venturing into new territory: career experiences of corporate venture capital … · 2014. 9....

VENTURING INTO NEW TERRITORY: CAREER EXPERIENCESOF CORPORATE VENTURE CAPITAL MANAGERS AND

PRACTICE VARIATION

GINA DOKKOUniversity of California, Davis

VIBHA GABAINSEAD, Singapore

When organizations adopt new practices, the practices are often modified to fit the newcontext. We argue that managers who implement new practices modify them, and thatthe extent of practice variation is determined by two types of these managers’ careerexperience: experience with the practice itself and experience that enables assessmentof the fit between the practice and the adopting firm. We test these arguments byobserving information technology firms’ modification of venture capital practices incorporate venture capital units. This study contributes to diffusion research by devel-oping and testing a framework for understanding the role of individuals in practicevariation.

The diffusion of practices has long been a centralconcern of organization theory. The social, institu-tional, and structural mechanisms through whichpractices diffuse (e.g., Abrahamson & Rosenkopf,1997; Davis & Greve, 1997; Palmer, Jennings, &Zhou, 1993), and the conditions under which theyare adopted or not (e.g., Ferlie, Fitzgerald, Wood, &Hawkins, 2005; Jonsson, 2009) are becoming wellunderstood. Yet, even in this rich literature, fewstudies have gone beyond the adoption event, in-stead assuming that practices are adopted uncriti-cally and in toto (Campbell, 2005: 54). In contrast,qualitative accounts of practice adoption suggestthat practices are not adopted wholesale but aretypically modified to fit a new context (Ansari,Fiss, & Zajac, 2010; Czarniawska-Joerges & Sevón,2005; Gaba & Meyer, 2008). These modifications

are important to understand, because they may en-able or impede persistence of a practice in partic-ular organizations (e.g., Strang & Jung, 2009), orthey may result in innovation to the practice itself(Ansari et al., 2010). Therefore, scholars need toextend theory about diffusion to understand thesources of practice variation.

Institutional theory is an important lens throughwhich researchers have examined practice diffusion(Strang & Meyer, 1993; Strang & Soule, 1998). Be-cause institutional processes favor isomorphism, thetheory plays down variation in organization-level ad-aptation to diffusing practices, and little attention hasbeen devoted to the adaptation of practices as they areimplemented in organizations (Lounsbury, 2008). Re-cent studies, however, acknowledge that heterogene-ity in organizational practices is institutionally andorganizationally shaped (Dacin, Goodstein, & Scott,2002; Lounsbury, 2008). For example, the presence ofcompeting institutional logics or of broader field-level organizations can lead to differences in how orwhich elements of practice are adopted (Lounsbury,2001, 2007). Within organizations, political contesta-tion and the framing of a practice can lead to variationin goal orientations and in how pervasively practicesare incorporated (Fiss & Zajac, 2004; Kennedy & Fiss,2009). These studies advance understanding of vari-ous field- and organization-level sources of variation;however, they do not address practice variation thatresults from individual-level action.

The translation perspective, offered by Scandina-vian institutionalism, portrays a process view ofpractice variation in which practices are posited to

We thank Matthew Bidwell, Beth Bechky, Diane Bur-ton, Gary Dushnitsky, Kathy Eisenhardt, Henrich Greve,Alan Meyer, and seminar participants at NYU Stern OT/Strategy Brownbag, West Coast Research Symposium onTechnology Entrepreneurship, the Wharton People andOrganizations Conference, and the Academy of Manage-ment meetings for their valuable comments. We alsothank our associate/action editor, Jim Combs, for hisgenerous support and guidance. Clare Lee, SpencerDoehlert, and Thijs Kwik provided valuable help withdata collection. The research reported in this paper wassupported by INSEAD research grant 2520-493R. Bothauthors contributed equally to this work.

Editor’s note: The manuscript for this article was ac-cepted for publication during the term of AMJ’s formereditor-in-chief, R. Duane Ireland.

� Academy of Management Journal2012, Vol. 55, No. 3, 563–583.http://dx.doi.org/10.5465/amj.2009.0909

563

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.

undergo change every time they are applied in anew context (Czarniawska-Joerges & Sevón, 2005;Sahlin & Wedlin, 2008). Case studies based on thisperspective show that individuals actively inter-pret and translate practices into a new context us-ing editing rules that are governed by the institu-tionalized beliefs and norms of the new context orrules that contextualize broad ideas into concreteconcepts that are easy to apply (e.g., Boxenbaum &Battilana, 2005; Morris & Lancaster, 2006). Al-though these studies’ primary focus is on the intraor-ganizational processes underlying adaptation, theyalso document how individuals actively interpretand translate practices to fit into a new local con-text. However, they do not provide a systematicbasis for making predictions about how individualsinfluence the extent of practice variation.

Therefore, existing theory does not address theextent of practice variation or provide a systematicway of understanding individuals’ influence onmodification of practices to fit the contexts ofadopting organizations. We address these gaps byfocusing on the role of individuals in practice vari-ation. The individuals who implement and managepractices evaluate the fit of the practices to a con-text, and these individuals can influence the extentof practice variation. We draw on organizationallearning theory that highlights individuals’ knowl-edge and cognition and their effects on how indi-viduals perceive problems and make choices(March & Simon, 1958; Miner & Mezias, 1996).Developments in this research stream show that theprior experiences of individuals are portable andcondition what they know and how they see theworld, which results in organization-level learning(Almeida, Dokko, & Rosenkopf, 2003; Wezel, Cat-tani, & Pennings, 2006), new product introduction(Rao & Drazin, 2002), and models of organizing(Beckman & Burton, 2008). Thus, we propose thatpractice variation is a function of the career expe-riences of the individuals who manage and imple-ment adopted practices. Implementing managers’career experiences supply knowledge and skillsand models of behavior or logics of action thataffect managerial action and organizational out-comes (Burton, Sorensen, & Beckman, 2002; Lant,Milliken, & Batra, 1992).

We develop a theoretical framework that distin-guishes between two relevant types of career expe-rience: experience conducting a practice itself, andexperience that enables assessment of the fit be-tween a practice and an adopting organization. Weargue that practice-specific and fit-specific careerexperiences provide individuals with specificknowledge and mental models that influence theextent to which practices are modified in adopting

organizations. We test hypotheses generated fromthis framework in the context of corporate venturecapital (CVC) units. CVC units are structurally dis-tinct units established within existing corporationsthat are dedicated to making external equity invest-ments in high-potential entrepreneurial start-ups(Gaba & Meyer, 2008). These units are modeled onthe venture capital practices of finding, funding,and guiding entrepreneurial start-ups. The imple-mentation of these practices in a corporate contextpresents challenges based on the goals and opera-tional methods of the practices as well as on idio-syncratic features of adopting firms, providing rea-sons for modifying these practices. In addition,differences in governance, compensation, and so-cial structures deter the free flow of labor fromindependent venture capital (IVC) firms into CVCunits, providing the variance in experiences of CVCmanagers that is needed to address our questions,and making CVC units a highly suitable context forstudying questions about individual-level effectson practice variation.

We contribute to the diffusion literature by de-veloping and testing theory about practice variationthat highlights the role of individuals who imple-ment practices and specifying how their career ex-perience influences the extent of practice variation.In doing so, we complement work on institutionaland organizational sources of practice variation byinvestigating its microfoundations.

PRACTICE VARIATION WITHIN CVC UNITS

Venture capital practices consist of selecting,funding through equity, and providing managerialsupport to young companies that have the potentialto grow exponentially (Gompers & Lerner, 2004).Established firms typically adopt venture capitalpractices by forming CVC units. In recent years,many technology firms have come to view CVCunits as a crucial part of their R&D and businessdevelopment activities (Chesbrough, 2003). Theseunits facilitate a coordinated and proactive ap-proach to external sources of new technologies tocomplement or even replace in-house R&D (Dush-nitsky & Lenox, 2005). The 1990s saw an unprece-dented diffusion of CVC units: by the year 2001over 300 U.S. firms had established corporate ven-turing units (Gaba & Meyer, 2008). Simultaneously,the corporate share of overall venture capital in-vestments rose from 2 percent in 1994 to 17 percentin 2000, according to Venture Economics (http://ventureeconomics.com/).

Investigating questions about adaptation and fitin a CVC context entails specifying areas in whichpractice variation could occur. We focus specifi-

564 JuneAcademy of Management Journal

cally on the venture capital practice of finding in-vestment targets—that is, the evaluation and selec-tion of start-ups, the critical first stage of theventure capital process (Gompers & Lerner, 2004).Adapting this practice to a corporate context canentail modification of the goal orientation and theoperational strategies concerning investments.First, venture capital practices are traditionallyconducted in independent venture capital partner-ships of professional investors whose central goalis to maximize financial returns on their invest-ments (Gompers & Lerner, 2004). To achieve thisgoal, IVCs typically invest in entrepreneurial start-ups that have innovative, but uncertain, productideas. The IVCs use their expertise and specializedskills to help bring products to market quickly and“cash out” at an optimal point (Metrick & Yasuda,2010). For CVC units, venture capital practices thatare oriented around identifying lucrative invest-ment opportunities are complicated by the overlayof a strategic imperative. Although firms may beenticed by the potential for a financial “home run,”most firms claim that their foremost objectives arestrategic: learning about new (and potentially dis-ruptive) technologies, gaining access to new mar-kets and business models, and identifying prospec-tive acquisition targets (Chesbrough, 2003). Thus,the goal orientation of CVC units can vary: they canbe oriented toward purely financial goals, as in thecase of IVCs; or they can be oriented toward strate-gic goals that contribute to an overall technology orbusiness strategy; or they can be oriented towardboth goals (MacMillan, Roberts, Livada, &Wang, 2008).

Second, IVCs follow certain typical operationalstrategies for selecting investment targets. Two im-portant operational choices that IVCs make are in-vestment timing and range of investment targets(Gompers & Lerner, 2004). Though IVCs can investin different stages of a start-up’s life, they tend toinvest early because they hold an advantage overother investors in the earlier stages. Early-stagestart-ups are often little more than an idea and afounding team, and IVCs have the specializedknowledge and skills needed to evaluate start-upsat this stage of development. In addition, IVCs tendto have extensive social networks that cross indus-try boundaries (Sorenson & Stuart, 2001), givingthem information about, and access to, a widerscope of investment opportunities than other inves-tors have. Given their expertise, IVCs concentrateinvestments where “information asymmetries” arelikely to be most significant and monitoring mostvaluable; specifically, they tend to follow opera-tional strategies of investing in early-stage compa-nies and in a variety of industries (Gompers, 1995).

