Content uploaded by Stephen Lloyd Jones
Author content
All content in this area was uploaded by Stephen Lloyd Jones on Oct 16, 2020
Content may be subject to copyright.
The evolution of cooperation in the face of
conflict: Evidence from the innovation
ecosystem for mobile telecom standards
development*
Stephen L. Jones Aija Leiponen
University of Washington Bothell Cornell University
Gurneeta Vasudeva
University of Minnesota
* Forthcoming in Strategic Management Journal (DOI: 10.1002/smj.3244). All authors contrib-
uted equally; authors listed in alphabetical order.
Abstract
Research summary. How does interfirm cooperation in innovation ecosystems evolve in the face
of conflict? We theorize that conflict propels firms to reconfigure cooperative relationships while
maintaining and even increasing cooperation with aggressors because cooperation is the primary
mechanism for value creation in such ecosystems. To empirically test our arguments, we study
patent litigation and subsequent cooperation between mobile telecommunications firms within
the 3GPP standards development organization. We find evidence of a dual cooperative strategy
in the face of conflict: while cooperation increases between litigants, defendants also enhance
cooperation with others to steer standards away from aggressors. We also highlight the contin-
gent roles of technological complementarities and relational position underpinning cooperation
after conflict. Our findings demonstrate that in innovation ecosystems, cooperation with adver-
saries persists despite conflict.
Managerial summary. Firms in innovation‐driven industries cooperate to develop interoperabil-
ity standards and compatible technologies. Yet, cooperative firms may disagree about what con-
stitutes fair, reasonable, and non‐discriminatory terms for licensing intellectual property. Thus,
conflict and patent litigation arise even as firms cooperate to build technologies and industry
standards. We find that in innovation ecosystems, firms commonly increase cooperative efforts
in response to conflict. Less‐connected firms or those with valuable complementary technologies
will likely expand cooperation than well‐connected firms or technological competitors. Well‐con-
nected firms may pursue alternative cooperative opportunities. We suggest that defendant firms'
managers can adopt a dual cooperative strategy: (a) identify private and shared benefits from
the joint development of complementary technologies with aggressors and (b) invest in alterna-
tive technological partnerships to influence the direction of future standards development.
Keywords: innovation ecosystems, interorganizational cooperation, standards development,
patent litigation, mobile telecommunications
Jones, Leiponen, and Vasudeva 2
1. Introduction
On August 13, 2020, Epic Games sued Apple
and Google for anticompetitive behavior re-
lated to their mobile platforms. Epic Games
had recently encouraged mobile players of its
hugely popular Fortnite game to buy their in-
game virtual goods directly from Epic’s web-
site, by-passing the mobile platforms and
their payment systems that automatically re-
sult in a 30 percent commission for Apple and
Google. Apple and Google, unsurprisingly,
threw Fortnite out of their application stores
as such off-platform sales are against their
platform rules. The next day, Epic Games filed
a thoroughly prepared 60-page document ar-
guing the case for antitrust enforcement. How
will the relationship of these long-time collab-
orators evolve going forward? Fortnite could
be permanently blocked out of Apple’s mobile
devices. Alternatively, by lowering its plat-
form commission for Epic Games, Apple and
Google could potentially need to renegotiate
fees with every major game developer, leading
to a huge loss of revenue. Thus, for all parties,
billions of dollars are at stake if they find no
cooperative solution. We argue that there is a
good chance that the parties will find a coop-
erative arrangement that allows Fortnite back
into the application market and even expands
the innovation opportunities for Epic Games
on mobile platforms. We empirically test our
arguments with data from a major standards-
development organization.
As the foregoing example illustrates, in-
novating firms are often bound together in
meta-organizations (Gulati, Puranam, &
Tushman, 2012) involving cooperative activi-
ties for technological products (Adner &
Kapoor, 2010; Jacobides, Knudsen, & Augier,
2006) and standards development (Farrell &
Saloner, 1988; Rysman & Simcoe, 2008; Rosen-
kopf, Metiu and George, 2001). Although coop-
eration within such innovation ecosystems
enables mutual gains through joint value cre-
ation, firms also face intense rivalry and ten-
sion over value appropriation (Casadesus-
Masanell & Yoffie, 2007; Ansari, Garud, & Ku-
maraswamy, 2016; Ranganathan, Ghosh, &
Rosenkopf, 2018). At times, firms’ ecosystem
interactions devolve into outright conflict,
which can result in litigation (Somaya, 2003;
Simcoe, Graham & Feldman, 2009). Yet, inno-
vation ecosystems thrive on cooperation. In
fact, many rapidly evolving ecosystems dis-
play both ample conflict and surprising
amounts of cooperation. However, little is
known about the impact of conflict on subse-
quent cooperation within innovation ecosys-
tems.
In this study, we explore when interde-
pendent firms in innovation ecosystems sus-
tain cooperation in the face of conflict with ri-
vals and complementors. In particular, we ex-
amine the cooperative development of tech-
nology when firms also engage in legal con-
flict. Our model of cooperation is informed by
the prisoner’s dilemma in evolutionary game
theory (Axelrod & Hamilton, 1981; Axelrod,
1984) and distinguishes among public, pri-
vate, and club benefits from cooperation (Bu-
chanan, 1965; Ostrom, 2000). In repeated pris-
oner’s dilemma games, it is in the long-term
interest of adversaries that are likely to inter-
act in the future to cooperate (Rapoport &
Chammah, 1965), but cooperation can break
down if the short-term benefits of defection
exceed the discounted long-term benefits of
cooperation (Axelrod, 1984). Based on these
principles, we theorize about the impact of pa-
tent litigation—which reflects disagreements
between firms about the value of the underly-
ing technology—on subsequent cooperation,
Evolution of cooperation in the face of conflict 3
and we examine contingencies for such coop-
eration to prevail. We argue that technological
and relational interdependence within an in-
novation ecosystem constrains strategic ac-
tions, making it costly to sever cooperation. As
a result, the greater the expectation of future
private and club benefits from cooperation,
the more firms will sustain cooperation even
after conflict.
Empirically, we document how patent lit-
igation propels firms to reexamine their tech-
nical cooperation and how this type of litiga-
tion also affects cooperation by firms not di-
rectly involved in the lawsuit. We find that the
defendants and plaintiffs of patent lawsuits
typically increase their cooperation in specific
technical areas following litigation filings re-
lated to those technical areas. Furthermore,
when firms are more technologically distant,
they can create mutually beneficial club ben-
efits and increase their cooperation after con-
flict. In contrast, defendant firms that have
abundant outside options for developing
other technologies owing to their central posi-
tion in the relational network do not increase
their cooperation. Defendants also expand co-
operation among themselves and with other
parties unrelated to the conflict, suggesting an
attempt to steer the standards development
away from the technologies controlled by the
plaintiff to attenuate interdependence and the
risk of future conflict.
The innovation ecosystem we study is the
standards development organization (SDO)
called 3GPP, the Third Generation Partnership
Project. It is recognized for developing global
mobile telecom standards. Firms participate
in SDOs like 3GPP to develop shared interop-
erable platform technologies and to increase
the adoption of their own technological inno-
vations (cf. Farrell & Saloner, 1988; Rysman &
Simcoe, 2008; Bar & Leiponen, 2014;
Vasudeva, Alexander & Jones, 2015). Coopera-
tion between firms occurs because private
benefits, such as capturing value by licensing
standard essential patents, depend on the col-
lective decisions concerning which technical
features are included in the standards (Farrell
& Simcoe, 2012). Some of the collaborative ac-
tivities create club benefits in the form of
complementary features that allow contrib-
uting firms to exclude others while enhancing
their own market positions.
Many of the participating firms possess
standard-essential patents (SEPs) upon which
technical contributions are based. Legal con-
flicts arise when firms that own SEPs believe
that other firms have failed to obtain the
needed licenses or compensate them appro-
priately, thus infringing on their intellectual
property (IP) (Layne-Farrar, Padilla, &
Schmalensee, 2007; Simcoe et al., 2009). Alter-
natively, licensee firms may file lawsuits if
they believe that the licensing arrangements
are too onerous, violating the 3GPP norm that
licensing terms be fair, reasonable, and non-
discriminatory (FRAND). Such disagreements
reveal the evolving “strategic stakes” for the
concerned parties: the technologies embodied
in SEPs are valuable to both the patent holders
and those who use them for their own prod-
ucts and innovation (Somaya, 2003).
Conflict manifested in litigation can make
continued technical cooperation between
firms difficult (Lanjouw & Schankerman,
2001; Shane & Somaya, 2007; Sytch & Tata-
rynowicz, 2014). Yet, as our findings show, co-
operation rather than defection can emerge as
the optimal strategy between adversaries en-
gaged in repeated interactions (Schelling,
1960; Axelrod & Hamilton, 1981; Axelrod, 1984;
Hirshleifer, 1991) in innovation ecosystems.
Jones, Leiponen, and Vasudeva 4
Our study makes three main contribu-
tions to the literature on the evolution of co-
operation in innovation ecosystems. First, our
study identifies how interfirm cooperation
changes due to conflict between ecosystem
members. We propose that cooperation be-
tween parties is driven by the expectation of
future private benefits and, in particular, club
benefits that arise from complementarities
between the contributors (Jacobides, Cen-
namo, & Gawer, 2018; Rothaermel & Boeker,
2008; Teece, 2018). Thus, distinct from static
accounts of cooperation in strategic alliances
(e.g., Parkhe, 1993; Khanna, Gulati, & Nohria,
1998; Arslan, 2018), our framework takes an
evolutionary approach to model technological
conflict and cooperation over time. In our
study, litigation captures conflict over existing
SEPs and related technologies, whereas
cooperation in standards development
influences how the technology will evolve in
the future.
Second, we present conditions under
which the affected firms are likely to increase
or decrease cooperation with the aggressor.
We highlight two contingencies—
technological distance and relational
position—that can alter the expectation of
future payoffs under repeated interactions
(cf. Ranganathan et al., 2018). In particular,
within our tight-knit innovation ecosystem,
technological distance creates conditions for
technological complementarities and lower
competitive pressure, enhancing potential
club benefits from cooperation. A relationally
central position, in contrast, lowers the cost of
defection because it enhances the defendant’s
alternatives for cooperative development. Our
theoretical framework sharpens our view of
the mechanisms driving the impact of conflict
on subsequent cooperation.
Third, while we are interested in
observing cooperation between adversaries,
the larger strategy question we address is how
conflict between two parties triggers broader
shifts in cooperative activity within the
innovation ecosystem (Paik & Zhu, 2016). We
find that defendants adopt a dual cooperative
strategy which entails cooperation with the
aggressor to build complementarities, while
also forming new ties with other participants
to potentially maneuver the technological de-
velopment away from the aggressor.
Finally, our study is empirically novel be-
cause we are able to isolate instances of con-
flict using SEP litigation and tightly link these
events to cooperative activities developing re-
lated technologies within 3GPP. We identify
SEP lawsuits directly related to the technical
specifications under development in a specific
committee and observe the pattern of cooper-
ation before and after the initiation of the law-
suit. These patterns include not only the
plaintiff and the defendant but also those in-
directly affected through the cooperative ef-
forts of the defendant. Such a methodological
approach that investigates how the litigation
action of one firm affects the cooperative out-
comes of the broader co-creation network is
important for understanding the dynamic in-
terplay of conflict and cooperation in innova-
tion ecosystems.
2. Theory and hypotheses
2.1 Cooperation between litigants
Cooperative activities generate three types of
benefits for firms participating in innovation
ecosystems: public benefits, club benefits, and
private benefits (Parkhe, 1993; Khanna et al.,
1998; Vasudeva & Teegen, 2011; Arslan, 2018;
Vasudeva, Leiponen, & Jones, 2020). Public
benefits accrue to everyone in the industry,
Evolution of cooperation in the face of conflict 5
independent of their contributions to the
shared platform. For example, any firm in the
industry could adopt and apply the 3GPP tech-
nical specifications for mobile telecommuni-
cations. Club benefits (Buchanan, 1965; Olson,
1965) are available to a subgroup of firms that
pool their technologies to create mutually
complementary products and services. The
profits from these innovations are shared
with partners that contribute to commercial-
izing the technologies. Private benefits are
claimed by an individual technology holder.
They might include royalty income from li-
censing SEPs and first-mover advantages in
subsequent product markets. Private benefits
and club benefits are obtained only when
firms contribute to technology development
because these contributions enable firms to
shape the technological system in a direction
that favors their own technology and market
positions.
