MULTIMARKET CONTACT AND RIVALRY OVER ...mason.gmu.edu/~hleeb/SMJ_17_MMC_Patent_Litigation.pdfIndeed, Jayachandran etal. (1999: p. 58) state that “regardless of multimarket contact,

Strategic Management JournalStrat. Mgmt. J., 38: 2508–2531 (2017)

Published online EarlyView 25 July 2017 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2676Received 9 September 2015; Final revision received 31 January 2017

Multimarket Contact and Rivalry overKnowledge-based Resources

Matt Theeke* and Hun LeeSchool of Business, George Mason University, Fairfax, Virginia

Research summary: Research shows that multimarket contact (MMC) reduces rivalry involvingdownstream activities. Yet, studies showing that MMC can increase the threat of imitation suggesta need to better understand how MMC affects upstream rivalry over knowledge-based resources.In this study, we argue that MMC increases rivalry over knowledge-based resources since thedeterrent threat of retaliation that typically leads to mutual forbearance in downstream activitieswill not be sufficient to restrain firms from protecting their knowledge from imitation in upstreamactivities. In support of these arguments we find that MMC increases the likelihood that a firminitiates patent litigation against a rival. This study suggests the relationship between MMC andrivalry may depend on the competitive domain and the type of resources over which firms arecompeting.

Managerial Summary: How does market overlap or MMC affect rivalry between two competi-tors? Prior studies have largely found that an increase in market overlap decreases rivalry in lessknowledge-intensive context because of the deterrent threat of retaliation. However, in this paper,we argue that an increase in market overlap may not reduce rivalry in more knowledge-intensivecontext because of heterogeneity in capabilities to protect knowledge. We find that a firm is morelikely to initiate patent litigation against a rival as market overlap increases. Our findings suggestthat the incentive to protect value across multiple product markets may surpass the motivation tocooperate with rivals and that managers should have a more nuanced view of how market overlapwith competitors affects rivalry in more knowledge-intensive contexts. Copyright © 2017 JohnWiley & Sons, Ltd.

Knowledge-based resources play an important rolefor firms in creating a competitive advantage (Grant,1996; Ndofor, Sirmon, & He, 2011). To achievecompetitive advantage from their knowledge, firmsneed to both create valuable innovations (McEvily& Chakravarthy, 2002; Nelson & Winter, 1982) andprotect the value of their knowledge when compet-ing against rival firms (Giarratana & Mariani, 2014;Liebeskind, 1996; Teece, 1986). This competitionoften occurs across multiple markets, increasing

Keywords: multimarket contact; rivalry; knowledge-basedresources; patent litigation; competitive dynamics*Correspondence to: Matt Theeke, School of Business, GeorgeMason University, 4400 University Drive, MSN 5F5, Fairfax, VA22030. E-mail: [email protected]

Copyright © 2017 John Wiley & Sons, Ltd.

multimarket contact (MMC) with rival firms (Fuen-telsaz & Gómez, 2006; Li & Greenwood, 2004).

Prior research on MMC has largely shown thatfirms can benefit from less intense rivalry as marketoverlap increases between two competitors—themutual forbearance hypothesis (Edwards, 1955;Jayachandran, Gimeno, & Varadarajan, 1999; Korn& Baum, 1999; Yu & Cannella, 2013). In par-ticular, prior studies emphasize the benefits ofreduced rivalry associated with MMC involvingdownstream activities such as new product andpricing actions (Kang, Bayus, & Balasubramanian,2010; Young, Smith, Grimm, & Simon, 2000) andmarket entry and exit decisions (Haveman & Non-nemaker, 2000; Korn & Baum, 1999). However,

Multimarket Contact and Rivalry 2509

other research suggests that competition can alsoencompass upstream activities related to firms’resources (Barney, 1986; Bergen & Peteraf, 2002;Capron & Chatain, 2008; Markman, Gianiodis, &Buchholtz, 2009). Upstream rivalry often entailshigher uncertainty than downstream rivalry, sinceactivities such as R&D rely on extensive searchand exploration (March, 1991; Nelson & Win-ter, 1982). Moreover, recent research suggests thatMMC heightens the threat of imitation that firmsexperience from rivals (Lieberman & Asaba, 2006),and studies show that MMC increases patenting(Greve & Mitsuhashi, 2004) and the extent to whichrival firms build on one another’s knowledge (Cas-sidy & Loree, 2001; Scott, 2001).

These studies suggest the possibility that firmsface a tension on the one hand to leverage knowl-edge across multiple markets but on the otherhand to protect the value of their knowledge frommultimarket rivals. In other words, a firm’s abilityto fully leverage the value from knowledge-basedresources may result in increasing MMC withrival firms, which can expose knowledge-basedresources to an increased threat of imitation. Hence,MMC encourages firms to be less rivalrous withdownstream activities but more rivalrous withupstream activities to protect knowledge-basedresources. Extant research on MMC has notfully addressed this tension or examined howMMC affects the decision to protect a firm’sknowledge-based resources from rivals.

With the goal of exploring this tension, this cur-rent study develops theory about how MMC affectsrivalry over knowledge-based resources involvingpatent litigation. Because MMC increases the threatof imitation (Lieberman & Asaba, 2006) at thesame time that heterogeneity exists in rivals’ abil-ity to protect their knowledge (Clarkson & Toh,2010; Teece, 1986), we argue that the motiva-tion to protect knowledge will tend to outweighthe deterrent effect of retaliation by the rival.Accordingly, we propose that the mutual forbear-ance effect arising from the threat of retaliationby a rival firm will not occur as MMC pushesfirms to protect knowledge-based resources. Toincrease confidence in our causal mechanisms, wedevelop contingencies related to: (a) a firm’s moti-vation to protect knowledge and (b) the deterrenteffect of retaliation by a rival. Hence, we arguethat the technological quality of the firm’s knowl-edge and the rival’s ability to protect its knowl-edge will respectively exacerbate and attenuate the

positive relationship between MMC and knowledgeprotection.

This study adds new insights about the relation-ship between MMC and mutual forbearance byshowing that this relationship may depend on thecompetitive domain and the nature of the resourcesover which firms are competing. In doing so, weadd to prior studies that have begun to unpack theboundary conditions affecting the MMC–mutualforbearance relationship (Yu & Cannella, 2013) byexamining factors such as market entry character-istics (Fan, 2010), the observability of strategies(Greve, 2008), and international market charac-teristics (Yu, Subramaniam, & Cannella, 2009).While we do not question the traditional mutualforbearance hypothesis found in downstreamactivities and in less knowledge-intensive contexts,we suggest a more nuanced relationship depend-ing on the domain. More specifically, we arguethat firms may become more protective of theirknowledge-based resources in upstream activitiesand in more knowledge-based contexts and thatMMC may lead to increasing rivalrous behavior.

This study also contributes to strategy researchon the defensive strategies that firms use to pro-tect proprietary technological assets. Studies havemade important strides advancing understandingof organizational and technological antecedents ofknowledge protection efforts (e.g., Hall & Ziedo-nis, 2007; Lanjouw & Schankerman, 2001; Poli-doro & Toh, 2011) and about the deterrent rolethat a firm’s reputation for protecting its resourcesplays in mitigating the leakage of knowledge andthe loss of employees (Agarwal, Ganco, & Ziedo-nis, 2009; Ganco, Ziedonis, & Agarwal, 2015). Weshow that MMC between rival firms may be anotherimportant determinant of a firm’s decision to pro-tect its knowledge-based resources. Furthermore,we also contribute to the resource base theory andcompetitive dynamics literature by developing the-ory about how downstream market contact betweenrivals shapes firms’ subsequent upstream competi-tive interactions in factor markets. We elaborate onthese contributions in the discussion section.

Theory and Hypotheses

Multimarket contact (MMC), which is essentiallythe extent of market overlap between two firms,has been shown to affect the intensity of rivalry

Copyright © 2017 John Wiley & Sons, Ltd. Strat. Mgmt. J., 38: 2508–2531 (2017)DOI: 10.1002/smj

2510 M. Theeke and H. Lee

or multimarket competition.1 The basic notionof MMC centers on Edwards’ (1955) mutualforbearance hypothesis. That is, the intensityof rivalry decreases between two firms as thenumber of markets that they compete or overlapin increases. Mutual forbearance occurs becausefamiliarity and deterrence increase as two firmscompete across more common markets. Familiarityenhances understanding of each other’s capabilitiesand actions while deterrence enhances the threat ofretaliation across multiple markets. Consequently,MMC encourages firms to tacitly cooperate witheach other for mutual benefit by deterring themfrom engaging each other aggressively.

By and large, prior studies have found support forthe mutual forbearance hypothesis—firms enjoyhigher performance and lower rivalry as MMCincreases (Baum & Korn, 1999; Jayachandran et al.,1999; Yu & Cannella, 2013). For example, numer-ous studies have shown a positive relationshipbetween MMC and performance such as mar-ket share stability (Heggestad & Rhoades, 1978),cost–price margins (Feinberg, 1985), and yields(Gimeno & Woo, 1999). Similarly, other stud-ies have shown that an increase in MMC reducesaggressive competitive behavior such as new prod-uct and pricing actions (Kang et al., 2010; Younget al., 2000) and market entry and exit (Fuentelsaz& Gómez, 2006; Haveman & Nonnemaker, 2000;Korn & Baum, 1999). While much of these stud-ies showed a linear relationship between MMCand reducing rivalry, a few studies that focusedon market entry demonstrated a curvilinear (orinverted U-shaped) relationship. That is, entry orrivalry increased at lower levels of MMC and thendecreased at higher levels of MMC. In total, thesestudies provide empirical evidence for the benefitsof MMC to firms in reducing rivalry.

Multimarket Contact and Rivalry overKnowledge-Based Resources

As mentioned previously, mutual forbearance, orrivalry restraint, results when competition acrossmore common markets increases the familiarity

1 Karnani and Wernerfelt (1985: p. 87) define multimarketcompetition as “a situation where firms compete against eachother simultaneously in several markets.” Yu and Cannella (2013:p. 76) note that “MMC simply indicates that two firms meet eachother in more than one market, while multimarket competitionrefers to both the existence of MMC and how it affects theinterfirm engagement, or rivalry.”

of rivals’ capabilities and actions and the deter-rent threat of retaliation across multiple markets(Edwards, 1955; Jayachandran et al., 1999). Hence,mutual forbearance occurs when multimarket com-petitors act less aggressively in anticipation thatrivals have the ability to counterattack. It is impor-tant to point out that that the mutual forbearancehypothesis implicitly assumes that rivals have sim-ilar action-response capabilities, since this is whatultimately gives rise to the threat of retaliation.Indeed, Jayachandran et al. (1999: p. 58) state that“regardless of multimarket contact, if one firm hasa resource advantage over its rival, it may not bemotivated to forbear from aggressive competitionbecause of the perception that it can outmaneuverthe competition, which cannot match its resources.”Similarly, Makadok (2010) proposes that the extentof competitive advantage determines whether a firmrefrains or engages a rival and demonstrates thatfirms with a resource disadvantage or small resourceadvantage prefer restrained rivalry, while firms witha significant resource advantage prefer aggressiverivalry. Hence, we argue below that MMC willnot restrain rivalry over knowledge-based resourcesbecause the incentives to protect knowledge willencourage firms with resource advantages to engagerivals.