CVC units can follow standard IVC investmentstrategies by investing early in start-ups, or theycan invest later, when the uncertainty about thestart-ups’ technologies and business models isgreatly reduced. Similarly, CVC units can vary fromIVCs in terms of investment range by restrictingtheir exploration of investment targets to a nar-rower range of start-ups that are more targeted tothe firm’s strategic needs.

Therefore, implementing the IVC practice ofevaluating and selecting investments in CVC unitscan entail adjustment in both the practice’s goalorientation and operational strategies.

THEORY AND HYPOTHESES

What are the sources of practice variation, and inwhat ways do practices vary? Institutional scholarsacknowledge the connection between institutionalcontext and variation in organizational behaviorand practices. For instance, Lounsbury (2007)found that competing logics within the mutualfund industry led to contestation over goal orienta-tion and variance in the use of professional moneymanagement firms. Alternatively, practice varia-tion can be shaped by organizational and technicalfit between diffusing practices and adopting organ-izations (Ansari et al., 2010; Meyer & Goes, 1988).Moreover, pressures for customization can be par-ticularly intense when innovative practices are ad-opted from a different organizational population(Gaba & Meyer, 2008). The translation perspectiveon practice adoption highlights the process bywhich practices are actively translated and editedto fit a local context (e.g., Czarniawska-Joerges &Sevón, 2005; Sahlin-Andersson, 1996). This per-spective draws on the concept of theorization (Gaba& Meyer, 2008; Nigam & Ocasio, 2010; Strang &Meyer, 1993), which entails “translating concretepractices into abstractions for export and then un-packing the abstractions into a (suitably modified)concrete practice upon arrival” (Strang & Soule,1998: 276). Critical to the translation process ishow practices become reinterpreted and then ap-plied in a new setting. In theory about the post-adoption processes that lead to variation, adoptingfirms locally “construct” the problems and needsthat trigger adoption in the first place, and thisconstruction is used to change practices to fit (Sah-lin-Andersson, 1996).

Therefore, practices can be modified as they areimplemented in organizations, leading to variationin the way practices are conducted. First, the pro-cess of translation can lead to changes in the goalsof a practice itself. For example, Boxenbaum andBattilana (2005) examined the diffusion of Ameri-

2012 565Dokko and Gaba

can-style diversity programs from the U.S. to Den-mark. They found that as these programs diffused,the goals of the program changed from increasingorganizational and individual effectiveness to en-hancing the integration of immigrants into the Dan-ish workforce. Likewise, Gaba and Meyer (2008)observed that the diffusion of venture capital prac-tices among publicly traded corporations was ac-companied by the modification of the practice’sgoals from pursuing purely financial returns to ac-cessing new and emerging technologies. Second,the operational methods of practices can also cometo vary as the practices are implemented in organ-izations (e.g., Morris & Lancaster, 2006; Zbaracki,1998). For example, though sophisticated statisticaltools are an important part of total quality manage-ment (TQM) practice, Zbaracki (1998) details howmanagers implementing TQM avoided statisticaltools that they found intimidating or difficult tounderstand, using them symbolically or margin-ally, or dropping them from the practice altogether.Similarly, Coburn (2004) found that public schoolteachers’ individual prior experiences with readinginstruction mediated between the adoption of newpractices for reading instruction and the routines,materials, and organization that constituted thepractices in their specific classrooms.

In sum, prior research has investigated field- ororganization-level sources of practice variation andhow practices can be translated and edited into thelanguage and logics of a new context. Despite theseadvances in scholars’ understanding of the sourcesof practice variation, individuals’ roles in modify-ing practices remain largely unaddressed. A num-ber of the studies cited above (i.e., Boxenbaum &Battilana, 2005; Coburn, 2004; Morris & Lancaster,2006; Zbaracki, 1998) are detailed qualitative ac-counts that suggest that individual-level effectsalso operate on practice variation. However, theydo not provide a systematic way of predicting howmuch individuals influence the extent of practicevariation.

Implementing Manager’s Career Experience andPractice Variation

Managers who perform practices are the inter-preters and editors of adopted practices (Boxen-baum & Jonsson, 2008; Sahlin-Andersson, 1996).They make choices about what aspects of practicesto implement and in what way practices will bemodified (i.e., what goals and operational strategieswill be pursued). Though an adopting organizationsets some of these directions, individuals bringtheir own expertise, preferences, and cognitions tothe task of implementing and operating a new prac-

tice, and these characteristics are a function of theircareer experience. There is ample evidence thatcareer experience affects what individuals know,the way in which they perceive the world, and howthey act. Organizational learning scholars considerprior career experience an important source oflearning, theorizing that experience carried by in-dividuals influences firm-level innovation orchange (Beckman & Burton, 2008; Rao & Drazin,2002; Singh & Agrawal, 2011) and individual andteam performance (Boh, Slaughter, & Espinosa,2007; Dokko, Wilk, & Rothbard, 2009). In doing so,these authors rely on two common mechanismsthat account for the effects of prior experience:specific knowledge and mental models.

Specific knowledge gained from individuals’prior jobs is one way in which career experiencecan affect practice variation. Specific knowledge,by which we mean knowledge and skill tailored toa particular environment or set of tasks, affects bothwhat individuals are able to do and how they de-fine problems (Carlile, 2004). First, specific knowl-edge enables the successful performance of tasksassociated with a practice or type of work (Argote,1999). For instance, entrepreneurs who have previ-ous experience with firm founding acquire specificknowledge about the practices of founding newventures that enables them to repeat the process(Shane & Khurana, 2003). In addition, career expe-rience can confer specific knowledge about socialenvironments. For example, firm-specific knowl-edge includes an understanding of a firm’s culture,politics, and routines that affects managerial ac-tions and outcomes (Groysberg, Lee, & Nanda,2008; Huckman & Pisano, 2006). Second, specificknowledge affects how managers understand prob-lems by creating a reference point for interpretingsituations in relation to their expertise. Elements ofa situation that are familiar are more salient, whichencourages individuals to interpret a problem interms of what they know (Lant et al., 1992). Mentalmodels are a second, related mechanism throughwhich career experience affects practice variation.Mental models reflect the way individuals see theworld: the causal models they hold about perfor-mance and their assumptions about what goals andbehavior are appropriate and valued (March & Si-mon, 1958). Individuals base these models on theirprior experiences (Dokko et al., 2009; Kraatz &Moore, 2002). The operation of mental models isprimarily nondeliberative (Gioia & Poole, 1984), insuch a way that taken-for-granted assumptions af-fect choices about what tasks and goals are mostimportant for an organization.

In explaining the extent of practice variation, wefocus on implementing managers’ career experi-


ence that is oriented either toward an adopted prac-tice, or toward its fit with the adopting firm.Though the career experience of implementingmanagers has not been directly applied to predict-ing the extent of practice variation, we expect themechanisms of specific knowledge and mentalmodels to operate in this relationship as well. Asindividuals implement and manage practices, theyinterpret problems, prioritize activities and goals,and conduct work according to their prior experi-ence. Our framework thus distinguishes betweentwo types of experiences held by implementingmanagers: practice-specific experience, defined asprior career experience performing a practice, andfit-specific experience, defined as prior career ex-perience that enables assessment of fit between apractice and organizational or technical aspects ofan adopting organization. We use this distinction tohypothesize the impact of implementing managers’prior career experiences on the extent of practicevariation, in terms of both the goal orientation andoperational strategies of the adopted practices.

Practice-Specific Experience and PracticeVariation

Career experience with a practice yields specificknowledge about the activities that comprise thepractice itself. This specific knowledge should leadimplementing managers to perceive the postadop-tion process as a problem of straightforward appli-cation of their previous experience with a practiceand to perform the activities in a way that is famil-iar to them (Dokko et al., 2009). Reinforcing thiseffect, these managers also tend to view their ac-

customed goals and familiar ways of performingthe practice as appropriate, because of the mentalmodels they hold from their experience with thepractice (Gioia & Poole, 1984; Kraatz & Moore,2002). Accordingly, implementing managers’ prac-tice-specific experience should lead to less varia-tion in both the goal orientation and the operationalstrategies of the practice as implemented in a newcontext.

For managers in CVC units, having prior experi-ence with venture capital practices in IVC firmsshould lead them to apply their experience in astraightforward way that minimizes practice varia-tion. That is, CVC managers with career experiencein IVCs should be oriented toward financial goals,as they are both knowledgeable about how to attainfinancial goals and conditioned to value thosegoals. Operationally, they are likely to apply theirknowledge about evaluating early-stage start-ups totheir strategy for an investment stage and to investin a relatively industry-agnostic way, as IVCs do,influencing the selection of investment targets ofthe CVC unit. This propensity to act in accordancewith prior experience should lead to less variationfrom standard IVC practice when implementingmanagers have IVC experience.

Hypothesis 1a. The higher the proportion ofimplementing managers who have practice-specific experience (i.e., career experience inIVCs), the less a practice’s goal orientation ismodified when it is adopted.

Hypothesis 1b. The higher the proportion ofimplementing managers who have practice-

FIGURE 1Implementing Managers’ Career Experience and Extent of Practice Variation


specific experience (i.e., career experience inIVCs), the less a practice’s operational strate-gies are modified when it is adopted.

Fit-Specific Experience and Practice Variation

In addition to practice-specific experience, im-plementing managers may also have experiencesthat can inform their assessment of the fit betweena practice and an adopting organization. Like deepunderstanding of the practice, deep understandingof an adopter’s objectives and social environmentor of its existing systems and technologies as theyrelate to an adopted practice can drive customiza-tion of a newly adopted practice (Ansari et al.,2010; Birkinshaw, Hamel, & Mol, 2008; Meyer &Goes, 1988). We distinguish two types of experi-ence that affect implementing managers’ assess-ment of the fit between a practice and an adoptingorganization. Organizational fit–specific experi-ence focuses on fit between a practice and organi-zational (i.e., strategic, cultural, and political) as-pects of an adopting organization, and technicalfit–specific experience focuses on fit between tech-nical aspects of a practice and systems and tech-nologies that an organization already uses.

Organizational fit–specific experience. Assess-ing the organizational fit of a practice entails con-sidering the practice in terms of its compatibilitywith the existing social environment and businessstrategies of an adopting firm. Organizational insid-ers have firm-specific knowledge that enables un-derstanding of the cultural, political, and strategiccontext that a practice must operate in (Groysberget al., 2008; Huckman & Pisano, 2006). Moreover,an individual who has been socialized into an or-ganization has internalized its culture (VanMaanen & Schein, 1979), so that the mental modelsassociated with organizational membership willalso drive assessments of fit (Dokko et al., 2009).Additionally, implementing managers who are al-ready organization members are more likely to un-derstand the impetus for adopting a practice andhow the practice fits with the organization’s overallbusiness strategy. Therefore, implementing manag-ers who have career experience in the adoptingorganization, regardless of what function they pre-viously performed, will likely see differences be-tween a canonical version of a practice and theobjectives and social environment of an organiza-tion that lead them to customize a practice.