If firms cease to cooperate after conflict,
they will only obtain public benefits subse-
quently, which could be an appealing option
so long as the private and club benefits from
cooperation are relatively small, and the costs
of cooperation are high. Thus, the structure of
the benefits and costs of cooperation is akin to
a prisoner’s dilemma scenario in which it is
optimal for the participants to cooperate in re-
peated games. However, participants in such
situations individually choose their actions
based on expected profits and may or may not
choose to cooperate (Axelrod, 1984). We illus-
trate these payoffs from cooperation when pa-
tent litigation conflict occurs with a prisoner’s
dilemma framework in an online appendix.
1
A prisoner’s dilemma is symmetric if both parties re-
ceive the same payoffs for the same potential out-
comes and is asymmetric if the payoffs differ.
Patent litigation implies that the technol-
ogy remains a critical strategic resource for
the patent holders and defendants (Somaya,
2003; Ziedonis, 2004). As Lerner and Tirole
(2015) note, ex ante—before adoption in a
standard—technology owners may be willing
to accept lower royalty arrangements to en-
sure that their technologies are incorporated,
but ex post they have an incentive to exploit
their expanded market power and increase
their royalty rates. Yet, defendants would not
infringe on the disputed technology and incur
the costs of litigation if the expected returns
from developing alternatives were higher
than the expected returns from navigating
current technological trajectory. Conse-
quently, ceasing to license from the patent
holder becomes an unappealing option, espe-
cially when the patented technology is essen-
tial for the standard. In such instances, we ar-
gue that a licensee will reconfigure its cooper-
ative strategies to manage its technological
dependence on the patent holder.
If the defendant must pay a higher licens-
ing fee to implement the standards, then the
higher fee will be required whether or not the
defendant cooperates with the patent holder
in further standards development. Since the
defendant must pay higher royalties anyway,
the firm might seek to appropriate more ben-
efit from the disputed technology. We suggest
that once a defendant is committed to the set
of technologies covered by essential patents, it
becomes lucrative for them to develop com-
plementary technological features that en-
hance club benefits for both parties, as shown
in the asymmetric prisoner’s dilemma model
1
in the online appendix.
Jones, Leiponen, and Vasudeva 6
Apart from the club benefits, a coopera-
tive approach may also confer private benefits
on the defendant by allowing the firm to con-
vert a one-way dependence on the plaintiff’s
technology to a two-way interdependence
(Jacobides et al., 2018), such that the plaintiff
also needs the defendant’s technology in fu-
ture generations of the product. When techno-
logical features are complementary (cf.
Milgrom & Roberts, 1990; Topkis, 1998), one
feature’s increasing value enhances the re-
turns to improving the other features (Gandal,
Kende, & Rob, 2000). Such co-specialization
fosters mutual forbearance and reduces the
possibility of future hold-up since both parties
become mutually dependent on each other’s
technological resources (Somaya, 2003; Ziedo-
nis, 2004; Teece, 2018).
Cooperation also generates private and
club benefits for the patent holder. As noted
above, club benefits accrue from two-way
complementarities. Private benefits accrue
because cooperation prevents the defendant
from entering into a “patent race” (Fuden-
berg, Gilbert, Stiglitz, & Tirole, 1983), thereby
safeguarding the stream of licensing reve-
nues. Building on the logic for managing rela-
tional conflict in repeated exchange settings
(e.g., Li & Matouschek, 2013), the patent
holder may even entice the defendant with
particularly appealing cooperative projects to
continue to cooperate and further enhance
the value of the technological platform. Thus,
the patent holder may alleviate conflict by of-
fering compelling cooperative initiatives.
To summarize, cooperation may become
the rational optimal response by firms that
experience conflict over IP. They may see
more benefits than costs in jointly developing
further enhancements of the contested tech-
nology. If the underlying technologies hold
strategic importance in the innovation ecosys-
tem due to widespread adoption, technology
users may look for opportunities to build com-
plementary gains. Moreover, the patent
holder will attempt to ensure continued use
and licensing of the technology through coop-
eration. We thus hypothesize that SEP litiga-
tion sustains or increases subsequent cooper-
ation between the litigating parties.
Hypothesis 1. After a patent litigation
event, a plaintiff and defendant will, ce-
teris paribus, sustain or increase cooper-
ation related to the patented technology
in an SDO.
Consistent with the evolutionary perspec-
tive on cooperation (Axelrod, 1984), Hypothe-
sis 1 implies that cooperation is likely to
emerge when conflicting parties are bound to-
gether within a common technological space,
such that the expectation of continuing inter-
actions amplifies the importance of the future
relative to the present. Two corollaries
emerge. First, parties should cooperate more
readily when they expect to create value in the
future, not only through private benefits from
potential technology licensing, but also
through joint creation of complementarities
resulting in club benefits. Second, there
should be few outside options for the defend-
ant, which increases the frequency of its inter-
actions with the aggressor and fosters mutual
dependence. Accordingly, in the hypotheses
that follow, we develop arguments for two
contingencies for cooperation to emerge: the
technological distance between the plaintiff
and defendant in the innovation ecosystem,
and the defendant’s relational position in the
ecosystem network.
Evolution of cooperation in the face of conflict 7
2.2 The contingent role of technological
distance between litigants
The preceding hypothesis assumes that con-
flicting firms can build technological comple-
mentarities. However, the firms’ characteris-
tics may make them more or less suited to do
so, which will likely influence the cooperative
response. If two firms are positioned for com-
plementary technology development, then in-
vestments by one firm will enhance the re-
turns for the other firm if it develops related
technologies.
Building on Dussauge, Garrette, and
Mitchell (2000), for firms to develop
complementary resources and capabilities
their resource endowments must differ and
they must make distinct but related
contributions to the standard. Accordingly,
we expect that firms that are technologically
more distant—that is, distinct in terms of
their technology portfolios—are likely to be
stronger complementors because they can
combine their technologies to create mutually
beneficial innovative features (Bar &
Leiponen, 2014; Fleming, 2001). Such
technological co-dependence reduces the
threat of holdup (Oxley, 1997; Pisano, 1989).
In contrast, the potential for joint value
creation is lower for firm dyads with a low
technological distance. Firms that are very
similar in terms of their technological
portfolios perceive few opportunities for novel
combinations. Although a minimal level of
technological overlap that provides
“absorptive capacity” is helpful for
cooperation (cf. Mowery, Oxley, & Silverman,
1996; Diestre & Rajagopalan, 2012), a very high
overlap (low technological distance) can
prevent novel recombinations. In our context
of study, since firms select into highly focused
technology development groups, we do not
expect absorptive capacity to be a major
hurdle in the process of technology
recombination (Fleming & Sorenson, 2001;
Vasudeva & Anand, 2011).
Technologically distant firms are also
more likely to cooperate after conflict because
they are less likely to engage in direct
competition. Because technological capabili-
ties evolve in a gradual and path-dependent
manner (Helfat, 1994; Stuart & Podolny, 1996),
it is difficult for a technologically distant firm
to make a sudden leap into another firm’s
technological territory, thereby making
knowledge-sharing less threatening. In con-
trast, firms within the same industry segment
with overlapping technological or commercial
domains are less likely to cooperate, especially
when the potential for the expropriation of IP
is high (Dushnitsky & Shaver, 2009). The
resulting cooperation between
technologically distant firms is akin to what
Dussauge et al. (2000: 104) call “link alliances”
that combine complementary skills while
allowing for private benefits in the form of
licensing. Hence, if conflict signals the
strategic value of the contested technology,
firms that are technologically distant yet stra-
tegically interdependent should respond more
cooperatively.
Hypothesis 2. After a litigation event,
subsequent cooperation in an SDO
between a plaintiff and a defendant will
be amplified by their technological
distance.
2.3 The contingent role of the
defendant’s relational position
The evolution of cooperation is determined by
the calculation of expected payoffs in the form
private and club benefits. These payoffs from
cooperation can change depending on the sur-
rounding market structure that determines
Jones, Leiponen, and Vasudeva 8
how likely it is that the adversaries will inter-
act with one another in the future (Axelrod,
1984). In particular, a defendant’s relative co-
operation payoffs with the plaintiff may de-
pend on its access to other firms in the co-cre-
ation network. Conflict may reduce a defend-
ant’s continued cooperation with the plaintff
when payoffs from developing new features
with other firms exceeds those from coopera-
tion with the plaintiff. Defendants with exist-
ing cooperative relationships with other firms
in the ecosystem are more likely to have such
opportunities. Consequently, the defendant’s
reliance on the plaintiff for mutual gains be-
comes proportionally smaller. Furthermore, a
more central relational position connotes
power and repute within the co-creation net-
work (Gulati, 1995), to build a coalition or
drive consensus in a direction that counter-
balances the power of the plaintiff in stand-
ards development (Ranganathan & Rosen-
kopf, 2014; Ranganathan et al., 2018; Simcoe,
2012).
A more central relational position thus
implies that the defendant has alternative co-
development relationships as outside options
that can overshadow opportunities for value
creation from cooperation with the plaintiff.
While there are always costs from ending and
reconfiguring cooperative arrangements, the
costs will be lower when there are already
many alternatives to the focal relationship.
Thus, the greater the number of outside op-
tions, the weaker will be the defendant’s in-
centives to respond cooperatively to the
changed technological opportunities vis-à-vis
the plaintiff. Therefore, the cooperative re-
sponse to conflict will become weaker for a de-
fendant with a number of other cooperative
ties within the co-creation network.
Hypothesis 3. After a litigation event, sub-
sequent cooperation between a plaintiff
and a defendant in an SDO will be dimin-
ished by a defendant’s more central rela-
tional position.
2.4 The defendant’s cooperation
strategy with other participants
The preceding arguments concerning the ad-
vantages of a strong relational position imply
that litigation by one party in an innovation
ecosystem may propel defendants to expand
their cooperation with other parties as a
buffer. In our context, a legal challenge by a
plaintiff might induce cooperation between
co-defendants and between defendants and
other parties in the standards development
arena. Thus, conflict may alter not only the co-
operative dynamics between adversaries, but
also influence their actions within the broader
ecosystem.
The goal of patent litigation is to appro-
priate value from a specific technological as-
set. As the patent holder attempts to appropri-
ate a larger share of the returns to innovation
through licensing fees, the defendant may be
encouraged to safeguard itself by circumvent-
ing the patented technology in question. Ac-
cordingly, the defendant may expand its ef-
forts and pursue other cooperative relation-
ships, as illustrated by the symmetric pris-
oner’s dilemma model in the online appendix.
Its cooperative efforts with others broaden its
opportunities for developing complementary
innovations, thus reducing the possibility of
getting “fenced in” by the plaintiff (Ziedonis,
2004). The less the defendant depends on the
plaintiff’s technologies for innovation oppor-
tunities, the better its bargaining position
with respect to the plaintiff. By building coali-
tions (Davis, 2016), the defendant could also
reduce the possibility of delays and hold-up
Evolution of cooperation in the face of conflict 9
that can determine success in standards de-
velopment and in downstream product mar-
kets (Farrell & Simcoe, 2012; Ranganathan et
al., 2018; Simcoe, 2012). Thus, cooperation
among co-defendants or among defendants
and other parties may increase following liti-
gation in an effort to move technological
standards away from the contested territory.
Consequently, a defendant may adopt a
dual strategy: while cooperating more with
the plaintiff to appropriate returns from com-
plements to the disputed technology, a de-
fendant may also build stronger cooperative
relationships with others to strengthen its re-
lational position or to further differentiate to
counteract future contests with the plaintiff.
Hypothesis 4a-b. Patent litigation will in-
crease subsequent cooperation in an SDO
(a) between co-defendants and (b) be-
tween defendants and other parties.
3. Data and method
3.1 Empirical context
3GPP technical contributions. We examine col-
laboration among firms in 3GPP, the key SDO
in the global mobile telecommunications sec-
tor. From 1999 through 2008, 3GPP developed
the 3G standard Universal Mobile Telecom-
munications System (UMTS). It then devel-
oped the 4G standard Long-Term Evolution
(LTE), with its first release in 2009 (Baron &
Gupta, 2018). 3GPP began planning its 5G
standard in 2016 (3GPP, 2018). Hundreds of
telecom firms such as carriers, handset man-
ufacturers, and chipset manufacturers are
3GPP members. Some actively contribute to
standards development, whereas others ob-
serve the standards process to keep abreast of
new advancements. We focus only on active
contributors.
Contributing to standards development
involves a substantial cost. Participating
firms’ standards development efforts are
closely connected with their product and
innovation strategies, so they bear the risk of
revealing their IP strategies (Rosenkopf et al.,
2001). They also pay SDO membership fees,
allocate engineering time and talent to
develop technical solutions, and support the
effort of engineers and managers to influence
the trajectory of standards development.
