innovation search of new ventures in a …yanzh/innovation search_smj 2010.pdf · service...

Strategic Management JournalStrat. Mgmt. J., 31: 88–109 (2010)

Published online EarlyView in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.806

Received 1 February 2008; Final revision received 29 July 2009

INNOVATION SEARCH OF NEW VENTURES IN ATECHNOLOGY CLUSTER: THE ROLE OF TIES WITHSERVICE INTERMEDIARIES

YAN ZHANG* and HAIYANG LIJones Graduate School of Business, Rice University, Houston, Texas, U.S.A.

In this study, we examine the relationships between new ventures’ ties with service intermediaries(i.e., technology service firms, accounting and financial service firms, law firms, and talentsearch firms) and their product innovation in the context of a technology cluster. Because serviceintermediaries sit at the intersection of many firms, organizations and industries, they maintainextensive networks in a cluster. We propose that new ventures’ ties with service intermediariesenable the ventures to plug into these networks and contribute to the ventures’ product innovationby broadening the scope of their external innovation search and reducing their search cost.Moreover, we argue that the positive relationships between new ventures’ ties with serviceintermediaries and their product innovation will become stronger when search in the networksin the cluster is more important to the ventures’ product innovation. Based upon a sample ofnew ventures in a technology cluster in China, our results support these arguments. Copyright 2009 John Wiley & Sons, Ltd.

INTRODUCTION

New ventures characteristically suffer from the lia-bility of newness because they have not establishedrelationships with outside actors and have lim-ited resources (Stinchcombe, 1965). For this rea-son, Schoonhoven, Eisenhardt, and Lyman (1990:177) have explicitly argued that successful prod-uct innovation is important for new ventures togain cash flow for financial independence, togain external visibility and legitimacy, and toimprove the likelihood of survival. This is espe-cially true in technology industries where dynamic

Keywords: ties with service intermediaries; product inno-vation; new ventures; technology cluster; emergingmarket∗ Correspondence to: Yan Zhang, Jones Graduate School of Busi-ness, Rice University, Houston, TX 77 005-1892, U.S.A.E-mail: [email protected]

market and technology changes require the ven-tures to constantly deal with environmental volatil-ity, to increase development speed, and to establishnew markets and technologies (Katila and Shane,2005; Li and Atuahene-Gima, 2001, 2002; Smith,Collins, and Clark, 2005).

The extant literature on innovation has high-lighted the importance of search (Cohen andLevinthal, 1990; Fleming, 2001; Katila, 2002;Katila and Ahuja, 2002; Rosenkopf and Nerkar,2001; Smith, et al., 2005). Innovation search isa problem-solving activity in which firms solveproblems through combining knowledge elementsto create new products (Katila, 2002). The knowl-edge space for firm innovation search has bothinternal sector (i.e., knowledge created within thesearching firm) and external sector (i.e., knowl-edge created by others) (Katila, 2002; Rosenkopfand Nerkar, 2001). Due to new ventures’ shorthistory and limited resources, their internal knowl-edge space can be limited. Thus, it is not surprising

Copyright 2009 John Wiley & Sons, Ltd.

Service Intermediaries and New Ventures Innovation Search in a Cluster 89

that the extant literature has emphasized the impor-tant role of new ventures’ external ties in theirinnovation, presumably because external ties canfacilitate their external innovation search. It hasbeen found that new ventures’ ties with estab-lished firms (e.g., Baum, Calabrese, and Silver-man, 2000; Shan, Walker, and Kogut, 1994; Stuart,2000) and universities and research institutes (e.g.,Baum et al., 2000) have positive impact on theirinnovation.

While external ties with established firms, uni-versities, and research institutes are important, newventures may face considerable difficulty in devel-oping ties with these prominent organizations.As noted by Stinchcombe (1965), new venturesmay find it difficult to access interfirm networksbecause they have a short history, lack of a provenperformance record, resources, legitimacy, and sta-tus. Indeed, recent research on innovation startsto pay attention to the role of service intermedi-aries (e.g., Hargadon and Sutton, 1997; Howells,2006; Wolpert, 2002). Service intermediaries referto professional service organizations that providefirms with supporting services in areas such asaccounting and finance, talent search, law, andtechnology services (e.g., technology commercial-ization and brokering). Because service interme-diaries are potentially available to all firms, theyare accessible to new ventures. More important,because service intermediaries sit at the intersec-tion of many firms, organizations, and industries,they can facilitate the exchange of informationabout innovation among firms (Wolpert, 2002: 78).The role of service intermediaries is particularlycritical in technology clusters. Bahrami and Evans(1995) argued that in technology clusters serviceintermediaries, together with private firms, univer-sities and research institutes, and venture capi-talists, form an industrial ecology that facilitatesand disciplines the development of technologyfirms. Saxenian (1990: 96) argued that by shar-ing costs and risks and pooling technical exper-tise service intermediaries allow Silicon Valley’sspecialist firms to continue to innovate and reactflexibly.

Despite the potential important role of serviceintermediaries in new venture innovation, theoret-ical or empirical work in this area is very lim-ited. To advance the literature, we examine howand under what conditions new ventures’ ties withservice intermediaries are related to their product

innovation in the context of a technology clus-ter. We define product innovation as the extentto which new ventures can successfully developnew products with superior quality and in a speedymanner for market penetration. Our premise is thatbecause service intermediaries sit at the intersec-tion of many firms, organizations, and industries,they maintain extensive networks of ties to differ-ent parts of the social system of a cluster. There-fore, new ventures’ ties with service intermediariesenable the ventures to plug into the networks andcontribute to the ventures’ product innovation bybroadening the scope of their external innovationsearch and reducing their search cost.

Furthermore, because ties with service inter-mediaries may contribute to new ventures’ prod-uct innovation by broadening their external searchscope and reducing their search cost, we wouldexpect that these positive effects will be strongerif search in the networks in the cluster in whichthe ventures reside is more important to the ven-tures’ product innovation. Based upon this argu-ment, we will examine the moderating roles ofventure managers’ perceived industry growth, theventures’ ties with foreign firms, and the ventures’export. These moderating effects, if supported, cangreatly substantiate and provide stronger supportto our argument on the effect of ties with serviceintermediaries on product innovation of new ven-tures.

Empirically, we will test these relationships witha sample of new ventures in a technology clus-ter in China. The emergence of a market econ-omy in China alters the opportunities and incentivestructures that have been shaped by a centrallyplanned economy and stimulates the growth ofnew ventures, particularly in technology industries(Li and Atuahene-Gima, 2001; Li and Hitt, 2006;Lu, 2000). However, China’s emerging economy isstill characterized by having volatile environmentsand a lack of institutions and strategic marketfactors to support business and innovation (Luo,2003). Because institutional voids often handicapentrepreneurial activities and innovation of newventures during economic transformation (Li andZhang, 2007), ties with service intermediaries inChina are deemed to play an important role. Thus,China’s emerging economy provides a rich contextto test the proposition that ties with service inter-mediaries can contribute to new ventures’ productinnovation.

Copyright 2009 John Wiley & Sons, Ltd. Strat. Mgmt. J., 31: 88–109 (2010)DOI: 10.1002/smj

90 Y. Zhang and H. Li

THEORY DEVELOPMENTAND HYPOTHESES

Ties with service intermediaries and newventure product innovation in a technologycluster

According to Porter (1998: 78), clusters refer to‘geographic concentrations of interconnected com-panies and institutions in a particular field.’ Clus-ters, particularly technology clusters, can drivethe direction and pace of innovation by mak-ing innovation opportunities more visible andproviding convenient access to complementaryresources and capabilities (Bahrami and Evans,1995; Porter, 1998; Saxenian, 1990; Zhang, Li,and Schoonhoven, 2009). However, the mere colo-cation of firms and institutions in a cluster onlycreates the potential for economic value and doesnot necessarily ensure its realization (Porter, 1998).For a residing firm to access important resourcesand information in a cluster and realize the poten-tial economic benefits offered by the cluster, thefirm has to be connected locally (Owen-Smith andPowell, 2004; Porter, 1998).

McEvily and Zaheer (1999: 1139) argue that inclusters, in addition to interfirm networks, firmsare also embedded in their ties with regional insti-tutions. Regional institutions not only provide spe-cific support services but also act as network inter-mediaries for interaction and information exchangeamong residing firms (McEvily and Zaheer, 1999).Similarly, Saxenian (1990) pointed out that SiliconValley is not only distinguished by the concentra-tion of skilled labor and firms, but also by a varietyof regional institutions—including Stanford Uni-versity, trade associations, and a myriad of special-ized consulting, market research, public relations,and venture capital firms. These regional institu-tions as intermediaries provide technical, financial,and networking services that firms usually can-not afford individually, and allow these firms tocontinue to innovate (Saxenian, 1990). As notedby Porter (1998), fostering ongoing relationshipswith service intermediaries in a cluster is an impor-tant way for firms to tap into the cluster’s assets.Empirically, McEvily and Zaheer (1999) foundthat the level of a firm’s ties with regional insti-tutions (i.e., regional industrial extension centers)in a cluster is positively related to the firm’sacquisition of competitive advantages. However,we know little about whether ties with a varietyof regional institutions (or service intermediaries)

may be related to new ventures’ product innovationand if so, under what conditions?

To advance this line of research, this studyempirically investigates the relationships betweennew ventures’ ties with service intermediaries andtheir product innovation in the context of a tech-nology cluster. Built upon the literature on inno-vation search (Cohen and Levinthal, 1990; Flem-ing, 2001; Katila, 2002; Katila and Ahuja, 2002;Rosenkopf and Nerkar, 2001), we propose that in atechnology cluster, new ventures’ ties with serviceintermediaries can contribute to their product inno-vation by broadening their external search scopeand reducing their search cost.

In today’s business environment, which is char-acterized by globalization, industry convergence,and rapid technology change, ‘no company issmart enough to know what to do with every newopportunity it finds, and no company has enoughresources to pursue all the opportunities it mightexecute’ (Wolpert, 2002: 80). This is why externalinnovation search becomes increasingly important;and, because of their limited internal resources, isparticularly important for new ventures. Throughexternal innovation search, new ventures becomeinnovative by capturing knowledge spillovers fromother firms and organizations (Katila, 2002). How-ever, new ventures can be handicapped in exter-nal innovation search. First, new ventures tend tohave narrow external search scope because theytypically have limited external contacts (Stinch-combe, 1965), and therefore rely upon their imme-diate personal networks (such as previous workingrelationships, voluntary connections, and kinshipand community ties) for identifying opportunities(Baker and Nelson, 2005; Starr and MacMillan,1990).

