Dynamic competition and antitrust: quick-look inferences from the analysis of big tech's R&D expenditure ratios

• Pages: 897-931
• Date: 01 June 2025
• Published date: 01 June 2025
• Author: Jorge Padilla,Douglas H. Ginsburg,Koren Wong-Ervin
• Subject Matter: Derecho Público y Administrativo

86 Antitrust Law Journal No. 3 (2025). Copyright 2025 American Bar Association. Reproduced

by permission. All rights reserved. This information or any portion thereof may not be copied

or disseminated in any form or by any means or downloaded or stored in an electronic

database or retrieval system without the express written consent of the American Bar

Association.

* Jorge Padilla is Senior Managing Director at Compass Lexecon and Senior Fellow of the

GW Competition & Innovation Lab, George Washington University. Douglas H. Ginsburg is

a judge on the U.S. Court of Appeals for the District of Columbia Circuit, Professor of Law

at the Antonin Scalia Law School at George Mason University, and former Assistant Attorney

General in charge of the Antitrust Division of the U.S. Department of Justice. Koren Wong-Ervin

is a Partner at Jones Day, a Senior Fellow at the George Washington Institute of Public Policy,

and former Attorney Advisor and Counsel for IP & International Antitrust at the Federal Trade

Commission. We have benefited from the excellent research assistance and comments of Joe

Perkins, Ram Saram, Suresh Kumar, and Macklin Willigan. We also acknowledge the comments

of Antara Dutta, Richard J. Gilbert, David Teece, and three anonymous referees. Jones Day

represents Google in a variety of matters. Compass Lexecon also represents clients who may

have an interest in the subject matter of this article—both clients who may be in favor of and

those who may be against the views expressed herein. This article was not funded or sponsored

in any way. The opinions set forth herein are the personal views or opinions of the authors;

they do not necessarily reflect the views or opinions of the organizations with which they are

associated or those organizations’ clients.

DYNAMIC COMPETITION AND ANTITRUST:

QUICK-LOOK INFERENCES FROM THE ANALYSIS

OF BIG TECH’S R&D EXPENDITURE RATIOS

J P

D H. G

K W-E*

INTRODUCTION ......................................... 898

I. THE DYNAMIC COMPETITION SCHOOL ................ 900

II. ASSESSING THE EVIDENCE AGAINST THE DYNAMIC

COMPETITION SCHOOL’S MAIN PROPOSITION ......... 906

A. M-S E E ................. 906

B. K A E ....................... 909

C. E  I, C,

 M ..................................... 910

D. M E .......................... 910

III. THE INNOVATION EFFORTS OF LEADERS

IN DYNAMICALLY COMPETITIVE MARKETS ........... 911

A. R   MAAMA ......................... 914

B. C A S ........................ 915

C. D ........................................ 915

CONCLUSION ........................................... 918

APPENDIX A: FINANCIAL YEARS OF FIRMS INCLUDED

IN DATA ............................................ 921

APPENDIX B: MAAMAs’ R&D INVESTMENT, COMPANY

ANALYSIS, 2000–2022 ................................ 922

APPENDIX C: FIGURES 2–10 .............................. 926

897

INTRODUCTION

This article examines the extent to which we should adapt the rules,

standards, and tests that apply to industries characterized by dynamic com-

petition, which we define as competition for the market through sequen-

tial winner-take-all races won through significant innovation as opposed

to competition in the market through competition on static price/output

metrics. More specifically, we examine whether, in markets structured by

innovation races, we should rely on static concentration measures such as

market-share thresholds. To answer this question, we consider two main

competing schools of thought.