In the context of CVC units, internally hired CVCmanagers should be more likely to modify practicesto fit an adopting organization. Understanding aCVC unit’s genesis and how it relates to the corpo-ration’s overall business strategy, culture, and ex-

isting power structures will lead these managers tocustomize the goal orientation and operationalstrategies of their CVC unit, resulting in greaterpractice variation. Specifically, implementingmanagers who previously held other jobs in theadopting firm are more likely to modify investmentgoals to be oriented toward strategic goals insteadof financial goals, and they are more likely to mod-ify operational strategies for investing by favoringlater-stage investing and investing in a narrowerrange of industries than IVCs would do.

Hypothesis 2a. The higher the proportion ofimplementing managers who have organiza-tional fit–specific experience (i.e., firm-specificcareer experience), the more a practice’s goalorientation is modified when it is adopted.

Hypothesis 2b. The higher the proportion ofimplementing managers who have organiza-tional fit–specific experience (i.e., firm-specificcareer experience), the more a practice’s oper-ational strategies are modified when it isadopted.

Technical fit–specific experience. Individualsimplementing adopted practices can also influencethe extent of practice variation if they have careerexperience that enables assessment of the technicalfit between a practice and an adopting organiza-tion. The technical dimension of fit concerns “thedegree to which the characteristics of a practice arecompatible with technologies already in use by po-tential adopters” (Ansari et al., 2010: 75). An organ-ization’s existing technologies may need to interactwith technical aspects of a practice, and specificknowledge and mental models coming from priorexperience that is related to these existing technol-ogies influence fit assessment, whether or not thetechnical experience was acquired within theadopting organization. For instance, the wide-spread adoption of internet e-commerce practicesby traditional corporations could be understood interms of organizational strategy, politics, and cul-ture, but it could also be considered in terms of afirm’s existing technical systems for sales, invento-ries, and distribution. In this case, an implementingmanager with experience in these technical sys-tems, such as an enterprise resource planning (ERP)system expert, would have a refined understandingof technical fit, whether or not the ERP experiencewas with the adopting firm. Moreover, day-to-daychoices that make up the operational strategies ofan adopted practice are especially conditioned bythe understanding of an organization’s technicalneeds (Dewar & Dutton, 1986). Therefore, we ex-pect implementing managers who have technical


fit–specific experience to be more likely to tailorpractices to fit because of their superior under-standing of the existing technologies and technicalneeds of the adopting firm and how they will in-teract with technical aspects of the practice.

In the case of CVC units, investments are typi-cally made in high-technology start-ups, and theability to understand an adopting firm’s technolog-ical base and assess fit with start-ups’ technologiescan be a source of practice variation. The better aCVC manager understands an adopting firm’s tech-nologies and R&D-related goals (e.g., by drawing onprior engineering experience), the less likely thatmanager will see an off-the-shelf version of a prac-tice as a fit. Therefore, CVC managers who haveengineering career experience will be more likelyto modify IVC practices in such a way that invest-ments are more strategically oriented and opera-tional strategies target later-stage investment andmore narrowly targeted industries.

Hypothesis 3a. The higher the proportion ofimplementing managers who have technicalfit–specific experience (i.e., engineering careerexperience), the more a practice’s goal orien-tation is modified when it is adopted at a CVC.

Hypothesis 3b. The higher the proportion ofimplementing managers who have technicalfit–specific experience (i.e., engineering careerexperience), the more a practice’s operationalstrategies are modified when the practice isadopted at a CVC.

It is important to note that managers can come tojobs having had multiple jobs and even multiplecareers (Arthur & Rousseau, 1996; Bidwell & Bris-coe, 2010); for example, CVC managers can haveboth IVC and engineering experience, or they canhave held engineering jobs inside and outside of anadopting firm in their careers.1 Conceptually, andin the empirical analysis, we consider these typesof experience in separate dimensions; for example,a manager with engineering experience within anadopting firm has both organizational and techni-cal fit-specific experience.

METHODS

Sample and Data

We constructed our sample using the 2000, 2001,and 2002 volumes of the Corporate Venturing Year-book and Directory, which lists firms with an active

CVC unit along with information about the year ofestablishment of the CVC unit. Though some ITfirms experiment with CVC activity prior to estab-lishing a formal unit, it is establishment—that is,the adoption of venture capital practices as a formaland staffed activity of the corporation—that is ourstarting event of interest, because it represents thestarting point for adaptation of IVC practices. Werestricted our sample to include only IT sectorfirms that had established CVC units.2 This proce-dure resulted in a sample of 93 IT firms with CVCunits. IT firms were an appropriate sample for test-ing our hypotheses, because firms in the IT sectormade more than three-quarters of all CVC invest-ments in the 1990s (Gaba & Meyer, 2008). More-over, focusing on IT firms exclusively allowed us tocontrol for unobserved heterogeneity at the indus-try level.

Next, we used the VentureXpert database to col-lect longitudinal data on the goal orientation andoperational strategies of all CVC units in our sam-ple until the year 2008. For CVC unit managers’prior experience, we collected biographical datathrough internet searches.

Because some data were missing from VentureX-pert, or biographical information was incomplete,our final sample is an unbalanced panel for 70 CVCunits over the period 1992–2008.3

Dependent Variables

We constructed and tracked each CVC unit’sportfolio of entrepreneurial start-ups to calculatetime-varying measures of CVC goal orientation andCVC operational strategies. We used VentureXpertto collect data on each CVC unit’s portfolio of in-vestments in U.S.-based entrepreneurial start-upsfrom the date of establishment of the unit up untilthe year 2008. The 70 firms in our final sampleinvested in 1,788 start-ups over the time period ofour study. VentureXpert classifies each start-up as

1 In our sample, only 50 percent of internal hires hadengineering career experience.

2 We used the National Science Foundation’s defini-tion of the IT sector (NSF, 2000) as comprising the fol-lowing five industry subsectors: (1) Office, Computingand Accounting Equipment (SIC code 357), (2) Commu-nications Equipment (SIC code 366) (3) Electronic Com-ponents (SIC code 367), (4) Communication Services (SICcodes 481–484, 489), (5) Computing and Data ProcessingServices (SIC code 737).

3 At each stage at which our sample was reduced, weperformed nonparametric Kolmogorov-Smirnov tests tocheck if the distribution of the firm-level variables sig-nificantly differed for the dropped firms as compared tothe final sample of 70 firms. In all cases, the null of nosignificant difference is not rejected.


“defunct,” “private,” “public,” and “subsidiary.”Each start-up enters a CVC unit’s portfolio on thedate of first investment and continues to be part ofthe portfolio as long as it is classified as active.Start-ups classified as public or subsidiary exit theportfolio the year after they went public or wereacquired. We used SDC’s Global Issues Database toidentify the year in which a start-up went publicand SDC’s Mergers & Acquisitions Database toidentify the year a start-up was acquired. For start-ups classified as defunct, we dropped them fromthe portfolio only if at least two years had elapsedsince the year they received their last round offinancing. Overall, of the 1,778 start-ups that re-ceived investments from our sample of CVC units,32 percent remained private, 11 percent went pub-lic, 40 percent were acquired, and 17 percent weredefunct.4

CVC goal orientation. To examine how muchCVC units modify the goals of venture capital prac-tices, we measured each CVC unit’s goal orienta-tion along two dimensions: financial and strategic.Financial orientation is the proportion of start-upsin a portfolio that resulted in an initial public of-fering (IPO) or an acquisition by a firm other thanthe parent of the focal CVC unit. For each CVC unitfor each year, our measure cumulates all IPOs andacquisitions and expresses this sum as a fraction ofthe cumulated entrepreneurial start-ups in the port-folio.5 IPOs and acquisitions by other firms arecash-generating events, and they represent an ori-entation toward financial goals that are not neces-sarily strategically relevant for the CVC unit’s par-ent firm, so we used these to represent anorientation toward financial goals. Since tradi-tional VC practices are financially oriented, a highfraction of IPOs and acquisitions in a CVC unit’sportfolio implies less modification of the originalpractice. The mean value for this variable is 0.28,which implies that the average CVC unit had 28percent of the start-ups it invested in either goingpublic or being acquired by other firms.

We measured strategic orientation as the cumu-lated number of entrepreneurial start-ups acquiredby the parent firm of a focal CVC unit as a propor-tion of all entrepreneurial start-ups in the CVCunit’s portfolio since the unit’s establishment (Ben-son & Ziedonis, 2010; Gaba & Meyer, 2008). As withour measure for financial orientation, we took theachievement of strategically relevant outcomes as ameasure of the extent of CVC managers’ orientationtoward strategic goals. Consequently, a higher frac-tion of start-ups acquired by a parent firm indicatesa divergence from the traditional financial goal ori-entation of IVCs. Though acquiring portfolio com-panies may not capture all strategic objectives ofCVC investment, and acquisitions may even de-stroy shareholder value for firms that do not havededicated CVC units (Benson & Ziedonis, 2010),the very act of acquisition is a visible and tangibleindicator of a CVC unit’s belief about the potentialstrategic benefits that could accrue from such anacquisition.

We constructed this variable using the SecuritiesData Corporation’s (SDC) Mergers & Acquisitionsdatabase to obtain a list of all private acquisitionsby the firms in our sample, then matched theseacquisitions with each CVC investment in the Ven-tureXpert database to include only those acquisi-tion targets in which the CVC unit had investedprior to acquiring it. Our summary statistics showthat the average CVC unit, in any given year, hadacquired 2 percent of the portfolio companies thatit invested in.6

The operationalizations of these dependent con-structs are imperfect in that they do not directlymeasure CVC units’ goal orientations and insteaduse investment outcomes as proxies. These proxiesrepresent the dependent constructs to the extentthat CVC units have varying goal orientations, andtheir investment outcomes reflect these orienta-tions. Surveys of CVC units support variation ingoal orientation. An Ernst & Young (2008) survey of37 CVC units found that 17 percent of them hadpurely strategic goals, 3 percent had purely finan-cial goals, and the remaining 80 percent had ablend of financial and strategic goals. Similarly, aU.S. NIST report describes a study of 48 CVC units,of which 15 percent had only financial goals, 15percent had only strategic goals, and 70 percenthad both financial and strategic goals (MacMillan

4 Of all the start-ups classified as defunct, 97 percenthad not received any funding for a period greater than orequal to five years. For the remaining 3 percent, wechecked through LexisNexis and internet sources andwere able to confirm that these were indeed defunctstart-ups. However, classifying them as private does notaffect our results qualitatively.