Second, new ventures can be handicapped inexternal innovation search for the reason of searchcost. Because new ventures have limited resources,even modest external search can be too costly forthem (in terms of both financial and organizationalresources and managerial time). Given these con-straints, it is critical for new ventures to balancethe needs and the costs of information search forproduct innovation in a dynamic and uncertainenvironment (Tushman and Nadler, 1978).

Ties with service intermediaries can help newventures balance their needs and costs of exter-nal innovation search. As an important feature ofthe local infrastructure of a cluster, service inter-mediaries uniquely sit at the intersection of many



firms and industries (Wolpert, 2002). Because theyinteract with a relatively large number of residingfirms and organizations, they maintain extensivenetworks of ties to different parts of the socialsystem of the cluster (McEvily and Zaheer, 1999;Saxenian, 1990). They act as repositories for infor-mation, knowledge, and opportunities in the cluster(Wolpert, 2002). If a new venture has close tieswith service intermediaries, the specialized net-work role of service intermediaries enables theventure to plug into the networks in the cluster,through which the venture becomes (indirectly)tied with different parts of the social system ofthe cluster. In other words, ties with service inter-mediaries provide an entry ticket for new venturesto join rich networks in the cluster.

More specifically, ties with service intermedi-aries can broaden new ventures’ external searchscope. Through their interactions with a relativelylarge number of firms and organizations in a clus-ter, service intermediaries can gather and channelinformation regarding what products other firmsoffer, what resources and capabilities they have,what problems they face in product innovation, andhow they solve problems. As illustrated by Wolpert(2002: 78), ‘If company A, for instance, needs out-side capabilities to commercialize a technology,it could ask its intermediary to find it a partner.The intermediary would share the information withother intermediaries in its search for an appropriatecollaborator. . ..’1 In other words, service interme-diaries’ specialized network role enables new ven-tures to conduct broad external innovation searchin the cluster rather than searching only in theirimmediate personal networks.

A broadened external search scope can con-tribute to new ventures’ product innovation in threemajor ways. First, innovation is an information-intensive activity. Information regarding otherfirms’ product offerings and innovation activitiescan make innovation opportunities more visible tonew ventures (Ahuja, 2000). Second, a broadenedexternal search scope can enrich a new venture’sknowledge pool and provide more choices forthe venture to solve problems (Katila and Ahuja,2002). There is a limit to the number of new prod-ucts that can be created by using the same set ofknowledge elements. Search with a broad scope

1 Wolpert (2002: 78) further noted, ‘The intermediaries can betrusted to maintain confidentiality because if they ever violatedthe terms of an arrangement no company would hire them again.’

can increase a venture’s new product introductionby adding new elements in its knowledge pooland thus improve the possibility for it to find newuseful combinations of these elements (Katila andAhuja, 2002). Third, a broadened external searchscope can also help new ventures locate externalcomplementary resources and capabilities that arecritical for their product innovation (Porter, 1998;Wolpert, 2002). Indeed, new ventures have uniqueadvantage in benefiting from a broad external inno-vation search. Because new ventures do not havespecialized structures and routines that may pre-vent them from thinking about new uses of existingresources (Katila and Shane, 2005: 816), they areable to integrate and recombine various knowledgeelements in external knowledge space to createtheir new products.

Ties with service intermediaries can also reducecosts associated with locating external sources ofinformation, knowledge, and specialized expertisethat are critical for product innovation (McEvilyand Zaheer, 1999). Given their extensive net-works, ties with service intermediaries constituteinformation channels that reduce the amount oftime and investment required for new ventures togather external information and knowledge (Shaneand Cable, 2002). New ventures with close tieswith service intermediaries may source what theyneed more rapidly, and thus enhance their capac-ity, speed, and flexibility in product innovation.Reduced search cost can further allow the ven-tures to dedicate more resources to innovation andquicken the process. As Bahrami and Evans (1995:68) observed, ‘a sophisticated service infrastruc-ture allows start up firms to focus on their cho-sen steeple of expertise, rather than dissipate theirenergies across a broad range of peripheral or sup-porting activities.’

This study examines four specific types of ser-vice intermediaries whose significance in tech-nology entrepreneurship has been highlighted inthe prior literature and that are also relevant toChina—the empirical context of this study. Theyare: accounting and financial service firms, lawfirms, talent search firms, and technology servicefirms.2 For example, Atwell (2000) argued that in

2 Although other organizations such as trade associations andgovernment agencies may also provide services to new ven-tures, their functions involve other components (e.g., lobbyingand regulation) and thus are outside the scope of this study. Fur-ther, although venture capitalists are a critical type of service



Silicon Valley, accountants not only play their tra-ditional roles as auditors and tax advisors, but alsoserve as consultants in structuring deals for firms.Bahrami and Evans (1995) noted that account-ing firms have specialized high-technology prac-tices and services designed for start-up and high-technology companies. Likewise, financial servicefirms may not only offer new ventures financialservices but may also help the ventures get toknow other firms that control financial resources,which can, in turn, increase product innovationspeed (Schoonhoven et al., 1990). Suchman (2000)documented that in Silicon Valley, law firms offerboth general business advice and legal guidanceand also help the ventures structure deals. Wolpert(2002: 82) stated that law firms often learn aboutbest practices, ideas for new inventions, and newways of doing business from competing and non-competing companies.

Talent search firms are used extensively by start-up and maturing firms to augment managementteams and recruit new talents (Bahrami and Evans,1995). They can facilitate the migration of peo-ple across firms, helping build social networks thattranscend firms. This enables new ventures to gainaccess to the knowledge generated by other firmsand become acquainted with other firms’ strate-gic moves (Boeker, 1997). Moreover, technologyservice firms provide services in technology com-mercialization and brokering. Through technologyservice firms, new ventures can commercializetheir innovation and obtain funding to supporttheir future innovation. Technology service firmsalso supply new ventures with previously unknowninformation regarding other firms’ technologiesand innovation, which has the potential to helpthe ventures in the development of new products(Bessant and Rush, 1995; Hargadon and Sutton,1997). Further, technology service firms help newventures quickly locate external sources of com-plementary technologies and specialized expertisethat are critical to implement their product innova-tion (Howells, 2006).

While these service intermediaries provide dif-ferent services, they share a common feature intheir specialized network role. In this study, we arenot arguing that specific services offered by these

intermediaries in Silicon Valley and its like (Saxenian, 1990;Suchman, 2000), the venture capital industry is at the emergingstage in China; thus, venture capitalist firms are not included inthis study.

intermediaries will be directly used in new ven-tures’ product innovation.3 Rather, we argue thatthe specialized network role of service intermedi-aries makes it possible for new ventures with closeties to them to broaden their external search scopeand reduce their search cost. Commonly, all tieswith different types of service intermediaries bringin directory information, networking opportuni-ties, and intermediary introduction. Distinctively,they render different information, knowledge, andadvice in respective professional fields, which inturn help the ventures better establish and developa more aligned interdependence between innova-tion and other supporting activities. Although weare not aware of any studies that specifically exam-ine service intermediaries in the Chinese context,we expect that the effect of ties with service inter-mediaries will be strong there. Because Chineseventures face a higher level of resource constraintscompared with their Western counterparts (Li andAtuahene-Gima, 2001), they have a stronger needto search in external knowledge space. Consid-ering that strategic factor markets have not yetbeen well developed in China, ties with serviceintermediaries can play an important role in newventures’ external search. Furthermore, Chineseculture strongly favors ties or guanxi. Ties withexternal actors have been an important strategy forChinese firms to overcome their resource obsta-cles (Li and Zhang, 2007; Luo, 2003; Peng andLuo, 2000). Particularly, the four types of serviceintermediaries examined in this study are critical toentrepreneurial activities and innovation in Chinaas they, respectively, promote or curtail the crucialconstraints faced by new ventures (i.e., financial,human capital, legal, and technological services).Therefore, we propose that:

Hypothesis 1: Ties with service intermediaries—(a) technology service firms, (b) accounting andfinancial service firms, (c) law firms, and (d)

3 Specific services provided by service intermediaries probablymay not be directly used in new ventures’ product innovationsince these services and product innovation fall in differentdomains. However, because service intermediaries are channelsof access to resources and information of various kinds, we arenot excluding the possibility that ties with service intermediariesmay help access complementary assets that do contribute directlyto product innovation.



talent search firms—will be positively related tonew ventures’ product innovation.4

The contingent value of ties with serviceintermediaries

The logic underlying Hypothesis 1 is that ties withservice intermediaries contribute to new ventures’product innovation by broadening their externalsearch scope and reducing their search cost. Thislogic suggests that the value of ties with serviceintermediaries in product innovation will dependon the extent to which new ventures rely uponsearch in the networks within the cluster for theirproduct innovation. The reason of this is that whennew ventures rely upon search in the networksto locate information, knowledge, and specializedexpertise for their product innovation, their tieswith service intermediaries matter more to theirproduct innovation by affecting the scope of theirexternal search and their search cost. In contrast,when new ventures have alternative sources ofinnovation search and rely less upon the networksin the cluster for innovation search, the value ofties with service intermediaries in product innova-tion will become limited. Based upon this argu-ment, we examine the moderating effects of threecontextual variables: perceived industry growth,new ventures’ ties with foreign firms, and newventures’ export.

The moderating role of perceived industry growth

Perceived industry growth refers to the level ofmanagers’ perceived growth of their principalindustry within the last three years (Li andAtuahene-Gima, 2002). Managers’ perceivedindustry growth affects the level of managerialdiscretion—or latitude of action—that managersperceive to have in their strategic decision making(Hambrick and Finkelstein, 1987). High industrygrowth is associated with more market opportuni-ties, greater competitive variation, and expandedoptions for firms (Datta, Guthrie, and Wright,2005; Hambrick and Finkelstein, 1987). Thus,managers who perceive high industry growth tendto believe that they have greater opportunities fordecision making freedom. Also, industry growth

4 In empirical models, we will examine the combined measureof ties with multiple service intermediaries as well as the tieswith each of these service intermediaries.

affects the level of resource munificence in theenvironment. Resources are more available inhigh growth industries (McDougall et al., 1994;Romanelli, 1989) and therefore managers who per-ceive high industry growth are less concerned withresource constraints.

