Amicus Curiae Brief — Ohio v. Am. Express Co., 138 S. Ct. 355 (2017) (No. 16-1454)

No. 16-1454

IN THE

Supreme Court of the Anited States

STATE OF OHIO, et al.,

Petitioners,

v.

AMERICAN EXPRESS COMPANY, et al.,

Respondents.

ON WRIT OF CERTIORARI TO THE UNITED STATES

COURT OF APPEALS FOR THE SECOND CIRCUIT

BRIEF FOR THE STATES OF NEW YORK, ALASKA,

CALIFORNIA, DELAWARE, HAWAI‘1, INDIANA, KENTUCKY,

MAINE, MASSACHUSETTS, MINNESOTA, MISSISSIPPI,

NEW MEXICO, NORTH CAROLINA, OREGON,

PENNSYLVANIA, SOUTH CAROLINA, WASHINGTON,

AND WISCONSIN, AND THE DISTRICT OF COLUMBIA,

AS AMICI CURIAE IN SUPPORT OF PETITIONERS

ERIC T. SCHNEIDERMAN

Attorney General

State of New York

BARBARA D. UNDERWOOD*

Solicitor General

BEAU W. BUFFIER STEVEN C. WU

Bureau Chief Deputy Solicitor General

ELINOR R. HOFFMANN JUDITH N. VALE

Deputy Bureau Chief Senior Assistant

JEREMY R. KASHA Solicitor General

Assistant Attorney General 120 Broadway

Antitrust Bureau New York, NY 10271

(212) 416-8020

barbara. underwood@ag ny gov

*Counsel of Record

(Addit ional counsel listed on signature page)

i

QUESTION PRESENTED

Section 1 of the Sherman Act prohibits

unreasonable restraints of trade. The petitioner

States and the United States (together, the

government“) allege that respondent American

Express Company (“Amex”) violated § 1 through anti-

steering rules, which prohibit merchants from

encouraging Amex cardholders to use different, lower-

fee credit cards. Applying the rule of reason, the

district court held that the government had proven a

prima facie violation by establishing that the anti-

steering rules imposed anticompetitive effects (such as

a price increase) on merchants. On appeal, the Second

Circuit reversed, holding that to prove a prima facie

violation the government had to show not only

anticompetitive effects on the merchant side but also

insufficient procompetitive benefits on the cardholder

side. The question presented is:

Under the rule of reason, did the district

court's finding that Amex’s anti-steering

provisions stifled price competition on the

merchant side of Amex’s credit-card platform

suffice to prove anticompetitive effects and

thereby shift to Amex the burden of

establishing any procompetitive benefits from

the anti-steering rules?

ul

TABLE OF CONTENTS

Page

INTEREST OF AMICI CURIAE . . 1

s. adieanninansbokaamnesindatiodiniations 3

JJ rr 4

B. The Government's Antitrust Complaint

GEER SS AA—A————— eC 6

C. The District Court’s Opinion 7

D. The Second Circuit’s Opinion 12

SUMMARY OF ARGUMENT . eee 13

ET dicniaetbichcmtdincsetsditnincreninhsentsensetsbmcsienens 17

I. The Traditional “Reasonable

Interchangeability” Rule for Market

Definition Applies to Two-Sided Platforms...... 17

A. The Services That Amex Sells to

Merchants Are a Separate Market from

the Services It Sells to Cardholders. s 17

B. Limiting the Relevant Market to One

Side of a Platform Does Not Preclude

Consideration of Impacts on the Other

. III ccccsnacssesnecsnsceccesuseoncones 18

1. The relevant market is one side of the

platform, but the court may consider

the competitive effects on that market

caused by a restraint’s effects on the

other side of the platform 20

2. Courts should not consider a

restraint’s asserted benefit on the

other side of a platform unless it

causes a procompetitive effect in the

EXE 22

iii

Page

ILA Plaintiff Satisfies Its Prima Facie Burden

by Demonstrating an Anticompetitive

Effect—Such as a Price Increase—on One

Side of a Two-Sided Platform 27

A. A Plaintiff Should Not Be Required to

Predict and Disprove All Possible

Procompetitive Benefits to Satisfy Its

Prima Facie Burden of Demonstrating an

Anticompetitive Effect 28

B. Maintaining the Proper Assignment of

Burdens Ensures That Potentially

Procompetitive Conduct Is Not Chilled by

Antitrust Enforcement Efforts. .................. 32

— c —j—ç— — 34

iv

TABLE OF AUTHORITIES

Cases Page(s)

Bell Atl. Corp. v. Twombly, 550 U.S. 544 (2007)...... 31

Board of Trade of Chicago v. United States, 246

Gee ree ee 20

California Dental Ass’n v. FTC, 526 U.S. 756

EEA 7,28,31,33

Eastman Kodak Co. v. Image Tech. Servs., Inc.,

II 19

Federal Trade Commission b. Actavis, Inc., 133

N. CC 29

Federal Trade Commission v. Indiana Fed n of

Dentists, 476 U.S. 447 (1986) . . 7

Geneva Pharm. Tech. Corp. v. Barr Labs. Inc.,

386 F.3d 485 (2d Cir. 2004) ees 7

Leegin Creative Leather Prods., Inc. v. PSKS,

pm! 8S Fok, eee 20,28

National Collegiate Athletic Ass’n v. Board of

Regents of the Univ. of Okla., 468 U.S. 85

c 7,28, 29,30

National Soc y of Pro Engers v. United States,

II . 20,23

Times-Picayune Publ’g Co. b. United States, 345

Deer 24

United States v. Charlotte- Mecklenburg Hosp.

Auth., 248 F. Supp. 3d 720 (W. D. N. C. 2017) 27

United States v. E.. du Pont de Nemours & Co.,

III 17

United States v. Grinnell Corp., 384 U.S. 563

dd 17

— . 17,21

Page(s)

United States v. Topco Assocs. Inc., 405 U.S. 596

1 25,26

United States v. Visa U.S.A., Inc., 163 F. Supp.

e 23

United States v. Visa CU. S. A., Inc., 344 F. 3d 229

e 18

Miscellaneous Authorities

Areeda, Phillip E. & Herbert Hovenkamp,

Antitrust Law: An Analysis of Antitrust

Principles and Their Application vol. 7 (3d ed.

e 8.31.32

Areeda, Phillip E., Herbert Hovenkamp & John

L. Solow, Antitrust Law: An Analysis of

Antitrust Principles and Their Application

EE 17

Armstrong, Mark, Competition in Two-Sided

Markets, 37 RAND J. of Econ. 669 (2006)............. 4

Carlton, Dennis W., & Alan S. Frankel,

Transaction Costs, Externalities, and “Two-

Sided” Payment Markets, 2005 Colum. Bus. L.

e 24

Delegation of the U.S. to the Competition

Committee, Roundtable on Two-Sided

Markets: Note by the Delegation of the United

States, OECD No. DAF/COMP/W1(2009)68

(2009), at https://www.ftc.gov/sites/default/

files/attachments/us-submissions-oecd-and-

other-international-competition-fora/

roundtabletwosided. pd 4.5.24

vi

Page(s)

Filistrucchi, Lapo, Damien Geradin & Eric van

Damme, Identifying Two-Sided Markets

(Tilburg Univ., TILEC Discussion Paper No.

