HOW DO SOCIAL DEFENSES WORK? A RESOURCE- …rkatila/new/pdf/KatilaAMJ2014.pdfhow do social defenses work? a resource-dependence lens on technology ventures, venture capital investors,

HOW DO SOCIAL DEFENSES WORK? A RESOURCE-DEPENDENCE LENS ON TECHNOLOGY VENTURES, VENTURE

CAPITAL INVESTORS, AND CORPORATE RELATIONSHIPS

BENJAMIN L. HALLENUniversity of Washington

RIITTA KATILAStanford University

JEFF D. ROSENBERGERWealthfront

Interorganizational relationships offer many potential benefits, but they also expose firmsto dangers, such as misappropriation, which pull partners apart. This tension betweencollaboration and competition is central to tie formation, especially for young technologyfirms that have both a high need for resources and high appropriability of their ownresources. Prior work has examined legal and timing defenses that enable interorganiza-tional ties by such low-power firms; we focus here on social defenses. In a longitudinalstudy of equity tie formation between young firms and established corporate “sharks,”spanning five technology-based industries and 25 years, we unpack the effects of socialdefenses and find, intriguingly, that centrally positioned third parties are a particularlypowerful social defense and that third-party social defenses are especially significantwhen more traditional defenses are unavailable. We thus offer the insight that third-partychaperones (central venture capital investors) play a key role in helping young firms tomitigate and navigate vulnerabilities while mobilizing resources.

They took our ideas and tried to claim them as theirown, and they tried to crush a little company.

(Infoflows founder, describing the young firm’sonce-promising partnership with established

firm Corbis, quoted in Lohr, 2010)

Firms depend on their environments forresources—but, as resource-dependence scholarspoint out, environments are not always dependable(Pfeffer & Salancik, 1978). As a consequence, firmsengage in an array of tactics to maintain controlover the resources they need and to reduce depen-dence on others. Empirical research has confirmedseveral key tactics that can be used to gain controlover significant resources, including mergers andacquisitions (Casciaro & Piskorski, 2005), ties withthe government (Hillman et al., 2009), and offers ofboard seats to representatives of key resource hold-ers (Pfeffer, 1972).

Recently, a stream of research on resource depen-dence has identified interorganizational ties as aprevalent and important tactic with which to re-duce dependence (Gulati, 1995b). Ties offer accessto resources and alter how a firm is perceived byothers (Podolny, 1993). But ties also produce newdependencies and often shift relative power, suchas when partners compete for control over valuedresources and one party seeks to force the other toaccept an unfavorable power imbalance (Emerson,1962). These outcomes are particularly likely when

The authors thank seminar audiences at the WhartonTechnology Conference, the London Business School,the University of Texas at Austin, the European Groupfor Organizational Studies, the INSEAD Network Evolu-tion Conference, the West Coast Research Symposium,the Smith Entrepreneurship Conference, and the review-ers and audience participants at the Academy of Man-agement and Strategic Management Society annual meet-ings. We are particularly grateful to associate editor TimPollock and our anonymous referees, and to Kathy Eisen-hardt, Lori Rosenkopf, Melissa Schilling, and Anil Guptafor their helpful feedback and comments. The researchsupport of the National Science Foundation (Grants SBE-0915236 and IIA-1305078 for the second author) and theStanford Technology Ventures Program is gratefully ac-knowledged. The first author was on faculty at LondonBusiness School at the time of the paper’s acceptance andgratefully acknowledges their support.

1078

� Academy of Management Journal2014, Vol. 57, No. 4, 1078–1101.http://dx.doi.org/10.5465/amj.2012.0003

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download, or email articles for individual use only.

the prospective partners are also competitors. Com-petitors have reasons to collaborate for shared ben-efit, but often simultaneously behave opportunisti-cally, misappropriating intellectual property oradversely altering a partner’s agenda for their ownbenefit (Agarwal, Croson, & Mahoney, 2010; Gulati& Singh, 1998).

Prior research has examined one particularlychallenging situation in which a potential partner(often termed a “shark”) is both particularly attrac-tive and particularly dangerous—a scenario thatoften characterizes tie formation between youngfirms and potentially rivalrous established corpo-rations (see Diestre & Rajagopalan, 2012; Katila,Rosenberger, & Eisenhardt, 2008). Specifically, thiswork examines young firms’ decisions to form part-nerships with corporations—that is, to “swim withsharks”—in preference to safer partners likely tooffer somewhat less attractive resources, but alsofewer risks of misappropriation.1 Katila et al.(2008), for example, found that young technologyfirms are likely to choose sharks when they needthe particular resources that these partners canuniquely provide and when they can defend theirown resources well against misappropriation. Morerecently, Diestre and Rajagopalan (2012) looked atthe “sharks dilemma” in the context of research anddevelopment (R&D) ties between young biotechnol-ogy firms and large pharmaceutical corporations.They focused on the corporate shark’s ability andincentives to appropriate knowledge, and found thatyoung firms are more likely to form relationshipswhen corporate sharks are less dangerous.

Central to the sharks dilemma is the resource-dependence concern of power imbalance (Pfeffer &Salancik, 1978). Resource interdependencies drawfirms into the relationship and create mutual de-pendence, but the young partner may begin to loseits leverage once its unique technology or marketopportunity has been revealed fully enough to bemisappropriated. In the absence of relevant de-fenses, the established partner may become thenear-exclusive holder of power in the relationship.To prevent such imbalances, prior research onsharks notes that young firms undertake classicpower-balancing operations (Emerson, 1962): they

form coalitions (for example enlisting the power ofthe state via legal defenses), reassess their resourceneeds and adjust their goals (perhaps delaying the tievia timing defenses), and identify alternative partnerswhose interests are more genuinely interdependent(that is, find less dangerous sharks). In sum, the re-search draws attention to deterrents to tie formationand demonstrates that, for young firms, the decisionto enter into a relationship entails not merely gettingaccess to resources, but also protecting the firm’s ownresources from exploitation.

Despite these significant insights, scholarly workon the sharks dilemma focuses on specific defensesavailable only to some, but not all, ventures. It isunspecific, for example, about how relationshipsform when suitable defenses are lacking, such aswhen legal and timing defenses are unavailable, or,when the most dangerous sharks also have the bestresources. Another unexamined question is the in-fluence of social context, because social defensesprovided by a firm’s existing partners—which canprevent misappropriation in subsequent relation-ships—may be particularly relevant for young firmswith limited access to other defenses. It also re-mains unexplored whether all social defenses areequally valuable, or whether some types are partic-ularly helpful for young firms—even potentiallyable to substitute for other defenses. Given these un-resolved questions, this study asks: how do socialdefenses influence young firms’ tie formation withcorporate partners (sharks)?

We tested the effects of social defenses usingquantitative data from 1979 to 2003 on 700 youngU.S. firms in five technology-intensive industriesthat made decisions about entering into corporateventure capital (CVC) relationships. We also ana-lyzed data on 11,132 corporate venture dyads, tofurther understand which specific corporate part-ners were chosen, and gathered qualitative evi-dence, including interviews with entrepreneurs,corporate investors, and venture capitalists, todeepen our understanding. Our analysis centers oncorporate investment (equity) relationships thatcome about because of young ventures’ need for theresources that corporations possess and despite theoften high risk of misappropriation of the ventures’own resources that makes defenses salient in theserelationships. Further, we focus on third-party so-cial defenses that draw on young firms’ existingventure capital (VC) investors and on their networkconnections. VC investors are an effective source ofsocial defenses owing to their ownership in theyoung firm, the prevalence of past and future inter-

1 In nature, the relationship between sharks and theirsmall-fish partners is both dangerous and beneficial:small fish eat parasites off the shark’s skin—a joint ben-efit—in turn gaining protection (from mid-sized fish) andtransportation (a free ride), although they also run therisk of being eaten themselves (Clark & Nelson, 1997).

2014 1079Hallen, Katila, and Rosenberger

actions among VC investors and CVC investors, andtheir ability to relay both information and relation-ship opportunities within their investment networks.

This paper contributes to our theoretical grasp ofan understudied aspect of resource dependencemanifested at the critical juncture of relationshipformation. Resource-dependence studies typicallyfocus on adjusting dependencies post-formation,ignoring the reality that young firms need to antic-ipate resource misappropriation prior to forming atie. We highlight social defenses that borrow thepower of third parties on behalf of young and oth-erwise low-power firms when they form new ties.More broadly, we contribute to the recent researchstream on low-power firms and their ability to over-come constraints at tie formation (Ahuja, 2000b;Hallen, 2008). Our contribution is to introduce so-cial defenses as a way in which to mitigate powerloss and thus lower barriers to tie formation, espe-cially for young firms.

We also contribute to the network literature’ssearch for “mechanisms . . . by which firms over-come the fears associated with partnerships” (Gu-lati, 2007: 49). In particular, we unpack the rangeand influence of social defenses, and show theprimary role of centrally positioned third parties infacilitating tie formation. We also find that geo-graphical proximity has an unexpected aligning ef-fect, with nearby third parties able to reinforce theirpreferences about “safer” partners more stronglythan can distant third parties. Finally, we show thatsocial defenses are particularly important whenother defenses are weak or absent; thus social de-fenses enable relationships that would otherwisefail to form. Taken together, our results suggest abroader and more nuanced view of social defenses.

RESEARCH CONTEXT: TECHNOLOGYVENTURES AND THEIR FUNDING PARTNERS

Prior literature on venture financing provides afoundation for understanding the resource-depen-dence challenges relevant to our study. One streamfocuses on young firms’ funding relationships withVC investors. Research reveals that entrepreneursform equity relationships with VC investors be-cause they need both financial resources and ad-vice on formulating and executing strategy (Cox,Katila, & Eisenhardt, 2014). In turn, VC investorsenter into these relationships both to provide finan-cial returns to their institutional investors and be-cause they view themselves as particularly capableof helping entrepreneurial firms to reach successful

exit events (Wasserman, 2012). In fact, VC inves-tors’ self-perceived identities often center on theirrole as co-creators of entrepreneurial firms (Graeb-ner & Eisenhardt, 2004; Hellmann, 2000). In keep-ing with these observations, those VC investorswhom we interviewed frequently used the term“we” to refer to themselves as members of the ven-ture team, and to indicate that their interests andthose of the venture were the same. Prior researchon venture financing confirms this alignment: VCinvestors typically align their own incentives withthose of entrepreneurs (Bitler, Moskowitz, & Viss-ing-Jørgensen, 2005), contract out agency risks viaterm sheets2 (Kaplan & Strömberg, 2004), and mon-itor their investments by occupying positions onyoung companies’ boards (Sapienza & Gupta,1994). While board representation and interactioncollectively result in VC investors having a degreeof authority over young companies, they also con-strain investors’ behavior by subjecting them to fidu-ciary responsibilities (Gulati & Singh, 1998). Overall,since VC investors seek financial returns via venturesuccess and stake their professional identities oncoaching successful ventures, their strategic align-ment with entrepreneurs is typically high.

A related stream of research focuses on youngenterprises’ funding relationships with CVC inves-tors. Young firms typically form CVC relationshipsto gain access to corporate resources (Dushnitsky &Lenox, 2006; Park & Steensma, 2012). Some of theentrepreneurs that we interviewed received largecash infusions from corporations; others soughtphysical resources (such as access to manufactur-ing or testing facilities, or sales channels) or, in thewords of one interviewee, “[investors] who couldhelp with customers and who would be useful froma business perspective”—that is, resources typi-cally not readily available elsewhere. Corporationsin turn sometimes formed such ties to use excessresources (for example, idle capacity in a saleschannel or at a manufacturing plant), but they wereparticularly eager to do so to acquire insight intopotentially disruptive technologies (Tushman & An-derson, 1986; Utterback, 1994). Via interactions withventures (such as pre-investment due-diligence dis-cussions and post-investment liaison relationships),

2 A “term sheet” is a legal document that specifies aninvestor’s funding terms. Commonly utilized terms suchas founder vesting, liquidation preferences, and the useof equity (vs. debt) ensure that both entrepreneurs andVC investors are primarily motivated to achieve a favor-able exit event (i.e., acquisition or initial public offering).