5 Our measure would be subject to bias if the propor-tion of companies that had an IPO or acquisition wereexpressed as a fraction of the number of active companiesin a CVC unit’s portfolio. By disregarding the companiesthat became defunct, we would overstate the degree offinancial orientation.

6 Since both financial and strategic orientation vari-ables are expressed as proportions, right censoring is lessof an issue. Firms that have a longer time to exit a port-folio company also have had a large number of portfoliocompanies at risk of exit.


et al., 2008). We included separate dependent vari-ables for financial and strategic goal orientation toallow for split orientation. Moreover, though in-vestments are subject to review and approval bytheir corporate parent, CVC unit managers selectthe set of potential investment targets to be evalu-ated, and perform due diligence and make recom-mendations. Therefore, CVC units have orienta-tions toward particular goals, and on average,orientation toward financial (strategic) goalsshould be associated with financially (strategically)oriented outcomes, making our dependent vari-ables suitable proxies for goal orientation. In addi-tion, these measures have the advantage of beingindependently observable and measurable, andthey are also time-varying, which allowed us toestimate coefficients with greater precision usingpanel methods. Despite being suitable, using proxymeasures introduces the possibility of measure-ment error. However, measurement error in the de-pendent variable does not bias estimates. It simplyresults in higher asymptotic variance that inflatesstandard errors (Wooldridge, 2002).

CVC operational strategies. We focused on twooperational strategies that are closely associatedwith traditional venture capital practices: the pro-portion of early-stage investments in a CVC unit’sportfolio and the diversity of investments in theCVC unit’s portfolio. VentureXpert classifies in-vestment in entrepreneurial start-ups according tothe following stages: “seed,” “early stage,” “expan-sion,” “later stage,” “acquisition,” and “public mar-ket.” Proportion of early-stage investments is thenumber of investments classified as seed or earlystage expressed as a proportion of all start-ups in aCVC unit’s portfolio. In our sample, 20 percent ofall investments in start-ups were classified as earlystage ones. Next, to measure the range of industriesinvested in by a CVC unit, we calculated the sec-toral diversification of investments. We identifiedthe industry sector of each start-up receiving CVCfunds using the Venture Economics IndustryClassification (VEIC), a VentureXpert proprietaryindustry classification scheme.7 From this, wecalculated a Herfindahl-Hirschman index of con-centration as

HHIit ��j Pijt2 ,

where Pitj is the proportion of start-ups in firm i’sCVC unit in industry-sector j at time t. Since this is

an inverse measure of diversification, to facilitateinterpretation of coefficients, we recoded it as (1 –HHIit).

Independent Variables

The major predictor variables in this study per-tain to CVC managers’ career experience. We col-lected biographical data on CVC personnel usinginternet-based searches conducted between No-vember 2008 and July 2010. The Corporate Ventur-ing Yearbook and Directory that we used to con-struct our firm-level sample also identifies thenames of key personnel in the CVC units. The 93CVC units in our sample were associated with 340individuals, whom we uniquely identified usingtheir names and the name of their CVC unit. Al-though biographical information is not universallyor consistently available online, we were able tofind at least some biographical information for 91percent of them (311 individuals). Commonsources of biographies were firm websites, Securi-ties and Exchange Commission filings, and profes-sional networking sites such as LinkedIn. Individ-uals could be missing from our sample becausepublished biographies could not be found, or be-cause their names were so common they could notbe uniquely identified (e.g., “Mike Smith”). In ad-dition, multiple biographical sources were foundfor many of the individuals in the database. Thoughmost of the sources contained the same informationabout prior experience, when sources containedunique information, we recorded them separately.We recorded 773 separate biographies for the 311managers.

For each manager, we reconstructed work his-tory, listing each job separately and recording datesof employment or chronological ordering whereavailable. We listed a separate job record for eachdifferent job title. This coding resulted in 1,565separate job records. Because we collected datawell after the 2000–02 sample time frame, we iden-tified whether these jobs occurred before or afterthe CVC job identified in the Corporate VenturingYearbook and Directory. Of the 1,565 jobs we iden-tified, 845 jobs preceded the CVC jobs, 375 jobswere held after the CVC job, and the remaining 345job records represented the CVC job.8 The 845 priorjobs were coded to capture different types of expe-rience; this coding was the basis of our indepen-dent variables. Note that because coding is done at

7 Since these portfolio companies are not publiclytraded, they are not required to report traditional indus-try classifications such as SIC or NAICS codes. See Du-shnitsky and Shaver (2009) for more details.

8 Some individuals had multiple job titles during theirtenure in a CVC unit; e.g., one person went from “busi-ness development manager” to “investment manager.”


the job level, a single CVC manager could havemultiple types of experience, i.e., he or she couldhave practice-specific and both types of fit-specificexperience.

Where dates of employment for the CVC job wereavailable, we developed variables that captured theexperience of managers in each CVC unit over time.Start and end dates for CVC unit jobs were availablefor 229 individuals (74 percent of the 311 manag-ers). An additional 41 managers’ biographies listedeither the start or end date for their CVC job. Com-mon reasons for end dates to be available whilestart dates were not are that biographies tend to bemore specific about recent jobs than about jobs inthe more distant past, or that the job immediatelyfollowing the CVC job provided information on itsstart date. For these reasons, we supplemented thesample by assuming that the manager started in theCVC job the first year his or her name appeared inthe Directory. If start dates were available but enddates were missing, the reasons were less clear, asthis condition generally resulted from the absenceof recent biographies on the manager. Therefore,we assumed that the manager left the CVC job justafter the last year his or her name appeared in theDirectory. With these assumptions, the analysissample increased to 270 managers (87 percent ofthe 311 managers).

Practice-specific career experience. We opera-tionalized practice-specific experience as the pro-portion of personnel in a CVC unit who had priorexperience in independent venture capital (IVC)firms. As part of the job-level coding, we coded IVCexperience for each person by examining job titlesand employers for prior jobs. For example, a gen-eral partner position at Frontier Ventures wascoded as an IVC job. We created annual panels forproportions of CVC personnel with practice-spe-cific career experience using the CVC job start andend dates.

Fit-specific career experience. We operational-ized fit-specific career experiences in two ways: asfirm-specific experience and as engineering expe-rience. CVC managers who had prior career expe-rience in an adopting firm were considered to havefirm-specific career experience. In our data, thesewere all internal hires—that is, the job immediatelypreceding their first CVC job was in the adoptingfirm. The variable fit-specific experience—organi-zational measures the proportion of CVC personnelwho were internal hires in the CVC units. Next,personnel with an engineering background haverelevant experience to assess the technical fit be-tween CVC practices and the adopting firm. Fit-specific experience—technical measures the pro-portion of CVC personnel who have engineering

work experience or an engineering degree. For bothmeasures, we created annual proportions using theCVC job start and end dates for the managers.

Controls

Firm-level controls. We included a number offirm-level controls that could account for the goalorientation and operational strategies of the CVCunits. First, we controlled for firm size, using dataon firm sales. Larger firms have more resources thatcan be allocated to CVC units, which may result inmore investment alternatives or higher-quality em-ployees. On the other hand, large firms are alsoencumbered by structural constraints that can dis-courage the pursuit of risky courses of action (Au-dia & Greve, 2006). Though we do not predict adirection for firm size’s effect, we control for it.Second, firms that exhibit better performance mayattract higher-caliber personnel, who might adaptpractices differently from lower-caliber personnel.Better-performing firms also tend to have a greaterappetite for risk (March & Shapira, 1987), whichcould influence their operational strategies. There-fore, we controlled for firm performance as incomebefore extraordinary items plus depreciation.Third, firm age is associated with increases in ratesof innovation (Sorensen & Stuart, 2000). However,older firms can also find it difficult to keep up withexternally generated technical changes and exhibitgreater inertia (Abernathy & Utterback, 1978). Firmage was the age of a firm in years at the time of CVCunit establishment. Finally, firms with greater slackresources are more likely to experiment in pursuitof new opportunities (Levinthal & March, 1981)and may find it easier to invest in riskier early-stagestart-ups, and to invest in a broader range of port-folio start-ups. Accordingly, we controlled for firmslack, measured as a firm’s current ratio (the ratio ofcurrent assets to liabilities), which represents the liq-uid resources uncommitted to liabilities (Bromi-ley, 1991).9

CVC unit controls. We also controlled for CVCunit characteristics that might influence CVC goalorientation or operational strategies. For example,

9 We also tested alternate measures of slack: absorbedslack, measured as the ratio of sales, general, and admin-istrative expenses to sales; and potential slack, measuredas the debt-to-equity ratio. Other controls we consideredincluded an alternate measure of firm size (measured asfirm assets), firm R&D intensity, and dummies for a firm’sprimary SIC codes intended to capture technological dif-ferences. None of these variables were significant, andwe did not include them in the final specification toconserve degrees of freedom.


CVC units with more extensive operations andthose that have been in operation longer are morelikely to have invested in a broader range of indus-try sectors. Therefore, we controlled for CVC expe-rience by including the cumulated number of start-ups each CVC unit had invested in since itsestablishment year. CVC duration was a count ofthe number of years since the date of the first in-vestment made by a CVC unit. Since seven firms inour sample had invested in portfolio companiesprior to the year of CVC unit establishment listed inthe Directory, we used the time elapsed since thedate of first investment (rather than the year ofestablishment) to construct this variable.10

Firm-specific fixed effects and temporal con-trols. Multiple omitted unobserved variables havethe potential to confound hypothesis testing. Weaccounted for such influences by employing amethodology that explicitly accounts for firm fixedeffects. Since the venture capital industry goesthrough episodic boom and bust cycles (Gompers &Lerner, 2004), we included time dummies in allspecifications to capture conditions in the overallpublic equity market.

Table 1 provides the summary statistics and cor-relations for the variables.

Empirical Methodology

In our empirical estimation, we needed to ad-dress two issues. First, since all of our dependentvariables were measured using a firm’s portfolio ofstart-ups, changes in CVC practices may not bereadily apparent unless a CVC unit significantlyalters its portfolio. More importantly, there may be“path dependence,” wherein decisions made by theCVC unit to change CVC practices going forwardare conditional on the characteristics of the currentCVC investment portfolio. Second, a shift in CVCunit personnel is likely to manifest itself as changesin CVC practices only over time, since identifyingand investing in entrepreneurial start-ups is a time-consuming process. This, in turn, implies the needto distinguish between short- and long-term effectsof changes in CVC unit personnel.