In knowledge-intensive contexts, the ability todevelop heterogeneous resources is a particularlyimportant determinant of competitive success(DeCarolis & Deeds, 1999; Grant, 1996; Nelson& Winter, 1982) and studies show that imitationposes a significant threat to the value of hetero-geneous knowledge-based resources (Peteraf,1993). MMC can increase familiarity of a focalfirm’s heterogeneous resources leading to a greaterimitation threat from rivals, which increases a focalfirm’s incentives to protect its knowledge. Yet,at the same time resource heterogeneity suggestsdifferences in firms’ abilities to take actions toprotect knowledge-based resources, which meansthat rivals may or may not be able to retaliateagainst a firm’s efforts to protect its knowledge.Consequently, when firms compete across multiplemarkets, the deterrent threat of retaliation that typ-ically leads to mutual forbearance in downstreamrivalry will not be sufficient to restrain firms fromprotecting their knowledge-based resources fromimitation in upstream rivalry.

An evolving stream of research suggests thatMMC can increase rivalry over knowledge-basedresources. Imitation and knowledge spillovers



can undermine the competitive value ofknowledge-based resources and can increasedue to competitive proximity (Giarratana & Mari-ani, 2014; Lippman & Rumelt, 1982). Liebermanand Asaba (2006: p. 377) note that MMC “increasesthe likelihood of rivalry-based imitation, since itexpands the domains where imitation can occurand raises the probability that firms respond toeach other in kind.” In line with these arguments,Anand, Mesquita, and Vassolo (2009) show a pos-itive relationship between MMC and explorativemarket entry as measured by R&D investmentsin a particular market segment. They argue thatknowledge-based resources such as R&D invest-ments encourage mimetic behavior to managecompetitive uncertainty; hence, MMC leads toincreased knowledge-based rivalry. Likewise, ina study of the shipbuilding industry, Greve andMitsuhashi (2004: p. 17) argue and find support thatgreater MMC will increase incentives to innovateand encourage firms “to be more active in taking outpatents in that market, as they expect higher profitsfrom each patent.” Besides patenting activities,two studies of the semiconductor and chemicalindustries indicate that MMC increases firms’tendency to engage in the citation of other firms’patents (Cassidy & Loree, 2001; Scott, 2001).On the whole, these studies demonstrate that inknowledge-intensive contexts, MMC increases thethreat of imitation by rivals. These studies divergefrom the typical findings that MMC reduces rivalryand underscore the need to better understand howMMC pushes firms to engage in knowledge-basedrivalry to protect their knowledge-based resourcesfrom imitation.

A firm’s ability to profit from knowledge-basedresources depends both on the value of its knowl-edge and on whether the firm has the ability toprevent imitation by rivals (Teece, 1986). Researchshows that firms can use different mechanismsto protect their knowledge-based resources, suchas secrecy, lead time, complementary assets, andpatents (James, Leiblein, & Lu, 2013; Levin,Klevorick, Nelson, & Winter, 1987). To protecttheir knowledge, strategy research also recognizesthat firms may have to take competitive actionsagainst rivals by initiating patent litigation (Pisano,2006; Polidoro & Toh, 2011). Studies show thatheterogeneity exists in firms’ ability to protecttheir knowledge-based resources. For example,since obtaining a patent presupposes that a firm hasestablished the novelty of its invention (Dam, 1994;

Scotchmer & Green, 1990), a firm in possession ofa patent often has a unique ability to initiate patentlitigation.

The heterogeneity in rival firms’ ability toprotect their knowledge-based resources createsasymmetry in their ability to initiate and respond torivals’ actions. As a result, the threat of retaliationby a rival that leads to mutual forbearance becomesless likely. For example, Liebeskind (1996: p. 94)proposes that “some firms are able to protect theirknowledge from expropriation or imitation moreeffectively than other firms.” Likewise, the role ofcompetitive asymmetry in firms’ effort to protectknowledge-based resources is illustrated by anexcerpt from an article by a UCLA expert on intel-lectual property law: “Every time a patent holderthinks about filing a patent lawsuit, the first questionto ask is whether the intended target will be able toretaliate”(Lichtman, 2012: para. 2). This asymme-try to initiate and respond to competitive actionsmay enable firms that possess a superior ability toprotect knowledge-based resources as a source ofRicardian action (Grimm, Lee, & Smith, 2006).As Capron and Chatain (2008) argue, firms thatpossess property-based resources, such as exclusiverights to a technology, are more likely to initiateactions against rivals since rivals do not possessthose resources used to counterattack them. Con-sequently, unlike other types of downstream com-petitive actions, such as pricing moves and productpositioning where rivalry is restrained preciselybecause of the imminent threat of retaliation fromrivals with similar action–response capabilities, thedeterrent effect of retaliation by a rival is dampenedas a result of the greater heterogeneity that existsin rivals’ ability to initiate and respond to actionsaimed at protecting knowledge from imitation.

Differences in action–response capabilitiesbetween downstream and upstream actions areillustrated by prior research. For instance, researchshows a 28% overall likelihood of retaliation todownstream competitive moves (i.e., pricing, route,ticketing, collaboration, service, and cargo-related)in the airline industry (Marcel, Barr, & Duhaime,2011: p. 127). Studies also indicate that firms willrespond to a rival’s pricing actions at a higher ratethan non-pricing actions (Montgomery, Moore, &Urbany, 2005; Venkataraman, Chen, & MacMillan,1997). Specifically, 36% of managers surveyed inanother study indicated that they would add pricepromotions in response to competitors’ price pro-motions (Steenkamp, Nijs, Hanssens, & Dekimpe,



2005: p. 47). Other research shows a 93% competi-tive response rate to new product introductions byincumbents in the consumer and industrial goodsindustries (Kuester, Homburg, & Robertson, 1999:p. 101). The relatively high retaliatory responses topricing and new products compare to an averageretaliation rate of approximately 10% for patentlitigation countersuits across the computer andresearch medicine industries (Somaya, 2003:pp. 25–26).2 Given this disparity, the deterrentthreat of MMC may be more pronounced involvingdownstream pricing and product actions, leadingto the typical mutual forbearance. Conversely, thedeterrent threat of MMC may be weaker involvingupstream efforts to protect knowledge using patentlitigation, leading to less restraint.

In sum, we argue that increasing MMC pushesfirms to engage in knowledge-based rivalryagainst competitors to protect the value of theirknowledge-based resources. That is, a firm’s abilityto capture value from knowledge-based resourcesresults in increasing MMC with rival firms, whichraises familiarity and can expose knowledge-basedresources to an increased threat of imitation. Moreimportantly, when firms compete across multiplemarkets, the deterrent threat of retaliation that typ-ically leads to mutual forbearance in downstreamrivalry will not be sufficient to restrain firms fromprotecting their knowledge-based resources fromimitation in upstream rivalry. Hence, it follows that:

Hypothesis 1 (H1): There is a positive rela-tionship between multimarket contact and thelikelihood that the focal firm will protect itsknowledge-based resources by initiating a patentlawsuit against a rival firm.

To increase confidence in our argument thatincreasing MMC leads firms to protect the valueof their knowledge-based resources, because theincentives to protect knowledge-based resourcesamid multimarket competition will tend to out-weigh the threat of retaliation by a rival firm, wedevelop two contingencies that build on these mech-anisms. In H2, our first contingency focuses on thefocal firm’s incentives to protect knowledge-based

2 Somaya’s (2003: p. 25) 8% countersuit rate for computer firmsand 1.5% for medical firms result in a 6% weighted average.The bias corrected rates (footnote 10, p. 26) of 14% and 3%respectively result in a 10% weighted average.

resources amid multimarket competition by argu-ing that the technological quality of the focal firm’sknowledge should exacerbate the positive relation-ship between MMC and knowledge protection. InH3, our second contingency focuses on the threatof retaliation by a rival firm by arguing that the rivalfirm’s ability to protect its knowledge should atten-uate the positive relationship between MMC andknowledge protection.

Contingent Effect of the Technological Qualityof the Focal Firm’s Knowledge

In H1, we argue that MMC increases the likeli-hood that a firm will protect its knowledge-basedresources because of the increased threat of imi-tation that it faces from multimarket competitors.If MMC increases a firm’s motivation to protectits resources from imitation by rivals, then factorsthat further increase a firm’s motivation to protectits knowledge-based resources should exacerbatethis positive effect. Prior research shows that thetechnological quality of a firm’s knowledge-basedresources enhances the value of a firm’s innova-tions (Trajtenberg, 1990) and increases its share-holder value (Hall, Jaffe, & Trajtenberg, 2005).We argue that the technological quality of thefocal firm’s knowledge will positively moderate therelationship between MMC and the focal firm’sdecision to protect its knowledge-based resourcesfrom a rival.

When deciding whether or not to respond tothe increased threat of imitation that results fromMMC, the focal firm will take into account thevalue of its knowledge-based resources. If the tech-nological quality of the focal firm’s knowledge islow, a focal firm will be less motivated to pro-tect its knowledge from the increased threat ofimitation that results from multimarket competi-tion. Conversely, if a focal firm possesses higherquality knowledge-based resources, its motivationwill be even greater to protect the value of theseresources from imitation by a multimarket rival.Therefore:

Hypothesis 2 (H2): The technological qualityof the focal firm’s knowledge-based resourceswill exacerbate the positive relationship betweenMMC and the likelihood that the focal firmwill protect its knowledge-based resources byinitiating a patent lawsuit against a rival firm.



Contingent Effect of Rival Firm’s Abilityto Protect Its Knowledge

If as we argued above in H1, MMC creates incen-tives to protect knowledge-based resources that tendto outweigh the deterrent threat of retaliation, thenfactors that increase a rival’s ability to retaliate byprotecting its own knowledge should also mod-erate the relationship between MMC and knowl-edge protection. As mentioned previously, researchshows that competitive asymmetry between rivalscan undermine deterrence by making firms morelikely to initiate actions against rivals who lack theresources used to attack them (Capron & Chatain,2008; Grimm et al., 2006). If indeed MMC tendsto increase rivalry over knowledge-based resources,in part, because of asymmetry in rivals’ ability toprotect their knowledge, then factors that tilt thebalance by making rivals more or less able to takeactions to protect their knowledge should moder-ate the effect that MMC exerts on knowledge pro-tection. A recent article from the Washington Postillustrates how competitors’ ability to retaliate withpatent lawsuits has limited how aggressively Nokiauses patent litigation against its rivals:

Nokia lays claim to an estimated 30,000 util-ity patents and 8,500 design patents, accord-ing to Microsoft. Last year, it filed lawsuitsagainst device manufacturer HTC in both Ger-many and the United States, accusing HTC ofinfringing on dozens of Nokia patents. Yet,Nokia has been deterred from asserting itspatents too aggressively by the risk of retal-iation by competitors. (Fung, 2013)

When deciding whether or not to take actionsto protect knowledge due to the increased threatof imitation that results from MMC, a focal firmwill consider the possible responses associated withretaliation by the rival firm. If a focal firm faces amultimarket competitor that has a lower ability toprotect its knowledge, this will reduce the likelihoodthat the rival firm will initiate a counterattackagainst the focal firm in one or more of the marketsin which the firms overlap, thus making the focalfirm even more likely to protect its knowledge.Conversely, if a focal firm faces a multimarketcompetitor that has a greater ability to protect itsknowledge, this will increase the likelihood thatthe rival firm will initiate a counterattack againstthe focal firm in one or more of the markets in

which the firms overlap, thus making the focalfirm less inclined to protect its knowledge becauseof the expected costs associated with retaliation.Consequently:

Hypothesis 3 (H3): The rival firm’s ability to pro-tect its knowledge-based resources will attenu-ate the positive relationship between MMC andthe likelihood that the focal firm will protect itsknowledge-based resources by initiating a patentlawsuit against a rival firm.