Working groups (WGs) in which the ma-
jority of standards development takes place
roll up into three technical specifications
groups (TSGs): Radio-Access Network (RAN)
maintains specifications for how mobile
phones connect to the wireless network; Core
Network and Terminals (CT) maintains spec-
ifications for the core network and its inter-
faces, protocols for mobile phones, and speci-
fications for SIM cards; and Service and Sys-
tem Aspects (SA) maintains specifications for
the overall system architecture. Table 1 de-
scribes the technical domain, contributors,
and litigation activity of each WG.
New specifications begin when WG
members propose new features as work items.
To address new requirements, members sub-
mit technical contributions to the WGs for
consideration. These contributions are ap-
proved through consensus or voting at WG
meetings (Baron & Gupta, 2018). Contribu-
tions are submitted by individual firms or
groups of firms. Hundreds of contributions
are required to create one technical specifica-
tion, and each WG maintains 30 to 100 tech-
nical specifications. In our analysis, we focus
on these technical contributions because they
are the finest-grained unit that captures tech-
nical cooperation between firms. We focus on
contributions jointly authored by two or more
firms.
Jones, Leiponen, and Vasudeva 10
Standard-essential patent licensing and
litigation. In technical contexts, litigation usu-
ally involves a plaintiff enforcing their IP
rights that the defendant has arguably in-
fringed. A lawsuit often follows a “cease and
desist” letter informing the potential infringer
that unless certain behavior (e.g., marketing a
product that contains the allegedly infringing
technology) is discontinued, it will be sued.
This legal challenge initiates a negotiation to
discover the basis of the challenge and deter-
mines whether the parties can agree to a set-
tlement. A settlement implies that the parties
2
Case reference 1:10-cv-00457, Northern Illinois Dis-
trict Court
agree on the value and infringement of the un-
derlying IP. However, if the parties disagree,
they will go to court. Such situations can arise
because a firm has not licensed from a patent
holder or because an existing license has ex-
pired and the firms could not reach a new li-
censing agreement. For example, Motorola
filed a suit against Research in Motion (RIM)
in 2010, claiming that RIM infringed on
Motorola’s patents concerning encrypted data
packet transfer.
2
The parties settled without
going to trial.
Table 1. Summary of 3GPP data
Working
Group
Technology Domain
Contributors
Litigation
Firms
Documents
Filings
Litigants
SEPs
Core Network & Terminals (CT)
CT1
Specifications for interfaces between Radio Access
Network and Core Network (Iu-CS, Iu-PS)
69
2,761
4
8
8
CT3
Specifications for interfaces between Core Network and
external networks
43
948
0
0
0
CT4
Specifications for Core Network
51
2,234
2
6
2
CT6
Smart card applications
45
208
0
0
0
Radio Access Network (RAN)
RAN1
Specifications for interface between User Equipment
and Node B (Uu)
89
10,273
35
38
54
RAN2
Architecture for User Equipment to Node B interface
95
12,488
26
33
40
RAN3
Specifications for interfaces between Node B and Radio
Network Controller (RNC) and between RNCs (IuB, IuR)
72
6,218
10
14
12
RAN4
Radio frequency aspects, repeaters, and radio perfor-
mance
114
13,084
4
7
6
RAN5
Specifications for conformance testing of the User
Equipment interface
79
8,696
0
0
0
Service and System Aspects (SA)
SA1
Requirements for services and features of the system
98
2,808
9
12
10
SA2
System architecture (except for RAN)
103
3,777
11
16
14
SA3
Requirements and architecture for system security
60
1,248
6
9
5
SA4
Specifications for audio and video codecs
56
604
8
16
18
SA5
Requirements and architecture for overall system man-
agement
40
1,884
1
4
1
Notes. Observation window is 2005-Q2 to 2012-Q3. Technical reports and discussion documents are included in the
document count; change requests, liaison requests, and withdrawn or unknown documents are excluded. SEPs =
standard essential patents.
Evolution of cooperation in the face of conflict 11
Patent litigation can also occur when
plaintiffs believe defendants are demanding
unreasonable and unfair licensing terms.
3GPP requires firms to license patents under
FRAND terms. However, the nature of FRAND
is loosely defined in SDO bylaws, and the liti-
gants may disagree about the value of the li-
censed SEPs. For example, Samsung filed a
complaint in 2007 against Interdigital alleging
anticompetitive licensing of its 3G technology,
claiming Interdigital was violating FRAND
terms.
3
In either situation, patent litigation high-
lights disagreement about the validity and
value of IP rights. The payoff to the owners of
standard-essential patents is substantial be-
cause firms throughout the industry need li-
censes to make products under the standard.
The FRAND licensing terms permit SDOs to set
standards while mitigating the threat of hold-
up over the proprietary technology. Thus, a
3
Case reference 1:07-cv-00167, Delaware District Court
member firm trades away its right to refuse li-
censing to anyone willing to pay FRAND roy-
alties so that its patented technology can be
included in the worldwide standard. However,
despite that FRAND terms generate broad
public benefits as well as private benefits to
patent holders, determining the value of the
IP can be contentious.
Table 2 lists the 3GPP firms that were
most often involved in SEP litigation during
the sample period. Most litigation was settled
out of court, but some went to trial. A number
of suits also spurred countersuits. SEPs being
litigated often affected more than one WG be-
cause the technology underlying an SEP can
touch specifications across WGs.
3.2 Sample
Data sources. Our sample comprises data from
four sources. First, our sample of technical
Table 2. Standard essential patent litigation: Most litigious firms
Litigant role
Case outcome
Suits with
counter-
suits
WGs af-
fected
Firm
Plaintiff
Defendant
Went to
trial
Settled or
dismissed
Unknown
Apple
4
9
1
8
4
6
10
Motorola
1
6
0
6
1
2
7
Samsung Electronics
3
4
0
4
3
4
6
Motorola Mobility
3
3
0
6
0
4
5
Nokia
1
5
0
6
0
0
8
Cisco Systems
3
1
0
1
3
0
8
Interdigital
3
1
0
4
0
2
6
Motorola Solutions
3
1
0
2
2
0
1
Qualcomm
2
2
1
2
1
2
8
SPH America
3
1
0
4
0
2
1
Broadcom
2
1
1
2
0
2
5
Hewlett-Packard
2
1
0
2
1
0
1
Research In Motion
1
2
0
3
0
0
7
Sony-Ericsson
1
2
0
3
0
1
3
Other firms (38 total)
17
33
9
33
8
4
11
Notes. Table lists firms involved in 3 or more cases filed between 2005-Q2 and 2012-Q3.
Jones, Leiponen, and Vasudeva 12
contributions was provided by the Searle Cen-
ter on Law, Regulation, and Economic Growth
at Northwestern University (Baron & Spulber,
2018). The Searle Center collected this data
from 3GPP.org. Second, we collected a sample
of all known SEPs from ETSI.org. The Euro-
pean Telecommunications Standards Insti-
tute (ETSI) works closely with 3GPP to publish
globally applicable telecommunications
standards. Using ETSI, firms declare their pa-
tents that they deem essential to specific tech-
nical specifications. Third, we collected SEP
and non-SEP litigation data between 3GPP
firms from Lex Machina, which mines and ag-
gregates litigation data in the U.S. (Most case
filings dealing with telecommunications pa-
tent disputes are litigated in the U.S.) Fourth,
we downloaded patent citation data from
Thompson Reuters for declared SEPs and alli-
ance data for firms in our dataset.
The 3GPP contribution data spans from
2005-Q2 through 2012-Q3. This time window
covers late contributions to the 3G standard
and early to middle contributions to the 4G
standard (Release 7 through 11). The contribu-
tion data were extracted from contribution
documents and included (a) the assigned WG,
(b) the submission date to a WG meeting for
consideration and approval, and (c) the firms
that authored the contribution. For each WG
and quarter, we counted the number of docu-
ments coauthored by each dyad. The total
number of dyadic coauthoring observations
was 1.4 million.
We connected 3GPP contributions and
SEPs through the technical specification num-
ber and firm name published in the ETSI data.
Thus, we can link SEPs to specific working
groups. Further, we added SEP litigation from
Lex Machina using the patent publication
number (which matches to SEPs in the ETSI
data) and firm names documented in the case
filing. Non-SEP litigation was matched by firm
name. Finally, we added patent forward-cita-
tions from Thompson Reuters using the pa-
tent publication number, and matched alli-
ance data collected from SDC Platinum by
firm name.
Panel creation. To test our hypotheses, we
used a subset of the data. In H1 through H3, we
are interested in plaintiff-defendant dyads; in
H4, we focus on co-defendant and defendant-
other dyads. The plaintiff-defendant and co-
defendant dyads include firms named in the
litigation. The defendant-other dyads include
an ego who was a defendant and an alter who
was not party to the case. While other dyad
combinations are possible, we focus only on
defendant dyads where our arguments and
hypotheses are the most apposite. These ob-
servations constitute the set of treated dyads.
We matched the treated dyads with similar
untreated dyads. We detail our matching pro-
cess in the Model Specification subsection.
Each dyad is observed for nine WG-quar-
ters: four quarters before filing, the filing pe-
riod, and four quarters after filing. The panel
included 287 unique firms; 8,377 unique dy-
ads; 17 quarters with filings; 14 WGs; and
217,296 observations. For H1 through H3, we
focused only on plaintiff-defendant dyads,
which were a small subset of the panel includ-
ing 148 unique dyads and 2,304 observations.
(Most observations in the panel are defend-
ant-other dyads.)
3.3 Variables
Dependent variable. Our variable of interest is
the number of contribution documents coau-
thored by a dyad within a quarter. It is calcu-
lated as a simple count of documents by a
dyad in a given WG at a given quarter. There
are different types of contribution documents
(Baron & Gupta, 2018); we include technical
Evolution of cooperation in the face of conflict 13
reports and proposals and discussion docu-
ments because they best capture new tech-
nical collaborations. They represent the fore-
front of new standards development and are
often produced as technical specifications are
emerging. Firms gain value from producing
these documents because they establish a
technical trajectory for the standards and bol-
ster the value of the IP on which they are built.
We omit change request documents because
they are primarily backward looking and re-
flect prior cooperative efforts. Coauthored
change requests often represent past rather
than new cooperative strategies.
Independent variables. Our independent
variables include the litigation filing treat-
ment, dyad type (plaintiff-defendant, co-de-
fendant, and defendant-other), and year pre-
and post-filing. Treated dyads are coded as 1;
matched untreated observations are coded as
0. Post-filing is set to 1 for quarters +1 to +4; 0
otherwise.
Moderators. The technological distance
variable was calculated as the Euclidean dis-
tance between the dyad members’ technolog-
ical profiles within a WG. A firm’s technologi-
cal profile within a WG was a vector of cumu-
lative document counts. Each vector element
held the cumulative number of documents au-
thored by a firm in a technical specification
assigned to the WG up to the given quarter. A
dyad’s technological distance could vary from
one WG to another. Defendant’s relational po-
sition was calculated as the defendant’s de-
gree centrality within a WG which was calcu-
lated as the sum of coauthoring ties with other
firms in a WG, excluding the plaintiff. For this
4
For instance, let A and B be the firms in a focal dyad
and let C and D be other firms in the network. If C co-
authored 1 document with A in a given WG-quarter
and 0 documents with B, then C would not provide
third party closure ( ). If D coauthored 9
measure, a coauthoring tie had a value of 1 if
the defendant had coauthored with the firm
within the prior year; 0 otherwise. Because
control observations do not have litigation at-
tributes, defendant’s relational position was
not calculated for them. This second modera-
tor only applies to treated observations.
Control variables. We included working
group dummies and quarter dummies to ac-
count for differences in levels of coauthoring
in different technological areas and over time.
We included an off-quarter filing control to
adjust for multiple treatments over time. For
instance, if a dyad was affected by litigation
+4 quarters after the focal treatment (because
another case was filed), then the dyad off-
quarter filing variable was set to 1 at the +4
quarters observation. This isolated the focal
treatment from the effect of filings in other
observation quarters.
We controlled for attributes of WG dyads.
We included third-party closure to account
for the constraints third parties imposed on
the focal dyads. We calculated it as the geo-
metric mean of coauthoring counts for the
prior year for ties between dyad members and
a third-party; these means were then summed
for all third parties.
4
We included the com-
bined citation-weighted patents within a dyad
and the standard deviation of citation-
weighted patents to account for the total tech-
nological strength of the firms and the differ-
ence in their technological strength, respec-
tively. Citation-weighted patents were calcu-
lated as the 5-year forward citations for SEPs
associated with the WG and owned by the
firms in the dyad. We included a dummy for
documents with A and 1 document with B, then clo-
sure by D would be . The total third-party
closure for the A-B dyad would be for the
given WG-quarter.