1080 AugustAcademy of Management Journal

corporations gained understanding of new technolo-gies and markets that the ventures had explored (Du-shnitsky & Lenox, 2005a; Wadhwa & Kotha, 2006).Thus, from the corporate perspective, investment rela-tionships with new firms offered both financial gain and(perhaps even primarily) access to new resources—especially technology.

Yet the same factors that promote the perceptionof resource interdependencies between venturesand corporations make corporations a more danger-ous type of partner—that is, a shark. Corporations,the technologies, customers, or manufacturing ca-pabilities of which are particularly attractive toyoung ventures, may also have strong incentivesto misappropriate young firms’ own innovations orto alter young firms’ agendas to their own advan-tage (Dushnitsky & Lenox, 2006; Kale, Singh, &Perlmutter, 2000). For example, one of our corpo-rate interviewees noted that investments in ven-tures had enabled his firm to “influence [ventures’]activities in directions that are of interest to us, farbeyond what is afforded us in typical R&D con-tracts.” Similarly, a well-regarded venture capital-ist observed that “corporations have other motivesfor doing venture capital . . . motives [that] aren’tparticularly well aligned with the founders, manag-ers, and financial investors. So there’s always tensionin a corporate venture investment, and it’s not alwayshealthy” (Wilson, 2008). This tension stands in sharpcontrast to the alignment of other investors, such asVC investors, who typically invest for purely finan-cial reasons and structure their investments to aligntheir interests closely with those of the young firm(Sahlman, 1990). Furthermore, unlike VC investors,few CVC investors occupy board seats as part of theirinvestment relationships; as a consequence, the latteravoid the board-related fiduciary responsibilities thatwould require them to act in a young firm’s interestand preserve the flexibility to pursue their own stra-tegic interests. Overall, our fieldwork and prior re-search on the sharks dilemma indicate that relation-ships between young firms and CVC investors, ascompared to VC investors, entail both a sharp extratension between attractive corporate resources andpossible misappropriation of the young firm’s ownresources.

SOCIAL DEFENSES AND CORPORATEPARTNERS

When contemplating CVC ties, young firms bal-ance gaining corporate resources against defendingtheir own resources. Most organizational theory re-

search on corporate sharks focuses on formal legalinstruments and partnering with known andtrusted actors as defenses against resource misap-propriation (e.g., Diestre & Rajagopalan, 2012; Ka-tila et al., 2008). Network theorists, by contrast,have examined the ability of third parties to facili-tate trust in relationships and to reduce misappro-priation (Bae & Gargiulo, 2004; Burt, 2005). Thirdparties are interesting because their influence is aby-product of network structure; it is not dyad-specific (Gulati, 1995a; Li, Eden, Hitt, & Ireland,2008). As a consequence, social defenses providedby existing third-party partners may be particularlyattractive to young firms, and to those new to asetting that have yet to develop trust and affectthrough dyadic interactions. The literature pointsto several mechanisms by which third parties canact as social defenses. We focus on two—disciplin-ing and aligning—which are particularly relevantto tie formation. We then present hypotheses abouthow each may influence young firms’ willingnessto form ties with corporate partners.

Social Defenses

Effects of social defenses: Disciplining. Thirdparties can influence tie formation because theyrepresent a threat of “disciplining” opportunisticpartners. For example, third parties may: (a) termi-nate current ties or avoid future ties with the of-fending party (Ahuja, 2000a; Burt, 2005: 105–111);or (b) broadcast allegations of opportunistic behav-ior that could damage the offending party’s reputa-tion (Gulati, 1995b; Soda, Usai, & Zaheer, 2004).Recipients of broadcasts may in turn participate insuch disciplining, either to protect themselves frommisappropriation or out of a sense of obligation tothe third party. Disciplining is thus a form of repu-tational threat with broad future implications(Raub & Weesie, 1990). Viewed through the lens ofpower balancing, disciplining gives the low-powerfirm a way in which to influence its partner’s rep-utation, and thus increases the high-power firm’smotivation to behave well within the relationship.Disciplining is thus a classic power-balancing tac-tic: because the threat of discipline puts its reputa-tion at risk, the high-power actor becomes moremotivationally invested in the goals of the other-wise low-power actor, and as a result its depen-dence in the relationship increases (Emerson,1962). In sum, the threat of disciplining reducespower imbalance in relationships.


Core to the third party’s effectiveness at disci-plining is its position in the overall network. Spe-cifically, when third parties are: (a) more desirablepartners themselves; and (b) better able to broad-cast allegations, the partners in the relationshiprealize that they have much to lose from behavingopportunistically (Pollock, 2004; Portes & Sensen-brenner, 1993). The core features of third partiesare nicely captured by centrality—that is, by theextent to which the third party is connected towell-connected others (Bonacich, 1987). Centrallyplaced third parties are more desirable partnersthemselves because they occupy valuable networkpositions: they are implicitly of high quality andregarded well by others, and consequently conferprestige on those firms with which they connect(Podolny, 1993; Pollock, Chen, Jackson, & Ham-brick, 2010; Pollock & Gulati, 2007). Central thirdparties are also likely to possess better platformsfrom which to broadcast allegations and calls fordisciplining, both because their allegations are aptto be perceived as credible and noteworthy (Benja-min & Podolny, 1999), and because the short net-work paths between themselves and others in thefield accelerate the spread of information (Gulati &Gargiulo, 1999; Podolny, 2001). Thus relationshipsare more likely to form in the presence of thirdparties with better ability to discipline misbehav-ing partners.

Effects of social defenses: Aligning. Third par-ties can also influence tie formation by helping toidentify better aligned dyads. For example, thirdparties may suggest partners, or characteristics ofpartners, that would reduce the threat of opportu-nistic behavior in a relationship. From the power-balancing perspective (Emerson, 1962), aligningidentifies mutually beneficial relationships anddiscourages those relationships in which opportu-nistic behavior is especially likely. If the point ofdisciplining is to enable the low-power actor toincrease its power by grabbing the coattails of athird-party chaperone, the point of third-partyaligning is to enhance its power by “setting up agood blind date” with desirable prospective part-ners with mutual dependence (Emerson, 1962).

The third party’s effectiveness at aligning is de-termined by how frequently it interacts with thefocal venture, and thus how well it knows its needsand vulnerabilities. Such familiarity is importantin two ways. First, third parties that interact fre-quently with the focal partner are likely to be moreinterested in, and better able to identify, alignedrelationships. They know the firm well, and can

tailor their recommendations about potential rela-tionships and well-aligned matches (Gulati, 1995b;Sorenson & Stuart, 2001). Second, frequent interac-tion makes recommendations more influential, be-cause third parties have more opportunities to ex-plain the reasoning behind their views on potentialpartners. As a whole, the logic of aligning suggeststhat third-party connections enable relationshipsby helping to identify partners whose interests arewell aligned.

Social Defenses and the Formation ofCorporate Relationships

The hypotheses that follow focus on existing re-lationships with VC investors as social defensesand on the effects of such defenses on young firms’subsequent ties with CVC investors. We examinetwo significant characteristics of VC investors: theircentrality, and their proximity to the young firm.The centrality of a young firm’s VC investors cap-tures the investors’ position in the overall VC syn-dication network, and therefore their desirabilityand network reach. Central VC investors, such asSequoia Capital and New Enterprise Associates, areviewed as highly desirable syndicate partners byother investors (including CVC investors) and arewell positioned to influence industry informationflows. Geographic proximity between the youngfirm and its VC investors captures the familiarityand strength of the immediate tie, because moreproximate investors are likely to interact more fre-quently with a young firm; thus they have a closerand more influential relationship. We will considerhow these two characteristics give rise to socialdefenses and change power dynamics in the ven-ture–CVC relationship.

We propose that young firms backed by a centralVC investor are likely to form more corporate invest-ment relationships for two reasons. First, the merepresence of a central venture capitalist as an investoris likely to discourage opportunistic behavior, be-cause corporate investors in the venture do not wantto jeopardize future relationships with a high-statusVC investor. In particular, more central VC investorstypically invest in higher-quality ventures (Hallen,2008), making such investors particularly attractivefuture syndication partners for corporate investors.Thus we expect young firms to anticipate that thepresence of central VC investors will motivate goodbehavior, and therefore to be more confident aboutforming corporate relationships.


Second, we expect that central VC investors canmore effectively tarnish the reputations of misbe-having corporate partners, because they enjoyshorter paths to more organizations, and because,as high-status investors, their allegations are per-ceived as particularly credible and noteworthy. Inother words, investors that are more central in theco-investment (syndication) network are likely tohave more effective platforms from which to broad-cast, giving their young firms a larger pool of de-fenders when a corporate partner misbehaves. Cen-tral VC investors are also motivated to sanctionmisbehaving corporate investors because their ownfinancial performance (and their access to futurecapital) is often tied to their ventures’ ability toretain control of intellectual property and to reachsuccessful exit events. In fact, developing a “repu-tation for toughness” (see Agarwal et al., 2009) canprotect a VC investor’s current and future invest-ments. VC investors compete with other investorsfor the most promising venture investment oppor-tunities and they will be viewed favorably as inves-tors if they are ready to reveal corporate misbehav-ior. Overall, we argue that central VC investors’interests in maintaining their own desirability asfuture partners, and their ability and incentives tobroadcast misbehavior to the co-investment net-work, make them “chaperones” that facilitate cor-porate relationships by lowering the power of cor-porations relative to young firms.

Of course, tie formation also depends on the in-terests of the corporate partner—in other words, itdoes indeed “take two to tango.” From the corpo-rate perspective, young firms backed by a centralVC investor are attractive because they are morelikely to be of high quality (Pollock & Gulati, 2007;Sørensen, 2007). So even though CVC investorsmay recognize that they will be more constrained,such young firms are likely to be more attractiveand conspicuous investment targets for them. Thuswe propose as follows:

Hypothesis 1. Entrepreneurs with more centralVC investors will form more investment rela-tionships with corporations.

Field evidence also suggests that VC investorsthat are geographically near to their portfolio firmsare more actively involved with them and under-stand their needs better (Lee, Pollock, & Jin, 2011;Sorenson & Stuart, 2001). By contrast, the formaland often superficial nature of board meetings mayinhibit the ability of remote VC investors to offer

rich advice and the ability of entrepreneurs to ac-cept it (Garg, 2013).

We propose that young firms with VC investorsthat are geographically more proximate are likely toengage more corporate partners. First, we arguethat, because they are nearby, local investors paymore attention to, and are more highly motivated tobe involved with, their portfolio companies; there-fore they are likely to be able to suggest betteraligned partners. In other words, a local VC inves-tor with a rich understanding of a young firm’sresources and vulnerabilities can offer customizedadvice about potential fit with particular CVC in-vestor partners. Also, local VC investors are avail-able to explain the reasoning behind their recom-mendations, making them more influential fromthe young firm’s standpoint. “You can’t have peo-ple just glancing at the board book two days beforethe board meeting,” explained one entrepreneurwhom we interviewed. “You need somebody whois meeting with you a couple of times a week . . .[at] a breakfast meeting . . . who is going to stay intune with what is going on in your head . . . [with]the things they can anticipate because they areconnected enough to what is going on.” Overall,geographical proximity to VC investors tends toincrease interaction, mentoring, and knowledgeflow as a result of more frequent planned interac-tion and more serendipitous interaction (Sapienza& Gupta, 1994). By contrast, distant VC partners arelikely to be less helpful, less aware that such deci-sions are under consideration (especially early on),and perhaps also less motivated and less knowl-edgeable for the purposes of assessing alignmentwith potential CVC partners. Thus, owing to theirgreater influence and better tailored advice, proxi-mate VC investors will make young firms feel saferabout pursuing corporate relationships by suggest-ing dyads in which power will be better balanced.

Firms backed by local VC investors can also beattractive for corporate investors. Local VC inves-tors are able to offer richer and stronger assess-ments of the young firm’s quality (Sorenson &Stuart, 2001), and are more likely to allocate dis-proportionate attention to endorsing nearby, ratherthan distant, firms (Levinthal & March, 1993), andthese nuanced recommendations are likely to at-tract more potential corporate partners. In contrast,young firms without local VC investors may receiveweaker and less precise endorsements, and maythus appear less attractive. We therefore proposethe following:


Hypothesis 2. Entrepreneurs with more geo-graphically local VC investors will form moreinvestment relationships with corporations.

