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 United States

STATE OF OHIO, ET AL.

Petitioners,

¥V.

AMERICAN EXPRESS COMPANY & AMERICAN EXPRESS

TRAVEL RELATED SERVICES COMPANY, INC.,

Respondents.

ON PETITION FOR A WRIT OF CERTIORARI

TO THE UNITED STATES COURT OF APPEALS

FOR THE SECOND CIRCUIT

BRIEF OF 28 PROFESSORS OF ANTITRUST LAW AS

AMICI CURIAE SUPPORTING PETITIONERS

ERIC F. CITRON

Counsel of Record

GOLDSTEIN & RUSSELL, P.C.

7475 Wisconsin Ave.

Suite 850

Bethesda, MD 20814

(202) 362-0636

ec@goldsteinrussell.com

TABLE OF CONTENTS

rE IE MINIT cites edecedtnsnanenncenssucdesasiensiecnescones 1

SUMMARY OF ARGUMENT. ................::cccccsecssceeseeees 3

I. The District Court Correctly Applied Rule-Of-

Reason Analysis To The Facts It Found. ............. 5

A. How the rule of reason works. .....................+. 5

1. Step one and “market power’ ................. 5

A EERSTE Te ee OE 10

RRR TAN a ret ee ee OE ee 11

B. The district court’s analysis......................... 12

Il. The Second Circuit’s “Two-Sided-Market”

Innovations Are Fundamentally Unsound. ....... 16

A. The relevant market cannot encompass

both sides of Amex’s platform...................... 17

B. “Netting” or “balancing” competitive

effects across both sides of a two-sided

platform is fundamentally unsound............ 20

1. The Second Circuit's “netting”

exercise iS EFTOMCOUS. ................cccceeeee0es 20

2. A special “net” price or benefit rule

for two-sided platforms is an

unnecessary invitation to error in the

SPIES iccuacsieicdatidcabsaginiaitdadstosclanstdutesti 27

IED Satciiatiituniniinseeminmieninndnsianesmncesintbicins 35

33

TABLE OF AUTHORITIES

Cases

Brown Shoe Co. v. United States,

Fe NB CE raierineevetintvisehcosscesotecndensseuass 17, 33

Cal. Dental Ass’n v. F.T.C.,

i ee En 5

Eastman Kodak v. Image Technical Servs., Inc.,

ee rey Se i rcieitcincttctcnctecidissitcusicovmstie’ 6, 7,17

F.T.C. v. Actavis, 133 S. Ct. 2223 (2013) ................000. 1

F.T.C. v. Indiana Fed’n of Dentists,

og eS fe 7, 13, 24

F.T.C. v. Phoebe Putney Health Sys.,

ES oe Ree ane 1

Hanover Shoe v. United Shoe Machinery Corp.,

Be ey ie ai cctesncictnccnscnesteccecedilinembinensamiaitivn 35

Illinois Brick Co. v. Illinois,

ERE a eee eR td 35

Illinois Tool Works v. Independent Ink,

ee Ii ciciuiciictnioictetatpiensbienavscticenspaddenioicle 9

Kimble v. Marvel Entm’t,

EIR Tear CC 1

Leegin Creative Prods. v. PSKS,

Re eG ee CE cicsccebessessceatsttesnetsimennipiisssnciiesiiabdans 1, 16

Nat'l Soc. of Prof. Eng’rs v. United States,

ae is I eipniticedisnniesticetciecstielins snd dpsiensetn dee 25

NCAA v. Bd. Of Regents of Univ. of Okla.,

gS Et enn 11, 24, 30, 33

Novell, Inc. v. Microsoft Corp.,

731 F.3d 1064 (10th Cir. 2013)................cccccscccceees 1

iii

Pac. Bell Tel. Co. v. Linkline Commce’ns Inc.,

BBB U.S. 438 (2009) .............cceccccceecceeeeeerereeeeeeeeeeennees 1

Times-Picayune Publ’g Co. v. United States,

BOE UB, BOE (1GBB) .....ccccsseeeeverscscercescesosscresseseseesees 22

Todd v. Exxon Corp.,

275 F.3d 206 (2d Cir. 2001) .........ccccccceeceeceeeeeeneneneees 7

United States v. E.1. DuPont de Nemours & Co.,

SEL UD. BIT (1BBG) nccccrcccsseveccsesscccccscccsscsesssvesseessess 13

United States v. Microsoft,

253 F.3d 34 (D.C. Cir, 2001) ........cccccccccceeeeeeeennnneenes )

United States v. Socony-Vacuum Oil Co.,

910 U.S. 160 (1940) ......cccccerecssrscrsevssssecrrererersssesseees 24

Verizon Comme’ns v. Law Offices of Curtis V.

Trinko, 540 U.S. 398 (2004) ..........ccccccccceeeneeenneeeereres 1

Other Authorities

D.H. Ginsburg & J. Wright, Dynamic Analysis

and the Limits of Antitrust Institutions,

78 Antitrust LJ. 1 (2012) .......ccccccereeeeerereeeeeeeeeneees 34

Evans & Schmalensee, /ndustrial Organization

of Markets with Two-Sided Platforms,

3 COMPETITION POL’Y INT'L 150 (2007) ..............+ 21

Frank Easterbrook, The Limits of Antitrust,

63 Tex. L. Rev. 1 (1984) ........ccccccceerereeeeeeeereeeeneennees 27

Herbert Hovenkamp, The Rule of Reason

(Fla. L. Rev. forthcoming 2018)... )

Louis Kaplow, Why (Ever) Define Markets?,

124 Harv. L. Rev. 437 (2010) ...........ccccccccecceereeeennnees 8

Michael Carrier, The Rule of Reason: An

Empirical Update For The 21st Century,

16 Geo. Mason L. Rev. 827, 827-29 (2009).............. 5

iv

Rochet & Tirole, Two—Sided Markets: A Progress

Report, 37 RAND J. Econ. 645 (2006) .................. 21

US DOJ & FTC, Horizontal Merger Guidelines

Gi iisicciervemncsoneniumiicbeanidbinidmlasidiasiidinidemaasabememens 13, 34

William Landes & Richard Posner, Market

Power in Antitrust Cases, 94 Harv. L. Rev. 937

Phillip Areeda & Herbert Hovenkamp, Antitrust

Law: An Analysis of Antitrust Principles and

their Application (4th ed. 2017).........ccccccccees passim

INTEREST OF AMICI’

Amici are 28 antitrust scholars who write to

share their disinterested perspective with the Court.

Their respective backgrounds are compiled in the ad-

dendum, but it suffices to note that they include

many leaders in the field, including the author of the

treatise most often relied upon by this Court’s anti-

trust opinions. See, e.g., Kimble v. Marvel Entm’t,

135 S. Ct. 2401, 2415 (2015) (Alito, J. dissenting);

F.T.C. v, Actavis, 133 8. Ct. 2223, 2227 (2013) (Brey-

er, J.); Pac. Bell Tel. Co. v. Linkline Comme’ns Inc.,

555 U.S. 438, 453 (2009) (Roberts, C.J.); Leegin Crea-

tive Prods. v. PSKS, 551 U.S. 877, 894 (2007) (Ken-

nedy, J.); F.T.C. v. Phoebe Putney Health Sys., 568

U.S. 216, 228 (2013) (Sotomayor, J.); Verizon

Comme’ns v. Law Offices of Curtis V. Trinko, 540

U.S. 398, 411 (2004) (Scalia, J.); Novell, Inc. v. Mi-

crosoft Corp., 731 F.3d 1064, 1070 (10th Cir. 2013)

(Gorsuch, J.) (citing Phillip Areeda & Herbert

Hovenkamp, Antitrust Law: An Analysis of Antitrust

Principles and their Application (“Areeda &

Hovenkamp”)).’

Amici’s interest in this case runs deep for two

reasons. First, this Court has encountered few op-

portunities to offer substantial guidance to the lower

courts on how to apply the rule of reason. It is thus

' No party’s counsel authored this brief in any part and no

one other than amici or their counsel made any contribution to

the preparation or submission of this brief. All relevant parties

have provided blanket consent to briefs of amici curiae.

* All citations to the treatise are to the online edition that

includes the 2017 supplement, and made by 4 number.

2

particularly important that the Court’s opinion here

make that doctrine—which already confounds many

lower courts—clearer rather than murkier. Second,

the ideas introduced by the Second Circuit’s decision

below are particularly pernicious; they are small but

potentially serious fractures to important, structural

bones of antitrust doctrine that will cause lasting

harm if not properly set by this Court.

As the district court’s careful analysis showed,

existing doctrines already provide accurate answers

to questions like those this case posed. And as is of-

ten true, the introduction of special rules to deal with

individual cases or isolated phenomena will tend to

do much more harm than good. Special rules are ob-

viously much more harmful if they are themselves

inaccurate—as the Second Circuit’s were. But

whether theoretically accurate or not, they also in-

troduce occasions for lower courts to miss the forest

for a tree they have misunderstood and that was nev-

er necessary to plant.