2012-008 2012), at

http://ssrn.com/abstract=200866 1 .................. 24,25

Hesse, Renata B., & Joshua H. Soven, Defining

Relevant Product Markets in Electronic

Payment Network Antitrust Cases, 73

. 18

Hovenkamp, Herbert, The Rule of Reason (Univ.

of Penn. Law Sch. Inst. for Law and Econ.,

Research Paper No. 17-28 2017), at

http://ssrn.com/abstract=28859 16 ........ 28,29,31,33

N.Y. Att'y Gen., Press Release, A.G.

Schneiderman Announces Multi-State

Settlement with NFL Permanently Barring

League-Wide Mandatory Ticket Price Floor

(Nov. 15, 2016), at https://ag.ny.gov/press-

release/ag-schneiderman-announces-multi-

state-settlement-nfl-permanently-barring-

—T—T—TT—T—T—T—————— 5 1

N. V. Att'y Gen., Press Release, A. G.

Schneiderman Announces Settlement with

Nation’s Largest Mall Operator To Stop

Anticompetitive Tactics at Woodbury

Common Outlet Center (Aug. 21, 2017), at

https://ag.ny.gov/press-release/ag-

schneiderman-announces-settlement-nations-

largest -mall- operator - sto 2

Rochet, Jean-Charles, & Jean Tirole, Platform

Competition in Two-Sided Markets, 1 J. Eur.

1 ccc 5,6

vii

Page(s)

Rochet, Jean-Charles, & Jean Tirole, Two-Sided

Markets: A Progress Report, 37 RAND J. of

Rysman, Marc, The Economics of Two-Sided

— — 4-6,24

Werden, Gregory J., Antitrust’s Rule of Reason:

Only Competition Matters, 79 Antitrust L.A.

PE —ↄ—I—I—:, — ̃— 28

INTEREST OF AMICI CURIAE

This appeal raises important legal questions about

how to apply the rule of reason to alleged anticompe-

titive conduct involving a “two-sided” platform—..e.,

an intermediary that brings together two distinct

groups of its customers to interact with each other.

Amici States of New York, Alaska, California,

Delaware, Hawai'i, Indiana, Kentucky, Maine,

Massachusetts, Minnesota, Mississippi, New Mexico,

North Carolina, Oregon, Pennsylvania, South

Carolina, Washington, Wisconsin, and the District of

Columbia have a direct interest in ensuring that they

may continue to effectively enforce antitrust laws to

preserve vigorous competition in industries that may

involve two-sided (or multisided) platforms.

The States have been increasingly active in

enforcing the antitrust laws in sectors that may

involve two-sided platforms, which have become more

common with the rise of computer technology and the

internet. For example, New York, Florida, Massachu-

setts, Ohio, Pennsylvania, and the District of

Columbia entered into a settlement with the National

Football League resolving antitrust claims arising

from the NFL’s policy of requiring all teams to impose

a price floor on secondary tickets sold over NFL-

sanctioned ticket-exchange websites—two-sided plat-

forms that unite ticket buyers with ticket owners who

want to resell their tickets.“ And New York recently

|! N.Y. Att'y Gen. Press Release, AG. Schneiderman

Announces Multi-State Settlement with NFL Permanently

Barring League-Wide Mandatory Ticket Price Floor (Nov. 15,

2016) (internet). (For sources available on the internet, URLs are

listed in the table of authorities.)

2

entered into a settlement with Simon Property Group,

the operator of an outlet center that connects retailers

and shoppers, to prohibit Simon from using its

contracts to restrict retailers from opening a nearby

store—an anticompetitive restraint that thwarted the

development of competing outlet stores.”

Amici States recognize that the economics of two-

sided platforms may differ from the economics of

traditional buyer-seller relationships, and that such

economic differences should be properly taken into

account in any application of traditional antitrust

principles. But the court of appeals erred here in its

approach to two-sided platforms, distorting bedrock

antitrust law based on the mistaken belief that it was

required to do so to account for the interdependent

nature of the two sides of the platform at issue—a

credit-card network that connects merchants and

cardholders. First, the court of appeals erroneously

defined the relevant market as the entire platform,

combining into a single market the services Amex sells

to merchants and the services Amex sells to card-

holders—even though the well-established law on

market definition would deem these two types of

services to be in separate markets. Second, the court

of appeals incorrectly required the government to

demonstrate a net anticompetitive effect across the

entire platform at the prima facie stage, rather than

an anticompetitive effect (such as a price increase) on

just one side of the platform.

2 N.Y. Att'y Gen., Press Release, AG. Schneiderman

Announces Settlement with Nations Largest Mall Operator To

Stop Anticompetitive Tactics at Woodbury Common Outlet

Center (Aug. 21, 2017) (internet).

3

These holdings threaten to undermine amici

States’ antitrust enforcement efforts by placing an

unwarranted burden on plaintiffs seeking to challenge

unreasonable restraints in industries that may involve

two-sided platforms, such as the credit-card industry.

Contrary to the court of appeals’ reasoning below,

well-established antitrust doctrines involving market

definition and burden shifting fully account for the

distinct economics of two-sided platforms and the

different ways in which firms operating two-sided (or

multisided) platforms may compete. The amici States

have a substantial interest in establishing the correct

principles for analyzing unreasonable restraints of

trade, in order to prevent disruption to their antitrust

enforcement efforts.

STATEMENT

In this antitrust enforcement action, the United

States and several States (together, the government“)

alleged that respondents American Express Company

and American Express Travel Related Services

Company (“Amex”) violated federal antitrust law by

contractually barring merchants that accept Amex

credit cards from encouraging their customers to use

a different credit card that is less expensive for the

merchants to accept. After a bench trial, the United

States District Court for the Eastern District of New

York (Garaufis, J.) held Amex liable for unreasonably

restraining trade in violation of § 1 of the Sherman

Act. On appeal, the United States Court of Appeals for

the Second Circuit reversed and entered judgment for

Amex.

4

A. Two-Sided Platforms

By connecting merchants and cardholders to

complete financial transactions, Amex operates a

business that economists call a “two-sided platform.”

See Jean-Charles Rochet & Jean Tirole, Two-Sided

Markets: A Progress Report (“Progress Report’), 37

RAND J. of Econ. 645, 645-46 (2006).* Although many

firms act as intermediaries between other businesses

and customers, a distinct feature of two-sided

platforms is that they connect two different groups of

the firm’s own customers and facilitate an interaction

between them. See Delegation of the U.S. to the

Competition Committee, Roundtable on Two-Sided

Markets: Note by the Delegation of the United States

(U.S. Roundtable Note”), at 2, OECD No.

DAF/COMP/WD(2009)68 (2009).

The manner in which the two sides of a platform

interact may depend on the specific customer groups

and businesses involved in each platform. But many

two-sided platforms share certain common features.

In particular, although the two groups of customers

often purchase different products at different prices

from the firm operating the platform, the value that

each customer group derives from the platform may

depend on the other group’s usage of the platform. See

id.; see also Progress Report, supra, at 645-67. With

respect to credit-card platforms, for example, a credit

card is more valuable to a retail customer when more

merchants accept that card at their stores. In turn, a

credit-card network is more valuable to merchants

See also Marc Rysman, The Economics of Two-Sided

Markets, 23 J. of Econ. Perspectives 125 (2009); Mark Armstrong,

Competition in Two-Sided Markets, 37 RAND J. of Econ. 669

(2006).

5

when more retail customers use that card to make

purchases. See U.S. Roundtable Note, supra, at 2.