Thus far we have examined VC centrality and VCproximity separately, but have not discussed howthey will influence relationship formation in com-bination. Because many young firms have both cen-tral VC investors and proximate ones, these dynam-ics are germane. We propose that centrality andproximity are complements, in the sense that theinfluence of VC centrality on tie formation riseswith increase in VC proximity. For example, VCsyndicates that possess both qualities may have aneven better understanding of both sides of thecoin—that is, they might meet frequently and in-formally with young firms to give rich advice aboutdesirable partners, and may also be particularlyable to discipline if a partner misbehaves, owing totheir centrality and network reach. Furthermore,the guidance of VC investors that are both proxi-mate and central may be particularly salient andmeaningful if it is delivered via proximate, infor-mal interaction with central high-status investors(Hovland, Janis, & Kelley, 1953; Magee & Galinsky,2008). In sum, having more local VC investors mayshift the influence of VC centrality from a puredisciplining defense toward a complementary com-bination of disciplining and aligning—potentiallyallowing young firms to become increasingly com-fortable engaging with more corporate sharks.

From the corporate investor perspective, thebacking of central and local VC investors may alsoenhance the attractiveness of a young firm. Specif-ically, this combination may better assure CVC in-vestors of a young firm’s quality and thus enhanceits prospects, with proximity enhancing the rich-ness and reliability of a venture capitalist’s adviceand endorsement of the venture, and centralitymaking it more likely that a VC investor has takena hand in developing a high-quality young firm andwill continue to do so. Overall, we therefore pro-pose that centrality and proximity are comple-ments, such that their combination will further in-crease the number of a young firm’s ties withcorporations:

Hypothesis 3. The more local its VC investors,the stronger the relationship between the cen-trality of a young firm’s VC investors and thenumber of corporate investment relationshipsthat it forms.

Our first hypotheses have focused on social de-fenses, but research on the sharks dilemma showsthat several other defenses also enable corporaterelationships. Katila et al. (2008) found that strongtrade secret regimes make corporate relationshipsmore likely, because trade secrets are an agreeabledefense for both the young firm and the corpora-tion: trade secrets provide young firms with a legalright to damages in the event of misappropriation,while assuring corporate investors that a youngfirm’s insights have not been broadly disclosed.Katila et al. (2008) also found that forming relation-ships at a later stage in a venture’s development(termed a “timing defense”) makes corporate rela-tionships more likely: as young firms mature, theirinnovations become harder to imitate and misap-propriate, thus reducing the likelihood of detri-mental shifts in power. For instance, early-stageboard meetings are likely to focus on technologydecisions; later-stage meetings focus more on mar-keting and the sales pipeline, and thus are apt toreveal less about the technology. Additionally, cor-porate investors are also likely to prefer investingin older ventures because they gain access to moreadvanced technologies (Cox et al., 2014). Overall,both trade secrets and timing have been found to bemutually agreeable defenses for young firms andtheir corporate investors.

But prior literature has left unexplored how tiesform when secrecy and timing defenses are weak orunavailable. Legal defenses, such as trade secrets,are typically enforceable only in certain industries(Heeley, Matusik, & Jain, 2007; Levin, Klevorick,Nelson, Winter, Gilbert, & Griliches, 1987). Simi-larly, timing defenses are less available to early-stage firms and often entail opportunity costs, be-cause they delay access to the resources that CVCrelationships can provide.

We propose that, although these traditionallystudied “formal” defenses (Dushnitsky & Lenox,2005a; Katila et al., 2008; Teece, 1986) may be themost salient, social defenses are likely to becomeespecially valuable when these other defenses areweak or unavailable. In other words, because socialdefenses are a by-product of network structure,they can be effective even when other defensesare not readily or fully available. In particular, cen-tral or proximate third-party partners (and theirrelated skills in disciplining and aligning) maycompensate for the lack of a strong trade secretregime and may facilitate access to corporate re-sources. Trade secret regimes are tied to an indus-try, whereas social defenses are not. Thus, for ven-


tures in weak trade secret regimes, social defensesvia central and proximate partners may provide acompetitive advantage over same-industry rivalsthat lack such defenses. Similarly, social defensesthat borrow the power of central VC partners areavailable even for early-stage firms and may beparticularly significant for them, because they lackthe timing defense that is available to mature firms.In the same way, the aligning quality of nearby VCinvestors can compensate for the lack of timingdefense, and may even be particularly useful be-cause local investors can make the partner matchwithout giving away insights into the venture’stechnology.

From the corporate perspective, third-party so-cial defenses may also replace traditional defensesto some extent. Just as trade secret protections mayattractively convey to corporate investors that ayoung firm’s intellectual property has not beenwidely revealed (Katila et al., 2008), so too maycentral or nearby VC investors provide substituteguarantees that the venture’s inventions and tech-nologies have not been widely leaked and theirvalue not captured by others in the industry. Inaddition, while corporations generally prefer toform relationships with later-round ventures, thequality and value of which has been more thor-oughly established (Katila et al., 2008), having cen-tral or nearby VC investors may provide alternativesignals of quality and rich endorsement that help toattract corporations to young ventures. We thuspropose the following:

Hypothesis 4a. The weaker its secrecy de-fenses, the stronger the relationship betweenthe centrality of a young firm’s VC investorsand the number of corporate investment rela-tionships that it forms.

Hypothesis 4b. The weaker its timing defenses,the stronger the relationship between the cen-trality of a young firm’s VC investors and thenumber of corporate investment relationshipsthat it forms.

Hypothesis 4c. The weaker its secrecy de-fenses, the stronger the relationship betweenthe proximity of a young firm’s VC investorsand the number of corporate investment rela-tionships that it forms.

Hypothesis 4d. The weaker its timing defenses,the stronger the relationship between the prox-imity of a young firm’s VC investors and the

number of corporate investment relationshipsthat it forms.

METHODS

Sample and Data Sources

We analyzed the formation of CVC relationshipsby VC-backed, high-technology firms over the 25-year period between 1979 and 2003. We chose VC-backed firms because such backing indicates hightechnology and market potential, and likely inter-est on the part of CVC investors, and because suchyoung firms are thus apt to have a choice in theirformation of ties, including the option of forgoingcorporate ties entirely.

Our sample was drawn from the population ofU.S. high-technology firms that received initialventure funding between 1979 and 1995. We beganthe sample in 1979, the year in which the U.S.Department of Labor clarified its “prudent man”rule explicitly to allow pension-fund managers toinvest in high-risk assets, including VC (Hochberg,Ljungqvist, & Lu, 2007). This regulatory changegreatly increased the available supply of venturefunding (Gompers & Lerner, 2001). We sampledventures founded through to 1995, but continueddata collection for funding rounds occurringthrough to 2003. Since ventures typically take be-tween five and seven years after their first round toexperience an exit event (Fenn, Liang, & Prowse,1997), this extended data collection allowed for amore complete perspective on tie formation byyoung firms.

Our unit of analysis is the funding round. Ven-ture investments are made in a series of discreterounds because investors generally stage their in-vestments to coincide with substantial advances ina venture’s progress (Hallen & Eisenhardt, 2012;Sahlman, 1990). Data were collected for each ven-ture’s funding rounds until an exit event or un-til 2003.

Our key data source was the VentureXpert data-base. These data, collected by the National VentureCapital Association, provide an accurate descrip-tion of U.S. venture financing (Kaplan, Sensoy, &Strömberg, 2002; Lerner, 1995), including detailedinformation about ventures, their VC and corporateinvestors, and funding rounds. The database is alsowell suited to our study in allowing for our unusu-ally long sampling period and for a focus on VC-backed technology ventures, for which CVC inves-


tors often offer both substantial benefits andsubstantial risks.

We triangulated the VentureXpert data on financ-ing rounds with the VentureOne database and withLexisNexis media articles. VentureXpert data weredrawn from investor surveys; VentureOne data,from entrepreneur surveys; LexisNexis data, fromarchived corporate press releases and media cover-age. Using these complementary data sources al-lowed us to draw on multiple informants abouteach event, increasing the accuracy of our data. Weconstructed the sample by first sampling venturerounds from VentureXpert, and then corroboratingrounds and specific investments via VentureOne.We used the same process to find data missing inVentureXpert. For data that remained missing orwhich was inconsistent in the two databases, weexamined media coverage and press releases fromLexisNexis. These efforts added information toroughly 20% of rounds, producing uniquely re-fined and comprehensive data on investments inventures in five industries over the course of25 years.

Our round-level sample is a stratified randomsample of 700 ventures drawn from the populationof high-technology ventures that received initialfunding between 1979 and 1995. The sample rep-resents approximately 11% of U.S. technology ven-tures funded during the time period. We stratifiedour sample by year and by five broad technology-industry groups designated by VentureXpert (“bio-technology,” “electronics,” “communications,”“software,” and “medical products”), and includedabout 140 firms in each industry. The most preva-lent two-digit Standard Industrial Classification(SIC) codes were 73 (21%), 28 (20%), 36 (17%), 38(13%), 35 (9%), and 48 (9%). The sampled venturesraised 18,168 investments, including more than1,200 corporate investments, in 4,073 fundingrounds.

To complement the round-level data on ventures,we collected an extensive dataset on characteristicsof venture investors and their networks. Consistentwith past studies, the network data were collectedfrom VentureXpert, and included all five-year net-work-history windows and all professional inves-tors from 1974 through to 2003. The data included3,622 investors (VC and CVC) based in the UnitedStates and active during this period, and their163,736 venture investments. To triangulate andexpand the data on corporate investors (for dyadanalysis), we used the Compustat, Hoover’s, andSecurities Data Corporation databases. Overall, the

data collection required considerable effort and re-sulted in a unique dataset that is significantlyricher than commercially available datasets.

We also conducted several background inter-views to better understand the context. First, weinterviewed technology entrepreneurs, partners atVC firms and CVC arms of corporations, business-unit managers, a lawyer specializing in technologyventures, and an angel investor. Second, we drewon systematic fieldwork data, tracking both inves-tors and entrepreneurs on the round-by-roundfundraising strategy of nine ventures (Hallen,2007). Finally, we read news articles on corporateinvestments to further understand these relation-ships, and co-taught several master’s and execu-tive-level venture-financing courses with VCpartners.

Measures

Dependent variable. Our dependent variable isthe number of corporate venture investors that par-ticipated in a given round of a young firm’s fund-ing.3 We coded an investor as corporate if it was anonfinancial firm that purchased private equity(excluding loans and public offerings). We ex-cluded subsidiaries of banks and insurance compa-nies in order to focus on CVC investors that couldoffer complementary resources, and which mightalso be likely to engage in misappropriation. Cor-porate VC investors were coded using companydirectories, annual reports, and databases on publiccompanies (such as Compustat and Worldscope).Two researchers independently coded the data,with the aid of a computer program that matchedinconsistent spellings and repeat investments. Weincluded both U.S. and foreign corporate investorsto make our coverage more comprehensive thanthat of studies focused exclusively on U.S. inves-

3 A count variable was used because each additionalcorporate investor increases potential exposure to misbe-havior. Furthermore, CVC investments exhibit a differentand less path-dependent empirical pattern than a ven-ture’s VC investments. Venture capital investors typi-cally invest in follow-on rounds, since prior VC investorstend both to follow on and to bring in new additional VCinvestors to set the valuation. For instance, approxi-mately 43% of the funding rounds in our data includedmore VC investors than the previous round. By contrast,only 4% of the funding rounds had more CVC investorsthan the previous round—and 76% of funding roundshad fewer CVC investors than the previous round.


tors (e.g., Dushnitsky & Lenox, 2005a). Overall, wecoded more than 1,200 corporate investments.

We also recognize that some corporations aremore likely than others to be direct rivals of theyoung firm. Although our study (and the depen-dent variable) focuses on the question of whether ornot relationships form with corporate sharks, het-erogeneity of corporate partners may also influencethe specific partner choice. By conducting dyadanalysis of specific types of corporate partner, wethen started to explore this interesting direction forfuture work.