We accounted for these concerns by including alagged dependent variable as an independent re-gressor in each of our empirical models. This ap-proach has several methodological advantages.First, it allows us to explicitly account for persis-tence and path dependence in CVC practices. Sec-ond, it allows us to estimate both a short-term con-temporaneous effect for changes in theindependent variable and a long-term steady-stateeffect. Third, the lagged dependent variable helpsus to deal with the autocorrelation in panel datathat can lead to incorrect estimates of the standarderrors (Wooldridge, 2002). Finally, the lagged de-pendent variable can be recast as a geometricallyweighted summary measure of all lagged CVC unitbackground variables whose effects decline overtime (Greene, 2000). Therefore, it also allowed us tomodel past unobserved shocks as affecting future

10 In addition to these controls, we also examined thebusiness education background of the personnel in CVCunits. We found that CVC units with a greater proportionof personnel with business degrees were more strategi-cally oriented and less financially oriented and, althoughless diversified, more likely to invest in early-stage com-panies. However, this control does not affect any of ourmain results in terms of the direction, significance, ormagnitude of the coefficients.

TABLE 1Summary Statistics and Correlationsa

Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12

1. Financial orientation 0.28 0.192. Strategic orientation 0.02 0.04 –.003. Proportion of early-stage investments 0.16 0.18 –.10 .014. Sectoral diversification 0.76 0.25 .03 .06 –.015. Practice-specific experience 0.15 0.29 .06 –.11 .10 –.126. Fit-specific experience—Organizational 0.63 0.40 .09 –.01 .04 .11 –.277. Fit-specific experience—Technical 0.43 0.40 .09 –.01 .11 –.13 –.07 .158. Firm ageb 3.05 0.74 .10 .03 .02 .19 .03 .09 .069. Firm sizeb 7.90 2.38 .07 .07 .17 .27 .03 –.12 –.09 .37

10. Firm slack 0.69 0.64 –.06 .06 –.02 –.18 –.17 .03 .18 –.10 –.4311. Firm performance 0.78 2.69 .09 .08 .09 .19 .01 .12 .04 .38 .38 –.0512. CVC experienceb 2.86 1.27 .25 .09 –.06 .74 –.04 .04 –.15 .40 .39 –.20 .3613. CVC duration 7.32 5.61 .37 .05 –.10 .25 .01 .02 .02 .58 .36 –.09 .35 .44

a n � 375.b Logarithm.


CVC practices, controlling for time-varying omittedvariables to a large extent.

For each of the four dependent variables, weestimated an equation of the form

yit � �yit � 1 � �Xit � �Zit � �i � �it;

t � 1, 2, ..., T;|�| 1 (1)

where � is the autoregressive parameter on thelagged dependent term yit – 1, �i is the firm-specificfixed effect, �s are the coefficients of interest, X is avector of our three main independent variables,and Z is the vector of controls. In this specification,� is the short-run effect of an independent variable,

and�

�1�� is the long-run effect of the same vari-

able.11

Simple ordinary least squares (OLS) coefficientestimates are biased by the presence of �i (the firmfixed effect), while the fixed effects within-firmestimator, is consistent only in panels where thetime dimension T is large; the transformed laggeddependent variables are correlated with the trans-formed error term, but this correlation goes to zeroonly for large T’s. In our data set, this is not thecase, because we have on average six years of datafor each firm. Therefore, we use the Arellano-Bonddifference estimator (Arellano & Bond, 1991; Holtz-Eakin, Newey, & Rosen, 1988), which is designedfor dynamic panels with a short time dimension (T 10) and a larger firm dimension. The Arellano-Bond difference estimator employs the techniqueof generalized method of moments (GMM), whichhas an additional advantage over least-squarestechniques for proportion measures: it does notrequire the normality of the residuals, so it can beeasily applied to dependent variables that are pro-portion measures or are censored (as is the case forthe Herfindahl index), since GMM yields consis-tent and asymptotically normal estimates (Hansen,1982). Bond (2002) provides a detailed expositionof the methodology.

Estimation proceeds by first differencing thedata, which eliminates the firm-specific effects �i

from the model. First-differencing gives12

�yit � ��yit � 1 � ��Xit � ��Zit � ��it. (2)

Even though first-differencing eliminates thefixed effect, we still needed to account for two

endogeneity concerns. First, �yit � 1 is correlatednecessarily with ��it since yit � 1 is correlated with�it � 1. Second, our key independent variables maybe endogenous (Xit may be correlated with �it

which, in turn, implies that �Xit is correlated with��it), even when we account for time-invariantfixed effects. For instance, consider a firm that haspurely strategic objectives for its CVC unit. At somepoint, it might suddenly be inclined toward finan-cial objectives when it observes a buoyant IPO mar-ket or the significant financial returns that IVCs aregenerating by taking their portfolio companies pub-lic. It may then be inclined to change both thepersonnel staffing of its CVC unit and its invest-ment goals. The time-invariant firm fixed effectswould not pick up this shift in objectives. CVCincentive compensation schemes could also affectinvestment strategies and goals, and these couldalso be a source of endogeneity if they varied overtime for a given firm. Third, we had Directoryinformation for only three years, so our unit-levelcareer background measures did not include po-tential CVC managers whose tenure did not coverthose three years. We did collect actual start andend dates for each manager’s CVC job and usedthem to calculate the career background variablesfor our panel data. However, to the extent thatthere are managers who do not appear in the data,or errors or omissions in the biographical data wecollected measurement error could be present inthe career background variables; such error couldbias estimates if the omissions were nonrandom.To deal with all of these endogeneity concerns for�Xit (as well as for �yit � 1) and to account forpotential measurement error in the main inde-pendent variables, we employed instrumentalvariables. We identified valid instruments—vari-ables that are orthogonal (independent) to thetransformed error terms but are correlated withthe potentially troublesome variables. Arellanoand Bond (1991) showed that, provided the errorterm �it is serially uncorrelated and T � 3, thedifferenced-lagged dependent variable �yit � 1

may be instrumented by appropriately lagged lev-els of yi. They showed that the following momentconditions are valid:

E<yit � s��it= � 0;s � 2;t � 3, 4, ..., T.

For the independent variables, we allowed Xit tobe correlated with past error terms �it � s and thecurrent error term �it, but uncorrelated with futureerror terms �it � s, for s � 1. From Arellano and Bond(1991), this implies that values of X’s, lagged twoperiods or more, are available as instruments. In

11 In the long-run steady-state yit � yit�1 � y* so y* �

�y* � �Xit � �Zit � �i � �it ) y* X �

�

�1��12 An important advantage of first-differencing is that

it does not use multiple degrees of freedom to account forfixed effects.


turn, this implies that the following additional mo-ment conditions apply:13

E<Xit � s��it= � 0;s � 2;t � 3, 4, ..., T.

Arellano (2003) showed that an overfitting biascaused by instrument proliferation in dynamicpanels may be a concern. To mitigate this concern,we used only two lags as instruments in the mo-ment conditions. To control for heteroskedasticity,we reported results using the two-step GMM esti-mator and employed a finite-sample correction tothe two-step covariance matrix (Windmeijer, 2005).Finally, we performed an array of specificationtests suggested in Arellano and Bond (1991) to ex-amine the validity of our instruments.

In sum, the Arellano-Bond difference estimatorallows us to estimate a dynamic model using thelagged variable as a regressor, account for unob-served firm fixed effects via first-differencing, andresolve the endogeneity problems inherent in oursetting, and it is easily applied to truncated orbounded dependent variables. We used the “xta-bond” command in STATA for estimating Equation2 for each dependent variable.

RESULTS

Main Analysis

Table 2 shows estimates for models of CVC goalorientation (models 1 and 2) and CVC operationalstrategies (models 3 and 4). For practice-specificcareer experience, Hypothesis 1a predicts that CVCunits with higher proportion of CVC managers withIVC experience will be more oriented toward finan-cial goals. Model 1 shows that an increase in thefraction of CVC unit personnel with IVC experienceleads to greater financial orientation—that is, anincrease in start-ups’ exits via IPO or acquisition.At the same time, model 2 shows that these CVCunits are also less likely to be strategically oriented,in that their parent firms are less likely to acquirethe start-ups they have invested in. Thus, we ob-serve less modification of the goal orientation of theadopted practice, supporting Hypothesis 1a. Interms of effect size, we find that for goal orienta-tion, a one standard deviation increase in the pro-portion of CVC managers with IVC experience in-creases the proportion of exits via IPO and third-party acquisitions by 0.01 in the short run and by0.013 in the long run. For the median firm, this isequivalent to a 3.1 percent increase in financial

orientation in the short run and a 4.0 percent in-crease in the long run. In contrast, a one standarddeviation increase in IVC experience reduces stra-tegic orientation by 1.96 percent in the short runand by 4.17 percent in the long run. The largecoefficients on the lagged dependent variables forstrategic orientation imply that the impact of a onestandard deviation change in the independent vari-ables is nearly twice as large in the long run as theshort run.

Next, models 3 and 4 provide estimates for CVCoperational strategies. Hypothesis 1b states that thehigher the proportion of CVC managers with IVCexperience, the less the operational strategies oftheir CVC unit will be modified. Models 3 and 4provide strong support for Hypothesis 1b, showingthat an increase in the proportion of IVC experiencein a CVC unit results in an increase in the propor-tion of early-stage investments and in the industrydiversity of investments. Such operational strate-gies are aligned with IVC practices, indicating lessmodification in the operational strategies of theadopted practice. Our estimates imply that a onestandard deviation increase in IVC backgroundleads over time to an increase in the respectiveproportions of early-stage investments and sectoraldiversification of 0.016 and 0.01 in the short runand of 0.021 and 0.02 in the long run. For themedian firm, this is equivalent to more than a 15percent increase in the proportion of early-stageinvestments and a 1.8 percent increase in diversi-fication in the long run.

For organizational fit–specific career experience,Hypothesis 2a predicts that CVC units with ahigher proportion of CVC managers with adoptingfirm–specific experience are more likely to modifythe goals of the practice. We find that an increase inthe proportion of internal hires in a CVC unit re-sults in lower financial orientation (model 1) and ahigher strategic goal orientation (model 2), therebysupporting Hypothesis 2a. Here, a one standarddeviation increase in the proportion of internalhires reduces the proportion of exits via IPOs andacquisitions by 0.04 in the short run and by 0.05 inthe long run, and it raises the proportion of ownacquisitions by 0.04 in the short run and by 0.08 inthe long run. For the median firm, this amounts toa 15 percent reduction in financial orientation anda 4 percent increase in strategic orientation in thelong run.

Hypothesis 2b states that CVC units with a higherproportion of managers who have adopting firm–specific experience are more likely to modify theoperational strategies of their CVC unit. Model 3shows that an increase in the proportion of internalhires leads to an increase in the proportion of early-

13 Using lagged terms reduced the number of observa-tions to 375 and entailed the loss of two firms from oursample.


stage start-up investments, which is consistentwith standard IVC practice, thus contradicting Hy-pothesis 2b. However, model 4 also shows that CVCunits invest in a narrower range of industry sectors,as hypothesized. Together, we find mixed supportfor Hypothesis 2b. A one standard deviation in-crease in proportion of internal hires results in a 6percent increase in the proportion of early-stageinvestments and a 1.7 percent decline in diversifi-cation in the long run for the median firm.