Data and Methods

To examine the influence that MMC has on thelikelihood that a firm will protect its knowledgefrom a rival, it is important to identify an empir-ical context where we can observe both interfirmrivalry across different markets and firms’ actionsto protect knowledge-based resources. The medicaldevices industry is a suitable context to test our the-ory since competition and efforts to protect knowl-edge are observable for a broad set of companies.Firms in the medical devices industry frequently usepatents to protect their knowledge from imitation byrivals (Levin et al., 1987), and research shows thathealth and drug firms use patent litigation to protecttheir innovations (Lanjouw & Schankerman, 2001;Polidoro & Toh, 2011). In addition, the medicaldevices context has been used as the empirical set-ting for other strategy studies involving competitionamong innovation-intensive firms (e.g., Mitchell,1989; Wu, 2013). Thus, the medical devices contextis appropriate to test our theory since we are able toobserve both interfirm rivalry and a firm’s efforts toprotect its knowledge-based resources.

Research involving MMC suggests that thefirm–competitor dyad is the appropriate level ofanalysis to examine interfirm rivalry (e.g., Baum& Korn, 1996; Chen, 1996; Yu & Cannella, 2007).Importantly, the medical devices industry offersa fine-grained means of assessing the degree towhich pairs of rival firms compete across differentproduct categories since these firms have to applyfor Food and Drug Administration (FDA) approvalfor medical devices introduced in different productcategories. Thus, in addition to allowing us toobserve knowledge protection involving rival firms,this setting is also appropriate because it allows usto observe MMC between pairs of rival firms.



We study firms’ actions taken to protect knowl-edge from rivals by examining the patent lawsuitsthat were filed by medical devices firms. To helpensure that we include a representative set of firmsthat are active in the medical devices industry, wedevelop our sample based on standard industrialclassification (SIC) codes and consider firms inCOMPUSTAT in SIC code 384 (Chatterji & Fab-rizio, 2014). After using the approach describedabove to identify firms in the medical devices indus-try at hazard of filing a patent lawsuit, we supple-ment the firm-level product data gathered from theFDA 510K and PMA databases with patent datafrom the Harvard Patent Database (Lai, D’Amour,Yu, Sun, & Fleming, 2011). The patent and prod-uct measures are developed using the year in whichthe firm applied for a patent that was subsequentlygranted by the U.S. Patent and Trademark Officeor the year in which it filed for FDA approval ofa medical device that was approved by the FDA.Our final sample includes 164 firms from the med-ical devices industry that held at least one patent.After identifying the firms at hazard of initiatingpatent litigation, we follow the approach used inprior literature and restricted our analysis to includefirm–rival dyads where the two firms meet in atleast one common product market (Boeker, Good-stein, Stephan, & Murmann, 1997). To help ensurethat both firms are actively developing products inthe markets where they overlap, we consider prod-uct code overlap during a 5-year window basedon the year in which the firms applied for FDAapproval of a new device or supplemental changesto an existing device. We next use the Lex MachinaIP Litigation Database to identify all patent lawsuitswhere each of these firms was listed as a plaintiff ordefendant to determine whether a patent lawsuit wasfiled by the focal firm against a specific rival firm.Patent litigation data are available as of 2000. Weuse data on firms’ patent applications made through2006 that were granted by the end of 2010. Accord-ingly, our analysis period considers the effect ofMMC at years 2000 through 2006 on patent lawsuitsin observation years 2001–2007.

Dependent Variable

Consistent with prior literature, we use patent litiga-tion to study the protection of knowledge (Clarkson& Toh, 2010; Polidoro & Toh, 2011). One signifi-cant advantage of measuring knowledge protectionusing patent litigation, instead of other mechanisms

(e.g., complementary assets and secrecy) in a dyadicstudy like this one, is the ability to observe a firm’sefforts to protect knowledge from a specific rival(Chen & Miller, 2012; Markman, Espina, & Phan,2004). Our dichotomous dependent variable is setto “1” if the focal firm filed a patent lawsuit againsta specific rival firm in observation year t and to “0”otherwise. Other variables correspond to year t− 1.

Independent Variables

Multimarket contact (H1) is assessed based on thenumber of FDA product categories in which thefirm and its rival both had products approved duringa 5-year window from the year of application.3

Alternate measures based on 4- and 6-year windowsproduced consistent results. Some prior studies haveused scaled measures of MMC (e.g., Baum &Korn, 1999; Gimeno & Woo, 1996; Young et al.,2000). To facilitate interpretation of the results, weuse market overlap. However, we ran models withscaled measures of MMC and found that the effectof MMC on knowledge protection is robust both formeasures of MMC that are scaled by the number ofproduct categories in which the focal firm competesand by the number of competitors in the categorieswhere the firms overlap.

The technological quality of the focal firm’sknowledge-based resources (H2). Prior literaturehas shown that a firm’s forward patent citationscapture the technological quality of its knowledge(Trajtenberg, 1990). The technological quality ofthe focal firm’s knowledge-based resources is mea-sured using the focal firm’s 5-year average numberof citations per patent, where the focal firm’s num-ber of citations received during the previous 5 yearswas divided by the number patents that receivedcitations during the previous 5 years (Ganco et al.,2015). We also checked robustness using a mea-sure of technological quality where the number ofcitations received during the previous 5 years wasdivided by the focal firm’s total cumulative numberof patents, instead of only those cited, and foundrobust results.

Rival firm’s ability to protect itsknowledge-based resources (H3) is measuredusing the rival firm’s yearly number of patentlawsuits as plaintiff scaled by the rival firm’s totalnumber of patent applications, since rival firms

3 Models using a measure based on approval year instead ofapplication year generated robust results.



that have initiated patent litigation have demon-strated the ability to protect their knowledge-basedresources. We also checked robustness using analternative measure of the rival firm’s ability toprotect its knowledge based on the rival firm’snumber of patents in the firm’s technologicalclasses, because research shows that a patentportfolio provides an arsenal which can deter firmsfrom initiating patent lawsuits (Somaya, 2012).Models using this alternate measure generatedrobust results.

Control Variables

The analysis controls for a number of differentsources of unobserved heterogeneity across firmsthat may influence both MMC and a firm’s decisionto protect its knowledge from a rival. Firm sizemay influence the degree to which firms operate inmultiple markets and larger firms are more likelyto initiate patent litigation (Hall & Ziedonis, 2007).We control for the size of the focal firm and the rivalfirm based on each firm’s number of employees.Since more profitable firms may be more likely toinitiate patent litigation and to engage in MMC, wealso control for the focal firm’s and the rival firm’syearly net income.

Strategy research shows that similarity incapabilities can affect market entry (Fuentelsaz& Gómez, 2006). Since resource similarity mayalso increase the likelihood of patent litigation, wecontrol for resource similarity based on the numberof overlapping primary technological classes whichthe focal firm and the rival firm have both usedfor their patents during the previous 5 years. Also,firms with more patents may be more likely tocompete across multiple markets, and researchconfirms that litigation risk depends on the size ofa firm’s patent portfolio (Lanjouw & Schankerman,2004). To account for this influence, we control forthe focal firm’s number of active patents (i.e., thoseapplied for in the last 20 years) in the technologicalclasses in which the rival firm has patented. Simi-larly, we also control for the rival firm’s number ofactive patents in the technological classes in whichthe focal firm has patented. The number of claimsmade in the firm’s patents can affect the propensityto litigate (Lanjouw & Schankerman, 2001). Sincefirms operating across multiple markets may alsomake more extensive claims in their patents, wecontrol for the average number of claims madein the focal firm’s and rival firm’s patents during

the last 5 years. Furthermore, a firm’s decisionto utilize a general versus niche product strategymay influence both its willingness to compete withrivals in multiple markets and its propensity toprotect its knowledge-based resources. To helpcapture this potential alternative explanation, weinclude control variables that capture the degree towhich the focal and the rival firm’s products areconcentrated or dispersed across different productcategories. Specifically, we use a Blau indexto assess the level of dispersion across productcategories during the previous 5 years.

Research also shows that the firm’s previouspatent lawsuits can influence the likelihood that itinitiates patent litigation (Polidoro & Toh, 2011). Toaccount for prior litigation, we include a control forthe focal firm’s yearly number of patent lawsuits asplaintiff scaled by its total patent applications. Wealso include controls for the yearly number of patentlawsuits where the focal firm and rival firm werelisted as defendant. Moreover, threats to the spe-cific knowledge domains in which the firms developproducts may affect both MMC and litigation. Inthe medical device industry, FDA product over-sight is divided among advisory committees that aregrouped by medical specialty (FDA, 2015). Sincethese advisory committees are organized around thespecialized medical knowledge needed to approvedifferent types of devices, adding dummies based onadvisory committees helps us control for the differ-ent knowledge domains in which each of the firmshave developed products. Accordingly, we includedummy variables for the advisory committees (e.g.,Cardiology, Immunology Devices, MicrobiologyDevices, Radiology, etc.) that have overseen theapproval of the focal firm’s products to account forsources of time–invariant unobserved heterogene-ity related to the knowledge domains that may affectMMC and litigation. Finally, we include year dum-mies to capture time specific factors that may influ-ence MMC and litigation.

Model Specification

Following the approach used in other recent strategystudies that have examined dichotomous outcomes,we use linear probability models to test the hypothe-ses (e.g., Conti, Gambardella, & Mariani, 2013).Linear probability models allow for a straightfor-ward interpretation of contingency hypotheses andproduce marginal effects that are consistent with



logistic regression (Angrist & Pischke, 2009). Wemean centered the independent variables in thereported models to reduce multi-collinearity thatcan occur in moderated regressions (Aiken & West,1991), and our reported models include firm-fixedeffects to help account for sources of time-invariantheterogeneity across firms. The use of firm-fixedeffects in conjunction with our lagged variables alsohelps to mitigate potential endogeneity concerns(Bettis, Gambardella, Helfat, & Mitchell, 2014).