We argue that the positive relationships betweenties with service intermediaries and new ventureproduct innovation will become weaker when per-ceived industry growth is high rather than low.First, as noted above, when perceived industrygrowth is high, new venture managers tend tobelieve that there exists a wide range of oppor-tunities and options that their ventures can pursuefor product innovation, thus, reducing their moti-vation to search what other firms are doing andhow they are doing it, and discouraging them tofollow other firms’ actions in their own ventures’product innovation. Also, when perceived industrygrowth is high, it appears that the industry residesin a very early phase of an industry life cycle.In this situation, managers tend to observe a highlevel of causal ambiguity about innovation (Reedand DeFillippi, 1990) regarding which directionand option of innovation has a better chance tosucceed, which can further discourage new ven-tures to follow and imitate other firms’ innovationactivities. Thus, when perceived industry growthis high, new ventures tend to pursue their ownway of product innovation, and external search inthe networks in the cluster in which they residebecomes less important in directing their productinnovation.

Second, when perceived industry growth is high,new venture managers are less concerned withresource constraints (McDougall et al., 1994;Romanelli, 1989). Managers tend to feel that exter-nal resource holders are willing to support inno-vation activities and more money will chase newventures with innovation opportunities. In this sit-uation, new ventures become less concerned aboutthe cost of external innovation search. For thesereasons, when perceived industry growth is high,the value of ties with service intermediaries interms of both broadening new ventures’ exter-nal search scope and reducing their search costdecreases. Accordingly, the positive relationshipsbetween ties with service intermediaries and newventure product innovation will become weaker.

One may argue that ties with service inter-mediaries could become more important to newventures’ product innovation in a faster growing



industry so as to better leverage such opportuni-ties via support from service intermediaries. Aswe argued earlier, ties with service intermediariescontribute to product innovation not because theirspecific services can be directly used for innova-tion, but because their specialized network roleenables new ventures to broaden external searchscope and reduce search cost. When perceivedindustry growth is high, although the services pro-vided by these intermediaries are still important,the information regarding other firms’ innovationactivities channeled by ties with these intermedi-aries becomes less important to the focal venture’sproduct innovation. Consistent with our argument,Li and Atuahene-Gima (2002) found that per-ceived industry growth is negatively related to Chi-nese ventures’ external collaboration. Therefore,we propose that:

Hypothesis 2: The positive relationships betweenties with service intermediaries—(a) technologyservice firms, (b) accounting and financial ser-vice firms, (c) law firms, and (d) talent searchfirms—and new ventures’ product innovationwill be weaker as perceived industry growth ishigher.

The moderating role of ties with foreign firms

Given the significant gaps between emerging econ-omy firms and foreign entrants in technologyand marketing capabilities, emerging economyfirms often engage in learning through partnershipswith foreign entrants (Hitt et al., 2000). Indeed,ties with foreign firms are considered to be oneof the most prominent ties in this context (Liand Atuahene-Gima, 2002). In particular, priorresearch on new ventures in emerging economieshas highlighted the importance of ties with foreignfirms through the adoption of agency businessesfor foreign firms, defined as a downstream alliance‘by which a firm markets and distributes a foreignfirm’s products and services’ in their home coun-try (Li and Atuahene-Gima 2002: 470). Bruton andRubanik (1997) noted that it is common for Rus-sian ventures to import and wholesale the productsof foreign firms. Lu (2000) also observed that Chi-nese ventures tend to build this type of downstreamalliance with foreign partners for growth. Li andAtuahene-Gima (2002) examined the antecedentsand consequences of this type of alliance and foundthat it is positively related to new venture financial

performance but negatively related to new ventureinnovation in China.

We argue that the positive relationships betweenties with service intermediaries and product inno-vation will become weaker if new ventures adoptagency businesses for foreign firms, because newventures often act as a domestic agent by focus-ing on distributing the foreign firms’ products andservices in the domestic market (Li and Atuahene-Gima, 2002). Most innovation work has been doneby the foreign partners, and the role of new ven-tures mainly focuses on distribution and sales.Thus, search of other firms’ innovation activitiesin the cluster in which they reside is not importantto them. Therefore, the value of their ties with ser-vice intermediaries in product innovation becomesless important to these ventures.

Moreover, even if new ventures that adoptagency business for foreign firms conduct prod-uct innovation, their ties with the foreign partnersrepresent an important source for their innovationsearch. Previous studies have shown that new ven-tures’ ties with established firms can provide criti-cal resources and capabilities to new ventures andhave a positive impact on their innovation (Baumet al., 2000; Shan et al., 1994; Stuart, 2000). Sim-ilarly, in the context of China’s emerging econ-omy, new ventures’ ties with foreign firms canserve as channels of access to resources of variouskinds. Being an agent for foreign firms to distributeand market their products in China enables Chi-nese ventures to learn from their foreign partnersand gain access to the foreign partners’ technol-ogy, management, and marketing skills that couldpotentially be used for the ventures’ own productinnovation (Li and Atuahene-Gima, 2002). Fur-ther, the adoption of agency business for foreignfirms in China usually is a profitable business (Liand Atuahene-Gima, 2002), which can help theventures accumulate financial resources to supporttheir product innovation. Since ties with foreignfirms through adopting agency business representan important alternative source for new ventures’external innovation search, they can decrease thevalue of ties with service intermediaries in prod-uct innovation. Based on the above discussion, wepropose that:

Hypothesis 3: The positive relationships betweenties with service intermediaries—(a) technologyservice firms, (b) accounting and financial ser-vice firms, (c) law firms, and (d) talent search



firms—and new ventures’ product innovationwill be weaker when the ventures adopt agencybusiness for foreign firms than when they do not.

The moderating role of new venture export

Whether or not a new venture exports reflects itsmarket focus. If a new venture does not export, itfocuses on developing new products to penetratedomestic markets. If it does export, its focus shifts,at least partially, to international markets. We arguethat the positive relationships between ties withservice intermediaries and new venture productinnovation will become weaker when new venturesexport than when they do not. When new venturesdo not export, they emphasize reaping benefitsfrom pent up indigenous demands in the domes-tic markets (Luo and Park, 2001). In this situation,what products domestic customers want and whatproducts other firms offer for domestic markets arecritical in deciding new ventures’ product inno-vation direction and speed. These ventures tendto closely interact with and follow domestic cus-tomers and domestic competitors for innovationsearch. As a result, ties with service intermediariesplay an important role in their external innovationsearch and are critical for these ventures to under-stand domestic market demands and competition.

In contrast, if new ventures export part or all oftheir products to international markets, they haveboth need and opportunity to search in interna-tional markets for their product innovation. Theyneed to understand what customers in internationalmarkets want and what products international com-petitors offer. Their participation in internationalmarkets also provides an opportunity for them toobserve international competitors’ product offer-ings and imitate them in their own product innova-tion—or, in other words, they can capture knowl-edge spillover in international markets. Moreover,since domestic markets of an emerging economylike China typically lag behind international mar-kets in terms of technology advancement, designsophistication, and quality standards, new ven-tures can easily modify their products that aredeveloped for international markets for domesticmarkets (however, it would be very difficult forthem to modify their products that are developedfor domestic markets for international markets).Therefore, new ventures’ participation in interna-tional markets through export provides them an

important alternative source for their external inno-vation search, which may decrease the value of tieswith service intermediaries in their product inno-vation. Therefore, we propose:

Hypothesis 4: The positive relationships betweenties with service intermediaries—(a) technologyservice firms, (b) accounting and financial ser-vice firms, (c) law firms, and (d) talent searchfirms—and new ventures’ product innovationwill be weaker when the ventures export thanwhen they do not.

METHODOLOGY

Sample and data collection

In 2001, we collected our data in the Pearl RiverDelta Torch High-Technology Geographical Clus-ter in Guangdong Province, one of the largest tech-nology clusters in China. We randomly selected500 new ventures (eight years old or younger) asthe sample frame from a list of technology firmscompiled by the Department of Science and Tech-nology in Guangdong Province. All of the venturesare manufacturing firms, and they met the threecriteria used to define a technology firm in China(Li and Atuahene-Gima, 2001): (1) its foundingteam must be composed of engineers or scien-tists, (2) thirty percent or more of its employeesmust be technical employees, and (3) it must spendthree percent or more of total sales on research anddevelopment (R&D).

The questionnaire was originally prepared inEnglish and then translated into Chinese by twoscholars competent in both languages and withsubstantial research experience in the subject areain China. To avoid cultural bias and ensure valid-ity, the Chinese version was back-translated intoEnglish, and we paid special attention to detect-ing any misunderstandings that might arise due totranslation. The questionnaire was further pretestedwith eight founding members of six new ventures.Items that were identified as being problematicwere revised or eliminated.

We administered the questionnaire on-site. Atrained research assistant scheduled appointments,presented the respondents with the questionnaires,answered general questions, and collected the com-pleted questionnaires. A major concern in surveyresearch is that a single respondent may answer



all of the questions in a consistent manner that canresult in common method bias (Podsakoff et al.,2003). To reduce the potential for this bias, wedesigned the questionnaire to include two sepa-rate parts. Part I contained questions regardingenvironments, the adoption of agency business forforeign firms, export, and product innovation. PartII contained questions regarding ties with serviceintermediaries and organizational demography. Foreach venture, we invited two top managers toparticipate, with one completing Part I and theother completing Part II, on a random basis. Sucha design can also require the respondents to fillout a shorter version of the questionnaire (i.e.,only one part of the questionnaire), thus reducingtheir response burden and improving their responseaccuracy.

We obtained data from 202 firms, with aneffective participation rate of 40.4 percent. Of theresponding ventures, 44 percent were in the elec-tronic information industry, 24 percent in the newenergy and new material industries, 21 percent inthe pharmaceutical and bioengineering industries,and 11 percent in the integrated optical-mechanicaland electric products industries. To assess the non-response bias, we compared the responding ven-tures with the nonresponding ventures and foundno significant differences in venture size and age.