Independent Variables

Social defenses. We constructed two measuresof social defenses. First, we measured VC centralityby means of the degree of embeddedness of theventure’s VC investors in the field-level formal co-investment network (Hochberg et al., 2007; Walker,Kogut, & Shan, 1997). We used eigenvector central-ity (Bonacich, 1972), which is an attractive measurebecause it weights each tie by the connected VCinvestor’s centrality, and therefore captures therange of an investor’s direct connections and therecursive network reach of those connections. Thuseigenvector centrality captures the desirability (sta-tus) of the VC investor as a current and futuresyndicate partner, and of the threat of discipliningembodied in that investor’s ability to broadcast al-legations of a CVC investor’s misappropriation toothers. We followed Lee et al. (2011) to calculateeigenvector centrality from a binary adjacency ma-trix, placing a tie between two VC investors if theyco-invested during the five-year window ending inthe year prior to the round. This network was con-structed using all investors active during the focalfive-year window. We scaled each year’s centrali-ties between 0 and 1 to allow better inter-year com-parisons. As in prior work, ventures without VCinvestors or with investors unconnected to theprimary network component were assigned a cen-trality of 0 to indicate that they did not have third-party VC ties as social defenses.4 We then calcu-lated the eigenvector centrality scores for all of theventure’s VC investors (that is, VC investors thathad invested in the venture in any past round) and

took the maximum score as the measure (Hallen,2008). Maximum score has several advantages.First, aggregating centralities would probably over-count network reach. Investors new to a deal oftensyndicate with habitual syndication partners, sug-gesting that all VC investors entering a given roundhave redundant network reach. Similarly, using anaverage measure of centrality is undesirable becauseit would dilute the measure if ventures were to haveboth highly central and less central investors.

We measured VC proximity by identifying thegeographical distances between the venture and allof its VC investors in any past round, and taking aninverse of the shortest distance. We used the dis-tance to the closest VC investor because Sorensonand Stuart (2001) found that remote investors oftenrely on those nearby to provide primary monitoringand mentoring to a venture (average distance pro-duced results that were similar). Distance in mileswas calculated using longitudes and latitudes deter-mined from the zip codes of the firms’ headquarters.Low values of this variable (approaching 0) indicatedistant VC investors; high values indicate local VCinvestors. We assigned a value of 0 to ventures with-out VC investors from any past round (akin to distantinvestors), because such ventures would not enjoythe social defenses of a proximate VC investor.

Secrecy defenses. Consistent with prior findingsthat young firms use trade secrets (not patents) as alegal defense against misappropriation (Katila etal., 2008), we measured the strength of trade secretdefenses in a venture’s industry. Trade secrets rep-resent an exclusive source of protection for a di-verse range of intellectual property—from know-how, to recipes, to customer lists—as long as firmsactually keep them secret (Epstein, 2004). Since thestrength of trade secret protection varies across in-dustries, we measured secrecy defense in a ven-ture’s industry using the Carnegie Mellon Survey ofIndustrial R&D, matching industries at the three-digit SIC code level at which the survey was con-ducted (Cohen, Nelson, & Walsh, 2000). These datahave been used extensively in prior studies and areconsidered the primary source on comparative ap-propriability (Arora & Ceccagnoli, 2006; Gulati &Singh, 1998). Furthermore, the collection date ofthese data (1994) falls at the approximate midpointof our time range, 1979–2003. Our specific measureis the percentage of product inventions for whichsurveyed R&D managers in an industry believedtrade secrets to be effective at preventing appropri-ation. In our sample, the values ranged from about34% in electronic components and accessories (SIC

4 Only 1% of our sample ventures’ first rounds in-volved solely a corporate investor and no VC investors.Excluding these firms from the sample did not influenceour results.


367) to almost 62% in metalworking machines andequipment (SIC 354).

Timing defenses. We measured a venture’s tim-ing defense by investment round (first, second, etc.)and logged this measure to reduce skew. Invest-ment round is an appropriate measure because itreflects the technology and customer-developmentmilestones that the venture has reached—that is,those “proofpoints” that make investors more will-ing to invest in a new round and simultaneouslymake misappropriation more difficult (Hallen &Eisenhardt, 2012; Katila & Mang, 2003). To furtherisolate the effects of timing defenses, we also con-trolled for a venture’s age and resource needs, asdescribed in detail below. We additionally testedan alternative measure of timing defenses using anordinal measure of venture development stage ob-tained from the VentureOne database; this alterna-tive measure was available for only about half ofthe rounds, but it confirmed our original results.

Controls

Since it is resource needs that push young firmsto form CVC relationships, we measured a ven-ture’s need for greater capital infusions by roundsize of the focal funding round (in US$000s). Be-cause entrepreneurs determine the size of a roundby balancing their firm’s capital requirementsagainst unnecessary ownership dilution from ex-cessive funding, this is an effective measure of afirm’s capital needs. We used the producer priceindex (PPI) to adjust the variable for inflation andlogged it to mitigate skew.

We also included controls to account for a ven-ture’s need for the complementary resources thatCVC investors provide, using Compustat data. First,we controlled for the need for manufacturing re-sources through manufacturing intensity, measuredas the capital intensity of the venture’s industry, sinceventures in highly capital-intensive industries arelikely to need greater manufacturing assets to pro-duce their products. Our measure of capital intensitywas the average ratio of fixed assets to sales in theventure’s industry in the focal year.

Second, because ventures with high marketingresource needs may seek corporate investors fortheir brand names, market knowledge, and distri-bution channels (Basu, Phelps, & Kotha, 2009;Khaire, 2010), we measured marketing intensity bymeans of the average ratio of advertising expendi-tures to sales in the venture’s industry. Industry-level measures were appropriate because comple-

mentary resource needs tend to be quite consistentin a given industry, but differ across industries(Katila & Shane, 2005), and because it is difficult tocollect venture-level data in a large-sample, multi-decade study such as ours.

We measured both manufacturing and marketingintensity at the industry level, using the granularfour-digit SIC codes.

Because prior work suggests that social connec-tions or similarity between focal partners facilitatetie formation (Gulati & Gargiulo, 1999), we con-trolled for several such factors. We included corpo-rate background as a dummy variable indicatingwhether a venture was a spinoff of a corporation(value � 1), because spinoffs may inherit connec-tions and knowledge from the parent that can makecorporate ties more likely (Agarwal, Echambadi,Franco, & Sarkar, 2004). And since tie formationmay be path-dependent (Schoonhoven & Ro-manelli, 2001), we included prior CVC investors asa binary measure with a value of 1 if any of theventure’s prior rounds included a CVC investor.

To control further for the interest of corporateinvestors in a venture (venture attractiveness), weincluded a control for the venture’s technology as-sets (Stuart, Hoang, & Hybels, 1999), measured bythe number of ultimately successful patent appli-cations that each venture filed yearly. We thenweighted these counts by the forward citations thateach patent received in the seven years subsequentto issue in order to capture differences in the qual-ity of inventions. To account better for the aging ofventures and the related accumulation of experi-ence (beyond that for which we controlled directlywith the timing defense variable), we also controlledfor venture age using data from VentureXpert on thenumber of months between a venture’s founding andthe focal round (logged).

Because prior work suggests that corporations aregenerally less likely to invest in industries in whichpatents are a strong defense (Dushnitsky & Lenox,2005b), we controlled for the strength of patentprotection in a venture’s industry. To assess patentstrength, we consulted the Carnegie Mellon Surveyof Industrial R&D (Cohen et al., 2000) and used asour measure the percentage of product inventionsfor which patents are considered an effective pro-tection mechanism in the industry.

Because the availability of funding may varyacross environments and influence the pursuit ofcorporate investors, we also controlled for theavailability of VC funding and for VC firm densityin the region. We measured VC availability by the


total annual inflation-adjusted investment on thepart of VC firms in our five focal industries, asreported by VentureXpert (in US$ billions). Wemeasured the VC richness of the region as the ven-ture’s location relative to that of many potential VCinvestors. Based on Sorenson and Audia (2000), weconstructed the measure as the sum of the inversedistances in miles between the focal venture i andVC firm j, where j was drawn from the set of allU.S.-based VC investors whose window of activityincluded the year of the focal round (that is, thoseVC investors that invested for the first time eitherbefore or during the focal year and whose most recentinvestment was either during or after the focal year).The more VC firms there were in proximity to theyoung firm, the higher the value of the variable. Dis-tance was calculated using longitudes and latitudesfrom zip codes, and logged to reduce skew:

ln(�[1 ⁄ (dij � 1)] � 1)

Finally, we included dummy variables to controlfor any unobserved industry effects in a venture’sindustry and calendar year dummies to capturepotential fluctuations in the economy that couldinfluence corporate tie formation.

Statistical Method

We used negative binomial regressions becauseour dependent variable, the number of CVC rela-tionships formed in a given funding round, ex-hibited overdispersion. To control for ventureheterogeneity, we used the generalized estimating

equations (GEE) regression method, which ac-counts for autocorrelation that arises because eachventure was included in the sample repeatedlyacross multiple rounds (Liang, Zeger, & Qaqish,1986). In contrast to random effects, the GEEmethod does not require the strong assumption thatunobserved venture-specific effects are uncorre-lated with the regressors. We used the exchange-able correlation structure to capture within-subjectcorrelation, since it permitted us to include ven-tures that raised only one round of investment, andused the negative binomial distribution and linkfunctions. Finally, we also ran Poisson GEE regres-sions with similar results.

RESULTS

Table 1 reports descriptive statistics and correla-tions. Our ventures raised an average of betweenfour and five rounds, with the size of an averageround being US$4 million. Rounds involved, onaverage, four investors. The correlations betweenthe independent variables were low and all vari-ance inflation factors (VIFs) were below the tradi-tional threshold of 10 (the highest being 4.3), sug-gesting that the reported estimates are unlikely tosuffer from multicollinearity bias.

Main Findings

Table 2 presents the results for the GEE negativebinomial regression analysis predicting the numberof CVC investment relationships formed in a round.

TABLE 1Descriptive Statistics and Correlations

Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14

1. Number of CVC investors 0.29 0.642. VC centrality 0.48 0.34 0.073. VC proximity 0.11 0.25 0.01 0.174. Secrecy defense 49.46 5.61 0.08 0.02 0.015. Timing defense 1.24 0.76 0.05 0.66 0.19 0.036. Round size 6.73 2.86 0.19 0.00 �0.02 0.03 �0.087. Venture age 3.56 1.23 0.03 0.41 0.15 �0.03 0.65 �0.038. Manufacturing intensity 0.62 0.47 0.02 �0.12 �0.04 0.05 �0.03 0.06 �0.019. Marketing intensity 0.04 0.04 0.05 �0.05 0.04 0.20 0.01 0.06 �0.02 0.00

10. Corporate background 0.09 0.28 0.05 0.05 �0.01 0.08 0.02 0.02 �0.07 �0.06 �0.0211. Prior CVC investors 0.40 0.49 0.13 0.44 0.11 0.06 0.55 �0.03 0.38 �0.07 �0.06 0.0412. Technology assets 11.73 45.46 0.10 0.10 0.00 0.07 0.05 0.06 0.04 0.00 0.04 0.06 0.0013. Patent strength 40.37 9.79 �0.03 0.01 0.10 0.05 0.00 �0.02 �0.01 �0.02 0.15 �0.11 �0.02 0.0214. VC availability 1.26 4.06 0.07 0.02 0.00 �0.04 0.11 0.17 0.10 0.18 0.04 �0.03 0.04 0.00 �0.0115. VC richness of region 1.91 1.02 0.05 0.05 0.47 0.08 0.09 0.05 0.10 0.09 0.22 �0.05 0.02 0.05 0.11 0.11

Note. Correlations above 0.03 are significant at the p � .05 level.