We strongly believe this is true with respect to

the Second Circuit’s (inaccurate) rules regarding so-

called “two-sided markets.”* These business models

are certainly important to scholars of business strat-

egy—and increasingly so as their prevalence grows in

today’s technological environment. But the prevail-

* Because, as explained infra p.28-31, the language of

“two-sided markets” is confounding for antitrust purposes, we

believe it clearer to refer to the general phenomenon as a two-

sided “platform” or “business model.” The “two-sided-market”

terminology having caught on below, however, we use the terms

interchangeably here.

3

ing view among antitrust economists and scholars is

that two-sidedness does not require changing the set-

tled rules of antitrust law, which can already analyze

competitive issues in such “markets” with straight-

forward doctrinal tools that provide accurate results.

SUMMARY OF ARGUMENT

At bottom, this case concerns two critical, relat-

ed, and mistaken ideas that the Second Circuit intro-

duced into antitrust law and economics, and that this

Court should reject. One is that a “relevant market”

for antitrust purposes must (or even can) include

“both sides” of a two-sided platform; the other is that

competitive harms and benefits must be (or even can

be) “netted” across relevant markets in rule-of-reason

cases. These ideas are not only wrong, but in direct

tension with core principles animating antitrust law.

In contrast, the district court’s opinion straicht-

forwardly applies settled doctrine—accepting the

facts as found, we believe it properly concluded that

there was an antitrust violation in this case by de-

termining that (1) American Express (Amex) had suf-

ficient market power to successfully impose a re-

straint eliminating horizontal price competition

among its competitors, and (2) this led, in turn, to

several anticompetitive effects, including supra-

competitive prices for merchant network services.

While Amex was free to introduce evidence of offset-

ting procompetitive benefits, the district court found

that it did not, in part because the “benefits” it point-

ed t. could not be deemed legitimate or procompeti-

tive. We believe any proper rule-of-reason analysis

should reach the same outcome on the facts as found,

and it is evident that the Second Circuit only did oth-

4

erwise by indulging its two mistaken ideas about

“two-sided” antitrust markets.

This brief thus proceeds in two parts. Part I ex-

plains certain basic tools of antitrust analysis—both

how and why we use them in rule-of-reason cases—

using the district court’s legally sound analysis as a

guide. Part II explains the critical errors in the Sec-

ond Circuit’s two-sided-market analysis: Section II.A

explains briefly why multiple “sides” of a multi-sided

platform or business model cannot be included in the

same relevant market; Section II.B explains that the

Second Circuit’s “netting” exercise is unsound and

ultimately antithetical to the basic purposes of the

Sherman Act. Section II.B.ii also includes an expla-

nation of how existing rule-of-reason doctrine permits

courts to consider relevant arguments about two-

sided platforms in a case like this one, while the Sec-

ond Circuit’s novel approach, even if sound, substan-

tially increases the risk of lower-court error.

Were we to boil all that down to a few sentences,

however, we would just say this. Efficient allocation

of resources in the economy depends on un/istorted

competitive price signals. The correct “balance” be-

tween, say, merchant prices and consumer rewards

in related or even interdependent markets is thus de-

termined by competition itself; that role cannot be

filled by a judicial scale-balancing exercise balancing

different market participants’ welfare because that

lies well beyond any judge’s (or economist’s) ken.

That Amex manipulated this balance through a re-

straint that virtually eliminated horizontal price

competition is the antitrust problem—not a defense—

because antitrust law endeavors to preserve competi-

tion on all sides of any business’s operations, and let

5

the free market’s invisible hand take the wheel from

there.

I. The District Court Correctly Applied Rule-

Of-Reason Analysis To The Facts It Found.

A. How the rule of reason works.

The rule of reason, which is correctly laid out in

petitioners’ briefs, U.S. Br. 20-22; Ohio Br. 20-22, is a

three-step heuristic courts have developed to deter-

mine the lawfulness of an alleged restraint with po-

tentially ambiguous competitive effects. The plaintiff

begins by showing that the restraint is plausibly an-

ticompetitive; the defendant responds by proffering

any legitimate, procompetitive justification it has for

the restraint; and the plaintiff responds by attempt-

ing to prove that any proffered justifications are ei-

ther factually false (i.e., the restraint does not serve

them) or at least available through “a substantially

less restrictive alternative.” Areeda & Hovenkamp

71502. Even if there are no less restrictive alterna-

tives, courts may still find that the restraint is harm-

ful on balance to competition in the relevant mar-

ket—though even this in-market balancing exercise

by courts is (appropriately) rare. See, e.g., Michael

Carrier, The Rule of Reason: An Empirical Update

For The 21st Century, 16 Geo. Mason L. Rev. 827,

827-29 (2009) (finding balancing step reached in unly

2.2% of cases from 1999 to 2009, and 4% of cases from

1977 to 1999). Below, we briefly elaborate these

three steps and the tools antitrust law uses to con-

duct them.

1. Step one and “market power”

The first step—the plaintiffs showing that the

restraint is “prima facie anticompetitive,” Cal. Dental

6

Ass'n v. F.T.C., 526 U.S. 756, 771 (1999)—can be

made in one of two ways. Both implicate, albeit dif-

ferently, the critical concept of market power. The

“direct method” involves direct proof of an actual an-

ticompetitive effect, which proves a fortiori that the

defendant has sufficient market power to bring that

effect about. The “indirect method” involves inde-

pendent proof of both market power and a restraint

that plausibly harms competition.

This Court has said that “market power” is the

power to get a purchaser “to do something that he

would not do in a competitive market."* Eastman

Kodak Co. v. Image Technical Servs., Inc., 504 U.S.

451, 464 (1992). More rigorous economic definitions

are available—the most accepted being the power to

profitably raise price above marginal cost, see Wil-

liam Landes & Richard Posner, Market Power in An-

titrust Cases, 94 Harv. L. Rev. 937, 939 (1981)—and

that definition is frequently helpful in contexts like

merger analysis. But in a case like this, the defini-

tion from Eastman Kodak expresses the critical con-

ceptual point.

Market power matters in a restraint case like

this one because, in its absence, merchants them-

selves could fight anticompetitive contract terms like

Amex’s non-discrimination provisions (NDPs) by

simply diverting their business to competitors. Anti-

trust law prefers trusting competition itself to protect

the market in this way, which competition will do in

* Notably, no coercion need be involved; the “power” comes

solely from the presence (and elasticity) of consumer demand

and the lack of substitute sources of supply.

7

the absence of market power. As Justice Scalia put it,

certain potentially anticompetitive restraints “are

completely without force when the participants lack

market power.” Eastman Kodak, 504 U.S. at 488

(dissent) (citing Areeda & Hovenkamp); see also

F.T.C. v. Indiana Fed’n of Dentists, 476 U.S. 447, 460

(1986) (“[T]he purpose of the ... market power [in-

quiry] is to determine whether an arrangement has

the potential for genuine adverse effects on competi-

tion|.]”). Thus, in a restraint case, the question is

whether the defendant has sufficient market power

to make the alleged restraint effective.

In cases involving the “direct metiod,” like this

one, the market-power showing is effectively sub-

sumed by the plaintiffs direct proof that the alleged

restraint is causing anticompetitive effects. In that

case, we know the defendant has sufficient market

power to generate those effects because it is generat-

ing them. Thus “proof of actual detrimental effects’

... can ‘obviate the need for an inquiry into market

power.” Indiana Fed’n, 476 U.S. at 460-61 (citing

Areeda & Hovenkamp) (emphasis added). In fact, it

can establish that the requisite market power exists

more directly “than calculations of elusive market

share(s]” or other indirect inquiries. Todd v. Exxon

Corp., 275 F.3d 191, 206 (2d Cir. 2001) (Sotomayor,

J.). Notably, as the United States emphasizes, U.S.

Br. 23-34, this case involved very strong proof under

the direct method—especially the evidence that

Amex’s rules entirely suppress horizontal price com-

petition among Amex and its competitors. See

Pet.App. 204a-207a & n.43 (discussing Discover tes-

timony); id. 197a-198a (discussing Amex’s own

acknowledgements of lack of pricing competition).