For technical fit–specific experience, Hypothesis3a predicts that an increase in the proportion ofCVC managers with engineering experience willlead to more strategically oriented CVC units. Wefind that an increase in the proportion of personnelwith an engineering background in a CVC unitleads to a lower financial and a higher strategic goalorientation, supporting Hypothesis 3a. Our esti-mates imply that a one standard deviation increaseresults in a 6 percent reduction in financial orien-

TABLE 2Arellano-Bond Estimates of the Impact of Career Background on CVC Goal Orientation and Operational Strategiesa

Variables

Model 1 Model 2 Model 3 Model 4

Goal Orientation Operational Strategies

FinancialOrientation,

t

StrategicOrientation,

t

Proportion ofEarly-StageInvestments,

t

SectoralDiversification,

t

Practice-specific experience 0.03** –0.27** 0.05* 0.05**(0.00) (0.05) (0.03) (0.01)

Fit-specific experience—Organizational –0.10** 0.10** 0.02** –0.03*(0.02) (0.02) (0.00) (0.01)

Fit-specific experience—Technical –0.04** 0.17** 0.07** –0.04**(0.01) (0.06) (0.01) (0.00)

ControlsFirm age 0.00 0.01 –0.03 –0.19**

(0.06) (0.13) (0.05) (0.07)Firm size –0.01** –0.03* 0.02** 0.01**

(0.00) (0.01) (0.00) (0.00)Firm slack 0.01** 0.29** 0.02** 0.02**

(0.00) (0.08) (0.00) (0.00)Firm performance 0.00 0.00 -0.00 –0.00

(0.00) (0.01) (0.00) (0.00)CVC experience –0.03** –0.00** –0.04** 0.18**

(0.01) (0.00) (0.01) (0.01)CVC duration 0.04** 0.05 0.00 –0.03**

(0.00) (0.56) (0.00) (0.01)Financial orientation (t – 1) 0.22**

(0.02)Strategic orientation (t – 1) 0.53**

(0.02)Proportion of early-stage investments (t – 1) 0.22**

(0.02)Sectoral diversification (t – 1) 0.10**

(0.01)Number of observations 375 375 375 375Number of firms 68 68 68 68Overall model fit �2 (df) 6,398.75**

(25)148.72**(25)

2,175.57**(25)

3,118.35**(25)

Improvement in model fit �2 (df � 3) 1,466.10** 110.44** 294.80** 735.61**Specification test p-valuesSerial correlation: AR(1) test 0.01 0.01 0.02 0.68Serial correlation: AR(2) test 0.80 0.30 0.06 0.15Overidentification test 1.0 1.0 1.0 1.0

a Robust standard errors are in parentheses. All columns include year dummies; all variables are in first-differences, which account forfirm-fixed effects. The lagged differenced dependent variable and the career background variables are instrumented by their higher laggedlevels. The last three rows report specification tests for the Arellano-Bond difference estimator.

* p � .05** p � .01


tation and a 7.3 percent increase in strategic orien-tation in the long run for the median firm.

In Hypothesis 3b, we predict that operationalstrategies of CVC units are more likely to be modi-fied in CVC units with a higher proportion of CVCmanagers having engineering experience. Like ourinternal hires hypothesis, Hypothesis 3b receivedmixed support. An increase in the proportion ofmanagers with engineering experience in a CVCunit leads to a higher proportion of investments inearly-stage start-ups (model 3), suggesting no vari-ation from standard IVC practice on this particularoperational dimension. However, model 4 showsthat these CVC units also invest in a diverse rangeof start-ups, as predicted in Hypothesis 3b. Here, aone standard deviation increase in the proportionof engineers results in an increase (decrease) in theproportion of early-stage ventures (diversification)by 0.02 (0.01) in the short run and by 0.03 (0.02) inthe long run. For the median firm, this reflects a 27percent increase in the proportion of early-stagecompanies in the long run and a 1.9 percent de-crease in diversification in the long run.

In all four models, we observed that the laggeddependent variables are positive and significant,showing persistence in CVC goal orientation andoperational strategies. Each of the models in Table2 reports a Wald test for overall model fit showingthat all variables are jointly significant in models1–4, and each model is a significant improvementover a constant-only model. The next Wald test(improvement in model fit) shows that the threecareer experience variables result in a significantimprovement of model fit compared to a baselinemodel that includes only the control variables. Fi-nally, the last three rows in Table 2 report p-valuesfrom three specification tests suggested by Arellanoand Bond (1991). The tests indicate that the origi-nal error term �it is serially uncorrelated, since thep-values exceed 0.05 for each of the AR(2) tests, sothat the moment conditions are well-specified. Sar-gan overidentification tests suggested that our in-struments are orthogonal to the error term and sup-ported the validity of our instruments.

Robustness Checks

In the main analysis, we tracked both the portfo-lio of each CVC unit and the CVC background vari-ables from the date of establishment of the unit.However, as explained earlier, there is a possibilityof measurement error for our independent variablesof interest. We accounted for this by using thelagged dependent variable as a regressor and byinstrumenting each of the career background vari-ables; however, as a robustness check, we reesti-

mated the models, restricting the time span of thedata to 2000-02. This restriction addressed mea-surement error in the CVC background variablesbut introduced the possibility that the portfolioover the period 2000-02 might have been set inmotion by managers who left prior to 2000. The useof the lagged dependent variable alleviated thisconcern to a large extent, since it recognized thatthe characteristics of the portfolio in the year 2000are a function of the portfolio’s characteristics inthe previous year. Our results (available from theauthors upon request) are robust to the restrictedtime frame.

We also split our sample into two subgroupsbased on the median age of a CVC unit. Theresults for the subsamples did not reveal anyconsistent differences, and our results continuedto be supported in both subsamples, suggestingthat the effect of career experience is similar forboth newer CVC units and older CVC units.

DISCUSSION

The primary objective of this study was to in-vestigate the effect of individuals’ career experi-ences on the extent of practice variation. Theinstitutional perspective on practice adoption,with its theoretical emphasis on isomorphism,has only started to go beyond adoption decisionsto examine postadoption variation. Recent stud-ies have suggested field-level and organizationalinfluences that lead to practice variation (e.g.,Kennedy & Fiss, 2009; Lounsbury, 2001, 2007);however, qualitative case studies of practiceadoption have suggested that individuals whoimplement and manage practices also play animportant role in their interpretation and trans-lation (Boxenbaum & Battilana, 2005; Morris &Lancaster, 2006; Zbaracki, 1998). The currentstudy extends theory on practice variation byproviding a systematic way of predicting the in-fluence of individuals on the extent of practicevariation. We develop a theoretical frameworkthat specifies practice-specific and fit-specificprior career experiences of implementing manag-ers as predictors of how much practices will varyafter adoption. Further, our analytical approachestimates a dynamic model with firm fixed ef-fects, accounts for endogeneity concerns in ourkey independent variables, and uses instru-mented lagged dependent variables to capturepersistence in CVC practices (Arellano & Bond,1991). This, in turn, enables us to show individ-ual-level effects over and above any organization-level effects.


We conceptualized the extent of practice varia-tion in terms of both goal orientation and opera-tional strategies of practices. Drawing on organiza-tional learning theory, we argued that differenttypes of career experiences confer specific knowl-edge and mental models that in turn influencepractice variation. Practice-specific experienceleads to a more faithful replication of an adoptedpractice, while fit-specific experience enables as-sessment of fit between the practice and theadopter, leading to greater modification of the ad-opted practice. We tested hypotheses based on thesearguments by examining the implementation of IVCpractices by CVC managers from 70 CVC units from1992 through 2008. Our analysis provides broad sup-port for our predictions and for the usefulness ofexamining individual effects on practice variation us-ing a framework of practice- and fit-specific careerexperience.

The Role of Individuals in Practice Variation

First, we find that CVC units staffed with man-agers having practice-specific experience (that is,IVC experience), show a financial goal orienta-tion and a propensity to invest in early-stagestart-ups and in diverse industries, in keepingwith the IVC practice blueprint (Gompers & Ler-ner, 2004). For fit-specific experience, we inves-tigated two types of career experience that wouldenable assessment of organizational and techni-cal fit between practice and adopter: firm-spe-cific experience and engineering experience. Wefound that CVC units staffed with managers hav-ing these types of experience resulted in modifi-cation of the goal orientation of the practice fromfinancial to strategic, as expected. However, re-sults for the modification of operational strate-gies were mixed. Although increasing firm-spe-cific experience and increasing engineeringexperience in a CVC unit were significantly asso-ciated with investments in a narrower range ofindustries, both of these types of experience wererelated to investments in early-stage start-ups.That is, having career experience that enablesassessment of fit between a practice and anadopting organization is strongly related to prac-tice variation for the outcome of goal orientation,but only for one measure of operational strategy.

The mixed results for the effect of fit-specificcareer experience on operational strategies suggestthat a more nuanced view of operational strategiesunderlying practices is in order. Why wouldn’t alloperational strategies be subject to the same pres-sures for practice variation? One possible explana-tion is that operational strategies may not be clas-

sifiable into a single category, but should perhapsbe considered along additional dimensions. For ex-ample, certain operational strategies may be so cen-tral to the overall conception of a practice thatimplementing managers do not consider alterna-tives. Venture capital is distinct from other forms offinancing because VCs specialize not only in pro-viding funds to young start-ups, but also in guidingthem through early stages of development (Gomp-ers & Lerner, 2004). Accordingly, for venture capi-tal investment, the operational strategy of early-stage investing may be integral to the practice ofselecting investment targets. Alternatively, certainoperational methods of a practice could be equallyapplicable to different contexts and may requireminimal or no modification. Because CVC activityis often used to complement in-house R&D, theoperational strategy of investing in early-stage start-ups could accelerate access to emerging technolo-gies and markets and enable increased knowledgecreation and a higher rate of technological innova-tion. Future research should further explore differ-ent types of operational strategies and their adap-tation to new contexts.