Since using linear models to examine discreteoutcomes can yield biased and inefficient estimates(Greene, 2003), we checked robustness using logis-tic regression and rare events logistic regression(King & Zeng, 2001). These models produced fullyrobust results for the relationship between MMCand litigation, thereby alleviating such concerns.We also ran supplemental logistic regression mod-els to conduct split sample econometric tests ofmarginal effects at high and low levels of the con-tingency variables (Hoetker, 2007; Wiersema &Bowen, 2009), and these models also generatedrobust results for the contingencies.4

Results

Table 1 reports the descriptive statistics. In total,we identify 45 instances where a focal firm ini-tiated patent litigation against a rival firm in theobservation year. Although patent litigation is alow frequency event, it can have billions of dollarsof financial impact (Sherry & Teece, 2004), makingit consequential especially to knowledge-intensivefirms. All of the pair-wise correlations in Table 1fall below the threshold of 0.8, and the onlypair-wise correlations exceeding 0.7 involvedcontrol variables for patents and employees.Accordingly, we ran a separate model droppingthe control variable for patents, and the reportedresults remained fully robust. Moreover, to furtherinvestigate the possibility that multicollinearitymay be affecting the reported results, we also ranvariance inflation factors (VIF) for the variables inTable 1. The variance inflation factors (Avg. VIF:1.81; Max VIF: 3.57) were well below the criticallevel of 10 (Kennedy, 2003). Taken together, thesedifferent tests help to mitigate concerns aboutmulticollinearity.

4 The logit models used for the split sample econometric test ofmoderators excluded advisory class dummies.

Table 2 reports linear probability models of influ-ences on patent litigation. Starting with the controlsin model 2, we find that the level of technologicalresource similarity based on overlapping technolog-ical classes exerts a positive and significant effect onlitigation propensity. This evidence is in line withprior studies showing that technological similar-ity increases the probability of infringement litiga-tion (Lanjouw & Schankerman, 2001). Importantly,other control variables also offer insight into thestrategic considerations underlying a firm’s decisionto protect its knowledge-based resources. We findevidence that firms approach this decision strate-gically by considering attributes of the rival firm.Specifically, we find that the focal firm is morelikely to initiate patent litigation against a rivalfirm with higher quality technological resources andwhen a rival firm has been sued for patent infringe-ment. Moreover, we find that the focal firm is lesslikely to initiate a patent lawsuit against a rival firmwith greater product diversity. This evidence sug-gests that firms are particularly likely to initiatepatent litigation against rivals with more valuabletechnologies, more concentrated product offeringsand those with prior history as a defendant againstclaims of patent infringement.

After considering these other influences, we nextturn to our discussion of the main tests of ourtheoretical predictions. Our theoretical argumentslinking MMC and patent protection build on priorstudies showing that MMC increases knowledgespillovers and imitation by competitors (e.g., Cas-sidy & Loree, 2001; Lieberman & Asaba, 2006;Scott, 2001). However, to increase confidence inthis mechanism we conducted our own supplemen-tal analysis using the incidence rate ratio from anegative binomial model with the same set of con-trol variables reported in Table 2 to assess whetherMMC increases the extent to which a rival firmbuilds on the knowledge of the focal firm. As MMCincreases by one market between the focal firmand the rival firm, this analysis reveals that therate at which the rival firm cites the focal firm’spatents increases by more than 8%. In line with priorresearch and consistent with our theoretical argu-ments, this additional evidence from our empiricalcontext lends support to our argument that MMCincreases the extent to which the rival firm imitatesthe focal firm’s knowledge.

The second element underlying our argumentslinking MMC and patent protection involves the rel-atively lower risk of retaliation involving upstream



Tabl

e1

Des

crip

tive

Stat

isti

csan

dC

orre

lati

onM

atri

x

Var

iabl

eM

ean

S.D

.M

inM

ax1

23

45

67

89

1011

1213

1415

1617

18

1L

itiga

tion

0.00

70.

080

12

Mul

timar

ketc

onta

ct1.

711.

931

210.

163

Foca

lfirm

’ste

chno

logi

cal

qual

ityof

know

ledg

e(c

itatio

nspe

rpa

tent

)

11.1

07.

040

430.

020.

10

4R

ival

’sab

ility

topr

otec

tits

know

ledg

e(l

awsu

itsas

plai

ntif

f%

pate

nts)

0.01

0.03

01

−0.

01−

0.04

−0.

02

5Fo

calfi

rmsi

ze(t

hous

ands

ofem

ploy

ees)

5.40

8.69

0.00

437.8

0.04

0.25

0.19

−0.

04

6R

ival

firm

size

(tho

usan

dsof

empl

oyee

s)

5.40

8.69

0.00

437.8

0.04

0.25

0.01

−0.

08−

0.08

7Fo

calfi

rm’s

net

inco

me

($m

illio

ns)

169.

2755

2 .44

−3,

577

2,80

20.

030.

170.

14−

0.02

0.60

−0.

06

8R

ival

firm

’sne

tin

com

e($

mill

ions

)

169.

2755

2.44

−3,

577

2,80

20.

020.

170.

02−

0.05

−0.

060.

60−

0.04

9R

esou

rce

sim

ilari

ty(t

echn

olog

ical

clas

sov

erla

p)

2.45

2.70

031

0.10

0.44

0.08

−0.

060.

380.

380.

230.

23

10Fo

calfi

rm’s

pate

nts

inri

val’s

tech

nolo

gica

lcl

asse

s

141.

4037

2.99

03,

614

0.06

0.30

0.20

−0.

030.

73−

0.01

0 .56

−0.

010.

48

11R

ival

firm

’spa

tent

sin

foca

lfirm

’ste

chno

logi

cal

clas

ses

141.

4037

2.99

03,

614

0.05

0.30

0.05

−0.

06−

0.01

0.73

−0.

010.

560.

480.

05

12Fo

calfi

rm’s

aver

age

clai

ms

per

pate

nt21.4

48.

700

71.8

0.00

0.03

0.15

0.02

0.04

−0.

030.

070.

000.

090.

070.

01


2518 M. Theeke and H. LeeTa

ble

1co

ntin

ued V

aria

ble

Mea

nS.

D.

Min

Max

12

34

56

78

910

1112

1314

1516

1718

13R

ival

firm

’sav

erag

ecl

aim

spe

rpa

tent

21.4

48.

700

71.8

0.00

0.03

0.02

−0.

01−

0.03

0.04

0.00

0.07

0.09

0.01

0.07

−0.

01

14R

ival

firm

’ste

chol

ogic

alqu

ality

ofkn

owle

dge

(cita

tions

per

pate

nt)

11.1

07.

040

430.

040.

100.

12−

0.08

0.01

0.19

0.02

0.14

0.08

0.05

0.20

0.02

0.15

15Fo

calfi

rm’s

prod

uct

dive

rsity

0.68

0.30

00.

990.

020.

260.

08−

0.06

0.49

0.01

0.27

−0.

010.

260.

300.

04−

0.10

−0.

050.

02

16R

ival

firm

’spr

oduc

tdi

vers

ity

0.68

0.30

00.

990.

020.

260.

02−

0.10

0.01

0.49

−0.

010.

270.

260.

040.

30−

0.05

−0.

100.

080.

14

17Fo

calfi

rm’s

law

suits

aspl

aint

iff

asa

%of

pate

nts

0.01

0.03

01

0.00

−0.

04−

0.08

−0.

01−

0.08

−0.

04−

0 .05

−0.

02−

0.06

−0.

06−

0.03

−0.

010.

02−

0.02

−0.

10−

0.06

18Fo

calfi

rm’s

law

suits

asde

fend

ant

0.73

1.27

07

0.04

0.22

0.24

−0.

020.

69−

0.07

0.50

−0.

050.

300.

50−

0.01

0.06

−0.

020.

030.

39−

0.01

−0.

04

19R

ival

firm

’sla

wsu

itsas

defe

ndan

t

0.73

1.27

07

0.06

0.22

0.03

−0.

04−

0.07

0.69

−0.

050.

500.

30−

0.01

0.50

−0.

020.

060.

24−

0.01

0.39

−0.

02−

0.06


Multimarket Contact and Rivalry 2519Ta

ble

2L

inea

rP

roba

bili

tyE

stim

ates

ofIn

fluen

ces

onL

itig

atio

nP

rope

nsit

y

Mod

el1

Mod

el2

Mod

el3

Mod

el4

Mod

el5

Con

trol

sM

ain

effe

ctM

ain

effe

ct&

H2

Mai

nef

fect

&H

3Fu

llm

odel

Mul

timar

ketc

onta

ct(H

1>

0)0.

007

.000

0.00

6.0

000.

006

.000

0.00

5.0

00(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

Mul

timar

ketc

onta

ct×

Foca

lfir

m’s

tech

nolo

gica

lqu

ality

ofkn

owle

dge

(H2>

0)

0.00

04.0

000.

0004

.000

(0.0

00)

(0.0

00)

Mul

timar

ket

cont

act×

Riv

al’s

abili

tyto

prot

ecti

tskn

owle

dge

(H3<

0)

−0.

24.0

00−

0.24

.000

(0.0

61)

(0.0

61)

Foca

lfirm

’ste

chno

logi

cal

qual

ityof

know

ledg

e(c

itatio

nspe

rpa

tent

)

−0.

0002

.77

−0.

0003

.64

−0.

0002

.72

−0.

0003

.59

−0.

0002

.67

(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

Riv

al’s

abili

tyto

prot

ecti

tskn

owle

dge

(law

suits

aspl

aint

iff

%pa

tent

s)

0.00

08.9

8−

0.00

4.9

1−

0.00

4.9

2−

0.15

.00

−0.

16.0

0(0.0

35)

(0.0

35)

(0.0

35)

(0.0

51)

(0.0

51)

Foca

lfirm

size

(tho

usan

dsof

empl

oyee

s)−

0.00

2.1

5−

0.00

2.1

9−

0.00

1.2

8−

0.00

2.1

6−

0.00

1.2

4(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

Riv

alfir

msi

ze(t

hous

ands

ofem

ploy

ees)

0.00

0003

.99

0.00

01.6

90.

0001

.79

0.00

01.6

10.

0001

.71

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Foca

lfirm

’sne

tinc

ome

($m

illio

ns)

0.00

0001

.81

0.00

0001

.84

0.00

0001

.88

0.00

0000

5.8

90.

0000

003

.93

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Riv

alfir

m’s

neti

ncom

e($

mill

ions

)−

0.00

0001

.68

−0.

0000

01.6

7−

0.00

0001

.63

−0.

0000

01.7

2−

0.00

0001

.67

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Res

ourc

esi

mila

rity

(tec

hnol

ogic

alcl

ass

over

lap)

0.00

3.0

00.

002

.01

0.00

2.0

00.