8

In “indirect method” cases, market power must

be proven independently. Typically, after identifying

the alleged restraint, the factfinder will define the

relevant market in which that restraint operates

(discussed infra pp.12-13, 17-20), and then attempt to

infer market power from the defendant’s estimated

market share and other factors. Particularly in re-

straint cases, this inquiry is a rough proxy at its very

best; sufficient market power to insist on some re-

straints can arise from a small share of a very con-

centrated market (like Amex has here), and can other

times be absent even when market share is very high

(for example, when entry is easy). See, e.g., Areeda &

Hovenkamp 9515 (“A firm could have substantial

market power without accounting for all or even most

of a market. By the same token, the power of a firm

with a dominant market share might be very high or

negligible[.|”); Louis Kaplow, Why (Ever) Define Mar-

kets?, 124 Harv. L. Rev. 437 (2010) (expressing skep-

ticism that market definition truly aids in assessing

market power). Some restraints can also be easier to

impose with less unilateral market power because (as

here) they are good for one’s own competitors as well,

and lead to reinforcing behavior in oligopolistic set-

tings, see, e.g., Areeda & Hovenkamp 99530b, 574a—

as happened when Visa and MasterCard adopted

their own NDPs. Pet.App. 180a. It is thus frequently

preferable to look for direct proof of market power,

which may include (among other things) a defend-

ant’s ability to raise prices repeatedly without losing

share, very high and persistent economic profits, or

ready success in thwarting counter-parties’ efforts to

resist the imposition of disfavored restraints. See

Pet.App. 158a-163a (detailing considerable evidence

in this regard); accord United States v. Microsoft, 253

9

F.3d 34, 58 (D.C. Cir. 2001) (pointing to Microsoft's

pricing conduct and pattern of exclusionary behavior

as direct evidence of monopoly power).

Before moving on, it is critical to emphasize that

there is nothing illegal about having market power.

The Sherman Act does not condemn its mere acquisi-

tion or existence; it is, in this context, only an indica-

tion that an allegedly anticompetitive restraint might

work. Put otherwise, “[w|hile market power is a nec-

essary condition for an anticompetitive restraint un-

der the rule of reason, it is never a sufficient condi-

tion.” Herbert Hovenkamp, The Rule of Reason p.17

(Fla. L. Rev. forthcoming 2018) (“Rule of Reason”),

https://goo.gl/4CL5jx.

The Second Circuit thus erred quite seriously

when it held that Amex’s “cardmember insistence”

could have no bearing on its market power, particu-

larly because it derived from “cardmember satisfac-

tion” and so couldn’t be a bad thing. Pet.App. 45a-

48a. In this case, it was not relevant where Amex’s

market power came from; the law cares only whether

it exists, and so might make a putatively anticompet-

itive restraint possible. Accordingly, if merchants so

wanted Amex that they felt bound to accept it, that

demonstrates market power without regard to why

that dynamic exists.° The district court was thus

° Relatedly, note that creating “worthwhile” products often

creates market power while even a patent “monopoly” on other

products would not. The patent-holder on a cure for cancer

would have enormous market power; the patent-holder on a

new headache remedy likely wouldn’t, given competing alterna-

tives. See Areeda & Hovenkamp 9518e; Illinois Tool Works v.

Independent Ink, 547 U.S. 28 (2006).

10

quite right to regard “cardholder insistence” as strong

evidence of market power, and the Second Circuit

quite wrong to reject it while expressly noting, sever-

al times, that “cardholder insistence is precisely what

makes accepting Amex cards worthwhile for mer-

chants to do.” See id. That which makes “accepting

Amex cards worthwhile” is exactly what gives Amex

market power.

2. Step two

Once the plaintiff makes its prima facie showing

under either the direct or indirect method, the bur-

den shifts to the defendant to proffer any legitimate,

procompetitive justifications it may have for the re-

straint. This burden shift follows from basic princi-

ples of proof that are not specific to antitrust. “The

defendant, being the author of the restraints, is in a

better position to explain why they are profitable and

in consumers’ best interests.” Areeca & Hovenkamp,

41505.

It is important to note, however, that not every

justification that generates welfare somewhere in the

economy—even by lowering prices—can be deemed

“legitimate” or “procompetitive.” Instead, this re-

quirement must exclude (at least) one type of justifi-

cation. Because promoting competition is the goal of

the antitrust laws, “defendants’ expectation of profit”

from a restraint must come “from something other

than a restriction of competition” itself. Rule of Rea-

son, p.22. Put otherwise, “[a]n effective defense must

be able to show that a practice has social benefits

that do not depend on the exercise of market power.”

Id. p.23 (emphasis added). Accordingly, at an abso-

lute minimum, a defendant’s justification cannot be

that it will use a restraint to generate monopoly rents

11

or restrict output in one aspect of its business and

then reinvest that revenue elsewhere—including, of

course, in its own bottom line.

As an example, this Court in NCAA v. Board Of

Regents of University of Oklahoma, 468 U.S. 85, 116-

17 (1984), did not permit the NCAA to defend a re-

striction on televising college football games on the

theory that it would “protect live attendance.” That

justification rested on the view that exercising mar-

ket power and restricting output (i.e., limiting broad-

casts) would lead to benefits elsewhere in the econo-

my, and so was “inconsistent with the basic policy of

the Sherman Act.” Jd. As this Court put it, “the

Rule of Reason does not support a defense based on

the assumption that competition itself is unreasona-

ble.” Id. at 117 (quoting Nat’ Soc’y of Prof Engr’s v.

United States, 435 U.S. 679, 696 (1978)). In contrast,

this Court explained that restrictions aimed at pro-

tecting “the integrity of [the product] as ... distinct

and attractive,” or promoting the NCAA’s non-

competition-related ends (like preserving a “revered

tradition of amateurism in college sports”) could be

legitimate, so long as they rested on evidence that the

restriction pro-competitively increased output in the

relevant market, rather than on the impermissible

intuition that restricting output was somehow “pro-

competitive.” See id. at 116-17, 120.

3. Step three

If legitimate justifications for the restraint are

offered, the burden shifts back to the plaintiff to

demonstrate that a “substantially less restrictive al-

ternative” could achieve the same benefits. Areeda &

Hovenkamp 41502. “By this stage of the controversy,

most cases will be resolved. If not—and rarely—the

12

harms and benefits must be compared to reach a net

judgment whether the challenged behavior is, on bal-

ance, reasonable.” Jd. Notably, this balancing occurs

within the relevant market, see infra pp.20-27; the

factfinder is not asked to balance the welfare of one

set of consumers against another’s somewhere else.

B. The district court’s analysis

The district court applied this settled framework

to the facts as it found them, which are amply de-

scribed in petitioners’ briefs. See U.S. Br. 2-12, 23-

34; Ohio Br. 1-13. Without unduly revisiting those

facts, the following section briefly reviews the district

court’s legal analysis of them to demonstrate how the

rule of reason properly operates in practice.

The court began by defining the relevant mar-

ket. Pet.App. 11la-148a. This was the appropriate

first step for two reasons. First, because the plain-

tiffs purported to carry their prima facie burden

through both the direct and indirect method of proof,

a market definition was needed to assess Amex’s

market power using its market share and other fac-

tors. Second, and more important here, a basic map-

ping of the relevant market is necessary to set the

scope of the ensuing inquiry, even under the direct

method. It is true that, because the “purpose of the

inquiries into market definition and market power is

to determine whether an arrangement has the poten-

tial for genuine adverse effects on competition, proof

of actual detrimental effect[s]” can be “sufficient to

support a finding that the challenged restraint was

unreasonable even in the absence of elaborate market

analysis.” Ind. Fed’n, 476 U.S. at 461. Even then,

however, courts must ultimately establish the basic

boundaries of the market in which the restraint op-

13

erates.° To both ends, market definition must be tied

to the set of substitutes reasonably available to the

market participants at issue if it is to serve its pur-

pose. See infra pp.17-20 (elaborating this point);

United States v. EJ. duPont de Nemours & Co., 351

U.S. 377, 395 (1956) (“In considering what is the rel-

evant market ... no more definite rule can be declared

than that” it consists of “commodities reasonably in-

terchangeable by consumers for the same purpos-

es|.!”).

Here, the district court engaged in rigorous mar-

ket definition, looking to both well-established eco-

nomic tests and a pragmatic analysis of competitive

realities supported by the testimony of market partic-

ipants. It found that the relevant market encom-

passed “general purpose credit and charge card net-

work services” offered to merchants. That market is

highly concentrated; those services are supplied by

only three major players (Amex, Visa, and Master-

Card) and one minor one (Discover). Among other

things, the district court declined to include debit-

card services within the relevant market by deter-

mining that debit- and credit-card services were not

* In cases that use the “direct method,” it may not be nec-

essary to start with market definition because “\e)vidence of

competitive effects can inform market definition, just as market

definition can be informative regarding competitive ef-

fects.” US DOJ & FTC, Horizontal Merger Guidelines §4 (2010).

See also id. (further noting that “analysis need not start with

market definition” because “although evaluation of competitive

alternatives available to customers is always necessary at some

point in the analysis,” some analytic tools “to assess competitive

effects do not rely on market definition”).

14

“reasonably interchangeable” and merchants would

not substitute one for the other. Pet.App. 122a-143a.

