emerging market firms' …haiyang/smj_2679 early view.pdf · 2721 examine the direct link...

Strategic Management JournalStrat. Mgmt. J., 38: 2704–2725 (2017)

Published online EarlyView 28 July 2017 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2679Received 2 May 2014; Final revision received 4 April 2017

Emerging Market Firms’ Internationalization: How DoFirms’ Inward Activities Affect Their OutwardActivities?

Haiyang Li,1,* Xiwei Yi,2 and Geng Cui3

1 Jesse H. Jones Graduate School of Business, Rice University, Houston, Texas2 Guanghua School of Management, Peking University, Beijing, China3 Department of Marketing & International Business, Lingnan University,Hong Kong

Research summary: In this study we examine how an emerging market firm’s inward internationalactivities (“inward activities”) are related to its outward international activities (“outwardactivities”) by focusing on the role of the firm’s gain from its inward activities. On the one hand,drawing upon the organizational learning perspective, we propose that a firm’s gain from inwardactivities may facilitate its outward activities through improving its resource fungibility. On theother hand, we draw upon the prospect theory to propose that a firm’s gain from inward activitiesmay hinder its outward activities by discouraging the firm’s top managers from taking risks thatare inherent in outward activities. With detailed data from a sample of manufacturing firms inChina, we find empirical support for both lines of arguments.

Managerial summary: Are emerging market firms with higher inward gain more likely to engagein outward internationalization activities? We argue that it depends upon how a firm uses its gainfrom inward activities. If the firm can improve its resource fungibility (particularly organizationalresource fungibility) from its inward gain, it is more likely to engage in outward activities. Ifthe firm cannot improve its resource fungiblity, the answer is no. Our findings suggest that foremerging market firms, internationalization is not just a path toward new markets; instead, itreflects how these firms exploit and explore what they have learned from their interactions withforeign firms at home in foreign markets. Therefore, managers must think more strategically ondeveloping (organizational) resource fungibility from their inward activities. Copyright © 2017John Wiley & Sons, Ltd.

After years of being investment destinations fortraditional multinational companies (MNCs) indeveloped markets, emerging market countries inthe past decade have become increasingly activein outbound foreign direct investment (FDI).Is the internationalization of emerging marketfirms different from that of MNCs in developed

Keywords: inward activities; outward activities; resourcefungibility; emerging market; internationalization*Correspondence to: Haiyang Li, Jesse H. Jones Graduate Schoolof Business, Rice University, 6100 Main Street, Houston, TX77005-1892. E-mail: [email protected]

Copyright © 2017 John Wiley & Sons, Ltd.

markets? While these two groups of firms sharemany aspects in their internationalization processes(Dunning, Kim, & Park, 2008), one key differenceis well accepted. That is, since emerging marketfirms’ internationalization is relatively recent, theyhave benefited tremendously from the presence ofinward FDI in their home markets (Buckley, Clegg,& Wang, 2002; Gu & Lu, 2011; Luo & Tung, 2007;Y. Zhang, Li, Li, & Zhou, 2010). More specifically,inward international activities (hereafter “inwardactivities”) such as distributing foreign productsor forming a joint venture with a foreign firm pro-vide opportunities for domestic firms in emerging

Emerging Market Firms’ Internationalization 2705

markets to acquire important resources and capabil-ities from foreign firms (Steensma, Tihanyi, Lyles,& Dhanaraj, 2005; Y. Zhang et al., 2010), whichcan help them adapt to international business envi-ronments and facilitate their outward internationalactivities (hereafter “outward activities”).

The key argument underlying this line of researchis that it is the resources and capabilities that afirm acquires from its inward activities that facilitateits outward activities. However, previous studieshave not made a distinction between what a firmdoes in terms of inward activities and what it gainsfrom its inward activities (e.g., Buckley et al., 2002;Karlsen, Silseth, Benito, & Welch, 2003; Luo &Tung, 2007). Rather, it is assumed that to the extenta firm engages in inward activities, it will gainresources and capabilities from such activities. Notethat not all firms benefit to the same extent fromtheir inward activities. Therefore, it is crucial toseparate a firm’s gain from inward activities fromthe inward activities in which it participates. Moreimportantly, the relationship between a firm’s gainfrom inward activities and its outward activitiesmay not be as straightforward as previously argued.Indeed, Li, Li, and Shapiro (2012) have shown thatan emerging market firm’s gain from its inwardactivities may decrease the firm’s propensity toparticipate in outward activities because of the riskynature of these activities.

In this study, we aim to address these issuesby revisiting the link between a firm’s inward andoutward activities and focusing on the role of thefirm’s gain from inward activities. A firm’s gainfrom inward activities (hereafter “inward gain”)refers to the benefits (e.g., capital, technologies, andknowledge) that the firm achieves from engagingin inward activities. We propose that inward gainhas two possible effects on a firm’s participation inoutward activities. On the one hand, drawing uponthe organizational learning perspective (Helfat &Peteraf, 2003; March, 1991), we argue that a firm’sinward gain will improve its resource fungibility,which will further increase the firm’s participationin outward activities. Resource fungibility refers tothe extent to which a firm’s resources and capa-bilities can be deployed in multiple geographicaland country settings (Oviatt & McDougall, 1994;Sapienza, Autio, George, & Zahra, 2006). On theother hand, drawing upon the prospect theory (Kah-neman & Tversky, 1979), we argue that controllingfor the effect of resource fungibility, a firm’s inwardgain may induce its top managers to frame their

situation as a gain domain. As such, these top man-agers may behave in a risk-averse manner and areless likely to engage in risky activities such as out-ward activities (Carpenter, Pollock, & Leary, 2003).

We tested these arguments with detailed data col-lected from 314 manufacturing firms in China’semerging market. We found that a firm’s inwardgain is positively related to the firm’s outward activ-ities through improving the firm’s organizationalresource fungibility (but its technological resourcefungibility does not seem to matter in this rela-tionship). Moreover, after controlling for this indi-rect relationship, we found that a firm’s inwardgain has a negative direct relationship with its out-ward activities. This direct negative relationship ismore salient for high-risk outward activities (e.g.,overseas mergers and acquisitions and establish-ing wholly owned overseas subsidiaries) than forthose with low risk (e.g., franchising). This negativerelationship is also stronger for firms with low-riskpropensity than for those with high-risk propensity.

Overall, our findings suggest that there are threepossible paths through which a firm’s inward activ-ities may be related to its outward activities: (a) afirm’s inward activities have a direct positive rela-tionship with its outward activities; (b) a firm’sinward gain has an indirect positive relationshipwith the firm’s outward activities through improv-ing its organizational resource fungibility; and (c)a firm’s inward gain has a direct negative relation-ship with the firm’s participation in outward activi-ties, particularly if those activities involve great riskand/or if the firm has low-risk propensity. Our the-oretical arguments and empirical findings provide amore complete picture of the inward–outward con-nection of emerging market firms’ internationaliza-tion and make important contributions to both thestrategy and the international business literature.

Theory and Hypotheses

Research Background

A firm’s internationalization comprises two impor-tant processes: inward activities and outwardactivities (Gu & Lu, 2011; Johanson & Vahlne,1977; Korhonen, Luostarinen, & Welch, 1996;Luo & Tung, 2007). Inward activities refer to theactivities that domestic firms engage in with foreignfirms in their home market, including technologyimports, product imports, product agency business,original equipment manufacturing (OEM), inward

Copyright © 2017 John Wiley & Sons, Ltd. Strat. Mgmt. J., 38: 2704–2725 (2017)DOI: 10.1002/smj

2706 H. Li, X. Yi, and G. Cui

franchising and licensing, and joint ventures. Out-ward activities refer to the activities that domesticfirms conduct in overseas markets, includingfranchising and licensing to foreign firms, overseasmergers and acquisitions, and establishing whollyowned subsidiaries in overseas markets.

In comparing recent outward activities ofemerging market firms with those of developedmarket firms, Dunning et al. (2008, p. 177) wrote:“perhaps most important of all, unlike yesterday’sdeveloped country TNCs (transnational compa-nies), today’s emerging market TNCs rarely havethe firm specific ownership advantages (notablyorganizational and management skills) to ensuresuccess in their outward FDI” (italics added bythe authors). In this sense, emerging market firms’outward activities are disadvantaged because theydo not have ownership advantages, which arehighlighted in Dunning’s (1993) OLI (Ownershipadvantage, Locational advantage, and Internaliza-tion) paradigm as well as in Johanson and Vahlne’s(1977) stage model for internationalization.

Scholars have argued that the inclusion of inwardactivities in the internationalization framework canhelp elucidate emerging market firms’ outwardactivities given their ownership disadvantage (Gu& Lu, 2011; Luo & Tung, 2007). Different fromthe internationalization path of developed mar-ket firms, emerging market firms have benefitedtremendously from inward activities with foreigncompanies (Luo & Tung, 2007; Y. Zhang, Li, & Li,2014). These inward activities may generate knowl-edge spillovers, which allow emerging market firmsto ascertain how to conduct businesses internation-ally even before actually investing overseas.

Several empirical studies have examined the rela-tionship between inward and outward activities inthe context of emerging markets. For example,Young, Huang, and McDermott (1996) used casestudies of Chinese multinationals and found thatthrough inward internationalization, Chinese firmshave accumulated considerable financial and oper-ational assets and improved technological and oper-ational assets, thus strengthening their outwardexpansion activities. Using cross-sectional data,Buckley et al. (2002) found that Chinese firms couldlearn advanced technologies from their foreign part-ners through inward activities, which improves theirexport performance. In addition, Gu and Lu (2011)used cross-country, industry-level data on venturecapital investments worldwide from 1985 to 2007to examine how inward investments from the home

country influence outward investments from thehost country. Their results suggest that even thoughemerging market firms do not have the advantagesor resources to expand internationally, their homemarket internationalization through inward invest-ments of foreign firms can facilitate future growthinto global markets.

