Replacing the Structural Presumption

• Author: Louis Kaplow
• Position: Harvard University and National Bureau of Economic Research
• Pages: 565-627

REPLACING THE STRUCTURAL PRESUMPTION

L

OUIS

K

APLOW

*

A party challenging a horizontal merger in the United States is said to bene-

fit from a structural presumption. Under this rebuttable presumption, the chal-

lenged merger is deemed to be sufficiently likely to substantially lessen

competition, without the challenger having to prove anticompetitive effects, if

the merger would significantly increase concentration in a highly concentrated

market. This structural presumption is associated with the Supreme Court’s

decision in Philadelphia Bank

1

and is instantiated in modified form in the U.S.

Horizontal Merger Guidelines.

2

It is also a subject of significant contemporary

debate

3

and provides the foundation for antitrust reform proposals that are

* Harvard University and National Bureau of Economic Research. I am grateful to the editors

and reviewers, Dale Collins, Scott Hemphill, Devesh Raval, Steven Shavell, and workshop par-

ticipants at Harvard, USC, and USC-Cambridge Virtual Antitrust for helpful discussions and

comments; Alex Blutman, Bryan Poellot, and Alexi Stocker for research assistance; and Harvard

University’s John M. Olin Center for Law, Economics, and Business for financial support. In

addition, I thank numerous lawyers and economists, both in the antitrust agencies and outside,

who have shared their experiences and thereby given me a better sense of how actual practice

may deviate from statements in agency guidelines and court opinions. This article is part of a

larger project, “Rethinking Merger Policy.” Disclaimer: I consult on antitrust matters, and my

spouse is a lawyer who mostly represents financial services firms.

1

United States v. Phila. Nat’l Bank, 374 U.S. 321, 363 (1963) (“[A] merger which produces

a firm controlling an undue percentage share of the relevant market, and results in a significant

increase in the concentration of firms in that market, is so inherently likely to lessen competition

substantially that it must be enjoined in the absence of evidence clearly showing that the merger

is not likely to have such anticompetitive effects.”).

2

See U.S. Dep’t of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines 7–19

(2010) [hereinafter U.S. Merger Guidelines]; see also Guidelines on the Assessment of Horizon-

tal Mergers Under the Council Regulation on the Control of Concentrations Between Undertak-

ings, 2004 O.J. (C 31) 5, ¶¶ 14–21 [hereinafter EU Merger Guidelines]. Although much of this

article’s analysis is applicable to any formulation of the structural presumption, for concreteness

and practical relevance it often focuses on the version stated in the text—looking to a significant

increase in concentration in a highly concentrated market—which is in fact reflected in the U.S.

Horizontal Merger Guidelines and many modern cases. As quoted in the preceding footnote, the

triggering language in Philadelphia Bank instead refers to the merged firm having an undue

share (which is omitted in many modern formulations) and a significant increase in concentration

(which is followed), without mentioning overall market concentration.

3

See, e.g., Peter C. Carstensen, The Philadelphia National Bank Presumption: Merger Analy-

sis in an Unpredictable World, 80 A

NTITRUST

L.J. 219 (2015); Douglas H. Ginsburg & Joshua

565

566 A

NTITRUST

L

AW

J

OURNAL

[Vol. 84

advanced in a recent U.S. House Majority Staff Report on Big Tech and in a

broader Senate bill.

4

The allure of the structural presumption is easy to appreciate. The predic-

tion of a merger’s anticompetitive effects is a costly, time-consuming, com-

plex, and uncertain undertaking. Antitrust enforcement agencies, particularly

early in the investigation of large numbers of proposed mergers, would like to

be able to use proxies, screens, and other shortcuts to provide a provisional

indication of which mergers are likely to be anticompetitive and thus warrant

further scrutiny. Courts likewise stand to benefit from simplification, even

after completion of a full trial, given the aforementioned difficulties, judges’

lack of expertise, and the absence of independent analytical resources.

Unfortunately, although these sensible objectives might seem to be ad-

vanced through the use of a structural presumption, they cannot be achieved

in practice because, upon analysis, the contemplated methodology does not

work in principle. This article explains how the structural presumption is fun-

damentally flawed because of its own internal illogic, its sharp conflict with

the economic analysis of anticompetitive effects, and the unintelligibility of

its associated legal framework. The structural presumption’s failure even as a

preliminary screening device a fortiori renders it unsound as a basis for actual

decision-making. It is therefore necessary to replace the structural presump-

tion—and dangerous to extend and enshrine it as currently proposed.