Having defined the market, the district court un-

dertook the market-power analysis required for at

least the indirect method of establishing a prima fa-

cie case. After analyzing Amex’s market share in

combination with a set of other relevant factors in-

cluding entry barriers, the high degree of industry

concentration, and the “amplifying effect” of Amex’s

“highly insistent or loyal cardholder base,” it found

that Amex had market power. Pet.App. 150a-165a;

supra pp.9-10. It also relied on direct indicia of mar-

ket power, including Amex’s ability to repeatedly

raise prices without losing share, id. 165a-180a, and

on-point testimony establishing that even the largest

national merchants cannot drop Amex in response to

its NDPs—though some have tried (and rapidly

failed). E.g., Pet.App. 162a-164a.

After finding that Amex had market power, the

district court concluded its step-one analysis by also

finding direct proof of anticompetitive effects. It

found that the challenged NDPs virtually eliminated

horizontal price competition for merchant network

services, impeded entry by low-cost business models,

stifled innovation, and raised prices for both mer-

chants and their customers. Pet.App. 19la-2l4a. It

thus held (correctly, on these facts) that plaintiffs

had satisfied their prima facie burden through both

the direct and indirect method. At a minimum, it is

certain as a matter of antitrust law that a firm’s ex-

ercise of market power to frustrate price competition

among its competitors “to the point of near irrele-

vance,” Pet.App. 195a, at least plausibly causes anti-

competitive harm.

15

The district court then turned to Amex’s prof-

fered legitimate justifications at step two. Pet.App.

228a-258a. It began by carefully parsing Amex’s ar-

gument that the NDPs and their restrictions on

steering were necessary to sustain Amex’s “spend-

centric” business model. Jd. 230a-240a. The district

court determined that, at bottom, this was an argu-

ment that the high merchant discount rates sus-

tained by the NDPs were necessary for Amex to offer

more lucrative rewards to cardholders, effectively

“shiftling) the bulk of interbrand competition in the

credit and charge card industry to the cardholder side

of the platform.” Jd. 234a-241a. The court found this

argument both factually unsupported and fundamen-

tally incompatible with the Sherman Act. Among

other things, it (correctly) cautioned that Amex

seemed to be defending the viability of its own busi-

ness model rather than competition as such, Pet.App.

235a, and that “the law does not permit American

Express to decide on behalf of the entire market

which legitimate forms of interbrand competition

should be available and which should not.” Pet.App.

240a; see supra p.10 (noting that suppression of com-

petition is not a “legitimate” or “procompetitive” justi-

fication).

Because Amex’s proffered justifications were

deemed both unsupported and illegitimate, the court

did not proceed to step three. Instead, it properly

concluded that the profound effect that the NDPs

have on horizontal price competition in the relevant

market for merchant network services—and the re-

sulting supracompetitive prices imposed on mer-

chants by Amex and its competitors—constituted a

violation of the Sherman Act.

16

In our view, this was a prototypical and straight-

forward application of rule-of-reason analysis that

appears to have reached the correct result—at least

on the facts as the district court found them. In fact,

any analysis leading to a contrary result is necessari-

ly suspicious. While vertical restraints like Amex’s

frequently have only limited effects on interbrand

competition (and may promote it), see, e.g., Leegin,

551 U.S. at 890, the district court found here that

Amex’s NDPs had profoundly restricted horizontal

price competition not just between Amex and its

competitors, but among all competing providers.

With the NDPs in place and secured by Amex’s mar-

ket power, none of Amex’s competitors had any incen-

tive to lower price; merchants’ inability to steer cus-

tomers to lower-cost cards (or even give customers

truthful information about the cards’ merchant fees)

had severed the connection between prices to mer-

chants and their quantities of card usage. Pet.App.

228a. In fact, Discover had testified—credibly, from

an economic perspective—that it simply could not

gain share by cutting merchant prices, and so rapidly

raised them instead. Pet.App. 206a. The court found

that supracompetitive prices and other anticompeti-

tive effects had in fact resulted. Pet.App. 191a-214a.

This established a violation of the Sherman Act, and

certainly sufficed for a prima facie case.

Il. The Second Circuit’s “Two-Sided-Market”

Innovations Are Fundamentally Unsound.

Given the district court’s sound analysis, the

Second Circuit’s reversal was erroneous. But the

much more critical point here is that the reasons the

Second Circuit gave for reversing—both of which

sounded in the “two-sidedness” of Amex’s business—

17

are highly problematic. Neither is logically sound

and both threaten to undermine basic principles of

antitrust law and economics. It is imperative the

Court reject both.

A. The relevant market cannot encompass

both sides of Amex’s platform.

As noted above, a relevant market for antitrust

purposes is defined by the identification of reasonable

substitutes for the product at issue. This can be done

with relatively rigorous economic analysis where nec-

essary—including measurements of cross-elasticities

of demand among putative substitutes—but it can

often be done with a more practical analysis of “com-

mercial realities” and participant behavior. See, e.z.,

Eastman Kodak, 504 U.S. at 482; Brown Shoe Co. v.

United States, 370 U.S. 294, 325 (1962); DuPont, 351

U.S. at 394.

Either way, the products that Amex sells on the

two “sides” of its platform do not exhibit the charac-

teristics of “reasonably interchangeable” substitutes

in any respect. Merchants do not buy and have no

use for the services sold to cardholders and cardhold-

ers do not buy and have no use for the services sold to

merchants. There is zero cross-elasticity of demand:

A merchant cannot switch to purchasing cardholder

services in response to an increase in the price of

merchant services, and vice versa. Far from being

substitutes, these services act more as complements

and so cannot belong in the same relevant market.

See Areeda & Hovenkamp 9565 (explaining that the

Court of Appeals “incorrectly concludled| that the

relevant market in which to consider American Ex-

press’s anti-steering rules was not limited to the

market for network |merchant] services but also in-

18

cluded consumers. ... [T]hose two groupings are not

substitutes for one another but rather behave more

as complements.”). Separate markets do not become

a single relevant market for antitrust purposes simp-

ly because one defendant sells two services as part of

the same platform. See id. (Second Circuit “was ap-

parently misled by the fact that Amex obtained reve-

nue from two sources, ... but the fact that a firm ob-

tains its profits from two different, non-substitutable

groups does not serve to place the|m] ... into the

same relevant market”).

As the United States correctly notes, U.S. Br. 35-

40, there is in fact no logical way to include two dif-

ferent “sides” of a company’s platform or business

model in one antitrust market. Among other things,

they do not include the same participants—a point

the district court emphasized in rejecting this argu-

ment. Pet.App. 119a. On the merchant “side,” Amex

competes only with Visa, MasterCard, and Discover

in a highly concentrated market; on the cardholder

“side,” Amex competes with Discover, Chase, Capital

One, Citibank, and thousands of other banks that is-

sue credit cards. Visa has a 45% share on the mer-

chant “side” and essentially a 0% share in issuing

cards or providing cardholder rewards—the individu-

al banks do that. Were it an important consideration

here, how would Visa’s “combined” share of the mer-

chant/cardholder market be calculated? Is it 45%,

22.5%, 0% or something else altogether? There is no

logical or practicable way to provide an answer.

Relatedly, but perhaps more importantly, at-

tempting to combine these two sides of Amex’s busi-

ness into one antitrust relevant market (however

that might be done) would undermine the very rea-

19

sons for performing a market definition in the first

place. We define markets in antitrust analysis to de-

termine whether market power exists in the relation-

ship between two sets of economic counterparties; we

want to know whom the market participants facing a

restraint or price increase might look to in an effort

to discipline the anticompetitive behavior of a seller

through competition itself. One need not measure

cross-elasticity of demand to see that a merchant dis-

satisfied with Amex’s imposition of the NDPs on any-

one accepting Amex cards cannot respond by taking

out a new credit card with Chase or Capital One.

Indeed, generally, the competitive dynamics that

determine the effect of a restraint or price increase

that a company imposes on customers in one aspect

of its business are not predictably affected one way or

another by the nature of competition in that compa-

ny’s other market relationships. This is to say that

effects occurring in other, closely interrelated mar-

kets may ameliorate an anticompetitive effect in the

restrained market, they may exacerbate it, or (as

likely happened here) they may have no plausible ef-

fect at all because of how the restraint itself muffles

competitive price signals. Thus, even if there are

cases where we have good reason to believe that out-

of-market effects cause “feedback” effects in the re-

strained market, see infra p.31, the solution cannot

be to lump two markets for non-substitutes into the

same “market.”