Examining the pattern of results across models,we also find that differing goal orientations forCVC managers might result in outcomes that needto be traded off against each other. Specifically,we expected the proportion of implementingmanagers with practice-specific experience topositively affect financial goal orientation, butwe also found that it negatively affects strategicgoal orientation. Similarly, we found that al-though fit-specific experience positively affectsthe strategic goal orientation of CVC units, it alsonegatively affects their financial orientation. Wedid not hypothesize such a trade-off; however,there is some conceptual support for it. Manage-rial attention is a limited resource (Simon, 1976),so restriction of attention to a financial orienta-tion might preclude attention to strategic con-cerns, and vice versa. Even if CVC units couldhave only one orientation, strategic and financialgoal orientations are not necessarily at odds. Astrategically valuable acquisition could be finan-cially neutral or even provide financial benefit,or investments that pay off with financial returnscan yield strategically valuable technologicalknowledge. Moreover, the negative and signifi-cant relationship between our variables for prac-tice-specific experience and strategic orientationsuggests a tension between practice expertise andthe needs of adopting firms. Hiring practice ex-perts would seem to be a sensible way of imple-menting new practices, yet our results suggestthat they may exhibit cognitions and behaviors


more suitable to an IVC context than to a corpo-rate context. Future work could explore the rela-tionship between financial and strategic orienta-tions, goals, and outcomes to understand theconditions under which trade-offs occur.

Exploring Practice Variation across Levels ofAnalysis

Our study extends theory on practice adoptionand variation by focusing on individual-level de-terminants. We confined our interest to individ-ual-level determinants to the extent that we con-trolled extensively for contextual elements,including fixed effects that removed stable char-acteristics of the adopting organizations fromconsideration, time-varying organizational andunit controls, and lagged dependent variablesthat accounted for persistence and path depen-dency in CVC practices. By holding organiza-tional context strongly constant, we have shownthat individual-level effects operate over andabove any effects at higher levels of analysis.However, organizational and institutional factorshave also been shown to affect practice variation,and we missed an opportunity to consider howthese determinants at multiple levels of analysismay interact. For example, Lounsbury (2001)found that field-level ecological organizations in-fluenced the way in which colleges implementedrecycling programs. Similarly, field-level organi-zations such as professional associations maychange how implementing managers apply theircareer experiences to adopted practices. For in-stance, professional associations for venture cap-italists may provide a vehicle for CVC managerswith engineering backgrounds to meet and learnfrom IVCs, counteracting the influence of theirfit-specific experience. On the other hand, suchan association could positively interact with IVCbackgrounds to reduce variation in venture cap-ital practices even further.

Another way that individual- and organization-level factors could interact concerns existing or-ganizational practices. In their application ofagency theory to CVC investment, Dushnitskyand Shapira (2010) found that higher-powered,incentive-based compensation schemes are asso-ciated with earlier-stage investment for CVCunits. Organizational compensation schemesmay interact with career background in interest-ing ways. High-powered incentive compensationis congruent with IVC practices, so CVC manag-ers with fit-specific experience may instigate lesspractice variation in goal orientation and opera-tional strategies in organizations that use such

incentives. An implication for studies of practicevariation is that existing practices of adoptingorganization may be more or less aligned with thesource of the adopted practice, reducing the needto vary practices to fit.

Limitations and Directions for Future Study

We have theorized that prior experience con-ditions specific knowledge and mental modelsthat drive the extent of practice variation. How-ever, our archival methods in this study did notallow for direct examination of the individual-level mechanisms that drove variation. Ratherthan mechanisms that involve learning and cog-nition, mechanisms having to do with other in-dividual-level characteristics might actually bethe causal factor. For instance, prior experienceconditions social identity and social networks aswell as knowledge and mental models, and thepropensity to modify practices might be a func-tion of whom CVC managers have strong socialties to or identify with, as opposed to what theyknow from their own past experiences. Researchto tease apart these individual-level mechanismswould be challenging to conduct, but it wouldinvolve direct contact with individuals to under-stand the way they think about their prior careersand how they interact with others. Next, uniquefeatures of the practice we studied may haveinfluenced our findings. VC practices are knownto be complex and poorly theorized (Gaba &Meyer, 2008), so that implementers in a corporatecontext need to translate the broad directives ofthe practices into operational actions. Thus, ourresults may not generalize to practices that aremore thoroughly theorized or that entail feweropportunities for job crafting or role innovation(Van Maanen & Schein, 1979; Wrzesniewski &Dutton, 2001). Finally, our research aggregatedthe experiences of personnel in CVC units, allow-ing us to separately examine different dimen-sions of experience for all personnel, but it ne-glects potentially interesting effects of group-level variables, such as the heterogeneity offunctional experience or prior employers, both ofwhich have been shown to affect organization-level outcomes (Beckman, Burton, & O’Reilly,2007; Higgins & Gulati, 2006). Though CVC unitsdo not function as interdependent teams in theway that top management teams or cross-func-tional teams do, examining how the backgroundsof CVC managers interact within a unit might beinteresting for future study.


Career Backgrounds, Human Capital, andCorporate Venture Capital

Our study primarily contributes to the under-standing of practice variation; however, the find-ings also have implications for research on theportability of human capital across organiza-tional boundaries. Studies in organizationallearning find that individuals bring their humancapital with them when they move to a new em-ployer, adding diverse knowledge and perspec-tives that result in innovation (Rao & Drazin,2002; Song, Almeida, & Wu, 2003). Our findingssuggest that career experiences outside of a cur-rent firm may actually lead to less innovation inpractices if a new hire has prior experience withthe practice. Second, human capital studies haveemphasized firm boundaries as an impediment tothe portability of experience from one context toanother (Groysberg et al., 2008; Huckman & Pi-sano, 2006). Our findings indicate that experi-ence in a practice or occupation also confers spe-cific knowledge that is portable across firmboundaries. Our emphasis on career experienceis also consistent with upper echelons theory,which addresses the functional backgrounds ofCEOs as sources of the knowledge and mentalmodels that influence strategic change in organ-izations (Boeker, 1997; Geletkanycz & Hambrick,1997). We provide evidence that the influence ofcareer experience on organizational outcomescan be extended beyond upper echelons to sub-units and lower hierarchical levels of firms. Fi-nally, our attention to the whole careers of indi-viduals suggests interesting extensions of currentwork on individual experience and organiza-tional outcomes. Rather than treating individualsas belonging to a single occupation or employer,we capture experiences that reflect the complex-ity and diversity of modern careers (Bidwell &Briscoe, 2010).

In addition, recent research in corporate venturecapital has provided insights into the role that CVCunits play in the business and innovation strategiesof firms (Dushnitsky & Lenox, 2006); however, rel-atively little is known about the determinants ofCVC unit investment practices. Our study supple-ments the CVC literature by showing that the careerbackgrounds of individuals managing CVC unitsare consequential for how these units conduct ven-ture capital practices.

Our study presents additional evidence thatthe specific characteristics of individuals shapethe activities and outcomes of the firms that em-ploy them. Rather than presenting an image ofabstract organizational actors, we bring individ-

uals into the foreground as important players inpostadoption practice variation. Specifically, in-dividuals’ career experiences influence the ex-tent to which adopting firms modify practices.Diffusing practices are not necessarily adoptedwithout adjustment; they can be transformed bythe managers who implement them.

REFERENCES

Abernathy, W. J., & Utterback, J. M. 1978. Patterns ofindustrial innovation. Technology Review, 80(7):40–47.

Abrahamson, E., & Rosenkopf, L. 1997. Social network ef-fects on the extent of innovation diffusion: A computersimulation. Organization Science, 8: 289–309.

Almeida, P., Dokko, G., & Rosenkopf, L. 2003. Startupsize and the mechanisms of external learning: In-creasing opportunity and decreasing ability? Re-search Policy, 32: 301–315.

Ansari, S. M., Fiss, P. C., & Zajac, E. J. 2010. Made to fit:How practices vary as they diffuse. Academy ofManagement Review, 35: 67–92.

Arellano, M. 2003. Panel data econometrics. Oxford,U.K. & New York: Oxford University Press.

Arellano, M., & Bond, S. 1991. Some tests of specificationfor panel data: Monte-Carlo evidence and an appli-cation to employment equations. Review of Eco-nomic Studies, 58: 277–297.

Argote, L. 1999. Organizational learning: Creating, re-taining, and transferring knowledge. Boston: Klu-wer Academic.

Arthur, M. B., & Rousseau, D. M. 1996. The boundary-less career: A new employment principle for a neworganizational era. New York: Oxford UniversityPress.

Audia, P. G., & Greve, H. R. 2006. Less likely to fail: Lowperformance, firm size, and factory expansion in theshipbuilding industry. Management Science, 52:83–94.

Beckman, C. M., & Burton, M. D. 2008. Founding thefuture: Path dependence in the evolution of top man-agement teams from founding to IPO. OrganizationScience, 19, 3–24.

Beckman, C. M., Burton, M. D., & O’Reilly, C. 2007. Earlyteams: The impact of team demography on VC fi-nancing and going public. Journal of Business Ven-turing, 22: 147–173.

Benson, D., & Ziedonis, R. H. 2010. Corporate venturecapital and the returns to acquiring portfolio com-panies. Journal of Financial Economics, 98: 478–499.

Bidwell, M., & Briscoe, F. 2010. The dynamics of interor-


ganizational careers. Organization Science, 21:1034–1053.

Birkinshaw, J., Hamel, G., & Mol, M. J. 2008. Managementinnovation. Academy of Management Review, 33:825–845.

Boeker, W. 1997. Executive migration and strategicchange: The effect of top manager movement onproduct-market entry. Administrative ScienceQuarterly, 42: 213–236.

Boh, W. F., Slaughter, S. A., & Espinosa, J. A. 2007.Learning from experience in software development:A multilevel analysis. Management Science, 53:1315–1331.

Bond, S. 2002. Dynamic panel data models: A guide tomicro data methods and practice. Portugese Eco-nomic Journal, 1: 141–162.

Boxenbaum, E., & Battilana, J. 2005. Importation as inno-vation: Transposing managerial practices acrossfields. Strategic Organization, 3: 355–383.

Boxenbaum, E., & Jonsson, S. 2008. Isomorphism, diffu-sion and decoupling. In K. Sahlin-Andersson, R.Greenwood, C. Oliver, & R. Suddaby (Eds.), TheSage handbook of organizational institutionalism:218–242. Los Angeles & London: Sage.

Bromiley, P. 1991. Testing a causal model of corporaterisk-taking and performance. Academy of Manage-ment Journal, 34: 37–59.

Burton, M. D., Sorensen, J. B., & Beckman, C. 2002. Com-ing from good stock: Career histories and new ven-ture formation. In M. Lounsbury & M. Ventresca(Eds.), Research in the sociology of organizations:Social structure and organizations revisited, vol.19. Greenwich, CT: JAI Press.

Campbell, J. L. 2005. Where do we stand? Commonmechanisms in organizations and social movementsresearch. In G. F. Davis, D. McAdam, W. R. Scott, &M. Zald (Eds.), Social movements and organizationtheory. New York: Cambridge University Press.