Theoretical Model

In this study, we revisit the link between a firm’sinward and outward activities by focusing on therole of the firm’s inward gain. While outwardactivities in general are risky to any firms (Car-penter et al., 2003; Hitt, Hoskisson, & Kim, 1997),this is particularly true in the case of emergingmarket firms. As previously noted, compared withtraditional MNCs in developed countries, emergingmarket firms do not have ownership advantagesthat can help mitigate the liability of foreignnessin international markets (Zaheer, 1995). Moreover,emerging market firms typically face an underde-veloped financial system at home that can seriouslylimit their access to financial capital for overseasinvestment (Aulakh, Rotate, & Teegen, 2000).These firms also lack the legitimacy of their homecountry origins (Gubbi, Aulakh, Ray, Sarkar, &Chittoor, 2009; Luo & Tung, 2007) and thus havelimited credibility in overseas markets. In addition,because most emerging markets have only recentlyopened to the global economy, there is a scarcepool of managerial talent that can help these firms’participation in international operations (Aulakhet al., 2000; Ramachandran & Pant, 2010). Equallyimportant, isolated by cultural, administrative, geo-graphic, and economic distance from internationalmarkets (Ghemawat, 2001), emerging market firmsmay also have difficulties in comprehending foreignmarkets.

Considering the risk that is inherent in outwardactivities for emerging market firms, we argue thatemerging market firms’ inward gain may affecttheir outward activities in two different ways. Onthe one hand, if a firm’s inward gain can improveits capabilities for international operations, the riskinherent in outward activities should be of lessconcern to the firm, and as a result, the firmis more likely to explore business opportunitiesin international markets. In other words, a firm’sinward gain may facilitate its outward activitiesby improving its capabilities for undertaking suchactivities. On the other hand, due to the risk inherent



Inward gain

Resource fungibility: - Organizational vs.

Technological (H2)

(H4)

Outward activities - High-risk vs. Low-risk

Path I(a) H1(+)

Path I(b) H1 (+)

Path II H3 (-)

Control variables

Inward activities

Firm riskpropensity

Inward gain (H4)

Outwar- High-risk

Path I(a)H1(+)

Path I(b)H1 (+)

Path II H3 (-)

C

Firm rm iskpropensity

H5 (+)

Figure 1. The conceptual model.

in outward activities, a firm’s inward gain maydiscourage the firm’s top managers from taking therisk. In short, a firm’s inward gain may hinder itsoutward activities by reducing its motivation forundertaking such activities.

In the following sections, we draw upon the orga-nizational learning perspective (Cyert & March,1963; Helfat & Peteraf, 2003; Sapienza et al., 2006)and the prospect theory (Kahneman & Tversky,1979), respectively, to develop these two lines ofarguments. Figure 1 presents the overall theoret-ical model of this study. Considering that previ-ous studies have primarily focused on the directlink between inward and outward activities, we alsoinclude this link in our model to control for alterna-tive explanations.

Path I. Inward Gain, Resource Fungibility,and Outward Activities: An OrganizationalLearning Perspective

The organizational learning perspective empha-sizes the importance of experiential learning (Cyert& March, 1963; Huber, 1991; March, 1991), aprocess by which organizations as collectivesderive meaning from direct observations or inter-actions with their environments (i.e., learning fromexperience). In the emerging market context, priorresearch argues that a firm’s engagement in inwardactivities facilitates the learning of foreign advancedtechnologies, establishes routines for international

operations, and cultivates a compatible organiza-tional culture (Guthrie, 2005; Luo & Tung, 2007)that may be used for the firm’s outward activities.

We argue that a firm’s inward gain may stim-ulate its outward activities through improving itsresource fungibility. Resource fungibility repre-sents a firm-level capability that enables the firmto deploy existing resources to new endeavors, thusreducing the cost of new investment (Oviatt &McDougall, 1994; Vassolo, Anand, & Folta, 2004).The idea that fungible resources can create benefitsfor operations in multiple market domains is alsoreflected in the resource-based view. For example,Penrose (1959) argued that when the same resourcecan be used for different purposes or in differentways, more diverse services could be provided.

Outward activities often require firms to dealwith unfamiliar environmental settings. Thus, suchactivities intrinsically incur a higher level of com-plexity and risk and consume significant resources(Carpenter et al., 2003). In emerging markets, firmsare generally characterized with limited resources,and their resource acquisition activities are oftenconstrained because local strategic factor marketsare underdeveloped (Y. Zhang & Li, 2010). Bydeploying resources acquired from inward activi-ties to outward activities, emerging market firmscan lower their initial investment of outward activ-ities to an affordable extent. Also, prior researchhas shown that high-resource fungibility increasesfirms’ propensity to engage in experimentation



because deploying the same resource for alter-native purposes reduces the cost of experimen-tation, whereas low-resource fungibility inhibitsexperimentation by increasing the cost of develop-ing or implementing new organizational processes(Autio, George, & Alexy, 2011). Outward activi-ties are largely comparable to experimentation sinceoutward activities often necessitate trial-and-errorlearning, which implies executing various routinesuntil successful outcomes are finally discovered(Sapienza et al., 2006). As argued by Hitt et al.(1997) and Sapienza et al. (2006), a firm mustadjust its resource configurations and organizationalprocess routinization to support subsequent inter-national activities. Thus, as resource fungibilityencourages experimentation, it may also encouragea firm’s outward activities.

Previous research has provided support for ourargument. For example, Guthrie (2005) arguedthat inward partnership with foreign firms trans-fers Western management practices to China, thusmaking Chinese firms’ managerial systems suitablefor international observance. Luo and Tung (2007)argued that inward activities, ranging from importand OEM to alliances and joint ventures, preparethe indigenous firms to adjust their resource con-figurations to international operations. In addition,as Anand and Delios (2002) noted, technology isintrinsically fungible across borders compared withdownstream resources, such as distribution systemsand sales force. Thus, technological gains can helpdomestic firms improve their technological capabil-ities that can be easily deployed across borders.

High levels of resource fungibility may be par-ticularly crucial for emerging market firms’ out-ward activities because these firms tend to be moreresource-constrained than developed country firms(Hitt, Li, & Worthington, 2005). High levels ofresource fungibility also allow these firms to shareor reallocate their limited resources across mul-tiple geographic markets (Sapienza et al., 2006).In summary, we argue that a firm’s gain from itsinward activities may facilitate its outward activ-ities by improving its capabilities for undertakingsuch activities. Therefore, we propose the followinghypothesis.

Hypothesis 1 (H1): A firm’s inward gain ispositively related to the firm’s outward activitiesthrough improving its resource fungibility.

We further make a distinction between orga-nizational resource fungibility and technologicalresource fungibility. Organizational resourcefungibility refers to the extent to which a firm’sorganization-related resources and capabilities(e.g., organizational structure, culture, and man-agement system) are adaptable across multiplegeographical regions, while technological resourcefungibility refers to the extent to which a firm’stechnological resources and capabilities (e.g.,manufacturing technologies and new productdevelopment technologies) are adaptable acrossgeographical regions. Although both are important,organizational resource fungibility may be morecritical than technological resource fungibility foremerging market firms’ outward activities.

There are two primary reasons. First, while alloutward activities of emerging market firms requiresome organizational resources, not all of theiroutward activities require technological resources.Take overseas acquisitions as an example. Emergingmarket firms would need organizational resourcesto negotiate the deals and manage the integrationprocesses. As Rui and Yip (2008) have noted,Chinese firms’ overseas acquisitions are closelyrelated to the growing entrepreneurship and man-agement skills embedded in these firms. However,these firms do not necessarily need technologicalresources for overseas acquisitions. This is becausemany emerging market firms use internationalexpansions, particularly overseas acquisitions, as a“spring board” to seek advanced technologies fromdeveloped markets (Luo & Tung, 2007).

Second, relative to technological resources,which may be used across country boundaries(Anand & Delios, 2002), organizational resourcesare more specific to individual countries andcultures. Note that emerging market firms generallyface significant challenges in overseas marketsincluding reconciling disparate national- andcorporate-level cultures, organizing globally dis-persed complex activities, and integrating home andhost country operations (Luo & Tung, 2007). Sinceemerging market firms typically lack internationalexperience and organizational expertise in handlingthese issues (Ghemawat, 2001), developing orga-nizational resource fungibility becomes imperativefor them to participate in outward activities. Fungi-ble organizational routines, processes, and culturescan help coordinate activities within the firms andacross borders and facilitate the internationalization



process (Sapienza et al., 2006). Thus, we proposethat,

Hypothesis 2 (H2): The relationship as proposedin H1 is stronger when a firm’s organizationalresource fungibility is improved than when itstechnological resource fungibility is improved.

Path II. Inward Gain and Outward Activities:A Prospect Theory Perspective

Prospect theory describes how individuals choosebetween probabilistic alternatives and evaluatepotential losses and gains (Kahneman & Tversky,1979). The central premise of this theory is thatindividuals tend to frame alternatives either as again domain or as a loss domain in comparison toa reference point, and they show tendencies towardbeing more aggressive and risk-seeking in the lossdomain but more conservative and risk-averse inthe gain domain. While this theory was initiallydeveloped at the individual level, several studieshave provided strong support for the implicationsof its basic propositions to firm behaviors. Forexample, using extensive data on U.S. firms,Fiegenbaum and Thomas (1988) consistentlyfound that a large majority of firms are risk-aversewhen experiencing performance gain (above thetarget level) and risk-seeking when experiencingperformance loss (below the target level). Lehner(2000), using different measures of risk and longertime frames, found similar findings.

As previously noted, outward activities are riskyfor emerging market firms. To the extent that a firmhas gained from its inward activities, its top man-agers tend to perceive the situation as a gain domain.In this situation, according to the prospect theory,managers tend to behave in a risk-averse and conser-vative manner because they now have something tolose. Managers may also become concerned aboutthe possibility of losing their current gain (Sitkin &Pablo, 1992) if they take the risk that is inherent inoutward activities. Thus, top managers of firms withgreater inward gain are more likely to adopt a con-servative attitude toward outward activities. There-fore, according to the prospect theory, we argue thatto the extent that emerging market firms have gainedfrom their inward activities, their motivation for par-ticipating in outward activities may be reduced. AsLi et al. (2012, p. 280) argued, “Since emergingmarket firms can benefit from inward FDI in theirhome markets, inward FDI represents an important

alternative source for them to acquire technologicalknowledge.” Empirically, with a sample of Chi-nese manufacturing firms, Li et al. (2012) foundthat inward FDI from countries with advanced tech-nologies decreases the propensity of Chinese firmsto invest in those countries for technology seek-ing. Based on the abovementioned discussions, weexpect that after controlling for the resource fun-gibility effect, a firm’s inward gain may have adirect and negative relationship with the firm’s par-ticipation in outward activities, because its inwardgain constructs a gain domain that may reduce itstop managers’ motivation for outward activities.Therefore,

Hypothesis 3 (H3): A firm’s inward gain has adirect negative relationship with the firm’s out-ward activities, after controlling for its indi-rect relationship with outward activities throughresource fungibility.