In sum, we care about market definition in re-

straint cases because it sets the appropriate scope of

the rule-of-reason inquiry—establishing the sphere in

which the coupling of market power and an anticom-

petitive practice can cause the kind of distortion and

20

welfare loss that matters to antitrust law. If compe-

tition across that relationship has been harmed or

destroyed, that is sufficient to cause resources to be

diverted inefficiently both within that relationship

and across related parts of the economy in ways that

untainted price signals would avoid. Defining the

market more broadly to include other aspects of the

restraining party’s business will lead to overlooking

those distortionary, anticompetitive effects, and en-

sure that market definition itself fails to answer the

questions for which it is asked.

B. “Netting” or “balancing” competitive ef-

fects across both sides of a two-sided

platform is fundamentally unsound.

The Second Circuit’s holding that a plaintiffs

prima facie case must also show “net” anticompetitive

effects across both sides in a case involving a two-

sided platform is also pernicious. Both logically and

practically, this is deeply unsound. Logically, it ig-

nores that the balance of prices across the platform’s

“sides” should be set by competition, not skewed by

competitive restraints and then excused by ad hoc

judicial balancing. And practically, it asks courts to

perform complex analyses they will find confusing or

impracticable, not the least of which is attempting to

pin down the ill-defined phenomenon of “two-sided

markets” itself.

1. The Second Circuit’s “netting” exercise is

erroneous.

Though hard to define, see infra pp.28-31, two-

sided platforms are certainly a recognizable phenom-

enon, and increasingly important to students of busi-

ness strategy and industrial organization. The dis-

21

trict court concluded that Amex was a two-sided plat-

form because it “sells different products or services to

two separate yet interrelated groups of customers

who, in turn, rely on the platform to intermediate

some type of interaction between them.” Pet.App.

77a-78a (citing Evans & Schmalensee, Industrial Or-

ganization of Markets with Two-Sided Platforms, 3

COMPETITION POL’y INT’L 150 (2007); Rochet & Tirole,

Two-Sided Markets: A Progress Report, 37 RAND J.

Econ. 645 (2006). For now, we adopt this as a work-

ing definition.

Note, however, that it is quite broad and encom-

passes ubiquitous business models like newspapers

(which sell to and unite readers and advertisers),

brokers of all stripes (who sell to and unite buyers

and sellers), cable companies (which sell to and unite

content providers and viewers), and many more.

From a competition-policy perspective, these indus-

tries do share certain characteristi~s: For example,

there is the “cross-platform network effect” the dis-

trict court identified, where “the number of agents or

the quantity of services bought on one side ... affects

the value that an agent on the other side of the plat-

form can realize.” Pet.App. 79a. But these business-

es also vary widely and there may be very little sali-

ence to their two-sidedness in any given case—

especially relative to other traditional considerations

like market concentration and barriers to entry.’

’ Indeed, the most salient aspect of two-sided platforms

from an antitrust perspective may be that their network effects

make entry by new platforms difficult, increasing the prospect

of durable market power.

22

In any event, the ubiquity of two-sided business

models means this Court has confronted them before

and rejected the Second Circuit’s special rule. In

Times-Picayune Publ’g Co. v. United States, 345 U.S.

594 (1953)—which concerned the sale of advertising

space by a newspaper—this Court was faced with a

classic two-sided platform: The paper sells ad space

to advertisers on one side and reporting to readers on

the other. Recognizing that “every newspaper is a

dual trader in separate though interdependent mar-

kets,” id. at 610 (emphasis added), this Court held

that, because the restraint at issue was applied only

in one of them, the decisive question was whether the

defendant had economic dominance in that market

alone. Jd. In other words, this Court held, in a case

where the defendant operated a two-sided platform,

that each side represented a “separate ... market”

and that injuring competition in the restrained mar-

ket alone was sufficient to violate the Sherman Act.

This Court’s analytical approach was correct then

and remains so today; it need only adhere to this

precedent to correctly decide this case.

Times-Picayune aside, however, amici would still

strenuously urge this Court not to approve of any

“netting” or “balancing” analysis across relevant

markets—even if they are “both sides” of a two-sided

platform—because that exercise is fundamentally in-

consistent with the antitrust laws’ core purposes. See

Areeda & Hovenkamp {41505 (criticizing Second Cir-

cuit’s “conclusion that when a restraint is alleged in a

two-sided market, a prima facie case requires the

plaintiff to allege net harm aggregated across both

sides”). This idea finds its most damaging expression

in the Second Circuit’s notion that Amex should be

23

allowed to use its NDPs to obstruct price competition

and keep merchant prices high because “a reduction

in revenue that Amex earns from merchants’ fees

may decrease the optimal level of cardholder bene-

fits.” Pet.App. 49a-50a. Antitrust law and policy

should not even indulge arguendo a defendant’s ex-

cuse that it is robbing Peter to pay Paul; basic anti-

trust policy requires that “competition should choose

the optimal mix of revenue between the two sides”—

not Amex’s near-total obstruction of horizontal com-

petition among Amex and its competitors on one side

or the other. See Areeda & Hovenkamp, 9562e.

For this reason, while we believe this Court

should reverse the Second Circuit and affirm the dis-

trict court’s judgment, we would urge the Court not to

rely in any respect on the district court having al-

ready “balanced” the benefits and harms across

Amex’s merchant and cardholder markets, and/or

having concluded that Amex still harmed competition

in the market as the Second Circuit defined it. Peti-

tioners understandably advance these alternative ar-

guments as litigants, but we believe these analyses

are ultimately unintelligible and should not be en-

couraged even as alternative considerations for fu-

ture factfinders. Ultimately, they can only confuse

the correct analysis.

Notably, Amex’s putatively “procompetitive” jus-

tification that high merchant prices lead to more re-

wards and competition for cardholders is plainly of

the illegitimate form described above, supra p.15—it

clearly depends on the exercise of market power to

work. Amex’s justification for its restraint is simply

that it will extract monopoly rents from merchants in

order to use (some of) them to entice new cardholders

24

to its platform. This is an unmitigated negative from

an antitrust perspective. As the United States cor-

rectly explains, U.S. Br. 45-46, antitrust law must

reject the distortionary effects of dictating prices

through restraints rather than competition because it

disrupts the “central nervous system of the economy,”

United States v. Socony-Vacuum Oil Co., 310 U.S.

150, 226 n.59 (1940).

To put the same point differently, it is important

that lower courts not confuse mere “lower prices”

somewhere in the economy with a “procompetitive

effect.” In important respects, the reality can be the

exact opposite: By “disruptling] the proper function-

ing of the price-setting mechanism”—that is, by using

a restraint to increase prices over here and lower

them over there—a practice necessarily undermines

the competitive process and so can violate the rule of

reason “even absent proof that it resulted in higher

prices.” Indiana Fed’n, 476 U.S. at 461-62. Accord-

ingly, this Court has made clear that “[{a] restraint

that has the effect of reducing the importance of con-

sumer preference in setting price” is inconsistent

“with thle] fundamental goal of antitrust law,”

NCAA, 468 U.S. at 107 & n.30, and that conduct that

“impedes the ordinary give and take of the market-

place,’ and substantially deprives the customer of ‘the

ability to utilize and compare prices” adversely af-

fects competition. Nat'l Soc’y, 435 U.S. at 692-93. In

the end, the goal is not to ensure that somebody bene-

fits from an alleged restraint; rather, it is to ensure

that the challenged restraint is not disrupting compe-

tition in its market, causing a misallocation of re-

sources to or from other areas of the economy.

25

Indeed, indulging a “netting” or “balancing” ap-

proach across two-sided platforms would immediately

vitiate the rationale for the best-known rule in all of

antitrust law—the per se proscription against price

fixing. Cartelists almost always have a story as to

why their price increases or output restrictions are a

net positive for the economy: They alleviate a supply

glut, keep failing firms in business, increase wages,

or minimize waste. Antitrust law accepts none of

these excuses not because they could never be true,

but because we are confident that the distortionary

effects on price signals are bad, even if there are (no

doubt) some parties throughout the economy who

benefit from the cartelists’ behavior. See Socony-

Vacuum Oil, 310 U.S. at 226 n.59. (“Whatever eco-

nomic justification particular price-fixing agreements

may be thought to have, the law does not permit an

inquiry into their reasonableness. They are all

banned because of their actual or potential threat to

the central nervous system of the economy.”). The

law would no more accept those justifications for hor-

izontal price fixing in the context of two-sided plat

forms; courts would certainly condemn the elimina-

tion of horizontal price competition the district court

found here had it been created by agreement among

Amex and its competitors. The distortionary effect of

the cartel would be an antitrust policy problem with-

out regard to whether fully 100% of the resulting

rents are passed over to cardholders. And the very

same principle explains why netting or balancing

across the platform is inappropriate here, too.