TABLE 2GEE Negative Binomial Regression Analysis of the Number of Corporate Investors

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

Intercept �1.14*** �1.03*** �2.09*** �1.05*** �1.38*** �1.03*** �0.16**(0.11) (0.12) (0.17) (0.12) (0.16) (0.11) (0.07)

DefensesVC centrality 0.28*** 0.05 0.26*** 0.21** 0.28*** 0.12**

(0.09) (0.10) (0.09) (0.10) (0.09) (0.05)VC proximity �0.08 0.01 �0.08 �0.11 �0.41***

(0.11) (0.12) (0.12) (0.14) (0.09)Secrecy defense 0.02*** 0.01 0.02*** 0.01* 0.02*** 0.02***

(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)Timing defense 0.16*** 0.02 0.15** 0.16** 0.16*** 0.09***

(0.06) (0.06) (0.06) (0.06) (0.06) (0.03)Social defenses � Other defensesVC centrality � VC proximity �0.72* �0.71***

(0.37) (0.19)VC centrality � Secrecy defense �0.00 �0.05***

(0.01) (0.01)VC centrality � Timing defense �0.25** �0.38***

(0.10) (0.05)VC proximity � Secrecy defense 0.01 0.05***

(0.02) (0.01)VC proximity � Timing defense 0.05 0.16*

(0.16) (0.09)Firm controlsRound size 0.15*** 0.14*** 0.14*** 0.14*** 0.13*** 0.14*** 0.01*

(0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.00)Venture age �0.09*** �0.16*** �0.00 �0.15*** �0.13*** �0.15*** 0.02

(0.02) (0.02) (0.03) (0.02) (0.02) (0.02) (0.02)Manufacturing intensity 0.35*** 0.35*** 0.37*** 0.34*** 0.32*** 0.34*** �0.01

(0.08) (0.09) (0.07) (0.09) (0.10) (0.09) (0.07)Marketing intensity �3.38*** �1.78*** �0.27 �1.96*** �1.34** �1.85*** �2.27***

(0.71) (0.57) (0.62) (0.60) (0.67) (0.58) (0.61)Corporate background 0.32*** 0.27*** 0.18** 0.27*** 0.31*** 0.27*** 0.20***

(0.07) (0.07) (0.07) (0.07) (0.08) (0.07) (0.07)Prior CVC investors 0.42*** 0.23*** 0.36*** 0.24*** 0.20*** 0.23*** 0.35***

(0.05) (0.06) (0.06) (0.06) (0.06) (0.06) (0.03)Technology assets 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00**

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)Patent strength 0.00 0.00 0.00 0.00 0.00 0.00 �0.01***

(0.00) (0.01) (0.00) (0.01) (0.01) (0.00) (0.00)Industry controlsVC availability �0.00 �0.00 0.00 �0.00 0.00 �0.00 �0.01

(0.01) (0.01) (0.00) (0.01) (0.01) (0.01) (0.01)VC richness of region 0.05 0.06* 0.06** 0.06* 0.07* 0.06* 0.13***

(0.03) (0.03) (0.03) (0.04) (0.04) (0.04) (0.02)Biotechnology 0.04 �0.01 0.36*** 0.01 0.29** �0.01 �1.05***

(0.10) (0.10) (0.11) (0.10) (0.13) (0.10) (0.08)Electronics 0.07 0.01 0.38*** 0.02 0.26** 0.01 �1.08***

(0.10) (0.11) (0.11) (0.11) (0.13) (0.11) (0.07)Communications �1.29*** �1.11*** �0.74*** �1.09*** �0.85*** �1.11*** �1.69***

(0.26) (0.27) (0.23) (0.28) (0.29) (0.27) (0.24)Software �0.30*** �0.13 0.24** �0.12 0.12 �0.13 �1.24***

(0.11) (0.12) (0.11) (0.12) (0.14) (0.12) (0.10)Year fixed effects Y Y Y Y Y Y Yn 4,073 4,073 4,073 4,073 4,073 4,073 4,073Chi-squared 1,210.58 1,024.39 837.92 997.35 640.80 1,030.22 1,622.79

Note. Standard errors are in parentheses; 700 ventures; 4,073 funding rounds.* p � .10

** p � .05*** p � .01; two-tailed tests


To interpret the interactions, we graphed estimatedCVC relationship formation in Figures 1 and 2 us-ing Model 7 from Table 2. Because multicollinear-ity across main effects and their interaction termscan be an issue, we centered the variables prior tocalculating the interactions (Cronbach, 1987).

Model 1 in Table 2 includes the control variables.We find that more corporate investment relation-ships form when a venture has high financial re-source needs and a prior relationship with a corpo-ration (that is, spinoff or prior corporate rounds). Inline with the strategic motivations of corporate in-vestors, we also find that ventures with better tech-nology assets are likely to have more corporatepartners.

In Hypothesis 1, we argued that ventures withmore central VC investors will form more corporateties. Model 2 adds VC centrality and the secrecyand timing defense variables. As expected, the co-efficients for both the secrecy and timing defensevariables are positive and significant (p � .05 orlower). The coefficient for VC centrality is positiveand significant in Model 2 (p � .01) and in the fullModel 7 (p � .05), supporting the hypothesis.5

In Hypothesis 2 we argued that ventures withnearby VC investors will acquire more corporateinvestors. The coefficient for VC proximity is un-expectedly negative, but not significant in Model 3,and negative and significant in Model 7 (p � .01).We will discuss this unexpected finding later inthe paper.

In terms of effect size, a 1 SD increase in VCcentrality is associated with a 4.2% increase in thenumber of corporate investors in a round.6 To il-lustrate the practical significance of these effects, ifa venture with a relatively low-centrality VC port-folio receives an investment from a high-centralityVC investor (such as Sequoia Capital), its likeli-hood of acquiring a CVC investor in the next roundincreases by more than 10%. Similarly, a 1SD in-crease in VC proximity is associated with a 9.9%decrease in the number of CVCs in a round.

In Hypothesis 3, we proposed that the positiveeffect of central VC investors on corporate relation-ships will be amplified when those VC investorsare also nearby. As with Hypothesis 2, the resultinginteraction is unexpectedly negative and margin-ally significant in Model 4 (p � .10), and negativeand strongly significant (p � .01) in Model 7. Tobetter explain this effect, we illustrate the findingsin Figure 1, which indicates that the effect is asym-metric and that greater proximity actually alterscentrality’s slope from positive to negative whenVC investors are closer than average to the venture.In order to test for the significance of the interac-tions at key levels of VC proximity, we followed theprocedure of Hilbe (2011). The test shows thatthe coefficient for the interaction is significant atthe minimum and maximum levels of proximity(using a p � .05 threshold), but not significant atthe mean level of proximity. Overall, these resultsconfirm that VC investors’ centrality has a positiveeffect on CVC tie formation when VC investors areremote, but that this effect sharply declines andturns negative when those investors are nearby. Interms of practical significance, a shift from low- tohigh-centrality VC investors increases the expectednumber of CVC investors in the next round by 16%if all of the VC investors are very distant geograph-ically. In contrast, if a current VC investor isnearby, the same shift in VC centrality decreasesthe expected number of CVC investors by 34%. Weexpand on this unexpected result later in the paper.

To test Hypotheses 4a–d—that social defensescompensate for traditional defenses—we assessedthe interactions of social defenses with secrecy andtiming defenses in Models 5 and 6, and in the fullModel 7. Our findings again confirm the hypothe-ses for the social defenses of central VC investors.The interactions of VC centrality with secrecy andtiming defenses are negative and significant in thefull model (p � .01; p � .01), and the interactions ofVC proximity with the same variables are signifi-cant, but positive given the negative main effect, inModel 7 (p � .01; p � .1). Overall, the results showthat VC centrality comes into play and influencesthe acquisition of corporate investors particularlywhen legal and timing defenses are weak or un-available. Intriguingly, a similarly hypothesized in-terplay of nearby VC investors and these other de-fenses does not pan out, and the effect of local VCinvestors is rather to pull partners further awayfrom one another when traditional defenses areweak. We return to this unexpected result later inthe paper.

5 We use the full Model 7 of Table 2 to interpret theeffect of centrality because omitting any key explanatoryvariables (such as centrality’s interaction with proximity)can lead to bias in the estimation of the remaining pa-rameters (Kennedy, 1998: 103).

6 Effect size is calculated as the exponent of a 1SDchange in an explanatory variable multiplied by the co-efficient in Model 7 of Table 2-e.g., for centrality,e^ (0.121 � 0.343) � 1.042.


Figure 2 illustrates the interaction effects of Hy-potheses 4a–d. It shows that central VC investorsenable corporate relationships when ventures are atan early stage and lack trade secret protection. Forinstance, a shift from less-central to more-centralVC investors produces a 99% increase in the ex-pected number of CVC investors in the next roundwhen secrecy defenses are low and a 58% increasewhen timing defenses are low. The correspondingeffect is a 9% increase at mean levels of secrecy andtiming.

Further, Figure 2 illustrates that the negative re-lationship between VC proximity and CVC tie for-mation is unexpectedly most pronounced whenventures lack strong timing and secrecy defenses.For example, a shift from remote to nearby VCinvestors decreases the expected number of CVCinvestors in the next round by 59% when secrecydefenses are weakest (versus a 15% increase whensecrecy defenses are strongest). Likewise, the sameshift from remote to nearby VC investors decreasesthe expected number of CVC investors by 45%when timing defenses are weakest (versus an 8%decrease when timing defenses are strongest). Asabove, we verified that all interactions were signif-icant (p � .05) at the various levels depicted inFigure 2, such as the minimum, mean, and maxi-mum levels of secrecy and timing (Hilbe, 2011).

Finally, Figure 2 intriguingly indicates a negativerelationship between VC centrality and more CVCinvestors at very high levels of secrecy or timingdefenses. In contrast, prior literature and our ownarguments suggested a flat (or slightly positive) re-lationship. One likely explanation is that CVC in-vestors may avoid ventures with multiple strong,overlapping defenses in the belief that expropriat-ing value from such a relationship would be toodifficult. Overall, the results confirm that socialdefenses—and central VC ties in particular—stepin to influence tie formation with corporate sharkswhen legal and timing defenses are weak.

Additional Analyses

Corporate partner heterogeneity. Our core hy-potheses center on the question of whether ven-tures form ties with sharks, but it is also intriguingto ask with whom they form such relationships. Inorder to examine which specific corporate partnersare likely to be chosen, we conducted a dyad-levelanalysis to ask why a tie formed between a partic-ular pair of firms and not between another pair(Table 3). This analysis focused on better under-standing the heterogeneity of corporate partners—that is, how social defenses influence engagingwith more or less dangerous corporate investors

FIGURE 1Interaction of VC Centrality and VC Proximity (Hypothesis 3)

0.00

0.05

0.10

0.15

0.20

0.25

0 0.2 0.4 0.6 0.8 1

Est

imat

ed N

um

ber

of C

VC

s

VC Centrality (Min to Max)

Max Proximity

Mean Proximity

Min Proximity

Note. All estimates are based on Model 7 in Table 2. Estimates are at the mean value of all variables, except those reported as varying.The vertical axis has been positioned to indicate the mean level of VC centrality.


(Diestre & Rajagopalan, 2012). The results of theseadditional analyses confirm that our main findingson social defenses are robust.

In our analysis, we included both venture–CVCdyads that actually formed and all dyads that couldhave formed, but did not, in a particular round.Dyads were created using our sample ventures andthe set of all public U.S. corporate investors thatinvested in any of the industries that we examined,resulting in 11,132 dyad observations in total. Toestimate the dyad models, we used a conditionallogit regression grouped by round.

To examine the influence of CVC heterogeneityon tie formation, we first looked at the situation inwhich the corporate partner is potentially danger-ous because it possesses the technical abilities touse and misappropriate the young firm’s technol-ogy. Because highly R&D-intensive corporationstend to have the technical abilities to better absorba young firm’s innovations, we interacted corporateR&D expenditures with social defenses. The coeffi-cient for corporate R&D expenditures is positiveand significant, as expected, but the interactionswith VC centrality and VC proximity are not. Theresults thus show that although technology-inten-sive corporations are more likely to engage in a

relationship, social defenses do not seem to com-pensate for the potential danger posed by suchpartners.