To further validate the logic underlying Hypoth-esis 3, we follow Shimizu’s (2007) suggestionof taking contextual factors into account whenapplying the prospect theory at the organizationallevel. More specifically, we argue that while, ingeneral, inward gain may be related to risk-averseand conservative managerial behaviors accordingto the prospect theory, the negative relationshipbetween inward gain and outward activities maydepend upon the risk levels of outward activities(i.e., high-risk vs. low-risk outward activities) aswell as the firms’ own risk propensity. We developspecific hypotheses in the following sections,which help verify the underlying assumption ofHypothesis 3 regarding risk and, if supported, maylend additional support to the prospect theory logicunderlying Hypothesis 3.

High-risk vs. low-risk outward activities. Dif-ferent types of outward activities involve differentlevels of resource commitment and risk (Johanson& Vahlne, 1977). For example, exporting, franchis-ing, or establishing sales subsidiaries and R&Dfacilities in overseas markets tend to require lessresource commitment than building joint ventures,acquiring firms, or establishing wholly ownedproduction subsidiaries in overseas markets. Rela-tively, the former group of activities requires fewerresources and involves lower risk than the lattergroup of activities. In particular, most emergingmarket firms have low managerial capabilities and



weak technological competences. These firms willface greater challenges when becoming involvedin joint ventures, mergers and acquisitions, andwholly owned production subsidiaries as opposedto conducting exporting, franchising, or establish-ing sales or R&D facilities. Therefore, we expectthat for outward activities with low risk, the directnegative relationship between inward gain andoutward activities will be less salient, while thenegative direct relationship between inward gainand outward activities will be stronger for outwardactivities with high risk. Thus, we propose that,

Hypothesis 4 (H4): The direct negative relation-ship between inward gain and outward activitiesas proposed in H3 is stronger for high-risk out-ward activities than for low-risk outward activi-ties.

Firm risk propensity. According to theprospect theory, a firm’s inward gain may discour-age its outward activities by reducing the firm’stop managers’ motivation for outward activitiesbecause in a gain domain, the managers tend tobehave in a more risk-averse manner (Kahneman& Tversky, 1979). However, firms differ in theirrisk propensity, which refers to a firm’s proclivityto engage in risky projects and its managers’preferences for bold vs. cautious actions to achievefirm objectives (Miller, 1983). We expect that sincea firm’s risk propensity affects its top managers’willingness to take risks, it may moderate the directnegative relationship between the firm’s inwardgain and its participation in outward activities.

For firms with low-risk propensity, their topmanagers are more likely to have a strong senseof loss aversion (Levy, 1996), an attitude thatvalues what they have more than comparable thingsthey do not have (Tversky & Kahneman, 1991).Thus, top managers of these firms are less willingto take additional risks that would be associatedwith outward activities. In contrast, firms withhigh-risk propensity have a greater appetite for riskyactions, including overseas expansions, which willreduce their top managers’ concerns of possible lossassociated with outward activities. Therefore, weargue that the direct negative relationship betweeninward gain and outward activities will be moresalient for firms with low-risk propensity than forthose with high-risk propensity. Based on the abovearguments, we propose the last hypothesis.

Hypothesis 5 (H5): The direct negative relation-ship between inward gain and outward activi-ties as proposed in H3 is stronger for firms withlow-risk propensity than for firms with high-riskpropensity.

Methods

Sample and Data Collection

The data used in this study was collected througha survey in China. We conducted the survey in2010 with the help of a major global research com-pany and its proprietary panel of firm executives.The research company regularly uses this panelto perform surveys to generate reports on industrydevelopment and managerial problems facing thesefirms.

In this study, we focused on four major manu-facturing sectors—food and beverage, textile andapparel, electric machinery and electronics, andtransportation equipment manufacturing. We chosenot to include firms in natural resources, financeand banking, and business services because basedon our interviews, outward activities in the first twosectors often reflect government agendas and out-ward activities in business services often requiredifferent (lower) levels of resource commitmentthan those in manufacturing sectors.

In the research company’s proprietary list, therewere 54,282 firms in these four target industries,which represented 33.6% of all firms in these indus-tries at the national level (there were 161,772 firmsin these industries in China based on the IndustryEnterprises Survey by China’s National StatisticalBureau in 2009).1 Considering the large number offirms in the proprietary list, it is reasonable to arguethat the list has a high level of representativeness offirms in the target industries in China. For budgetaryreasons, we selected 1,500 firms from the list witha random sample of 375 firms from each of the fourindustries (375*4).

With the help of the research company, we sentquestionnaires to the 1,500 firms. At the beginning

1 The database of Industrial Enterprises Survey contains the mostcomprehensive information about firms in China (Tian, 2007; Y.Zhang et al., 2010, 2014). By law, all firms in China are requiredto cooperate with the National Statistical Bureau and submit theirbasic and financial information to the bureau (Chang & Xu, 2008).The aggregation of firm-level information collected by the bureauis published in the official China Statistics Yearbooks.



of the questionnaire, we asked whether the companyhad any international business activity. Only thosewith international business activity were invited tocomplete the questionnaire. We adopted this screen-ing procedure because not all firms are interestedin international businesses and our research ques-tions on the linkages between inward and outwardactivities would not be relevant to firms that are notinterested in international businesses. Possible lim-itations of this sampling approach will be furtherdiscussed in the discussion section.

To minimize the bias of common method vari-ance, we designed the questionnaire by followingthe procedures suggested by Podsakoff, MacKen-zie, Lee, and Podsakoff (2003). We carefully con-structed the questionnaire items based on the lit-erature in English and then translated the itemsinto Chinese to make them as simple, specific,and concise as possible. Comprehension problemscaused by item complexity or ambiguity may haveinduced respondents to develop their own idiosyn-cratic meanings, which may have resulted in com-mon method bias (Podsakoff et al., 2003). Itemwording and terminology in Chinese were refinedaccordingly to ensure the validity and appropri-ateness of the measures in the Chinese context.The Chinese version was then back-translated intoEnglish and we paid special attention to any misun-derstandings that might have arisen due to transla-tion. On the questionnaire, we assured respondentsthat (a) the collected information is for academicuse only, (b) there were no right or wrong answersto the questions, and (c) respondent anonymity wasguaranteed.

The questionnaire was sent to top executivesof the sampled firms who were familiar withtheir firms’ international activities. After 2 weeks,we made follow-up phone calls and sent emailreminders to those who had not responded. Finally,314 completed questionnaires were collected,resulting in a response rate of 21%. At the end ofthe survey, our research assistants independentlyplaced phone calls to a random 10% of the respon-dents to verify their identities and responses tothe survey. All of the respondents confirmed thatthey had participated in the survey. To check fornonresponse bias, we compared the respondents’firm size and industry profiles with those ofnonrespondents. We found no statistically signif-icant differences. We also used t tests to examineif differences existed in terms of the means of keyconstructs between early and late respondents as

recommended by Armstrong and Overton (1977)and we found no statistically significant differ-ences. These results indicate that nonresponse biaswas not an issue in the study.

Four weeks after completion of the major sur-vey, we sent all of the respondents a shorter ver-sion of the questionnaire that only included themeasures of the key constructs for validation pur-poses. We received responses from 146 firms inthe second-round survey. These data were used toexamine the reliability of the survey in the firstround and to conduct a robustness check of our find-ings. To examine the reliability of the first-roundsurvey, we ran a Pearson’s correlation test to eachpaired-item and found that for every paired-item,the correlation coefficient was significant and above0.9, indicating a high reliability of our data. Wealso calculated the aggregated scores for inwardactivities and outward activities, and found thatfirst-round scores and second-round scores werehighly correlated (0.95 and 0.96, respectively).

Measurement

Table 1 lists the measures for all of the multi-itemconstructs, each of which will be discussed in detailbelow. We averaged the items to create the scoresfor each of the constructs.

Outward activities. We draw upon the work byKorhonen et al. (1996) and Luo and Tung (2007)and asked each respondent to evaluate the degreeto which his/her firm had engaged in various typesof outward international activities. These activitiesincluded: (a) franchising in overseas markets, (b)exporting products to overseas markets, (c) formingalliances or joint ventures in overseas markets, (d)acquiring or merging firms in overseas markets, (e)establishing wholly owned production subsidiariesin overseas markets, (f) establishing wholly ownedsales subsidiaries in overseas markets, and (g)establishing wholly owned R&D facilities inoverseas markets.