Jones, Leiponen, and Vasudeva 14
no authoring by a dyad member, which was
coded as 1 if either firm in the dyad had not
authored documents in the WG up to the given
quarter. This dummy isolated instances of no
defined technological distance. Finally, we in-
cluded the number of alliances between dyad
members and the number of non-SEP-related
cases where dyad members were co-litigants
(i.e., co-defendants or co-plaintiffs) or oppos-
ing litigants. These variables account for ways
in which interactions outside the context of
3GPP might affect litigation and coauthoring.
Note that we observed no publicly reported al-
liances between plaintiffs and defendants
over the observation window, but there were
alliances between defendants and other WG
members; thus, the effect of alliances only ap-
pears in our models that include defendant-
other dyads.
We also controlled for attributes of the
SEP-related litigation. We included the num-
ber of 3GPP firms that were listed as litigants
in open cases and the number of SEPs listed in
open cases within a WG-quarter. This ac-
counted for changes in the overall level of liti-
gation within a WG. We also included a
dummy for whether the litigation was part of
a suit-countersuit. A willingness to counter
litigation may indicate particular value in the
underlying technologies being litigated,
which may change the way in which firms re-
spond when litigation is filed.
3.4 Model Specification
We used a difference-in-differences approach
to test our hypotheses. The treatment we ob-
served was a plaintiff filling a lawsuit against
defendants’ alleged SEPs infringement or un-
fair licensing terms. This litigation treatment
is not purely exogenous to the system because
the treatment occurs only after negotiations
or other remedies break down. Thus, there
can be selection effects into the treatment
group. To mitigate this concern, we matched
litigating dyads with a comparable set of non-
litigating dyads through the quarter in which
a suit was filed. We then used the paired treat-
ment and control observations to estimate the
effect of SEP litigation on subsequent coau-
thoring. Because we only included lawsuits
containing SEPs (and excluded unrelated suits
such as those related to products), we isolated
the effect of litigation directly related to
standards development. Thus, we can be more
confident that the change in cooperation post
filing is directly related to the IP conflict.
Our observed response was the coauthor-
ing of dyad in working group at filing
quarter and pre- or post-filing (
). To simplify the notation, we rep-
resent the combined index simply as in-
dex . We modeled coauthoring between
plaintiffs and defendants as a negative bino-
mial distribution with mean and variance,
, where is the
dispersion parameter that allows the variance
to be greater than the mean. We used a nega-
tive binomial model instead of a linear or Pois-
son model because our response was a non-
negative count variable with over-dispersion.
The mean was modeled as:
,
where: are year-quarter dummies, esti-
mates the difference in coauthoring between
treated and untreated dyads ( when is
a treated dyad); estimates the coauthoring
difference-in-differences ( when is a
treated dyad and ); is a column vector
of parameter estimates for the control varia-
bles in row vector (including dummy var-
iables for working group); and is the error
term.
Evolution of cooperation in the face of conflict 15
Unlike a linear model where each term is
added to the model, the terms in a negative bi-
nomial model enter multiplicatively. This can
be seen when we exponentiate the function:
.
For a model such as ours, the average
treatment effect of the treated (ATT) would
normally be the statistic of interest. The ATT
would be appropriate if the average was indic-
ative of the treatment effect across the various
dyads. However, the level of participation in
coauthoring across the sample of dyads varies
widely, and it is likely that the treatment ef-
fect will differ based on the level of participa-
tion. More-active firms will experience a
larger effect of litigation, whereas less-active
firms will experience a smaller effect of litiga-
tion. Thus, we focus on the average treatment
rate on the treated instead of the average ab-
solute effect on the treated. The rate estimate
better reflects the effect of the treatment than
the absolute value because of the heterogene-
ous activity in the sample. Assuming a con-
stant rate for the sample allows less-active
firms to have smaller absolute treatments and
more-active firms to have greater absolute
treatments. We define the average treatment
rate for the treated () as:
,
where and are the potential outcomes
with and without the treatment, respectively.
5
When a litigating dyad is treated, then
;
counterfactually, if a litigating dyad were not treated,
then . Thus,
the average treatment rate on the treated is simply
raised to the power:
.
When the moderators (technological distance and rela-
tional position) are added to the model, the treatment
rate is multiplied by , where and
The average treatment rate for the treated is
estimated as .
5
Causal inference. Our interest is the effect
of a litigation filing (the treatment) on a
dyad’s subsequent coauthoring. Accurately
estimating the causal effect is complicated by
the observational nature and complexity of
the data. First, litigating dyads are not inde-
pendent of other dyads, and the effect of liti-
gation may spill over within the network. We
addressed the problem of non-independent
treatments by explicitly modeling the dyad
types of interest affected by litigation and by
removing other dyad types from the treat-
ment group. Specifically, we demarcate plain-
tiff-defendant, co-defendant, and defendant-
other dyads as treated observations. Other dy-
ads within the network for a given WG-quar-
ter that could be affected by the treatment
were excluded from estimation.
Second, litigants are not randomly dis-
tributed within the sample. It is often the
more active participants in a WG who contrib-
ute frequently that are litigants. Furthermore,
there may be time-variant reasons for plain-
tiffs to both litigate and co-author with de-
fendants. We addressed the problem of sam-
ple selection using a matching procedure. We
chose a matching procedure over a two-step
selection model because we have many in-
formative covariates but lack an appropriate
instrument. The matching procedure allowed
are the values of two moderators and and are the
parameters for the and interactions, re-
spectively. The moderators possess only nonnegative
values. Thus, when is zero the treatment rate is
, holding constant; when is greater than
zero, then the treatment rate is
(again holding constant), such that the treatment
rate will rise or fall depending on whether is positive
or negative, respectively. The effect of is similarly
calculated, holding constant.
Jones, Leiponen, and Vasudeva 16
us to match treated dyads with untreated dy-
ads that were similar on observable charac-
teristics. The standard assumption for a
matching model such as ours is that the treat-
ment assignment is independent of the poten-
tial outcome given the observable covariates
(Stuart, 2010).
We matched exactly the quarter in which
the litigation occurred and then used nearest
neighbor matching. We chose nearest neigh-
bor matching because it is easy to understand
and execute, supports exact matching, and has
been shown to be effective among alternatives
choices (Austin, 2014). We used the shortest
Mahalanobis distance to find comparable ob-
servations. We chose Mahalanobis distance
matching because it is shown to approximate
a fully blocked experimental design (Iacus,
King, and Porro, 2011) and recent work has
questioned the effectiveness of propensity
score matching (King and Nielsen, 2019). We
matched without replacement and selected
one untreated observation for every treated
observation. The pool of untreated observa-
tions included dyads that were untreated and
were in untreated WG-quarters. We excluded
all dyads within a treated WG-quarter and
firm pairs in untreated WGs that were treated
in a different WG. We excluded these observa-
tions from the untreated pool because spillo-
vers could affect such dyads, making them un-
desirable matches. For purposes of matching,
the pre-treatment period was the four quar-
ters prior to a filing and the filing quarter.
We used the following covariates in the
matching procedure: pre-treatment coauthor-
ing in each of the 5 quarters; third-party clo-
sure (calculated based on the pretreatment
period); dyad members’ combined citation-
weighted patents; standard deviation of cita-
tion-weighted patents; technological distance;
no authoring by a dyad member; number of
alliances between dyad members; number of
non-SEP-related cases where dyad members
were co-litigants; and the number of non-SEP-
related cases where dyad members were op-
posing litigants. Other control variables were
not included because they only applied to
treated dyads (e.g., litigation attributes).
24,783 treated dyads were matched to dy-
ads in a pool of 739,121 untreated observa-
tions. Diagnostics showed that the standard-
ized difference (Stuart, 2010) between treated
and untreated observations improved (i.e., de-
creased) due to matching. The standardized
difference (a heuristic for matching suffi-
ciency) for all matched attributes was less
than 0.25. Thus, the matching procedure sub-
stantially improved the similarity of the non-
litigation dyads to the litigation dyads.
3.5 Descriptive Analysis
To understand why firms litigated and coau-
thored, we conducted preliminary descriptive
analysis by aggregating the data to the WG-
firm-quarter level (32,550 observations with
287 unique firms). We used logit models to
predict (a) the odds of suing another firm in
the WG, (b) the odds of being sued by a WG
firm, and (c) the odds of coauthoring with an-
other firm. We examined attributes of firms
and WGs in quarter to predict the response
variables in .
The litigation analysis revealed that firms
were more likely to file a lawsuit (i.e., be a
plaintiff) if they owned substantial IP related
to the WG and if they were already plaintiffs
in non-SEP-related suits. On average, a firm’s
odds of filing a lawsuit increased 15 percent
for every one thousand citations to the firm’s
WG-related patents. (Firms with patents aver-
aged 1,742 citations in a WG.) Additionally, a
firm’s odds of filing increased 4 percent for
every additional non-SEP-related suit in
Evolution of cooperation in the face of conflict 17
which they were a plaintiff. (Firms averaged 2
suits.) This suggests that firms are apt to as-
sert IP rights if they own substantial IP and
have a capability to employ litigation as a
means of assertion.
The two predictors also predicted a firm’s
likelihood of being sued (i.e., being a defend-
ant). This is not surprising with the substan-
tial number of countersuits within this con-
text. However, firms were less likely to be sued
if they had a higher percentage of their tech-
nical activity in the WG or if the overall num-
ber of SEPs being litigated in a WG increased.
On average, the odds of being sued decreased
10 percent with a 10 percent rise in a firm’s
share of its total contributions belonging to
the WG. (Firms averaged 11% of their total
contributions in a given WG.) This may sug-
gest a liability of newness: firms that are
building out their contributions in a WG may
become targets of litigation. Additionally, the
odds of being sued decreased 6 percent for
every additional SEP being litigated in the WG.
(WGs averaged 5 SEPs being litigated per
quarter.) This suggests that firms in a WG may
become more cautious after SEPs become en-
tangled in litigation.
The coauthoring analysis revealed that
firms were more likely to coauthor with other
firms in a WG if they owned IP in the WG, if a
larger percentage of their technical activity
was in the WG, and if they had a history of co-
authoring and a more central relational posi-
tion. On average, the odds of a firm coauthor-
ing increased 33 percent for every one thou-
sand patent citations. The odds increased 6
percent with a 10 percent rise in a firm’s share
of its total contributions belonging to the WG.
The odds increased 19 percent for every 100
additional cumulative documents. (Firms av-
eraged 86 cumulative documents.) And the
odds increased 16 percent for each additional
coauthoring partner (i.e., expansion of rela-
tional position; firms averaged 5 partners).
Overall, the coauthoring analysis suggests
that activity in the WG—whether measured
by cumulative documents, relational position,
IP, or firms’ share of technical activity—is re-
lated to higher coauthoring. Interestingly, IP
was related to litigation and coauthoring;
however, cumulative coauthoring and rela-
tional position were not related to litigation,
and firms’ share of technical activity in the
WG was related in the opposite direction. This
suggests litigation and coauthoring have dif-
ferent underlying drivers, although both are
connected to the presence of IP.
Note, this analysis is merely descriptive
and intended to elucidate the data we exam-
ine. We also present descriptive statistics—
means, standard deviations, and correla-
tions—in Table 3.
4. Results
Figure 1 illustrates the difference-in-differ-
ences effect for litigating dyads. The solid lines
depict the quarter-to-quarter coauthoring for
litigating and matched non-litigating dyads.
The dashed lines depict the average coauthor-
ing pre- and post-filing for these same dyads.
An inspection of the pre-filing trends in the
figure suggests they do not violate the parallel
trends assumption. Further, we tested
whether the level of coauthoring for any pre-
filing quarters differed. The t-values ranged
from −0.95 to 1.50, suggesting that the parallel
trends assumption was appropriate. Figure 1
shows that after a litigation event, coauthor-
ing in non-litigating dyads fell slightly while
coauthoring in litigating dyads increased sub-
stantially. Thus, consonant with H1, Figure 1
suggests that filings had a positive effect on
coauthoring between defendants and plain-
tiffs.
Jones, Leiponen, and Vasudeva 18
The models in Table 4 test the differences
illustrated in Figure 1. Model 1 presents a sim-
ple difference-in-differences model with no
controls. The results show that a case filing in-
creased coauthoring between defendants and
plaintiffs, which supports H1. The average
treatment rate for the litigating dyads was 3.1
( = 1.12; p = .001); that is, the average rate of
coauthoring increased three-fold after a case
filing. Model 2 adds the controls, which ad-
justs the treatment rate lower to 2.1 ( = 0.72;
p = .01). Prior to litigation, a defendant and
plaintiff averaged about one document per
year. The treatment rate in Model 2 indicates
that litigation increased coauthoring to about
two documents per year, on average.