Part I examines the structural presumption’s internal logic, emphasizing its

fatal reliance on market definition. This dependence is lethal both because of

market definition’s incoherence and because the very need for market defini-

tion contradicts the central point of the presumption. To explain the latter, the

presumption is triggered when concentration (market share) measures are suf-

ficiently high. Courts and agencies have long required, however, that such

concentration be assessed only in a so-called relevant market, which is one

that is chosen after completion of the market definition process. But how is

one to choose the best market definition? In principle and to a significant

degree in practice, this choice is made based on evidence that helps to predict

D. Wright, Philadelphia National Bank: Bad Economics, Bad Law, Good Riddance, 80 A

NTI-

TRUST

L.J. 377 (2015); Herbert Hovenkamp & Carl Shapiro, Horizontal Mergers, Market Struc-

ture, and Burdens of Proof, 127 Y

ALE

L.J. 1996 (2018); John Kwoka, The Structural

Presumption and the Safe Harbor in Merger Review: False Positives or Unwarranted Con-

cerns?, 81 A

NTITRUST

L.J. 837 (2017); Steven C. Salop, The Evolution and Vitality of Merger

Presumptions: A Decision-Theoretic Approach, 80 A

NTITRUST

L.J. 269 (2015); Sean P. Sullivan,

What Structural Presumption?: Reuniting Evidence and Economics on the Role of Market Con-

centration in Horizontal Merger Analysis, 42 J. C

ORP

. L. 403 (2016).

4

See M

AJ

. S

TAFF

R

EP

.

AND

R

ECOMMENDATIONS

, S

UBCOMM

.

ON

A

NTITRUST

, C

OM

.

AND

A

D-

MIN

. L.

OF THE

C

OMM

.

ON THE

J

UDICIARY

, I

NVESTIGATION OF

C

OMPETITION IN

D

IGIT

. M

KTS

.

391–99 (2020) [hereinafter H

OUSE

M

AJORITY

S

TAFF

R

EPORT

]; Competition and Antitrust Law

Enforcement Reform Act of 2021, S. 225, 117th Cong. (Feb. 4, 2021) [hereinafter Senate Bill].

2022] R

EPLACING THE

S

TRUCTURAL

P

RESUMPTION

567

anticompetitive effects. Hence, we have come full circle. The judge assesses a

battle of experts and other evidence relating to anticompetitive effects in order

to define the market, in order to measure concentration, in order to see if the

structural presumption is triggered, in order that we can then presume an-

ticompetitive effects without actually having to consider them. That the struc-

tural presumption is patently illogical at its core is largely ignored.

Part I further demonstrates that matters are worse for reasons related to

deep flaws in the market definition process, however well one attempts to

undertake it. (This is so unless one renders the matter moot via complete re-

verse engineering: that is, choosing the market definition that ratifies an out-

come determined entirely on other grounds.) Market definition throws away

information, redefined markets are useless for analysis, and market shares

cannot coherently be interpreted in the manner that Supreme Court cases de-

mand and merger assessments purport to do. An immediate corollary is that,

when information is particularly scarce—such as when a competition agency

screens merger filings to identify those deserving further scrutiny—market

definition and the structural presumption are counterproductive. One can ill

afford to discard information when it is especially meager to begin with, and

there is no force capable of suspending the laws of logic when a decision

maker would find it convenient to do so.

Part II relates the structural presumption to decades of economics research

on the prediction of mergers’ likely anticompetitive effects. Because the pre-

sumption relies on concentration measures, the focus of this Part’s analysis is

on when and how they—or other market share information—may illuminate

the analysis. In light of the aforementioned defects with market definition, one

must proceed carefully. As it happens, certain market share information is

sometimes relevant if the shares are in “narrow” markets that align with par-

ticular economic models, regardless of whether such markets would be “rele-

vant” under existing protocols.

The possible use of concentration or market share information is considered

with respect to the standard types of anticompetitive effects from horizontal

mergers: unilateral effects with homogeneous goods, unilateral effects with

differentiated products, and coordinated effects. Three key lessons emerge.

First, the correct method of analysis—and thus the relevance, if any, of infor-

mation on market shares—differs greatly across these settings. Hence, the

one-size-fits-all structural presumption—and closely related methods in

merger guidelines in the United States, European Union, and elsewhere—is a

nonstarter. Second, even when market shares in a particular model are rele-

vant, other factors (notably, the elasticity of demand) are also relevant and

quite important, so market share information alone—which is all the structural

presumption considers—cannot give even an approximate indication of an-

ticompetitive effects in any setting. Third, the concentration or market share