intersection was termed the "overlap area."

These "overlap customers," according to Dr. Israel, were the customers most likely to suffer harm

from the merger, because these were the customers who would be left with one less alternative

supplier after the merger. Exhibit 40 from Dr. Bresnahan's report, which is reproduced below,

shows Dr. Israel's methodology in the Omaha, Nebraska, area. The blue-dotted circle corresponds

63

to Sysco's 75 percent draw area, and the green-dotted circle corresponds to USF's. The dark gray

area corresponds to the "overlap customers." DX-01359, Ex. 40.

EXHIBIT 40

OISTRlBUTION CENTERS LOCATED NEAR THE FTC'S CONTESTED LOCAL AREAS

OMAHA, NE/COUNCIL BLUFFS, IA

@ MaJO~ tkuad\r.e.r

{fl' US Fc~?j-s BrnadlF1e 0 MaJf..i[ 6ruadlaHJ~' w~~i; Sa~es f}a:a

'W>W l5% USF Draw l\r~a \'DQ);t rrnt~ radiu!>J

In his third step, Dr. Israel identified the broadline distributors who could compete for the

customers in the overlap area. To do this, Dr. Israel drew circles around each overlap customer

using the 75 percent draw radius. This created a larger circle that moved the outer boundaries of

the overlap area by the same radius as the 75 percent draw area, which is represented by the light

gray area in Exhibit 40 above. According to Dr. Israel's analysis, the light gray area is the area to

which customers can practically tum for alternative sources of broadline distribution. All of the

64

competitors located within the light gray area were factored into Dr. Israel's local market share

computations.

Defendants attack Dr. Israel's "circle drawing exercise" as "arbitrary" and not reflective of

industry realities. Defs.' Opp'n Br. at 27. Specifically, they assert that Dr. Israel's methodology

is flawed because it assumes that competitors will drive no greater distance than Sysco's or USF's

75 percent draw radius to serve customers. Defendants point to competitor declarations and

testimony showing that in many of the 32 local markets in which the FTC claims Defendants have

a dominant market share, competitors are willing to, and do, drive distances greater than the

75 percent draw radius to compete for and deliver to customers.

Notwithstanding this criticism, the court finds that there is nothing inherently "arbitrary"

about Dr. Israel's methodology in defining the local markets. To the contrary, given the absence

of an industry standard for defining a local market, Dr. Israel's methodology provides a practical

approach and solution to an otherwise thorny problem. Dr. Israel's premise in defining these

markets-that driving distance matters-is amply supported by the record and common sense.

Customers who are farther away from a distribution center cost more to service. Longer distances

correspond to, among other things, higher gas usage, more labor hours, and increased wear and

tear on trucks. Given that the geographic market need not be defined by "metes and bounds,"

Conn. Nat 'l Bank, 418 U.S. at 669 (citation omitted) (internal quotation marks omitted),

Dr. Israel's 75 percent draw methodology identifies "the area of competitive overlap, [where] the

effect of the merger on competition will be direct and immediate," Phila. Nat'l Bank, 374 U.S. at

357. See also Conn. Nat'l Bank, 418 U.S. at 670 n.9 (remanding to the district court to define the

local market and observing that the "federal bank regulatory agencies define a bank's service area

as the geographic area from which the bank derives 75% of its deposits"). The court therefore

65

concludes that the relevant local geographic markets are the areas of overlap resulting from

Dr. Israel's 75 percent draw methodology.

Ultimately, what really troubles Defendants about Dr. Israel's "circle drawing exercise" is

not the resulting geographic areas, but what those areas mean for calculating Defendants' local

market shares. The court considers those arguments in the next section.

II. THE PROBABLE EFFECTS ON COMPETITION

Having concluded that the FTC has carried its burden of establishing a relevant market-

both a nationwide market for broadline foodservice to national customers and various local

markets for broadline foodservice to local customers-the court turns next to "the likely effects of

the proposed [merger] on competition within [those] market[s]." SwedishMatch, 131 F. Supp. 2d

at 166. As the Court of Appeals explained in Heinz, the government "must show that the merger

would produce 'a firm controlling an undue percentage share of the relevant market, and [would]

result[ ] in a significant increase in the concentration of firms in that market."' 246 F.3d at 715

(quoting Phi/a. Nat'/ Bank, 374 U.S. at 363). "Such a showing establishes a 'presumption' that

the merger will substantially lessen competition." Id. (citation omitted).

The Court of Appeals has held that the FTC can establish its prima facie case by showing

that the merger will result in an increase in market concentration above certain levels. Id. "Market

concentration is a function of the number of firms in a market and their respective market shares."

Arch Coal, 329 F. Supp. 2d at 123. A common tool used to measure changes in market

concentration is the Herfindahl-Hirschmann Index (HHI). Heinz, 246 F.3d at 716; see also Merger

Guidelines § 5.3. HHI figures are "calculated by summing the squares of the individual firms'

market shares," a calculation that "gives proportionately greater weight to the larger market

shares." Merger Guidelines§ 5.3. "Sufficiently large HHI figures establish the FTC's prima facie

66

case that a merger is anti-competitive." Heinz, 246 F.3d at 716. The Merger Guidelines, which

provide "a useful illustration of the application ofHHI," FTC v. PPG Indus., Inc., 798 F.2d 1500,

1503 n.4 (D.C. Cir. 1986), state that a market with an HHI above 2,500 is considered "highly

concentrated"; a market with an HHI between 1,500 and 2,500 is considered "moderately

concentrated"; and a market with an HHI below 1,500 is considered "unconcentrated," Merger

Guidelines§ 5.3. Furthermore, a merger that results in "highly concentrated markets that involve

an increase in the HHI of more than 200 points will be presumed to be likely to enhance market

power." Id. In Heinz, the Court of Appeals recognized that an increase in HHI by 510 points

"creates, by a wide margin, a presumption that the merger will lessen competition." 246 F.3d at

716.

