Cross‐licensing and competition

• DOI: http://doi.org/10.1111/1756-2171.12248
• Author: Yassine Lefouili,Doh‐Shin Jeon
• Date: 01 September 2018
• Published date: 01 September 2018

RAND Journal of Economics

Vol.49, No. 3, Fall 2018

pp. 656–671

Cross-licensing and competition

Doh-Shin Jeon∗

and

Yassine Lefouili∗

We analyze the competitive effects of bilateral cross-licensing agreements in a setting with many

competing firms. We show that firms can sustain the monopoly outcome if they can sign un-

constrained bilateral cross-licensing contracts. This result is robust to increasing the number of

firms who can enter into a cross-licensing agreement.We also investigate the scenario in whicha

cross-licensing contractcannot involve the payment of a royalty by a licensee who decides ex post

not to use the licensed technology. Finally, policy implications regarding the antitrust treatment

of cross-licensing agreements are derived.

1. Introduction

A cross-license is an agreement between two firms that allows each to use the other’s patents

(Shapiro, 2001; R´

egibeau and Rockett, 2011). Cross-licensing has long been a common practice.

For instance, Taylor and Silberston (1973) report that cross-licensing accounts for a significant

share of all licensing arrangements in many industries: 50% in the telecommunications and

broadcasting industry, 25% in the electronic components industry, 23% in the pharmaceutical

industry, etc.1Cross-licensing is therefore likely to have an impact on competition in a large

number of sectors.

Cross-licensing agreements involve both technological and monetary transfers. Technolog-

ical transfers are generally perceived as procompetitive: they can result in goods being produced

at lower costs by potentially more firms. These transfers are particularly useful in Information

Technology (IT) industries, such as the semiconductor and mobile phone industries, where the

∗ToulouseSchool of Economics, University of Toulouse Capitole; dohshin.jeon@gmail.com, yassine.lefouili@tse-fr.eu.

We thank the Co-Editor, Ben Hermalin, and two anonymous referees for very helpful comments and suggestions.

We are also grateful to Reiko Aoki, Claude d’Aspremont, Paul Belleflamme, David Besanko, Jan Boone, Bernard

Caillaud, Zhijun Chen, Jacques Cr´

emer, Andrew Daughety, Vincenzo Denicolo, David Encaoua, Pierre Fleckinger,

Alberto Galasso, Gautam Gowrisankaran, Bertrand Gobillard,Renato Gomes, Tadashi Hashimoto, Roman Inderst, Bruno

Jullien, Byung-Cheol Kim, Laurent Linnemer,Patrick Rey, Jacob Seifert, Sa¨

ıd Souam, Thibaud Verg´

e, TakuroYamashita,

and participants in various conferences and seminars for valuable comments and discussions. The financial support of

the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme

(grant agreement no. 670494), and the NET Institute, http://www.NETinst.org,is g ratefullyacknowledged.

1In particular, cross-licensing in the semiconductor industry has received much attention in the literature (Grindley

and Teece,1997; Hall and Ziedonis, 2001; Galasso, 2012).

656 C2018, The RAND Corporation.

JEON AND LEFOUILI / 657

intellectual property rights necessary to market a product are typically held by a large number of

parties, a situation known as a patent thicket (Shapiro, 2001, 2007; Department of Justice [DOJ]

and Federal TradeCommission [FTC], 2007; Galasso and Schanker man, 2007).2Monetary trans-

fers, however, can be anticompetitive. More specifically, high per-unit royalties can allow firms

to sustain high prices.

The following natural question arises: do cross-licensing partners have incentives to agree

on high royalties? The existing literature providesan answer to this question in a duopoly setting:

in that case, two firms can sign a cross-licensing agreement that specifies royalties high enough

to replicate the monopoly profit (Katz and Shapiro, 1985; Fershtman and Kamien, 1992). This

monopolization result can be generalized in a straightforward way to a setting with more than

two firms signing a multilateral agreement involving all of them (see Section 2).

However, in practice, we often see bilateral cross-licensing in industries with more than two

firms. In this setting, would any pair of firms agree on high royalties that might weaken their

competitive positions vis-`

a-vis their rivals? We build a model to investigate whether bilateral

cross-licensing agreements can still allow firms to sustain the monopoly outcome in this scenario.

We consider N(≥3) symmetric Cournot oligopolists selling a homogeneous good and

owning one patent each. Firms can get access to the technologies coveredby their rivals’ patents

through cross-licensing agreements, before competing in a product market. We suppose that,

given a set of patented technologies a firm has access to, its marginal cost is constant, and that

the larger the set, the smaller the marginal cost.

We assume that cross-licensing contracts are private, that is, their terms are observable only

to the parties signing them, and focus on bilaterally efficient agreements. A set of cross-licensing

agreements is said to be bilaterally efficient if each agreement maximizes the joint profit of the

pair of firms who sign it, given all other agreements.3Note that a firm’s overall profit is composed

of the profit it makes from selling its product and the revenues generated by the licensing of its

technology.

In Section 2, we analyze our baseline model, in which firms can sign unconstrained bilateral

cross-licensing agreements. We focus on symmetric equilibria, where any two distinct firms sign

a cross-licensing contract and every firm pays the same royalty to any other firm. Two firms in a

given coalition can indirectly affecttheir joint output through the royalties they charge each other.

When deciding these royalties, they take into account twoopposite effects: the coordination effect

and the royalty-saving effect. The former captures the fact that the twofir ms have joint incentives

to restrict their joint output below its noncooperative equilibrium level,and the latter refers to the

idea that the coalition’s marginal cost is lower than each of its member’s marginal cost, because

the royalties the two firms charge each other are internal transfers within the coalition. The

royalty-saving effect provides incentives to reduce the royalties charged byeach fir m to the other

one, whereas the coordination effect provides incentives to increase these royalties. We show

that these two effects cancel out when the (symmetric) per-unit royalty is equal to the one that

maximizes the industry profit. This implies that the monopoly outcome can be sustained through

bilaterally efficient cross-licensing agreements.

In Section 3, we extend our analysis in two directions. First, we consider an environment

in which cross-licensing agreements can be signed by coalitions of any size. We show that our

monopolization result still holds in such a setting. Second, weprovide an extension of the baseline

model, which incorporates ex post usage constraints. In the baseline model, firm ipays to firm

jthe royalty specified in their cross-licensing agreement regardless of whether the former uses

the latter’s patented technology. This can lead to royalties that are higher than the cost reduction

derived from the use of a given licensed technology. However, competition authorities usually

2According to FTC (2011), “[t]he IT patent landscape involves products containing a multitude of components,

each covered by numerous patents. [. .. ] This contrasts with the relationship between products and patents in the

pharmaceutical and biotech industries where innovation is generally directed at producing a discrete product coveredby

a small number of patents.”

3A more precise statement is provided in Definition 1.

C

The RAND Corporation 2018.