Moreover, while the Second Circuit suggests that

Amex passes on high merchant fees to cardholders as

part of its business model, this phrasing somewhat

26

obscures reality. Amex is a profit-maximizing firm,

not a wealth-redistribution engine; to the extent it

increases rewards or decreases fees to cardholders, it

is only because competition for cardholders makes it

so. Notably, that market is healthy: There are thou-

sands of rival firms issuing cards and competing to

win a share of consumers’ wallets, entry is relatively

easy, and there is constant innovation in offers and

rewards models. In contrast, the market is quite un-

healthy on the merchant “side” because it is both

quite concentrated and restricted by Amex’s NDPs.

Eliminating the NDPs will help to heal that market

and so will undoubtedly affect both Amex’s and its

competitors’ bottom lines, because they will now have

to compete for both cardholders and merchants. But

given the robustness of existing competition over

cardholders, there may be no substantial decreases in

reward >xpenditures at all. And even if there are,

antitrust policy prefers to have two healthy markets

rather than one, because that leads to more efficient

resource allocation as between them.

Importantly, none of this is to say that Amex

should be constrained in choosing the pair of prices it

wants to charge merchants and cardholders in its

own business model. The sole point is that it should

not be free to use its market power to prevent mer-

chants from fostering price competition among Amex

and its competing card networks and benefitting from

the lower prices other competing networks might of-

fer merchants as a result. Nor should Amex be free

to choose for consumers whether they prefer to forego

Amex’s rewards in favor of other inducements mer-

chants might offer them for using the merchants’ fa-

vored cards. See Areeda & Hovenkamp {41505

27

(“|Clompetition is what determines how revenue is

assessed with respect to each side. Some card issuers

pursue a strategy of obtaining high market fees while

offering more generous terms to customers, while

others do the opposite. [Amex’s] policy effectively

made customers indifferent to merchant charges and

to the extent those charges could be expected to be

higher, restrained competition|.!”). Contrary to the

Second Circuit’s view, Amex cannot have “a legiti-

mate interest” in restricting free-market forces. The

interaction of those forces—not Amex’s NDPs—must

be allowed to determine the optimal level of both

merchant prices and cardholder rewards.

2. A special “net” price or benefit rule for two-

sided platforms is an unnecessary invita-

tion to error in the lower courts.

It is widely accepted that antitrust law must be

implemented by an imperfect system that forces diffi-

cult economic judgments on lay judges and juries,

and that it must therefore account for the risk of er-

rors that harm competition in its pursuit of consumer

welfare. See, e.g., Frank Easterbrook, The Limits of

Antitrust, 63 Tex. L. Rev. 1 (1984). That insight ap-

plies with special force to the inauguration of a new

set of rules for two-sided markets that would require

courts to attempt to create “net prices” or to aggre-

gate and balance competitive benefits and harms

across a two-sided platform. Even if this were theo-

retically possible, it is certainly impracticable and

likely to create repeated errors in the lower courts

that will contribute to market-wide inefficiency.

For example, efforts to balance competitive ef-

fects across relevant markets or generate “net” prices

face intractable “commensurability” problems. The

28

district court found that Amex itself was unable to

propose a workable net price measurement that ac-

counted for prices on both sides of its platform,

Pet.App. 184a, and that result is hardly surprising.

Amex extracts merchant discount fees in dollars and

pays out rewards in “miles,” “points,” progress toward

status rewards, and other non-monetary benefits.

Amex also charges ever-changing fees and interest

rates that in part reflect the consumer “price” for

card usage. There is thus no ready way to even ap-

proximate the per-transaction dollar price to card-

holders for using their Amex card, let alone “balance”

a cardholder’s price against a merchant’s price for

two different services.

Creating a special rule that permits cross-market

balancing of benefits and harms for “two-sided mar-

kets” will also lead to vexing questions about how

even to define which markets are “two-sided.” Based

on various (easily confused) definitions, so-called two-

sided markets might encompass anything from busi-

nesses with plain-vanilla, vertical supply chains to

what Amex calls its “two-sided transaction market.”

See Pet.App. 78a. These definitions appear to us or-

thogonal to underlying antitrust principles, and so

should not be turned into important doctrinal boxes

that come with different rules and defenses. Instead,

as we explain below, the underlying antitrust princi-

ples can themselves be used, through ordinary anti-

trust analysis, to capture whatever is special about a

business model’s two-sidedness in any given case.

Begin by noting that two-sidedness is ultimately

a description of a business model, not a “market” at

all. Uber, Lyft, taxi cabs, and typical livery services

all compete directly in the market for riders, but

29

their business models are different, and there may or

may not be substitution among them in the driver or

labor market. A rule that permits or requires cross-

platform balancing for cases involving “two-sided

markets” is thus inherently confusing: Does it apply

whenever the defendant operates a two-sided plat-

form, when its closest competitors are also two-sided

platforms, when some competitors are also platforms,

or on some other basis? Why can Uber try to excuse

a restraint that injures its drivers based on benefits

to riders when a livery or cab company could not?

Note, also, that every business has far more than

two “sides.” Apple and Google have similar platforms

for selling online music, but also compete vigorously

for specialized labor, real estate in Silicon Valley,

mobile operating system usage, and more. Suppose

Google imposes a restraint on performers selling mu-

sic in the Google Play store and defends it as generat-

ing higher profits it can use to sell its Pixel phones

more cheaply in competition with iPhones, or to set

its salaries for programmers higher so as to be more

competitive in that market. What makes these exer-

cises in cross-market balancing—which clearly vio-

late the Court’s focus on individual markets in Times-

Picayune and NCAA—any different from Amex’s ar-

gument here?

Next, imagine an online consignment operation

for consumer electronics that returns a discounted

portion of every sale to the original owner, while of-

fering buyers a reward for every transaction. So far

as we can tell, this is a “two-sided transaction mar-

ket” that mirrors Amex’s business model quite pre-

cisely. Pet.App. 78a. But even froin a colloquial per-

spective, what “market” does this consignment busi-

30

ness occupy? Is it a consumer-electronics market? A

used consumer-electronics market? A consignment-

based, used consumer-electronics sales market? An

online electronics resale platform market? Merely

adverting to the company’s two-sided business model

tells us nothing about its markets, or whether it qual-

ifies for a special cross-market balancing rule.

The natural answer in terms of antitrust market

definition, of course, is that putting this business in a

market depends on what the case is about. The rele-

vant market depends on why we care: If our online

consignment operation is merging with a brick-and-

mortar retailer of new electronics, it may cause unac-

ceptable concentration in the market for consumer

electronic sales (where both parties compete), but

probably not in the market for repurchasing used

electronics from original owners (because only one

operates there). It may also cause no problems at all,

depending on the nature of substitution between new

portant point, however, is that the antitrust answer

to the question of what market our consignment

business occupies may lie on either side of a compa-

ny’s “platform” and may encompass business models

with ordinary vertical supply chains or not, all de-

pending on the case or restraint at issue. For similar

reasons, there is no way to know ex ante whether the

two-sidedness of a defendant's platform will affect

market definition, market power, anticompetitive ef-

fects, or anything else antitrust law cares about. The

two-sided-market category is the tail, not the dog.

That said, the United States is certainly correct

that the two-sidedness of a company’s business can

be relevant in any given case. U.S. Br. 50-54. Its

31

relevance, however, is already captured by the ordi-

nary tools of antitrust analysis as the district court

applied them. Clarifying how those tools work—

rather than adding new, hard-to-implement rules for

new, hard-to-define categories—is far more likely to

help the lower courts avoid serious errors.

For example, it is already understood that when

two aspects of a company’s business model are closely

related or “interdependent,” that may mean that

competitive effects or anticompetitive distortions on

one side of its business will strongly affect the other

over time, and vice versa, causing “feedback” effects

that may ameliorate or exacerbate the original dis-

tortion. See, eg., Areeda & Hovenkamp {562e.

These “dynamic” considerations or “feedback” effects

may constrain or reinforce anticompetitive exercises

of market power; it depends entirely on industry-

specific context. In any case, however, a defendant is

free to argue that its market power in one market is

illusory in fact because the close interrelationship be-

tween that market and another disciplines its ability

to raise prices or impose anticompetitive conditions

in the relevant market.

But, importantly, if that argument is presented,

it requires no special two-sided-market rules to ana-

lyze; all it needs are the conventional tools for decid-

ing whether market power exists or not. And partic-

ularly where (as here) the proof of market power and

anticompetitive effects is established through direct

evidence, there is no reason to worry that we have in-

correctly assessed a defendant’s power in the relevant

market by ignoring a price effect somewhere else in

the economy. We know the requisite market power

exists because the anticompetitive effect occurred.

32

Indeed, notice that Amex could not possibly

make such an argument here, because the very anti-

competitive effect of its restraint is to sever the con-

nection between the merchant and cardholder sides

of its platform. Competition for cardholder transac-

tions does not benefit merchants (or constrain the ex-

ercise of market power over them) because merchants

lack the ability to steer cardholders away from cards

that charge them a higher price. The two sides of the

Amex platform may be interdependent in some ways,

but—given the NDPs—they are not interdependent

in the important sense that competition for cardhold-

ers will (or even can) feed back into improved pricing

conditions or competitive dynamics on the merchant

side. The restraint itself prevents the feedback.