When the value that each customer group

derives from a two-sided platform is interconnected in

this manner, the company operating the platform may

need to obtain and retain sufficient numbers of

customers on each side of the platform to ensure that

the platform functions effectively. Jean-Charles

Rochet & Jean Tirole, Platform Competition in Two-

Sided Markets, 1 J. Eur. Econ. Ass’n 990, 990 (2003).

For example, if too few merchants accept a credit-card

company’s card, consumers would likely have little

incentive to use that card. And if hardly any

consumers use a specific credit card, merchants would

likely have little incentive to accept that card brand.

To solve this problem, where one group of customers

is more difficult to attract than the other group, the

platform may structure its pricing to account for this

difference. U.S. Roundtable Note, supra, at 2-3. For

example, a shopping mall charges retailers rent to

access the mall but likely allows shoppers to enter

without charge. Credit-card companies have likewise

long charged merchants higher prices for using credit-

card services than are charged to cardholders, who

may pay an annual card fee but usually pay no

transaction fees and may even pay negative prices if

they receive rewards from using a card. See id. at 3.

Pricing thus may look “unusual in two-sided markets”

because pricing on “one side of the market depends not

only on the demand and costs that those consumers

bring but also on how their participation affects

participation” and pricing on the other side of the

platform. Rysman, supra, at 129. Moreover, pricing is

not the only competitive factor that may differ in two-

sided markets because strategies for “innovation,

6

advertising, and quality investment can also take on

added dimensions in a two-sided market.” Id. at 135.

B. The Government’s Antitrust Complaint

Against Amex

In this case, the government alleges that Amex

illegally stifled competition on the merchant side of its

credit-card platform—also known as the “network-

services market” or the “merchant-services market.”

The government challenged Amex’s “anti-steering

rules,” which are private contractual provisions that

bar merchants that accept Amex cards from

encouraging their customers to use a different credit

card. (Pet. App. 66a-68a.) Amex’s anti-steering rules

thus prohibit a merchant from indicating that it

prefers another credit card or from offering consumers

an incentive (such as a price discount) to use a

different credit card. For example, the rules preclude

a merchant from telling a customer that it “prefers

Visa”; providing a discount to customers who use

Discover cards; or posting a sign displaying the fees it

pays to use each credit-card network. (Pet. App. 95a-

96a, 100a.) Absent the anti-steering rules, a merchant

might want to encourage its customers to use a non-

Amex card brand because Amex often charges mer-

chants higher fees than other credit-card companies

charge. (Pet. App. 68a.)

The government alleged that the anti-steering

rules caused actual anticompetitive effects in the

merchant-services market by essentially eliminating

the credit-card companies’ incentives to compete for

merchants. (See Pet. App. 68a.) For example, the

government alleged that the anti-steering rules had

raised merchant fees above competitive levels,

prevented credit-card companies from offering lower

7

merchant fees, and stifled innovation. (See Pet. App.

196a-197a, 203a, 212a-214a.)

C. The District Court’s Opinion

Amex’s anti-steering rules are a form of restraint

that is analyzed for antitrust purposes under the rule

of reason. Courts follow a four-step burden-shifting

framework to apply the rule of reason. First, the

plaintiff has the initial burden to make a prima facie

showing that the challenged restraint is anticompeti-

tive in the relevant market.‘ See National Collegiate

Athletic Ass n v. Board of Regents of the Univ. of Okla.

(“NCAA”), 468 U.S. 85, 104-13 (1984). Second, if the

plaintiff carries its burden, the burden shifts to the

defendant to establish a procompetitive justification

for the restraint in the relevant market. See id. at 113;

California Dental Ass’n v. FTC, 526 U.S. 756, 788

(1999) (Breyer, J. concurring in part & dissenting in

part). Third, where the defendant satisfies its burden,

the plaintiff may show that the defendant’s procompet-

itive benefits could have been achieved through less

restrictive means, i.e., actions that are less damaging

to competition. Fourth, if no less-restrictive alter-

natives exist, the plaintiff bears the burden of

demonstrating that the anticompetitive harms

outweigh the procompetitive gains. See 7 Phillip E.

As an alternative to establishing directly that a restraint

caused actual anticompetitive effects, a plaintiff may prove

anticompetitive effects indirectly by establishing that the

defendant imposed a restraint and possessed sufficient market

power to impair competition. See FTC v. Indiana Fed’n of

Dentists, 476 U.S. 447, 460-61 (1986); Geneva Pharm. Tech. Corp.

v. Barr Labs. Inc., 386 F.3d 485, 509 (2d Cir. 2004). This market-

power route to establishing anticompetitive effects is not at issue

here. See infra at 9 n.5.

8

Areeda & Herbert Hovenkamp, Antitrust Law: An

Analysis of Antitrust Principles and Their Application

430 (3d ed. 2010) (describing framework’s four steps).

Applying this framework here, the district court

determined after a seven-week bench trial that the

anti-steering rules violate § 1 of the Sherman Act.

(Pet. App. 66a-72a.)

Market Definition: The court first determined that

the relevant market was the market for merchant

“network services”’—i.e., the services that enable

merchants to process credit-card payments. The court

rejected Amex’s argument that the relevant market

should be a market for credit-card “transactions,”

which would include not only merchant services but

also the distinct services that Amex provides to

cardholders—such as the provision of credit and

cardholder rewards for using an Amex card. The court

explained that merchant services and cardholder

services are not interchangeable, and thus could not

be collapsed into a single, platform-wide market under

traditional antitrust doctrines. (Pet. App. 113a-121a.)

The court emphasized that limiting the relevant

market to include only merchant services would not

require the court to ignore entirely the cardholder side

of the platform. Rather, the court explained that

because the merchant and cardholder sides of the

platform are “inextricably linked” (Pet. App. 118a), the

court would consider “the two-sided features” of the

platform as part of the rule of reason analysis (Pet.

App. 121a).

Anticompetitive Effects: The district court next

concluded that the government had satisfied its prima

facie burden of showing that Amex’s anti-steering

rules had caused actual anticompetitive effects in the

9

market for merchant services.“ Specifically, the court

concluded that the rules had increased prices to

merchants above competitive levels, blocked other

firms from using a lower-fee business model, and

stifled innovation. (Pet. App. 203a-214a.)

The court determined that the anti-steering

rules caused these anticompetitive effects by exploiting

interconnections between the merchant and card-

holder sides of the network. As the court explained,

the rules prohibited merchants from responding to

increases in Amex’s fees by encouraging customers to

use lower-fee cards. (Pet. App. 194a-197a.) The rules

did so by controlling merchants’ interactions with

customers at the “point of sale“ the critical moment

when the merchant side and cardholder side of the

platform meet to complete a credit-card transaction.

(Pet. App. 195a.) At that moment, the merchant’s

decision to use a particular credit-card company’s

services—and thus to pay that company’s per-

transaction fee—is effectively controlled by the

customer’s decision to use a specific card. The court

determined that prohibiting merchants from steering

customers to use lower-fee cards deprived the

merchants of “any meaningful ability” to switch their

own use of network services in response to lower fees

offered by a particular card network. (Pet. App. 196a.)

As a result, the district court found that the anti-

steering rules impeded horizontal competition among

the credit-card networks because the networks lacked

5 In the alternative, the district court held that the govern-

ment had satisfied its prima facie burden by demonstrating that

Amex possessed sufficient market power to cause anticompetitive

effects. (Pet. App. 148a-191la). This market power finding was

reversed by the Second Circuit and is no longer at issue.