The profiles of the two respondents from eachventure are listed in the Appendix. These respon-dents had positions of managing director/chiefexecutive officer (CEO), business manager, orproduct manager in the ventures. The mean of theirinvolvement in strategic decision making was 4on a five-point scale (1 = very uninvolved and5 = very highly involved). Average working expe-rience in their current industries was 6.2 years.These data suggest that the respondents were expe-rienced and knowledgeable about the issues understudy, which increased our confidence in the qual-ity of the data. The two respondents from eachventure did not differ significantly in terms oftheir positions in the firm, average working expe-rience in the current industry, or average participa-tion level in strategic decision making. However,they differed in terms of age and education. Inorder to check the consistency between the judg-ments of the two respondents in each venture, weexamined the correlations between their responses(Nunnally and Bernstein, 1994) with regard to(1) whether the firm has adopted agency businessfor foreign firms, and (2) R&D spending on new

products (these two questions were answered byboth respondents). The correlations of these twoquestions between the two respondents are 0.83(p < 0.001) and 0.71 (p < 0.001), respectively.The high levels of interrater reliability suggest thatresponses from different individuals in the sameventure are consistent, providing validation of oursubjective measures.

Measures

All of the perceptual measures were rated basedon a five-point Likert scale (1 = not at all and5 = very much). For multi-item constructs, weaveraged the items to create the scores for theconstructs.

Dependent variable

We used five items to measure product innovation.In order to provide an anchor for the respondentsto answer the questions, we asked them to ratethe extent to which their ventures were success-ful relative to their major competitors in terms of(1) frequently introducing new products, (2) beingfirst in new product introductions in the mar-ket, (3) quickly launching new products into themarket, (4) developing new products with supe-rior quality, and (5) using new products to pene-trate markets. This scale highlights the two keycomponents used for assessing the performanceof new products: quality and speed (Brown andEisenhardt 1995), which have also been empha-sized in Lu’s (2000) study of Chinese technologyfirms. We collected data on the percentage of totalsales attributed to new products over the past threeyears, and its correlation with product innovationwas 0.26 (p < 0.001), which provided evidence tosupport the external validity of this construct.

Predictors

To measure ties with service intermediaries, weasked the respondents to indicate the extent towhich their ventures had close relationships with(1) technology service firms (i.e., technology com-mercialization and brokering), (2) accounting andfinancial service firms, (3) law firms, and (4) talentsearch firms. McEvily and Zaheer (1999) useda similar approach to measure firms’ ties with



regional institutions in a cluster. We created a com-posite measure of ties with multiple service inter-mediaries by averaging these four items. We alsoused ties with each of the four types of serviceintermediaries as a predictor in empirical models.This subjective measure captures the extent of newventures’ ties with service intermediaries. Our useof these general questions rather than the name-generator approach to measuring ties is consistentwith previous studies on ties in the Chinese con-text (e.g., Luo, 2003; Peng and Luo, 2000; Xin andPearce, 1996). As Peng and Luo noted, the name-generator approach is not effective in the Chinesecontext because ‘ties were regarded as a personaland business secret, and some respondents werereluctant to disclose such contacts’ (Peng and Luo,2000: 491).

Perceived industry growth was measured bythree items adapted from McDougall et al. (1994).The respondents were asked to indicate the extentto which they agreed with the following state-ments regarding their principal industry in the lastthree years: (1) There has been high growth indemand in this industry; (2) This industry offeredmany attractive opportunities for future growth;and (3) Growth opportunities in this industry havebeen abundant. It would be better if we coulduse archival data of industry growth to validateour subjective ratings. However, the industry dataavailable in China is at such a gross level that onewould lose any significant meanings since most ofthe sampled firms would fall into very few cat-egories. Furthermore, the high-technology indus-tries we studied are emerging in China and mostof the ventures in these industries are relativelysmall; thus, it is not surprising that even venturesin the same industry category may focus on dif-ferent niches of the industry. Therefore, industry-level archival data become less relevant in ourresearch setting. Indeed, our use of perceptual mea-sures is consistent with a strong research tradi-tion that suggests that environmental conditionsaffect firm decisions only when they are perceivedor interpreted by decision makers (e.g., Li andAtuahene-Gima, 2002; Luo and Park, 2001; Weick,1969). As Luo and Park argued, ‘Firms interpretthe same environmental attributes differently andthus respond with different strategies. Therefore,the objective reality of the physical environmentmay be less important in determining organiza-tional actions’ (Luo and Park, 2001: 153).

To measure the adoption of agency business forforeign firms, we asked the respondents to indicatewhether his or her firm has been involved in anybusiness as an agent to market and distribute a for-eign firm’s products and services in China (Li andAtuahene-Gima, 2002). The variable was dummycoded (1 = yes, 0 = no). To measure export, weasked the respondents to indicate whether his orher firm has exported products out of China. Thevariable was also dummy coded (1 = yes, 0 = no).

Controls

Following Li and Atuahene-Gima (2002), we con-trolled for the following variables: venture age(i.e., the number of years since founding), found-ing team size (number of founding team membersin the venture), venture size (the natural log ofthe number of full-time employees in 2000), for-eign invested venture (1 = yes, 0 = no), andventure origin (independent venture = 0 versuscorporate venture = 1). We also controlled forenvironmental uncertainty, which was measuredby three items adapted from Miller (1987). Theseitems reflect the speed of change and predictabilityabout technology and product market conditions inthe ventures’ principal industries. The respondentswere asked to rate the degree to which they agreedwith the following statements regarding their prin-cipal industry over the past three years: (1) It hasbeen difficult to forecast how technologies willchange in this industry, (2) Competitors’ actionshave been highly unpredictable, and (3) Productmarket conditions have been changing very fast.Finally, because our sample included ventures infour industries, we created three industry dummyvariables with the integrated optical industry as thebase group.

Adequacy of the measures: reliability, validity,and common method variance

We took several steps to ensure data validity andreliability. As noted earlier, we pretested the sur-vey with eight founding members of six new ven-tures. In the questionnaire itself, we used previ-ously validated measurement items whenever pos-sible to help ensure the validity of our measures.We assessed the reliability of the multi-item con-structs with Cronbach’s alpha, and all scales exceptenvironmental uncertainty had reliabilities greaterthan the recommended 0.70 (see Table 1). The



Table 1. Construct measurement and confirmatory factor analysis results

Item description summary Standardizedloading

t-value

Product innovation (Brown and Eisenhardt, 1995; and Lu, 2000) (a = 0.90)Rate the extent to which your venture was successful relative to your major competitors in terms of the following:

1. Frequently introducing new products 0.64 8.932. Being first in new product introductions in the market 0.61 8.063. Quickly launching new products into the market 0.75 11.084. Developing new products with superior quality 0.69 10.025. Using new products to penetrate markets 0.70 9.95

Ties with service intermediaries (newly developed items) (a = 0.80)Indicate the extent to which your venture had close relationships with the following:

1. Technology service firms 0.74 11.912. Accounting and financial service firms 0.90 15.593. Law firms 0.89 15.124. Talent search firms 0.63 9.23

Perceived industry growth (Li and Atuahene-Gima, 2001) (a = 0.78)Rate the degree to which each of these statements describes your principal industry over the last three years:

1. There has been high growth in demand in this industry. 0.70 9.882. This industry offered many attractive opportunities for future growth. 0.70 9.893. Growth opportunities in this industry have been abundant. 0.82 11.75

Environmental uncertainty (Miller, 1987) (a = 0.63)Rate the degree to which each of these statements describes your principal industry over the last three years:1. It has been difficult to forecast how technologies will change in this industry. 0.67 7.082. Competitors’ actions have been highly unpredictable. 0.55 6.243. Product market conditions have been changing very fast. 0.60 6.64

Model fit indexχ 2 = 130.28, p =.00; χ 2/df = 1.61; GFI = 0.92; CFI = 0.95; IFI = 0.95; RMSEA = 0.05

alpha for environmental uncertainty is 0.63, whichis generally acceptable for questionnaire scales.Considering the face validity of this factor andthe strong factor loadings, we believed that it wasreasonable to use this factor as a control in thesubsequent analysis (c.f. Li and Atuahene-Gima,2002).

We conducted a confirmatory factor analysis toassess the convergent and discriminant validity ofthe multi-item constructs. As presented in Table 1,results of the confirmatory factor analysis indi-cated that the measurement model fits the data well(χ 2 = 130.28, p =0.00; χ 2/df = 1.61, goodness-of-fit index (GFI) = 0.92, comparative fit index(CFI) = 0.95, incremental fit index (IFI) = 0.95,root mean square error of approximation (RMSEA)= 0.05), all of which confirmed the unidimension-ality of each construct in the model. Convergentvalidity is observed when the path coefficientsfrom the latent constructs to their correspond-ing manifest indicators are statistically significant(i.e., t > 2.0) (Anderson and Gerbing, 1988). All

items loaded significantly on their correspondinglatent construct, with the lowest t-value being 6.24,thereby providing evidence of convergent validity.

We used two methods to assess discriminantvalidity. First, because 95 percent confidence bandsaround the φs did not contain a value of one,we concluded that the constructs possessed dis-criminant validity (Anderson and Gerbing, 1988).Second, we conducted a chi-square difference testfor all of the multi-item constructs in pairs to seeif they were distinct from one another. The pro-cess involved collapsing each pair of constructsinto a single model and comparing its fit with thatof a two-construct model (Anderson and Gerbing,1988). In every case, a two-factor model had a bet-ter fit than a single-factor model, thus supportingthe discriminant validity of the constructs.

As in most survey research, the use of per-ceptual data for both independent and dependentvariables in this study may raise the concernsof common method variance. Although a numberof studies have suggested that common method



variance may not be as much of an artifact asis commonly assumed (e.g., Avolio, Yammarino,and Bass, 1991; Spector, 1987), Podsakoff et al.,(2003). Based on a careful examination of the lit-erature, Podsakoff et al. (2003: 900) argued thatcommon method variance is often a problem inbehavioral research and researchers should imple-ment measures to control for it. Following Pod-sakoff et al.’s (2003) recommendation, we haveintegrated both procedural methods and statisti-cal techniques to reduce the potential of commonmethod variance. Regarding procedural methods,as noted earlier, we used two respondents ratherthan a single respondent in each venture for datacollection. As Podsakoff et al. (2003: 887) noted,collecting the measures of variables from differentsources ‘makes it impossible for the mind set ofthe source or rater to bias the observed relationshipbetween the predictor and criterion variable.’ Also,we assured the respondents that their answers wereconfidential and that there were no right or wronganswers to the questions in the survey. Theseprocedures likely reduce the respondents’ evalu-ation apprehension and ‘make them less likely toedit their responses to be more socially desirable,lenient, acquiescent, and consistent with how theythink the researcher wants them to respond’ (Pod-sakoff et al. 2003: 888).