Defendants can also potentially argue that the

feedback effect is one that improves the competitive

process operating on both aspects of its two-sided

platform at once through a conventional step-two,

rule-of-reason argument. We take this to be the

United States’ suggestion that the defendant can

show “at the second step” that the restraint is “rea-

sonably necessary to achieve legitimate pro-

competitive benefits in a closely related and interde-

pendent market.” U.S. Br. 52 (emphasis added).*

This entails a frankly difficult showing that the “dy-

namic” effects of a restraint are ultimately good for

* Formally, this does not entail balancing of out-of-market

“benefits” with in-market “harms”—an analysis the United

States itself rejects. U.S. Br. 41, 45-46. Instead, it entails look-

ing to out-of-market effects in interdependent markets because

they may improve (or reflect improved) competition in the rele-

vant market itself.

33

the restrained party in the medium-run in a way the

antitrust laws can accept. Often, this requires show-

ing that a product could not exist without the re-

straint. And, importantly, that showing needs to be

made for the product—not just the defendant’s ver-

sion of it—to ensure that this argument does not re-

duce to the view that “competition should be re-

strained because it would hurt my ability to compete

with my competitors (and their potentially superior

business models).” That argument, of course, is the

one this Court has rightly rejected as unacceptable

throughout antitrust law. See supra pp.10-11 (dis-

cussing NCAA); Brown Shoe, 370 U.S. at 320 (anti-

trust law is “concernied] with the protection of com-

petition, not competitors”) (emphasis added).®

Ultimately, we believe the consideration of out-

of-market benefits in cases involving two-sided plat-

forms (or other interdependent markets) is at best an

oblique way of getting to the simpler question of

whether competition continues to protect or ultimate-

ly benefits the restrained parties, and while such

“dynamic” analyses might theoretically prove profita-

ble, we doubt the game is worth the several boxes of

candles it will take the lower courts to play it. The

“out-of-market benefit” inquiry tends in the direction

of weighing the welfare of one set of consumers

against the welfare of another, which is what anti-

trust law seeks to avoid in favor of reliance on compe-

* In any event, the general condition on this showing is the

one emphasized at the beginning, supra pp.10-11, and that the

Court must make clear: A legitimate justification must not de-

pend on the existence of market power to work. See Areeda &

Hovenkamp 41505.

34

tition itself. Meanwhile, the hypothetical causal

chain that makes an out-of-market benefit procom-

petitive overall is typically attenuated and will be

hard for lower courts to follow, so this Court should

regard the risk of lower-court error as both high and

very costly. Because the ultimate concern remains on

the avoidance of competitive distortions within any

properly defined set of market relationships—that is,

within a relevant market—the better course is for

this Court to simply instruct the lower courts to look

for evidence of off-setting benefits solely within the

relevant market itself."° That is particularly so be-

cause the kind of multi-step causal tracing exercise

involved in the consideration of out-of-market bene-

fits is precisely the kind of complexity this Court’s

antitrust rules have endeavored to avoid. See, e.g.,

Illinois Brick Co. v. Illinois, 431 U.S. 720 (1977);

Hanover Shoe, Inc. v. United Shoe Machinery Corp.,

392 U.S. 481 (1968).

Whatever the Court says in this regard, however,

it should make clear that it cannot help Amex for the

reasons the district court gave below. Amex’s fun-

damental argument is that it must prevent steering

away from its cards in order to keep its rewards high

*° Accord Douglas H. Ginsburg & Joshua Wright, Dynamic

Analysis and the Limits of Antitrust Institutions, 78 Antitrust

L.J. 1 (2012) (concluding that courts should hesitate in incorpo-

rating “dynamic effects” analysis, even if theoretically useful,

because of high risk of error); cf. Horizontal Merger Guidelines

§10 n.14 (noting that agencies properly consider any anticom-

petitive effect in a relevant market sufficient to challenge a

merger, and consider “efficiencies not strictly in the relevant

market” only as a matter of prosecutorial discretion).

35

and support its business model. That argument de-

pends on using an exercise of market power to raise

prices and direct the benefits elsewhere, and imposes

a restraint that ensures that those benefits cannot

redound to merchants’ benefit through the competi-

tive process. It is thus neither legitimate nor pro-

competitive, and so the district court properly reject-

ed it. Pet.App. 234a-24la. This Court should thus

reverse the Second Circuit’s decision and, for the

sake of clarity in future cases, affirm the sound appli-

cation of the rule of reason by the district court.

CONCLUSION

This Court should reverse the Court of Appeals

and affirm the judgment of the district court.

Respectfully sub.aitted,

Eric F. Citron

Counsel of Record

GOLDSTEIN & RUSSELL, P.C.

7475 Wisconsin Ave.

Suite 850

Bethesda, MD 20814

(202) 362-0636

ec@goldsteinrussell.com

December 14, 2017

ADDENDUM

la

ADDENDUM

Identity of Amici Curiae

The amici listed below are distinguished

antitrust law professors and scholars. University

affiliations are listed only for purposes of

identification. Listed professors are acting only in

their individual capacities and do not purport to

represent the views of their universities.

Herbert Hovenkamp, James G. Dinan

Professor at the Law School and the Wharton

School of the University of Pennsylvania. He

has been the Rockefeller Foundation Fellow,

Harvard Law School; Fellow of the American

Council of Learned Societies, Harvard Law

School; Faculty Scholar, University of Iowa;

Presidential Lecturer, University of lowa; and

the recipient of the University of Iowa

Collegiate Teaching Award. He is the senior

surviving author of Antitrust Law (formerly

with Phillip Areeda & Donald Turner),

currently 22 volumes.

Harry First, Charlies L. Denison Professor of

Law at New York University School of Law

and Co-Director of the law _— school's

Competition, Innovation, and Information Law

Program. From 1999-2001 he served as Chief

of the Antitrust Bureau of the Office of the

Attorney General of the State of New York.

Professor First is the co-author of the casebook

Free Enterprise and Economic Organization:

Antitrust (7th Ed. 2014). He was twice a

2a

Fulbright Research Fellow in Japan and

taught antitrust as an adjunct professor at the

University of Tokyo. Professor First is a

contributing editor of the Antitrust Law

Journal, foreign antitrust editor of the

Antitrust Bulletin, a member of the executive

committee of the Antitrust Section of the New

York State Bar Association, and a member of

the advisory board and a Senior Fellow of the

American Antitrust Institute.

Einer R. Elhauge, Petrie Professor of Law at

Harvard Law School, where he writes and

teaches on Antitrust Law and Economics.

Professor Elhauge is author of U.S. Antitrust

Law & Economics, co-author of Global

Antitrust Law & Economic, co-author of

Antitrust Law, Vol X with Areeda, Elhauge &

Hovenkamp, editor of the Research Handbook

on the Economics of Antitrust Law, and the

author of articles on antitrust law and

economics that have won awards and appeared

in peer-reviewed economics journals and top

law reviews. He is also President of Legal

Economics, LLC, former FTC Special

Employee on Antitrust Issus, member of the

editorial board for the Competition Policy

International, and member of the advisory

boards for the Journal of Competition Law &

Economics and for the Social Sciences

Research Network on Antitrust Law & Policy.

3a

Eleanor M. Fox, Walter J. Derenberg

Professor of Trade Regulation at New York

University School of Law. She was awarded

an inaugural Lifetime Achievement Award in

2011 by the Global Competition Review for

“substantial, lasting, and transformational

impact on competition policy and practice.” She

received the inaugural award for outstanding

contributions to the competition law

community in 2015 by the Academic Society

for Competition Law, the world network of

academic law and economic competition

experts.

Stephen Calkins, Professor of Law, Wayne

State University. Professor Calkins is the

author of one of the seminal Antitrust text

books — Antitrust Law: Policy and Practice (4th

ed. 2008) (with C. Paul Rogers III, Mark R.

Patterson and William R. Andersen). He is also

the author of Antitrust Law and Economics in

a Nutshell (5th ed. 2004) (with Ernest Gellhorn

and William Kovacic) and served as a co-editor

of the ABA Antitrust Section, Consumer

Protection Law Developments (2009). Professor

Calkins is a life member of the American Law

Institute, a fellow of the American Bar

Foundation and a member of the advisory

boards for the American Antitrust Institute,

Sedona Conference and National State

Attorneys Genera! Program Advisory Project

at Columbia Law School. For the American

Bar Association, he has served on the Councils

4a

of the Sections of Administrative Law and

Regulatory Practice and the Section of

Antitrust Law (two, three- year terms). He is a

former chair of the Association of American

Law School's Antitrust and Economic

Regulation Committee.