10

any incentive to offer merchants lower fees. (Pet. App.

194a-197a, 202a-203a.)

Procompetitive Effects: Because the government

had established actual anticompetitive effects, the

district court shifted the burden to Amex to demon-

strate that its anti-steering rules had procompetitive

effects in the merchant-services market. (Pet. App.

228a.) Among other arguments, Amex asserted two

primary procompetitive justifications. First, Amex

argued that the anti-steering rules allowed Amex to

compete against other credit-card networks for

cardholders’ business. Amex contended that the anti-

steering rules protected its ability to impose higher

fees on merchants, and that those fees in turn funded

robust cardholder rewards and other benefits that

allowed Amex to compete with the larger credit-card

networks (MasterCard and Visa) for cardholders. (See

Pet. App. 229a-234a; Defs.’ Post-Trial Br. (Defs. Br.“)

3-4, 21-25, United States v. American Express Co., No.

10-cv-4496 (E.D.N.Y.), ECF No. 605; Defs.’ Propose.

Findings of Fact & Conclusions of Law (“Defs.

Proposed Findings”) 29-36, American Express, No. 10-

cv-4496, ECF No. 612.)

Second, Amex asserted that the anti-steering

rules allowed Amex to offer significant value, not only

to cardholders, but also to merchants. Amex argued

that its cardholder rewards program and other

cardholder benefits allowed merchants to gain access

to high-spending Amex customers. (See Defs. Proposed

Findings, supra, at 36-40.) Amex further argued that,

absent the anti-steering rules, merchants could free-

ride on the benefits of Amex’s investments in

cardholder rewards while avoiding the higher

merchant fees needed to fund those programs. (See

Pet. App. 255a-256a; Defs. Br., supra, at 25-26.)

11

In analyzing both of these arguments, the

district court questioned whether it could consider the

anti-steering rules’ purported procompetitive impacts

on the market for cardholder services at all. The court

noted that procompetitive effects in one market cannot

be used to offset anticompetitive effects in a separate,

interrelated market—here, the market for merchant

services. (Pet. App. 238a-24 la.)

The court nonetheless concluded that, even if

such “cross-market balancing is appropriate” (Pet.

App. 239a), the anti-steering rules were not reasonably

necessary to achieve the procompetitive effects that

Amex had raised (Pet. App. 239a-240a). The court

found that Amex’s business model was unlikely to

collapse if the anti-steering rules were removed

because merchants and cardholders would still want

to use Amex’s services if Amex truly delivered services

supericr to those of other credit-card networks. (Pet.

App. 242a-245a.) The court also determined that the

rules were not needed to prevent free-riding because

rational cardholders would continue to use their Amex

cards so long as the rewards offered by Amex were

greater than any benefits provided by merchants for

using a different card. (Pet. App. 257a.) The court

concluded that even without the rules, Amex would

thus continue to have an incentive to invest in its

cardholder rewards program. (Pet. App. 257a.) Finally,

the court determined that even if the anti-steering

rules were necessary to achieve some procompetitive

effects, any such procompetitive gains did not “offset,

much less overcome,” the rules’ anticompetitive effects

on the network-services market. (Pet. App. 229a; see

also Pet. App. 239a-240a.)

12

D. The Second Circuit’s Opinion

On appeal, the Second Circuit reversed and

directed the district court to enter judgment in favor

of Amex. (Pet. App. 4a.) Two holdings by the Second

Circuit are relevant here.

First, the Second Circuit held that the district

court had improperly defined the relevant market by

excluding the customer side of Amex’s two-sided

platform. The Second Circuit determined that the

proper market was the entire platform, including both

cardholders and merchants (Pet. App. 50a), because

concluding otherwise would improperly ignore “the

two markets’ interdependence” and the anti-steering

rules’ impacts on the cardholder side (Pet. App. 34a-

35a.)

Second, the Second Circuit concluded that the

government had failed to carry its prima facie burden

of establishing that Amex’s anti-steering rules caused

actual anticompetitive effects in the platform-wide

market it had defined. The court held that the

government was required to prove a net anticompeti-

tive harm across the entire platform—for example, by

demonstrating that the anti-steering rules caused

Amex to charge a supracompetitive “net price” to

merchants and cardholders (Pet. App. 49a); reduced

output by lowering the number of credit-card

transactions; or reduced the net quality of services

Amex provides to merchants and cardholders (Pet.

App. 5la-52a).

The court determined that the government had

failed to establish any such platform-wide net

anticompetitive harm, and that the burden to prove

procompetitive effects hud thus never shifted to Amex.

While the court did not question the district court's

13

finding that the anti-steering rules had increased

prices for merchants, the court found that the

government had failed to demonstrate a higher “net

price” that took into account the benefits that Amex

provided to cardholders. (Pet. App. 49a; see Pet. App.

53a.) The court thus faulted the district court for

improperly focusing “on the interests of merchants

while discounting the interests of cardholders.” (Pet.

App. 54a.) The court further reasoned that the

government had failed to satisfy its prima facie

burden, notwithstanding the price increase for

merchants, because the overall number of credit-card

transactions had increased rather than decreased, and

because Amex had improved the quality of its

cardholder services. (Pet. App. 52a-53a.)

SUMMARY OF ARGUMENT

Two-sided platforms present unique features that

courts may appropriately consider in applying

established antitrust doctrines. But the court of

appeals here committed two critical legal errors in

applying antitrust principles to Amex’s two-sided

credit-card platform. The court of appeals incorrectly

concluded (I) that the relevant market must be

defined as the entire platform, rather than as separate

markets for merchant services and cardholder services;

and (2) that at the prima facie stage, the government

must establish a net anticompetitive effect that harms

both merchants and cardholders, rather than an

anticompetitive effect (such as a price increase) on

only the merchant side of the platform.®

As explained supra at 7 n.4, as an alternative to direct proof

of an actual anticompetitive effect, an antitrust plaintiff may

14

The Second Circuit distorted bedrock antitrust

principles to reach both of these conclusions, based on

the unfounded concern that it was required to do so to

account for interdependencies between the merchant

and cardholder sides of Amex’s platform. But long-

standing antitrust doctrines already ensure tuat

courts may properly consider interdependencies

between the two sides of a platform. The Second

Circuit's novel and disruptive approaches to two-sided

platforms are unnecessary and threaten to undermine

valid antitrust enforcement.

I. This Court has long defined the relevant market

for antitrust purposes as including only products that

are reasonably interchangeable for each other. The

court of appeals failed to apply this controlling principle

here when it combined into a single market the

services that Amex sells to merchants and the services

Amex sells to cardholders. These two types of services

are not interchangeable because a merchant that buys

Amex’s merchant services cannot switch to purchasing

Amex’s cardholder services in response to Amex

increasing merchant fees.

The Second Circuit mistakenly defined the market

as the entire platform based on the concern that

separating the markets for merchant services and

cardholder services would improperly ignore intercon-

nections between these two sides of Amex’s platform.

But defining the relevant market as the market for

merchant services would not preclude consideration of

satisfy its prima facie burden by demonstrating that a defendant

has imposed a restraint of trade and possesses sufficient market

power to suppress competition in a properly defined market

affected by the restraint. This independent route to establishing

anticompetitive effects is not at issue here.

15

the anti-steering rules’ impacts on the cardholder side

of the platform, as the court of appeals reasoned.