002

.01

0.00

2.0

1(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)



ble

2co

ntin

ued

Mod

el1

Mod

el2

Mod

el3

Mod

el4

Mod

el5

Con

trol

sM

ain

effe

ctM

ain

effe

ct&

H2

Mai

nef

fect

&H

3Fu

llm

odel

Foca

lfirm

’spa

tent

sin

riva

l’ste

chno

logi

calc

lass

es0.

0000

07.2

30.

0000

004

.94

−0.

0000

01.8

30.

0000

01.9

3−

0.00

0001

.85

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Riv

alfir

m’s

pate

nts

info

cal

firm

’ste

chno

logi

cal

clas

ses

−0.

0000

03.5

9−

0.00

001

.10

−0.

0000

1.1

3−

0.00

001

.05

−0.

0000

1.0

7(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Foca

lfirm

’sav

erag

ecl

aim

spe

rpa

tent

−0.

0002

.40

−0.

0002

.45

−0.

0002

.39

−0.

0002

.45

−0.

0002

.39

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Riv

alfir

m’s

aver

age

clai

ms

per

pate

nt−

0.00

01.3

9−

0.00

01.2

9−

0.00

01.2

6−

0.00

01.2

8−

0.00

01.2

5(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Riv

alfir

m’s

tech

olog

ical

qual

ityof

know

ledg

e(c

itatio

nspe

rpa

tent

)

0.00

04.0

10.

0004

.04

0.00

04.0

30.

0004

.02

0.00

04.0

2(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

(0.0

00)

Foca

lfirm

’spr

oduc

tdi

vers

ity0.

006

.66

0.00

3.8

20.

004

.80

0.00

3.8

40.

003

.83

(0.0

15)

(0.0

15)

(0.0

15)

(0.0

15)

(0.0

15)

Riv

alfir

m’s

prod

uctd

iver

sity

−0.

01.1

7−

0.01

.00

−0.

01.0

0−

0.01

.00

−0.

01.0

0(0.0

04)

(0.0

04)

(0.0

04)

(0.0

04)

(0.0

04)

Foca

lfirm

’sla

wsu

itsas

plai

ntif

fas

a%

ofpa

tent

s−

0.03

.42

−0.

04.3

6−

0.04

.37

−0.

04.3

6−

0.04

.36

(0.0

41)

(0.0

40)

(0.0

40)

(0.0

40)

(0.0

40)

Foca

lfirm

’sla

wsu

itsas

defe

ndan

t0.

001

.45

0.00

1.5

40.

001

.55

0.00

1.5

10.

001

.51

(0.0

02)

(0.0

02)

(0.0

02)

(0.0

02)

(0.0

02)

Riv

alfir

m’s

law

suits

asde

fend

ant

0.00

3.0

10.

002

.06

0.00

2.0

60.

002

.06

0.00

2.0

6(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

(0.0

01)

Con

stan

t0.

05.0

50.

06.0

10.

06.0

10.

06.0

10.

06.0

1N

umbe

rof

obse

rvat

ions

6,33

06,

330

6,33

06,

330

6,33

0

Stan

dard

erro

rsin

pare

nthe

ses,

p-va

lues

inita

lics.

Incl

udes

year

dum

mie

s,ad

viso

rycl

ass

dum

mie

san

dfir

mfix

edef

fect

s.



patent litigation actions as compared to the down-stream actions that have been the focus of most priorMMC studies. As mentioned previously, prior stud-ies show that while retaliation is relatively commonfor downstream actions like new product introduc-tions (Kuester et al., 1999) and pricing actions (Mar-cel et al., 2011; Steenkamp et al., 2005), retaliationis less common for knowledge protection effortsinvolving patent lawsuits. Going beyond this evi-dence from prior literature, we also examined theincidence of retaliation involving patent litigationin our own empirical context. Fully in line with thelow incidence of countersuits from prior research(Somaya, 2003), we find that approximately 9% oflitigation pairs involved a potential retaliatory coun-tersuit.5 Hence, in support of our arguments, we findevidence both that MMC increases the threat of imi-tation and that retaliation is relatively uncommoninvolving patent lawsuits.

Having provided evidence of the proposed routethrough which the hypothesized relationship devel-ops, we turn next to the empirical tests of thehypotheses. In H1, we argue that there is a posi-tive relationship between MMC and the likelihoodthat the focal firm will protect its knowledge-basedresources by initiating patent litigation against arival firm. In support of H1, model 2 in Table 2shows that when MMC increases between the focalfirm and the rival firm, the focal firm is more likelyto initiate a patent lawsuit against the rival firm(𝛽 = 0.007, p= .000).6 To better understand the eco-nomic impact that MMC has on knowledge pro-tection, we examined the change in the probabilityof initiating a patent lawsuit at different levels ofMMC. The probability of litigation shown in thedescriptive statistics in Table 1 is .007, and this cor-responds to an average level of MMC of 1.7. Usingthe coefficient reported in model 2 (𝛽 = 0.007) andholding other variables at their mean values, wenote that a one standard deviation (1.9) increasein MMC increases the probability of litigation by.013 (i.e., 0.013=MMC 𝛽 of 0.007×MMC stan-dard deviation of 1.9) and thereby results in morethan a 100% increase versus the baseline probabil-ity that the focal firm will initiate patent litigation

5 The 9% is the percent of dyads with litigation where a rivalinitiated litigation against the focal firm in the same year that thefocal firm filed a lawsuit against that rival firm. It becomes 15.6%if we also consider countersuits in the subsequent year.6 Both our theory development and prior MMC studies inknowledge-based contexts made us expect a positive relationshipbetween MMC and litigation. In line with expectations, we findno evidence of curvilinear effects.

against a rival firm. The magnitude of this changein probability demonstrates that MMC significantlyaffects a firm’s decision to protect its knowledge.

In H2, we argued that the technological qualityof the focal firm’s knowledge-based resources willexacerbate the positive relationship between MMCand knowledge protection. In support of H2, model3 in Table 2 shows that the interaction betweenMMC and the technological quality of the focalfirm’s knowledge is positive (𝛽 = 0.0004, p= .000),which means that the technological quality of thefocal firm’s knowledge increases the likelihood thatthe focal firm initiates patent litigation against amultimarket competitor. This shows that as the tech-nological quality of the focal firm’s knowledgeincreases, the focal firm is even more likely to pro-tect its knowledge from a multimarket competitor.

In H3, we argued that the rival firm’s abil-ity to protect its knowledge-based resources willattenuate the positive relationship between MMCand knowledge protection. Also, in support ofH3, model 4 in Table 2 shows that the interactionbetween MMC and the rival firm’s ability to pro-tect its knowledge is negative (𝛽 =−0.24, p= .000),which means that the rival firm’s ability to protectits knowledge reduces the likelihood that the focalfirm initiates patent litigation against a multimarketrival. This model shows that when deciding whetherto take actions to protect knowledge from a multi-market competitor that the focal firm considers therival firm’s ability to retaliate, such that the focalfirm is less likely to initiate patent litigation whenthe rival firm has greater ability to initiate a coun-terattack against the focal firm.

In conjunction with the tests in the reportedmodels based on interactions, we also graphicallyexamined the relationship between the MMC andlitigation at high and low levels of the contingencyvariables. We examined the effects of MMC onlitigation for the top 10% of each contingencyvariable versus the bottom 90%.7 The graphsshow fitted lines for the relationship between theMMC and litigation at high and low levels of thecontingency variables.

In line with the empirical test for H2 discussedabove, Figure 1 shows a steeper slope for the fit-ted line corresponding to higher levels of techno-logical quality. A split sample econometric test of

7 Graphs and tests with high and low split at mean or medianvalues also offer support of H2 and H3.



0.0

5.1

.15

.2

0 5 10 15 20

Low Technological Quality

High Technological Quality

PR(L

itiga

tion)

MMC

Figure 1. Contingent effect of focal firm’s technological quality (H2).

-.05

0.0

5.1

.15

0 5 10 15 20MMC

PR(L

itiga

tion)

Low rival's ability to protect its knowledge

High rival's ability to protect its knowledge

Figure 2. Contingent Effect of rival firm’s ability to protect its knowledge (H3).

the difference in the MMC coefficients from lin-ear probability models run at these same low andhigh levels of firm’s technological quality revealsthat coefficients are significantly different (t= 3.1)in the direction hypothesized in H2. Figure 2 alsooffers support for H3 and shows divergent slopes forthe fitted lines for high and low levels of rival’s abil-ity to protect its knowledge. Importantly, Figure 2shows that when only considering observations withhigh levels of rival’s ability to protect its knowledge(i.e., where we restrict the analysis to only considerthose observations where the rival firm is in the top

10%), the relationship between MMC and litigationbecomes negative, thereby reverting to the classicMMC mutual forbearance relationship. This graphprovides some additional evidence to support theargument that MMC may not result in mutual for-bearance because the average threat of retaliationinvolving upstream rivalry may simply be too low todeter the focal firm from initiating patent litigationagainst the rival firm. We also ran a split sampleanalysis comparing these same high and low levelsof the rival firm’s ability to protect its knowledge.Consistent with the graph, the coefficient for the



relationship between MMC and litigation changesfrom a positive coefficient of 0.008 at low levels ofthe rival’s ability to protect its knowledge to a nega-tive coefficient of −0.003 at high levels of the rival’sability to protect its knowledge (t=−3.8). Again,the findings are consistent with our argument thatthe average retaliation threat associated with patentlitigation may not be sufficient to restrain rivalry,which is why the MMC relationship reverts to thenegative one shown in prior MMC literature onlyat high levels of the rival firm’s ability to protectits knowledge, when the rival firm is most able toretaliate. Taken together, the empirical tests com-bined with this additional evidence offer consistentsupport for both of the contingency hypotheses.

Alternative Explanations for Results

Table 3 reports robustness tests developed toinvestigate potential alternative explanations to acausal relationship between MMC and knowledgeprotection. While our use of lagged variables inconjunction with firm fixed-effects helps to addressconcerns about endogeneity, we have developed avariety of robustness tests and supplemental analy-ses to investigate potential alternative explanationsthat may interfere with interpretation of the results.

One potential concern is that while the regres-sions include firm fixed effects, there may be otherreasons related to the firm’s patents, locations ormarkets that are correlated with MMC and a firm’sdecision to initiate patent litigation. For example, itis possible that the relationship between MMC andknowledge protection may be driven entirely byhigh value patents or highly interconnected patents.If this were the case, then the effect of MMC onknowledge protection may result only from theinclusion of observations related to high valueor highly interconnected patents. To investigatethis possibility, we examine whether the positiveeffect of MMC on knowledge protection effortsremained significant when excluding observationswhere the focal firm has a higher than averagelevel of forward citations or higher than averagepercentage of self-citations. In other words, the firstanalysis excludes dyad-year observations wherethe focal firm has an above average level of forwardcitations (i.e., focal firm’s technological quality),and the second analysis excludes dyad-year obser-vations where the focal firm has an above averagelevel of self-citations. Models 1 and 2 in Table 3show that the effect of MMC on patent litigation

remained significantly positive when restrictingthe analyses by only considering observationswith below average levels of forward citations orthose with below average levels of focal firm’spercentage of self-citations. Therefore, we do notfind evidence that the hypothesized relationship isdriven exclusively by the inclusion of observationswhere focal firms have high value patents or bymore interconnected patents.