We then examined the situation in which thecorporation is potentially more dangerous becauseits market incentives to misappropriate the youngfirm’s technology are higher. We measured marketincentives by means of a dummy variable that in-dicated whether the venture and the corporationoperated in the same industry (business related-ness) because the risk of opportunism is greater insuch situations (Dushnitsky & Shaver, 2009). Here,the results confirmed that same-industry sharks areparticularly likely to engage in relationships and,most intriguingly, that social defenses compensatefor the potential danger posed by these same-indus-try partners—that is, that relationships with moredangerous same-industry sharks are particularlylikely when a venture is backed by more central VCinvestors and more unlikely when the VC investorsare nearby. The results thus strongly corroborateour main findings on the influence of social de-fenses on tie formation with sharks. Notably, theseresults hold when we control for prior dyadic rela-tionships between a particular VC investor and aparticular CVC investor (that is, closure of the VC–

FIGURE 2Interactions of Social and Other Defenses (Hypotheses 4a–d)

0.000.050.100.150.200.250.300.35

0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1Est

imat

ed N

um

ber

of C

VC

s


a: VC Centrality x Secrecy Defense

Max Secrecy

Mean Secrecy

Min Secrecy

0.000.050.100.150.200.250.300.35

Est

imat

ed N

um

ber

of C

VC

s

VC Proximity (Min to Max)

c: VC Proximity x Secrecy Defense

Max Secrecy

Mean Secrecy

Min Secrecy

0.000.050.100.150.200.250.300.35

Est

imat

ed N

um

ber

of C

VC

s


b: VC Centrality x Timing Defense

Max Timing

Mean Timing

Min Timing

0

0.05

0.1

0.15

0.2

0.25

Est

imat

ed N

um

ber

of C

VC

s

VC Proximity (Min to Max)

d: VC Proximity x Timing Defense

Max Timing

Mean Timing

Min Timing

Note. All estimates are based on Model 7 in Table 2. Estimates are at the mean value of all variables, except those reported as varying.The vertical axis has been positioned to indicate the mean level of VC centrality/VC proximity.


CVC–venture triad). Overall, these analyses showthat when corporations have strong market incen-tives to misappropriate (that is, when they are es-pecially shark-like), third-party social defensesmatter even more.

Social defenses. We also tried several alternativemeasures of VC centrality (and the related disci-plining ability) to assess the robustness of our re-sults. For example, the results were robust to: (a)the number of ventures that the VC investor hadtaken public in the preceding five years; and (b) thenumber of ventures in which it had invested (Lee etal., 2011), both of which capture the VC investor’sdesirability as a future syndicate partner and ability

to broadcast misbehavior within co-investmentnetworks.

We also ruled out the possibility that an endorse-ment effect of central VC investor ties provides analternative explanation for our findings. If this werethe case, we would expect a positive interaction ofthe traditional defenses with the VC-centrality de-fense in Table 2. This would be because ventureswould enjoy the protection of the other defensesplus the ability to better attract CVC investors as aresult of a central VC investor’s endorsement. Theobserved negative interactions in Table 2 fail tosupport this alternative explanation and are consis-tent with VC centrality as a social defense.

TABLE 3Conditional-Logit Analysis of the Likelihood of Tie in a Dyad Round†

Model 1 Model 2 Model 3 Model 4

Dangerousness of corporate investorCorporate R&D (ln) 0.07** 0.08** 0.08** 0.08**

(0.03) (0.03) (0.03) (0.03)Same-industry CVC investor 0.57*** 0.57*** 0.56*** 0.56***

(0.14) (0.14) (0.14) (0.14)Hypothesized interactionsVC centrality � Corporate R&D (ln) 0.11 0.12

(0.10) (0.10)VC proximity � Corporate R&D (ln) 0.13 0.14

(0.14) (0.14)VC centrality � Same-industry CVC investor 0.78* 0.83**

(0.41) (0.41)VC proximity � Same-industry CVC investor �0.90* �0.93*

(0.54) (0.55)ControlsCorporate-venture distancea (ln) �0.05** �0.05** �0.05** �0.05**

(0.02) (0.02) (0.02) (0.02)Prior corporate-venture tie 4.79*** 4.79*** 4.82*** 4.82***

(0.22) (0.22) (0.22) (0.22)Strength of common third-party tiesb 0.68*** 0.68*** 0.69*** 0.69***

(0.16) (0.16) (0.16) (0.16)VC centrality � Strength of common third-party ties �1.53*** �1.59*** �1.51*** �1.57***

(0.37) (0.38) (0.37) (0.38)Joint centralityc (CVC–VC) 2.91*** 2.58*** 3.04*** 2.72***

(0.83) (0.85) (0.83) (0.86)CVC centrality �1.92** �1.68** �2.09*** �1.85**

(0.80) (0.82) (0.80) (0.82)Chi-squared 953.71 956.30 959.61 962.56

Note. Standard errors are in parentheses; n � 11,132 dyads.a Geographical distance between a venture and a corporate investor in miles.b Measured by number of CVC–VC ties over prior five years.c Accounts for mutual attraction of central VC investors and central CVC investors; calculated following Gulati and Gargiulo (1999).† We exploit the matched nature of the sample design, utilizing a fixed effect for each venture-round case. This approach produces

unbiased and efficient estimates of effects conditional on both round-level factors and the decision to accept CVC. The tradeoff is that weare unable to include variables that are invariant at the dyad level (VC proximity, VC centrality).

* p � .10** p � .05

*** p � .01; two-tailed tests


We also conducted additional analyses to ex-plore our unexpected findings on VC proximity.While we cannot, of course, completely reject thepossibility that VC proximity has “the wrong sign”in our results (a negative, rather than the hypothe-sized positive, effect), several tests suggest that thisis unlikely. As described in detail below, weruled out problems related to multicollinearityand measurement, and omitted variables thatmay all cause a change in the sign of a coefficient(Kennedy, 2005).

We first checked whether collinearity is an issueby obtaining the VIFs for all of the variables (Men-ard, 2002). All were less than the recommendedcutoff value (for example the VIF for VC proximitywas 1.90), thus allaying concern about multicol-linearity. We then examined the correlations ofproximity with the other variables, but all were low(Table 1). The only exception is the moderate(r � .47) correlation between proximity and the VCrichness of the region variables. Consequently, wemade an effort to reduce multicollinearity by re-moving the common variance in the covariate (thatis, the VC richness of the region) by residualizing:we predicted VC richness of the region by VC prox-imity and used the residual as an instrument inplace of the original VC richness of the region con-trol in a separate analysis. We also omitted theother explanatory variables that may be collinear(such as VC centrality), but VC proximity alone alsohad a negative coefficient. Overall, these tests con-firmed our original results on VC proximity.

We also ruled out measurement issues. The re-sults were robust to using alternative versions ofVC proximity (for example, when we replaced amaximum score of the variable with the mean) andto an alternative sampling method (dyad analysis—see Table 3). We also added a binary control vari-able that indicated whether or not the venture waslocated in Silicon Valley or the Boston area (sincesuch ventures have a large number of potential VCinvestors nearby) and may have different decision-making calculus regarding CVC investors, whichcould influence our results. The original results onVC proximity persisted. Additionally, because theeffect may be primarily driven by situations inwhich the most central VC investors are also themost proximate, we added a three-way interactionof centrality and proximity with a new binary vari-able that took a value of 1 if the most central VCinvestor was also proximate (that is, based within40 miles of the venture, equivalent to the distance

between Silicon Valley and San Francisco). Ouroriginal results were again supported.

We then turned to examining our underliningassumptions about VC proximity and the relatedaligning logic of social defenses. An alternative ex-planation that is parsimonious with our findings isthat VC partners are cautious about corporate in-vestors in general and that proximate VC investorsare likely to be particularly influential in steeringventures away from CVC investors. Like parentswho have more control over their kids when theystill live at home than when they are away at col-lege, local VC investors may be better able to mon-itor and advise young firms, and perhaps to stopthem from doing something that the investor per-ceives as ill-advised. Greater VC proximity may thus,as theorized, transform the effect of VC centrality byproducing an aligning defense, but this aligning de-fense may, on aggregate, push young firms away from(rather than toward) CVC investors.

To further examine this alternative argument, weregressed the effects of VC proximity and its inter-action with centrality on the likelihood of addinganother VC (rather than a CVC) investor. Indeed,rather than the negative interaction coefficient inthe original models predicting CVC relationships,the coefficient for the interaction predicting an-other VC relationship was now positive and signif-icant at the p � .1 level, supporting our revisedtheory that local VC investors may indeed pushyoung firms toward VC, rather than CVC, partners.

We also analyzed our qualitative evidence fromfieldwork to further refine this new theory. Ourinterviews indicated that many VC investors didindeed have reservations about corporations asventure investors. “They [VC investors] believe thatbringing in a CVC [investor] . . . isn’t helpful, be-cause it may give you a label of being in the pocketof the corporation,” said one interviewee. “There isalso an element of, ‘Don’t know you; not preparedto do a deal with you.’” A VC investor furtherasserted, “We do not want the venture to become [abig corporation’s] R&D arm.”

The therapeutics venture Neurex provides a goodillustration of such aligning effects. Based in MenloPark, California, Neurex was initially funded by theVC firm Biovest Partners and Sutter Hill Ven-tures—the latter being both local and central—andin a second round by the VC firms Accel, Mayfield,New Enterprise Associates, Sequoia, and Greylock.Consistent with our finding that ventures with cen-tral and proximate VC investors generally avoidCVC investors, Neurex continued to raise funds


solely from VC investors in subsequent rounds un-til its initial public offering in 1993. By contrast,Vivid Semiconductor in Chandler, Arizona, tookits initial funding in 1995 from the distant, butcentral, VC firms Institutional Venture Partnersand Kleiner Perkins Caufield & Byers (both ofMenlo Park). It later added corporate investorAmkor Electronics, illustrating the positive influ-ence of centrality when VC investors are distant.Overall, evidence suggests that many VC inves-tors reinforce their preference to limit CVC in-vestments in young firms that are nearby, thusfurther corroborating the VC-proximity effect thatwe found.

DISCUSSION

Prior literature has drawn attention to the “swim-ming with sharks” dilemma—that is, the fact thatpotential corporate partners are both particularlyattractive and particularly dangerous. This paperhas looked at young firms’ formation of ties withthese sharks by elaborating both how resource de-pendencies push partners into relationships, andhow rivalrous misappropriation and related risks ofpower imbalance pull them away. Our contribu-tions have been: (a) to unpack the role of socialdefenses as a particularly accessible strategy foryoung firms to protect against potentially opportu-nistic partners; and (b) to examine the interplaybetween social and other types of defenses.

Contributions to Resource-Dependence Theory

Our core contribution is to resource-dependencetheory (Hillman et al., 2009; Pfeffer & Salancik,1978). We extend the theory by highlighting re-source-dependence and power-balance consider-ations at the pivotal juncture of tie formation,rather than during the relationship, as is typical ofresource-dependence work. In particular, our workcontributes to the emerging literature on tie forma-tion by low-power firms. The motivation to studylow-power firms and their strategies follows fromthe insight that many traditional resource-depen-dence strategies, including mergers, joint ventures,and partnering with trusted local partners, are un-available to low-power firms, either because theirpowerful partners are likely to resist power-balanc-ing operations (Casciaro & Piskorski, 2005) or be-cause they lack a history of trustworthy long-termrelationships. By looking at social defenses thatthird-party VC investors provide, we thus respond

to the calls of Wry, Cobb, and Aldrich (2013: 445) toleverage resource-dependence theory’s insightsabout power “beyond dyadic relationships,” and tore-examine the overlooked “complex and intercon-nected environments that Pfeffer and Salancik[originally] theorized [about]” (Wry et al. 2013:463). Overall, our approach harkens back to theoriginal concept of power networks outlined inEmerson (1962): social defenses provide protec-tions by enabling low-power firms effectively toborrow the power of powerful third parties intheir environments.