The moderating effect of technological
distance and relational position is presented
in Model 3 and illustrated in Figure 2. Dyads
with greater technological distance increased
coauthoring more after a filing than those that
Table 3. Descriptive statistics
Mean
S.D.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(1)
Coauthored documents
0.18
1.25
(2)
Plaintiff-defendant dyad
0.01
0.10
0.00
(3)
Co-defendant dyad
0.01
0.12
0.04
-0.01
(4)
Defendant-other dyad
0.98
0.15
-0.03
-0.66
-0.74
(5)
Post-filing
0.44
0.50
0.01
0.00
0.00
0.00
(6)
Litigating dyad
0.50
0.50
0.07
0.00
0.00
0.00
0.00
(7)
Technological distance
0.17
0.51
0.05
-0.01
-0.01
0.01
0.00
-0.02
(8)
Defendant relational position
0.08
1.61
0.02
0.43
-0.01
-0.29
0.00
0.05
0.00
(9)
No authoring by dyad member
0.69
0.46
-0.18
-0.01
-0.01
0.02
0.00
0.01
-0.51
-0.03
(10)
Dyad members' citation-weighted pa-
tents
1.51
3.37
0.07
0.02
0.00
-0.01
0.00
0.14
0.00
0.03
(11)
Std. dev. of citation-weighted patents
1.02
2.29
0.05
0.01
-0.01
0.00
0.00
0.14
0.01
0.02
(12)
Dyad third-party closure
1.74
6.85
0.49
0.01
0.04
-0.04
0.00
0.13
0.10
0.05
(13)
Dyad alliances
0.01
0.13
0.01
-0.01
-0.01
0.01
0.00
0.01
-0.01
0.00
(14)
Dyad non-SEP cases, co-litigants
1.40
7.67
0.04
0.00
0.03
-0.03
0.02
0.03
-0.01
0.00
(15)
Dyad non-SEP cases, opposing litigants
0.06
0.33
0.03
0.14
0.03
-0.12
0.02
0.02
0.00
0.08
(16)
WG litigants in open cases
10.11
10.32
0.09
0.00
0.02
-0.01
0.12
0.48
-0.09
0.04
(17)
WG SEPs in open cases
6.87
6.35
0.07
0.00
0.01
0.00
0.06
0.52
-0.09
0.03
(18)
Off-quarter filing
0.04
0.20
0.00
0.01
0.00
-0.01
0.08
0.10
-0.03
0.01
(19)
Suit-countersuit
0.12
0.32
0.00
-0.01
-0.03
0.03
0.00
0.37
-0.08
0.00
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(10)
Dyad members' citation-weighted pa-
tents
-0.03
(11)
Std. dev. of citation-weighted patents
-0.02
0.98
(12)
Dyad third-party closure
-0.30
0.10
0.08
(13)
Dyad alliances
-0.02
0.02
0.02
0.01
(14)
Dyad non-SEP cases, co-litigants
-0.04
-0.02
-0.03
0.06
0.18
(15)
Dyad non-SEP cases, opposing litigants
-0.04
0.02
0.00
0.07
0.02
0.18
(16)
WG litigants in open cases
0.09
0.15
0.15
0.22
0.00
0.03
0.03
(17)
WG SEPs in open cases
0.10
0.16
0.16
0.18
0.00
0.02
0.02
0.90
(18)
Off-quarter filing
0.06
0.07
0.07
0.03
0.01
0.02
0.02
0.20
0.18
(19)
Suit-countersuit
0.12
-0.03
-0.03
0.04
0.01
0.04
0.01
0.18
0.23
0.17
Notes. Total observations = 217,296; defendant-plaintiff dyads = 2,304; Co-defendant dyads = 2,934; defendant-
other dyads = 212,058.
Evolution of cooperation in the face of conflict 19
were less distant ( = 1.97; p = 0.002), which
supports H2. Panel A in Figure 2 illustrates the
predicted mean response and the lower bound
of the 95% confidence interval, which was cal-
culated using nonparametric paired (treat-
ment with control) bootstrapping with 10,000
iterations. When technological distance was
low ( = 0.00) and defendant relational po-
sition was held constant ( = 14), the aver-
age treatment rate was 2.2; when technologi-
cal distance was high ( = 0.30), the treat-
ment rate rose to 4.1. Thus, the coauthoring re-
sponse to litigation increased with greater
technological distance. When technological
distance was greater than 0.07, the lower
bound of the 95% confidence interval was
above 1. (Coauthoring increases post-filing
when the rate is above 1.)
When the defendant’s relational position
was on the periphery of the network (mean-
ing the defendant coauthored with few other
firms), then the defendant coauthored more
with the plaintiff post-filing. Plaintiff-defend-
ant coauthoring post-filing decreased when a
defendant was more central ( = −.05, p =
.002), which supports H3. Panel B in Figure 2
illustrates the predicted response. When rela-
tional position was low ( = 0) and techno-
logical distance was held constant
( = 0.14), then the treatment rate was 5.5;
when relational position was high ( = 30),
then the treatment rate dropped precipitously
to 1.4. Thus, the coauthoring response for de-
fendants with many other coauthoring ties
was lower.
The treatment rate for plaintiff-defend-
ant, co-defendant, and defendant-other dyads
was estimated by Model 4. The average treat-
ment rate for plaintiff-defendant dyads was
3.7 (est. = 1.31, p < .001), which was somewhat
higher than the Model 2 estimate. The average
treatment rate was 2.1 for co-defendant dyads
(est. = 0.73, p = .001) and 2.5 for defendant-
other dyads (est. = 0.91, p < .001), which sup-
ports H4a and H4b. Co-defendants averaged
about two coauthored documents per year
prior to a case filing; a filing increased coau-
thoring with another defendant to about four
documents per year, on average. A defendant
coauthored an average of just less than one
document per year with another member of
the working group prior to a case filing; a fil-
ing increased coauthoring to two to three doc-
uments per year. On average, a defendant
would coauthor with 75 other WG members.
By multiplying the dyadic annual average pre-
filing rate by the average number of defend-
ant-other dyads, we found that a defendant
Figure 1. Technical document coauthoring pre-
and post-filing for defendant-plaintiff dyads.
Solid lines depict quarter-by-quarter
coauthoring; dashed lines depict average
coauthoring pre- and post-filing. The darker
(blue) line depicts coauthoring between
litigating dyads; the lighter (yellow) line depicts
coauthoring between matched control dyads.
Jones, Leiponen, and Vasudeva 20
Table 4. Negative binomial difference-in-differences models
Model 1
Model 2
Model 3
Model 4
Est.
S.E.
Est.
S.E.
Est.
S.E.
Est.
S.E.
Baseline
-2.09
(0.36)
Post filing
-0.46
(0.31)
Litigating dyad ()
0.51
(0.58)
-0.27
(0.33)
-1.04
(0.41)
-0.33
(0.33)
Litigating dyad post-filing ()
1.12
(0.35)
0.72
(0.29)
1.43
(0.64)
1.31
(0.29)
Technological distance moderator ()
Tech. distance * Post-filing
-0.80
(0.48)
Tech. distance * Litigating dyad
-1.59
(0.18)
Tech. distance * Litigating dyad post-filing ()
1.97
(0.62)
Defendant relational position moderator ()
Relational position * Litigating dyad
0.05
(0.01)
Relational position * Litigating dyad post-filing ()
-0.05
(0.01)
Co-defendant dyad
Relative baseline
0.71
(0.20)
Co-defendant dyad
-0.81
(0.41)
Co-defendant dyad post-filing
0.73
(0.23)
Defendant-other dyad
Relative baseline
0.09
(0.23)
Defendant-other dyad
-0.30
(0.15)
Defendant-other dyad post-filing
0.91
(0.10)
Controls
Technological distance
0.45
(0.17)
0.69
(0.22)
-0.10
(0.11)
No authoring by dyad member
-3.81
(0.70)
-3.95
(0.68)
-2.01
(0.16)
Dyad members' citation-weighted patents
0.11
(0.02)
0.11
(0.03)
0.08
(0.03)
Std. dev. of citation-weighted pa-
tents
-0.06
(0.04)
-0.09
(0.04)
-0.08
(0.04)
Dyad third-party closure
0.08
(0.02)
0.05
(0.02)
0.12
(0.01)
Dyad alliances
0.39
(0.15)
Dyad non-SEP cases, co-litigants
0.09
(0.03)
0.10
(0.03)
0.00
(0.00)
Dyad non-SEP cases, opposing litigants
-0.24
(0.21)
-0.30
(0.20)
-0.01
(0.08)
WG litigants in open cases
-0.10
(0.04)
-0.10
(0.04)
0.00
(0.02)
WG SEPs in open cases
0.16
(0.07)
0.18
(0.08)
-0.05
(0.03)
Off-quarter filing
-0.03
(0.46)
-0.17
(0.50)
0.11
(0.09)
Suit-countersuit
0.05
(0.31)
0.03
(0.33)
-0.44
(0.26)
WG dummies
Included
Included
Included
Quarter dummies
Included
Included
Included
Dispersion parameter
18.4
2.8
2.6
6.9
Akaike information criterion (AIC)
1,811
1,413
1,407
109,444
Observations
2,304
2,304
2,304
217,296
Notes. Cluster robust standard errors (clustered on firm and partner) in parentheses. The effects of the moder-
ators are similar whether they are modeled separately or together, so Model 3 presents the moderators to-
gether. Defendant relational position only applies to treated dyads, so there is no ‘relational position x post fil-
ing’ interaction.
Evolution of cooperation in the face of conflict 21
was, on average, co-authoring about 70 docu-
ments per year pre-filing and 175 documents
per year post-filing.
The average treatment rates did not differ
for the three types of dyads. In Model 4, the
estimated rate for defendant-plaintiff dyads
was greater than the rate for defendant-other
dyads, which was greater than the rate for co-
defendant dyads. However, a likelihood ratio
test revealed that the coefficients did not dif-
fer statistically (χ2 = 2.90; d.f. = 2; p = 0.23).
4.1 Supplemental analyses
Learning as a competing argument. One com-
peting argument posits that learning opportu-
nities may cause increased cooperation after
litigation. According to this argument, the dis-
pute arises because the litigants lack under-
standing of one another, which would compel
them to cooperate to learn more about one an-
other, and hence improve the process of
standard development in the working group.
In this scenario, litigants without a history of
cooperation should cooperate more after a
conflict event than litigants with a history of
cooperation. To test this possibility, we reran
Model 2 with the cumulative count of coau-
thored documents (which captures repeated
cooperation) as a moderator. If the coefficient
for the moderation effect was negative, then it
would support this argument. However, the
cumulative count moderation was positive
(est. = 0.03; p = 0.28). This is counter to what
we would expect if learning was driving the
results.
Identification and threats to exogeneity.
To account for the non-randomness of the lit-
igation treatment, we used a difference-in-dif-
ferences approach with a matching procedure
and various controls as described above. How-
ever, there may still be time-varying unob-
served characteristics that affect litigation
and coauthoring. (Such characteristics are
more likely to bias the plaintiff-defendant dy-
ads than the defendant-other dyads because
the “other” firms are not party to the conflict
Figure 2. Predicted effect of litigation on plaintiff-defendant coauthoring moderated by (A)
technological distance and (B) defendant relational position. Predicted effects are based on Model
3. The litigation effect is measured as a rate, so values above 1 increase coauthoring, whereas
values below 1 decrease coauthoring. C.I. = confidence interval.
Jones, Leiponen, and Vasudeva 22
that culminates in a lawsuit.) In searching for
a valid instrument, we discovered two U.S. Su-
preme Court cases
6
within our time window
that would likely affect the propensity to liti-
gate but not co-authorship. However, when
we used the court cases as instruments, we
found that they only weakly predicted litiga-
tion. We did not use them because weak in-
struments could lead to a more severe bias.
To address the concern of bias, we turned
to proxies that could account for the most
likely and problematic unobserved con-
founder. Our primary concern in the context
of 3GPP is that a change in the value of the
technology within the WG (which we do not
observe) could lead to litigation as well as
greater cooperation. In this scenario, not ac-
counting for a change in value could lead to
overstatement of the effect of litigation. To ac-
count for this possibility, we constructed two
proxies that captured the change in value of
the underlying IP. First, we calculated the to-
tal number of documents authored within a
WG-quarter. This not only included technical
reports and discussion documents, but all
types of contributions and change requests.