We also take steps to investigate alternativeexplanations related to the focal firm’s geographiclocations and the markets in which it operates thatmay be correlated with MMC and with the prob-ability of initiating patent litigation. To do so, wedevelop falsification tests. Given our argument thatMMC increases the likelihood that a focal firm willprotect its knowledge-based resources from imita-tion, we would not expect to find a positive relation-ship between MMC and litigation based solely onthe average level of MMC in either the focal firm’smarkets or its geographic locations. To explore thispossibility, in lieu of the MMC variable used inthe reported models, we devised falsification testswhere we constructed an MMC variable using theaverage level of MMC for the other firms (exclud-ing the focal firm) operating in those markets orgeographic locations. Hence, our falsification testsexamine whether, contrary to our dyad-level argu-ments about the relationship between MMC and lit-igation, the average level of MMC (excluding thefocal firm) in the product markets or geographiclocations listed in the firm’s FDA filings will pro-duce the same positive effect on litigation. Mod-els 3 and 4 of Table 3 report falsification tests forgeographic locations and markets. The fact that theaverage MMC for markets and geographic locationsfails to significantly predict patent litigation offerssome additional evidence in support of the argumentthat MMC between a focal firm and its rivals is driv-ing upstream rivalry involving patent litigation.

Another possibility is that the direction of therelationship between MMC and litigation maybe reversed, such that more litigious firms maybe particularly likely to compete with rivals in agreater number of product markets. To help usunderstand whether the firm’s willingness to pro-tect its knowledge-based resources may be drivingboth MMC and litigation, we ran a supplementalanalysis based on a subsample of observationswhere we have excluded cases in which the focalfirm has previously initiated a patent lawsuit duringthe analysis period. Mode1 5 from Table 3 shows



ble

3L

inea

rP

roba

bili

tyE

stim

ates

ofIn

fluen

ces

onL

itig

atio

nP

rope

nsit

y

Mod

el1

Mod

el2

Mod

el3

Mod

el4

Mod

el5

Bel

owav

erag

eB

elow

aver

age

Fals

ifica

tion

test

-Fa

lsifi

catio

nte

st-

Exc

ludi

ngTe

chno

logi

cal

qual

ityIn

terc

onne

cted

ness

Geo

.loc

atio

nsM

arke

tsL

itigi

ous

firm

s

Mul

timar

ketc

onta

ct(H

1>

0)0.

004

.000

0.00

5.0

00−

0.00

8.4

90.

003

.59

0.00

4.0

00(0.0

0)(0.0

0)(0.0

1)(0.0

1)(0.0

0)Fo

calfi

rm’s

tech

nolo

gica

lqu

ality

ofkn

owle

dge

−0.

002

.20

0.00

1.3

4−

0.00

02.7

8−

0.00

02.7

2−

0.00

03.6

3(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)R

ival

’sab

ility

topr

otec

tits

know

ledg

e(l

awsu

itsas

plai

ntif

f%

pate

nts)

−0.

01.7

2−

0.01

.76

0.00

1.9

70.

001

.98

−0.

01.8

4(0.0

4)(0.0

4)(0.0

4)(0.0

4)(0.0

3)

Foca

lfirm

size

(tho

usan

dsof

empl

oyee

s)

0.00

2.4

4−

0.00

2.2

2−

0.00

2.1

4−

0.00

2.1

40.

01.1

9(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

1)

Riv

alfir

msi

ze(t

hous

ands

ofem

ploy

ees)

0.00

01.7

80.

0001

.79

0.00

0004

.99

0.00

0001

.997

−0.

0003

.19

(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)Fo

calfi

rm’s

neti

ncom

e($

mill

ions

)−

0.00

004

.14

−0.

0000

04.4

50.

0000

01.8

40.

0000

01.8

1−

0.00

007

.19

(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)R

ival

firm

’sne

tinc

ome

($m

illio

ns)

0.00

0003

.28

0.00

0002

.49

−0.

0000

01.6

9−

0.00

0001

.67

−0.

0000

001

.97

(0. 0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)R

esou

rce

sim

ilari

ty(t

echn

olog

ical

clas

sov

erla

p)

0.00

1.1

50.

0002

.77

0.00

3.0

00.

003

.00

0.00

2.1

4(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)

Foca

lfirm

’spa

tent

sin

riva

l’ste

chno

logi

cal

clas

ses

−0.

0000

3.5

1−

0.00

001

.39

0.00

001

.23

0.00

001

.23

0.00

01.4

2(0.0

0)(0.0

0)(0.0

0)(0.0

0 )(0.0

0)


Multimarket Contact and Rivalry 2525Ta

ble

3co

ntin

ued

Mod

el1

Mod

el2

Mod

el3

Mod

el4

Mod

el5

Bel

owav

erag

eB

elow

aver

age

Fals

ifica

tion

test

-Fa

lsifi

catio

nte

st-

Exc

ludi

ngTe

chno

logi

cal

qual

ityIn

terc

onne

cted

ness

Geo

.loc

atio

nsM

arke

tsL

itigi

ous

firm

s

Riv

alfir

m’s

pate

nts

info

calfi

rm’s

tech

nolo

gica

lcla

sses

−0.

0000

1.2

2−

0.00

001

.20

−0.

0000

03.5

8−

0.00

0003

.59

0.00

0000

4.9

5(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)

Foca

lfirm

’sav

erag

ecl

aim

spe

rpa

tent

−0.

0004

.38

−0.

0003

.32

−0.

0003

.37

−0.

0002

.43

0.00

02.3

4(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)R

ival

firm

’sav

erag

ecl

aim

spe

rpa

tent

−0.

0003

.04

−0.

0001

.70

−0.

0001

.39

−0.

0001

.40

−0.

0002

.17

(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)R

ival

firm

’ste

chol

ogic

alqu

ality

ofkn

owle

dge

0.00

03.1

60.

0002

.20

0.00

04.0

10.

0004

.01

0.00

01.4

3(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)Fo

calfi

rm’s

prod

uct

dive

rsity

0.01

.72

−0.

01.5

40.

01.7

20.

01.7

30.

01.4

4(0.0

2)(0.0

2)(0.0

2)(0.0

1)(0.0

2)R

ival

firm

’spr

oduc

tdi

vers

ity−

0.01

.02

−0.

01.0

2−

0.01

.18

−0.

01.1

7−

0.00

5.3

0(0.0

1)(0.0

0)(0.0

0)(0.0

0)(0.0

0)Fo

calfi

rm’s

law

suits

aspl

aint

iff

asa

%of

pate

nts

−0.

05.2

0−

0.04

.34

−0.

03.4

3−

0.03

.41

(0. 0

4)(0.0

4)(0.0

4)(0.0

4)

Foca

lfirm

’sla

wsu

itsas

defe

ndan

t−

0.00

2.3

30.

001

.77

0.00

1.4

90.

001

.47

−0.

001

.85

(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)R

ival

firm

’sla

wsu

itsas

defe

ndan

t0.

002

.22

0.00

3.0

20.

003

.01

0.00

3.0

10.

0002

.87

(0.0

0)(0.0

0)(0.0

0)(0.0

0)(0.0

0)C

onst

ant

0.04

.14

0.07

.01

0.06

.06

0.04

.14

−0.

004

.76

Num

ber

ofob

serv

atio

ns3,

667

4,60

06,

282

6,33

02,

872

Stan

dard

erro

rsin

pare

nthe

ses,

P-va

lues

inita

lics.

Mod

els

incl

ude

year

dum

mie

s,ad

viso

rycl

ass

dum

mie

san

dfir

mfix

edef

fect

s.



that we continue to find a positive relationshipbetween MMC and litigation, even when excludingthese firms that might be most inclined towardinitiating patent litigation.

Finally, we also conducted a supplemental analy-sis using a matching approach to explore the pos-sibility that alternative explanations arising fromnon-random assignment of observations may bedriving the results (Bettis et al., 2014). For eachfirm-year observation, we looked for a firm-rivaldyad with high MMC and matched it with afirm–rival dyad with low MMC where the focalfirm was the same across both of these dyads. Toensure that the focal firm was paired with a sim-ilar rival across each of these dyads, we matchedeach of these firm–rival dyads based on the levelof resource similarity and based on whether or notthe rival firm was a defendant. We continue to find apositive relationship between MMC and patent lit-igation (𝛽 = 0.006, p= .000), even when using thismore stringent matching approach where the focalfirm is used as its own control group. While it isarguably impossible to fully rule out all endogene-ity concerns, our use of firm fixed effects along withan extensive list of control variables in conjunctionwith these supplemental analyses and falsificationtests help to increase confidence in the results.

Discussion and Conclusion

Scholars contend that MMC increases familiaritywith and the threat of retaliation against rivalsand, in turn, reduces the intensity of rivalry (i.e.,mutual forbearance hypothesis) entailing down-stream competitive actions such as pricing andmarket entry. But more recent research finds thatMMC in knowledge-intensive industries gives riseto opportunities for firms to manage competitiveuncertainty through imitation and to capture morevalue across multiple markets, which often reducesthe incentive to mutually forbear. In this study, weargue that MMC increases the likelihood that a firmwill initiate patent litigation against a rival firm toprotect its knowledge from imitation. Moreover, wealso argue that a firm will be more likely to pro-tect its knowledge when it possesses higher qual-ity knowledge-based resources, but less likely toprotect its knowledge when the rival has a greaterability to take retaliatory actions against the firm toprotect its own knowledge.

In support of our arguments, we find that asthe number of overlapping FDA product codes inthe medical devices industry between a firm andits rival increases, that the focal firm was morelikely to initiate a patent lawsuit against the rivalfirm. We also find support for both contingencyhypotheses. When the focal firm had a greaternumber of forward citations it was even more likelyto initiate patent litigation against a multimarketrival. Conversely, as the rival firm’s number oflawsuits as plaintiff as a proportion of its patentsincrease, the focal firm was less likely to initiate apatent lawsuit against a multimarket competitor.

Our results do not imply that the traditionalmutual forbearance hypothesis associated withMMC found in earlier studies is flawed. But ourresults do suggest that effects of MMC on rivalrymay differ based on the nature of resources (i.e.,more versus less knowledge-based resources) anddomain of the competitive action (i.e., upstreamversus downstream). In other words, the earlierstudies focused on downstream activities and lessknowledge-based resources have largely foundthat MMC decreases rivalry. More recent research,including this current study, focused on upstreamactivities and knowledge-based resources andshowed that MMC can encourage rivalry as firmsare motivated to protect their knowledge-basedresources from imitation.