Regarding statistical techniques, we used Har-man’s one-factor test to check for the presence ofcommon method variance (Podsakoff and Organ,1986). Significant common method variance wouldresult in one general factor accounting for themajority of covariance in the variables. We sub-jected all of the key variables to a factor analysis(the four multi-item constructs in Table 1). Theanalysis resulted in four factors with eigenvaluesgreater than one, with the first factor accountingfor only 27 percent of total variance. This resultsuggests that common method variance is unlikelyto have caused any significant relationships amongvariables in our study. Further, we followed thelatent variable approach suggested by Podsakoffet al. (2003: 894) to control for the effects of anunmeasured latent methods factor. We allowed theitems to load on their theoretical constructs, as wellas on a latent common methods variance factor.Then we examined the significance of the struc-tural parameters both with and without the latentcommon methods variance factor in the model.We found that all significant relationships heldafter controlling for the latent common methods

variance factor, providing evidence that commonmethod variance is not an issue in this study (Li,Bingham, and Umphress, 2007).

Furthermore, our interaction hypotheses, if sup-ported, can provide additional evidence that com-mon method variance is not an issue in this study.As Aiken and West (1991) pointed out, supportedinteraction hypotheses are less subjective to thecommon method variance because it is unlikelythat respondents would have an ‘interaction-basedtheory’ in their minds that could systematicallybias their responses to produce these results. Inthis study, since we used two respondents ineach venture, we do not believe that the twodifferent respondents would collectively have an‘interaction-based theory’ in their minds to pro-duce the significant interaction results.

Finally, one potential question of the surveymethodology is how to compare responses acrossrespondents. We have addressed this issue to theextent possible in this paper. First, as discussedearlier, we have taken several procedural methodsto ensure that the respondents are knowledgeableand experienced, thus capable of providing validand accurate assessment. Second, we used two dif-ferent respondents in each venture. The high levelsof interrater reliability suggest that the respon-dents have used their venture’s real situations asthe anchor to answer survey questions. Third, itis possible that managers’ perceptions and actionsmay vary systematically across firms with differ-ent attributes. To take into account this possibil-ity, we have controlled for several important firmattributes as discussed above.

Data analysis

Adoption of agency business for foreign firms wasused as a moderating variable in this study. How-ever, not all the ventures have the chance of adopt-ing agency businesses for foreign firms. Hence,new ventures with agency business for foreignfirms may differ from those without this business.In this study, we used Heckman’s selection modelto account for this self-selection effect (Heckman,1979; Shaver, 1998). In essence, Heckman’s selec-tion model is a two-stage procedure (Heckman,1979). In the first stage, probit regression was usedto estimate the probability that a venture adoptsagency business for foreign firms as a function ofventure age, founding team size, venture size, for-eign owned venture, venture origin, environmental



uncertainty, perceived industry growth, export, andthree industry dummies.

Based upon the results of the first-stage model,we predicted and saved the value for the inverseMill’s ratio (λi). The inverse of Mill’s ratio isthe monotone decreasing function of the probabil-ity that a venture is selected into agency businessfor foreign firms. The inverse of Mill’s ratio wasthen included as a regressor in the second-stagemodels to estimate a venture’s product innovation(Heckman, 1979; Shaver, 1998). This two-stageprocedure can generate consistent and asymptoti-cally efficient estimates (Heckman, 1979).

RESULTS

Table 2 presents the means, standard deviations,and correlations among the variables examinedin the study. Table 3 presents the results of theregression models. Model 1 only included con-trols.5 Model 2 added the main effects of ties withmultiple service intermediaries and the moderatingvariables: perceived industry growth, the adoptionof agency business for foreign firms, and export.Model 3 added the interaction terms. To reducethe potential problem of multicolinearity, predic-tor and moderator variables (except those that aredummy variables) were mean-centered prior to thecreation of interaction terms (Aiken and West,1991). To examine ties with each type of serviceintermediaries, we repeated Models 2 and 3 for tieswith technology service firms in Models 2a and 3a,for ties with accounting and financial service firmsin Models 2b and 3b, for ties with law firms inModels 2c and 3c, and for ties with talent searchfirms in Models 2d and 3d, respectively.

Hypothesis 1 proposes that ties with serviceintermediaries are positively related to productinnovation. The results in Model 2 suggest that tieswith multiple service intermediaries are positiveand significant (b = 0.15, p < 0.05). This predictoraccounts for two percent of variance in productinnovation (p < 0.05), and its one unit increasewould increase product innovation by 0.15 units.

5 Heckman’s two-stage models do a particularly good job asestimation when there is at least one variable that may beconsidered an ‘instrument’ that is a good predictor in the firststage but not in the second stage of the model. In this study,the industry dummy variables were included as the ‘instrument’variables in the first stage but not in the second stage of themodel. Ta

ble

2.D

escr

iptiv

est

atis

tics

and

Pear

son

corr

elat

ion

mat

rixa

(N=

202)

Var

iabl

esM

ean

S.D

.1

23

45

67

89

1011

1213

14

1.Pr

oduc

tin

nova

tion

2.86

1.03

—2.

Tie

sw

ithm

ultip

lese

rvic

ein

term

edia

ries

2.48

1.03

0.26

—3.

Tie

sw

ithte

chno

logy

serv

ice

firm

s2.

471.

240.

270.

90—

4.T

ies

with

acco

untin

g/fin

anci

alse

rvic

efir

ms

2.22

1.11

0.20

0.80

0.68

—5.

Tie

sw

ithla

wfir

ms

2.41

1.23

0.25

0.92

0.80

0.66

—6.

Tie

sw

ithta

lent

sear

chfir

ms

2.83

1.24

0.15

0.79

0.59

0.42

0.66

—7.

Perc

eive

din

dust

rygr

owth

3.57

0.80

0.05

0.23

0.22

0.22

0.22

0.13

—8.

Age

ncy

busi

ness

for

fore

ign

firm

s0.

470.

50−0

.07

0.04

0.10

−0.1

00.

070.

060.

15—

9.E

xpor

t0.

430.

500.

290.

340.

280.

320.

310.

240.

000.

05—

10.

Ven

ture

age

5.25

2.88

−0.0

1−0

.04

−0.0

4−0

.03

0.00

−0.0

60.

010.

070.

04—

11.

Foun

ding

team

size

4.95

3.92

0.08

−0.0

7−0

.09

−0.0

8−0

.09

0.02

0.00

−0.1

3−0

.09

0.02

—12

.V

entu

resi

ze3.

501.

650.

340.

060.

030.

020.

020.

14−0

.01

0.07

0.20

0.11

0.31

—13

.Fo

reig

nin

vest

edve

ntur

e0.

110.

310.

16−0

.08

−0.0

8−0

.01

0.05

−0.1

0−0

.06

−0.0

80.

21−0

.03

0.13

0.19

—14

.V

entu

reor

igin

0.30

0.45

0.09

0.06

0.06

−0.0

50.

080.

11−0

.12

−0.0

90.

05−0

.01

0.04

0.20

0.05

—15

.E

nvir

onm

enta

lun

cert

aint

y3.

330.

770.

070.

070.

060.

070.

100.

00−0

.02

0.03

0.07

−0.1

3−0

.05

−0.1

2−0

.03

0.05

aC

orre

latio

nsth

atar

e0.

14or

larg

erar

esi

gnifi

cant

atth

ele

vel

ofp

<0.

05.



Tabl

e3.

Res

ults

ofO

LS

regr

essi

ona,

b(N

=20

2)

Var

iabl

esM

odel

1M

odel

2M

odel

3M

odel

2aM

odel

3aM

odel

2bM

odel

3bM

odel

2cM

odel

3cM

odel

2dM

odel

3d

Con

trol

sV

entu

reag

e−0

.04

−0.0

2−0

.03

−0.0

2−0

.02

−0.0

2−0

.05

−0.0

2−0

.04

−0.0

2−0

.02

Foun

ding

team

size

0.00

−0.0

3−0

.03

−0.0

2−0

.02

−0.0

4−0

.05

−0.0

3−0

.02

−0.0

4−0

.03

Ven

ture

size

0.33

∗∗∗

0.34

∗∗∗

0.35

∗∗∗

0.33

∗∗∗

0.36

∗∗∗

0.34

∗∗∗

0.34

∗∗∗

0.34

∗∗∗

0.35

∗∗∗

0.33

∗∗∗

0.34

∗∗∗

Fore

ign

inve

sted

vent

ure

0.10

0.06

0.06

0.07

0.07

0.05

0.04

0.06

0.05

0.05

0.05

Ven

ture

orig

in0.

02−0

.03

−0.0

8−0

.03

−0.0

6−0

.02

−0.0

7−0

.04

−0.0

8−0

.03

−0.0

4E

nvir

onm

enta

lun

cert

aint

y0.

12†

0.11

†0.

16∗

0.10

0.16

∗0.

12†

0.14

∗0.

11†

0.14

∗0.

12†

0.14

∗

Inve

rse

ofM

ill’s

ratio

(λ)

−0.0

50.

060.

090.

030.

090.

060.

110.

050.

080.

050.

05

Pre

dict

ors

Mul

tipl

ese

rvic

eA

.Tec

hnol

ogy

B.A

ccou

ntin

gan

dC

.Law

firm

sD

.Tal

ent

sear

ch.

inte

rmed

iari

esse

rvic

efir

ms

finan

cial

serv

ice

firm

sC

.Law

firm

sfir

ms

Tie

sw

ithse

rvic

ein

term

edia

ries

(TSI

)0.

15∗

0.17

∗∗0.

18∗∗

0.20

∗∗0.

10†

0.14

∗0.

16∗∗

0.19

∗∗0.

050.

07Pe

rcei

ved

indu

stry

grow

th0.

060.

070.

050.

060.

070.

090.

060.

050.

080.

08A

genc

ybu

sine

ssfo

rfo

reig

nfir

ms

−0.1

0−0

.09

−0.1

1†−0

.10

−0.0

9−0

.09

−0.1

0−0

.11†

−0.1

0−0

.10

Exp

ort

0.17

∗0.