The first school of thought is what we refer to as the “Dynamic Competition

School,” which takes the position that, in markets structured by innovation

races, the economic tests that rely on the use of static concentration meas-

ures are fundamentally flawed.1 This is because, with dynamic competition,

firms continually try to develop leapfrogging innovations. These firms may

have small or nonexistent market shares until one succeeds, but during that

time, the incumbent is driven to innovate in an attempt to keep from being

displaced.2

For the authors of the Dynamic Competition School, using static concentra-

tion measures to guide antitrust enforcement and merger control may lead to

more fragmented markets and, possibly, though not necessarily, lower prices

in the short term, but it likely will chill innovation and reduce consumer wel-

fare in the long run, since consumers benefit more from innovation than from

low prices.3 In their opinion, competition agencies should not obsess over

existing market structure; rather, they should ensure that the assets needed

to innovate are not monopolized by incumbents, so that the markets remain

contestable and the incumbents are compelled to operate under the threat of

disruptive innovation and entry. This is because the possibility of unfettered

competition for the market is sufficient to ensure that market outcomes are

beneficial to consumers, even if the incumbent maintains a dominant or a

monopoly position over time.

1 See, e.g., David S. Evans & Richard Schmalensee, Some Economic Aspects of Antitrust

Analysis in Dynamically Competitive Industries, 2 I P’ & E. 1, 18 (2002);

see also David J. Teece & Mary Coleman, The Meaning of Monopoly: Antitrust Analysis in

High-Technology Industries, 43 A B., 801, 824 (1998).

2 See Evans & Schmalensee, supra note 1, at10–18.

3 See Frank Easterbrook, Ignorance and Antitrust, in A, I, 

C 119, 122–23 (Thomas M. Jorde & David J. Teece eds., 1992) (“An antitrust

policy that reduced prices by 5 percent today at the expense of reducing by 1 percent the annual

rate at which innovation lowers the cost of production would be a calamity. In the long run a

continuous rate of change, compounded, swamps static losses.”).

898 A L J [Vol. 86

The second school of thought, which we refer to as the “Mainstream

School,” recognizes that the threat of preemptive innovation may, in principle,

have an impact on existing market participants, but it takes the position that

any impact is unlikely to be determinative in the markets in which leading

online platforms operate because those markets are also characterized by fac-

tors such as network effects, economies of scale and scope, and switching

costs that make entry ineffective.4 In short, this second school seems to

argue that a market should not be considered dynamically competitive unless

we observe fluctuations in market share or, in other words, only if we observe

“action–reaction”—i.e., when the market leader tends to lose market share

over time and the market never tips to monopoly.5

The Dynamic Competition School disagrees, explaining that, in markets

structured in terms of innovation races, we see stable or even increasing market

shares and a lot of innovation and other positive market outcomes. That is,

dynamic industries may be characterized by “increasing dominance”—i.e.,

the market leader’s share grows over time and the market eventually tips to

monopoly—rather than action–reaction. In fact, in markets in which new entry

is won through significant innovation, incumbents are likely to invest signifi-

cant amounts in research and development (R&D) to protect their rents. This,

according to the Dynamic Competition School, is important and explains why

potential entrants can have a significant influence on market behavior such

that even high shares that are stable over the course of several years do not

alone indicate a lack of competition.

This article focuses on the key question of how to determine whether the

threat of preemptive innovation is conditioning incumbents. We do this by

examining the R&D expenditures of 25 companies in five sectors, including

the so-called big tech companies, software and business-to-business (B2B),

and pharma. We examine expenditures in absolute terms and, more impor-

tantly, normalized by revenues.

4 See, e.g., D. C E P, U D C 37, 41

(2019), assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/

file/785547/unlocking_digital_competition_furman_review_web.pdf; Carl Shapiro, Protecting

Competition in the American Economy: Merger Control, Tech Titans, Labor Markets, 33 J. E.

P. 69 (2019); Jonathan B. Baker, Finding Common Ground Among Antitrust Reformers,

84 A L.J. 705 (2022). The Unlocking Digital Competition report was coauthored by

Professor Jason Furman, Professor Diane Coyle, Dr. Amelia Fletcher, Philip Marsden, and

Derek McAuley. D. C E P, supra, at7.

5 See generally John Vickers, The Evolution of Market Structure When There Is a Sequence of

Innovations, 35 J. I. E. 1 (1986). For example, a duopoly market will be characterized

by action–reaction when the market shares of the duopolists tend to fluctuate around 50% or, in

other words, when they alternate as market leaders.

2025] D C  A 899