We use this as a proxy because authoring of all
types is apt to increase if the value of IP is in-
creasing. Moreover, an increase in value is
likely to attract new firms to the WG. Thus, we
created a count of new firms entering each
WG. We reran model 2 with the proxies. Both
total WG authoring ( = 0.06; p = .02) and new
entrants ( = 0.12; p = .003) positively pre-
dicted coauthoring. Still, the litigation effect
remained consistent: in the model with the
proxies was 0.76 vs. 0.72 originally. These re-
sults enhanced confidence in our models.
6
KSR International Co. v. Teleflex Inc., 2007, and Ebay
v. MercExchange, 2006.
The role of plaintiff attributes. A defend-
ant’s reaction to litigation within the plaintiff-
defendant dyad may not only be based on the
attributes of the defendant (e.g., relational po-
sition) or the dyad (e.g., technological dis-
tance), but also on the attributes of the plain-
tiff. It is plausible that coauthoring after liti-
gation is more likely when a plaintiff has
greater IP to protect and a stronger litigation
capability, which could make the club benefits
more appealing to the defendant. To explore
this possibility, we considered the moderating
effects of: (a) a plaintiff’s IP, measured as the
plaintiff’s citation-weighted patents, and (b)
its litigating ability, proxied by the number of
non-SEP-related cases initiated by the plain-
tiff. We added these moderators to Model 2
and found that a plaintiff’s IP led to higher co-
authoring post litigation, but general litigat-
ing ability did not. However, upon including
the technological distance and defendant’s re-
lational position moderators in Model 3, the
effect of the plaintiff’s IP disappeared.
We further considered the effect of plain-
tiffs and defendants being entwined in other
litigation not related to SEPs. Thus, we added
to Model 3 a moderator for the number of
non-SEP-related cases where the plaintiff and
defendant were opposing litigants as well as
the proxy for plaintiff’s general litigating abil-
ity. This model revealed two opposite effects.
First, dyads coauthored more post-litigation
when plaintiffs possessed a demonstrated liti-
gation ability. The average treatment rate for
plaintiff-defendant dyads was 1.3 when the
plaintiff had no non-SEP-related litigation,
whereas the treatment rate was 3.5 when the
plaintiff had three open non-SEP-related
Evolution of cooperation in the face of conflict 23
cases (holding the other moderators con-
stant). In contrast, the post-filing coauthoring
rate lowered when plaintiff-defendant dyads
were engaged in litigation unrelated to SEPs.
When the dyad had no non-SEP-related litiga-
tion, then the average treatment rate was 5.7.
However, when the dyad was engaged in two
non-SEP-related cases, the average treatment
rate was 0.2. This model with the additional
moderators fit the data better than Model 3
(χ2 = 12.85; d.f. = 5; p = 0.02). Overall, the first
moderator reinforces our arguments: plain-
tiffs with a litigation capability will use it to
enforce patent rights and push defendants to-
ward greater cooperation. The second moder-
ator suggests a boundary condition: when a
plaintiff and defendant are litigating technol-
ogy unrelated to 3GPP SEPs, they may not
choose to cooperate more within 3GPP.
Plaintiff versus defendant ownership of
SEPs. A plaintiff may own SEPs and sue a de-
fendant for patent infringement or a defend-
ant may own SEPs and get sued by a plaintiff
for unreasonable or discriminatory licensing
terms. While both types of suits suggest disa-
greement about the value of the underlying
technology, one may drive the observed re-
sults more than the other. In our dataset,
plaintiffs held the SEPs in 75 percent of the
cases, and defendants held the SEPs in 25 per-
cent of the cases. We reran Model 2 and split
our results by whether the plaintiff or the de-
fendant possessed the SEPs. We found the
treatment rate was 4.0 when defendants held
SEPs ( = 1.39; p = .004) and the rate was 2.0
when plaintiffs held SEPs ( = 0.67; p < .02).
However, a likelihood ratio test revealed that
this alternative model did not fit better than
Model 2 (χ2 = 2.04; d.f. = 2; p = 0.36). This sug-
gests that cooperation will increase whether
the defendant or plaintiff holds patents.
Persistence of the litigation effect. We ex-
amined whether the effect of litigation is tem-
porary or persists over time. To test this, we
added four more quarters of data so that we
had eight quarters of coauthoring post-filing.
We then added a variable to Model 4 that was
1 for treated dyads in the second-year post-fil-
ing ( ); 0 otherwise. The existing
treatment × post-filing variable, , was also
coded as 1 for all eight post-filing quarters. A
near-zero estimate for the new parameter
would suggest a persistent effect from year 1
to 2, whereas a negative or positive coefficient
would indicate a diminishing or growing ef-
fect, respectively. For defendant-plaintiff dy-
ads, the rate of coauthoring dropped 34 per-
cent from the first-year to the second-year
post-filing (est. = −0.42; p = .44); for co-de-
fendant dyads, the rate of coauthoring
dropped 68 percent (est. = −2.16; p = .06); and
for defendant-other dyads, the rate of coau-
thoring rose 16 percent est. = 0.15; p = .41).
Only the co-defendant change in coauthoring
from year 1 to 2 was notable. Further analysis
showed that by the second year, co-defendant
rates of coauthoring were similar to pre-treat-
ment, but plaintiff-defendant dyads remained
2.2 times higher than pre-treatment and de-
fendant-other dyads remained 2.5 times
higher. Thus, except for co-defendants, litiga-
tion’s influence on coauthoring continued to
persist.
5. Discussion and conclusions
Our study examines the evolution of coopera-
tion in the face of conflict about valuable in-
tellectual property. In innovation ecosystems,
cooperation occurs frequently. In 3GPP, firms
participate in standards development to
jointly create technical specifications for mo-
bile telecommunications. However, conflict is
also frequent, and patent litigation regularly
Jones, Leiponen, and Vasudeva 24
hampers firms’ ability to derive value from
their IP or products. In light of this phenome-
non, we asked how firms respond to conflict
within an interdependent technological land-
scape.
Similar to how cooperation might evolve
in economic, political, or biological ecosys-
tems (Axelrod & Hamilton, 1981; Axelrod,
1984), we find that firms bound together in an
innovation ecosystem continue to cooper-
ate—and even increase their cooperation—
following a conflict. We argue that this occurs
because essential IP (the source of conflict)
cannot be bypassed without significant tech-
nology development, a costly and unappealing
option. Since defecting is a poor alternative,
adversarial firms engage in cooperation after
conflict to enhance private and club benefits
derived from complementary assets (Teece,
2018; Jacobides et al., 2018). Such a cooperative
approach also reduces the risk of future hold
up and prevents technological substitution by
rivals (Arora, Fosfuri, & Gambardella, 2001;
Polidoro & Toh, 2011).
The extent to which contesting firms ben-
efit from cooperation depends on whether
their technological profiles are complemen-
tary. If litigating firms are more technologi-
cally similar or less distant, they have fewer
joint opportunities and pose greater competi-
tive threats to one another. In contrast, more
technologically distant firms (within the eco-
system) have greater potential for comple-
mentarities that generate club benefits. Also,
the cooperative benefits for litigants is rela-
tive to other opportunities in the co-creation
network. If a defendant holds a central net-
work position, it enjoys more outside options
and becomes less inclined to cooperate with
the aggressor. Further, while a firm may
choose to cooperate with an aggressor, it may
in tandem cooperate with others to move the
technological trajectory away from the con-
tested IP and safeguard against future attacks
(Ziedonis, 2004).
5.1 Contributions
Our study contributes to four conversations in
the strategic management literature. First,
our study is among the first within strategic
management to focus on outright conflict be-
tween firms and its subsequent implications
for cooperation. This omission probably owes
to the fact that while interfirm conflict is fre-
quent, it is rarely well-documented, making it
difficult to examine. As a result, conflict is
viewed as a perpetual state of disagreement
(cf. Sytch & Tatarynowicz, 2014). In contrast,
we trace a more nuanced set of strategic ac-
tions that present conflict as an evolutionary
aspect of competition. Our key insight is that
the effect of conflict on subsequent coopera-
tion in innovation ecosystems is more com-
plex than the extant literature on interfirm
cooperation would lead us to believe. Conflict
does not simply lead to firms severing cooper-
ative ties (Baker Faulkner, & Fisher, 1998;
Greve, Baum, Mitsuhashi, & Rowley, 2010;
Polidoro, Ahuja, & Mitchell, 2011). Depending
on their technological and relational positions
in the ecosystem, parties in conflict can use
cooperative arrangements to build comple-
mentarities or redirect technological develop-
ment to their advantage. Moreover, we
address the question of how conflict between
two firms triggers broader shifts in the
technological landscape. Our finding that
firms reconfigure their interactions with
third parties post-conflict offers insights on
the dynamic ramifications of conflict within
innovation ecosystems (Helfat & Raubitschek,
2018).
Second, patent litigation is a special kind
of strategic conflict where one party demands
Evolution of cooperation in the face of conflict 25
concessions from another party via the legal
system. According to Somaya (2003), such de-
mands may arise from disputes about asset
valuations and the expected evolution of tech-
nologies and markets. The parties may also
hold asymmetric information and divergent
stakes in terms of their technological or mar-
ket positions. As a result, the parties may use
the legal process to sort out disagreements.
Our study builds on this view of litigation as a
strategic process. We conceptualize the filing
of a lawsuit as indication of changes in the
parties’ strategic stakes and describe the im-
plications for their cooperation. Prior litera-
ture has viewed litigation as a socially acrimo-
nious process (e.g., Sytch & Tatarynovicz,
2014). In contrast, we view litigation through
a strategic lens as an element of high-technol-
ogy competition that illustrates the parties’ at-
tempts to capture value from the complex and
rapidly evolving innovation ecosystem.
Third, litigation in high-technology eco-
systems is often the result of interdependen-
cies created by complementary technologies
(cf. Adner & Kapoor, 2010; Jacobides et al.,
2006; Teece, 2018). Jacobides et al. (2018)
highlight the importance of understanding
the direction and strength of complementari-
ties in order to design robust ecosystem gov-
ernance that enables collaboration and value
capture while incentivizing participation. We
extend their framework by demonstrating the
influence of technological and relational in-
terdependence on cooperative outcomes after
conflict. Our study generates novel insights
about the relationships between complemen-
tarities, cooperation, and value capture in in-
novation ecosystems: under strong comple-
mentarities, strategies to create joint value en-
tice cooperation, but outside options can
weaken the cooperative response and keep fu-
ture attempts at appropriation in check.
Fourth, while we build on prior studies in
cooperative strategy and game theory (e.g.,
Parkhe, 1993; Brandenburger & Stuart, 1996;
Casadesus-Masanell & Yoffie, 2007; Arslan,
2018), our conceptual framework introduces
the central role of club goods in innovation
ecosystems. Extant research focuses on public
and private benefits only, and conflates public
and club benefits under the notion of common
benefits (Arslan 2018). We argue that when
technological inputs are complementary
requiring continuous investment, club
benefits provide major incentives for
cooperation. By solely focusing on private
benefits as the driver of cooperation, one may
miss the importance of club benefits to
sustain broader collective action.
Finally, from an empirical standpoint,
our study examines repeated interaction in an
interdependent setting. Repeated interaction
is an important aspect of competitive strategy
(Gulati, 1995), but we are not aware of studies
with long panels of cooperative activity inter-
spersed with disagreement and potential co-
operative failure (cf. Johnston & Waldfogel,
2002). Further, our setting allows us to match
firms operating in similar circumstances, ex-
cept for the litigation disturbance. Our study
is unique in tracing the impact of conflict on
closely related cooperative activities. Few
studies have achieved such a close corre-
spondence between the conflict domain and
strategic actions associated with it. We can
generate exceptionally specific insights com-
pared to studies with firm-level and annual
observations of firm behavior and outcomes
where the list of possible confounding factors
is excessive.
5.2 Limitations and future research
Patent litigation is primarily about licensing
negotiations for using existing IP, which is
Jones, Leiponen, and Vasudeva 26
backward-looking. However, with repeated
interactions, litigation can also involve for-
ward-looking strategic behavior (e.g., Macchi-
avello & Morjaria, 2015). That means a plain-
tiff’s litigation strategies can be tied to its sub-
sequent cooperation, confounding our efforts
to isolate the litigation effect. Nonetheless, to
our knowledge, the trigger to initiate litigation
stems from the evolving value of the underly-
ing technology. Thus, we believe that our re-
search design of matching highly similar dy-
ads around the quarter of a lawsuit filing is
able to reasonably identify the impact of liti-
gation on cooperation, even though the
broader technology market may influence
both litigation and cooperation decisions. In
particular, the quarterly difference-in-differ-
ences design is able to isolate the immediate
impact of filing from the general, gradual ap-
preciation of SEPs. Furthermore, including
additional proxies for demand for the technol-
ogy in the marketplace does not substantially
change our results.