Rather, under a proper market definition, the court

would be able to consider the anti-steering rules’

impacts on the cardholder side to the extent that such

impacts affect competition on the merchant side. Such

an inquiry both accords with long-standing antitrust

analysis and ensures that the competitive realities of

a two-sided platform are fully addressed.

Defining the relevant market as including only

merchant services also protects against another error

in the Second Circuit's reasoning, namely, the

suggestion that any benefit on one side of a platform

will always be relevant to evaluating the reasonable-

ness of a restraint on the other side of the platform—

irrespective of whether the benefit actually affected

competition on the side of the platform at issue. Such

a blanket rule does not comport with the rule of

reason’s fact-intensive inquiry into a restraint's

competitive effects in the relevant market. Because

different industries may involve different types of two-

sided (or multisided) platforms, the procompetitive

significance of an out-of-market benefit on the

relevant side of a platform will vary depending on the

particular platform and restraint at issue. The

antitrust inquiry should thus remain focused on

whether, under the facts of each case, a procompetitive

effect on one side of a platform affects competition in

the relevant market on the other side of the platform.

Il. The court of appeals’ incorrect definition of the

market as including both merchant and cardholder

services caused the court to conclude that the

government's prima facie burden to establish an

anticompetitive effect required showing that the anti-

steering rules imposed a net anticompetitive harm to

16

both merchants and cardholders. But the established

burden-shifting framework applied under the rule of

reason provides that where the government establishes

an anticompetitive effect, the burden shifts to the

defendant to show potential procompetitive effects

from its restraint. The Second Circuit's “net harm”

approach drastically distorts this framework by

requiring the government, at the prima facie stage, to

predict and disprove any potential procompetitive

effects.

The burden-shifting framework was developed to

avoid placing precisely such an undue burden on

antitrust plaintiffs. Because a defendant possesses in-

depth knowledge about its own industry and motiva-

tions for implementing a restraint, the burden should

remain with the defendant to demonstrate procompet-

itive effects from its conduct. And contrary to the

Second Circuit’s concern, maintaining a proper

allocation of burdens will not chill potentially

procompetitive conduct in industries involving two-

sided platforms. The burden-shifting framework

already protects such conduct by providing a

defendant with a full and fair opportunity to establish

that its conduct causes procompetitive effects on the

side of the platform at issue.

17

ARGUMENT

I. The Traditional “Reasonable

Interchangeability” Rule for Market

Definition Applies to Two-Sided Platforms.

A. The Services That Amex Sells to

Merchants Are a Separate Market from

the Services It Sells to Cardholders.

For more than fifty years, this Court, legal

scholars, and economists have defined the relevant

market for antitrust purposes as including only

products that are reasonably interchangeable with the

product over which the defendant is alleged to have

exercised an illegal restraint of trade. See, e.g., United

States v. E. I. du Pont de Nemours & Co., 351 U.S. 377,

404 (1956); see also United States v. Grinnell Corp.,

384 U.S. 563, 571-72 (1966); 2B Phillip E. Areeda,

Herbert Hovenkamp & John L. Solow, Antitrust Law:

An Analysis of Antitrust Principles and Their

Application 235-38, 389-93 (4th ed. 2014). This

“reasonable interchangeability rule, E. I. du Pont, 351

U.S. at 404, properly identifies the range of products

that consumers can buy from other firms instead of

the defendant's product in response to defendant’s

conduct. The relevant market has long been defined in

this way because reasonably interchangeable products

constrain the defendant's ability to engage in

anticompetitive conduct: purchasers will simply buy

another product rather than continue to purchase the

defendant’s product subject to defendant's anticompet-

itive restraint. See United States v. Microsoft, 253 F.3d

34, 51-52 (D.C. Cir. 2001) (en banc) (per curiam);

Areeda, Hovenkamp & Solow, supra, at 266-67.

18

Applying this bedrock principle of market

definition here, the services that Amex sells to mer-

chants and the services that Amex sells to cardholders

are in separate antitrust markets because the two

types of services are not interchangeable. As the

petitioner States and the United States aptly explain,

merchants that purchase Amex’s merchant services

cannot switch to buying Amex’s cardholder services—

such as the extension of credit and provision of

rewards for using an Amex card—in response to Amex

increasing its merchant fees. See Br. for the Pet'rs &

Resp’ts Nebraska, Tennessee, and Texas 53-55; Br. for

the United States 35-37. Likewise, a cardholder

unhappy with an increase in the interest rate charged

by Amex cannot switch to purchasing network

services. Accordingly, merchant services “constitute a

distinct market, separate from the market for

cardholder services. See United States v. Visa U.S.A.,

Inc., 344 F.3d 229, 239 (2d Cir. 2003); see also Renata

B. Hesse & Joshua H. Soven, Defining Relevant

Product Markets in Electronic Payment Network

Antitrust Cases, 73 Antitrust L.J. 709, 726-27 (2006)

(credit-card network may be able to exercise market

power over only one side of two-sided platform).

B. Limiting the Relevant Market to One

Side of a Platform Does Not Preclude

Consideration of Impacts on the Other

Side of the Platform.

The Second Circuit violated the well-

established principles of market definition by

collapsing the market for merchant services and the

market for cardholder services into a single platform-

wide market for credit-card transactions. The court

applied this novel platform-wide approach based on a

19

concern that dividing the two sides of Amex’s network

into separate markets would improperly ignore

interdependencies between the merchant and

cardholder sides of the platform. (See Pet. App. 34a-

35a, 39a-40a, 49a.)

The court’s concern was unfounded. Separating

the merchant and cardholder sides of Amex’s network

into distinct markets would not preclude all considera-

tion of the anti-steering rules’ competitive impacts on

the cardholder side. To the contrary, courts applying

the rule of reason may consider evidence about a

restraint’s impacts outside of the market in question

when such impacts affect competition in the relevant

market. Applying the traditional doctrine of market

definition thus does not preclude the factual inquiry

into cardholder benefits discussed by the court of

appeals here. By contrast, the court of appeals’ merger

of the distinct merchant- and cardholder-services

markets improperly constrains such a factual inquiry

by mistakenly presuming that any cardholder benefit

caused by Amex’s anti-steering rules would be relevant

to justifying the higher prices that those rules inflict

on merchants—irrespective of the nature or strength

of the interconnections between the two sides of the

platform.

The appropriate way for this Court to resolve

these competing concerns is to hold that (a) the

relevant market is limited to merchant-side services,

and (b) courts may look to the anti-steering rules’

effects on the cardholder-services market, but only to

the extent that cardholder benefits affect competition

in the relevant market for merchant services. In this

way, the market definition respects the “commercial

realities” of two-sided platforms, see Eastman Kodak

Co. v. Image Tech. Serus., Inc., 504 U.S. 451, 482 (1992),

20

by permitting a factual inquiry into the related

cardholder-services market, while limiting any such

inquiry to only the relevant aspects of that separate

market.

1. When the relevant market is one

side of the platform, a court may

consider the competitive effects on

that market caused by a restraint’s

effects on the other side of the

platform.