To test H4, we categorized outward activitiesinto high-risk and low-risk based on resource com-mitment and risk that each activity may involve.Low-risk outward activities include: (a) franchisingin overseas markets, (b) exporting products tooverseas markets, (c) establishing wholly ownedsales subsidiaries in overseas markets, and (d)establishing wholly owned R&D facilities in over-seas markets. High-risk outward activities include:(a) forming strategic alliances or joint ventures in



Table 1Coefficients, Z-statistics, and Reliability for the Measurement Modela

Item descriptionsStandardized

coefficientStandard

error Reliability AVE

Outward activities b 0.9180 0.6212Rate the degree to which your firm is involved in the following activities in oversea markets:

Franchising in oversea markets 0.822 /Exporting products to overseas markets 0.479 0.047***Forming alliances or joint ventures in overseas markets 0.850 0.056***Acquiring or merging firms in overseas markets 0.831 0.058***Establishing wholly owned production subsidiary in overseas markets 0.841 0.055***Establishing wholly owned sales subsidiaries in overseas markets 0.774 0.052***Establishing wholly owned R&D subsidiaries in overseas markets 0.851 0.054***

Inward activities b 0.8320 0.5002Rate the degree to which your firm is involved in the following activities in the domestic market:

Importing new products and/or new services from foreign countries 0.642 /Acting as agents for selling foreign companies’ products 0.664 0.117***Introducing foreign capital (including foreign venture capital) 0.762 0.109***Forming joint ventures with foreign firms in China 0.814 0.129***Introducing production lines or manufacturing equipment from foreignfirms

0.636 0.079***

Inward gainb 0.8330 0.5003Rate the degree to which your firm has benefited from inward international activities along the following aspects:

The firm has obtained sufficient foreign capital 0.699 /The firm has acquired advanced manufacturing technologies 0.735 0.082***The firm has acquired advanced product development technologies 0.748 0.090***The firm has acquired advanced management knowhow and managerialtalent

0.721 0.085***

The firm has acquired a large amount of overseas market knowledge andcustomer information

0.627 0.081***

Resource fungibilityc 0.8733 0.5362Rate your firm’s resource deployment from the following aspects:

The firm’s overall resources can adapt well to overseas markets 0.650 /The firm’s organizational structure can adapt well to overseas markets 0.784 0.115***The firm’s organizational culture can adapt well to overseas markets 0.778 0.119***The firm’s product manufacturing technologies can adapt well tooverseas markets

0.697 0.106***

The firm’s product development technologies can adapt well to overseasmarkets

0.670 0.104***

The firm’s human resource management system can adapt well tooverseas markets

0.800 0.126***

Industry competition intensityc 0.7995 0.5067Evaluate the accuracy of the following statements about your key industry

There is fierce competition in the industry 0.897 /There are frequent promotions in the industry 0.581 0.082***There are frequent price wars in the industry 0.709 0.085***There is a high level of product homogeneity in the industry 0.618 0.088***

N = 314.a Significance levels: p< .001***; p< .01**; p< .05*; p< 1†.b 7-point scale: 0= none; 6= a lot.c 7-point scale: 1= totally disagree; 7= totally agree.

overseas markets, (b) acquiring or merging firmsin overseas markets, and (c) establishing whollyowned production subsidiaries in overseas markets.

Inward activities. To measure inward activities,we asked each respondent to indicate the degree towhich his/her firm was involved in various types

of inward activities, including (a) importing newproducts or new services from foreign countries,(b) acting as agents for selling foreign companies’products, (c) introducing foreign capital (includingforeign venture capital), (d) forming joint ven-tures with foreign companies in China, and (e)



introducing product lines or manufacturingequipment from foreign countries.

Inward gain. This construct was measured byasking the respondent to indicate the degree towhich his/her firm had benefited from inward activ-ities compared with competitors in the follow-ing dimensions: (a) the firm has obtained suffi-cient foreign capital, (b) the firm has acquiredadvanced manufacturing technologies, (c) the firmhas acquired advanced product development tech-nologies, (d) the firm has acquired advanced man-agement knowhow and managerial talent, and (e)the firm has acquired a large amount of overseasmarket knowledge and customer information.

Resource fungibility. Drawn upon the work byOviatt and McDougall (1994) and Sapienza et al.(2006), we measured resource fungibility by ask-ing the respondent to indicate the degree to whichhis/her firm’s resources could be easily applicable tocross-border deployment. We focused on six dimen-sions: (a) overall resources, (b) organizational struc-ture, (c) organizational culture, (d) product man-ufacturing technologies, (e) product developmenttechnologies, and (f) human resource managementsystem. To test H3, we further categorized resourcefungibility into technological resource fungibilityand organizational resource fungibility. Techno-logical resource fungibility includes the followingtwo dimensions: (a) product manufacturing tech-nologies and (b) product development technolo-gies. Organizational resource fungibility includesthe following three dimensions: (a) organizationalstructure, (b) organizational culture, and (c) humanresource management system.

Firm risk propensity. Based on Miller (1983),this construct was measured by asking each respon-dent to indicate the degree to which his/her firmfavored risk-taking. We used two items: (a) Our firmhas a strong culture for adventure and risk-takingand (b) we believe that a strategy emphasizingstability is more suitable for our firm than astrategy emphasizing adventure (reverse-coded).We calculated the sum of these two items (afterreverse-recoding the second one) and divided thesample into two groups (high-risk vs. low-riskpropensity firms) based on this variable’s samplemedian value.

Control variables. To tease out alternativeexplanations, we controlled for several firm-and industry-level factors in our analyses. Wecontrolled for firm size because larger firms may bemore likely to engage in outward activities. Firm

size was measured by the natural log of the numberof employees in a firm. Firm age was controlled forbecause according to Johanson and Vahlne’s (1977)stage model for internationalization, older firms aremore likely to engage in outward activities. Firmage was measured by the natural log of the numberof years since a firm’s inception. R&D intensitywas controlled for because previous literaturesuggests that firms that invest heavily in R&Dare more likely to take the risk to go international(Filatotchev & Piesse, 2009). R&D intensity wasmeasured by asking the respondent to indicate theaverage percentage of a firm’s R&D investment toits sales revenue in the past 3 years.

International experience of the top managementteam (TMT) may also play an important role ina firm’s internationalization decision (Carpenter,Sanders, & Gregersen, 2001). Following Samb-harya (1996), we measured this variable by usingthe average number of years that TMT membershave spent abroad on assignment and/or in highereducation, or in an international division. Wedefined TMT members as including the highestlevel of management—the chairman, chief execu-tive officer, president, and chief operating officer—as well as the next highest tier (Wiersema & Bantel,1992).

At the industry level, we controlled for industrycompetition intensity, which was measured by fouritems: (a) the degree of fierce competition in theindustry, (b) the frequency of price wars in theindustry, (c) the frequency of promotions in theindustry, and (d) the degree of product homogeneityin the industry.

Structural Equation Method

We followed Anderson and Gerbing’s (1988)recommendation and used structural equationmodeling (SEM) to test our hypotheses. We choseSEM rather than multiple regressions for thefollowing reasons. First, it has been argued thatSEM is particularly effective when testing modelsthat contain latent constructs that are measuredwith multiple indicators (Steensma & Lyles, 2000).In our model, a number of variables are latent con-structs measured with multiple indicators. Second,an important difference between SEM and multipleregressions is that SEM has a unique ability tosimultaneously examine a series of dependencerelationships (Shook, Ketchen, Hult, & Kacmar,2004). In our model, the theoretically related and



intertwined relationships between inward activities,inward gain, resource fungibility, and outwardactivities can be better tested by using SEM than bymultiple regressions. Third, SEM differs from mul-tiple regressions in that SEM can simultaneouslyanalyze multiple dependent variables (Joreskog,Sorbom, du Toit, & du Toit, 1999). In our study,testing H4 required comparing the effects ofpredicting variables on two dependent variables(low-risk outward activities and high-risk outwardactivities). This kind of comparison can be betterdone with SEM than with multiple regressions.

We conducted a two-step analysis. In the firststep, we conducted a confirmatory factor analysisto test the construct validity of the latent variableswe defined, including both convergent validityand discriminant validity. In the second step, weestablished the structural model based on themeasurement model verified in the first step. Thestructural model tested our hypotheses. In addi-tion, SEM enables the comparison of alternativemodels that helps to identify causal relationships(James, Mulaik, & Brett, 1982). Therefore, thistwo-step approach allowed us to evaluate theconstruct validity in the measurement model andassess the adequacy of the proposed theoreticalrelationships in the structural model, respectively(Bollen, 1989).

Results

Measurement Model

We validated the measurements by following thestandard procedures (Anderson & Gerbing, 1988).First, for each of these constructs, we conducted aprincipal component factor analysis with varimaxrotation. The results demonstrate that for each ofthe constructs, only the first eigenvalue is greaterthan one, leading support to the unidimensionalityof these constructs (Anderson & Gerbing, 1988).Second, we conducted a confirmatory factor anal-ysis by including five multi-item constructs pre-viously proposed (i.e., outward activities, inwardactivities, inward gain, resource fungibility, andindustry competition intensity).2 We then examined

2 “Firm’s risk propensity” was not included in the measurementmodel because it was not used as a latent construct. Instead, asdiscussed in the Methods section, we divided the sample into twogroups (high-risk propensity firms vs. low-risk propensity firms)based on this variable’s sample median value.

the fit to the data of the five-factor model in whichthe five latent variables were assessed by differentsets of indicators. The measurement model fitsthe data well by demonstrating a series of goodmodel fit indexes (𝜒2 = 724.87, Degree of Freedom(DF)= 314, Chi-Square (CMIN)/DF= 2.31, Com-parative Fit Index (CFI)= 0.91, Root Mean SquareError of Approximation (RMSEA)= 0.060).

Table 1 presents the standardized coefficients foreach item, composite reliability, and average vari-ance extracted (AVE) for each construct. The resultsshow that the five latent variables perform well interms of their convergent validity. Convergent valid-ity was demonstrated through the statistical signif-icance of the measurement model’s parameter esti-mates using a Z-ratio. The Z-ratio is calculated bydividing the parameter estimate by its standard error(Widaman, 1985) and the Z-ratio greater than 1.96is considered significant at the .05 level. We alsoexamined convergent validity through the size offactor loadings. All of the factor loadings are greaterthan 0.40, showing support for convergent valid-ity (Ford, MacCallum, & Tait, 1986). Further, wecalculated the AVE, which measures the amountof variance captured by a specific construct rela-tive to the amount of variance attributable to themeasurement model. The AVEs for the constructsrange from 0.5002 to 0.6212, all of which are abovethe 0.5 cut-off value (Fornell & Larcker, 1981).Finally, we calculated the composite reliability foreach latent variable by dividing the squared sum ofthe individual standardized loadings by the sum ofthe variance of their error terms and the squared sumof the individual standardized loadings. The valueof composite reliability ranges from 0.80 to 0.92,which all exceed the threshold value of 0.70 (Nun-nally, 1978). These results suggest that the measure-ment model shows adequate internal consistency,indicating strong convergent validity.