Limitations

Before turning to the implications of this study,it is important to acknowledge some limitationsrelated to our measures, empirical setting, and dataavailability. One set of limitations involving ourchoice of measures and empirical setting may affectthe generalizability of the findings. Prior literaturehas drawn attention to a variety of ways that firmscan protect their knowledge, such as secrecy, com-plementary assets and non-compete agreements(James et al., 2013; Levin et al., 1987; Marx, Strum-sky, & Fleming, 2009). In this study, we focus onknowledge protection using patent litigation in partbecause of the prevalence of patent protection inthe medical devices industry, but primarily becausepatent litigation enables us to observe a focal firm’sefforts to protect knowledge directed at a specificrival firm in a way that is not feasible using thesealternative means of knowledge protection. It is,however, important to acknowledge that initiatingpatent litigation is an extreme and costly means of



protecting technological assets. While this may, onthe one hand, add some conservatism to our tests,it also suggests the possibility that the effects ofMMC on efforts to protect knowledge may differdepending on the particular knowledge protectionmechanisms that firms utilize. Therefore, futureresearch may be needed to assess whether MMCalso affects secrecy or investments in complemen-tary assets and whether MMC induces firms to maketrade-offs between these alternative mechanisms.Generalizability concerns may also extend to ourchoice of empirical context. Because the medicaldevices industry is a highly knowledge-intensivesetting where firms typically rely on patent protec-tion, it is also possible that these findings may notextrapolate to less knowledge-intensive contexts orto contexts where firms do not rely as extensivelyon patent protection. Similarly, while prior researchhas consistently demonstrated that MMC is animportant strategic consideration across somecontexts (e.g., airlines, banking), it may be lessrelevant in other industry contexts. Although theeffects of MMC on downstream rivalry have beenshown to generalize across a variety of industrycontexts (Yu & Cannella, 2013), there is a needfor additional studies to investigate whether theeffects of MMC on upstream rivalry apply in lessknowledge-intensive contexts.

Another set of limitations center around dataavailability. As mentioned previously, an impor-tant advantage of using the medical devices set-ting as the empirical context for this study is thefact that FDA data provide a fine-grained wayto observe MMC across product categories andupstream rivalry involving patent litigation. How-ever, these advantages are accompanied by somelimitations that also need to be acknowledged. First,we do not have access to downstream sales infor-mation by product category, which means that weare not able to observe downstream rivalry or toassess the strategic significance of product mar-kets. Hence, in this current study we are not ableto examine whether MMC leads firms to maketrade-offs between upstream and downstream com-petitive actions or whether the technology domainmay provide firms with an opportunity to compen-sate for downstream effects of MMC. Second, wedo not have the ability to observe rival firms’ prod-uct offerings across geographic markets. Therefore,we caution that future work may be needed to under-stand whether product and geographic MMC have

similar effects on a firm’s decision to protect itsupstream knowledge.

Implications for Research and Practice

This study offers important contributions toresearch on multimarket competition. Recentstudies have shed light on a number of importantboundary conditions affecting the MMC–mutualforbearance relationship (Yu & Cannella, 2013).These studies have drawn attention to moderatorsinvolving market entry characteristics (Fan, 2010),the observability of strategies (Greve, 2008), andinternational market characteristics (Yu et al.,2009). This current study adds new insights aboutthe determinants of mutual forbearance by showingthat the relationship between MMC and rivalry maydepend on the competitive domain and the natureof the resources over which firms are competing.Scholars underscore the importance of valuable andheterogeneous resources to a firm’s effort to achievecompetitive advantage (Crook, Ketchen, Combs,& Todd, 2008; Peteraf, 1993; Peteraf & Barney,2003). Importantly, in this study we show thatresource value and heterogeneity in firms’ abilityto respond to competitive actions can compel firmsnot to forbear but to attack even in the presence ofMMC knowing rivals cannot effectively retaliate.Thus, our study suggests that firms may face atrade-off between the strategic benefits associatedwith valuable and heterogeneous knowledge-basedresources and the increased need to protect suchresources from multimarket competitors.

This study also contributes to strategy researchon defensive strategies that firms use to protect pro-prietary technological assets. Studies have madeimportant strides advancing understanding of orga-nizational and technological antecedents of patentlitigation (e.g., Hall & Ziedonis, 2007; Lanjouw& Schankerman, 2001; Polidoro & Toh, 2011;Somaya, 2012) and about the role that a firm’s rep-utation for toughness in patent enforcement playsin mitigating the leakage of knowledge and theloss of employees (Agarwal et al., 2009; Gancoet al., 2015). This current study adds to priorresearch on the defensive strategies devised to pro-tect knowledge-based resources by showing that thestructure of the competitive environment may beanother important determinant of a firm’s decisionto initiate patent litigation against a rival. There isan opportunity for future research to integrate theseinsights with prior work by examining how MMC



affects a firm’s ability to protect technological assetsusing its reputation for toughness in patent enforce-ment. For example, an important question pertainsto how multimarket competition affects knowledgeprotection efforts based on reputational deterrence.Future research could investigate whether MMCamplifies or diminishes the deterrent value associ-ated with firm’s reputation in its efforts to protectthe firm’s resources from imitation or substitution.

Scholars have also highlighted the need forgreater theoretical integration between resourcebased theory and competitive dynamics (e.g., Chen& Miller, 2012). We add to a number of recent stud-ies that have examined the interplay between firmresources and competitive actions (e.g., Clarkson& Toh, 2010; Ndofor et al., 2011; Polidoro & Toh,2011). By developing theory that integrates compe-tition in both factor markets (i.e., knowledge pro-tection) and product markets (i.e., MMC), we showthat competitive contact in downstream productmarkets between rivals can also shape firms’ subse-quent upstream competitive interactions. A promis-ing path for future research involves examining howMMC affects other forms of upstream rivalry. Forexample, scholars could examine how MMC influ-ences rival firms’ decisions to hire one another’s sci-entists and how such talent poaching affects the sub-sequent leakage of knowledge between rival firmsthat compete in multiple markets. Furthermore,scholars can also extend this research by inves-tigating whether firms make trade-offs betweenupstream and downstream rivalry. For example, itis possible that the technology domain may providean opportunity to compensate for other downstreameffects that MMC produces.

Lastly, our study also offers important managerialimplications. This study underscores the increasedimportance of building strong capabilities to protectknowledge when firms compete in multiple mar-kets. We also suggest that managers should havea more nuanced view of how MMC with rivalsaffects competition. In most instances, firms expectto benefit from reduced competition as market over-lap increases. However, the incentive to protect thevalue associated with upstream resources acrossmultiple product markets may, at times, overshadowthe motivation to cooperate between firms. Hence,managers should be mindful of these differencesbetween upstream and downstream actions as theydevelop strategies to achieve and sustain competi-tive advantage from knowledge-based resources.

Acknowledgement

We are extremely grateful to Associate Editor GlennHoetker and two anonymous reviewers for theirvaluable feedback and guidance, which greatlyenhanced our paper.

References

Agarwal, R., Ganco, M., & Ziedonis, R. H. (2009).Reputations for toughness in patent enforcement:Implications for knowledge spillovers via inventormobility. Strategic Management Journal, 30(13),1349–1374.

Aiken, L. S., & West, S. G. (1991). Multiple regression:Testing and interpreting interactions. Newbury Park,CA: Sage.

Anand, J., Mesquita, L. F., & Vassolo, R. S. (2009). Thedynamics of multimarket competition in explorationand exploitation activities. Academy of ManagementJournal, 52, 802–821.

Angrist, J. D., & Pischke, J. (2009). Mostly harmlesseconometrics: An empiricists companion. Princeton,NJ: Princeton University Press.

Barney, J. B. (1986). Strategic factor markets: Expecta-tions, luck, and business strategy. Management science,32(10), 1231–1241.

Baum, J. A. C., & Korn, H. J. (1996). Competitivedynamics of interfirm rivalry. Academy of ManagementJournal, 39, 255–291.

Baum, J. A. C., & Korn, H. J. (1999). Dynamics ofdyadic competitive interaction. Strategic ManagementJournal, 20, 251–278.

Bettis, R., Gambardella, A., Helfat, C., & Mitchell, W.(2014). Quantitative empirical analysis in strategicmanagement. Strategic Management Journal, 35(7),949–953.

Bergen, M., & Peteraf, M. A. (2002). Competitor identi-fication and competitor analysis: A broad-based man-agerial approach. Managerial and Decision Economics,23(4–5), 157–169.

Boeker, W. J., Goodstein, J. S., Stephan, J., & Murmann,J. P. (1997). Competition in a multimarket environment.Organization Science, 8, 126–142.

Capron, L., & Chatain, O. (2008). Competitorsresource-oriented strategies: Acting on competi-tors’ resources through interventions in factor marketsand political markets. Academy of ManagementReview, 33, 97–121.

Cassidy, C. M., & Loree, D. (2001). Dynamics of knowl-edge transfer among multimarket competitors. In B.JAC & H. R. Greve (Eds.), Advances in strategic man-agement (Vol. 18, pp. 141–174). London, England:Emerald.

Chatterji, A. K., & Fabrizio, K. R. (2014). Using users:When does external knowledge enhance corporateproduct innovation? Strategic Management Journal,35(10), 1427–1445.



Chen, M.-J. (1996). Competitor analysis and interfirmrivalry: Toward a theoretical integration. Academy ofManagement Review, 21, 100–134.

Chen, M.-J., & Miller, D. (2012). Competitive dynam-ics: Themes, trends, and a prospective researchplatform. Academy of Management Annals, 6(1),135–210.

Clarkson, G., & Toh, P. K. (2010). Keep out’signs: The roleof deterrence in the competition for resources. StrategicManagement Journal, 31(11), 1202–1225.

Conti, R., Gambardella, A., & Mariani, M. (2013).Learning to be Edison: Inventors, organizations, andbreakthrough inventions. Organization Science, 25(3),833–849.

Crook, T. R., Ketchen, D. J., Combs, J. G., & Todd,S. Y. (2008). Strategic resources and performance: Ameta-analysis. Strategic Management Journal, 29(11),1141–1154.

Dam, K. W. (1994). The economic underpinnings of patentlaw. Journal of Legal Studies, 23, 247–271.

DeCarolis, D. M., & Deeds, D. L. (1999). The impactof stocks and flows of organizational knowledge onfirm performance: An empirical investigation of thebiotechnology industry. Strategic Management Jour-nal, 20(10), 953–968.

Edwards, C. D. (1955). Conglomerate bigness as asource of power. In the National Bureau of EconomicsResearch conference report, Business Concentrationand Price Policy. Princeton, NJ: Princeton UniversityPress.

Fan, T. P. C. (2010). De novo venture strategy: Archincumbency at inaugural entry. Strategic ManagementJournal, 31, 19–38.

FDA. (2015). Medical devices advisory commit-tee. http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/MedicalDevices/MedicalDevicesAdvisoryCommittee/default.htm.

Feinberg, R. M. (1985). “Sales at risk”: A test of the mutualforbearance theory of conglomerate behavior. Journalof Business, 58, 225–241.