Allowing a court to consider arguments that an

alleged restraint’s impact on one side of a two-sided

platform affects competition in the relevant market on

the other side of the platform accords with established

antitrust principles and long-standing practice. The

touchstone of the rule of reason is the competitive

effect of the challenged restraint and whether it

suppresses or promotes competition. See National

Soc’y of Prof Eng’rs v. United States, 435 U.S. 679,

688 (1978); see also Leegin Creative Leather Prods.,

Inc. v. PSKS, Inc., 551 U.S. 877, 886 (2007) (“In its

design and function the rule [of reason] distinguishes

between restraints with anticompetitive effect that

are harmful to the consumer and _ restraints

stimulating competition that are in the consumer's

best interest.”). This analysis has long encompassed

consideration of a wide variety of factors—such as

historical conditions, applicable regulations, and other

“facts peculiar to the business“ that illuminate

whether the defendant’s conduct has anticompe itive

or procompetitive effects in the relevant market.

Board of Trade of Chicago v. United States, 246 U.S.

231, 238 (1918). Competitive effects or legitimate

justifications that arise in a separate but closely

21

interrelated market may likewise be considered, so

long as they cause competitive effects in the market at

issue.

Courts are familiar with considering the ways in

which a defendant’s conduct affects competition in the

relevant market through interconnections between

that market and a distinct but related market. For

example, in Microsoft Corp., several States and the

United States proved at trial that Microsoft unlawfully

maintained its monopoly in the market for Intel-

compatible computer operating systems (“operating-

system market”) through, among other things,

anticompetitive actions in the separate but inter-

related market for internet browser software. See 253

F.3d at 47. In reaching this conclusion, the D.C.

Circuit rejected Microsoft's argument that its actions

in the browser market were irrelevant because there

were separate markets for browsers and operating

systems. See id. at 53-54, 78-79. Notwithstanding the

separate markets, the court determined that inherent

interconnections between these markets allowed

Microsoft to maintain its monopoly in the operating-

system market through actions it took in the browser

market specifically, Microsoft had suppressed usage

of other companies’ browsers and thereby prevented

development of competing software applications that

would have increased competition in the operating-

system market. Jd. at 60-62. Microsoft's browser-

related conduct thus contributed to its anticompetitive

monopolization of the operating-system market.

The parties in this proceeding presented similar

evidence below about the feedback loop between the

cardholder impacts of Amex’s anti-steering rules and

competition on the merchant market at issue here. For

example, the government argued that the anti-steering

22

rules prevented consumers from obtaining (or even

perceiving) any benefit from the lower-cost credit

cards offered by Amex’s competitors, and thus

restrained price competition in the merchant-services

market by eliminating any incentive for the credit-

card networks to compete to offer lower merchant fees.

(See Pls. Post-Trial Mem. 1, 5-7, 9, 11, 14-15, American

Express, No. 10-cv-4496, ECF No. 606; see also Pet.

App. 191la-198a.) Amex, in turn, argued that the anti-

steering rules permitted it to offer robust benefits to

cardholders that encouraged them to spend more

money at the merchants’ businesses—thus improving

competition in the merchant-services market. (See

Pet. App. 86a-88a; see also Defs. Br., supra, at 23-25;

Defs. Proposed Findings, supra, at 226-33, 241-48.)

Adhering to traditional principles of market

definition thus fully permits courts to consider

whether a restraint’s impact on one side of a platform

causes actual and direct competitive effects in the

relevant market on the other side of a platform. As the

record shows, such evidence may both support and

undermine a plaintiffs claim of an unreasonable

restraint on competition in the relevant market. The

court of appeals thus had no basis to reject traditional

market-definition principles here.

2. Courts should not consider a

restraint’s asserted benefit on the

other side of a platform unless it

causes a procompetitive effect in

the market at issue.

The Second Circuit’s erroneous market definition

led it to believe that any procompetitive benefit on the

cardholder side could be considered in the competitive-

effects analysis—irrespective of whether the benefit

23

had an effect on competition in the relevant merchant-

services market. (See Pet. App. 5la (criticizing

government for failing to consider anti-steering rules’

effects on “all Amex consumers on both sides of the

platform”); Pet. App. 54a (criticizing district court for

“discounting the interests of cardholders”).) Amex

likewise argued below that its anti-steering rules

could be justified solely by their alleged procompeti-

tive effects on the cardholder side of the platform. See

supra at 10.

Such reasoning expands too far the scope of

potential procompetitive justifications relevant to

analyzing a two-sided platform under the rule of

reason. As this Court has explained, the rule of reason

“does not open the field of antitrust inquiry to any

argument in favor of a challenged restraint.” National

Soc’y of Profl Eng’rs, 435 U.S. at 688 (emphasis

added). In particular, a defendant may not simply

point to some perceived benefit in a separate market

to justify its anticompetitive actions in the market in

question. See United States v. Visa U.S.A., Inc., 163 F.

Supp. 2d 322, 330 (S.D.N.Y. 2001) (vigorous competi-

tion “to offer lower interests rates” and incentive

programs to cardholders could not be used to offset

“reduced price competition in the network services

market”). Rather, the rule of reason “focuses directly

on the challenged restraint’s impact on competitive

conditions” in the relevant market. National Soc’y of

Prof Eng’rs, 435 U.S. at 688.

To be sure, two-sided platforms that exhibit close

interconnections between the groups on either side of

the platform would ordinarily aid a party’s attempt to

establish how impacts on one side of a platform affect

competition on the other side. But because not all two-

sided platforms are alike, the strength and importance

24

of “the network interactions between the two sides” of

a platform will vary depending on the platform and

market participants, as well as the specific restraint

at issue. U.S. Roundtable Note, supra, at 3; see Dennis

W. Carlton & Alan S. Frankel, Transaction Costs,

Externalities, and “Two-Sided” Payment Markets,

2005 Colum. Bus. L. Rev. 617, 630 (2005). Establishing

the relevant connection is thus a fact-intensive inquiry

that will turn on the particular features of the

platform and restraint in question—features that are

likely to differ from platform to platform and restraint

to restraint.

For example, platforms that are used to complete

financial transactions between two customer groups

(“transactional platforms”) differ significantly from

other types of platforms. Transactional platforms—

such as the credit-card payment platform at issue

here—often exhibit strong, two-way interconnections

between each side’s use of the platform. See Lapo

Filistrucchi, Damien Geradin & Eric van Damme,

Identifying Two-Sided Markets 4 (Tilburg Univ.,

TILEC Discussion Paper No. 2012-008 2012) (internet).

In other cases, however, the interactions between

the two sides of a platform “may be insignificant or

may not be relevant for a particular antitrust issue” or

a particular restraint. U.S. Roundtable Note, supra, at

3. For example, a newspaper is generally considered

to be a two-sided platform that unites readers and

advertisers. See Rysman, supra, at 128; see Times-

Picayune Publ’g Co. v. United States, 345 U.S. 594, 610

(1953). But a transaction between a reader and an

advertiser does not occur each time that a reader

purchases a newspaper or an advertiser buys adver-

tising space. And while an advertiser likely wants to

purchase more advertising from a newspaper that has

25

more readers, a reader does not necessarily want to

purchase a newspaper because it has more

advertisers—in fact, a reader might prefer a news-

paper with fewer advertisers. See Filistrucchi et al.,

supra, at 4. As a result, although advertising revenue

may allow the newspaper to charge a lower price to

readers, the interdependencies between the two

groups are more attenuated than in a transactional

platform. See id. at 6-7.