Table 2 summarizes the descriptive statis-tics and correlations among the constructs ofthe study. Apparently, some constructs such asinward activities and inward gain are conceptuallyrelated and empirically correlated in a substantialway. To deal with this issue, we followed theapproach suggested by Tanriverdi and Venka-traman (2005) and compared our measurementmodel (𝜒2 = 724.87, DF= 314, CMIN/DF= 2.31,CFI= 0.91, RMSEA= 0.060) with alternative mea-surement models, respectively: (a) modeling inwardactivities and inward gain as one construct insteadof separate constructs (𝜒2 = 832.22, DF = 318,



Table 2Summary Statistics and Correlation Matrixa

Variables Mean SD 1 2 3 4 5 6 7 8 9

1. Outward activitiesb 3.48 1.442. Inward gainb 4.62 0.93 0.533. Resource fungibilityc 5.45 0.93 0.51 0.414. Firm risk propensityd 0.50 0.50 0.07 0.03 0.065. Inward activitiesb 3.83 1.31 0.56 0.47 0.50 0.076. Firm age 2.85 0.59 0.04 −0.01 −0.02 0.01 −0.027. Firm size 7.15 1.25 0.08 0.06 −0.02 −0.01 −0.02 0.258. International

experience of TMT8.44 6.43 0.18 0.18 0.17 0.04 0.20 0.12 −0.01

9. R&D intensity 9.38 3.87 0.24 0.19 0.14 0.03 0.15 0.01 −0.07 −0.0210. Industry competition

intensityc5.53 0.93 0.22 0.20 0.21 0.04 0.15 −0.07 −0.01 0.02 −0.08

Note. N = 314; TMT = top management team.a Correlations with absolute values larger than 0.21 are significant at the level of p< .05 (two-tailed tests).b 7-point scale: 0= none; 6= a lot.c 7-point scale: 1= totally disagree; 7= totally agree.d Coded as a dummy for subgroup analysis.

CMIN/DF= 2.62, CFI= 0.80, RMSEA= 0.078);and (b) modeling resource fungibility as part ofthe inward gain construct (𝜒2 = 949.87, DF= 318,CMIN/DF= 2.99, CFI= 0.77, RMSEA= 0.086).The results show that our measurement modelhas a better model fit than the two alternativemodels, thus supporting the notion that inwardactivities, inward gain, and resource fungibility areindependent constructs.

Moreover, following X. Zhang and Bartol(2010), we conducted additional analyses toestablish the discriminant validity of these con-structs. We assessed their discriminant validity bycomparing the chi-square differences between aconstrained confirmatory factor model (in whichthe interfactor correlation is set to 1) and an uncon-strained model (in which the interfactor correlationis free). Specifically, we compared the fit of theunconstrained five-factor measurement model inwhich all covariances between latent variables wereunconstrained, with the fit of alternative models inwhich the number of factors was reduced by settingdifferent combinations of covariance betweenlatent variables equal to one. Our results showthat the five-factor model fits the data better thanall other models in which one or more covariancewas set equal to one (Δ𝜒2 ranges from 37.13 to188.30, p< .01 in all cases), suggesting strongdiscriminant validity among the constructs. Finally,we evaluated discriminant validity of the constructsby examining whether the AVE of each construct

was greater than the squared correlation betweentwo constructs (Fornell & Larcker, 1981). TheAVEs are all higher than the squared correlations,which indicate good discriminant validity at theconstruct level.

The use of survey data may raise the concernof common method variance bias because theindependent variables and the dependent variablewere from the same respondents. We conductedseveral analyses to address this potential issue.First, we constructed Harman’s one-factor test. Theresult revealed five factors with eigenvalues greaterthan 1. Second, we employed statistical methodsto address this concern. Podsakoff et al. (2003)recommended a confirmatory factor analysis oncompeting models to test the severity of the con-cern. That is, if the method variance is a significantproblem, a single-factor model should then fit thedata as nicely as a more complex model. We foundthat our theoretical model containing multifac-tors (𝜒2 = 724.87, DF= 314, CMIN/DF= 2.31,CFI= 0.91, RMSEA= 0.060) yielded asignificantly better fit of the data than the simplemodel (𝜒2 = 1788.26, DF= 324, CMIN/DF= 5.52,CFI= 0.68, RMSEA= 0.120). Third, we used thedata from the second round of the survey to validateour data from the first round and to conduct arobustness check, which yielded highly consistentfindings. These results will be reported in the“Robustness checks” section.



Table 3Path Coefficients and Model Fit Indexes of Core Modelsa,b

Model 1 Model 2 Model 3

PathsFirm age→Outward activities 0.01 0.02 0.02Firm size→Outward activities 0.14** 0.13** 0.13**International experience of TMT→Outward activities 0.02 −0.01 −0.01R&D intensity→Outward activities 0.15** 0.10* 0.06Industry competition intensity→Outward activities 0.11* 0.04 0.06Inward activities→Outward activities 0.66** 0.48** 0.65**Inward activities→ Inward gain 0.65** 0.67**Inward activities→Resource fungibility 0.09 0.05Path I(a): Inward gain→Resource fungibility 0.62** 0.67**Path I(b): Resource fungibility→Outward activities 0.32** 0.45**Path II: Inward gain→Outward activities −0.37**Model fit index𝜒2 413.31 981.10 971.01Degree of freedom 168 428 427CMNI/DF 2.46 2.29 2.27CFI 0.82 0.89 0.90RMSEA 0.068 0.064 0.063

Note. N = 314; TMT = top management team.a The coefficients are all standardized regression coefficients.b Significance levels: p< .01**; p< .05*; p< .1† (two-tailed tests.)

Structural Model

Table 3 reports the results of the structural mod-els. Model 1 is the basic model that includescontrol variables only. The data has a goodfit with the model (Browne & Cudeck, 1993)(𝜒2 = 413.31, DF= 168, CMIN/DF= 2.46,CFI= 0.82, RMSEA= 0.068). The path frominward activities to outward activities is statis-tically significant (b= 0.66, p< .01), suggestingthat a firm’s inward activities have a significantlypositive relationship with its outward activities.The path from firm size to outward activities is sta-tistically significant (b= 0.14, p< .01), suggestingthat larger firms are more likely to participate inoutward activities. In addition, R&D intensity hasa significantly positive relationship with outwardactivities (b= 0.15, p< .01). Furthermore, theindustry competition intensity has a significantand positive relationship with outward activities(b= 0.11, p< 0.05).

Compared with Model 1, Model 2 (𝜒2 = 981.10,DF= 428, CMIN/DF= 2.29, CFI= 0.89,RMSEA= 0.064) adds the path from inwardgain to outward activities through resource fungi-bility. Two other paths are also added. Specifically,the path from inward activities to inward gain isadded because it is natural to conjecture that a

firm’s inward gain should come from its participa-tion in inward activities. Also, the path from inwardactivities to resource fungibility is added in orderto ensure that the relationship between inwardgain and resource fungibility does not reflect thepossible relationship between inward activities andresource fungibility.

H1 proposes that a firm’s inward gain has a posi-tive relationship with its outward activities throughimproving its resource fungibility. As shown inModel 2, inward gain has a significantly posi-tive relationship with resource fungibility [PathI(a)] (b= 0.62, p< .01) and resource fungibilityhas a significantly positive relationship with out-ward activities [Path I(b)] (b= 0.32, p< .01). Theseresults support H1.

H2 proposes that the relationship as stated in H1is stronger when a firm’s organizational resourcefungibility is improved than when its technologicalresource fungibility is improved. This hypothesis istested by the results reported in Model 4 in Table 4.As shown in this model (𝜒2 = 899.87, DF= 394,CMIN/DF= 2.28, CFI= 0.90, RMSEA= 0.063),technological resource fungibility does not havea significant relationship with outward activities(b=−0.12, n.s.) whereas organizational resourcefungibility has a significantly positive relationshipwith outward activities (b= 0.54, p< .01). We



Table 4Differences between Technological and Organizational Resource Fungibilitya,b

Model 4

PathsFirm age→Outward activities 0.01Firm size→Outward activities 0.11**International experience of TMT→Outward activities 0.01R&D intensity→Outward activities 0.06†Industry competition intensity→Outward activities 0.08Inward activities→Outward activities 0.63**Inward activities→ Inward gain 0.67**Inward activities→Technological resource fungibility −0.04Inward activities→Organizational resource fungibility 0.11Path I(a):Path I(a.1): Inward gain →Technological resource fungibility 0.70**Path I(a.2): Inward gain →Organizational resource fungibility 0.59**Path I(b):Path I(b.1): Technological resource fungibility→Outward activities −0.12Path I(b.2): Organizational resource fungibility→Outward activities 0.54**Path IIInward gain→Outward activities −0.31*Model fit index𝜒2 899.87Degree of freedom 394CMNI/DF 2.28CFI 0.90RMSEA 0.063| Critical ratio of Path I(b.2) vs. Path I(b.1) | 2.18*

Note. N = 314; TMT = top management team.a The coefficients are all standardized regression coefficients.b Significance levels: p< .01**; p< .05*; p< .1† (two-tailed tests).

further compared these two coefficients by usingthe critical ratio of the parameters following Byrne(2010). The critical ratio with an absolute valuegreater than 1.96 for differences between thesetwo paths is considered to be significant at p< .05(Byrne, 2010). In our model, the absolute valueof the critical ratio of the coefficients of the twopaths is 2.18 (larger than 1.96), indicating that theeffect of organizational resource fungibility is sig-nificantly stronger than the effect of technologicalresource fungibility on a firm’s outward activities,thus supporting H2.

H3 states that after controlling for the indirectrelationship between inward gain and outwardactivities through resource fungibility, inward gainhas a direct negative relationship with outwardactivities. To test H3, Model 3 of Table 3 addsthe direct link between inward gain and out-ward activities (Path II) (𝜒2 = 971.01, DF= 427,CMIN/DF= 2.27, CFI= 0.90, RMSEA= 0.063).

This model shows that inward gain has a signifi-cantly negative relationship with outward activities(Path II) (b=−0.37, p< .01), thus supporting H3.

H4 proposes that the relationship as pro-posed in H3 is stronger for high-risk outwardactivities than for low-risk outward activi-ties. We tested this hypothesis by Model 5 inTable 5 (𝜒2 = 955.47, DF= 418, CMIN/DF= 2.29,CFI= 0.90, RMSEA= 0.063). As shown in thismodel, the direct relationship between inward gainand low-risk outward activities is negative andsignificant (b=−0.27, p< .05) whereas the directrelationship between inward gain and high-riskoutward activities is also negative and significant(b=−0.45, p< .01). The absolute value of thecritical ratio of the coefficients of these two pathsis 2.12 (larger than 1.96), indicating that inwardgain has a stronger negative direct relationshipwith high-risk outward activities than with low-riskoutward activities, thus supporting H4.