19∗∗

0.17

∗0.

17∗

0.19

∗∗0.

21∗∗

0.17

∗0.

18∗∗

0.21

∗∗0.

22∗∗

TSI

∗pe

rcei

ved

indu

stry

grow

th−0

.19∗∗

−0.1

9∗∗−0

.24∗∗

∗−0

.15∗∗

−0.0

9†T

SI∗

agen

cybu

sine

ssfo

rfo

reig

nfir

ms

−0.1

6∗∗−0

.17∗∗

−0.1

1∗−0

.15∗∗

−0.1

0†T

SI∗

expo

rt−0

.12∗

−0.0

9†−0

.10†

−0.1

7∗∗−0

.03

Fva

lue

4.80

∗∗∗

4.88

∗∗∗

5.46

∗∗∗

5.16

∗∗∗

5.77

∗∗∗

4.55

∗∗∗

4.98

∗∗∗

4.95

∗∗∗

5.63

∗∗∗

4.40

∗∗∗

3.79

∗∗∗

Rsq

uare

d0.

150.

220.

290.

230.

300.

210.

270.

220.

300.

200.

22A

djus

ted

Rsq

uare

d0.

120.

180.

240.

190.

250.

160.

220.

180.

240.

150.

16In

crea

sed

Rsq

uare

d—

0.07

∗∗0.

07∗∗

∗0.

08∗∗

∗0.

07∗∗

∗0.

06∗∗

0.06

∗∗∗

0.07

∗∗∗

0.08

∗∗∗

0.05

∗0.

01

Sign

ifica

nce

leve

l:†

p<

0.10

;∗ p

<0.

05;

∗∗p

<0.

01;

∗∗∗

p<

0.00

1(o

ne-t

aile

dte

sts

for

hypo

thes

ized

effe

cts

and

two-

taile

dte

sts

for

cont

rols

).a

Stan

dard

ized

coef

ficie

nts

repo

rted

.b

InM

odel

s2a

–3d

,w

ete

sted

ties

with

each

ofth

ese

rvic

ein

term

edia

ries

sepa

rate

ly.

Bec

ause

ofsp

ace

limit,

inth

ista

ble

ties

with

serv

ice

inte

rmed

iari

esre

ferr

edto

ties

with

mul

tiple

serv

ice

inte

rmed

iari

esin

Mod

els

2an

d3,

ties

with

tech

nolo

gyse

rvic

efir

ms

inM

odel

s2a

and

3a,

ties

with

acco

untin

gan

dfin

anci

alse

rvic

efir

ms

inM

odel

s2b

and

3b,

ties

with

law

firm

sin

Mod

els

2can

d3c

,an

dtie

sw

ithta

lent

sear

chfir

ms

inM

odel

s2d

and

3d.



Similarly, the results in Models 2a, 2b, and 2csuggest that ties with technology service firms(b = 0.18, p < 0.01), ties with accounting andfinancial service firms (b = 0.10, p < 0.10), andties with law firms (b = 0.16, p < 0.01) arepositive and significant. But ties with talent searchfirms are not significant (b = 0.05, n.s., Model2d). Thus, Hypothesis 1 is supported for ties withtechnology service firms, accounting and financialservice firms, and law firms, but not for ties withtalent search firms.

Hypothesis 2 proposes that the positive rela-tionships between ties with service intermediariesand product innovation are weaker when perceivedindustry growth is higher. The results in Model3 suggest that the interaction term between per-ceived industry growth and ties with multipleservice intermediaries is negative and significant(b = −0.19, p < 0.01). To facilitate interpreta-tion, we plotted this significant interaction effect inFigure 1. In order to create this figure, all variablesin Model 3 except ties with multiple service inter-mediaries and perceived industry growth were con-strained to means. Ties with multiple service inter-mediaries and perceived industry growth took thevalues of one standard deviation below and abovethe mean. As shown in Figure 1, ties with multipleservice intermediaries have a positive relationshipwith product innovation when perceived industrygrowth is low; but this positive relationship disap-pears when perceived industry growth is high.

Figure 1. Ties with multiple service intermediaries andproduct innovation—the moderating role of perceived

industry growth.

The size of the interaction effect is reflected inthe difference between the R2 of models with-out vs. with the interaction term (Jaccard, Tur-risi, and Wan, 1990). The models with and with-out the interaction of perceived industry growthand ties with multiple service intermediaries are0.22 and 0.26, respectively, which means that theinteraction effect accounts for four percent (p <

0.01) of variance in product innovation. Jaccardet al. (1990) suggested that a significant interac-tion effect should be further analyzed and inter-preted as a conditional effect on the main effect.We calculated the simple slopes of product inno-vation on ties with multiple service intermediariesconditional on the (high vs. low) levels of per-ceived industry growth (Aiken and West, 1991).When perceived industry growth is high, the sim-ple slope is −0.02 (n.s.), which means that changein ties with multiple service intermediaries hasno impact on product innovation. When perceivedindustry growth is low, the simple slope is 0.40(p < 0.001), which means that one unit increasein ties with multiple service intermediaries wouldincrease product innovation by 0.40 units.

The results in Models 3a, 3b, 3c, and 3d suggestthat perceived industry growth’s interactions withties with technology service firms (b = −0.19,p < 0.01), with ties with accounting and financialservice firms (b = −0.24, p < 0.001), with tieswith law firms (b = −0.15, p < 0.01), and withties with talent search firms (b = −0.09, p < 0.10)are negative and significant. The plots and slopeanalyses of these interaction effects are consistentwith those discussed above (available from theauthors upon request). Hence, these results supportHypothesis 2.

Hypothesis 3 proposes that the positive rela-tionships between ties with service intermediariesand product innovation are weaker for new ven-tures adopting agency business for foreign firmsthan for those without this business. The results inModel 3 suggest that the interaction term betweenadoption of agency business for foreign firms andties with multiple service intermediaries is negativeand significant (b = −0.16, p < 0.01). This inter-action effect accounts for two percent (p < 0.05)of variance in product innovation. This significantinteraction effect is plotted in Figure 2 (the adop-tion of agency business for foreign firms took thevalues of one and zero). As shown in Figure 2, tieswith multiple service intermediaries have a posi-tive relationship with product innovation in new



ventures that do not adopt agency business for for-eign firms. Its simple slope is 0.36 (p < 0.001),meaning that one unit increase in ties with multi-ple service intermediaries would increase productinnovation by 0.36 units. In contrast, this positiverelationship disappears in new ventures that adoptagency business for foreign firms (its simple slopeis 0.00, n.s.).

The results in Models 3a, 3b, 3c, and 3d suggestthat the interactions of adoption of agency businessfor foreign firms with ties with technology ser-vice firms (b = −0.17, p < 0.01), with ties withaccounting and finance service firms (b = −0.11,p < 0.05), with ties with law firms (b = −0.15, p< 0.01), and with ties with talent search firms (b= −0.10, p < 0.10) are negative and significant.The plots and slope analyses of these significantinteraction effects are consistent with those dis-cussed above. Hence, Hypothesis 3 is supported.

Hypothesis 4 proposes that the positive relation-ships between ties with service intermediaries andproduct innovation are weaker for new venturesthat export than for those that do not. The results inModel 3 suggest that the interaction term betweenexport and ties with multiple service intermediariesis negative and significant (b = −0.12, p < 0.05).This interaction effect accounts for two percent (p< 0.05) of variance in product innovation. Thissignificant interaction effect is plotted in Figure 3(export took the values of one and zero). As shownin Figure 3, ties with multiple service interme-diaries have a positive relationship with product

Figure 2. Ties with multiple service intermediaries andproduct innovation—the moderating role of adopting

agency business for foreign firms.

innovation in new ventures that do not export. Itssimple slope is 0.24 (p < 0.01), meaning that oneunit increase in ties with multiple service interme-diaries would increase product innovation by 0.24units. In contrast, this positive relationship disap-pears in new ventures that export (its simple slopeis −0.08, n.s.).

The results in Models 3a, 3b, and 3c suggest thatthe interactions of export with ties with technologyservice firms (b = −0.09, p < 0.10), with tieswith accounting and finance service firms (b =−0.10, p < 0.10), and with ties with law firms(b = −0.17, p < 0.01) are negatively related toproduct innovation. The plots and slope analyses ofthese significant interaction effects are consistentwith those discussed above. But the interactionof export with ties with talent search firms is notsignificant (b = −0.03, n.s.). Hence, Hypothesis 4received modest support.

DISCUSSION AND CONCLUSION

In this study, we examined the relationshipsbetween new ventures’ ties with service intermedi-aries and their product innovation. With data on asample of new ventures from a technology clusterin China, we found that ties with a variety of ser-vice intermediaries (i.e., technology service firms,accounting and financial service firms, and lawfirms) have significant positive relationships withnew ventures’ product innovation. We also foundthat these positive relationships disappear whenmanagers’ perceived industry growth is higher,

Figure 3. Ties with multiple service intermediaries andproduct innovation—the moderating role of export.



when new ventures adopt agency businesses forforeign firms, and when the new ventures export.These results suggest that the value of ties withservice intermediaries in product innovation cannotbe determined in a vacuum, but rather is contextspecific, depending upon the extent to which newventures rely upon search in the networks of thecluster in which they reside for their product inno-vation.

The insignificant result related to ties with talentsearch firms deserves further exploration because itis in contrast to Bahrami and Evans’s (1995) obser-vation that talent search firms play a significantrole in technology startups in Silicon Valley. Onepossible interpretation is that different from Sili-con Valley, the labor market in China has not beenwell developed. Professional talent search firms arestill new and they may not be able to help newventures plug into the networks in the cluster inwhich the ventures reside. We acknowledge this isspeculative. However, we believe that it is impor-tant for future research to explore how institutionalcontexts may shape the extent to which ties withservice intermediaries matter to new ventures.