A. Concentration in the National Broadline Customer Market

1. Dr. Israel's National Broadline Customer Market Shares Calculations

In some cases the merging parties' market shares and post-merger HHis are seemingly

uncontroversial. See, e.g., Staples, 970 F. Supp. at 1081-82; H&R Block, 833 F. Supp. 2d at 71-

72. Not so here. Because there are no industl)'-recognized market shares for national broadline

customers, the FTC tasked Dr. Israel with calculating the market shares and the HHis. Not

surprisingly, Defendants vigorously contested his methodology and conclusions.

Dr. Israel calculated Defendants' national customer shares as follows. As his first step, he

identified Defendants' individual sales to national broadline customers, i.e., the numerator for the

market share calculation. Those sales figures came directly from the parties' "national" customer

designations: for Sysco, its sales to CMU customers, and for USF, its sales to national customers.

Next, Dr. Israel determined the total sales by all broadline distributors to national

customers, i.e., the denominator for the national share calculation. Again, because there is no

67

industry-recognized figure for such sales, Dr. Israel estimated them. He did so in two ways. First,

he aggregated the national sales of the three principal competitors for national customers-Sysco,

USF, and DMA-and added in another share equal to DMA's. This total comprised the

denominator for his "baseline" shares calculation. PX09350-074. The addition of another DMA-

sized share to the denominator was premised on his observation from the RFP/bidding data that

the size of sales to national customers by all broadliners other than Sysco, USF, and DMA was

about the same as DMA's.

Dr. Israel also used a second method to calculate the total sales to national customers. He

aggregated the national sales reported by the largest 16 broadliners, including DMA and MUG, in

response to the FTC's civil investigative demands. This data is referred to as CID data. Dr. Israel

ran several "sensitivities" on this sum, adding in sales to account for variations in CID responses

(e.g., some distributors did not segregate "national" from total sales). Dr. Israel also aggregated

the national sales of Sysco, USF, DMA, and MUG, plus an estimate of national sales for all other

responding distributors based on the assumption that each distributor's national-local sales ratio

was the same as Defendants' ratio. Dr. Israel's various approaches yielded a total national

broadline sales estimate of $28 to $30 billion. Hr'g Tr. 1177-78; see also PX09060-006 (PFG

business plan estimating the size of the national customer market to be approximately $20 billion).

As his last step, Dr. Israel adjusted his market shares to account for the divestiture to PFG.

The chart below reflects Dr. Israel's post-merger, post-divestiture market share and HHI

calculations. For his "baseline" calculation, Dr. Israel determined that the parties' post-merger

national broadline customer market share would be 71 percent with an HHI increase of nearly

2,000 points. His CID data-based calculations, shown as (i) through (vi) in the chart, also yielded

high post-merger shares and significantly increased HHis. Dr. Israel's most conservative

68

approach, in which he assumed that the top 16 broadliners had national to local sales ratios that

were equal to Defendants' ratio of such shares-( iv) in the chart below-resulted in a post-merger

market share of 59 percent and an Hill increase of 1,500 points. PX09350-186, Table 18.

Table 18

Shares of Sales to ::Xational Broadline Cu.<;tomer1;, After Accounting for the Proposed

Diwstiture

Poo;t-Di.vestiture Shares Post-Div.:'Stitme HJ-Il's

l-Eil 2 H'r'.J

Ba'ieline ·n °o )Jl9 1.966

(i) N mional 0

6S o 4.935 l.953

!ii) National~ Imputed National 4.549 1.799

{iii) National - Regional 66% 4.614 1.822

(i\"} ?\mional-;- sy~tenb 4,_l,...,i

~

1.643

(v) Nmional + Reg1om1l + System<; 61°0 4.oii7 1590

{vi) Partie~· Raiil> of:National 59°0 3.809 1500

2. Defendants' Arguments

Defendants raise a host of objections to the reliability of Dr. Israel's methodology and

calculations. They contend that his use of their "national" sales in the numerator was arbitrary

because, as discussed above, not all of Defendants' "national" sales are to customers with a multi-

regional footprint. The inclusion of those sales, they contend, overstated Defendants' national

market share. They also argue that Dr. Israel's numerator included some sales to systems-like

customers, such as to Five Guys, but his denominator excluded competitors' systems sales. This

asymmetry, they assert, also resulted in an overstatement of Defendants' share. They further

contend that the denominator used in Dr. Israel's "baseline" calculation is unreliable because it

relies on the flawed RFP/bidding data set. And, finally, they argue that the denominator in the

CID data calculation excludes over $30 billion in sales-though the source of this number is

69

unclear. 24 They contend that these errors in developing the numerator resulted in biased market

share calculations.

None of these arguments ultimately persuade the court that Dr. Israel's methodology or his

market shares and HHI calculations are unreliable. The FTC need not present market shares and

HHI estimates with the precision of a NASA scientist. The "closest available approximation"

often will do. PPG, 798 F.2d at 1505 (citation omi

This text is long and has been trimmed here. Open the source document for the complete record.