Dynamic competition with network externalities: how history matters

• Author: Bruno Jullien,Hanna Halaburda,Yaron Yehezkel
• DOI: http://doi.org/10.1111/1756-2171.12304
• Published date: 01 March 2020
• Date: 01 March 2020

RAND Journal of Economics

Vol.51, No. 1, Spring 2020

pp. 3–31

Dynamic competition with network

externalities: how history matters

Hanna Halaburda∗

Bruno Jullien∗∗

and

Yaron Yehezkel∗∗∗

We consider dynamic competition among platforms in a market with network externalities. A

platform that dominated the market in the previousperiod becomes “focal” in the current period,

in that agents play the equilibrium in whichthey join the focal platform whenever such equilibrium

exists. Yet when faced with higher-quality competition, can a low-quality platform remain focal?

In the finite-horizon case, the unique equilibrium is efficient for “patient” platforms; with an

infinite time horizon, however, there are multiple equilibria whereeither the low- or high-quality

platform dominates. If qualities are stochastic, the platform with a better average quality wins

with a higher probability, even when its realized quality is lower, and this probability increases

as platforms become more patient. Hence, social welfare may decline as platforms become more

forward looking.

1. Introduction

A traditional concern in markets with network externalities is that the “wrong” platform may

dominate due to consumers’ miscoordination. Our main research question is whether markets can

∗New York University; hhalaburda@gmail.com.

∗∗Toulouse School of Economics, CNRS, Toulouse.

∗∗∗Coller School of Management, Tel-Aviv University;yehezkel@tauex.tau.ac.il.

For helpful comments and discussion, wethank Gar y Biglaiser,Tim Bresnahan, Luis Cabral, Carlo Cambini, Annamaria

Conti, Jacques Cr´

emer, Renato Gomes, Andrei Hagiu, Pedro Pereira, Markus Reisinger, and Andre Veiga as well as

participants at the Fourth Annual Searle Center Conference on Internet Search and Innovation, the 11th Workshop on

Media Economics, the Third ICT Conference in Munich, the Multi-Sided Platforms Workshopat NUS, the Conference on

the Economics of ICT in Paris, the ZEW Conference on the Economics of ICT in Mannheim, the 8th Tallerin Industrial

Organization, the 2015 Conference on Economics of Intellectual Property,Software and the Internet in Toulouse, the NYU

Stern Economics of Strategy Workshop, the CRES Strategy Conference in St. Louis, and the ESMT Workshop on the

Economics of Platforms in Berlin. Wegratefully acknowledge financial support from the NET Institute, www.NETinst.org,

the Henry Crown Institute of Business Research in Israel, and ANR under grant ANR-17-EURE-0010. This project has

also received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research

and innovation programme (grant agreement no. 670494).

C2020, The RAND Corporation. 3

4/THE RAND JOURNAL OF ECONOMICS

correct such inefficient outcomes in a dynamic competition between far-sighted platforms, when

consumers’ expectations concerning the dominant platform depend on the history. We confirm

this intuition for long finite horizon, but identify two novel factors that may prevent efficient

outcome when the horizon is infinite.

Platform competition typically involves both network effects and repeated interaction. We

often observe that a platform that was dominant in the recent past has the advantage of customers’

favorable expectations, meaning that customers expect that this platform will also attract other

customers in the current period. Weshall refer to such a platfor m as a focal platform. Forexample,

Apple’s success with the iPhone 4 resulted in pre-orders for its iPhone 5, exceeding two million

within a day of its launch in September 2012—even though there were not yet any applications

that could take advantage of the phone’s new features. Moreover, analysts predicted that 50 million

users would buy the new smartphone within three months of its launch.1A similar dynamics was

in evidence for the iPhone 6’s release, as sales topped four million in the first 24 hours.2Analysts

reported that this trend continued with the iPhone 7’srelease.3In contrast, neither Blackber ry nor

Windowsphones enjoyed a comparable advantage during this period. Even though the Blackberry

phones—the Q10 and the Z10—received glowing reviews, the absence of positive expectations

made it difficult for the firm to gain substantial market share: application developerswere skeptical

about the phone’s ability to attract users; sales were indeed sluggish, which was due in no small

part to the paucity of available apps.4We can ascribe the developer’s skepticism and the resulting

lack of apps to the Blackberry platform’s recent history.

Yet even dominant platforms can lose market share, despite winning in the past, when faced

with a higher-quality competitor. In the market for smartphones, for instance, Nokia dominated

the early stage (along with RIM) with smartphones based on a physical keyboard. Apple then

revolutionized the industry by betting on its new operating system, which featured touch-screen

technology. A few years later, Samsung managed to gain substantial market share (though not

market dominance) by betting on smartphones with large screens. The supplanting of industry

leaders was likewise a common theme in the market for video game consoles. Nintendo, Sony,

and Microsoft alternated as the market leader (Hagiu and Halaburda, 2009). Thus, platforms are

sometimes able to overcome the market’s unfavorable expectations. These examples raise the

question of when is it profitable for a platform facing unfavorable position to invest in capturing

the market, and when it is profitable for a platform facing a favorableposition to invest in retaining

the market, if each firm knows that its current strategy may affect its future position.

If the impact on the future is ignored, then it may not be profitablefor the nonfocal platfor m

to overcome unfavorable expectations, even if it can offer higher quality than the focal platform.

The reason is that network effects provide the focal platform with a short-term competitive

advantage: in a one-time interaction, a focal platform can use its position to dominate the market,

even when competing against a higher-quality platform. Yet we can expect that, in the long run, if

platforms are forward looking, then a high-quality but presently nonfocal platform can overcome

its expectations disadvantage because it can afford short-term losses in order to become focal in

the future. At the same time, a low-quality forward looking focal platform also has an incentive

to invest in maintaining its dominant market position. So, even though the nonfocal platform has

a quality advantage, the focal platform has an expectations advantage.

In this article, we ask whether a low-quality platform currently benefiting from its focal

position can continue to dominate the market in a dynamic setting—that is, when platforms

account for the future benefits of capturing the market today.For example, would Apple continue

to dominate the market for tablets if competitors (Samsung’s Tab, Microsoft’s Surface, etc.)

1Ryan Faughnder and Adam Satariano, “AppleiPhone 5 Pre-Orders Top 2 Million, Doubling Record,” Bloomberg

(2012). Available at: (www.bloomberg.com/news/2012-09-17/apple-iphone-5-pre-orders-top-2-million-double-prior-

record-1-.html).

2www.cnet.com/news/apple-iphone-6-iphone-6-plus-preorders-top-4m-in-first-24-hours.

3www.ft.com/content/dc40349c-8d2b-345c-b9c3-20dfe41a9992.

4On the quality and launch of the Blackberry phones, see Austen (2012) and Bunton (2013).

C

The RAND Corporation 2020.