The effect of litigation on cooperation be-
tween a defendant and other parties including
co-defendants is relatively less confounded by
the plaintiff’s strategies. If a defendant coop-
erates with other parties after being sued, the
defendant’s is more likely responding to the
litigation and not the plaintiff’s pre-existing
strategy. Nevertheless, as with all empirical
strategy research, the decisions of individual
firms are influenced by developments in the
marketplace, and ideally, we would like to test
the robustness of our results with an instru-
ment that affects the cost of litigation but is
unaffected by the market or other unobserva-
ble factors.
Relatedly, our study of cooperation after
conflict hinges on the assumption that
adversaries can calculate with some certainty
the expectation of future payoffs—be they
private or club benefits. However, uncertainty
about the evolution of the technology,
supporting industries and dominant players
can undermine such calculations.
Technological and commercial uncertainty is
high in the telecommunications industry
owing to the evolving generations of
technology which could usher in unforeseen
new players and radical innovations.
Moreover, conflict over technology in
standards development could be compounded
by conflict in other political and economic
domains. In particular, international conflicts
can arise in strategic industries and cast a
shadow on expected benefits from
cooperation between private actors.
5.3 Managerial implications
This study provides insights for managers of
high-technology firms innovating and com-
peting in innovation ecosystems. Conflict
drains resources, and firms need to be cogni-
zant of when to engage in it and when to tol-
erate or even accommodate aggression. For
example, engaging in patent wars or stand-
ards wars can be extremely costly. However,
when in conflict, firms that have cultivated
cooperative alternatives are better positioned
to strategically respond by diverting the tech-
nological development away from the aggres-
sor and towards less contested territory. Over-
all, such conflict need not be viewed as devas-
tating incidents of cooperative failure but as
instances of competition where expectations
about future payoffs have changed and where
firms can renegotiate their relationships to re-
fashion the technological trajectory. Occasion-
ally those negotiations fail and firms end up in
court. However, a swift reaction to cut off co-
operative activity may be misplaced. Instead,
it is more beneficial to assess whether it is ad-
vantageous in the long term to continue the
Evolution of cooperation in the face of conflict 27
existing arrangement or to divert investments
into new cooperative arrangements.
Acknowledgements
We wish to thank Scott Crawford, Robin Stitzing,
Kirti Gupta, Jussi Heikkilä, and Pekka Sääskilahti,
as well as participants at the Searle Roundtable on
Patents and Standards at Northwestern Univer-
sity, at the Wharton Technology Conference, at
Cornell University, at the Munich Summer Insti-
tute, and at the Whitman School of Management
at Syracuse University for their input on drafts of
this article.
References
3GPP.org (2019). Release 15. Retrieved from
http://www.3gpp.org/release-15.
Adner, R., & Kapoor, R. (2010). Value creation in
innovation ecosystems: How the structure of
technological interdependence affects firm
performance in new technology
generations. Strategic Management
Journal, 31(3): 306-333.
https://doi.org/10.1002/smj.821
Ansari, S., Garud, R., & Kumaraswamy, A. (2016).
The disruptor's dilemma: TiVo and the US
television ecosystem. Strategic Management
Journal, 37(9):1829-1853.
https://doi.org/10.1002/smj.2442
Arora, A., Fosfuri, A., & Gambardella, A.
(2001). Markets for technology: The economics
of innovation and corporate strategy.
Cambridge, MA: MIT press.
Arslan, B. (2018). The interplay of competitive
and cooperative behavior and differential
benefits in alliances. Strategic Management
Journal, 39(12): 3222-3246.
https://doi.org/10.1002/smj.2731
Austin, P. C. (2014). A comparison of 12
algorithms for matching on the propensity
score. Statistics in Medicine, 33(6), 1057-1069.
https://doi.org/10.1002/sim.6004
Axelrod, R. (1984). The Evolution of Cooperation.
Basic Books: New York.
Axelrod, R., & Hamilton, W. D. (1981). The
evolution of
cooperation. Science, 211(4489):1390-1396.
https://doi.org/10.1126/science.7466396
Baker, W. E., Faulkner, R. R., & Fisher, G. A.
(1998). Hazards of the market: The continuity
and dissolution of interorganizational market
relationships. American Sociological Review,
147-177. https://doi.org/10.2307/2657321
Bar, T., & Leiponen, A. (2014). Committee
composition and networking in standard
setting: The case of wireless
telecommunications. Journal of Economics &
Management Strategy 23(1):1-23.
https://doi.org/10.1111/jems.12044
Baron, J., & Gupta, K. (2018). Unpacking 3GPP
standards. Journal of Economics &
Management Strategy, 27(3), 433-461.
https://doi.org/10.1111/jems.12258
Baron, J., & Spulber, D. F. (2018). Technology
standards and standards organizations:
Introduction to the Searle Center Database.
Journal of Economics & Management Strategy,
27(3), 462-503.
https://doi.org/10.1111/jems.12257
Brandenburger, A. M. & Stuart Jr, H. W. (1996).
Value‐based business strategy. Journal of
Economics & Management Strategy, 5(1): 5-24.
https://doi.org/10.1111/j.1430-
9134.1996.00005.x
Buchanan, J. M. (1965). An economic theory of
clubs. Economica, 32(1): 1–14.
https://doi.org/10.2307/2552442
Casadesus-Masanell, R. & Yoffie, D. B. (2007).
Wintel: Cooperation and Conflict. Management
Science, 53(4): 584-598.
https://doi.org/10.1287/mnsc.1060.0672
Davis, J. P. (2016). The group dynamics of
interorganizational relationships:
Collaborating with multiple partners in
innovation ecosystems. Administrative Science
Quarterly, 61(4): 621–661.
https://doi.org/10.1177/0001839216649350
Diestre, L., & Rajagopalan, N. (2012). Are all
‘sharks’ dangerous? new biotechnology
ventures and partner selection in R&D
alliances. Strategic Management
Journal, 33(10), 1115-1134.
https://doi.org/10.1002/smj.1978
Dushnitsky, G., & Shaver, J. M. (2009).
Limitations to interorganizational knowledge
acquisition: The paradox of corporate venture
capital. Strategic Management Journal, 30(10),
1045-1064. https://doi.org/10.1002/smj.781
Jones, Leiponen, and Vasudeva 28
Dussauge, P., Garrette, B., & Mitchell, W. (2000).
Learning from competing partners: Outcomes
and durations of scale and link alliances in
Europe, North America and Asia. Strategic
Management Journal, 21(2), 99-126.
https://doi.org/10.1002/(SICI)1097-
0266(200002)21:2<99::AID-SMJ80>3.0.CO;2-G
Farrell, J., & Saloner, G. (1988). Coordination
through committees and markets. RAND
Journal of Economics, 19(2):235-252.
https://doi.org/10.2307/2555702
Farrell, J., & Simcoe, T. (2012). Choosing the rules
for consensus standardization. The RAND
Journal of Economics, 43(2), 235-252.
https://doi.org/10.1111/j.1756-2171.2012.00164.x
Fleming, L. (2001). Recombinant uncertainty in
technological search. Management
Science, 47(1), 117-132.
https://doi.org/10.1287/mnsc.47.1.117.10671
Fleming, L., & Sorenson, O. (2001). Technology as
a complex adaptive system: evidence from
patent data. Research policy, 30(7), 1019-1039.
https://doi.org/10.1016/S0048-7333(00)00135-
9
Fudenberg, D., Gilbert, R., Stiglitz, J., & Tirole, J.
(1983). Preemption, leapfrogging and
competition in patent races. European
Economic Review, 22(1), 3-31.
https://doi.org/10.1016/0014-2921(83)90087-9
Gandal, N., Kende, M. & Rob, R. (2000). The
dynamics of technological adoption in
hardware/software systems: The case of
compact disc players. RAND Journal of
Economics, 31(1): 43-61.
https://doi.org/10.2307/2601028
Greve, H. R., Baum, J. A., Mitsuhashi, H., &
Rowley, T. J. (2010). Built to last but falling
apart: Cohesion, friction, and withdrawal from
interfirm alliances. Academy of Management
Journal, 53(2), 302-322.
https://doi.org/10.5465/amj.2010.49388955
Gulati, R. 1995. Does familiarity breed trust? The
implications of repeated ties for contractual
choice in alliances. Academy of Management
Journal, 38(1), 85-112.
https://doi.org/10.5465/256729
Gulati, R., Puranam, P., & Tushman, M. (2012).
Meta‐organization design: Rethinking design
in interorganizational and community
contexts. Strategic Management Journal, 33(6),
571-586. https://doi.org/10.1002/smj.1975
Helfat, C. E. (1994). Evolutionary trajectories in
petroleum firm R&D. Management
Science, 40(12):1720-1747.
https://doi.org/10.1287/mnsc.40.12.1720
Helfat, C. E., & Raubitschek, R. S. (2018). Dynamic
and integrative capabilities for profiting from
innovation in digital platform-based
ecosystems. Research Policy, 47(8), 1391-1399.
https://doi.org/10.1016/j.respol.2018.01.019
Hirshleifer, J. (1991). The technology of conflict as
an economic activity. American Economic
Review (Papers and Proceedings), 81(2):130-
134.
Iacus, S. M., King, G., & Porro, G. (2011).
Multivariate matching methods that are
monotonic imbalance bounding. Journal of the
American Statistical Association, 106(493), 345-
361. https://doi.org/10.1198/jasa.2011.tm09599
Jacobides, M. G., Cennamo, C., & Gawer, A. (2018).
Towards a theory of ecosystems. Strategic
Management Journal, 39(8), 2255-2276.
https://doi.org/10.1002/smj.2904
Jacobides, M. G., Knudsen, T. & Augier, M.,
(2006). Benefiting from innovation: Value
creation, value appropriation and the role of
industry architectures. Research Policy, 35(8):
1200-1221.
https://doi.org/10.1016/j.respol.2006.09.005
Johnston, J. S., & Waldfogel, J. (2002). Does repeat
play elicit cooperation? Evidence from federal
civil litigation. The Journal of Legal
Studies, 31(1):39-60.
https://doi.org/10.1086/339468
Khanna, T., Gulati, R., & Nohria, N. 1998. The
dynamics of learning alliances: Competition,
cooperation, and relative scope. Strategic
Management Journal 19(3): 193-210.
https://doi.org/10.1002/(SICI)1097-
0266(199803)19:3<193::AID-SMJ949>3.0.CO;2-C
King, G., & Nielsen, R. (2019). Why propensity
scores should not be used for
matching. Political Analysis, 27(4), 435-454.
https://doi.org/10.1017/pan.2019.11
Lanjouw, J., Schankerman, M. (2001).
Characteristics of patent litigation: A window
on competition. RAND Journal of Economics,
32(1): 129-151. https://doi.org/10.2307/2696401
Layne-Farrar, A., Padilla, A. J., & Schmalensee, R.
(2007). Pricing patents for licensing in
standard-setting organizations: Making sense
Evolution of cooperation in the face of conflict 29
of FRAND commitments. Antitrust Law
Journal, 74(3), 671-706.
Lerner, J., & Tirole, J. (2015). Standard-essential
patents. Journal of Political Economy, 123(3),
547-586. https://doi.org/10.1086/680995
Li, J., & Matouschek, N. (2013). Managing
conflicts in relational contracts. American
Economic Review, 103(6), 2328-2351.
https://doi.org/10.1257/aer.103.6.2328
Macchiavello, R., & Morjaria, A. (2015). The value
of relationships: evidence from a supply shock
to Kenyan rose exports. American Economic
Review, 105(9), 2911-45.
https://doi.org/10.1257/aer.20120141
Milgrom, P., & Roberts, J. (1990). The economics
of modern manufacturing: Technology,
strategy and organization American Economic
Review, 80(3): 511-528.
Mowery, D. C., Oxley, J. E., & Silverman, B. S.
(1996). Strategic alliances and interfirm
knowledge transfer. Strategic management
journal, 17(S2), 77-91.
https://doi.org/10.1002/smj.4250171108
Oxley, J. E. (1997). Appropriability hazards and
governance in strategic alliances: A transaction
cost approach. The Journal of Law, Economics,
and Organization, 13(2), 387-409.
https://doi.org/10.1093/oxfordjournals.jleo.a0
23389
Olson, M. (1965). The logic of collective action.
Cambridge: Harvard University Press.