An approach to two-sided platforms that too

broadly allows a restraint’s procompetitive effects on

one side of the platform to justify its anticompetitive

effects in the relevant market on the other side of the

platform would improperly ignore such critical

differences between platforms. Such an approach would

also undermine proper enforcement in cases involving

two-sided platforms or similarly complex industries. If

the universe of relevant procompetitive factors is not

limited to factors that are proven to have procompeti-

tive effects in the relevant market, defendants could

easily justify misconduct that stifles competition in

one market by pointing to purported procompetitive

effects in some other market—so long as they could

plausibly claim that the two kets are connected in

some way. Such an expansive view of the relevant

procompetitive justifications would improperly allow

defendants to “sacrifice competition” in one market

“for greater competition” in another market, thereby

frustrating the goals of the antitrust laws: to preserve

vigorous competition in all markets. United States v.

Topco Assocs. Inc., 405 U.S. 596, 610-11 (1972).

Nearly any two-sided platform could adopt Amex’s

expansive theory of procompetitive effects to justify

even a naked restraint of trade on one side of the

platform. For example, an outlet mall that prohibited

26

retailers from opening another outlet store within

sixty miles could claim that it needs such a restraint

to ensure that sufficient shoppers visit its outlet mall,

thereby allowing the mall to compete with other

outlets—even though the restraint plainly inhibits

outlet malls’ competition for retailers. This Court

should not endorse such arguments, which, if

accepted, would effectively allow a defendant to

suppress competition in one market so long as it uses

its supracompetitive profits to benefit a different

group of customers in the process. See id. at 611-12

(explaining that defendants “keenly aware of their

own interests” are not permitted to determine “relative

values to society of competitive areas of the economy”).

Moreover, adopting the Second Circuit’s reasoning

could have widespread negative consequences for

antitrust enforcement in industries that involve

interrelated markets other than two-sided platforms.

Any firm in such an industry could try to justify its

restraints anticompetitive effects in the relevant

market by pointing to purported procompetitive

benefits from the restraint in a separate, interrelated

market. Such arguments would likely become

widespread because separate markets are often

interconnected with each other. For example, the

health care industry involves several separate but

interrelated markets involving health care providers,

insurance companies, and patients. But interconnec-

tions between the price of health care services and the

price of health insurance does not mean that the

medical services that hospitals sell to patients are in

the same market as the insurance plans that

insurance companies sell to insureds. And a hospital

should not be able to justify a restraint that raises

prices or decreases quality in the health care services

27

market by claiming, for example, that the restraint

increases innovation in the insurance market. (/

United States v. Charlotte-Mecklenburg Hosp. Auth.,

248 F. Supp. 3d 720, 728-32 (W.D.N.C. 2017) (govern-

ments stated § 1 claim against hospital based on

hospital’s contractual provisions prohibiting insurance

companies from encouraging patients to use lower-

cost health care services). Allowing such wide-ranging

effects to be cognizable would render unmanageable

an already complex rule of reason analysis.

Il. A Plaintiff Satisfies Its Prima Facie Burden

by Demonstrating an Anticompetitive

Effect—Such as a Price Increase—on One

Side of a Two-Sided Platform.

The Second Circuit’s erroneous definition of the

relevant market as including both merchant services

and cardholder services caused the court to conclude

that the government’s initial, prima facie burden to

establish an anticompetitive effect required showing

that the anti-steering rules imposed a “net harm” to

competition for both merchants and cardholders. (Pet.

App. 54a.) The Court should reject this approach and

hold that the government satisfied its prima facie

burden here by demonstrating that Amex’s anti-

steering rules stifle price competition in the market

for merchant services.

28

A. A Plaintiff Should Not Be Required to

Predict and Disprove All Possible

Procompetitive Benefits to Satisfy Its

Prima Facie Burden of Demonstrating

an Anticompetitive Effect.

Through decades of experience, the courts have

developed a burden-shifting framework to structure

the highly complex and fact-intensive rule of reason

inquiry. See Gregory J. Werden, Antitrust’s Rule of

Reason: Only Competition Matters, 79 Antitrust L. J.

713, 756-57 (2014). Under this framework, when the

plaintiff establishes that a restraint imposes an

anticompetitive effect in the relevant market, the

burden shifts to the defendant to establish a procom-

petitive effect from the restraint. See NCAA, 468 U.S.

at 104-13; California Dental Ass n, 526 U.S. at 788

(Breyer, J., concurring in part & dissenting in part).

When the defendant satisfies its burden, the plaintiffs

may show that the procompetitive benefits could be

achieved through less restrictive means. Finally, if the

plaintiff fails to establish such less-restrictive options,

the court determines whether the anticompetitive

harms outweigh the procompetitive gains. See Areeda

& Hovenkamp, supra, at 430.

This structured framework evolved to achieve a

“fair and efficient” system for distinguishing between

restraints that harm competition, which are

prohibited, and restraints that promote competition,

which are allowed. See Leegin Creative Leather Prods.,

551 U.S. at 898-99; see also California Dental Ass n,

526 U.S. at 793-94 (Breyer, J., concurring in part &

dissenting in part). The serial shifting of burdens

strikes an important balance between the litigants.

See Herbert Hovenkamp, The Rule of Reason 16 (Univ.

of Penn. Law Sch. Inst. for Law and Econ., Research

29

Paper No. 17-28 2017) (internet) (“Of all the

procedural issues involved in antitrust litigation

under the rule of reason, none are more critical than

questions about assignment of the burden of proof).

In particular, by giving plaintiffs the initial burden of

showing only anticompetitive effects at the prima facie

stage, and leaving to defendants the burden of

identifying procompetitive justifications, the burden-

shifting framework ensures that plaintiffs are not

required to predict and refute “every possible fact or

theory” that a defendant might assert to justify its

conduct—a herculean standard that would render

antitrust enforcement and litigations unfocused and

unmanageable. See FTC v. Actavis, Inc., 133 S. Ct.

2223, 2237 (2013); see also Hovenkamp, Rule of Reason,

supra, at 34 (discussing Actavis).

Here, the government satisfied its initial burden

at the prima facie stage by establishing that Amex’s

anti-steering rules had “the effect of raising the price

that merchants pay for credit-card services, see NCAA,

468 U.S. at 105, by eliminating any incentive for the

credit-card networks to compete to offer lower fees to

merchants. (See Pet. App. 195a-203a.) Such proof that

the anti-steering rules actually “operated to raise

prices” in the relevant market should have shifted to

Amex the “heavy burden of establishing an affirmative

defense which competitively justifies this apparent

deviation from the operations of a free market.” See

NCAA, 468 U.S. at 113; see also Hovenkamp, Rule of

Reason, supra, at 22 (discussing NCAA).

The court of appeals concluded otherwise, holding

that the government had not satisfied its prima facie

burden because it had not proven that the anti-

steering caused a net anticompetitive effect on both

merchants and cardholders. (See Pet. App. 5la-52a.)

30

The court appeared to give two reasons that the

government's focus on supracompetitive prices for

merchants was insufficient. First, the court reasoned

that the government should further have calculated

the “net” or “two-sided price” of credit-card transac-

tions by offsetting “the value or cost of the rewards

paid to cardholders” against the higher merchant fees

that funded those rewards. (Pet. App. 53a.) Second,

the court reasoned that the government should have

shown that the anti-steering rules also reduced the

overall volume of credit-card transactions or decreased

the quality of credit-card services. (Pet. App. 52a.)