Table 5Differences between High-Risk and Low-Risk Outward Activitiesa,b

Model 5

DV1: Outward activitiesrefer to low-risk activities

DV2: Outward activities referto high-risk activities

PathsFirm age→Outward activities 0.01 0.03Firm size→Outward activities 0.13** 0.09*International experience of TMT→Outward activities −0.02 −0.01R&D intensity→Outward activities 0.06 0.06Industry competition intensity→Outward activities 0.08† 0.05Inward activities→Outward activities 0.58** 0.67**Inward activities→ Inward gainc 0.68** 0.68**Inward activities→Resource fungibilityc 0.05 0.05Path I(a): Inward gain→Resource fungibility c 0.67** 0.67**Path I(b): Resource fungibility→Outward activities 0.46** 0.43**Path II: Inward gain→Outward activities −0.27* −0.45**Model fit index𝜒2 955.47Degree of freedom 418CMNI/DF 2.29CFI 0.90RMSEA 0.063| Critical ratio of Path II for DV2 and DV1 | 2.12*

Note. N = 314; TMT = top management team.a The coefficients are all standardized regression coefficients.b Significance levels: p< .01**; p< .05*; p< .1† (two-tailed tests).c These paths only appear once in the model because the variable of “outward activities” is not involved in these paths. We report thecoefficients of these paths under both DV1 and DV2 to show the whole picture on how results are different/the same for low-risk outwardactivities (DV1) and high-risk outward activities (DV2).

H5 proposes that the relationship as stated inH3 is stronger for firms with low-risk propen-sity than for those with high-risk propensity. Totest H5, we followed Byrne (2010) and Denis(2010) and divided the sample into two groupsbased on the median value of firm risk propen-sity. We then compared the coefficients of thesame path (Path II) in the two groups by usingthe critical ratio of the parameters based on az-test because the critical ratio should have a stan-dard normal distribution if the coefficients areassumed to be equal in the population (Byrne,2010; Denis, 2010). Model 6 in Table 6 presentsthe results of the pairwise parameter compari-son (𝜒2 = 1527.94, DF = 861, CMIN/DF= 1.78,CFI= 0.92, RMSEA= 0.050). As shown in thismodel, the direct relationship between inward gainand outward activities (Path II) is significantly neg-ative for firms with low-risk propensity (b=−0.47,p< .01), but it is not statistically significant forfirms with high-risk propensity (b=−0.19, n.s.).Moreover, the absolute value of the critical ratio

of the two coefficients is 2.05 (larger than 1.96),thus supporting H5. Since firm risk propensity isa continuous variable, as a robustness check, wecreated an interaction term between inward gainand firm risk propensity and run a regression anal-ysis. As shown in Table 7, the main effect ofinward gain on outward activities is significantlynegative (b=−0.203, p< .05). The interaction termbetween inward gain and firm risk propensity is pos-itively and significantly related to outward activi-ties (b= 0.210, p< .01), suggesting that the nega-tive effect of inward gain on outward activities isstronger when firm risk propensity is low ratherthan high. These findings provide additional supportfor H5.

Robustness Checks

One may argue that in the Chinese context,state-owned enterprises (SOEs) and private firmsmay behave differently in that SOEs’ outward activ-ities tend to reflect government agendas. To address



Table 6The Moderating Role of Firm Risk Propensitya,b

Model 6

Firm risk propensity

Low High

PathsFirm age→Outward activities 0.04 0.04Firm size→Outward activities 0.09** 0.11*International experience of TMT→Outward activities −0.01 −0.01R&D intensity→Outward activities 0.04 0.05Industry competition intensity→Outward activities 0.07 0.07Inward activities→Outward activities 0.62** 0.69**Inward activities→ Inward gain 0.66** 0.69**Inward activities→Resource fungibility 0.15 −0.02Path I(a): Inward gain→Resource fungibility 0.60** 0.72**Path I(b): Resource fungibility→Outward activities 0.60** 0.30**Path II: Inward gain→Outward activities −0.47** −0.19Model fit index𝜒2 1527.94Degree of freedom 861CMNI/DF 1.78CFI 0.92RMSEA 0.050| Critical ratio of Path I(a) between two groups | 0.01| Critical ratio of Path I(b) between two groups | 1.65†| Critical ratio of Path II between two groups | 2.05*

Note. N = 314; TMT = top management team.a The coefficients are all standardized regression coefficients.b Significance levels: p< .01**; p< .05*; p< .1† (two-tailed tests).

this issue, in supplementary analyses, we dividedthe sample into two groups—SOEs (N = 154) andprivate firms (N = 160). The results, as reported inthe Appendix S1, show that the positive indirectrelationship (i.e., inward gain→ resource fungi-bility→ outward activities) holds for both groups.However, the negative direct relationship (i.e.,inward gain→ outward activities) only holds forprivate firms (b=−0.33, p< .05), not for SOEs(b=−0.30, n.s.). (However, the critical ratio of thetwo coefficients is not statistically significant.)

To check the robustness of our results, we usedthe data obtained in the second round of the sur-vey (4 weeks after the first-round survey, N = 146)to measure outward activities, but kept the mea-sures of other constructs unchanged. We then repli-cated the models in Tables 3–6 (results availablefrom the authors upon request). The results are con-sistent with those reported in Tables 3–5 (for H1,H2, H3, and H4); however, the sample size was toosmall to perform the subgroup analysis in Table 6(for H5). We acknowledge that 4 weeks may notbe long enough to establish causal relationships.

Nevertheless, these results help to build the robust-ness of our findings.

Discussion and Conclusion

Complex Linkages Between Firm InwardActivities and Outward Activities

In this study we drew upon organizational learningtheory and prospect theory to examine how a firm’sgain from inward activities may be related to itsoutward activities in an emerging market context.With data from a sample of Chinese manufacturingfirms, we found empirical evidence to support bothlines of logic.

In support of the organizational learning logic,we found that a firm’s inward gain is positivelyrelated to its resource fungibility, which is fur-ther positively related to its outward activities.This indirect positive relationship is more salientwhen the firm’s organizational resource fungibilityis improved than when its technological resource



Table 7Robustness Check: Results of Regression Analysis of theModerating Role of Firm Risk Propensity

Variables Model 1 Model 2

Constant −3.255** −3.441**(0.540) (0.538)

Firm age 0.005 0.005(0.004) (0.004)

Firm size 0.112* 0.077(0.055) (0.056)

International experienceof TMT

−0.001 0.001(0.010) (0.010)

R&D intensity 0.014 0.017(0.017) (0.017)

Industry competitionintensity

0.045 0.014(0.072) (0.072)

Firm risk propensity 0.180* 0.198**(0.076) (0.076)

Resource fungibility 0.422** 0.419**(0.092) (0.091)

Inward gain −0.203* −0.134(0.088) (0.090)

Inward gain* Firm riskpropensity

0.210**(0.073)

F-value 7.12** 7.41**R-square 0.16 0.18Adjusted R-square 0.14 0.16N 314 314

Note. TMT = top management team.Significance levels: p< .01**; p< .05*; p< .1† (two-tailed tests).

fungibility is improved. In support of the prospecttheory logic, we found that a firm’s inward gainhas a direct negative relationship with its outwardactivities, especially for high-risk outward activi-ties, and for firms with low-risk propensity. Whilenot hypothesized, our results also show a directpositive relationship between inward activities andoutward activities. These results provide a rela-tively comprehensive portrait of the complex link-ages between a firm’s inward and outward activities.

Overall, according to our findings, there are threepossible paths through which an emerging marketfirm’s inward activities may be associated with itsoutward activities: (a) a firm’s inward activities havea direct positive relationship with its outward activ-ities; (b) a firm’s inward gain has an indirect posi-tive relationship with its outward activities throughimproving its resource fungibility (particularly itsorganizational resource fungibility); and (c) a firm’sinward gain has a direct negative relationship withits propensity for outward activities, particularly forthose activities with high risk.

These three paths seem to be in line with theawareness-motivation-capability (AMC) model inthe strategy research (Chen, Su, & Tsai, 2007).According to the AMC model, for a firm to takea competitive action (e.g., participating in outwardactivities), it needs to be aware of the opportunityand possess the motivation and capabilities to takethe action (Chen et al., 2007). Consistent with the“capability” component of the AMC model, ourfindings show that a firm’s inward gain can enhanceits resource fungibility, which can further facilitatethe firm’s outward activities (Path I). Our findingthat a firm’s inward gain has a direct negativerelationship with its outward activities (Path II) isconsistent with the “motivation” component of theAMC model. In other words, as a firm can gain fromits inward activities, its motivation for participatingin outward activities would be reduced.

Interestingly, after controlling for the “capabil-ity” and “motivation” factors, we still find a directpositive link between a firm’s inward activities andoutward activities. It is possible that a firm’s inwardactivities, per se, may increase the firm’s aware-ness of the opportunities for outward activities andthus direct the firm’s attention to overseas mar-kets. According to the AMC model, awareness ofcompetitive opportunities is required for a firm totake action to explore international opportunities(Chen et al., 2007). Inward activities can enhancethe firm’s awareness of overseas opportunities byconnecting the firm with foreign companies andallowing the firm to understand the “nuts and bolts”of foreign market activities (Karlsen et al., 2003).For example, inward activities may involve tripsto foreign countries, discussions and negotiationswith foreign partners, visits of foreign partners’plants and facilities, and investigation of foreignpartners’ overseas markets, all of which can be posi-tively related to the focal firm’s awareness of poten-tial business opportunities in overseas markets. AsGu and Lu (2011, p. 280) noted, “a local firmcan become internationalized in its home market,through the inward investments of foreign firms,and its interactions with those foreign firms.”

Contributions

Our study makes contributions to the literatureon emerging market firms’ internationalization inseveral important ways. First, previous studiesof the inward–outward relationship have mainlydrawn upon the organizational learning argument to



examine the direct link between inward activitiesand outward activities (e.g., Korhonen et al., 1996;Luo & Tung, 2007). Our findings suggest that thereare three possible ways through which a firm’sinward activities may be related to its outward activ-ities. Our study thus portrays a far more complexpicture of the inward–outward connection than pre-vious studies have suggested and thus could signifi-cantly advance our understanding of this importantquestion. Our theoretical arguments and empiricalfindings associated with prospect theory are partic-ularly intriguing as they point out the potential neg-ative effect of a firm’s inward gain on its outwardactivities, whereas previous studies have almost allsuggested a positive inward–outward connection(Gu & Lu, 2011; Luo & Tung, 2007; Welch &Luostarinen, 1993).