Carlile, P. R. 2004. Transferring, translating, and trans-forming: An integrative framework for managingknowledge across boundaries. Organization Sci-ence, 15: 555–568.

Chesbrough, H. 2003. Open innovation: The new imper-ative for creating and profiting from technology.Boston: Harvard Business School Press.

Coburn, C. E. 2004. Beyond decoupling: Rethinking therelationship between the institutional environment andthe classroom. Sociology of Education, 77: 211–244.

Czarniawska-Joerges, B., & Sevón, G. 2005. Global ideas:How ideas, objects and practices travel in a globaleconomy. Malmö; & Herndon, VA: Liber; Copenha-gen Business School Press.

Dacin, M. T., Goodstein, J., & Scott, W. R. 2002. Institu-

tional theory and institutional change: Introductionto the special research forum. Academy of Manage-ment Journal, 45: 45–56.

Davis, G. F., & Greve, H. R. 1997. Corporate elite net-works and governance changes in the 1980s. Amer-ican Journal of Sociology, 103: 1–37.

Dewar, R. D., & Dutton, J. E. 1986. The adoption of radicaland incremental innovations—An empirical-analy-sis. Management Science, 32: 1422–1433.

Dokko, G., Wilk, S. L., & Rothbard, N. P. 2009. Unpackingprior experience: How career history affects job per-formance. Organization Science, 20: 51–68.

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

Dushnitsky, G., & Lenox, M. J. 2006. When do incum-bents learn from entrepreneurial ventures? Corpo-rate venture capital and investing firm innovationrates. Research Policy, 34: 615–639.

Dushnitsky, G., & Shapira, Z. 2010. Entrepreneurial fi-nance meets organizational reality: Comparing in-vestment practices and performance of corporateand independent venture capitalists. Strategic Man-agement Journal, 31: 990–1017.

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

Ernst & Young. 2008. Global corporate venture capitalsurvey, 2008-09.

Ferlie, E., Fitzgerald, L., Wood, M., & Hawkins, C. 2005.The nonspread of innovations: The mediating role ofprofessionals. Academy of Management Journal,48: 117–134.

Fiss, P. C., & Zajac, E. J. 2004. The diffusion of ideas overcontested terrain: The (non)adoption of a shareholdervalue orientation among German firms. Administra-tive Science Quarterly, 49: 501–534.

Gaba, V., & Meyer, A. D. 2008. Crossing the organiza-tional species barrier: How venture capital practicesinfiltrated the information technology sector. Acad-emy of Management Journal, 51: 976–998.

Geletkanycz, M. A., & Hambrick, D. C. 1997. The externalties of top executives: Implications for strategicchoice and performance. Administrative ScienceQuarterly, 42: 654–681.

Gioia, D. A., & Poole, P. P. 1984. Scripts in organizationalbehavior. Academy of Management Review, 9:449–549.

Gompers, P. A. 1995. Optimal investment, monitoring,and the staging of venture capital. Journal of Fi-nance, 50: 1461–1489.


Gompers, P. A., & Lerner, J. 2004. The venture capitalcycle (2nd ed.). Cambridge, MA: MIT Press.

Greene, W. H. 2000. Econometric analysis (4th ed.).Upper Saddle River, NJ: Prentice Hall.

Groysberg, B., Lee, L.-E., & Nanda, A. 2008. Can they takeit with them? The portability of star knowledgeworkers’ performance. Management Science, 54:1213–1230.

Hansen, L. P. 1982. Large sample properties of general-ized-method of moments estimators. Econometrica,50: 1029–1054.

Higgins, M. C., & Gulati, R. 2006. Stacking the deck: Theeffects of top management backgrounds on investordecisions. Strategic Management Journal, 27:1–25.

Holtz-Eakin, D., Newey, W., & Rosen, H. S. 1988. Esti-mating vector autoregressions with panel data.Econometrica, 56: 1371–1395.

Huckman, R. S., & Pisano, G. P. 2006. The firm specificityof individual performance: Evidence from cardiacsurgery. Management Science, 52: 473–488.

Jonsson, S. 2009. Refraining from imitation: Professionalresistance and limited diffusion in a financial mar-ket. Organization Science, 20: 172–186.

Kennedy, M. T., & Fiss, P. C. 2009. Institutionalization,framing, and diffusion: The logic of TQM adoptionand implementation decisions among U.S. hospi-tals. Academy of Management Journal, 52: 897–918.

Kraatz, M. S., & Moore, J. H. 2002. Executive migrationand institutional change. Academy of ManagementJournal, 45: 120–143.

Lant, T. K., Milliken, F. J., & Batra, B. 1992. The role ofmanagerial learning and interpretation in strategicpersistence and reorientation—An empirical explo-ration. Strategic Management Journal, 13: 585–608.

Levinthal, D., & March, J. G. 1981. A model of adaptiveorganizational search. Journal of Economic Behav-ior & Organization, 2: 307–333.

Lounsbury, M. 2001. Institutional sources of practicevariation: Staffing college and university recyclingprograms. Administrative Science Quarterly, 46:29–56.

Lounsbury, M. 2007. A tale of two cities: Competinglogics and practice variation in the professionalizingof mutual funds. Academy of Management Journal,50: 289–307.

Lounsbury, M. 2008. Institutional rationality and prac-tice variation: New directions in the institutionalanalysis of practice. Accounting, Organizations andSociety, 33: 349–361.

Macmillan, I., Roberts, E., Livada, V., & Wang, A. 2008.Corporate venture captial [CVC]: Seeking innova-tion and strategic growth. NIST GCR 08-916, Na-tional Institute of Standards and Technology, U.S.Department of Commerce.

March, J. G., & Shapira, Z. 1987. Managerial perspectiveson risk and risk-taking. Management Science, 33:1404–1418.

March, J. G., & Simon, H. A. 1958. Organizations. NewYork: Wiley.

Metrick, A., & Yasuda, A. 2010. Venture capital and thefinance of innovation (2nd ed.). Hoboken, NJ: Wi-ley.

Meyer, A. D., & Goes, J. B. 1988. Organizational assimi-lation of innovations: A multilevel contextual anal-ysis. Academy of Management Journal, 31: 897–923.

Miner, A. S., & Mezias, S. J. 1996. Ugly duckling no more:Pasts and futures of organizational learning research.Organization Science, 7: 88–99.

Morris, T., & Lancaster, Z. 2006. Translating managementideas. Organization Studies, 27: 207–233.

Nigam, A., & Ocasio, W. 2010. Event attention, environ-mental sensemaking, and change in institutional log-ics: An inductive analysis of the effects of publicattention to Clinton’s health care reform initiative.Organization Science, 21: 823–841.

Palmer, D. A., Jennings, P. D., & Zhou, X. G. 1993. Lateadoption of the multidivisional form by large UnitedStates corporations—Institutional, political, andeconomic accounts. Administrative Science Quar-terly, 38: 100–131.

Rao, H., & Drazin, R. 2002. Overcoming resource con-straints by recruiting talent from rivals: A study ofrecruitment and product innovation in the mutualfund industry 1986-1994. Academy of ManagementJournal, 45: 491–507.

Sahlin, K., & Wedlin, L. 2008. Circulating ideas: Imita-tion, translation and editing. In K. Sahlin-Ander-sson, R. Greenwood, C. Oliver & R. Suddaby (Eds.),The Sage handbook of organizational institution-alism: 218–242. Los Angeles & London: Sage.

Sahlin-Andersson, K. 1996. The construction of organi-zational fields. In B. Czarniawska & G. Sevon (Eds.),Translating organizational change: 69–92. Berlin& NY: Walter de Gruyter.

Shane, S., & Khurana, R. 2003. Bringing individuals backin: The effects of career experience on new firmfounding. Industrial and Corporate Change, 12:519–543.

Simon, H. A. 1976. Administrative behavior: A study ofdecision-makingprocesses inadministrativeorgan-ization (3rd ed.). New York: Free Press.


Singh, J., & Agrawal, A. 2011. Recruiting for ideas: Howfirms exploit the prior inventions of new hires. Man-agement Science, 57: 129–150.

Song, J., Almeida, P., & Wu, G. 2003. Learning-by-hiring:When is mobility more likely to facilitate interfirmknowledge transfer? Management Science, 49: 351–365.

Sorensen, J. B., & Stuart, T. E. 2000. Aging, obsolescence,and organizational innovation. Administrative Sci-ence Quarterly, 45: 81–112.

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

Strang, D., & Jung, D. I. 2009. Participatory improvementat a global bank: The diffusion of quality teams andthe demise of a six sigma initiative. OrganizationStudies, 30: 31–53.

Strang, D., & Meyer, J. W. 1993. Institutional conditionsfor diffusion. Theory and Society, 22: 487–511.

Strang, D., & Soule, S. A. 1998. Diffusion in organizationsand social movements: From hybrid corn to poisonpills. In J. Hagan & K. Cook (Eds.), Annual review ofsociology, vol. 24: 265–290. Palo Alto, CA: AnnualReviews.

Van Maanen, J., & Schein, E. H. 1979. Toward a theory oforganizational socialization. In L. L. Cummings &B. M. Staw (Eds.), Research in organizational be-havior, vol. 1: 209–264. Greenwich, CT: JAI.

Venture Economics 2001. http://ventureeconomics.com/.

Wezel, F. C., Cattani, G., & Pennings, J. M. 2006. Com-

petitive implications of interfirm mobility. Organi-zation Science, 17: 691–709.

Windmeijer, F. 2005. A finite sample correction for thevariance of linear efficient two-step GMM estima-tors. Journal of Econometrics, 126: 25–51.

Wooldridge, J. 2002. Econometric analysis of cross sec-tion and panel data. Cambridge, MA: MIT Press.

Wrzesniewski, A., & Dutton, J. E. 2001. Crafting a job:Revisioning employees as active grafters of theirwork. Academy of Management Review, 26: 179–201.

Zbaracki, M. J. 1998. The rhetoric and reality of totalquality management. Administrative Science Quar-terly, 43: 602–636.

Gina Dokko ([email protected]) is an assistant profes-sor at the Graduate School of Management at the Univer-sity of California, Davis. She received her Ph.D. from theWharton School at the University of Pennsylvania. Herresearch addresses the consequences of individuals’ jobmobility and career histories for innovation, learning,and performance.

Vibha Gaba ([email protected]) is an assistant pro-fessor of organizational behavior at INSEAD, Singapore.She received her Ph.D. in management from LundquistCollege of Business at the University of Oregon. Hercurrent research includes corporate venture capital, rela-tionships between internal and external R&D, and organ-izational change and innovation.


Copyright of Academy of Management Journal is the property of Academy of Management and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

venturing into new territory: career experiences of corporate venture capital … · 2014. 9....

Documents