Ostrom, E. (2000). Collective action and the
evolution of social norms. Journal of Economic
Perspectives, 14(3):137-158.
https://doi.org/10.1257/jep.14.3.137
Paik, Y., & Zhu, F. (2016). The impact of patent
wars on firm strategy: Evidence from the
global smartphone industry. Organization
Science, 27(6), 1397-1416.
https://doi.org/10.1287/orsc.2016.1092
Parkhe, A. (1993). “Messy” research,
methodological predispositions, and theory
development in international joint
ventures. Academy of Management
Review, 18(2), 227-268.
https://doi.org/10.5465/amr.1993.3997515
Pisano, G. P. (1989). Using equity participation to
support exchange: Evidence from the
biotechnology industry. Journal of Law,
Economics & Organization, 5, 109.
Polidoro Jr, F., Ahuja, G., & Mitchell, W. (2011).
When the social structure overshadows
competitive incentives: The effects of network
embeddedness on joint venture
dissolution. Academy of Management
Journal, 54(1), 203-223.
https://doi.org/10.5465/amj.2011.59215088
Polidoro, F., & Toh, P.K. (2011). Letting rivals
come close or warding them off? The effects of
substitution threat on imitation
deterrence. Academy of Management
Journal, 54(2):369-392.
https://doi.org/10.5465/amj.2011.60263099
Ranganathan, R., & Rosenkopf, L. (2014). Do ties
really bind? The effect of knowledge and
commercialization networks on opposition to
standards. Academy of Management
Journal, 57(2), 515-540.
https://doi.org/10.5465/amj.2011.1064
Ranganathan, R., Ghosh, A., & Rosenkopf, L.
(2018). Competition–cooperation interplay
during multifirm technology coordination: The
effect of firm heterogeneity on conflict and
consensus in a technology standards
organization. Strategic Management Journal,
39(12), 3193-3221.
https://doi.org/10.1002/smj.2786
Rapoport, A., & Chammah, A. M. (1965). Sex
differences in factors contributing to the level
of cooperation in the Prisoner's Dilemma
game. Journal of personality and Social
Psychology, 2(6), 831-838.
https://doi.org/10.1037/h0022678
Rosenkopf, L., Metiu, A., & George, V. P. (2001).
From the bottom up? Technical committee
activity and alliance formation. Administrative
Science Quarterly, 46(4):748-772.
https://doi.org/10.2307/3094830
Rothaermel, F. T., & Boeker, W. (2008). Old
technology meets new technology:
Complementarities, similarities, and alliance
formation. Strategic Management
Journal, 29(1), 47-77.
https://doi.org/10.1002/smj.634
Rysman, M., & Simcoe, T. (2008). Patents and the
performance of voluntary standard-setting
organizations. Management science, 54(11),
1920-1934.
https://doi.org/10.1287/mnsc.1080.0919
Schelling, T. C. (1960). The Strategy of Conflict.
Cambridge, MA: Harvard University Press.
Jones, Leiponen, and Vasudeva 30
Shane, S., & Somaya, D. (2007). The effects of
patent litigation on university licensing efforts.
Journal of Economic Behavior &
Organization, 63(4), 739-755.
https://doi.org/10.1016/j.jebo.2006.05.012
Simcoe, T. S., Graham, S. J. H., & Feldman, M. P.
(2009). Competing on standards?
Entrepreneurship, intellectual property, and
platform technologies. Journal of Economics
and Management Strategy, 18(3): 775-816.
https://doi.org/10.1111/j.1530-
9134.2009.00229.x
Simcoe, T. (2012). Standard setting committees:
Consensus governance for shared technology
platforms. American Economic Review, 102(1):
305-336. https://doi.org/10.1257/aer.102.1.305
Somaya, D. (2003). Strategic determinants of
decisions not to settle patent litigation.
Strategic Management Journal, 24(1): 17-38.
https://doi.org/10.1002/smj.281
Stuart, E. A. (2010). Matching methods for causal
inference: A review and a look
forward. Statistical Science, 25(1):1-21.
https://doi.org/10.1214/09-STS313
Stuart, T. E., Podolny, J.M. (1996). Local search
and the evolution of technological
capabilities. Strategic Management
Journal, 17(S1):21-38.
https://doi.org/10.1002/smj.4250171004
Sytch, M., & Tatarynowicz, A. (2014). Friends and
foes: The dynamics of dual social structures.
Academy of Management Journal, 57(2): 585-
613. https://doi.org/10.5465/amj.2011.0979
Teece, D.J. (2018). Profiting from innovation in
the digital economy: Enabling technologies,
standards, and licensing models in the wireless
world. Research Policy, 47(8): 1367-1387.
https://doi.org/10.1016/j.respol.2017.01.015
Topkis, D. M. (1998). Supermodularity and
Complementarity. Princeton University Press:
Princeton New Jersey.
Vasudeva, G., Alexander, E.A., & Jones, S.L. (2015).
Institutional logics and interorganizational
learning in technological arenas: Evidence
from standard-setting organizations in the
mobile handset industry. Organization
Science, 26(3):830-846.
https://doi.org/10.1287/orsc.2014.0940
Vasudeva, G., & Anand, J. (2011). Unpacking
absorptive capacity: A study of knowledge
utilization from alliance portfolios. Academy of
Management Journal, 54(3), 611-623.
https://doi.org/10.5465/amj.2011.61968108
Vasudeva, G., Leiponen, A., & Jones, S. L. (2020).
Dear Enemy: The Dynamics of Conflict and
Cooperation in Open Innovation
Ecosystems. Strategic Management
Review, 1(2), 355-379.
http://dx.doi.org/10.1561/111.00000008
Vasudeva, G., & Teegen, H. (2011). Why do private
firms invest in public goods?.In Cross-Sector
Leadership for the Green Economy, Marcus, A.,
Shrivastava, P., Sharma, S., & Pogutz, S.
(eds). Palgrave Macmillan: New York; 263-276.
Ziedonis, R.H. (2004). Don't fence me in:
Fragmented markets for technology and the
patent acquisition strategies of
firms. Management Science, 50(6), 804-820.
https://doi.org/10.1287/mnsc.1040.0208
Evolution of cooperation in the face of conflict 31
ONLINE APPENDIX
A.1 Payoff structures in innovation ecosystems with intellectual property rights
Firms in innovation ecosystems create value through individual or cooperative contributions.
We illustrate firms’ decisions to contribute through two prisoner’s dilemma (PD) models. The
first is asymmetric and models the payoff matrix for cooperation or defection by an intellectual
property (IP) licensor and an IP licensee. The second is symmetric and models the payoff matrix
for two IP licensees. In the context of standards development, mutual cooperation signifies that
the two firms coauthor contributions, cooperation by one firm and defection by the other
signifies the cooperating firm authors contributions alone, and mutual defection signifies neither
firm authors contributions.
Contributions within an innovation ecosystem generates three types of outcomes: public
benefits, club benefits and private benefits. When technical development is open, public benefits
are available to everyone in the industry independent of their contributions. In a telecom
standards development organization (SDO), these might include technical features of the
communication system that are described in the technical specifications available to everyone.
Even though all firms can access public benefits, some firms must contribute to generate them.
If they do not contribute, then is reduced by percent such that the public benefit becomes
and . Private benefits are available to firms who make contributions to the
open development. Private benefits consist of opportunities to shape technical development in a
direction that is advantageous to the contributing firm. In SDOs, these might include
opportunities for first-mover advantages in commercializing the standardized products and the
creation of features particularly demanded in markets dominated by the firm (Bar & Leiponen,
2014). Club benefits K are available to firms who work jointly on technical development. They
arise for firms who develop standards that are complementary to the joint offerings of the firms.
In SDOs, this may occur if one firm’s technology is enhanced by a partner’s technology; joint
standards development then enhances the benefit to both. There is also a cost to standards
development when done individually. This cost is reduced by percent when joint development
occurs such that the cost is and .
Jones, Leiponen, and Vasudeva 32
A.2 IP licensor and licensee model
In the presence of IP rights, IP holders also receive a private benefit from licensing fee , and
licensees incur the cost of fee . As technologies evolve, the value of IP can be reduced if licensees
steer the trajectory of the technology away from the underlying IP used in the open standard. To
maintain its value, an IP licensor contributes to the standard to ensure it continues on a
trajectory that benefits the firm. Thus, if an IP licensor contributes alone or jointly, it maintains
a licensing fee ; if it does not contribute, its licensing fee falls to , where is the
percentage decline in licensing benefit and .
IP licensor
Cooperate
Defect
IP licensee
Coop.
Defect
The model is a prisoner’s dilemma (a) when the licensor’s and licensee’s defection payoff is
greater than its cooperation payoff in both the cases when the other firm cooperates and when
the other firm defects and (b) when, for both firms, the payoff for mutual cooperation is greater
than the payoff for mutual defection. If the licensor cooperates, the licensee’s defection payoff is
greater than its cooperation payoff when the joint development cost is greater than the combined
club and private benefits:
(1)
If the licensor defects, the licensee’s defection payoff is greater than its cooperation payoff when
the solo development cost is greater than the private benefits and the public benefits loss:
(2)
Equation 1 and 2 indicate that the licensee’s decision to defect or cooperate does not depend on
licensing fee . The licensee’s payoff for mutual cooperation is greater than its payoff for mutual
Evolution of cooperation in the face of conflict 33
defection when the club and private benefits and the public benefits loss are greater than the
cost of joint development and the licensing fee reduction:
. (3)
This game is a prisoner’s dilemma if the public benefits loss, , is greater than the licensing fee
reduction, , due to the constraint in equation 1.
After litigation, firms may find that club benefits are higher and the licensing fee may
increase. Since is not in equation 1 or 2, the increased licensing fee does not affect the licensee’s
choice to cooperate. In contrast, an increase in club benefits, , would increase the licensee’s
likelihood of cooperating if the licensor cooperates.
The licensor’s payoff is slightly different than the licensee’s payoffs. If the licensee
cooperates, the licensor’s defection payoff is greater than its cooperation payoff when the joint
development cost is greater than the combined club and private benefits and the licensing
revenue loss:
(4)
Equation 4 mirrors equation 1 except it includes the licensing revenue loss. If the licensee defects,
the licensor’s defection payoff is greater than its cooperation payoff when the solo development
cost is greater than the private benefits, the public benefits loss, and the licensing revenue loss:
(5)
The licensor’s payoff for mutual cooperation is greater than its payoff for mutual defection when
(6)
Equations 4 and 5 mirror equations 1 and 2, respectively, but they also include the licensing
revenue loss, . Equation 6 mirrors equation 3 except that is on the left hand side (mutual
cooperation). Thus, equations 4−6 indicate that the licensor is more motivated than the licensee
to cooperate to protect its licensing revenues.
Jones, Leiponen, and Vasudeva 34
If, as a result of litigation, club benefits and licensing revenues increase, then the licensor’s
likelihood of cooperating also grows. The licensor is discouraged from defecting because it would
lead to a larger licensing revenue loss, . The licensor also receives increased club benefits, , if
the licensee cooperates. Overall, the licensor benefits more than the licensee from cooperation,
but the cooperation payoffs increase for both after litigation.
A.3 IP licensees model
In a scenario with two licensees, we assume there is a third firm that is an IP licensor who
participates in standards development. However, IP licensees can individually or collectively act
to challenge the existing technological trajectory and attempt to steer the standards away from
the IP holder’s technology. If a licensee works alone, it can reduce the licensing fee by such that
licensing fee is . If licensees work cooperatively, they can jointly create club benefits
and further reduce the licensing fee by . Because we assume an IP holder is active in standards
development, is not reduced by mutual defection.
IP licensee
Cooperate
Defect
IP licensee
Coop.
Defect
If one licensee cooperates, the other’s defection payoff is greater than its cooperation payoff
when the joint development cost is greater than the club and private benefits and the licensing
fee reduction:
(7)
If one licensee defects, the other’s defection payoff is greater than its cooperation payoff when
the solo development cost is greater than the private benefits and the licensing fee reduction:
(8)
Evolution of cooperation in the face of conflict 35
The licensee’s payoff for mutual cooperation is greater than its payoff for mutual defection when
(9)
Equation 9 can be rewritten as , which, together with equation 7,
shows that the model is a prisoner’s dilemma when the club and private benefits and the
licensing fee reduction is less than the joint development cost but more than the joint
development cost less the licensing fee reduction.
After litigation, the licensing fee may increase, but the club benefits between the licensees,
, are not apt to increase because they are not based on the underlying technologies of the IP
holder. Since increases, the potential licensing fee reduction, , increases, which would
increase the likelihood that the licensees cooperate.