We also extend network theory—and the social-structural perspective on dealing with cooperationproblems in particular—by unpacking the charac-teristics of third parties that give rise to varyingforms of social defenses. First, we document therole of centrally positioned partners as a powerfulsocial defense. Our findings indicate that the dis-ciplining defense made possible by central VC in-vestors is unique among social defenses in that itpushes young firms to expand their portfolios ofrelationships to include corporate partners. Theunderlying mechanism here is that the threat ofdisciplining increases the high-power partner’s(CVC investor’s) motivational commitment to therelationship, because corporate partners have moreto lose if they behave opportunistically in the pres-ence of central third parties.

Our results also confirm the significance of VCinvestors’ geographical proximity and of the relatedaligning defense, but in an unexpected way: wefind that a young firm’s proximity to its VC inves-tors sharply reduces the number of corporatesharks with which it will partner, even when theVC investors are also central in their networks. Ourfurther analyses suggest that, rather than promot-ing CVC investors, local VC investors often causeventures to shun corporate investors in favor ofmore path-dependent portfolios—that is, “betteraligned” VC investors. Viewed through the lensof resource-dependence theory, this finding sug-gests that ventures may have a more balancedpower relationship with new VC investors thanwith new CVC investors, and that existing VCinvestors that are nearby are able to reinforcetheir preferences about future corporate investorsmore strongly than VC investors that are geo-graphically distant. Altogether, our findings in-dicate that aligning is a broader mechanism thanwas previously thought, and that it operates by(a) finding alternative sources for required re-sources (VC investors), and— even more signifi-


cantly—(b) adjusting goals so as to depend lesson certain resources (that is, the resources offeredby a CVC investor). Overall, while the impliciteffect of disciplining by central VC investors is tomake corporate investors more dependent on theventure relationship (by introducing a reputationthreat), the contrary effect of aligning keepsyoung firms from becoming too dependent oncorporate partners.

Further, our finding that aligning (proximity) ef-fects overshadow disciplining (centrality) effects inour setting is intriguing. It suggests the existence ofboundary conditions for social defenses and fortheir effectiveness at changing power dynamics.Prior work on disciplining, at both the dyadic andfield levels, tends to concentrate on the extremes ofsingle and highly repetitive relationships (and therelated strong disciplining effect), overlooking themiddle ground. But many CVC investors partici-pate in venture investing only occasionally7—andthus both they and VC investors may be unsureabout the level of repetition, thus influencing therelative threat that disciplining provides. For fu-ture work, sharks relationships offer a promisingpath to richer understanding of social defenses,their boundary conditions, and the conditionsunder which disciplining or aligning is likely todominate.

Our contributions to understanding resource de-pendence also have limitations that offer opportu-nities for future work. One limitation is that weexamine only one type of social defense, VC ties,although these investments account for only a frac-tion of a young firm’s possible relationships. Futureresearch on ventures’ other interorganizational re-lationships, such as strategic alliances and industryconsortia (see Alexy, George, & Salter, 2013;Prashantham & Birkinshaw, 2008; Wang & Zajac,2007), may uncover other sources and kinds ofdefenses, especially because each type of relation-ship is likely to be associated with a different un-derlying network structure (see Rosenkopf & Schil-ling, 2007). Another limitation is our focus onformation of relationships. Future work could con-tribute by comparing the efficacy of social defenseswith that of other defenses in enabling high-per-forming ties (see Tucci, Ginsberg, & Hasan, 2011).Because corporate venture relationships are per-ceived as relatively distant and exploratory (and

less path-dependent) ties that enable both parties totranscend local search, analyzing the contributionsof these ties to such performance outcomes asbreakthrough innovation would be an interestingavenue for future work (see Katila & Chen, 2008;Tushman & Anderson, 1986).

Another useful path for future work is to examinedisciplining and aligning as resource-dependencetactics more generally, in other types of situation.Beyond entrepreneurship, such defenses may berelevant in multiple situations in which the otherparty holds virtually all of the cards and thus fewtraditional resource-dependence strategies, includ-ing mergers, joint ventures, and board interlocks,are apt to be effective (for example, a smaller firmsupplying goods to Walmart or Amazon, or non-profits interacting with major donors). Such situa-tions are important to study: we are only beginningto understand how and with what tactics low-power organizations can attempt to level the play-ing field in these situations.

Contributions beyond Resource Dependence

Further, we show that social defenses are partic-ularly important when other defenses are unavail-able. Research has typically examined secrecy andtiming defenses in isolation from social defenses,but we unite these previously separate perspec-tives. First, we find that the effects of social de-fenses on tie formation are significant even whenwe control for more traditional defenses. Second,and even more significantly, we find that socialdefenses may take the place of traditional defenseswhen the latter are weak or nonexistent. This find-ing suggests an equifinality of defense mechanisms,such that firms with weak trade secret protection orlow timing defenses might instead gain protectionfrom social defenses, and vice versa. In contrast,the literature on social networks often highlightsthe path-dependent, “rich get richer” effects of ex-isting (and prominent) connections (Argote & Mi-ron-Spektor, 2011; Gulati & Gargiulo, 1999; Sine,Shane, & DiGregorio, 2003). Our theory and find-ings offer a different path, suggesting that thirdparties are most beneficial to—and influential on—firms that are otherwise more vulnerable. Thusthird-party social defenses have an equalizingeffect in helping less-privileged firms to gainresources.

We also contribute to understanding how entre-preneurs mobilize resources by building a portfolioof diverse ties (see Mitchell & Singh, 1992). Prior

7 In our data, CVC investors invest on average in onlyone of every four funding rounds.


literature demonstrates that the resource-mobiliza-tion strategies of young firms are strongly influ-enced by path-dependency (exploitation) and thatresources are obtained from prior employers orfrom other local, trusted partners. Our findingsshow, by contrast, that social defenses via earlythird-party ties may help young firms to acquireresources from diverse others in the environment(exploration). Thus whereas prior literature hasemphasized the importance of early connectionsfor gaining legitimacy (Stuart et al., 1999), our find-ings show that early connections (especially prom-inent ones) are also critical in engaging new part-ners with different resources (Katila & Ahuja,2002). In this way, our findings offer an expandedunderstanding of the importance of network ties toyoung firms and rich insight into the origins ofresource mobilization by entrepreneurial firms.

CONCLUSION

Tension between collaboration value and com-petitive misalignment is ubiquitous in interorgani-zational relationships. Resource-dependence theo-rists have identified a range of mechanisms formanaging such tensions, but it is less clear howyoung firms can protect themselves early on, whenrelationships form. Our research specifies the typesof third-party defenses that are particularly rele-vant for young firms entering into relationshipswith corporate sharks, and thus offers both a richerand more precise perspective on social defensesand a holistic perspective that integrates social,legal, and timing defenses.

REFERENCES

Agarwal, R., Croson, R., & Mahoney, J. 2010. The role ofincentives and communication in strategic alliances:An experimental investigation. Strategic Manage-ment Journal, 31: 413–437.

Agarwal, R., Echambadi, R., Franco, A. M., & Sarkar,M. B. 2004. Knowledge transfer through inheritance:Spin-out generation, development, and survival.Academy of Management Journal, 47: 501–522.

Ahuja, G. 2000a. Collaboration networks, structuralholes, and innovation: A longitudinal study. Admin-istrative Science Quarterly, 45: 425–455.

Ahuja, G. 2000b. The duality of collaboration: Induce-ments and opportunities in the formation of inter-firm linkages. Strategic Management Journal, 21:317–343.

Alexy, O., George, G., & Salter, A. 2013. Cui bono? The

selective revealing of knowledge and its implica-tions for innovative activity. Academy of Manage-ment Review, 38: 270–291.

Argote, L., & Miron-Spektor, E. 2011. Organizationallearning: From experience to knowledge. Organiza-tion Science, 22: 1123–1137.

Arora, A., & Ceccagnoli, M. 2006. Patent protection, com-plementary assets, and firms’ incentives for technol-ogy licensing. Management Science, 52: 293–308.

Bae, J., & Gargiulo, M. 2004. Partner substitutability, al-liance network structure, and firm profitability inthe telecommunications industry. Academy of Man-agement Journal, 47: 843–859.

Basu, S., Phelps, C., & Kotha, S. 2009. Towards under-standing who makes corporate venture capital in-vestments and why. Journal of Business Venturing,26: 153–171.

Benjamin, B. A., & Podolny, J. M. 1999. Status, quality,and social order in the California wine industry.Administrative Science Quarterly, 44: 563–589.

Bitler, M. P., Moskowitz, T., & Vissing-Jørgensen, A.2005. Testing agency theory with entrepreneur effortand wealth. Journal of Finance, 60: 539–576.

Bonacich, P. 1972. Factoring and weighting approachesto status scores and clique identification. Journal ofMathematical Sociology, 2: 113–120.

Bonacich, P. 1987. Power and centrality: A family ofmeasures. American Journal of Sociology, 92:1170–1182.

Burt, R. S. 2005. Brokerage and closure: An introduc-tion to social capital. New York: Oxford UniversityPress.

Casciaro, T., & Piskorski, M. J. 2005. Power imbalance,mutual dependence, and constraint absorption: Acloser look at resource dependence theory. Admin-istrative Science Quarterly, 50: 167–199.

Clark, E., & Nelson, D. 1997. Young whale sharks, Rhin-codon typus, feeding on a copepod bloom near LaPaz, Mexico. Environmental Biology of Fishes, 50:63–73.

Cohen, W. M., Nelson, R. R., & Walsh, J. P. 2000. Pro-tecting their intellectual assets: Appropriabilityconditions and why US manufacturing firms pat-ent (or not). NBER Working Paper No. 7552, Na-tional Bureau of Economic Research, Cambridge,MA.

Cox, E., Katila, R., & Eisenhardt, K. 2014. Who takes youto the dance? How funding partners influence in-novative activity in young firms. Working Paper,University of Washington, Washington, DC, andStanford University, Stanford, CA.

Cronbach, L. J. 1987. Statistical tests for moderator vari-


ables: Flaws in analyses recently proposed. Psycho-logical Bulletin, 102: 414–417.

Diestre, L., & Rajagopalan, N. 2012. Are all “sharks”dangerous? New biotechnology ventures and partnerselection in R&D alliances. Strategic ManagementJournal, 33: 1115–1134.

Dushnitsky, G., & Lenox, M. J. 2005a. When do incum-bents learn from entrepreneurial ventures? Corpo-rate venture capital and investing firm innovationrates. Research Policy, 34: 615–639.

Dushnitsky, G., & Lenox, M. J. 2005b. When do firmsundertake R&D by investing in new ventures? Stra-tegic Management Journal, 26: 947–965.

Dushnitsky, G., & Lenox, M. J. 2006. When does corpo-rate venture capital investment create firm value?Journal of Business Venturing, 21: 753–772.

Dushnitsky,G.,&Shaver, J.M.2009.Limitations to interor-ganizational knowledge acquisition: The paradox ofcorporate venture capital. Strategic ManagementJournal, 30: 1045–1064.

Emerson, R. M. 1962. Power-dependence relations.American Sociological Review, 27: 31–41.

Epstein, R. A. 2004. The constitutional protection oftrade secrets under the takings clause. University ofChicago Law Review, 71: 57–73.

Fenn, G. W., Liang, N., & Prowse, S. 1997. The privateequity market: An overview. Financial Markets, In-stitutions and Instruments, 6: 1–106.

Garg, S. 2013. Board-level strategic decision making: TheCEO’s perspective. Academy of Management Re-view, 38: 90–108.

Gompers, P. A., & Lerner, J. 2001. The money of inven-tion: How venture capital creates new wealth. Bos-ton: Harvard Business School Press.

Graebner, M. E., & Eisenhardt, K. M. 2004. The seller’sside of the story: Acquisition as courtship and gov-ernance as syndicate in entrepreneurial firms. Ad-ministrative Science Quarterly, 49: 366–403.

Gulati, R. 1995a. Does familiarity breed trust? The impli-cations of repeated ties for contractual choice inalliances. Academy of Management Journal, 38:85–112.

Gulati, R. 1995b. Social structure and alliance formationpatterns: A longitudinal analysis. AdministrativeScience Quarterly, 40: 619–652.

Gulati, R. 2007. Managing network resources: Alli-ances, affiliations and other relational assets. Ox-ford, U.K.: Oxford University Press.