The court of appeals’ “net harm” test misconstrues

the government’s prima facie burden. Both of the

court’s grounds for finding the government’s initial

showing defective required the government to prove

more than anticompetitive effects in the relevant

merchant-services market. But the additional show-

ings that the court would have required are more

appropriately seen as part of Amex’s burden to prove

procompetitive justifications. Under this Court's

precedents, it was Amex’s burden to demonstrate that

its anti-steering rules led to increased cardholder

benefits (to the extent that such benefits affect the

merchant-services market, see supra Point I.B.2), or

that the rules contributed to increased output or

quality in the merchant-services market. See NCAA,

468 U.S. at 113. By requiring the government to

disprove these procompetitive justifications at the

outset, the court of appeals wrongly collapsed the first

two steps of the burden-shifting inquiry into a single

step.

Mandating that the government predict and

disprove potential procompetitive effects simply to

show an initial anticompetitive effect in the market

31

would severely undermine enforcement. The courts

have placed the burden on defendants to establish

that their conduct creates procompetitive effects

because a defendant possesses in-depth knowledge

about its own industry, business model, and motiva-

tions for imposing a restraint. See Areeda &

Hovenkamp, supra, at 422-23; see also California

Dental Ass’n, 526 U.S. at 792 (Breyer, J., concurring

in part & dissenting in part) (explaining that

defendants have incentive to introduce evidence of

procompetitive justifications, if such evidence exists).

By contrast, a government enforcer usually does not

possess such information, and would thus be put in the

untenable position of needing to speculate as to every

possible rationalization that a defendant might assert

for its conduct.

Indeed, this information gap between the litigants

would pose particularly difficult hurdles for the

government during early stages of litigation. If

proving a “net anticompetitive effect” is required at

the prima facie stage, as the Second Circuit held, then

the government might need to plead such a “net effect”

plausibly in its complaint simply to survive a motion

to dismiss. See Hovenkamp, Rule of Reason, supra, at

17, 20. See generally Bell Atl. Corp. u. Twombly, 550

U.S. 544, 570 (2007) (discussing minimum pleading

standards). But the government often does not know

prior to bringing suit whether a restraint might create

procompetitive effects in the relevant market—let

alone whether the restraint has procompetitive

impacts on one side of a platform that redound to the

benefit of the market at issue. By contrast, because the

defendant is “in the best position to suggest the

benefits that might flow from [its] activities,” the

plaintiff and court are entitled to presume that no

32

such benefits exist if the defendant cannot establish

them. Areeda & Hovenkamp, supra, at 422.

Accepting the Second Circuit’s novel rule would

thus raise the risk of courts dismissing valid

enforcement actions at the outset, even where the

defendant’s conduct imposes an obvious anticompeti-

tive effect on one side of the platform. Such an

approach could significantly harm competition in

many sectors of the economy by shielding

anticompetitive behavior by dominant firms in

industries involving two-sided platforms.

B. Maintaining the Proper Assignment of

Burdens Ensures That Potentially

Procompetitive Conduct Is Not Chilled

by Antitrust Enforcement Efforts.

The court of appeals required the government to

disprove procompetitive effects at the prima facie

stage based on the concern that, absent such a prima

facie showing, antitrust lawsuits might chill genuinely

procompetitive activities involving two-sided platforms.

Specifically, the Second Circuit reasoned that ignoring

interconnections between the two sides of a platform

might prevent a firm from engaging in conduct that

enhances output on both sides of a platform because of

the fear that the firm’s conduct could have an

apparent anticompetitive effect (e.g., higher prices) on

one side of the platform. (Pet. App. 35a (stating that

focusing on competitive effects in only the merchant-

services market would penalize legitimate competitive

activities on the cardholder side “no matter how

output-expanding such activities may be”).)

The court of appeals’ concern about protecting the

incentives of firms to engage in procompetitive

33

activities is legitimate. But this concern is already

fully addressed by the established burden-shifting

framework. If a plaintiff establishes a prima facie case

of anticompetitive harm based on effects on one side of

a platform, the defendant has a full and fair

opportunity to establish that its conduct creates

procompetitive effects in the relevant market. Such

procompetitive effects might include that its restraint

enhances the platform’s output—an effect that could

benefit both sides of the platform. And if a defendant

establishes a procompetitive effect in the relevant

market, the parties may then attempt to show whether

or not this procompetitive effect can be achieved

through less restrictive means. See supra at 7-8. See

Hovenkamp, Rule of Reason, supra, at 25. The

established burden-shifting framework thus already

protects against antitrust analysis too abbreviated to

consider fully the competitive factors at play in a

market. See California Dental Ass n, 526 U.S. at 793-

94 (Breyer, J., concurring in part & dissenting in part).

34

CONCLUSION

The decision and order of the court of appeals

should be reversed and remanded for further

proceedings.

Respectfully submitted,

ERIC T. SCHNEIDERMAN

Attorney General

State of New York

BARBARA D. UNDERWOOD*

Solicitor General

BEAU W. BUFFIER STEVEN C. WU

Bureau Chief Deputy Solicitor General

ELINOR . HOFFMANN JUDITH N. VALE

Deputy Bureau Chief Senior Assistant

JEREMY R. KASHA Solicitor General

Assistant Attorney General barbara.underwood@ag.ny.gov

Antitrust Bureau

December 2017 * Counsel of Record

(Counsel listing continues on next page.)

JAHNA LINDEMUTH

Attorney General

State of Alaska

1031 W. Fourth Ave.

Suite 200

Anchorage, AK 99501

XAVIER BECERRA

Attorney General

State of California

1300 I St.

Sacramento, CA 95814

MATTHEW P. DENN

Attorney General

State of Delaware

Department of Justice

820 N. French St.

Wilmington, DE 19801

DOUGLAS S. CHIN

Attorney General

State of Hawai

425 Queen St.

Honolulu, HI 96813

CURTIS T. HILL, JR.

Attorney General

State of Indiana

200 W. Washington St.

Room 219

Indianapolis, IN 46204

35

ANDY BESHEAR

Attorney General

State of Kentucky

700 Capitol Ave., Suite 118

Frankfort, KY 40601

JANET T. MILLS

Attorney General

State of Maine

6 State House Station

Augusta, ME 04333

MAURA HEALEY

Attorney General

Commonwealth of

Massachusetts

One Ashburton Pl.

Boston, MA 02108

LORI SWANSON

Attorney General

State of Minnesota

102 State Capitol

75 Rev. Dr. Martin Luther

King Jr. Blvd.

St. Paul, MN 55155

JIM Hoop

Attorney General

State of Mississippi

Post Office Box 22947

Jackson, MS 39225

HECTOR H. BALDERAS

Attorney General

State of New Mexico

408 Galisteo St.

Santa Fe, NM 87501

JOSH STEIN

Attorney General

State of North Carolina

Department of Justice

114 W. Edenton St.

Raleigh, NC 27603

ELLEN F. ROSENBLUM

Attorney General

State of Oregon

1162 Court St. N. E.

Salem, OR 97301

JOSH SHAPIRO

Attorney General

Commonwealth of

Pennsylvania

36

ALAN WILSON

Attorney General

State of South Carolina

1000 Assembly St.

Columbia, SC 29201

BOB FERGUSON

Attorney General

State of Washington

800 Fifth Ave., Ste. 2000

Seattle, WA 98104

BRAD SCHIMEL

Attorney General

State of Wisconsin

Department of Justice

P.O. Box 7857

Madison, WI 53703

KARL A. RACINE

Attorney General

District of Columbia

Suite 650 North

Strawberry Square, 16th Fl. 441 4th St., NW

Harrisburg, PA 17120

Washington, DC 20001