Second, our study highlights the crucial roleof resource fungibility in the inward–outwardconnection. While prior research has emphasizedthe importance of resources and capabilities inthe inward–outward connection, few studies haveexplicitly examined the role of resource fungibilityin this connection. We fill this loophole by empir-ically demonstrating that a firm’s inward gain canfacilitate its outward activities through improvingits resource fungibility. As shown in Model 5 inTable 5, resource fungibility is equally importantfor both high-risk outward activities and low-riskoutward activities. Moreover, as shown in Model6 in Table 6, resource fungibility is important forboth firms with low-risk propensity and those withhigh-risk propensity (although its importance isgreater for firms with low-risk propensity thanfor those with high-risk propensity, which is notsurprising). These findings are important becausethey empirically validate the relevance of “resourcefungibility” to firms’ internationalization processas proposed in the literature (Autio, 2005; Oviatt& McDougall, 1994; Sapienza et al., 2006). AsSapienza et al. (2006, p. 926) specifically posited,“highly fungible firm resources can buffer costsand facilitate the development of capabilities topursue new market entry opportunities.”

More important, we made a distinction betweenorganizational and technological resource fungi-bility and found that it is organizational resourcefungibility rather than technological resourcefungibility that facilitates a firm’s outward activ-ities (Model 4 in Table 4). These findings areconsistent with Dunning et al.’s (2008) argumentthat emerging market firms lack firm-specific

ownership advantages (notably organizationaland management skills) to ensure success intheir outward FDI. We believe that our study isthe first empirical study that has systematicallyexamined the role of resource fungibility in firms’internationalization processes. Our findings areparticularly important for emerging market firms,as these firms tend to have fewer firm-specificresources or capabilities. Developing resourcefungibility—particularly organizational resourcefungibility—from inward activities thus becomescrucial to internationalization.

Third, while prior research has acknowledgedthat internationalization is risky, our study makesan extra effort in examining how risk may shapefirm internationalization. For the first time in theliterature, we incorporate the prospect theory topropose and support that a firm’s inward gain maydiscourage the firm from participating in outwardactivities by reducing its top managers’ motivationfor taking risk in outward activities. Also, sincenot all outward activities involve the same level ofrisk, we distinguish between outward activities withhigh risk and those with low risk and find that thediscouraging role of a firm’s inward gain is morelikely to hold for outward activities with high riskthan for those with low risk. We further distinguishfirms with high-risk propensity from those withlow-risk propensity and empirically show that thediscouraging role of a firm’s inward gain holdsonly for firms with low-risk propensity, not forthose with high-risk propensity. Indeed, our resultof the control variable, firm R&D intensity, is alsoconsistent with this line of logic. R&D and outwardactivities are both risk-taking activities (Filatotchev& Piesse, 2009) and it appears that firms favoringone also favor the other. Overall, our results showthat whereas a gain situation may reduce a firm’s topmanagers’ motivation for taking risks, this dependsupon the risk level of the potential outward activitiesas well as the firm’s own appetite for risk.

Our study also has important practical implica-tions by providing a comprehensive explanationof emerging market firms’ outward interna-tional activities. Thus, to answer the questionof whether emerging market firms with higherinward gain engage in greater or fewer outwardactivities, we would argue that it depends uponhow a firm uses its gain from inward activities.If the firm can improve its resource fungibility(particularly organizational resource fungibil-ity) from its inward gain, the indirect positive



relationship (inward gain→ resource fungibil-ity→ outward activities) would likely dominate thedirect negative relationship (inward gain→ outwardactivities). If the firm cannot improve its resourcefungiblity, the direct negative relationship wouldlikely dominate the indirect positive relationship.Moreover, the indirect positive relationship wouldlikely dominate the direct negative relationshipfor outward activities with low risk. The indirectpositive relationship would also likely dominatethe direct negative relationship for firms withhigh-risk propensity. It is critical to note thatfor emerging market firms, internationalizationis not just a path toward new markets; instead,it reflects how these firms exploit and explorewhat they have learned from their interactionswith foreign firms at home in new (foreign) mar-kets. It is thus important for managers to thinkmore strategically on developing their (organi-zational) resource fungibility from their inwardactivities.

Limitations and Future Research Directions

Like any investigation, our study has limitations thatshould be addressed in the future. First, our datawere collected from four manufacturing industriesin a single country. Moreover, as noted in the Meth-ods section, in order to ensure that our researchquestions were relevant to the firms that answeredthe questionnaire, we excluded firms that did nothave any international businesses at the time of thesurvey. As a result, our findings may not be gen-eralizable to firms that have not had internationalbusinesses. How such firms become interested ininternational businesses, of course, is an interestingquestion, yet it is beyond the scope of this study.

Second, we focused on the contingent roles of thetypes of outward activities and firms’ risk propen-sity in the negative relationship between inwardgain and outward activities. Other contingent fac-tors are also worth exploring. For example, wefound that the negative relationship between inwardgain and outward activities holds for private firmsbut not for SOEs. It is possible that governmentsupport may be an important contextual factor inthe way that high government support reduces topmanagers’ concerns on the potential loss of outwardactivities and thus weakens the negative relationshipbetween inward gain and outward activities.

Third, our arguments are based upon firms’general internationalization experience. It may be

interesting for future research to consider both thecountry origins of foreign partners in inward activ-ities and the target countries for outward activities.For example, would a firm’s gain from inwardactivities with foreign partners from Country A bemore relevant to the firm’s outward activities inCountry A than in Country B? Moreover, it is likelythat firms entering into less developed countriessuch as those in Southeast Asia, Africa, and SouthAmerican may face different levels and types of riskand resource requirements from those firms thatinvest in North America and Europe, and that thesefirms may also differ in their entry modes as well asrisk tolerance in internationalization. Also, the roleof TMT international experience could be furtherexplored. We measured the average internationalexperience of TMTs and did not find any significantrelationship between TMT international experienceand outward activities. Future research couldexamine the variance of international experienceamong TMT members. For example, a team of fivein which each member has 2 years of internationalexperience may be less effective than a team inwhich only the CEO has 10 years of experience(but the other four do not have any experience).

Fourth, while China is the largest emerging mar-ket, people may wonder if our conclusions wouldhold for firms in other emerging markets that maydemonstrate different characteristics. Given the het-erogeneity of emerging markets, further examina-tion of our hypotheses is encouraged in differentgeographic and country settings. Also, because ofthe cross-sectional nature of our research design, weare cautious in inferring the direction of causalityamong the key constructs even though we have useda second round of surveys to validate our findings.However, as Kenny (1979) argued, a careful studyof cross-sectional relationships before attempting tovalidate the findings via costly longitudinal stud-ies is a commonly accepted approach for build-ing causal relationships (c.f., Li & Atuahene-Gima,2002). We hope that this study will serve as a foun-dation for future research to adopt a longitudinalresearch design (with a lag of a year or longer) toverify our hypothesized relationships.

In conclusion, we have proposed and foundempirical support that emerging market firms’inward gains may have dual effects on theiroutward internationalization. Our arguments andresults provide a relatively complete picture onemerging market firms’ internationalization and



will encourage more future research effort on thisinteresting and important topic.

Acknowledgements

We would like to thank the Editor, Rich Bettis, andthe two anonymous referees for their constructivecomments and suggestions that have helped signifi-cantly improve the article. The article has benefitedsignificantly from Yan Anthea Zhang’s commentsand suggestions. Haiyang Li appreciates the supportfrom National Natural Science Foundation of China(Project Number: No. 71132007).

References

Anand, J., & Delios, A. (2002). Absolute and relativeresources as determinants of international acquisitions.Strategic Management Journal, 23(2), 119–134.

Anderson, J. C., & Gerbing, D. W. (1988). Structuralequation modeling in practice: A review and rec-ommended two-step approach. Psychological Bulletin,103(3), 411–423.

Armstrong, J. S., & Overton, T. S. (1977). Estimatingnon-response bias in mail survey. Journal of MarketingResearch, 14(3), 396–402.

Aulakh, P. S., Rotate, M., & Teegen, H. (2000). Exportstrategies and performance of firms from emergingeconomies: Evidence from Brazil, Chile, and Mexico.Academy of Management Journal, 43(3), 342–361.

Autio, E. (2005). Creative tension: The significance ofBen Oviatt’s and Patricia McDougall’s article ‘towarda theory of international new ventures’. Journal ofInternational Business Studies, 36(1), 9–19.

Autio, E., George, G., & Alexy, O. (2011).International entrepreneurship and capabilitydevelopment—qualitative evidence and futureresearch directions. Entrepreneurship Theory andPractice, 35(1), 11–37.

Bollen, K. A. (1989). Structural equations with latentvariables (). New York, NY: Wiley.

Browne, M. W., & Cudeck, R. (1993). Alternative ways ofassessing model fit. In K. A. Bollen & J. S. Long (Eds.),Testing structural equation models (pp. 136–162).Newbury Park, CA: Sage.

Buckley, P. J., Clegg, J., & Wang, C. (2002). The impactof inward FDI on the performance of Chinese manufac-turing firms. Journal of International Business Studies,33(4), 637–655.

Byrne, B. M. (2010). Structural equation modeling withAMOS: Basic concepts, applications, and program-ming (). New York, NY: Taylor and Francis Group.

Carpenter, M. A., Pollock, T. G., & Leary, M. M. (2003).Testing a model of reasoned risk-taking: Governance,the experience of principals and agents, and global strat-egy in high-technology IPO firms. Strategic Manage-ment Journal, 24(9), 803–820.

Carpenter, M. A., Sanders, W. G., & Gregersen, H. B.(2001). Bundling human capital with organizationalcontext: The impact of international assignment experi-ence on multinational firm performance and CEO pay.Academy of Management Journal, 44(3), 493–511.

Chang, S. J., & Xu, D. (2008). Spillovers and competitionamong foreign and local firms in China. StrategicManagement Journal, 29(5), 495–518.

Chen, M. J., Su, K.-H., & Tsai, W. (2007). Competitive ten-sion: The awareness-motivation-capability perspective.Academy of Management Journal, 50(1), 101–118.