Gulati, R., & Gargiulo, M. 1999. Where do interorganiza-tional networks come from? American Journal ofSociology, 104: 1439–1493.

Gulati, R., & Singh, H. 1998. The architecture of cooper-

ation: Managing coordination costs and appropria-tion concerns in strategic alliances. AdministrativeScience Quarterly, 43: 781–784.

Hallen, B. L. 2007. The origin of network positions: Howentrepreneurs raise funds. Doctoral Dissertation,Stanford University, Stanford, CA.

Hallen, B. L. 2008. The causes and consequences of theinitial network positions of new organizations: Fromwhom do entrepreneurs receive investments? Ad-ministrative Science Quarterly, 53: 685–718.

Hallen, B., & Eisenhardt, K. M. 2012. Catalyzing strate-gies and efficient tie formation: How entrepreneurialfirms obtain investment ties. Academy of Manage-ment Journal, 55: 35–70.

Heeley, M. B., Matusik, S. F., & Jain, N. 2007. Innovation,appropriability, and the underpricing of initial pub-lic offerings. Academy of Management Journal, 50,50: 209–225.

Hellmann, T. 2000. Venture capitalists: The coaches ofSilicon Valley. Working Paper, Stanford GraduateSchool of Business, Stanford, CA.

Hilbe, J. M. 2011. Negative binomial regression. Cam-bridge, U.K.: Cambridge University Press.

Hillman, A. J., Withers, M. C., & Collins, B. J.. 2009.Resource Dependence Theory: A Review. Journal ofManagement, 35: 1404–1427.

Hochberg, Y. V., Ljungqvist, A., & Lu, Y. 2007. Whomyou know matters: Venture capital networks andinvestment performance. Journal of Finance, 62:251–301.

Hovland, C. I., Janis, I. L., & Kelley, H. 1953. Communi-cation and persuasion. New Haven, CT: Yale Uni-versity Press.

Kale, P., Singh, H., & Perlmutter, H. 2000. Learning andprotection of proprietary assets in strategic alliances:Building relational capital. Strategic ManagementJournal, 21: 217–237.

Kaplan, S. N., Sensoy, B. A., & Strömberg, P. 2002. Howwell do venture capital databases reflect actualinvestments? Working Paper, University of ChicagoGraduate School of Business, Chicago.

Kaplan, S. N., & Strömberg, P. 2004. Characteristics, con-tracts, and actions: Evidence from venture capitalistanalyses. Journal of Finance, 59: 2177–2210.

Katila, R., & Ahuja, G. 2002. Something old, somethingnew: A longitudinal study of search behavior andnew product introduction. Academy of Manage-ment Journal, 45: 1183–1194.

Katila, R., & Chen, E. 2008. Effects of search timing onproduct innovation: The value of not being in syncwith rivals. Administrative Science Quarterly, 53:593–625.


Katila, R., & Mang, P. Y. 2003. Exploiting technologicalopportunities: The timing of collaborations. Re-search Policy, 32: 317–332.

Katila, R., Rosenberger, J. D., & Eisenhardt, K. M. 2008.Swimming with sharks: Technology ventures, de-fense mechanisms and corporate relationships. Ad-ministrative Science Quarterly, 53: 295–332.

Katila, R., & Shane, S. 2005. When does lack of resourcesmake new firms innovative? Academy of Manage-ment Journal, 48: 814–829.

Kennedy, P. E. 1998. A guide to econometrics. Oxford,U.K.: Blackwell.

Kennedy, P. E. 2005. Oh no! I got the wrong sign! Whatshould I do? Journal of Economic Education, 36:77–92.

Khaire, M. 2010. Young and no money? Never mind: Thematerial impact of social resources on new venturegrowth. Organization Science, 21: 168–185.

Lee, P. M., Pollock, T. G., & Jin, K. 2011. The contingentvalue of venture capitalist reputation. Strategic Or-ganization, 9: 33–69.

Lerner, J. 1995. Venture capitalists and the oversight ofprivate firms. Journal of Finance, 50: 301–318.

Levin, R. C., Klevorick, A. K., Nelson, R. R., Winter, S. G.,Gilbert, R., & Griliches, Z. 1987. Appropriating thereturns from industrial research and development.Brookings Papers on Economic Activity, 3: 783–831.

Levinthal, D. A., & March, J. G. 1993. The myopia oflearning. Strategic Management Journal, 14: 95–112.

Li, D., Eden, L., Hitt, M., & Ireland, D. 2008. Friends,acquaintances, or strangers? Partner selection inR&D alliances. Academy of Management Journal,51: 315–334.

Liang, K. Y., Zeger, S. L., & Qaqish, B. 1986. Longitudinaldata analysis using generalized linear models. Bi-ometrika, 73: 13–22.

Lohr, S. 2010. In a partnership of unequals, a start-upsuffers. The New York Times, July 18.

Magee, J., & Galinsky, A. 2008. Social hierarchy: Theself-reinforcing nature of power and status. Acad-emy of Management Annals, 2: 351–398.

Menard, S. 2002. Applied logistic regression. ThousandOaks, CA: Sage.

Mitchell, W., & Singh, K. 1992. Incumbents’ use of pre-entry alliances before expansion into new technicalsubfields of an industry. Journal of Economic Be-havior and Organization, 18: 347–372.

Ozcan, P., & Eisenhardt, K. 2009. Origin of alliance port-folios: Entrepreneurs, network strategies, and firm

performance. Academy of Management Journal,52: 246–279.

Park, H. D., & Steensma, H. K. 2012. When does corporateventure capital add value for new ventures? Strate-gic Management Journal, 33: 1–22.

Pfeffer, J. 1972. Size and composition of corporate boardsof directors: The organization and its environment.Administrative Science Quarterly, 17: 218–228.

Pfeffer, J., & Salancik, G. R. 1978. The external control oforganizations: A resource dependence perspec-tive. Stanford, CA: Stanford University Press.

Podolny, J. M. 1993. A status-based model of marketcompetition. American Journal of Sociology, 98:829–872.

Podolny, J. M. 2001. Networks as the pipes and prisms ofthe market. American Journal of Sociology, 107:33–60.

Pollock, T. G. 2004. The benefits and costs of underwrit-ers’ social capital in the U.S. IPO market. StrategicOrganization, 2: 357–388.

Pollock, T. G., Chen, G., Jackson, E. M., & Hambrick, D. C.2010. How much prestige is enough? Assessing thevalue of multiple types of high-status affiliates foryoung firms. Journal of Business Venturing, 25:6–23.

Pollock, T. G., & Gulati, R. 2007. Standing out from thecrowd: The visibility-enhancing effects of IPO-re-lated signals on alliance formation by entrepreneur-ial firms. Strategic Organization, 5: 339–372.

Portes, A., & Sensenbrenner, J. 1993. Embeddedness andimmigration: Notes on the social determinants ofeconomic action. American Journal of Sociology,98: 1320–1350.

Prashantham, S., & Birkinshaw, J. 2008. Dancing withgorillas: How small firms can effectively partnerwith MNCs. California Management Review, 51:6–23.

Raub, W., & Weesie, J. 1990. Reputation and efficiency insocial interactions: An example of network effects.American Journal of Sociology, 96: 626–654.

Rosenkopf, L., & Schilling, M. A. 2007. Comparing alli-ance network structure across industries: Observa-tions and explanations. Strategic EntrepreneurshipJournal, 13: 191–209.

Sahlman, W. A. 1990. The structure and governance ofventure-capital organizations. Journal of FinancialEconomics, 27: 473–521.

Sapienza, H. J. 1992. When do venture capitalists addvalue? Journal of Business Venturing, 7: 9–27.

Sapienza, H. J., & Gupta, A. 1994. Impact of agency risksand task uncertainty on venture capitalists–CEO in-


teraction. Academy of Management Journal, 37:1618–1632.

Schoonhoven, C., & Romanelli, E. (Eds.). 2001. The en-trepreneurship dynamic: Origins of entrepreneur-ship and the evolution of industries. Stanford, CA:Stanford University Press.

Sine, W., Shane, S., & DiGregorio, D. 2003. The haloeffect and technology licensing: The influence ofinstitutional prestige on the licensing of universityinventions. Management Science, 49: 478–496.

Soda, G., Usai, A., & Zaheer, A. 2004. Network memory:The influence of past and current networks on per-formance. Academy of Management Journal, 47:893–906.

Sørensen, M. 2007. How smart is smart money? A two-sided matching model of venture capital. Journal ofFinance, 62: 2725–2762.

Sorenson, O., & Audia, P. G. 2000. The social structure ofentrepreneurial activity: Geographic concentrationof footwear production in the United States, 1940–1989. American Journal of Sociology, 106: 424–461.

Sorenson, O., & Stuart, T. E. 2001. Syndication networksand the spatial distribution of venture capital invest-ments. American Journal of Sociology, 106: 1546–1588.

Stuart, T. E., Hoang, H., & Hybels, R. C. 1999. Interorgan-izational endorsements and the performance ofentrepreneurial ventures. Administrative ScienceQuarterly, 44: 315–349.

Teece, D. J. 1986. Profiting from technological innova-tion: Implications for integration, collaboration, li-censing and public policy. Research Policy, 15:285–305.

Tucci, C., Ginsberg, A., & Hasan, I. 2011. The influence ofcorporate venture capital investment on the likeli-hood of attracting a prestigious underwriter: An em-pirical investigation. Advances in Financial Eco-nomics, 14: 165–201.

Tushman, M. L., & Anderson, P. 1986. Technologicaldiscontinuities and organizational environments.Administrative Science Quarterly, 31: 439–465.

Utterback, J. 1994. Mastering the dynamics of innova-tion: How companies can seize opportunities in theface of technological change. Boston: Harvard Busi-ness School Press.

Wadhwa, A., & Kotha, S. 2006. Knowledge creationthrough external venturing: Evidence from the tele-communications equipment manufacturing indus-

try. Academy of Management Journal, 49: 819–835.

Walker, G., Kogut, B., & Shan, W. 1997. Social capital,structural holes and the formation of an industrynetwork. Organization Science, 8: 109–125.

Wang, L., & Zajac, E. J. 2007. Acquisition or alliance? Adyadic perspective on interfirm resource combina-tions. Strategic Management Journal, 28: 1291–1317.

Wasserman, N. 2012. The founder’s dilemmas. Prince-ton, NJ: Princeton University Press.

Wilson, F. 2008. Corporate venture capital. July 21.Blog post online at http://www.avc.com/a_vc/2008/07/corporate-ventu.html

Wry, T., Cobb, C., & Aldrich, H. 2013. More than a met-aphor: Assessing the historical legacy of resourcedependence and its contemporary promise as a the-ory of environmental complexity. Academy of Man-agement Annals, 7: 439–486.

Zott, C., & Huy, Q. N. 2007. How entrepreneurs usesymbolic management to acquire resources. Admin-istrative Science Quarterly, 52: 70–105.

Benjamin L. Hallen ([email protected]) is an assistant pro-fessor of management and organization at the Michael G.Foster School of Business at the University of Washing-ton. He received his PhD in management science andengineering from Stanford University. His research in-volves the study of how new actors strategically embedthemselves in market networks, with a particular focuson how entrepreneurs obtain investments and resourcesfor young ventures.

Riitta Katila ([email protected]) is a professor of man-agement science and engineering at Stanford University.She received her PhD in strategy from the University ofTexas at Austin. Her research interests include technol-ogy strategy, technology entrepreneurship, and organiza-tional learning and innovation.

Jeff D. Rosenberger ([email protected])is vice president of research at Wealthfront. He receivedhis PhD in management science and engineering fromStanford University. At Wealthfront, he is responsible forinvestment research, customer development for new in-vestment products, the Wealthfront investment seminarprogram, and work with the investment advisory board.


http://www.avc.com/a_vc/2008/07/corporate-ventu.html

http://www.avc.com/a_vc/2008/07/corporate-ventu.html

Documents

HOW DO SOCIAL DEFENSES WORK? A RESOURCE- …rkatila/new/pdf/KatilaAMJ2014.pdfhow do social defenses work? a resource-dependence lens on technology ventures, venture capital investors,