Fuentelsaz, L., & Gómez, J. (2006). Multipointcompetition, strategic similarity and entry into geo-graphic markets. Strategic Management Journal, 27,477–499.

Fung, B. (2013, September 3). Is Nokia about to become apatent troll? Washington Post.

Ganco, M., Ziedonis, R. H., & Agarwal, R. (2015). Morestars stay, but the brightest ones still leave: Job hoppingin the shadow of patent enforcement. Strategic Manage-ment Journal, 36(5), 659–685.

Giarratana, M. S., & Mariani, M. (2014). The rela-tionship between knowledge sourcing and fear ofimitation. Strategic Management Journal, 35(8),1144–1163.

Gimeno, J., & Woo, C. Y. (1996). Hypercompeti-tion in a multimarket environment: The role ofstrategic similarity and multimarket contact incompetitive de-escalation. Organization science, 7,322–341.

Gimeno, J., & Woo, C. Y. (1999). Multimarket contact,economies of scope, and firm performance. Academyof Management Journal, 42, 239–259.

Grant, R. M. (1996). Toward a knowledge-based theory ofthe firm [Winter Special Issue]. Strategic ManagementJournal, 17, 109–122.

Greene, W. H. (2003). Econometric analysis. Upper SaddleRiver, NJ: Prentice-Hall.

Greve, H. R. (2008). Multimarket contact and sales growth:Evidence from insurance. Strategic Management Jour-nal, 29(3), 229–249.

Greve, H. R., & Mitsuhashi, H. (2004). Multiunit organiza-tion and multimarket strategy: The dynamics of marketentry and commitment. Scandinavian Journal of Man-agement, 20, 9–30.

Grimm, C. M., Lee, H., & Smith, K. G. (2006). Strategy asaction: Competitive dynamics and competitive advan-tage. New York: Oxford University Press.

Hall, B. H., Jaffe, A., & Trajtenberg, M. (2005). Mar-ket value and patent citations. RAND Journal of Eco-nomics, 36(1), 16–38.

Hall, B. H., & Ziedonis, R. H. (2007). An empiri-cal analysis of patent litigation in the semiconduc-tor industry (Working paper). Berkeley: Universityof California.

Haveman, H. A., & Nonnemaker, L. (2000). Competitionin multiple geographic markets: The impact on growthand market entry. Administrative Science Quarterly, 45,232–267.

Heggestad, A. A., & Rhoades, S. A. (1978). Multi-market interdependence and local market competitionin banking. Review of Economics and Statistics, 60,523–532.

Hoetker, G. (2007). The use of logit and probit models instrategic management research: Critical issues. Strate-gic Management Journal, 28(4), 331–343.

James, S. D., Leiblein, M. J., & Lu, S. (2013). Howfirms capture value from their innovations. Journal ofManagement, 39(5), 1123–1155.

Jayachandran, S., Gimeno, J., & Varadarajan, P. R. (1999).The theory of multimarket competition: A synthesisand implications for marketing strategy. Journal ofMarketing, 63, 49–66.

Kang, W., Bayus, B. L., & Balasubramanian, S. (2010).The strategic effects of multimarket contact: Mutualforbearance and competitive response in the personalcomputer industry. Journal of Marketing Research, 47,415–427.

Karnani, A., & Wernerfelt, B. (1985). Multiple pointcompetition. Strategic Management Journal, 6(1),87–96.

Kennedy, P. (2003). A guide to econometrics (5th end ed.).Cambridge, MA: MIT Press.

King, G., & Zeng, L. (2001). Logistic regression in rareevents data. Political Analysis, 9(2), 137–163.

Korn, H. J., & Baum, J. A. C. (1999). Chance,imitative, and strategic antecedents to multimar-ket contact. Academy of Management Journal, 42,171–193.

Kuester, S., Homburg, C., & Robertson, T. S. (1999).Retaliatory behavior to new product entry. Journal ofMarketing, 63(4), 90–106.

Lai, R., D’Amour, A., Yu, A., Sun, Y., & Fleming, L.(2011). Disambiguation and co-authorship networks


http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/MedicalDevices/MedicalDevicesAdvisoryCommittee/default.htm




of the U.S. patent inventor database (1975-2010)(Working paper). Harvard University.

Lanjouw, J. O., & Schankerman, M. (2001). Characteris-tics of patent litigation: A window on competition.RAND Journal of Economics, 13, 129–151.

Lanjouw, J. O., & Schankerman, M. (2004). Protect-ing intellectual property rights: Are small firmshandicapped? Journal of Law and Economics, 47,45–74.

Levin, R., Klevorick, A., Nelson, R., & Winter, S. (1987).Appropriating the returns from industrial research anddevelopment. Brooking Papers on Economic Activity,18(3), 783–820.

Li, S. X., & Greenwood, R. (2004). The effect ofwithin-industry diversification on firm performance:Synergy creation, multi-market contact and marketstructuration. Strategic Management Journal, 25(12),1131–1153.

Lieberman, M. B., & Asaba, S. (2006). Why do firmsimitate each other? Academy of Management Review,31(2), 366–385.

Liebeskind, J. P. (1996). Knowledge, strategy, and thetheory of the firm. Strategic Management Journal,17(S2), 93–107.

Lichtman, D. (2012). Retaliation as patent strategy. TheMedia Institute. http://www.mediainstitute.org/IPI/2012/061212.php

Lippman, S. A., & Rumelt, R. P. (1982). Uncertainimitability: An analysis of interfirm differences in effi-ciency under competition. The Bell Journal of Eco-nomics, 13, 418–438.

Makadok, R. (2010). The interaction effect of rivalryrestraint and competitive advantage on profit: Why thewhole is less than the sum of the parts. ManagementScience, 56(2), 356–372.

Marcel, J. J., Barr, P. S., & Duhaime, I. M. (2011).The influence of executive cognition on competi-tive dynamics. Strategic Management Journal, 32(2),115–138.

March, J. G. (1991). Exploration and exploitation inorganizational learning. Organization science, 2(1),71–87.

Markman, G. D., Espina, M. I., & Phan, P. H.(2004). Patents as surrogates for inimitable andnon-substitutable resources. Journal of Management,30(4), 529–544.

Markman, G. D., Gianiodis, P. T., & Buchholtz, A. K.(2009). Factor-market rivalry. Academy of ManagementReview, 34(2), 423–441.

Marx, M., Strumsky, D., & Fleming, L. (2009). Mobil-ity, skills, and the Michigan non-compete experiment.Management Science, 55(6), 875–889.

McEvily, S. K., & Chakravarthy, B. (2002). The per-sistence of knowledge-based advantage: An empir-ical test for product performance and technologi-cal knowledge. Strategic Management Journal, 23(4),285–305.

Mitchell, W. (1989). Whether and when? Probabilityand timing of incumbents’ entry into emerging indus-trial subfields. Administrative Science Quarterly, 34,208–230.

Montgomery, D. B., Moore, M. C., & Urbany, J. E.(2005). Reasoning about competitive reactions:Evidence from executives. Marketing Science, 24(1),138–149.

Nelson, R. R., & Winter, S. G. (1982). The evolutionarytheory of economic change. Cambridge, MA: BelknapPress.

Ndofor, H. A., Sirmon, D. G., & He, X. (2011). Firmresources, competitive actions and performance: Inves-tigating a mediated model with evidence from thein-vitro diagnostics industry. Strategic ManagementJournal, 32(6), 640–657.

Peteraf, M. A. (1993). The cornerstones of competitiveadvantage: A resource-based view. Strategic Manage-ment Journal, 14, 179–191.

Peteraf, M. A., & Barney, J. B. (2003). Unraveling theresource-based tangle. Managerial and Decision Eco-nomics, 24, 309–323.

Pisano, G. (2006). Profiting from innovation and theintellectual property revolution. Research Policy, 35,1122–1130.

Polidoro, F., & Toh, P. K. (2011). Letting rivals come closeor warding them off? The effects of substitution threaton imitation deterrence. Academy of Management Jour-nal, 54, 369–392.

Scott, J. T. (2001). Designing multimarket-contact hypoth-esis tests: Patent citations and multimarket contactin the product and innovation markets of the chem-icals industry. In J. A. C. Baum & H. R. Greve(Eds.), Advances in strategic management (Vol. 18, pp.175–200). London, England: Emerald.

Scotchmer, S., & Green, J. (1990). Novelty and disclosurein patent law. The RAND Journal of Economics, 21,131–146.

Sherry, E., & Teece, D. (2004). Royalties, evolving patentrights, and the value of innovation. Research Policy, 33,179–191.

Somaya, D. (2003). Strategic determinants of decisionsnot to settle patent litigation. Strategic ManagementJournal, 24(1), 17–38.

Somaya, D. (2012). Patent strategy and management: Anintegrative review and research agenda. Journal ofManagement, 38, 1084–1114.

Steenkamp, J. B. E., Nijs, V. R., Hanssens, D. M.,& Dekimpe, M. G. (2005). Competitive reactions toadvertising and promotion attacks. Marketing Science,24(1), 35–54.

Teece, D. J. (1986). Profiting from technological inno-vation: Implications for integration, collaboration,licensing and public policy. Research Policy, 15,285–305.

Trajtenberg, M. (1990). A penny for your quotes: Patentcitations and the value of innovations. The Rand Jour-nal of Economics, 21(1), 172–187.

Venkataraman, S., Chen, M.-J., & MacMillan, I. C.(1997). Anticipating reactions: Factors that shapecompetitor responses. In G. S. Day & D. J. Reibstein(Eds.), Wharton on dynamic competitive strategy(pp. 198–219). New York, NY: John Wiley andSons, Inc.

Wiersema, M., & Bowen, H. (2009). The use of limiteddependent variable techniques in strategy research:


http://www.mediainstitute.org/IPI/2012/061212.php

http://www.mediainstitute.org/IPI/2012/061212.php


Issues and methods. Strategic Management Journal, 30,679–692.

Wu, B. (2013). Opportunity costs, industry dynamics, andcorporate diversification: Evidence from the cardio-vascular medical device industry, 1976-2004. StrategicManagement Journal, 34, 1265–1287.

Young, G., Smith, K. G., Grimm, C. M., & Simon, D.(2000). Multimarket contact and resource dissimilarity:A competitive dynamics perspective. Journal of Man-agement, 26, 1217–1236.

Yu, T., & Cannella, A. A. (2007). Rivalry between multina-tional enterprises: An event history approach. Academyof Management Journal, 50, 663–684.

Yu, T., & Cannella, A. A. (2013). A comprehensivereview of multimarket competition research. Journal ofManagement, 39, 76–109.

Yu, T., Subramaniam, M., & Cannella, A. A. (2009).Rivalry deterrence in international markets: Contin-gencies governing the mutual forbearance hypothesis.Academy of Management Journal, 52, 127–147.


Documents

MULTIMARKET CONTACT AND RIVALRY OVER ...mason.gmu.edu/~hleeb/SMJ_17_MMC_Patent_Litigation.pdfIndeed, Jayachandran etal. (1999: p. 58) state that “regardless of multimarket contact,