Cyert, R. M., & March, J. G. (1963). A behavioral theoryof the firm (). New York, NY: Wiley-Blackwell.

Denis, J. D. (2010). Multi-group analysis in AMOS. Data& Decision Lab, Department of Psychology, Universityof Montana, Missoula. Retrieved from http://Psychweb.Psy.Umt.Edu/Denis/Datadecision/Multigroup/amos_group.Html.

Dunning, J. H. (1993). Multinational enterprises andthe global economy (). Wokingham, England:Addison-Wesley.

Dunning, J. H., Kim, C., & Park, D. (2008). Old wine innew bottles: A comparison of emerging-market TNCstoday and developed-country TNCs thirty years ago. InK. Sauvant (Ed.), The rise of transnational cooperationfrom emerging markets: Threat or opportunity (pp.158–180). Cheltenham, England: Edward Elgar.

Fiegenbaum, A., & Thomas, H. (1988). Attitudes towardrisk and the risk–return paradox: Prospect theoryexplanations. Academy of Management Journal, 31(1),85–106.

Filatotchev, I., & Piesse, J. (2009). R&D, internationaliza-tion and growth of newly listed firms: European evi-dence. Journal of International Business Studies, 40(8),1260–1276.

Ford, J. K., MacCallum, R. C., & Tait, M. (1986). Theapplication of exploratory factor analysis in appliedpsychology: A critical review and analysis. PersonnelPsychology, 39(2), 291–314.

Fornell, C., & Larcker, D. F. (1981). Evaluating struc-tural equation models with unobservable variables andmeasurement error. Journal of Marketing Research, 18,39–50.

Ghemawat, P. (2001). Distance still matters. HarvardBusiness Review, 79(8), 137–147.

Gu, Q., & Lu, J. W. (2011). Effects of inward investmenton outward investment: The venture capital industryworldwide 1985–2007. Journal of International Busi-ness Studies, 42(2), 263–284.

Gubbi, S. R., Aulakh, P. S., Ray, S., Sarkar, M., &Chittoor, R. (2009). Do international acquisitions byemerging-economy firms create shareholder value? Thecase of Indian firms. Journal of International BusinessStudies, 41(3), 397–418.

Guthrie, D. (2005). Organizational learning and productiv-ity: State structure and foreign investment in the rise ofthe Chinese corporation. Management and Organiza-tion Review, 1(2), 165–195.

Helfat, C. E., & Peteraf, M. A. (2003). The dynamicresource-based view: Capability lifecycles. StrategicManagement Journal, 24(10), 997–1010.


http://psychweb.psy.umt.edu/denis/datadecision/multigroup/amos_group.html



http://books.google.com/books?hl=en&lr=&id=mhcWghgCx2kC&oi=fnd&pg=PA158&dq=outward+fdi+and+emerging+market+firms+and+dunning&ots=On-&uscore;hRVuPf&sig=xtRSJomlSXNezPe1BNS0c1qhWW4


Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). Inter-national diversification: Effects on innovation and firmperformance in product-diversified firms. Academy ofManagement Journal, 40(4), 767–798.

Hitt, M. A., Li, H., & Worthington, W. J. (2005). Emergingmarkets as learning laboratories: Learning behaviors oflocal firms and foreign entrants in different institutionalcontexts. Management and Organization Review, 1(3),353–380.

Huber, G. P. (1991). Organizational learning: The con-tributing processes and the literatures. OrganizationScience, 2(1), 88–115.

James, L. R., Mulaik, S. A., & Brett, J. M. (1982). Causalanalysis: Assumptions, models, and data (). BeverlyHills, CA: Sage.

Johanson, J., & Vahlne, J. E. (1977). The internation-alization process of the firm-a model of knowledgedevelopment and increasing foreign market com-mitments. Journal of International Business Studies,8(1), 23–32.

Joreskog, K. G., Sorbom, D., du Toit, S., & du Toit, M.(1999). LISREL 8: User’s reference guide (). Chicago,IL: Scientific Software International.

Kahneman, D., & Tversky, A. (1979). Prospect theory: Ananalysis of decision under risk. Econometrica: Journalof the Econometric Society, 47(2), 263–291.

Karlsen, T., Silseth, P. R., Benito, G. R., & Welch, L.S. (2003). Knowledge, internationalization of the firm,and inward–outward connections. Industrial Market-ing Management, 32(5), 385–396.

Kenny, D. A. (1979). Correlation and causality (). NewYork, NY: Wiley.

Korhonen, H., Luostarinen, R., & Welch, L. (1996). Inter-nationalization of SMEs: Inward-outward patterns andgovernment policy. Management International Review,36(4), 315–329.

Lehner, J. M. (2000). Shifts of reference points for framingof strategic decisions and changing risk-return associa-tions. Management Science, 46(1), 6–76.

Levy, J. S. (1996). Loss aversion, framing, and bargaining:The implications of prospect theory for internationalconflict. International Political Science Review, 17(2),179–195.

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

Li, J., Li, Y., & Shapiro, D. (2012). Knowledge seeking andoutward FDI of emerging market firms: The moderatingeffect of inward FDI. Global Strategy Journal, 2(4),277–295.

Luo, Y., & Tung, R. L. (2007). International expansion ofemerging market enterprises: A springboard perspec-tive. Journal of International Business Studies, 38(4),481–498.

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

Miller, D. (1983). The correlates of entrepreneurshipin three types of firms. Management Science, 29(7),770–791.

Nunnally, J. (1978). Psychometric theory (). New York,NY: McGraw-Hill.

Oviatt, B. M., & McDougall, P. P. (1994). Toward a theoryof international new ventures. Journal of InternationalBusiness Studies, 25(1), 45–64.

Penrose, E. T. (1959). The theory of the growth of the firm(). New York, NY: John Wiley.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Pod-sakoff, N. P. (2003). Common method biases in behav-ioral research: A critical review of the literature andrecommended remedies. Journal of Applied Psychol-ogy, 88(5), 879–903.

Ramachandran, J., & Pant, A. (2010). The liabilities of ori-gin: An emerging economy perspective on the costs ofdoing business abroad. In T. M. Devinney, T. Pedersen,& L. Tihanyi (Eds.), Advances in international man-agement: The past, present and future of internationalbusiness and management (pp. 231–265). New York,NY: Emerald.

Rui, H., & Yip, G. S. (2008). Foreign acquisitions byChinese firms: A strategic intent perspective. Journalof World Business, 43(2), 213–226.

Sambharya, R. B. (1996). Foreign experience of top man-agement teams and international diversification strate-gies of U.S. multinational corporations. Strategic Man-agement Journal, 17(9), 739–746.

Sapienza, H. J., Autio, E., George, G., & Zahra, S. A.(2006). A capabilities perspective on the effects ofearly internationalization on firm survival and growth.Academy of Management Review, 31(4), 914–933.

Shimizu, K. (2007). Prospect theory, behavioral theory,and the threat-rigidity thesis: Combinative effects onorganizational decisions to divest formerly acquiredunits. Academy of Management Journal, 50(6),1495–1514.

Shook, C. L., Ketchen, D. J., Hult, G. T. M., & Kacmar,K. M. (2004). An assessment of the use of structuralequation modeling in strategic management research.Strategic Management Journal, 25(4), 397–404.

Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualizing thedeterminants of risk behavior. Academy of ManagementReview, 17(1), 9–38.

Steensma, H. K., & Lyles, M. A. (2000). ExplainingIJV survival in a transitional economy through socialexchange and knowledge-based perspectives. StrategicManagement Journal, 21(8), 831–851.

Steensma, H. K., Tihanyi, L., Lyles, M. A., & Dhanaraj,C. (2005). The evolving value of foreign partnershipsin transitioning economies. Academy of ManagementJournal, 48(2), 213–235.

Tanriverdi, H., & Venkatraman, N. (2005). Knowledgerelatedness and the performance of multibusiness firms.Strategic Management Journal, 26(2), 97–119.

Tian, X. (2007). Accounting for sources of FDI technologyspillovers: Evidence from China. Journal of Interna-tional Business Studies, 38, 147–159.

Tversky, A., & Kahneman, D. (1991). Loss aversion inriskless choice: A reference-dependent model. Quar-terly Journal of Economics, 106(4), 1039–1061.

Vassolo, R. S., Anand, J., & Folta, T. B. (2004).Non-additivity in portfolios of exploration activities:A real options-based analysis of equity alliances in



biotechnology. Strategic Management Journal, 25(11),1045–1061.

Welch, L. S., & Luostarinen, R. K. (1993). Inward-outwardconnections in internationalization. Journal of Interna-tional Marketing, 1(1), 44–56.

Widaman, K. F. (1985). Hierarchically nested covari-ance structure models for multitrait-multimethod data.Applied Psychological Measurement, 9(1), 1–26.

Wiersema, M. F., & Bantel, K. A. (1992). Top manage-ment team demography and corporate strategic change.Academy of Management Journal, 35(1), 91–121.

Young, S., Huang, C., & McDermott, M. (1996). Inter-nationalization and competitive catch-up processes:Case student evidence on Chinese multinationalenterprises. Management International Review, 36(4),295–314.

Zaheer, S. (1995). Overcoming the liability of foreignness.Academy of Management Journal, 38(2), 341–363.

Zhang, X., & Bartol, K. M. (2010). Linking empoweringleadership and employee creativity: The influence ofpsychological empowerment, intrinsic motivation, andcreative process engagement. Academy of ManagementJournal, 53(1), 107–128.

Zhang, Y., & Li, H. (2010). Innovation search of newventures in a technology cluster: The role of ties withservice intermediaries. Strategic Management Journal,31(1), 88–109.

Zhang, Y., Li, Y., & Li, H. (2014). FDI spillovers over timein an emerging market: The role of entry tenure andbarriers to imitation. Academy of Management Journal,57(3), 698–722.

Zhang, Y., Li, H., Li, Y., & Zhou, L. (2010). FDI spilloversin an emerging market: The role of foreign firms’country origin diversity and domestic firms’ absorp-tive capacity. Strategic Management Journal, 31(9),969–989.

Supporting Information

Additional supporting information may be foundin the online version of this article:

Appendix S1. Path coefficents for SOEs vs. privatefirms.a,b


emerging market firms' …haiyang/smj_2679 early view.pdf · 2721 examine the direct link...

Documents