Tim Wu, Professor of Law, Columbia Law

School. Professor Wu has cauthored several

books, including "Network Neutrality

Broadband Discrimination" (2003), Who

Controls the Internet (2006), The Master

Switch (2010), and The Attention Merchants

(2016). Wu was a law clerk for Justice Stephen

Breyer and Judge Richard Posner, and has also

worked at the White House National Economic

Council, at the Federal Trade Commission, for

the New York Attorney General, and in the

Silicon Valley telecommunications industry.

Barak Richman, Edgar P. and Elizabeth C.

Bartlett Professor of Law and Professor of

Business Administration at Duke University.

He previously served as » law clerk to Judge

Bruce M. Selya of the Wuited States Court of

Appeals for the First Circuit, and from 1994-

1996 he handled international trade legislation

as a staff member of the United States Senate

Committee on Finance. He writes regularly on

issues related to economics and antitrust.

Professor Richman is the author of Stateless

Commerce, which was published by Harvard

University Press.

5a

Thomas Greany, Visiting Professor, UC

Hastings College of Law. Professor Greaney

was the Chester A. Myers Professor of Law and

Director of the Center for Health Law Studies

at Saint Louis University School of Law. Prior

to joining the SLU Law faculty, he served as an

Assistant Chief in the Department of Justice,

Antitrust Division, specializing in health care

antitrust litigation, and completed a visiting

professorship at Yale Law School.

Peter Carstensen, Fred W. & Vi Miller Chair

in Law Emeritus, University of Wisconsin Law

School. He previously served as an attorney in

the Antitrust Division of the United States

Department of Justice. Professor Carstensen

is also a Senior Fellow of the American

Antitrust Institute.

Spencer Weber Waller, Interim Associate

Dean for Academic Affairs, Professor and

Director for Consumer Antitrust Studies at

Loyola University of Chicago, School of Law.

Darren Bush, Professor of Law and Law

Foundation Professor, University of Houston

Law Center. Professor Bush served as a

co-author with Harry First and the late

John J. Flynn on the antitrust casebook

FREE ENTERPRISE AND ECONOMIC

ORGANIZATION: ANTITRUST (7th Ed.)

with Foundation Press.

6a

Chris Sagers, James. A. Thomas

Distinguished Professor of Law. He is a

member of the American Law Institute, a

Senior Fellow of the American Antitrust

Institute, and a leadership member of the ABA

Antitrust Section.

Robert H. Lande, Venable Professor of Law,

University of Baltimore School of Law.

Professor Lande is a co-founder and a Director

of the American Antitrust Institute, a past

chair of the AALS Antitrust Section, and has

held many positions in the ABA Antitrust

Section. He is also an elected member of the

American Law Institute.

Robin Feldman, Harry & Lillian Hastings

Professor of Law & Director of the Institute for

Innovation Law, U.C. Hastings College of Law.

Professor Feldman previously chaired the

Executive Committee of the Antitrust Section

of the American Association of Law Schools

and clerked for The Honorable Joseph Sneed of

the U.S. Court of Appeals for the Ninth

Circuit. She is also a Fellow of the American

Antitrust Institute.

Jeffrey Harrison, Huber C. Hurst Eminent

Scholar Chair in the Levin Collage of Law at

the University of Florida. He is the co-

author of Understanding Antitrust and

its Economic Implications, (6th ed.,

Matthew Bender, 2013) with E.T.

Sullivan.

7a

Professor Harrison's casebook on Law and

Economics is in the third edition. His Nutshell

on Law and Economics is in its sixth edition.

John B. Kirkwood, Professor of Law, Seattle

University School of Law. He is a Senior

Fellow of the American Antitrust Institute and

an Adviser to the Institute of Consumer

Antitrust Studies. Professor Kirkwood

previously directed the Planning Office, the

Evaluation Office, and the Premerger

Notification Program at the FTC's Bureau of

Competition in Washington, D.C. and later

managed cases and investigations at the

Northwest Regional Office.

Joshua P. Davis, Associate Dean for

Academic Affairs, Director of the Center for

Law and Ethics, Professor, and Dean's Circle

Scholar, University of San Francisco, School of

Law. Dean Davis is on the board for the

American Antitrust Institute, and he

previously served as a Fellow at the Center for

Applied Legal Studies at Georgetown

University Law Center and as the clerk to the

Hon. Patrick E. Higginbotham on the Fifth

Circuit Court of Appeals.

Norman W. Hawker, Professor of Finance

and Commercial Law, Western Michigan

University. He is also a Senior Fellow of the

American Antitrust Institute.

8a

Max Huffman, Professor of Law and Director

of Corporate and Commercial Law Graduate

Certificate program, University of Indiana,

Rober H. McKinney School of Law.

Warren Grimes, Associate Dean for Research

and Irving D. and Florence Rosenberg

Professor of Law, Southwestern Law School.

Dean Grimes is co-author of the definitive

antitrust law text for lawyers and law

students, The Law of Antitrust: An Integrated

Handbook with the late Professor Lawrence

Sullivan. Dean Grimes has chaired the Los

Angeles County Bar Association Antitrust and

Trade Regulation Section and is a member of

the Executive Committee, and he serves on the

Advisory Board of the American Antitrust

Institute.

Mark R. Patterson, Professor of Lav.,

Fordham University School of Law. Professor

Patterson has also been a visiting professor at

several law schools in the U.S. and at Bocconi

University in Milan. He was a co-author of

Antitrust Law: Policy and Practice (4th ed.

2008) (with C. Paul Rogers III, Stephen

Calkins, and William R. Andersen) and is the

author of the forthcoming book Antitrust Law

in the New Economy: Google, Yelp, LIBOR,

and the Control of Information (Harvard

2017).

9a

Marina Lao, Professor of Law, Seton Hall

Law. Professor Lao was previously awarded a

Fulbright Fellowship. She currently serves as

a member of the advisory board of the

American Antitrust Institute, and was Chair of

the Section of Antitrust and Economic

Regulation of the Association of American Law

Schools.

Michael A. Carrier, Professor of Law,

Rutgers Law School. Professor Carrier is a co-

author of the leading [P/antitrust treatise, JP

and Antitrust Law: An Analysis of Antitrust

Principles Applied to Intellectual Property Law

(2d ed. 2009, and annual supplements, with

Hovenkamp, Janis, Lemley, and Leslie). He is

a member of the Board of Advisors of the

American Antitrust Institute and is a past

chair of the Executive Committee of the

Antitrust and Economic Regulation section of

the Association of American Law Schools.

Edward Cavanagh, Professor of Law, St.

John’s University. Professor Cavanagh is

currently a member of the Council of the ABA

Antitrust Section. He has previously served as

co-chair of the ABA Antitrust Section Public

Service Committee. He has also served as co-

chair of the Antitrust Section’s Civil Practice

and Procedure Committee. Professor

Cavanagh is a past chair of the New York

State Bar Association Antitrust Section and

currently a member of its Executive Committee.

10a

Professor Cavanagh is a member of the

Association of the Bar of the City of New York

and has served on its Antitrust and Trade

Regulation Committee and its Federal Courts

Committee

Barak Orbach, Professor of Law and Director

of the Business Law program, University of

Arizona, James E. Rogers College of Law.

Professor Orbach is the author of a leading

casebook: Regulation: Why and How the State

Regulates (Foundation Press, 2012). Professor

Orbach previously served as an Advisor for

Law & Economics to the Israeli Antitrust

Commissioner.

Jon Baker, Research Professor of Law at

American University Washington College of Law.

Professor Baker served as the Chief Economist of

the Federal Communications Commission from

2009 to 2011, and as the Director of the Bureau of

Economics at the Federal Trade Commission

from 1995 to 1998. Previously, he worked as a

Senior Economist at the President’s Council of

Economic Advisers, Special Assistant to the

Deputy Assistant Attorney General for

Economics in the Antitrust Division of the

Department of Justice, an Attorney Advisor to

the Acting Chairman of the Federal Trade

Commission, and an antitrust lawyer in private

practice. Professor Baker is the co-author of an

antitrust casebook, a past Editorial Chair of

Antitrust Law Journal, and a past member of the

Council of the American Bar Association’s Section

of Antitrust Law. He has published widely in the

fields of antitrust law, policy, and economics.

lla

e Andrew Chin, Professor of Law, University

of North Carolina School of Law. Professor

Chin is the recipient of a Rhodes Scholarship

and a National Foundation Graduate

Fellowship. He clerked for Judge Henry H.

Kennedy Jr. of the United States District

Court for the District of Columbia and

assisted Judge Thomas Penfield Jackson and

his law clerks in the drafting of the findings of

fact in United States v. Microsoft Corporation.

e Thomas J. Horton, Professor of Law and

Heidepriem Trial Advocacy Fellow at the

University of South Dakota School of Law.