Contributions

Contributions to research on innovation search.Our findings contribute to a better understandingof innovation search, particularly in the contextof new ventures. The extant literature on inno-vation has highlighted the importance of search:search of both new and old knowledge, and searchof both familiar and unfamiliar knowledge (Flem-ing, 2001; Katila, 2002; Katila and Ahuja, 2002;Rosenkopf and Nerkar, 2001). While a combina-tion of old and familiar knowledge components cancontribute to firm innovation by increasing innova-tion reliability and predictability (Fleming, 2001;Katila and Ahuja, 2002), cumulative use of com-bination can hurt firm innovation (Fleming, 2001).Thus it is important for firms to search a greatervariety of knowledge. However, it is difficult forfirms to explore new regions of the essentiallyinfinite knowledge space (Fleming, 2001). This isparticularly true for new ventures that have limitedresources and external contacts. Thus, the impor-tance and the difficulty of search of a greater vari-ety of knowledge constitute a dilemma. Then whydo some new ventures perform better than othersin external innovation search? Our results provide

an explanation to this important question and sug-gest that new ventures may solve this dilemma bybetter using their ties with service intermediaries.Sitting at the intersection of many firms, organiza-tions, and industries, service intermediaries are in aunique position to help new ventures broaden theirexternal innovation search scope without involvingtoo much search cost.

Our findings also contribute to the understand-ing of context specificity of innovation search. Asnoted by Katila (2002), how firms search cannotbe studied in isolation from where they search.Our findings suggest that ties with service inter-mediaries have no relationships with new ven-tures’ product innovation when the ventures adoptagency business for foreign firms and when theyexport. When new ventures adopt agency businessactivities for foreign firms, they either do not domuch product innovation, or if they do, they haveopportunities to search in the knowledge space oftheir foreign partners. Similarly when new ven-tures export, they may be able to search in theknowledge space of international markets. In eitherof these situations, new ventures are less likely tosearch in the networks in the cluster in which theyreside, and thus their ties with service intermedi-aries become less relevant to their product inno-vation. Therefore, our findings provide evidenceto support the argument that how firms searchdepends upon where they search.

Contributions to research on external ties of newventures. To the best of our knowledge, this studyis the first to systematically examine the role ofties with service intermediaries in new ventureinnovation. This focus is different from previousstudies that have typically examined the role ofties with prominent organizations, such as ties withestablished firms (e.g., Baum et al., 2000; Shanet al., 1994; Stuart, 2000) and with universitiesand research institutes (e.g., Baum et al., 2000),in new venture innovation. While ties with theseprominent organizations may enable new ven-tures to access their resources and capabilities, tieswith service intermediaries enable new ventures tosearch a broader range of firms, organizations, andindustries. One may argue that service intermedi-aries only provide new ventures supporting ser-vices, which are neither rare nor inimitable. How-ever, it has to be recognized that the specializednetwork role of service intermediaries can con-tribute to product innovation by facilitating new



ventures’ external innovation search. Our resultsare in line with McEvily and Zaheer’s (1999) find-ing that linkages with regional institutions (i.e.,regional industrial extension centers) in a clus-ter contribute to firms’ acquisition of competitivecapabilities. Put together, these results suggest thatfuture research needs to look beyond new ventures’ties with a few prominent organizations and to paymore attention to a broader set of external ties inwhich the ventures are embedded.

Our study also contributes to this literature byexamining how different types of external ties mayjointly influence product innovation of new ven-tures, an area largely ignored in prior research.Arora and Gambardella (1990), for example, sug-gest that pharmaceutical firms’ ties with smallbiotechnology firms and with universities representcomplementary strategies for innovation. Baumet al. (2000) noted that multiple networks withsimilar partners may yield fewer benefits than net-works with differentiated partners because same-type networks can bring in redundant resourcesand information. However, these studies did notempirically test how different types of external tiesjointly affect new venture innovation. A relatedbut different study is the work by Ahuja (2000),which examines the joint effect of direct ties andindirect ties on firm (not specifically new ven-ture) innovation. He found that the impact ofindirect ties on a firm’s innovation output willbe lower when the firm has a greater number ofdirect ties. Our findings that ties with foreign firmsmitigate the value of ties with service intermedi-aries in new venture product innovation parallelAhuja’s (2000) findings. That is, the substitutioneffect may occur not only between direct ties andindirect ties (Ahuja, 2000) but also between dif-ferent types of direct ties, or more specificallybetween ties with foreign firms and ties with ser-vice intermediaries.

Contributions to research on emerging economies.In addition, our findings advance the researchon firm strategies in emerging economies. Recentstudies have suggested that firms in emergingeconomies such as China, Eastern Europe, and theformer Soviet republics have extensive reliance onexternal ties (e.g., Luo, 2003; Xin and Pearce,1996). Several scholars have examined firms’external ties with governmental officials (Li andAtuahene-Gima, 2001; Li and Zhang, 2007; Pengand Luo, 2000; Xin and Pearce, 1996), ties with

other firms (Peng and Luo, 2000), and ties withforeign firms (Li and Atuahene-Gima, 2002). Asvaluable as that work is, we add to the litera-ture by showing that ties with service interme-diaries represent an important strategy for newventures in China’s emerging economy. Becausestrategic factor markets and institutional infras-tructures to support entrepreneurial activities andinnovation have not yet been well developed inChina, ties with service intermediates can substi-tute such institutional voids and serve as a conduitfor a wide range of information, resources, andopportunities

Managerial implications. In both academic liter-ature and popular press, new venture managershave long been advised to establish external tieswith prominent organizations such as establishedfirms, universities, and research institutes. Ourstudy highlights the important role of ties with ser-vice intermediaries in new ventures. Although ser-vice intermediaries are potentially accessible to allventures, not all ventures are connected to serviceintermediaries to the same extent or benefit fromthem to the same degree. Our results demonstratethat close ties with service intermediaries can con-tribute to new ventures’ product innovation. Forexample, through close ties with service interme-diaries, new ventures can visualize new opportuni-ties, new ideas, and best practices for doing busi-nesses synthesized from information and insightsprovided by these intermediaries. Our findings sug-gest that new venture managers need to break outof the innovation box and learn how to leveragetheir ties with service intermediaries for the pur-pose of innovation search. Our results also showthat the value of ties with service intermediaries innew venture product innovation will depend uponthe extent to which the ventures rely upon search inthe networks in the cluster in which they reside fortheir product innovation. The more new venturesrely upon search in the networks in the cluster, themore valuable their close ties with service inter-mediaries. Overall, our study provides managerialinsights about how and when new ventures canbenefit from their ties with service intermediaries.

Limitations and future research directions

Our study has limitations that should be addressedin future research. First, this study is cross-sectional in design. While we have proposed



that ties with service intermediaries are positivelyrelated to new ventures’ product innovation, it ispossible that the direction of causality may bereversed. However, for the following reasons webelieve that the reversed causality is not a signif-icant concern in this study. (1) Findings in theextant literature generally support the idea thatexternal ties enhance innovation output rather thaninnovation output attracts external ties (Shan et al.,1994: 393). (2) The significant results of our inter-action hypotheses are difficult to interpret fromthe reversed causality and thus offer stronger sup-port to our theory (Simons and Peterson, 2000:109). (3) Our field interviews suggest that in manyChinese new ventures ties with service interme-diaries were in place prior to developing newproducts because these ties are more convenient,easier, and less costly to establish. Nonetheless,a longitudinal design, cross-validation of the find-ings, and more sources of data would enable usto further assess the causality of the hypothesizedrelationships.

Second, our sample only included new venturesin a single Chinese technology cluster. Zhang, Li,and Schoonhoven (2009) have noted that in Chinathere are different types of technology clusters withdifferent growth rates. Thus, replications of ourmodel with new ventures in different clusters inChina are needed for more confident generaliza-tion. Also, as we discussed earlier, the value ofties with service intermediaries is highly contextspecific. For example, because institutional voidsmay handicap Chinese firms’ innovation duringeconomic transformation, ties with service inter-mediaries play a crucial role in this context. Webelieve that it is important for future research toexplore how institutional and cultural contexts mayshape the extent to which ties with service interme-diaries matter to new ventures and even contributeto the growth of technology clusters on a cross-country basis.

Finally, we used perceptual scales to measurenew ventures’ ties with service intermediaries.Future research may operationalize such ties withother measures such as the number, frequency,and quality of these ties and the status of theservice intermediaries. For example, ties with high-end service intermediaries may add greater valueto innovation performance than ties with low-endintermediaries. Although we did not have detailedinformation about the status of service intermedi-aries, our field interviews suggest that the Chinese

government has tried to control the quality of theservice intermediaries within the technology clus-ters by certifying the qualifications of the interme-diaries. Also, future research could examine fac-tors explaining why Chinese technology venturesestablish ties with service intermediaries in the firstplace. It has been observed that Chinese firms ingeneral are reticent to value professional servicesin a way comparable to their Western counterparts.Then, why do some Chinese technology venturesvalue service intermediaries more and have closerties with them than others? This should be inter-esting to explore in the future.

While the literature has highlighted the criti-cal role of external ties in new ventures, mostof the studies have focused on ties with promi-nent organizations. Research has devoted far lessattention to exploring the role of ties with lessprominent organizations such as ties with ser-vice intermediaries. To this end, we take an ini-tial step by examining how new ventures’ tieswith service intermediaries are linked to theirproduct innovation in the context of a technol-ogy cluster. Our findings show that ties with ser-vice intermediaries contribute to product innova-tion in new ventures and that these relationshipsbecome stronger when the ventures rely more uponsearch in the networks of the cluster in whichthey reside for their product innovation. Theseresults suggest that despite their liability of new-ness, new ventures can create more from less byaccessing a wide range of information, resources,and opportunities through their ties with serviceintermediaries within the cluster. We hope thatour findings can contribute to a better under-standing of how new ventures use their externalties for external innovation search. We encouragefuture research to pay more attention to a broaderset of external ties in which new ventures areembedded.

ACKNOWLEDGEMENTS

We thank Editor Will Mitchell and the two anony-mous referees for their constructive suggestionsand insightful comments. The paper benefited sig-nificantly from the comments by Gautam Ahuja,Bert Cannella, Klaus Ulenbruck, MargaretheWiersema, and Wenpin Tsai.



REFERENCES

Ahuja G. 2000. Collaboration networks, structural holes,and innovation: a longitudinal study. AdministrativeScience Quarterly 45: 425–455.

Aiken LS, West SG. 1991. Multiple Regression: Testingand Interpreting Interactions . Sage: Newbury Park,CA.

Anderson JC, Gerbing DW. 1988. Structural equationmodeling in practice: a review and recommended two-step approach. Psychological Bulletin 103: 411–423.

Arora A, Gambardella A. 1990. Complementarity andexternal linkages: the strategies of the large firms inbiotechnology. Journal of Industrial Economics 38:361–379.

Atwell JD. 2000. Guiding the innovators: why accoun-tants are valued. In The Silicon Valley Edge, Lee C,Miller WF, Hancock MG, Rowen HS (eds). StanfordUniversity Press: Stanford, CA; 355–369.

Avolio BJ, Yammarino FJ, Bass BM. 1991. Identifyingcommon methods variance with data collected from asingle source: an unresolved sticky issue. Journal ofManagement 17: 571–587.

Bahrami H, Evans S. 1995. Flexible re-cycling and high-technology entrepreneurship. California ManagementReview 37: 62–89.

Baker T, Nelson RE. 2005. Creating something fromnothing: resource construction through entrepreneurialbricolage. Administrative Science Quarterly 50:329–366.

Baum JAC, Calabrese T, Silverman BS. 2000. Don’tgo it alone: alliance network composition andstartups’ performance in Canadian biotechnology.Strategic Management Journal , March Special Issue21: 267–294.

Bessant J, Rush H. 1995. Building bridges for innovation:the role of consultants in technology transfer. ResearchPolicy 24: 97–114.

Boeker W. 1997. Executive migration and strategicchange: the effect of top manager movementon product-market entry. Administrative ScienceQuarterly 42: 213–236.

Brown SL, Eisenhardt KM. 1995. Product development:past research, present findings, and future research.Academy of Management Review 20: 343–378.

Bruton GD, Rubanik Y. 1997. High technologyentrepreneurship in transitional economies: the Rus-sian experience. Journal of High Technology Manage-ment Research 8: 213–223.

Cohen WM, Levinthal DA. 1990. Absorptive capacity:a new perspective on learning and innovation.Administrative Science Quarterly 35: 128–152.

Datta DK, Guthrie JP, Wright PM. 2005. Human resourcemanagement and labor productivity: does indus-try matter? Academy of Management Journal 48:135–145.

Fleming L. 2001. Recombinant uncertainty in technolog-ical search. Management Science 47: 117–132.

Hambrick DC, Finkelstein S. 1987. Managerial discre-tion: a bridge between polar views of organizations.

In Research in Organizational Behavior (vol. 9), Cum-mings LL, Staw BM (eds). JAI Press: Greenwich, CT;369–406.

Hargadon A, Sutton RI. 1997. Technology brokeringand innovation in a product development firm.Administrative Science Quarterly 42: 716–749.

Heckman JJ. 1979. Sample selection bias as aspecification error. Econometrica 47: 153–161.

Hitt MA, Dacin MT, Levitas E, Arregle JL, Borza A.2000. Partner selection in emerging and developedmarket contexts: resource-based and organizationallearning perspectives. Academy of ManagementJournal 43: 449–467.

Howells J. 2006. Intermediation and the role ofintermediaries in innovation. Research Policy 35:715–728.

Jaccard J, Turrisi R, Wan CK. 1990. Interaction Effectsin Multiple Regression . Sage: Newbury Park, CA.

Katila R. 2002. New product search over time: past ideasin their prime? Academy of Management Journal 45:995–1010.

Katila R, Ahuja G. 2002. Something old, somethingnew: a longitudinal study of search behavior andnew product introduction. Academy of ManagementJournal 45: 1183–1194.

Katila R, Shane S. 2005. When does lack of resourcesmake new firms innovative? Academy of ManagementJournal 48: 814–829.

Li H, Atuahene-Gima K. 2001. Product innovationstrategy and the performance of new technologyventures in China. Academy of Management Journal44: 1123–1134.

Li H, Atuahene-Gima K. 2002. The adoption ofagency business activity, product innovation, andperformance in Chinese technology ventures. StrategicManagement Journal 23(6): 469–490.

Li H, Bingham J, Umphress E. 2007. Fairness from thetop: perceived procedural justice and collaborativeproblem solving in new product development.Organization Science 18(2): 200–216.

Li H, Hitt MA. 2006. Growth of new technologyventures in China: an introduction. In Growth ofNew Technology Ventures in China’s Emerging Market(New Horizons in International Business Series), Li H(ed). Edward Elgar Publishing Ltd.: Northampton,MA; 3–10.

Li H, Zhang Y. 2007. The role of managers’ politicalnetworking and functional experience in newventure performance: evidence from China’s transitioneconomy. Strategic Management Journal 28(8):791–804.

Lu QW. 2000. China’s Leap into the InformationAge: Innovation and Organization in the ComputerIndustry . Oxford University Press: New York.

Luo Y. 2003. Industrial dynamics and managerialnetworking in an emerging market: the case of China.Strategic Management Journal 24(13): 1315–1327.

Luo Y, Park SH. 2001. Strategic alignment and perfor-mance of market-seeking MNCs in China. StrategicManagement Journal 22(2): 141–155.

McDougall PP, Covin JG, Robinson RB Jr, Herron L.1994. The effects of industry growth and strategic



breadth on new venture performance and strategy con-tent. Strategic Management Journal 15(7): 537–554.

McEvily B, Zaheer A. 1999. Bridging ties: a sourceof firm heterogeneity in competitive capabilities.Strategic Management Journal 20(12): 1133–1156.

Miller D. 1987. The structural and environmentalcorrelates of business strategy. Strategic ManagementJournal 8(1): 55–76.

Nunally JC, Bernstein IH. 1994. Psychometric Theory(3rd edn). McGraw Hill: New York.

Owen-Smith J, Powell WW. 2004. Knowledge networksas channels and conduits: the effects of spillovers inthe Boston biotechnology community. OrganizationScience 15: 5–21.

Peng MW, Luo Y. 2000. Managerial ties and firmperformance in a transition economy: the nature ofa micro-macro link. Academy of Management Journal43: 486–501.

Podsakoff PM, MacKenzie SB, Lee J-Y, Podsakoff NP.2003. Common method biases in behavioral research:a critical review of the literature and recom-mended remedies. Journal of Applied Psychology 88:879–903.

Podsakoff PM, Organ DW. 1986. Self-reports in organi-zational research: problems and prospects. Journal ofManagement 12: 531–544.

Porter ME. 1998. Clusters and the new economicsof competition. Harvard Business Review 76:(November/December): 77–90.

Reed R, DeFillippi RJ. 1990. Causal ambiguity, barriersto imitation, and sustainable competitive advantage.Academy of Management Review 15: 88–102.

Romanelli E. 1989. Environments and strategies oforganization start-up: effects on early survival.Administrative Science Quarterly 34: 369–387.

Rosenkopf L, Nerkar A. 2001. Beyond local search:boundary-spanning, exploration, and impact in theoptical disk industry. Strategic Management Journal22(4): 287–306.

Saxenian A. 1990. Regional networks and the resurgenceof Silicon Valley. California Management Review33(1): 89–112.

Schoonhoven CB, Eisenhardt KM, Lyman K. 1990.Speeding products to market: waiting time to firstproduct introduction in new firms. AdministrativeScience Quarterly 35: 177–207.

Shan W, Walker G, Kogut B. 1994. Interfirm cooperationand startup innovation in the biotechnology industry.Strategic Management Journal 15(5): 387–394.

Shane S, Cable D. 2002. Network ties, reputation, andthe financing of new ventures. Management Science48: 364–381.

Shaver JM. 1998. Accounting for endogeneity whenassessing strategy performance: does entry modechoice affect FDI survival? Management Science 44:571–585.

Simons TL, Peterson RS. 2000. Task conflict andrelationship conflict in top management teams: thepivotal role of intragroup trust. Journal of AppliedPsychology 85: 102–111.

Smith KG, Collins CJ, Clark KD. 2005. Existing knowl-edge, knowledge creation capability, and the rate ofnew product introduction in high-technology firms.Academy of Management Journal 48: 346–357.

Spector PE. 1987. Method variance as an artifact inself-reported affect and perceptions at work: myth orsignificant problem? Journal Applied Psychology 72:438–443.

Starr JA, MacMillan IC. 1990. Resource co-optationvia social contracting: resource acquisition strategiesfor new ventures. Strategic Management Journal ,Summer Special Issue 11: 79–92.

Stinchcombe AL. 1965. Organizations and social struc-ture. In Handbook of Organizations , March JG (ed).Rand McNally: Chicago, IL; 142–193.

Stuart TE. 2000. Interorganizational alliances and theperformance of firms: a study of growth andinnovation rates in a high-technology industry.Strategic Management Journal 21(8): 791–811.

Suchman MC. 2000. Dealmakers and counselors: lawfirms as intermediaries in the development ofSilicon Valley. In Understanding Silicon Valley: TheAnatomy of an Entrepreneurial Region , Kenney M(ed). Stanford University Press: Stanford, CA, 71–97.

Tushman ML, Nadler DA. 1978. Information processingas an integrating concept in organizational design.Academy of Management Review 3: 613–624.

Weick KE. 1969. The Social Psychology of Organizing .Addison-Wesley: Reading, MA.

Wolpert JD. 2002. Breaking out of the innovation box.Harvard Business Review 80: (August): 77–83.

Xin K, Pearce JL. 1996. Guanxi: connections assubstitutes for formal institutional support. Academyof Management Journal 39: 1641–1658.

Zhang Y, Li H, Schoonhoven CB. 2009. Intercommunityrelationships and community growth in China’shigh technology industries 1988–2000. StrategicManagement Journal 30: 163–183.



APPENDIX: PROFILES OF THE TWO RESPONDENTS FROM EACH VENTURE IN OUR SURVEY

Profile First respondent Second respondent

Gender

Male 71.8% 69.6%Female 28.2% 30.4%

Age

29 or less 29.2% 58%30–40 47.5% 28%41–50 20.3% 12%51 and above 3.0% 2%

Level of education

Less than Bachelor Degree 33.2% 14.3%Bachelor Degree 55.9% 54.0%Master Degree 7.9% 23.8%Ph.D. Degree 3% 7.9%

Position

Managing director, CEO 29.2% 29.2%Business manager 21.3% 20.3%Functional manager 49.5% 50.5%

Average work experience in the current industry 6.3 years 6.0 years

Average participation level in strategic decision making 4.0 4.0(on a five-point scale) (on a five-point scale)


innovation search of new ventures in a …yanzh/innovation search_smj 2010.pdf · service...

Documents